Drug Hypersensitivity

Drug Hypersensitivity

Drug Hyper sensitivity

-

Editor

W.J. Pichler Bern 78 figures, 58 in color, 78 tables, 2007

Basel • Freiburg • Paris • London • New York • Bangalore • Bangkok • Singapore • Tokyo • Sydney

Prof. Dr. Werner J. Pichler Department of Rheumatology and Clinical Immunology/Allergology Inselspital, University of Bern CH–3010 Bern (Switzerland)

Library of Congress Cataloging-in-Publication Data Drug hypersensitivity / editor, W.J. Pichler. p. ; cm. Includes bibliographical references and index. ISBN 978–3–8055–8269–8 (hard cover : alk. paper) 1. Drug allergy. I. Pichler, Werner J., 1949– [DNLM: 1. Drug Hypersensitivity. WD 320 D794 2007] RC598.D7D78 2007 616.97‘58–dc22

2007007180

Bibliographic Indices. This publication is listed in bibliographic services, including Current Contents® and Index Medicus. Disclaimer. The statements, options and data contained in this publication are solely those of the individual authors and contributors and not of the publisher and the editor(s). The appearance of advertisements in the book is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements. Drug Dosage. The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any change in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. All rights reserved. No part of this publication may be translated into other languages, reproduced or utilized in any form or by any means electronic or mechanical, including photocopying, recording, microcopying, or by any information storage and retrieval system, without permission in writing from the publisher. © Copyright 2007 by S. Karger AG, P.O. Box, CH–4009 Basel (Switzerland) www.karger.com Printed in Switzerland on acid-free paper by Reinhardt Druck, Basel ISBN 978–3–8055–8269–8

tion of extremely high HLA-B associations with certain drugs and diseases, which already results in effective prevention of some forms of drug hypersensitivity reactions. The people interested in drug hypersensitivity come from many different areas. It is a truly interdisciplinary group which recently gained its own identity by holding two drug hypersensitivity meetings (in Bern 2004 and Liverpool 2006) where participants from different areas exchanged ideas. Most of the contributors to this book were present at these meetings and participated in the friendly but also hot discussions, not least because the group from Bern proposed a new concept on how T cells may be stimulated by drugs (pharmacological interaction with immune receptors, the p-i concept), which contradicted prevailing concepts quite radically. Future studies and results will show which of these ideas, which are also represented in the various chapters of this book, will prevail. If the p-i concept is confirmed (which I assume), we will have to learn what it means for the clinical symptoms, diagnosis and prediction of drug hypersensitivity reactions. But the impact of this new concept will go far beyond drug hypersensitivity alone, as it combines pharmacology with antigen-specific immu-

Foreword

nology, and will therefore also influence other areas of research and medicine. In this case the idea of also seeing in a drug hypersensitivity reaction an unintended ‘experiment’ of the treating physician would be confirmed, and the only positive side of drug hypersensitivity would become apparent, namely that medicine and science could learn from drug hypersensitivity. We need to approach these iatrogenic and thus often embarrassing diseases with an open mind, not frightened by the complexity of the issue, and not using its unpredictability as a permanent excuse to do nothing: It is not so much the disease which is embarrassing, but the handling of these unexpected side effects. Drug hypersensitivity reactions should and will become less bizarre and less frightening when physicians, scientists and the pharmaceutical industry work together to better understand these diseases. I would like to thank the many contributors to this book, they did an excellent job. I am also very thankful to Karger Publishers, in particular Thomas Nold, and my secretary Franziska Mitton who helped me to organize this book. Werner J. Pichler Bern, November 2006

IX

Contents

VIII Foreword

Epidemiology of Drug Allergies 2 Epidemiology and Causes of Drug Hypersensitivity Demoly, P. (Montpellier); Viola, M. (Rome); Rebelo Gomes, E. (Porto); Romano, A. (Rome/Troina) 18 Epidemiology and Causes of Severe Cutaneous Adverse Reactions to Drugs Mockenhaupt, M. (Freiburg)

Pathomechanisms, Genetics and Animal Models 34 Chemical-Induced Contact Hypersensitivity in the Mouse Model Martin, S.F. (Freiburg) 47 Lessons from Nickel Hypersensitivity: Structural Findings Weltzien, H.U. (Freiburg) 55 Drugs as Haptens, Antigens, and Immunogens Park, B.K.; Sanderson, J.P.; Naisbitt, D.J. (Liverpool) 66 The p-i Concept: Evidence and Implications Gerber, B.O.; Pichler, W.J. (Bern) 74 Tolerance Mechanisms to Small Molecular Compounds Cavani, A.; De Pità, O. (Rome) 84 HIV and Drug Hypersensitivity Pirmohamed, M. (Liverpool) 95 Abacavir Hypersensitivity Nolan, D.; Almeida, C.-A.; Phillips, E.; Mallal, S. (Perth) 105 Genetics of Severe Drug Hypersensitivity Reactions in Han Chinese Hung, S.-I.; Chung, W.-H.; Chen, Y.-T. (Taipei) 115 Nevirapine Hypersensitivity Shenton, J.M. (Toronto, Ont./Syracuse, N.Y.); Popovic, M. (Toronto, Ont./Basel); Uetrecht, J.P. (Toronto, Ont.)

V

129 Animal Models of Toxic Epidermal Necrolysis Azukizawa, H.; Itami, S. (Osaka) 140 Non-Clinical Testing Approaches for Drug Development: Possibilities and Limitations Kawabata, T.T. (Groton, Conn.); Piccotti, J.R. (Ann Arbor, Mich.) 151 Adverse Side Effects to Biological Agents Pichler, W.J. (Bern); Campi, P. (Florence)

Clinical Manifestations 168 Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation Pichler, W.J. (Bern) 190 Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins) Torres, M.J.; Mayorga, C.; Blanca, M. (Malaga) 204 Perioperative Anaphylaxis Dewachter, P. (New York, N.Y.) 216 IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics Campi, P.; Manfredi, M.; Severino, M. (Florence) 233 Hypersensitivity Reactions to Iodinated Contrast Media: An Update Christiansen, C. (Oslo) 242 Maculopapular Drug Eruptions Yawalkar, N. (Bern) 251 Drug-Induced Hypersensitivity Syndrome and Viral Reactivation Shiohara, T.; Takahashi, R.; Kano, Y. (Tokyo) 267 Clinic and Pathogenesis of Severe Bullous Skin Reactions: Stevens-Johnson Syndrome, Toxic Epidermal Necrolysis Allanore, L.; Roujeau, J.-C. (Créteil) 278 Drug Allergic Liver Injury Cerny, A.; Bertoli, R. (Lugano) 295 Drug-Induced Interstitial Nephritis Keller, M.; Spanou, Z.; Pichler, W.J. (Bern) 306 Blood Dyscrasias Caused by Hypersensitivity to Drugs Aster, R.H. (Milwaukee, Wisc.) 321 Allergy and Pseudoallergy to Drugs and Vaccines in Children Ponvert, C.; Hadj-Rabia, S.; Scheinmann, P. (Paris) 340 Hypersensitivity to Aspirin and Other NSAIDs: Mechanisms, Clinical Presentation and Management Szczeklik, A.; Niżankowska-Mogilnicka, E.; Sanak, M. (Kraków)

VI

Contents

Diagnosis 352 Approach to the Patient with a Drug Hypersensitivity Reaction – Clinical Perspectives Bircher, A.J. (Basel) 366 Place of Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions Barbaud, A. (Nancy) 380 In vitro Tests of T-Cell-Mediated Drug Hypersensitivity Beeler, A.; Pichler, W.J. (Bern) 391 In vitro Tests: Basophil Activation Tests Sanz, M.L. (Pamplona); Gamboa, P.M. (Bilbao); De Weck, A.L. (Pamplona)

Desensitization 404 Desensitization with Antibiotics Solensky, R. (Corvallis, Oreg.) 413 Drug Desensitization in Oncology: Chemotherapy Agents and Monoclonal Antibodies Castells, M. (Boston, Mass.) 426 Author Index 427 Subject Index

Contents

VII

Foreword

Drug hypersensitivity is a complex and still widely neglected topic. It is considered to be a difficult area of medicine, as many different drugs can cause hypersensitivity reactions, albeit each drug rather rarely. It can result in many different types of diseases, with unknown but presumably quite distinct pathomechanisms. Although almost every doctor has encountered some drug-allergic diseases, not many doctors have large experience with a certain drug and a certain type of drug hypersensitivity. As drug hypersensitivity characteristically appears unexpectedly, patient-oriented research is difficult. Because of its unpredictable nature it was distinguished from pharmacological reactions (type A) and termed a type-B reaction, whereby B was soon interpreted as meaning bizarre! These features did not contribute to making drug hypersensitivity a popular area of medicine or science, the chances of finding a breakthrough were simply considered to be too small. Until now, most books on this topic have focused on epidemiological aspects and listed the drugs and their side effects, but the immune pathogenesis, its relation to the clinical features of drug hypersensitivity diseases, immunogenet-

ic aspects, animal models, new diagnostic possibilities, treatments and desensitizations have not been discussed in detail. This book tries to change this approach and tackles these areas. It covers epidemiology, pathogenesis, clinics and diagnosis of these diseases. It reaches from animal models to clinical chapters on how to approach a patient with drug allergy and how to desensitize him. The aim is to provide a general view on this topic – so that the practicing physician, the allergologist, the pharmacologist, the epidemiologist, the geneticist, the immunotoxicologist, the safety officer of a pharmaceutical company, and the scientist interested in interactions of small molecules with the immune system, etc., will find relevant information on this now rapidly evolving field of medicine. This approach has become possible because during the last 10–15 years substantial development in this area has taken place. The important role of drug-specific T cells in these reactions has been deciphered; new animal models for certain forms of drug hypersensitivity diseases have been established, and new test systems have been developed to better define the incriminated drug. An important breakthrough was the identifica-

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 2–17

Epidemiology and Causes of Drug Hypersensitivity Pascal Demoly a
Marinella Viola b
Eva Rebelo Gomes d
Antonino Romano b, c a Exploration

des allergies et INSERM, Hôpital Arnaud de Villeneuve, University Hospital of Montpellier, Montpellier, France; b Department of Internal Medicine and Geriatrics, UCSC-Allergy Unit, Complesso Integrato Columbus, Rome, and c IRCCS Oasi Maria S.S., Troina, Italy; d Allergy Department, Hospital Maria Pia, Porto, Portugal

Abstract Drug hypersensitivity reactions (DHRs) are the adverse effects of drugs, taken at a dose which is tolerated by normal subjects, which clinically resemble allergy. There are few true epidemiological data on DHRs. The available information requires a cautious interpretation because the pathogenic mechanism has not been demonstrated by diagnostic tests. Both under- and over-diagnosis must be taken into account. DHRs may represent up to one third of adverse drug reactions, be life-threatening, require or prolong hospitalization, and entail changes in drug prescription. They concern more than 7% of the general population, and therefore are an important public health problem. A few risk factors have been pinpointed. Future progress in genetics, as well as well-designed epidemiological studies on hypersensitivity drug reactions, will be helpful in identifying patients at risk of developing such reactions, in particular severe ones, and in implementing early preventive measures. This review describes current data on the incidence, prevalence, mortality, and risk factors of these reactions. Copyright © 2007 S. Karger AG, Basel

Introduction

Drug hypersensitivity reactions (DHRs) represent adverse effects of drugs, taken at a dose which is tolerated by normal subjects, which clinically resemble allergy. Numerous reactions with

symptoms suggestive of an allergy are often erroneously considered to be real drug allergies. In any case, although DHRs are a frequent, almost constant worry for the prescribing physicians, there is no extensive epidemiological study. DHRs belong to type B adverse drug reactions (ADRs) [1–7]. The classical pharmacological classification of ADRs by Rawlins and Thompson [7] divides these into two major subtypes: type A reactions, which are dose-dependent and predictable, and type B, which are neither. The majority of ADRs are type A reactions. Type B reactions constitute approximately 10–15% of all ADRs (up to one third for some authors) and include DHRs. This classification was further extended to include other subtypes [8–11]. Although more accurate, the latter is too complex to use in everyday clinical practice. Drug allergic reactions, according to the Nomenclature Review Committee of the World Allergy Organization, refer to DHRs where a definite immunological mechanism, either IgE- or T-cell-mediated, is demonstrated [12]. ADRs that clinically resemble an allergy, but where an immunological process is not proven should be classified as non-immune DHRs [12]. This is impor-

tant because most of the available epidemiological studies to date refer to ADRs in general terms rather than drug allergy. In addition, drug allergy studies rely on clinical histories of a temporal relationship between administration of the suspect drug and symptoms/signs without in vivo or in vitro tests demonstrating drug-specific IgE- or T-cell-mediated mechanisms. This is due to the lack of standardized tests for many of these drugs and the limited use of drug provocation tests.

Prevalence and Incidence of Drug Hypersensitivity Reactions

DHRs are responsible for significant morbidity, mortality and socioeconomic costs that have yet to be fully calculated. Current epidemiological data have to be carefully evaluated as different studies use different populations (either adult or pediatric populations or both, inpatients or outpatients), different definitions of ADRs/drug allergy, and different methodologies and methods of data analysis. It should also be kept in mind that causality assessment (or drug imputability) relies mostly on clinical histories that are not accurate enough for the firm diagnosis of DHRs and cannot replace drug allergy testing [13]. Data on Hospital-Based Populations The Boston Collaborative Drug Surveillance Program collected information on all ADRs in 4,031 hospitalized patients during a 6-month period: 247 reactions were declared (an incidence of 6.1%), of which 41.7% were severe and 1.2% led to the patient’s death [14]. The majority (61.7%) of these reactions were unpredictable and therefore possibly allergic reactions, although this is not stated in the paper. The use of an automatic detection system in the LDS Hospital in Salt Lake City allowed the identification of 731 reactions among 36,653 hospitalized patients (only 12.3% were reported by the doctors in the hospital) during an 18-month period [15]. The incidence

Epidemiology and Causes of Drug Hypersensitivity

(1.8%) was lower than in the previous study; 13.8% were severe and 32.7% of an allergic nature. However, if the criteria for inclusion had been similar to those used in the Boston study, the incidence (2.8%) would still have been quite similar. Lazarou et al. [16] showed in a metaanalysis of 39 prospective USA studies from 1966 to 1996 that 15.1% of hospitalized patients suffered an ADR (6.7% severe) and that the incidence of drug-related hospital admissions ranged from 3.1 to 6.2%. Although Kvasz et al. [17] raised some questions about the methodology and the validity of the meta-analysis, many subsequent studies have reported similar data. A study by Fattinger et al. [18] analyzed 4,331 hospitalizations in two Swiss departments of internal medicine and found that clinically relevant ADRs occurred in 11% of the patients and that ADRs were the cause of admission in 3.3%. In the prospective French pharmacovigilance study by Olivier et al. [19], data from 671 patients admitted to an emergency department during a 4-week period led to the identification of 44 ADRs involving 41 patients, an incidence of 6.1% (table 1). In Singapore, a 2-year prospective study by Thong et al. [20] using a network-based electronic notification system where each case was verified by a trained allergist, detected 366 cases of reported drug allergy in a total of 90,910 inpatients. After review, 210 were classified as drug allergy. Cutaneous eruptions were the most common clinical manifestations (95.7%), systemic symptoms occurred in 30% of the cases, and serious adverse reactions such as Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN) and general exfoliative dermatitis occurred in 11 patients (5.2%). Antibiotics and anti-epileptic drugs accounted for 75% of the reactions. Thong et al. concluded that the incidence of drug allergy in hospitalized patients was 0.42%. In a Swiss paper by Hardmeier et al. [21], 481 (7.5%) ADRs occurred among 6,383 inpatients (4 allergic reactions to antibiotics in patients with known allergy to the same drug) and there were 2.9% ADR-re-

3

Table 1. Epidemiological data on hospital-based populations Ref.

Patients

ADRs

Incidence of ADRs %

Period

Type of ADRs Culprit drugs

Remarks

14

4,031

247

6.1

6 months

B (61.7%)

41.7% severe ADRs 1.2% fatalities

15

36,653

731

1.8

18 months

A (664 ADRs) analgesics, anti-infectious B and cardiovascular agents

13.8% severe ADRs 32.7% allergic ADRs

16

62,480

15.1 (ADRs) 6.7 (severe ADRs) 0.32 (fatalities)

1966–1996

A (76.2%) B

18

3,624

4,331

0.14 (fatalities)

1996–1998

mainly A

19

671

44 (in 41 6.1 patients)

4 weeks

A (33%)

20

90,910

366

0.42

1997–1999

210 cases of drug allergy

antibiotics antiepileptics

21

6,383

481

7.5

1996–2000

A (0.4%) B (7.2%)

antithrombotics, cardiovascular drugs, antibiotics, hypnotics and NSAIDs

22

171

7

4

20 days

A B (2 allergic reactions)

antibiotics

23

18,820

1,225

6.5

6 months

A (95%)

NSAIDs diuretics

24

13,294

48

0.36

2000–2001

B

antibiotics (penicillins)

all cutaneous ADRs 34% severe ADRs

25

8,437 children

222

2.6

2002

B (98%)

antibiotics NSAIDs

98% cutaneous ADRs

26

789

789

1/13,000 administrations

1999–2000

B

NMBAs, latex antibiotics

all anaphylactic reactions during anesthesia in France 63.6% IgE-mediated

27

83

83

1996–2001

B

NMBAs latex

all anaphylaxis during anesthesia in Norway, 71.1% IgE-mediated

28

68 children

68

1989–2001

B

NMBAs, latex, colloids, opiates, hypnotics

cross-reactivity amongst NMBAs in 23 of 30 children (76%)

29

337,647 injections

A B

RCM

severe ADRs: 0.22% (ionic RCM) and 0.04% (non-ionic RCM), 2 fatalities

31

11,898

132

1.1

1998–2004

mainly A

fluorescein

32

489,494

110

1–10/10,000 person-years

1998–2001

B

antiepileptics

1/2,100 surgeries 12.7 (ionic RCM) 3.1 (non-ionic RCM)

cancer chemotherapeutics, iloprost, cyclosporin A

48% possible ADRs 41% possible ADRs disease-unrelated 11% clinically relevant ADRs were the cause of hospitalization in 3.3% 71% possible ADRs 18% plausible ADRs 11% likely ADRs 95.7% (cutaneous symptoms) 30% (systemic symptoms) 5.2% (severe ADRs)

all ADRs were SJS and TEN, >90% of cases occurred within the first 63 days of therapy

ADRs = Adverse drug reactions; NMBAs = neuromuscular blocking agents; NSAIDs = non-steroidal anti-inflammatory drugs; RCM = radio contrast media; SJS/TEN = Stevens-Johnson syndrome/toxic epidermal necrolysis.

4

Demoly ⴢ Viola ⴢ Rebelo Gomes ⴢ Romano

lated hospital admissions. In a 20-day observational prospective study from an Italian university hospital, among 171 inpatients undergoing antibiotic treatment, 7 (4%) patients experienced an ADR, of which 2 (28%) (angioedema from piperacillin; skin rash from ceftriaxone) may have been allergic reactions [22]. Pirmohamed et al. [23] conducted a prospective study in two National Health Service hospitals in Merseyside, UK, comprising 18,820 patients aged 116 years admitted over a 6-month period, which found 1,225 (6.5%) admissions related to an ADR. Most ADRs (95%) were classified as type A reactions, and thus not related to drug hypersensitivity. However, non-steroidal anti-inflammatory drugs (NSAIDs) were again the most commonly implicated drugs (causing hemorrhagic complications, wheezing and dermatological reactions), followed by diuretics. In a 6-month prospective survey of cutaneous drug reactions in hospitalized patients from a French general hospital, a total of 48 cases assessed by dermatologists as cutaneous allergic reactions from drugs resulted in an incidence of 3.6 per 1,000 patients. Of these, 34% were considered severe and antibiotics, mainly penicillins, were the most commonly implicated drug [24]. A recent retrospective case control study from Singapore by Kidon and See [25] using the hospital inpatient electronic medical record found 222 (2.6%) patients reporting a previous ADR among 8,437 hospitalized children. Almost 70% of them involved the use of antibiotics (-lactams in 45%) and NSAIDs (18.5%); 98% of the reactions were cutaneous and could have been allergic in nature. When considering more specific clinical settings it is possible to find more accurate data. The French registry of anaphylaxis during general anesthesia identified 789 cases during a 2-year survey, giving an incidence of 1/13,000 administrations of general anesthesia. 518 (66%) were IgE-mediated, with neuromuscular blocking agents (NMBAs) (58.2%) (incidence of 1/6,500 NMBA injections) and antibiotics (15.1%) being

Epidemiology and Causes of Drug Hypersensitivity

the most common etiologic agents [26]. The latter are now a rising cause of anaphylaxis during general anesthesia. Similarly, a high prevalence (71.1%) of IgE-mediated reactions has been reported in a 6-year survey performed by a single allergy center in Norway. Among 83 cases of anaphylaxis related to general anesthesia, 93.2% were caused by NMBAs [27]. Suxamethonium was the most frequently involved substance, followed by rocuronium and vecuronium. The few reactions in which other allergies could be detected were mainly linked to latex (3.6%). Data in children are scarce. A 12-year survey at a French pediatric center reported 68 cases of children who suffered anaphylaxis during general anesthesia [28]. Through allergologic diagnostic procedures (skin tests and specific IgE assays), an IgE-mediated mechanism was demonstrated in 51 patients: 31 (60.8%) reacted to NMBAs, 14 (27%) to latex, 7 (14%) to colloids, 5 (9%) to opiates, and 6 (12%) to hypnotics. The NMBA vecuronium caused the largest number of reactions. Cross-reactivity among the NMBAs available in France was observed in 23 of 30 children (76%), particularly to vecuronium, atracurium and pancuronium. The estimated frequency of IgE-mediated anaphylactic reactions was 1 in 2,100 operations. In the study by Katayama et al. [29], of 337,647 injections of radio contrast media (RCM) in Japan, the incidence of adverse reactions was 12.7% (0.22% of severe reactions for ionic products and 0.04% for non-ionic products); two reactions were fatal (incidence of 0.0006%). The data concerning RCM have been reviewed recently [30]. It would appear that mild immediate reactions occur in 3.8–12.7% of patients receiving intravenous injections of high-osmolar, ionic RCM and in 0.7– 3.1% of patients receiving low-osmolar non-ionic RCM. Severe immediate reactions have been reported to occur with a frequency of 0.1–0.4% for ionic RCM and with a frequency of 0.02–0.04% for non-ionic ones. The frequency of non-immediate adverse reactions ranges from 0.5 to 23%. This large variation may be due to the difficulty

5

in verifying whether symptoms occurring hours or days after RCM exposure are actually caused by the RCM. When radiological examinations with use of RCM were compared with examinations without RCM, most non-immediate symptoms, except skin reactions, were found to be unrelated to the RCM administration. Thus, various types of exanthema seem to account for the majority of the RCM-induced non-immediate hypersensitivity reactions. Such eruptions have been reported to affect some 1–3% of RCM-exposed patients. With regard to fluorescein, an Australian review reported 132 adverse reactions among a total of 11,898 fluorescein angiograms performed between 1998 and 2004 [31]. Reactions were mainly nausea and vomiting, but dizziness, fainting, localized reactions, and urticaria also occurred. There were no serious adverse reactions or deaths recorded. As far as antiepileptic drugs are concerned, a recent study evaluated the risk of hospitalization for SJS and TEN in new users and estimated that it was low for carbamazepine, lamotrigine, phenytoin, and phenobarbital (ranging between 1 and 10 per 10,000 new users) and significantly lower for valproic acid. Furthermore, more than 90% of cases occurred within the first 63 days of antiepileptic use [32]. Data on Outpatients and General Population Epidemiological data on drug hypersensitivity in non-hospitalized subjects and the general population are even scarcer and are limited mainly to studies on antibiotic use (table 2). A prospective study of patients receiving monthly injections of penicillin G (for rheumatic fever) found 57 reactions in 1,790 patients (incidence of 3.2% of patients and 0.19% of injections), 4 cases of anaphylaxis (incidence of 0.2% of patients and 0.01% of injections), and 1 fatality (incidence of 0.05% of patients and 0.003% of injections) [33]. Apter et al. [34] led a retrospective cohort study using the UK General Practice Research Database. Records

6

of patients who received at least two prescriptions for penicillin at least 60 days apart were selected and examined for hypersensitivity reactions. A penicillin prescription was given to 3,375,162 patients (all ages), of which 6,212 (0.18%) experienced an allergic-like event. Of these, 48.5% were given a second prescription after the event and only 1.89% had another event, suggesting that penicillin hypersensitivity is less frequent in outpatients than in hospitalized patients and that most reactions resembling drug allergy are not drug related. On the other hand, although the difference in incidence is small in the group reporting a previous reaction, the risk of a second event increased 11.2 times. However, higher numbers are reported in the meta-analysis by Impicciatore et al. [35], where the reported incidence of ADRs in pediatric outpatients was 1.46%. In another review of 5,923 records from a private group pediatric practice in northern Virginia, cutaneous eruptions occurred in 7.3% of children who were given common oral antibiotics [36]. With regard to quinolones, 3 out of 3,200 students treated with ciprofloxacin to prevent meningococcal carriage experienced an anaphylactic reaction [37]. A cross-sectional survey of a general adult population from Porto, Portugal, found a global 7.8% (181/2,309) prevalence of self-reported drug allergy; 4.5% to penicillins or other -lactams, 1.9% to aspirin or other NSAIDs and 1.5% to other drugs. Most of the reported reactions were immediate (43%), occurred on the first day of treatment (78.5%), and involved the skin (63.5%), and thus could have been immunemediated [38]. Similar results were found among university students using a comparable methodology [39]. Between 2003 and 2004, a subsample of 282 general practitioners in the BEACH (Bettering the Evaluation and Care of Health) data collection program in Australia recorded patient responses to questions about ADRs [40]. From 8,215 encounters, doctors reported that 852 patients (10.4%) had experienced an ADR in the previous 6 months. Patients aged 1 45 years (ver-

Demoly ⴢ Viola ⴢ Rebelo Gomes ⴢ Romano

Table 2. Epidemiological data on outpatients and general population Ref.

Patients

Patients with ADRs

Incidence of ADRs %

Period

Type of ADRs

Culprit drugs Remarks

33

1,790 (32,430 injections)

57

3.2 (ADRs) 0.2 (anaphylaxis) 0.05 (fatality)

1988–1990

B

benzathine penicillin

patients with long-term drug prophylaxis for rheumatic fever 4 anaphylaxis, 1 fatality

34

3,375,162 (1st prescription) 2,017,957 (2nd prescription)

6,212

0.18

1987–2001

B

penicillin

3,509

1.89 with previous reactions 0.17 without

risk of DHRs is markedly increased in subjects with a previous reaction

35

11,513 children

192

1.46

36

5,923 children records

472 erythemas

7.3

37

3,200

8

1:1,000

38

2,309

181

7.8

39

2,150

40

8,215

852

not specified 1996

64 reactions definite/probable 1 severe ADR

B

antibiotics

none were severe

B

ciprofloxacin 3 anaphylaxis 3 erythemas 2 nausea and vomiting

2002

B

mainly -lactams

63.5% cutaneous ADRs

7.7

2001

B

antibiotics NSAIDs

mainly cutaneous ADRs

10.4

2003–2004

A B

>50% mild ADRs 35.8% moderate ADRs 10% severe ADRs risk factors: old age and female gender

ADRs = Adverse drug reactions; DHRs = drug hypersensitivity reactions; NSAIDs = non-steroidal anti-inflammatory drugs.

sus !45 years), children aged 1–4 years (versus older children), and female patients (versus male patients) were significantly more likely to have experienced an ADR. Most patients (83.5%) had experienced only one ADR, while 10.7 and 5.8% had experienced two or more events, respectively. For 71.9% of the patients, one reason for the most recent event was a recognized side effect, followed by drug sensitivity (12.4%) and allergy (11.0%). Over half of the patients were rated as having a ‘mild’ event, with 35.8% rated as ‘moderate’ and 10.0% as ‘severe’. General practitioners classified 23.2% of events as preventable, and 7.6% of events resulted in hospitalization.

Epidemiology and Causes of Drug Hypersensitivity

It is possible to conclude that DHRs represent up to one third of ADRs, which may affect 7% of the general population and up to 20% of hospitalized patients besides being responsible for as much as 8% of hospital admissions. Reactions are in most cases not declared, but reported numbers can also be inflated by the lack of a definite diagnosis. General Data on Mortality DHRs can be serious and life-threatening. All routes of administration are potentially lethal: oral, injectable (i.v., i.m., s.c., intralymphatic, intra-articular), inhaled, topical (intrauterine, rectal, cutaneous). The study by Lazarou et al. [16]

7

showed that 0.32% of hospitalized patients in the USA died from ADRs, an estimated 106,000 deaths for the year 1994, which would make them the fourth most frequent cause of death in that country. The proportion of allergic reactions in this study was not evaluated, but could be estimated to be around 23.8% (all severities). Pirmohamed et al. [23] found that the overall fatality rate related to ADRs was 0.15%, and calculated that ADRs that had led to hospital admissions were responsible for 5,700 deaths per year in England. In the study by Fattinger et al. [18] the estimated incidence (0.14%) of possible ADR-related deaths was similar. In the study by Hardmeier et al. [21], among 6,383 inpatients, 10 ADR-related fatalities were reported (0.16% of the included patients), corresponding to 3% of all deaths. A similar percentage (5%) of ADR-related deaths was reported by Juntti-Patinen and Neuvonen [41] in their study evaluating 1,511 in-hospital fatalities. In the study by Moore et al. [42] of 329 patients admitted to an internal medicine ward over 6 months, there were 31 patients (9.4%) with at least one ADR and a fatal reaction occurred in 4 of them (13%). Thong et al. [20] found that the mortality due to drug allergy was 0.09 per 1,000 hospitalizations. Anaphylactic shock is one of the severe reactions commonly associated with drug allergy fatalities. It is usually an IgE-mediated reaction and it is the most frightening and potentially lethal allergic event. Non-IgE-mediated anaphylactic shocks can also be drug related and equally dangerous. In the retrospective study by Kemp et al. [43], of 266 cases of anaphylaxis described by a private allergy practice in Memphis, drugs (20%) were the second most recognizable cause of reactions, with NSAIDs responsible for half of them. The lethal risk related to penicillin anaphylaxis has long been known. There is renewed interest in this subject because injectable antibiotics are responsible for a large percentage (15%) of anaphylaxis during anesthesia, making them the third most important cause after NMBAs and

8

latex [26]. Antibiotics take the first position in cases declared to the French Allergovigilance Network [44]. Based on a medical literature review to obtain prevalence estimates of anaphylaxis in the general population, Neugut et al. [45] calculated a rate of 0.7–10% for penicillins (the USA population at risk would be 1.9–27.2 million) and 0.22–1% for RCM (the USA population at risk would be 22,000–100,000). Matasar and Neugut [46] reported about 1,500 annual deaths from anaphylaxis in the USA. In the UK, where hospital admissions for acute anaphylaxis are increasing (from 56 per million in 1991 to 102 per million in 1995) [47], the work by Pumphrey [48] on deaths from anaphylaxis (1992–2001) shows that drugs are the leading cause (88 deaths out of 202) followed by food allergy and insect stings. In the complete 1992–1998 UK registry of fatal anaphylaxis, there were 164 fatalities, of which druginduced anaphylaxis constituted 39% of the cases, with 27 cases from anesthetics, 16 from antibiotics, and 8 from RCM [49]. In another recent analysis by Peng and Jick [50] using a general practice research database (1994–1999), 675 cases of anaphylaxis were reported; thus, the estimated incidence of 8.4 per 100,000 persons/year of anaphylaxis in the UK, with oral medicines being the second most common cause after insect stings. The recent 3-year Swiss study by Helbling et al. [51] led to the identification of 226 individuals diagnosed with 246 episodes of life-threatening anaphylaxis and 3 deaths. The authors calculated an annual incidence of 7.9–9.6 cases per 100,000 inhabitants per year of anaphylaxis and identified drugs as the second most common cause (18.1%) after hymenoptera stings. Van der Klauw et al. [52] analyzed 345 cases of probable anaphylaxis and 485 cases of possible anaphylaxis associated with drugs over a 20-year period in the Netherlands. In this study, the mortality from anaphylactic shock was 2.5% (21 cases of 830) and the implicated drugs were: RCM (8 deaths), dextran (3), glafenin (2), immunotherapy for allergy (2), protamine (1), penicillin (1), tetracosac-

Demoly ⴢ Viola ⴢ Rebelo Gomes ⴢ Romano

tide (1), metoprolol (1), erythromycin (1), and butylscopolamine with metimazole (1). In Italy, Cianferoni et al. [53] carried out a retrospective review of the clinical features of 113 episodes of anaphylaxis regarding 107 patients and resulting in admission to hospital. Drugs (NSAIDs and antibiotics) were the most frequent cause of reactions (49%), followed by hymenoptera stings (29%). From 1968 to 1990 the Danish Committee on Drug Administration [54] recorded 30 cases of fatal anaphylaxis, of which 8 were caused by RCM, 6 by penicillins, 5 by allergen extracts, 2 by NSAIDs and 1 by a NMBA (incidence of 0.3 cases of fatal anaphylaxis per million inhabitants per year). Anaphylaxis is not, however, the only cause of mortality due to allergic drug reactions. The serious, mostly drug-related delayed dermatological conditions SJS (ca. 10% mortality) and TEN (30% mortality), with estimated incidences of 0.4–1.2 and 1.2–6 per million people/year, respectively [55], are other examples of life-threatening reactions [56]. They occurred in 5.2% of the 210 drugallergic patients in the study by Thong et al. [20]. The multisystem organ hypersensitivity syndrome (10% mortality) and organ-specific involvement, including hepatitis, can also contribute to drug hypersensitivity-related mortality.

Risk Factors for Drug Hypersensitivity Reactions

Some risk factors related to drugs, treatment regimens, and patients (such as age, gender, concurrent illnesses, and previous reactions to related drugs), have been identified as having an important role in drug hypersensitivities. Drug-Related Aspects A large variety of drugs are currently used in everyday practice. However, those implicated in allergic reactions are a much smaller group. In order to be immunogenic or a complete allergen, it

Epidemiology and Causes of Drug Hypersensitivity

is believed that a substance must have a sufficient molecular weight (11,000 daltons); thus, most drugs behave as haptens and have to bind to carrier proteins to induce a specific immune response. -Lactams are intrinsically reactive (supporting the hapten concept of the pathophysiology of drug allergy), other drugs, like sulfamethoxazole, require previous conversion to a reactive intermediate (pro-hapten concept). Drug-related cytotoxicity may also be of importance in enhancing the immune response (danger concept). Although not reactive, some other drugs can still be immunogenic by direct noncovalent binding to immune receptors, mostly Tcell receptors (pharmacological interaction concept) [57]. The notion that the type of drug itself is an important risk factor for drug allergy even among the same therapeutic group can be illustrated by the review by Ibia et al. [36]. Based on the number of patients for whom each group of antibiotics was prescribed, the documented frequencies of reactions were: 12.3% for cefaclor, 8.5% for sulfonamides, 7.4% for penicillins, and 2.6% for other cephalosporins. This is also shown in another survey from the Boston Collaborative Drug Surveillance Program [58]. In this survey, which provided most of the data regularly used regarding the incidence of DHRs, the authors analyzed the incidence of cutaneous drug reactions in 15,438 hospitalized patients. There were 358 reactions reported and confirmed by a dermatologist, with an overall incidence of 2.3%. The number of reactions over the number of administrations for each drug was 5.1% for amoxicillin, 3.4% for cotrimoxazole, 3.3% for ampicillin, 2.1% for cephalosporins, 2% for erythromycin, 1.8% for penicillin G, and 0.4% for gentamycin. Another interesting aspect is the changing pattern of reactivity to certain drugs over time, which has been especially demonstrated with -lactams [59]. Traditionally, reactions to -lactams concerned mostly penicillin G, but in recent years reactions to amoxicillin and to cephalosporins have been increasing and their prevalence seems

9

to differ among populations and/or countries (e.g., anaphylaxis associated with -lactams appears to be rare in Central Europe and common in Southern Europe). Treatment Regimen The dosage of the drug and the mode of administration influence the frequency of reactions. It appears that intermittent and repeated administrations can be more sensitizing than an uninterrupted treatment [60]. This is supported by a recent publication by Cetinkaya and Cag [61] who studied 147 children who had received -lactams at least 3 times in the preceding 12 months without allergic reaction. A 10.2% frequency of positive skin tests to penicillin was found and the authors concluded that frequent use of -lactam antibiotics leads to sensitization. Pichichero and Pichichero [62], however, found no difference in the frequency of previous -lactam treatments in skin-test-negative and skin-test-positive children. With regard to the route of administration, the parenteral route is considered the most immunogenic, although strong data showing that the parenteral route is more immunogenic than the oral route are lacking. Topical administration to the skin (and not the mucosa) is an important sensitization pathway [63, 64]. Host-Related Factors Host-related factors can predispose patients to drug allergy, especially by acting on the way the drug is processed. Most studies show that women are more often affected than men (65–70 vs. 30– 35%) [65–67]. Differences can, however, depend on the age group [25], on the type of reaction (cutaneous reaction rates were 35% higher in females than males in the Bigby et al. [58] review), and on the culprit drug [38]. In a gemifloxacine unpublished study, 30% of women in child-bearing age reacted as compared to only 4% of males (http:// www.fda.gov). Estrogen production or use possibly plays a role. Subgroup analysis of the study

10

of Thong et al. [20] showed that hospitalized female patients were statistically significantly more likely to develop drug allergy than males, although there were no significant differences in the clinical manifestations and mortality between both genders. In addition, patients 165 years of age did not appear to be more at risk of developing drug allergy than those !65 years and there was no increase in the severity of allergic reactions or drug-related mortality [pers. commun., unpubl. data]. It is often reported that children are less affected than adults [68]. In the study on pediatric patients by Temple et al. [69], 565 ADRs were reported by a hospital surveillance program over a 6-year period (1994–1999), with an ADR rate of 0.85 per 100 admissions. Antibiotics were the most frequently implicated drug class (26.4%), although 2.8% of patients had a documented allergy to the medication suspected of causing the reaction. However, other studies on ADRs in pediatric populations report incidences similar to the ones in adults [35, 36]. In the 20-day observational prospective study of Mazzeo et al. [22], the rate of reactions (most of them probably not due to allergy) to antibiotics in inpatients was higher in children than in adults. An 8-month survey of ADR in a pediatric isolation ward by Weiss et al. [70] showed that among 68 ADRs detected in 46 (21.5%) of 214 patients, 7 (10%) were classified as severe and 50% were antibiotic-associated. Of all the ADRs, 14 (20.6%) were considered to be of an immunologic nature. The role of atopy is still under debate, but it does not seem to be a major risk factor [66, 67, 71]. The influence of atopy may, however, depend on the drug in question and it was reported to be a risk factor for NSAID hypersensitivity, especially when cutaneous reactions are present [72]. Some ethnic groups and genetic backgrounds seem to facilitate certain types of ADRs (table 3). Easterbrook et al.’s [73] study on epidemiological risk factors for hypersensitivity reactions to abacavir found the Caucasian race to be a risk factor

Demoly ⴢ Viola ⴢ Rebelo Gomes ⴢ Romano

Table 3. Genetic risk factors for drug hypersensitivity reactions Ref.

Population

Genetic predisposition

Drug-related clinical symptoms

25

222 hospitalized children

old age, male gender, Chinese descent, asthma and associated chronic illness

ADRs due to mainly -lactams and NSAIDs

white race and a high CD8 cell count at therapy initiation

abacavir hypersensitivity

East Asians were more likely to develop cough and hyperkalemia, and African-Americans to develop angioedema

ADRs associated with ACE inhibitors

African and Asian people

ACE inhibitors-induced cough

more frequently in Blacks (18.0%) than in Non-Blacks (3.2%)

self-reported chloroquine allergy

abnormal des-Arg(9)-BK degradation

ACE inhibitors-induced angioedema

73 74

2,225

76 77

Maputo, Mozambique

78 79

Australian

HLA-B*5701, HLA-DR7 and HLA-DQ3 incidence of HLA-B*5701 allele = 94.4%

abacavir hypersensitivity

80

Australian

HLA-B*5701 and Hsp70-Hom M493T alleles haplotypic polymorphism within the TNF promoter region (TNF–238A)

abacavir hypersensitivity

83

USA

HLA-B*5701 (incidence = 22.2%)

abacavir hypersensitivity

84

TNF promoter polymorphism (–308 position)

severity of CBZ hypersensitivity

85

English

HSP70 gene variants

CBZ-induced SJS/TEN

86#, 87

Han Chinese

HLA* 1502 gene, #incidence of this allele = 100%

CBZ-induced SJS/TEN

89

European

HLA* 1502 allele and Asian ancestry in 4/12 cases

CBZ-induced SJS/TEN

92

Han Chinese

HLA-B*5801

allopurinol-induced severe cutaneous allergic reactions

94

LTC4S promoter polymorphism

ASA-induced asthma

95

familiar aggregation inheriting the LTC4S –444C allele deletion of GSTM1

ASA-induced urticaria

96

Japanese

promoter –1993T>C polymorphism in the TBX21 gene, coding for T-bet

ASA-induced asthma

97

Korean

HLA-alleles DRB1*1302 and DQB1*0609, promoter polymorphisms of ALOX5 (–1708A>G), CysLTR1 (–634C>T) and FcR1 (–344C>T, –95T>C)

ASA-induced urticaria/angiodema

FcR1 (–190T>C) polymorphism of MS4A2 gene

ASA-induced asthma

Korean

CysLTR1 variants

ASA-induced asthma

98 99 100

ALOX5 Sp1 repeat polymorphism

severity of ASA-induced asthma

101

Chinese

E237G variant of FcR1 gene

IgE-mediated penicillin allergy

102

Chinese

IL-4RQ576R polymorphism

IgE-mediated penicillin allergy

103

Chinese

IL-4-IL-13-SNP polymorphisms

IgE-mediated penicillin allergy

104

Italian

polymorphisms of IL-13 (R130Q and –1055C>T variants) and IL-4RA (I50 V, S478P, and Q551R variants)

immediate allergic reactions to -lactams

105

Caucasian

Ile75Val variant of IL-4R gene two linked IL-10 promoter gene polymorphisms (–819C>T and –592C>A)

immediate allergic reactions to -lactams

ACE = Angiotensin-converting enzyme; ADRs = adverse drug reactions; ALOX5 = 5-lipoxygenase; BK = bradykinin; CBZ = carbamazepine; CYP2C9 = cytochrome P4502C9; CysLTR1 = cysteinyl LT receptor 1; GSTM1 = glutathione S-transferase M1; FcR1 = high-affinity IgE receptor  chain; FcR1 = high-affinity IgE receptor  chain; HSP = heat-shock protein; LTC4S = leukotriene C4 synthase; NSAIDs = non-steroidal antiinflammatory drugs; SJS/TEN = Stevens-Johnson syndrome/toxic epidermal necrolysis.

Epidemiology and Causes of Drug Hypersensitivity

11

for reactions. In a recent cohort study evaluating risk factors for ADRs associated with angiotensin-converting enzyme (ACE) inhibitors involving 2,225 people, of whom 19% had to discontinue therapy due to ADRs, African-Americans were found to be more susceptible to developing ACE-related angioedema than other ethnic groups [74]. This confirmed the results of other authors [75]. African and Asian people also appear to be at an increased risk for ACE inhibitorinduced cough [76]. In the study by Lunet et al. [77], self-reported chloroquine allergy in Maputo, Mozambique, was more frequent in Blacks (18.0%) compared with Non-Blacks (3.2%). In the study by Kidon et al. [25], Chinese descent, asthma, and associated chronic illness were all considered as independent risk factors for ADRs. These differences may be due to genetic polymorphisms that alter drug metabolism or immune responses in some individuals, leading to an increased susceptibility to certain drugs or to certain ADRs. For example, individuals with ACE genotype II are reported to have an increased risk for ACE inhibitor-induced cough [76] and in those with angioedema an abnormal degradation of some bradykinin metabolites has been described [78]. With regard to abacavir hypersensitivity, Mallal et al. [79, 80] demonstrated that it is linked to HLA-B*5701 and to its combination with a haplotypic Hsp70-Hom M493T variant. The authors also suggest that in Whites, genotyping for HLA-B*5701 should be performed before prescription of abacavir [79] and this strategy appears to reduce the incidence of abacavir hypersensitivity in their clinic [81]. An alternative flow cytometry method for HLA-B57 phenotyping using commercially available B17 monoclonal antibodies has been developed and represents a sensitive, rapid and cost-effective alternative to HLA typing as a screening assay prior to abacavir prescription [82]. The incidence of this allele in abacavir hypersensitive persons is high (94.4%) [79] in the Australian cohort, but

12

in other studies [83] it is lower (22.2%), but still significantly higher than average. A haplotypic polymorphism within the TNF promoter region (TNF–238A) may also affect the levels of TNF production influencing the severity of abacavir reactions [80]. Another TNF promotor polymorphism (–308TNF) was associated with a more severe course of carbamazepine hypersensitivity reactions [84]. Concerning carbamazepine hypersensitivity, recent studies have also found an association between SJS and TEN and the heat-shock protein (HSP70) gene variants [85], as well as the HLA-B*1502 gene [86, 87]. The incidence of this allele is high (100%) in the Han Chinese with carbamazepine-induced severe bullous skin reactions [86], but it is not in Whites [88, 89]. Interestingly, although only 4 out of 12 carbamazepine-induced SJS/TEN cases of the European study RegiSCAR had the HLAB*1502 allele, it is remarkable that all 4 had an Asian ancestry [89]. This shows that although this HLA region may contain important genes for SJS/TEN, this allele is not universal and ethnicity matters. In vitro data have suggested that these patients may have a detoxification defect [88]. The work by Bavdekar et al. [90] proposes that accumulation of toxic arene oxide metabolites, due to a defect in epoxide hydrolase-mediated detoxification, contributes to anticonvulsant hypersensitivity. However, no specific polymorphisms have been pinpointed in the epoxide hydrolase gene [88, 91]. Chen’s group [92] also demonstrated a strong association in Han Chinese between the genetic marker HLA-B*5801 allele and allopurinol-induced severe cutaneous allergic reactions. With regard to NSAID ADRs, it was recently suggested that NSAID-related gastrointestinal bleeding could be associated with two polymorphisms of cytochrome P4502C9 [93]. As far as NSAID hypersensitivity is concerned, a few gene polymorphisms have been found. In bronchial biopsies of patients with aspirin-induced asthma, an overexpression of leukotriene C4 synthase was reported. This could

Demoly ⴢ Viola ⴢ Rebelo Gomes ⴢ Romano

be partially explained by a genetic polymorphism in the promoter region of this enzyme [94]. In aspirin-induced urticaria, a familial aggregation inheriting this –444C allele of the leukotriene C4 synthase gene has also been found, as well as a deletion of the glutathione S-transferase M1 [95]. A Japanese study demonstrated that the promoter –1993T 1C polymorphism in the TBX21 gene, coding for T-bet, a Th1-specific transcription factor, is associated with a risk of aspirin-induced asthma [96]. The Korean cohort of aspirin hypersensitivity showed that HLA alleles DRB1*1302 and DQB1*0609 may be genetic markers of aspirin-induced urticaria/angioedema. The promoter polymorphisms of the 5-lipoxygenase (–1708A 1G), cysteinyl leukotreiene receptor 1 CysLTR1 (–634C1T) and high-affinity IgE receptor  chain FcR1 (–344C1T, –95T 1C) genes were also described [97]. Other studies of the same group suggest that the genetic variants of FcR1 (–190T 1C) [98], CysLTR1 [99] and tandem repeat in 5-lipooxygenase promoter are associated with aspirin-induced asthma [100]. Finally, concerning -lactam hypersensitivity, Qiao’s group showed that the E237G variant of FcR1 [101], IL-4RQ576R [102], IL-4-IL-13-SNP polymorphisms [103] may participate in the development of IgE-dependent penicillin allergy in a Chinese population. A recent study performed on an Italian population evaluated the association between immediate allergic reactions to -lactams, specifically penicillins and cephalosporins, and the polymorphisms of IL-13 (R130Q and –1055C1T variants) and IL-4RA (I50V, S478P, and Q551R variants). The combination of the less frequent allele of the IL-13 R130Q polymorphism with any of the predominant homozygous genotypes of the three polymorphisms of IL-4RA was more significantly associated with the risk of -lactam allergy (p = 0.0006, 0.0077, and 0.0041, respectively) than any other polymorphism considered alone (p = 0.1745, 0.0268, 0.1812, 0.0152, respectively). The same associations were ob-

Epidemiology and Causes of Drug Hypersensitivity

served with serum IgE levels (IL-13/IL-4RA variant combinations: p = 0.0009, 0.0007, 0.0020, respectively and each variant: p = 0.0201, 0.0021, 0.0531, and 0.0417, respectively). The combination of IL-4RA variants with 1055C1T polymorphism produced similar associations [104]. In a Caucasian population, Guglielmi et al. [105] found among atopic subjects two significant associations between immediate -lactam allergy in women and the Ile75Val variant of IL-4R gene and two linked IL-10 promoter gene polymorphisms (–819C1T and –592C1A). The time has now come to identify and analyze all relevant gene polymorphisms involved in drug hypersensitivity. Drug hypersensitivity appears to be more frequent in certain diseases and therefore, concomitant pathologies might play a role. For example, hypersensitivity reactions to NSAIDs are particularly frequent in some populations, such as asthmatics [106]. Interestingly, Ventura et al. [107] describe hypersensitivity to aspirin as a risk factor, along with female gender and atopy, for immediate reactions to glucocorticoids. HIV-infected patients suffer 10–100 times more from cutaneous reactions to drugs, especially to cotrimoxazole [108]. The frequency of drug hypersensitivity in this population ranges from 3 to 20%. A recent review by Temesgen and Beri [109] on drug allergy in HIV patients showed comprehensive data on reactions associated with antiretroviral drugs as well as with cotrimoxazole. Aminopenicillininduced exanthema in florid infectious mononucleosis may result from specific sensitization to the drug, although the exact role of Epstein Barr virus remains unknown [110, 111]. Finally, many factors can combine to lead to an ADR: drugs [26], infection [112], exercise, and food allergy [113] may interact with each other to induce a hypersensitivity reaction or increase its severity.

13

Conclusion

There are few epidemiological data on DHRs which affect up to 20% of hospitalized patients and up to 7% of outpatients. The available information, based predominantly on the epidemiology of ADRs, requires cautious interpretation, because these reactions are rarely accurately classified or definitively diagnosed. Both under-diagnosis (due to under-reporting [114, 115]) and over-diagnosis (due to the over-use of the term ‘allergy’ [116]) have also to be considered. Mis-

classification based on drug allergy histories may have consequences on individual treatment choices and lead to the use of more expensive and less effective drugs. Multicenter studies, both in hospital-based populations and in general populations, using the same methodologies and definitions would also be of great value in order to have an accurate global perspective about risk factors and possible regional differences and to allow the implementation of better preventive measures for patients.

References 1 WHO: International drug monitoring: the role of national centres. Tech Rep Ser No 498. Geneva, WHO, 1972. 2 Edwards IR, Aronson JK: Adverse drug reactions: definitions, diagnosis and management. Lancet 2000;356:1255– 1259. 3 Meyboom RH, Lindquist M, Egberts AC: An ABC of drug-related problems. Drug Saf 2000;23:99–105. 4 Stephens MDB: Definitions and classifications of adverse reaction terms; in Stephens MDB, Talbot JCC, Routledge PA (eds): The Detection of New Adverse Reactions, ed 4. London, Macmillan, 1998, pp 32–44. 5 Ross SD: Drug-related adverse events: a readers’ guide to assessing literature reviews and meta-analyses. Arch Intern Med 2001;161:1041–1046. 6 Nebeker JR, Barach P, Samore MH: Clarifying adverse drug events: a clinician’s guide to terminology, documentation and reporting. Ann Intern Med 2004;140:795–801. 7 Rawlins MD, Thompson JW: Pathogenesis of adverse drug reactions; in Davies DM (ed): Textbook of Adverse Drug Reactions. Oxford, Oxford University Press, 1977, p 10. 8 Aronson JK: Drug therapy; in Haslett C, Chilvers ER, Boon NA, Colledge NR, Hunter JAA (eds): Davidson’s Principles and Practice of Medicine, ed 19. Edinburgh, Elsevier Science, 2002, pp 147–163.

14

9 Grahame-Smith DG, Aronson JK: Adverse drug reactions; in The Oxford Textbook of Clinical Pharmacology and Drug Therapy. Oxford, Oxford University Press, 1984, pp 132–157. 10 Hartingan-Go KY, Wong JQ: Inclusion of therapeutic failures as adverse drug reactions; in Aronson JK (ed): Side Effects of Drugs, Annual 23. A Worldwide Yearly Survey of New Data and Trends in Adverse Drug Reactions. Amsterdam, Elsevier, 2000. 11 Aronson JK, Ferner RE: Joining the DoTS: new approach to classifying adverse drug reactions. BMJ 2003;327: 1222–1225. 12 Johansson S, Bieber T, Dahl R, et al: Revised nomenclature for allergy for global use: report of the Nomenclature Review Committee of the World Allergy Organization, October 2003. J Allergy Clin Immunol 2004;113:832– 836. 13 Hayward RA, Hofer TP: Estimating hospital deaths due to medical errors: preventability is in the eye of the reviewer. JAMA 2001;286:415–420. 14 Bates DW, Cullen DJ, Laird N, et al: Incidence of adverse drug events and potential adverse drug events. JAMA 1995;274:29–34. 15 Classen DC, Pestotnik SL, Evans RS, Burke JP: Computerized surveillance of adverse drug events in hospital patients. JAMA 1991;266:2847–2851.

16 Lazarou J, Pomeranz BH, Corey PN: Incidence of adverse drug reactions in hospitalized patients. A meta-analysis of prospective studies. JAMA 1998;279: 1200–1205. 17 Kvasz M, Allen E, Gordon MJ, et al: Adverse drug reactions in hospitalized patients: a critique of meta-analysis. Med Gen Med 2000;2:E3. 18 Fattinger K, Roos M, Vergeres P, et al: Epidemiology of drug exposure and adverse drug reactions in two Swiss departments of internal medicine. Br J Clin Pharmacol 2000;49:158–167. 19 Olivier P, Boulbes O, Tubery M, et al: Assessing the feasibility of using an adverse drug reaction preventability scale in clinical practice: a study in a French emergency department. Drug Saf 2002;25:1035–1044. 20 Thong BY, Leong KP, Tang CY, Chng HH: Drug allergy in a general hospital: results of a novel prospective inpatient reporting system. Ann Allergy Asthma Immunol 2003;90:342–347. 21 Hardmeier B, Braunschweig S, Carvallaro M, et al: Adverse drug events caused by medication errors in medical inpatients. Swiss Med Wkly 2004;134: 664–670. 22 Mazzeo F, Capuano A, Avolio A, Filippelli A, Rossi F: Hospital-based intensive monitoring of antibiotic-induced adverse events in a university hospital. Pharmacol Res 2005;51:269–274

Demoly ⴢ Viola ⴢ Rebelo Gomes ⴢ Romano

23 Pirmohamed M, James S, Meakin S, et al: Adverse drug reactions as cause of admission to hospital: prospective analysis of 18,820 patients. BMJ 2004; 329:15–19. 24 Fiszenson-Albala F, Auzerie V, Mahe E, et al: A 6-month prospective survey of cutaneous drug reactions in a hospital setting. Br J Dermatol 2003; 149:1018– 1022. 25 Kidon MI, See Y: Adverse drug reactions in Singaporean children. Singapore Med J 2004;45:574–577. 26 Mertes PM, Laxenaire MC, Alla F, Groupe d’Etudes des Réactions Anaphylactoïdes Peranesthésiques: Anaphylactic and anaphylactoid reactions occurring during anesthesia in France in 1999–2000. Anesthesiology 2003;99: 536–545. 27 Harboe T, Guttormsen AB, Irgens A, Dybendal T, Florvaag E: Anaphylaxis during anesthesia in Norway: a 6-year single-center follow-up study. Anesthesiology 2005;102:897–903. 28 Karila C, Brunet-Langot D, Labbez F, et al: Anaphylaxis during anesthesia: results of a 12-year survey at a French pediatric center. Allergy 2005;60:828– 834. 29 Katayama H, Yamaguchi K, Kozuka T, et al: Adverse reactions to ionic and nonionic contrast media. A report from the Japanese Committee on the Safety of Contrast Media. Radiology 1990;175: 621–628. 30 Brockow K, Christiansen C, Kanny G, et al: Management of hypersensitivity reactions to iodinated contrast media. Allergy 2005;60:150–158. 31 Kwan AS, Barry C, McAllister IL, Constable I: Fluorescein angiography and adverse drug reactions revisited: the Lions Eye experience. Clin Experiment Ophthalmol 2006;34:33–38. 32 Mockenhaupt M, Messenheimer J, Tennis P, Schlingmann J: Risk of StevensJohnson syndrome and toxic epidermal necrolysis in new users of antiepileptics. Neurology 2005;64:1134–1138. 33 International Rheumatic Fever Study Group: Allergic reactions to long-term benzathine penicillin prophylaxis for rheumatic fever. Lancet 1991;337:1308– 1310. 34 Apter AJ, Kinman JL, Bilker WB, et al: Represcription of penicillin after allergic-like events. J Allergy Clin Immunol 2004;113:764–770.

35 Impicciatore P, Choonara I, Clarkson A, et al: Incidence of adverse drug reactions in paediatric in-/out-patients: a systematic review and meta-analysis of prospective studies. Br J Clin Pharmacol 2001;52:77–83. 36 Ibia EO, Schwartz RH, Wiedermann BL: Antibiotic rashes in children: a survey in a private practice setting. Arch Dermatol 2000;136:849–854. 37 Burke P, Burne SR: Allergy associated with ciprofloxacin. Br Med J 2000;320: 679. 38 Gomes E, Cardoso MF, Praça F, et al: Self-reported drug allergy in a general adult Portuguese population. Clin Exp Allergy 2004;34:1597–1601. 39 Falcao H, Lunet N, Gomes E, et al: Drug allergy in university students from Porto, Portugal. Allergy 2003;58:1210. 40 Miller GC, Britth HC, Valenti L: Adverse drug events in general practice patients in Australia. Med J Aust 2006; 184:321–324. 41 Juntti-Patinen L, Neuvonen PJ: Drugrelated death in a university central hospital. Eur J Clin Pharmacol 2002;58: 479–482. 42 Moore N, Lecointre D, Noblet C, Mabille M: Frequency and cost of serious adverse drug reactions in a department of general medicine. Br J Clin Pharmacol 1998;45:301–308. 43 Kemp SF, Lockey RF, Wolf BL, Lieberman P: Anaphylaxis. A review of 266 cases. Arch Intern Med 1995;155:1749– 1754. 44 Moneret-Vautrin DA, Morisset M, Flabbee J, Beaudouin E, Kanny G: Epidemiology of life-threatening and lethal anaphylaxis: a review. Allergy 2005;60: 443–451. 45 Neugut AI, Ghatak AT, Miller RL: Anaphylaxis in the United States: an investigation into epidemiology. Arch Intern Med 2001;161:15–21. 46 Matasar MJ, Neugut AI: Epidemiology of anaphylaxis in the United States. Allergy Asthma Rep 2003;3:30–35. 47 Sheikh A, Alves B: Hospital admissions for acute anaphylaxis: time trend study. Br Med J 2000;320:1441. 48 Pumphrey R: Anaphylaxis: Can we tell who is at risk of a fatal reaction? Curr Opin Allergy Clin Immunol 2004;4: 285–290. 49 Pumphrey RS: Lessons for management of anaphylaxis from a study of fatal reactions. Clin Exp Allergy 2000; 30:1144–1150.

Epidemiology and Causes of Drug Hypersensitivity

50 Peng MM, Jick H: A population-based study of incidence, cause and severity of anaphylaxis in the United Kingdom. Arch Intern Med 2004;164:317–319. 51 Helbling A, Hurni T, Mueller UR, Pichler WJ: Incidence of anaphylaxis with circulatory symptoms: a study over a 3year period comprising 940,000 inhabitants of the Swiss canton Bern. Clin Exp Allergy 2004;34:285–290. 52 Van der Klauw MM, Wilson JH, Stricker BH: Drug-associated anaphylaxis: 20 years of reporting in the Netherlands (1974–1994) and review of the literature. Clin Exp Allergy 1996;26:1355– 1363. 53 Cianferoni A, Novembre E, Mugnaini L, et al: Clinical features of acute anaphylaxis in patients admitted to a university hospital: an 11-year retrospective review (1985–1996). Ann Allergy Asthma Immunol 2001;87:27–32. 54 Lenler-Petersen P, Hansen D, Andersen M, Sorensen HT, Bille H: Drug-related fatal anaphylactic shock in Denmark 1968–1990. A study based on notifications to the Committee on Adverse Drug Reactions. J Clin Epidemiol 1995; 48:1185–1188. 55 Roujeau JC, Stern RS: Severe cutaneous adverse reactions to drugs. N Engl J Med 1994;331:1272–1285. 56 Wolf R, Orion E, Marcos B, Matz H: Life-threatening acute adverse cutaneous drug reactions. Clin Dermatol 2005;23:171–181. 57 Britschgi M, von Greyerz S, Burkhart C, Pichler WJ: Molecular aspects of drug recognition by specific T cells. Curr Drug Targets 2003;4:1–11. 58 Bigby M, Jick S, Jick H, Arndt K: Druginduced cutaneous reactions. A report from the Boston Collaborative Drug Surveillance Program on 15,438 consecutive inpatients, 1975 to 1982. JAMA 1986;256:3358–3363. 59 Blanca M, Torres MJ, Mayorga C, et al: Immediate allergic reactions to -lactams: facts and controversies. Curr Opin Allergy Clin Immunol 2004;4: 261–266. 60 De Weck AL: Pharmacologic and immunochemical mechanisms of drug hypersensitivity. Immunol Allergy Clin North Am 1991;11:461–474. 61 Cetinkaya F, Cag Y: Penicillin sensitivity among children without a positive history for penicillin allergy. Pediatr Allergy Immunol 2004:15:278–280.

15

62 Pichichero ME, Pichichero DM: Diagnosis of penicillin, amoxacillin and cephalosporin allergy: reliability of examination assessed by skin testing and oral challenge. J Pediatr 1998;132: 137–143. 63 Juan WH, Yang LC, Hong HS: Acute generalized exanthematous pustulosis induced by topical lindane. Dermatology 2004;209:239–240. 64 McIlwain M, Primosch R, Bimstein E: Allergic reaction to intranasal midazolam HCL: a case report. Pediatr Dent 2004;26:359–361. 65 Barranco P, Lopez-Serrano MC: General and epidemiological aspects of allergic drug reactions. Clin Exp Allergy 1998;28(suppl 4):S61–S62. 66 Haddi E, Charpin D, Tafforeau M, et al: Atopy and systemic reactions to drugs. Allergy 1990;45:236–239. 67 Asero R: Detection of patients with multiple drug allergy syndrome by elective tolerance tests. Ann Allergy Asthma Immunol 1998;80:185–188. 68 Demoly P, Bousquet J: Epidemiology of drug allergy. Curr Opin Allergy Clin Immunol 2001;1:305–310. 69 Temple ME, Robinson RF, Miller JC, et al: Frequency and preventability of adverse drug reactions in paediatric patients. Drug Saf 2004;27:819–829. 70 Weiss J, Krebs S, Hoffmann C, et al: Survey of adverse drug reactions on a pediatric ward: a strategy for early and detailed detection. Pediatrics 2002;110; 254–257. 71 Ponvert C, Le Clainche L, de Blic J, et al: Allergy to -lactam antibiotics in children. Pediatrics 1999;104:45. 72 Sanchez-Borges M, Capriles-Hulett A: Atopy is a risk for non-steroidal antiinflammatory drug sensitivity. Ann Allergy Asthma Immunol 2000;84: 101–106. 73 Easterbrook PJ, Waters A, Murad S, et al: Epidemiological risk factors for hypersensitivity reactions to abacavir. HIV Med 2003;4:321–324. 74 Morimoto T, Gandhi TK, Fiskio JM, et al: An evaluation of risk factors for adverse drug events associated with angiotensin-converting enzyme inhibitors. J Eval Clin Pract 2004;10:499–509. 75 Brown NJ, Ray WA, Snowden M, Griffin MR: Black Americans have an increased rate of angiotensin-converting enzyme inhibitor associated angioedema. Clin Pharmacol Ther 1996;60:8– 13.

16

76 Dykewicz MS: Cough and angioedema from angiotensin-converting enzyme inhibitors: new insights into mechanisms and management. Curr Opin Allergy Clin Immunol 2004;4:267–270. 77 Lunet N, Falcao H, Sousa M, Bay N, Barros H: Self-reported food and drug allergy in Maputo, Mozambique. Public Health 2005;119:587–589. 78 Molinaro G, Cugno M, Perez M, et al: Angiotensin-converting enzyme inhibitor-associated angioedema is characterized by a slower degradation of desarginine(9)-bradykinin. J Pharmacol Exp Ther 2002;303:232–237. 79 Mallal S, Nolan D, Witt C, et al: Association between presence of HLAB*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet 2002;359:727–732. 80 Martin AM, Nolan D, Gaudieri S, et al: Predisposition to abacavir hypersensitivity conferred by HLA-B*5701 and haplotypic Hsp70-Hom variant. Proc Natl Acad Sci USA 2004;101:4180–4185. 81 Rauch A, Nolan D, Martin A, et al: Prospective genetic screening decreases the incidence of abacavir hypersensitivity reactions in the Western Australian HIV cohort study. Clin Infect Dis 2006;43:92–102. 82 Martin AM, Krueger R, Almeida CA, et al: A sensitive and rapid alternative to HLA typing as a genetic screening test for abacavir hypersensitivity syndrome. Pharmacogenet Genomics 2006;16:353–357. 83 Stekler J, Maenza J, Stevens C, et al: Abacavir hypersensitivity reaction in primary HIV infection. AIDS 2006;20: 1269–1274. 84 Pirmohamed M, Lin K, Chadwick D, Park BK: TNF promoter region gene polymorphisms in carbamazepinehypersensitive patients. Neurology 2001;56:890–896. 85 Alfirevic A, Mills T, Harrington P, et al: Serious carbamazepine-induced hypersensitivity reactions associated with the HSP70 gene cluster. Pharmacogenet Genomics 2006;16:287–296. 86 Chung WH, Hung SI, Hong HS, et al: Medical genetics: a marker for StevensJohnson syndrome. Nature 2004;428: 486.

87 Hung SI, Chung WH, Jee SH, et al: Genetic susceptibility to carbamazepineinduced cutaneous adverse drug reactions. Pharmacogenet Genomics 2006; 16:297–306. 88 Pirmohamed M: Genetic factors in the predisposition to drug-induced hypersensitivity reactions. AAPS J 2006;8: E20–E26. 89 Lonjou C, Thomas L, Borot N, et al: A marker for Stevens-Johnson syndrome…: ethnicity matters. Pharmacogenomics J 2006;6:265–268. 90 Bavdekar SB, Muranjan MN, Gogtay NJ, et al: Anticonvulsant hypersensitivity syndrome: lymphocyte toxicity assay for the confirmation of diagnosis and risk assessment. Ann Pharmacother 2004;38:1648–1650. 91 Gaedigk A, Spielberg SP, Grant DM: Characterization of the microsomal epoxide hydrolase gene in patients with anticonvulsant adverse drug reactions. Pharmacogenetics 1994;4:142–153. 92 Hung SI, Chung WH, Liou LB, et al: HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci USA 2005;102:4134–4139. 93 Martinez C, Blanco G, Ladero JM, et al: Genetic predisposition to acute gastrointestinal bleeding after NSAIDs use. Br J Pharmacol 2004;141:205–208. 94 Sanak M, Simon HU, Szczeklik A: Leukotriene C4 synthase (LTC4s) promoter polymorphism and risk of aspirin-induced asthma. Lancet 1997;350:1599– 1600. 95 Mastalerz L, Setkowicz M, Sanak M, Rybarczyk H, Szczeklik A: Familial aggregation of aspirin-induced urticaria and leukotriene C synthase allelic variant. Br J Dermatol 2006; 154:256– 260. 96 Akahoshi M, Obara K, Hirota T, et al: Functional promoter polymorphism in the TBX21 gene associated with aspirin-induced asthma. Hum Genet 2005; 117:16–26. 97 Kim SH, Ye YM, Lee SK, Park HS: Genetic mechanism of aspirin-induced urticaria/angioedema. Curr Opin Allergy Clin Immunol 2006;6:266–270. 98 Kim SH, Bae JS, Holloway JW, et al: A polymorphism of MS4A2 (-109T 1 C) encoding the -chain of the high-affinity immunoglobulin E receptor (FcR1) is associated with a susceptibility to aspirin-intolerant asthma. Clin Exp Allergy 2006;36:877–883.

Demoly ⴢ Viola ⴢ Rebelo Gomes ⴢ Romano

99 Kim SH, Oh JM, Kim YS, et al: Cysteinyl leukotriene receptor-1 promoter polymorphism is associated with aspirin-intolerant asthma in males. Clin Exp Allergy 2006;36:433–439. 100 Kim SH, Bae JS, Suh CH, et al: Polymorphism of tandem repeat in promoter of 5-lipoxygenase in ASA-intolerant asthma: a positive association with airway hyperresponsiveness. Allergy 2005;60:760–765. 101 Qiao HL, Yang J, Zhang YW: Specific serum IgE levels and FcRI genetic polymorphism in patients with penicillins allergy. Allergy 2004;59:1326– 1332. 102 Qiao HL, Yang J, Zhang YW: Relationships between specific serum IgE, cytokines and polymorphisms in the IL-4, IL-4R in patients with penicillins allergy. Allergy 2005;60:1053– 1059. 103 Yang J, Qiao HL, Dong ZM: Polymorphisms of IL-13 and IL-4-IL-13-SNPs in patients with penicillin allergies. Eur J Clin Pharmacol 2005;61:803– 809. 104 Guéant-Rodriguez RM, Romano A, Béri-Dexheimer M, et al: Gene-gene interactions of IL-13 and IL-4RA variants in immediate allergic reactions to -lactam antibiotics. Pharmacogenetics 2006 (in press).

105 Guglielmi L, Fontaine C, Gougat-Barbera C, et al: IL-10 promoter and IL4R gene polymorphisms are associated with immediate -lactam allergy in atopic women. Allergy 2006;61: 921–927. 106 Fahrenholz JM: Natural history and clinical features of aspirin-exacerbated respiratory disease. Clin Rev Allergy Immunol 2003;24:113–124. 107 Ventura MT, Muratore L, Calogiuri GF, et al: Allergic and pseudoallergic reactions induced by glucocorticoids: a review. Curr Pharm Des 2003;9: 1956–1964. 108 Coopman SA, Johnson RA, Platt R, Stern RS: Cutaneous disease and drug reactions in HIV infection. N Engl J Med 1993;328:1670–1674. 109 Temesgen Z, Beri G: HIV and drug allergy. Immunol Allergy Clin N Am 2004;24:521–531. 110 Webster AW, Thompson RA: The ampicillin rash. Lymphocyte transformation by ampicillin polymer. Clin Exp Immunol 1974;18:553–564. 111 Renn CN, Straff W, Dorfmuller A, AlMasaoudi T, Merk HF, Sachs B: Amoxicillin-induced exanthema in young adults with infectious mononucleosis: demonstration of drugspecific lymphocyte reactivity. Br J Dermatol 2002;147:1166–1170.

112 Mizukawa Y, Shiohara T: Virus-induced immune dysregulation as a triggering factor for the development of drug rashes and autoimmune diseases: with emphasis on EB virus, human herpesvirus 6 and hepatitis C virus. J Dermatol Sci 2000; 22:169– 180. 113 Moneret-Vautrin DA, Morisset M: Adult food allergy. Curr Allergy Asthma Rep 2005;5:80–85. 114 Backstrom M, Mjorndal T, Dahlqvist R: Under-reporting of serious adverse drug reactions in Sweden. Pharmacoepidemiol Drug Saf 2004;13:483–487. 115 Mittmann N, Knowles SR, Gomez M, et al: Evaluation of the extent of under-reporting of serious adverse drug reactions: the case of toxic epidermal necrolysis. Drug Saf 2004;27:477–487. 116 Messaad D, Sahla H, Benahmed S, et al: Drug provocation tests in patients with a history suggesting an immediate drug hypersensitivity reaction. Ann Intern Med 2004;140:1001–1006.

Dr. Pascal Demoly Exploration des allergies et INSERM Hôpital Arnaud de Villeneuve, University Hospital of Montpellier FR–34295 Montpellier Cédex 5 (France) Tel. +33 4 6733 6127, Fax +33 4 6704 2708 E-Mail [email protected]

Epidemiology and Causes of Drug Hypersensitivity

17

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 18–31

Epidemiology and Causes of Severe Cutaneous Adverse Reactions to Drugs Maja Mockenhaupt Dokumentationszentrum schwerer Hautreaktionen, Department of Dermatology, University Medical Center Freiburg, Freiburg, Germany

Abstract Toxic epidermal necrolysis (TEN), Stevens-Johnson syndrome (SJS), acute generalized exanthematous pustulosis (AGEP) and hypersensitivity syndrome (HSS), which is also called drug reaction with eosinophilia and systemic symptoms (DRESS), are severe cutaneous adverse reactions (SCARs) to drugs. Due to the high morbidity and mortality, they are a threat to the affected patient, the patient’s family and the treating physician. In addition, these reactions have a substantial impact on public health in general including healthcare providers, pharmaceutical companies, and regulatory agencies. In order to quantify the risk of certain drugs to induce these reactions, it is mandatory that they are well defined and separated from other conditions. Furthermore, reliable data sources for drug use of patients with SCARs as well as the general population are needed. For SJS and TEN, large epidemiologic studies have provided information on the incidence and demography, and also on risk factors. Data on demography and etiology have been obtained for AGEP, and a definition for HSS/DRESS has been developed. Copyright © 2007 S. Karger AG, Basel

Introduction

Severe cutaneous adverse reactions (SCARs) that are mainly induced by administration and intake of drugs are Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN), but also acute

generalized exanthematous pustulosis (AGEP) and hypersensitivity syndrome (HSS), nowadays more specifically called drug reaction with eosinophilia and systemic symptoms (DRESS). For several decades mainly case reports and case series were published, whereas more recently epidemiologic studies have been undertaken, making it possible to obtain reliable data on the incidence and demography as well as the etiology of SCARs. The pre-condition for these studies and the analysis of data was a consensus definition of the clinical classification of skin reactions in the spectrum of SJS and TEN. In addition, well-defined diagnostic criteria were required to study AGEP and HSS/DRESS.

Definition of Severe Cutaneous Adverse Reactions

SJS and TEN SJS and TEN are characterized by extensive detachment of the epidermis and erosions of the mucous membranes [1, 2 ; Allanore and Roujeau, pp 267–277]. Histopathology reveals subepidermal separation and necrotic keratinocytes either in wide dissemination or full-thickness necrosis

of the epidermis, which is considered to be the result of extensive apoptosis [2, 3]. Although the histopathology of erythema exsudativum multiforme (EEM) is almost identical with that of SJS and TEN, a consensus definition enables differentiation from EEM majus (EEMM, with mucosal involvement) based on the clinical presentation [4]. This is important for risk estimation of causes, as several studies have suggested that the etiology of EEMM differs from that of SJS and TEN. There is growing evidence that SJS and TEN are a single disease with common causes and mechanisms [5, 6]. The principal difference is the extent of detachment, limited to less than 10% of the body surface area in SJS, and widespread with more than 30% of the body surface area in TEN. Mortality is high and the patients who recover might be left with sequelae and sometimes long-lasting disability. AGEP AGEP is a pustular reaction which has been given a number of different designations in the literature such as toxic pustuloderma, pustular drug rash, pustular psoriasiform eruption with leukocytosis, or which has been classified as pustular psoriasis or another disease [7, 8]. The reaction is characterized by the sudden occurrence of dozens of sterile, non-follicular pinhead-sized pustules arising on an edematous erythema. The rash is commonly accentuated in the main folds. Additional skin symptoms can comprise edema of the face and unspecific lesions such as purpura, target-like lesions, vesicles or blisters [8–10]. Mucous membrane involvement is rare, usually mild, and in general restricted to one site (mostly oral). Cutaneous manifestations are often accompanied by systemic symptoms such as fever and leukocytosis. Histology shows subcorneal and/or intra-epidermal pustules, a sometimes pronounced edema in the papillary dermis and perivascular infiltrates consisting of neutrophils and sometimes some eosinophils [8, 9]. Psoriasiform changes are usually not present.

Epidemiology of SCARs

A very characteristic feature of this skin reaction is its clinical course. Skin symptoms usually arise rapidly (within a few hours) and resolve quickly (within a few days) without treatment. Complications are rare and occur mostly in elderly people or patients in a poor general medical condition [8, 11]. HSS/DRESS/DiHS The term HSS has long been used to summarize numerous severe drug reactions [12, 13]. In recent years efforts have been made to more accurately describe the pattern and identify criteria for a better definition of the disease. In this context the term drug reaction with eosinophilia and systemic symptoms (DRESS) was created as well as the expression drug-induced hypersensitivity syndrome (DiHS) [14, 15; Shiohara et al., pp 251– 266]. Long discussions between experts from different countries revealed that, whatever the denomination, the original syndrome is characterized by a variable combination of drug-induced immunologic background, later onset than other drug reactions, longer duration than common ‘drug rashes’, multi-organ involvement, lymphocyte activation (node enlargement, lymphocytosis and atypical lymphocytes in the circulation), eosinophilia, and frequent virus reactivation [15– 17]. Defining criteria for making the diagnosis of HSS/DRESS turned out to be specifically complicated because, besides its rather variable presentation, it is a diagnosis by exclusion. Main features like rash, fever, and organ involvement can also be attributed to a wide range of other causes such as infections, and to concomitant and preexisting diseases. Therefore, each symptom should always be thoroughly investigated for its relation to the syndrome. Not all symptoms and signs are always recognized, and asymptomatic systemic involvement such as eosinophilia and atypical lymphocytes are often not determined or determined too late, leading to their underestimation. In addition, partly due to the relatively long delay of the reaction after beginning drug use and its long du-

19

ration after withdrawal of the culprit drug, some signs and symptoms disappear before being recognized as drug-related [16].

Description of Data Sources

Large scale epidemiologic data on SCARs were obtained by several studies performed in Europe during the last two decades. After hospital-based retrospective studies over a period of 5 years had been performed in France and Germany in the 1980s [18, 19], a population-based registry on severe skin reactions in Germany was established in 1990 in order to ascertain all hospitalized cases of SJS, TEN and EEMM [20, 21]. In parallel, an international case-control study on SCARs – the so-called SCAR study – was undertaken in France, Germany, Italy, and Portugal between 1989 and 1995, focusing on SJS and TEN [22, 23]. Between 1997 and 2001 the European ongoing case-control surveillance of SCARs – also referred to as EuroSCAR – was conducted in Austria, France, Germany, Israel, Italy, and the Netherlands. Besides SJS and TEN, another type of drug-induced severe skin reaction was studied: AGEP [24, 25]. In 2003 the European registry on SCARs to drugs and collection of biological samples – also called RegiSCAR – was initiated. It operates in the same 6 countries as the EuroSCAR study and includes cases of SJS, TEN, AGEP and HSS/DRESS [26]. German Registry Cases are actively detected in a network of approximately 1,700 hospitals including all departments of dermatology and pediatrics, all burn units, and all departments of internal medicine with intensive care facilities. In addition, every hospital with more than 200 beds is monitored. Each participating department is asked to name a contact person to improve collaboration. Potential cases are reported by phone, fax or e-mail directly to the registry center. Inclusion criteria are

20

checked, and in case they are met, a physician arranges a visit in the treating hospital. In order to not miss cases, all departments receive quarterly letters addressed to the contact person. To each letter a pre-stamped postcard is added. The postcard should be sent back to provide information on unreported potential cases. If the postcard is not sent back, a telephone call is intended. For validation of the coverage rate the amount of returned postcards as well as all further contacts with the individual departments are considered. The registry and the German spontaneous reporting systems for adverse drug reactions exchange case reports on drug-induced cases of severe skin reactions. Since SJS and TEN require hospitalization, only very few patients should be missed because they are not hospitalized or even die before hospital admission due to their severe skin reaction. Thus, the registry is considered to be exhaustive for detection of SJS and TEN cases in Germany. The interview consists of questions regarding the current illness, demographic data, recent and past medical history, recent infections, and history of adverse drug reactions. Detailed information is obtained on medication use. A dermatologic expert committee with no information on exposures reviews all collected cases using clinical data, photographs and histopathology. Based on the consensus definition described before and with the help of a score, cases are classified as ‘definite’, ‘probable’ or ‘possible’ severe skin reactions or are excluded [20, 21]. SCAR and EuroSCAR Studies Information on patients with severe skin reactions is obtained in exactly the same way as described for the German registry. Interviewers are trained together and the same questionnaires and guidelines are used to collect information. Only prospectively ascertained and directly interviewed community cases of SJS and TEN (i.e., patients who developed the adverse reaction

Mockenhaupt

in the community and were admitted because of SCARs) were included in the SCAR and EuroSCAR studies. Cases of SJS and TEN ascertained by the German registry meeting the inclusion criteria for the case-control studies were included in the SCAR and EuroSCAR studies. In the other participating countries a network was set up specifically for these case-control studies. Cases were validated as described above, and controls were also checked for appropriateness of diagnoses and eligibility by a committee, and inappropriate controls were excluded [22–24]. RegiSCAR Study All cases of TEN, SJS, AGEP and DRESS requiring admission to one of the hospitals participating in the network are included in the study. The RegiSCAR study is not limited to so-called community cases of SCARs, but also includes socalled hospital cases (i.e., patients who developed the adverse reaction during hospitalization for another disease). The network of the German registry and the EuroSCAR study is used for the RegiSCAR study as described above. Besides interview of affected patients in the acute stage of the disease, a follow-up examination (interview and blood sample) takes place about 2 months later (8 8 2 weeks). One year after the adverse event a quality of life questionnaire (SF 36) with specific questions related to long-lasting sequelae of SCARs is sent to patients with SJS/TEN and DRESS [26]. Although there are variations in the focus of these studies (e.g., different types of reactions are investigated), the inclusion of cases (e.g., reactions occurring in the community, reactions occurring in hospital), and the methodological approach (e.g., case-control study, case registry, cohort study), they have several important issues in common. These include the careful documentation of clinical data, histopathology and clinical photographs as well as detailed medical history, especially on drug use and infections, by trained investigators (healthcare professionals). Further-

Epidemiology of SCARs

more, all cases of these prospective studies are systematically reviewed by an expert committee using a score to evaluate the diagnosis based on predefined criteria without knowledge of potential causes.

Incidence and Demographic Data

SJS and TEN The retrospective studies performed in the 1980s revealed incidence rates for TEN of 1.2 per 1 million inhabitants per year in France, and 0.93 per 1 million per year in Germany [18, 19]. However, the data were not primarily collected for epidemiologic purposes. In other countries, incidences between 1.4 and 6 per million person-years were reported for SJS and TEN [27–29]. The large variation in the incidence rate may be attributable to a smaller reference population and a difference in the case criteria. The population-based registry in Germany was operating in former West Germany and Berlin until the end of 1995, covering an average population of 65.76 million inhabitants. Based on that population an incidence of 1.53 per 1 million inhabitants per year could be calculated for SJS, SJS/TEN overlap and TEN together in 1991, whereas the overall incidence between 1992 and 1995 was 1.89 for SJS, TEN and their overlap, assuming that these reactions represent a single entity of different severity [20, 21]. Since the beginning of 1996, the new federal states of former East Germany are also included in the registration, now covering an average population of about 82 million inhabitants. Between April 1990 and December 2006, more than 5,000 cases of severe skin reactions were reported to the registry, 2,557 of which were interviewed and reviewed by the dermatologic expert committee. 2,193 were finally included as cases of severe skin reactions; 573 of these cases were diagnosed as EEMM, 778 as SJS, 487 as SJS/TEN overlap, 254 as TEN with maculae and as TEN on

21

large erythema according to the consensus definition. Concerning age distribution, more than 80% of EEMM patients are younger than 40 years in contrast to less than 40% of SJS patients. Approximately 75% of patients with SJS/TEN overlap and TEN are more than 40 years old. More men develop EEMM (almost 70%), whereas for SJS and TEN with maculae the distribution of gender is almost equal (slightly more females), and a female preponderance of around 65% can be observed in SJS/TEN overlap. Virtually no patient with EEMM died. In contrast the mortality is almost 10% for patients with SJS, approximately 30% for patients with SJS/ TEN overlap and almost 50% for patients with TEN. For SJS, SJS/TEN overlap and TEN together the mortality rate is almost 25% [2 ; unpublished updates of the German registry]. In order to evaluate mortality due to SCARs, time of death in relation to the onset of the reaction, age, underlying diseases, and the amount of skin detachment have to be considered. About 5% of the patients with SJS and TEN are HIV-infected. Not surprisingly, the distribution of age and gender differs from that of non-HIVinfected patients with SCARs, whereas the mortality rate and outcome are comparable [30]. AGEP The largest compilation of AGEP cases was obtained by the EuroSCAR study. However, reliable incidence data for AGEP are not available, as the case-control study did not work on a populationbased scale. Nevertheless, it seems that AGEP is more frequent in some European countries than in others, which might be explained by the availability of certain drugs known to induce AGEP. Among 150 potential cases of AGEP, 97 were finally validated. The mean age was 56 (range 4– 91) years. A female predominance could be observed with 80% of the patients being women. The death rate was around 4%. The majority of cases (78/97) derived from France [25].

22

HSS/DRESS Until recently, many publications comprised various adverse reactions under the denomination of HSS [12, 13]. This ‘merging’ and the fact that no epidemiologic studies were performed on the disease may explain why no reliable incidence data are available for HSS/DRESS. For the first time cases of HSS/DRESS were systematically ascertained based on certain inclusion criteria by the RegiSCAR study. To date, 98 potential cases have been ascertained, 38 of which were validated as ‘probable’ or ‘definite’ cases of HSS/DRESS in this ongoing project. Due to strict validation criteria, a substantial number was excluded from the analysis at this stage of the study. Since the major aim of collecting HSS/DRESS was to better characterize the disorder and separate it from other types of cutaneous adverse reactions, this approach seems to be appropriate. Preliminary data from the RegiSCAR study show a mean age of 47.4 (range 3–84) years, a male-female ratio of 19/19, and a mortality rate of 1/38 [unpublished data].

Methodology of Risk Estimation

Case Registries and External Reference Data In order to get an overview on how often a certain drug may be responsible for severe skin reactions, it is not sufficient to look only at the absolute number of cases that have been exposed to that specific drug prior to the onset of the adverse reaction. Furthermore, it is necessary to compare the absolute number of cases to all individuals who have taken the drug in a certain time period, e.g., 1 year. Because the number of people who take a certain drug is not known, prescription data in defined daily doses (DDD) are helpful as a reference for drug use. For this kind of approach it is necessary that the data are population-based, which is the case for the registry on severe skin reactions in Germany as well as for the prescription data available in Germany [31]. By comparing the number of cases of SCARs within 1 year

Mockenhaupt

after the use of a specific drug to the prescription data in DDD of that drug in the same year, drugbased incidences per 1 million DDD can be calculated, assuming a Poisson distribution for rare events [32]. This method for risk evaluation is a simple way to provide a crude estimate of drug risk. When estimating the risk among various drugs, the Hazard function of these drugs should be considered. Therefore, only drugs of the same or a similar indication should be compared. Nevertheless, a number of confounding factors has to be taken into account, e.g., the duration of the availability of a drug on the market, the duration of exposure to the drug, the missing of co-medication in the model, the possible interaction of different drugs, etc. Case-Control Approach of the SCAR and EuroSCAR Study A different method for risk evaluation of drugs is provided by the case-control design. In the socalled SCAR and EuroSCAR studies hospitalized patients with SJS, SJS/TEN overlap and TEN as well as controls were compared in terms of their drug use prior to the onset of the disease. For each case 3 controls were matched on age, gender, region and date of interview, among patients hospitalized for acute conditions including infections (e.g., pneumonia), trauma (e.g., fractures), and abdominal emergencies (e.g., appendicitis, ruptured ovarian cyst, strangulated hernia) that had not occurred as a complication of an underlying chronic disease. Drugs usually taken for a short time period, and those usually given for a longer time period, e.g., months or years, were analyzed separately. Whenever a patient started a drug he was considered exposed only if this drug was still taken in a ‘window’ of 7 days preceding the index day. For drugs with long elimination half-lives the exposure window was extended to 14 (oxicam-NSAIDs, allopurinol) or 21 days (phenobarbital) [22]. Univariate relative risks (RRs) and their 95% confidence intervals (CI) were estimated in terms

Epidemiology of SCARs

of odds ratios with standard methods. When there was no exposed control or case, exact methods were used to estimate 95% CI [33–35]. Multivariate RRs were estimated only when at least 3 cases and controls were exposed to the risk factor considered. Adjustment was performed on country, gender, age, exposure to certain drug groups, and any suspected non-drug confounding factor. In addition, the frequency of concomitant exposure to a ‘highly suspected’ drug, and the delay between beginning drug use and the onset of the adverse reaction were evaluated. 50–60% of case patients taking drugs without significant risk were exposed to at least one ‘highly suspected’ drug in parallel, in contrast to patients using drugs with a high risk. The time pattern of exposure was also very different. For ‘highly suspected’ drugs the beginning of exposure aggregated between 32 and 4 days before the onset of the adverse reaction, when there was no similar aggregation for non-associated drugs (fig. 1, 2) [24].

Etiologic Factors of SCARs

SJS and TEN In addition to various drugs and drug groups, infections, especially viral infections, such as herpes simplex, influenza, viral hepatitis, but also mycoplasma pneumonia infection have been reported to be etiologic factors of EEMM, SJS and TEN [5, 36]. Cases of SJS and TEN without any drug intake seem to occur very rarely. Furthermore, many patients with acute infections before their severe skin reaction also took anti-infective medication. Very often it is hard to determine whether the symptoms, e.g., oronasal soreness and conjunctival injection are signs of an acute infection or the beginning of the severe skin reaction itself. Therefore, it is crucial to determine the index day, i.e., the day of onset of the adverse reaction [20–22]. A possible interaction of infection and medication as well as the interaction of different drugs have not yet been clarified. To

23

Lamotrigine (n = 14) 9 8

Number of cases

7 6 5 4 3 2 1

Fig. 1. Time latency between beginning of drug use and onset of SJS/ TEN. Example of a ‘highly suspected’ drug.

0 1–7

date, there is no reliable in vitro or in vivo test to determine the link between a specific drug and the SCAR in a single case. Even by undertaking a provocation test with the suspected drug, the reaction could be induced again in not more than 1 of 10 patients with previous TEN [37]. Thus, the detection of the culprit drug mainly relies on the time interval between introduction of the drug and the onset of the adverse reaction, which on average has been reported to be 1–3 weeks [1]. With the population-based data of the German registry linked to nationwide prescription data, drug-based incidences were calculated for various NSAIDs. In addition, it was considered whether a certain drug had co-medication with a ‘highly suspected’ drug in the likely etiologic period of 2 weeks before the index day. The analysis revealed overall low risks for most NSAIDs except for those of the oxicam-type, which could be confirmed by data of the SCAR study [38]. As the risk of SJS and TEN is highest in new users of drugs, the use of denominators reflective of all users can lead to low estimates of risk associated with drugs. Instead of using prescription data as a total per year, the numbers of new users can be estimated from the numbers of dispensed prescriptions in Germany, the average prescribed

24

8–14

15–21

22–29 30–37 Time, days

38–45

46–56

>56

doses and the duration of use in Mediplus database (IMS Health), Germany, and assumptions which relate new use to growth in national dispensings. For this analysis data on all hospitalized cases of SJS and TEN with use of various antiepileptic drugs were obtained from the German registry [39]. More than 90% of SJS and TEN cases occurred in the first 63 days of drug use. Over 4 years, the increase in dispensings was 5% for carbamazepine, 65% for lamotrigine, 6% for phenobarbital, –16% for phenytoin, 26% for valproic acid. Across a range of assumptions about the frequency of incident use, the risk estimates vary between 1 and 10 per 10,000 new users for all antiepileptic drugs except for valproic acid for which estimates were consistently lower [39]. In the SCAR study, the crude RR was increased for drugs with a longer period of intake such as carbamazepine, phenobarbital, phenytoin, oxicam-NSAIDs, and allopurinol. For these drugs, the risk seems to be higher during the first 2 months of treatment. Concerning drugs usually taken for a short time, the risk was increased for cotrimoxazole and other anti-infective sulfonamides, aminopenicillins, quinolones, cephalosporins, and chlormezanone [23].

Mockenhaupt

Valproic acid (n = 16) 14

Number of cases

12

Fig. 2. Time latency between beginning of drug use and onset of SJS/ TEN. Example of a non-associated drug. * Cumulative data of all patients who took valproic acid for more than 8 weeks.

10 8 6 4 2 0 1–7

For sulfonamides, e.g., where only 1 control patient had been exposed to the drug, the calculated risk was infinite, and thus can hardly be compared to the risk of other anti-infectives. In that case, the DDD methodology described previously may be used supplementarily. Prescription data are not able to provide detailed information on drug use, but they allow an overall estimate of drug use within a defined population. Thus, they are useful for risk estimation of rare events and rare exposure to certain drugs. The case-control design enables performing a direct comparison between the drug use of case patients and control patients, but for analysis of rare exposure to a certain drug, a large number of control subjects is needed [40]. Since the case definitions of both the German registry of severe skin reaction as well as the international case-control study (SCAR) are based on the consensus definition published by BastujiGarin et al. [4] in 1993, comparison of the data is possible. The EuroSCAR study comprises more recent data on drug risks for SJS and TEN. 379 ‘community’ cases of SJS and TEN (i.e., patients who developed the adverse reaction outside the hospital and who were admitted because of symptoms of

Epidemiology of SCARs

8–14

15–21

22–29 30–37 Time, days

38–45

46–56

>100*

SCARs) and 1,505 controls, all with a determined index day and adequate information on exposures, were accepted. Among medications with prior alerts, two were highly associated with SJS or TEN: nevirapine and lamotrigine. Both shared the overall pattern of ‘highly suspected’ drugs (recent onset of use and infrequent co-medication with another ‘highly suspected’ drug) [24]. To avoid adverse reactions the manufacturer had proposed slow titration of the dosis (lead-in periods) for both agents, but obviously this did not work for severe skin reactions like SJS and TEN [41]. High risk was confirmed for all previously suspected drugs such as anti-infective sulfonamides (especially cotrimoxazole), allopurinol, carbamazepine, phenytoin, phenobarbital, and oxicam-NSAIDs, with the exception of valproic acid (univariate RR = 9.4 [3.9–23], multivariate RR = 2.0 [0.6–7.4]). Most ‘highly suspected’ drugs are usually taken on a long-term basis. Among cases exposed to these drugs, 85–100% had initiated their treatment less than 8 weeks before the reaction (table 1). The median latency time between the start of intake and the index day was less than 4 weeks (15 days for carbamazepine, 24 days for phenyto-

25

Table 1. Estimation of relative risks (RR) for drugs with a high risk of inducing SJS and TEN Drug, duration of use

Recently marketed drugs Nevirapine ≤8 weeks >8 weeks Lamotrigine ≤8 weeks >8 weeks Long-term marketed drugs Co-trimoxazole ≤8 weeks >8 weeks Other anti-infective sulfonamides1 ≤8 weeks >8 weeks Allopurinol ≤8 weeks >8 weeks Carbamazepine ≤8 weeks >8 weeks Phenytoin ≤8 weeks >8 weeks Phenobarbital ≤8 weeks >8 weeks Oxicam-NSAIDs2 ≤8 weeks >8 weeks

Case patients (n = 379)

Control patients (n = 1,505)

Univariate RR (95% CI)

Multivariate RR (95% CI)

Number of cases with use of other ‘highly suspected’ drugs started within 8 weeks

0 0 0 1 (7%) 1 (7%) 0

21 (5.5%) 20 1 14 (3.7%) 14 0

0 0 0 0 0 0

>22 >22 >0.1 >14 >14 n.d.

n.d.

24 (6.3%) 19 5

1 (0.1%) 0 1

102 (14–754) >20 21 (2.3–172)

n.d.

4 (17%) 0 4 (80%)

13 (3.4%) 13 0 66 (17.4%) 56 10 31 (8.2%) 29 2 19 (5.0%) 17 2 20 (5.3%) 17 3 11 (2.9%) 11 0

1 (0.1%) 1 0 28 (1.9%) 1 27 4 (0.3%) 0 4 3 (0.2%) 0 3 5 (0.3%) 1 4 7 (0.5%) 3 4

53 (7.0–410) 53 (7.0–410) n.d. 11 (7.0–18) 261 (36–G) 1.4 (0.7–3.0) 33 (12–95) >32 2.0 (0.4–121) 26 (7.8–90) >17 2.7 (0.4–16) 17 (6.2–45) 71 (9.4–532) 3 (0.7–13) 6.4 (2.5–17) 15 (4.1–54) 0 (0–6.2)

n.d.

0 0 0 7 (11%) 2 (4%) 5 (50%) 1 (3%) 0 1 3 (16%) 2 (12%) 1 3 (15%) 2 (12%) 1 (33%) 1 (9%) 1 (9%) 0

n.d.

18 (11–32) n.d. 0.9 (0.3–2.4) 72 (23–225) n.d. n.d. 17 (4.1–68) n.d. n.d. 16 (5.0–50) n.d. 2.4 (0.2–23) 16 (4.9–52) 50 (12–211) n.d.

1 Includes sulfasalazine (5 cases, 1 control), sulfadiazine (5 cases, 0 controls), sulfadoxine (2 cases, 0 controls), sulfafurazole (2 cases, 0 controls). 2 Includes meloxicam (2 cases, 2 controls), piroxicam (6 cases, 4 controls), tenoxicam (3 cases, 1 control).

in, 17 days for phenobarbital, 20 days for allopurinol), whereas it was much longer for drugs with no associated risk (above 30 weeks for valproic acid, ACE inhibitors and calcium channel blockers). For allopurinol, 56 of 66 exposed patients were recent users in contrast to only 1 of 27 controls. The univariate RR for recent use was 261

26

(36–infinite) versus a multivariate RR of 0.9 (0.3– 2.4) for long-term use. In general, no significant risk persisted after 8 weeks of use. A large number of drugs of common use, such as
-blockers, ACE inhibitors, calcium channel blockers, sulfonamide-related diuretics and sulfonylurea antidiabetics, insulin, and propionic

Mockenhaupt

Table 2. Estimation of relative risks (as OR) for drugs with a high risk of inducing AGEP1 Drug or coalition

Case patients (n = 97)

Control patients (n = 1,009)

Odds ratio

(95% CI)

Pristinamycin Aminopenicillins Quinolones (Hydroxy-)chloroquine Sulfonamides Terbinafine Diltiazem

10 (10%) 18 (19%) 9 (9%) 7 (7%) 4 (4%) 4 (4%) 7 (7%)

0 17 (2%) 5 (0.5%) 2 (0.2%) 0 0 10 (1%)

G 23 33 39 G G 15

26 10 8.5 8.0 7.1 7.1 5.0

1 2

Cases using other ‘highly suspected’ drugs starting within 8 weeks2, % G 54 127 191 G G 48

10 17 33 0 0 25 0

Multivariate odds ratio was used if at least 3 cases and 3 controls were exposed, otherwise univariate. Recent use of other ‘highly suspected’ drugs (i.e. any other drug listed in the table).

acid NSAIDs were not associated with a detectable risk to induce SJS or TEN [24]. AGEP In contrast to many small and one larger case series of more than 60 patients, the EuroSCAR study provided not only information on drugs causing the reaction, but in addition allowed risk estimation based on the case-control approach. Of the 97 patients with AGEP, 13 reported the use of a macrolide antibiotic in the week before onset of the skin reaction. Ten of these cases were associated with pristinamycin, a macrolide that was only marketed in France. Interestingly, the cutaneous reaction occurred after only 1 day of intake in 9 cases, and after 2 days in 1 case. Since no control was exposed to pristinamycin, no multivariate odds ratio (OR) could be provided, but the high lower bound of the univariate CI together with the small percentage of cases with recent concomitant exposure to highly suspected drugs suggest a high risk for pristinamycin (table 2) [25]. Other macrolide antibiotics were also associated with a probably lower but significant risk. For ampicillin and amoxicillin, which are often reported causes for AGEP [9, 10], the multi-

Epidemiology of SCARs

variate OR was 23 (CI 10–54). Exposure time was less than 15 days in all cases and often very short. The results were similar for quinolones. A total of 7 cases versus 2 controls were exposed to chloroquine or hydroxychloroquine leading to a univariate OR of 39 with a lower bound of the 95% CI of 8.0. None of these cases was exposed to another highly suspected drug suggesting that these antimalarials are a risk factor for AGEP. Four patients had taken anti-infective sulfonamides, and 4 were exposed to the antimycotic terbinafine. The lack of exposed controls does not allow calculating an OR but the lower bound of the confidence interval of 7.1 for both drugs suggests a high risk. The class of calcium channel blockers was present in 17 cases and 69 controls (multivariate OR 3.5 [CI 1.7–7.4]). Because there are several prior reports attributing AGEP to diltiazem, this drug was specifically looked at [42]. No other single calcium channel blocker was associated with a significant risk and the multivariate OR for the class excluding diltiazem was 1.9 (CI 0.7–5.3). No patient exposed to diltiazem was simultaneously exposed to another high-risk drug.

27

Two different patterns were observed when looking at the duration of treatment before reaction onset in all patients exposed to a ‘high-risk’ drug. For all exposures to antibiotics (41 cases), including sulfonamides, the median treatment duration was 1 day. In contrast, for all other associated drugs the median was 11 days. Further exploration of data did not provide an explanation to these different delays [25]. HSS/DRESS To date our knowledge about drugs inducing HSS/DRESS derives from case reports and case series. Often, the reaction itself was named after the drug that was considered to be culprit, e.g., allopurinol HSS, dapsone syndrome, anticonvulsant syndrome, etc. One of the manifold difficulties for evaluating drug risks in HSS/DRESS is the fact that the index day is hard to determine. Important first signs of the reaction may not be recognized, and the cutaneous eruption frequently indicating the onset of SJS, TEN and AGEP may occur later in the course of the disease, and may be transient or even absent in some cases. The RegiSCAR study has not yet been analyzed with regard to drug association in cases of HSS/DRESS, and thus risk estimates cannot be provided. However, in case reports and case series aromatic anticonvulsants such as phenytoin, carbamazepine, phenobarbital are most frequently reported as causes of HSS/DRESS. Furthermore, there are numerous reports on HSS/ DRESS after the use of minocycline, allopurinol, thalidomide, sulfonamides and other drugs [12, 14, 43]. It is also discussed that viral reactivation may play an important role in the development of HSS/DRESS, especially in recurrent cases and persisting reactions [15, 17; Shiohara et al., pp 251–266]. Nevertheless, it has not yet been clarified whether the reactivation of HHV-6 and other members of the human herpes virus family are part of the disease itself or should be interpreted as a complication.

28

‘Genetic Epidemiology’ of SCARs

A genetic susceptibility for TEN was suspected many years ago. At that time, different HLA loci were found in patients with TEN caused by antibacterial sulfonamides and TEN associated with oxicam-NSAIDs [44]. More recently a very strong association between carbamazepine-induced SJS in Han Chinese patients and HLA-B*1502 was observed [45; Hung et al., pp 105–114]. This study revealed that the incidence of SJS and TEN is probably much higher in Taiwan than in Europe, and that the majority of SJS and TEN cases in Taiwan is related to carbamazepine (approximately 25%), whereas this antiepileptic agent accounts for no more than 5% of cases in European countries [23, 24, 39]. In the RegiSCAR study 12 carbamazepine-induced cases of SJS and TEN could be identified. Among these only 4 had the HLAB*1502 allele, and, remarkably, these 4 patients were of East-Asian descent. In contrast, the sample of European carbamazepine-induced cases showed a variety of HLA-B alleles, but no specific allele or supertype was found with the complete association described by the Taiwanese group [46]. In a successive study a very strong association between HLA-B*5801 and allopurinol-induced SCARs, both SJS/TEN and HSS/DRESS was found in Chinese patients [47]. This result could only partly be confirmed for European patients with SJS or TEN [48]. These findings are very important because they clearly demonstrate that, first, the genetic predisposition to develop SCARs is highly associated with specific drugs, and, second, ethnicity matters much more than anticipated.

Further Considerations

The RegiSCAR study has established a cohort of patients with SJS and TEN as well as HSS/DRESS. Preliminary analysis revealed that long-lasting

Mockenhaupt

sequelae (e.g., after eye involvement) and overall disability are much more frequent than expected in patients who survived SJS or TEN. In terms of HSS/DRESS, complete recovery after a few weeks was observed, but also the recurrence of the disease over a period of more than 1 year [unpublished data]. The quality of life for patients and the economic impact for them, but also for health care providers and pharmaceutical manufacturers need to be evaluated. It is important that decisions made by regulatory agencies are based on thorough post-marketing surveillance of drugs. Each alert has to be carefully followed, before action is taken. Large scale epidemiologic studies on adverse reactions can help to quantify and minimize the risks of certain medications.

Conclusion

SCARs have to be separated from one another because they show substantial differences in terms of various epidemiologic data, genetic findings, and etiology. Thus, the spectrum of drugs caus-

ing SJS and TEN differs to a large extent from those inducing AGEP. Whereas ‘high-risk’ drugs for SJS/TEN are anti-infective sulfonamides, allopurinol, certain antiepileptic drugs (carbamazepine, lamotrigine, phenobarbital, phenytoin), nevirapine and oxicam-NSAIDs, AGEP is predominantly caused by pristinamycin, aminopenicillins, quinolones, (hydroxy-)chloroquine, sulfonamides, terbinafine and diltiazem. HSS/ DRESS can be induced by a number of drugs known to be related to SJS and TEN, such as aromatic anticonvulsants and allopurinol, but also other medications such as dapsone and minocycline. Further studies in the fields of epidemiology, post-marketing surveillance, risk estimation of drugs, genetic and immunologic investigations, quality of life studies of the affected patients, pharmacoeconomic studies as well as studies on the public health impact of SCARs are needed. Therefore, it is important to continuously watch for various data sources appropriate to investigate SCARs, especially because the drug market is growing rapidly and SCARs are too rare to be detected in preclinical studies.

References 1 Guillaume JC, Roujeau JC, Revuz J, Penso D, Touraine R: The culprit drugs in 87 cases of toxic epidermal necrolysis (Lyell’s syndrome). Arch Dermatol 1987;123:1166–1170. 2 Mockenhaupt M, Norgauer J: Cutaneous adverse drug reactions: StevensJohnson syndrome and toxic epidermal necrolysis. Allergy Clin Immunol Int 2002;14:143–150. 3 Rzany B, Hering O, Mockenhaupt M, Schröder W, Goerttler E, Ring J, Schöpf E: Histopathological and epidemiological characteristics of patients with erythema exsudativum multiforme majus (EEMM), Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN). Br J Dermatol 1996; 135:6–11.

Epidemiology of SCARs

4 Bastuji-Garin S, Rzany B, Stern RS, Shear NH, Naldi L, Roujeau JC: Clinical classification of cases of toxic epidermal necrolysis, Stevens-Johnson syndrome, and erythema multiforme. Arch Dermatol 1993;129:92–96. 5 Auquier-Dunant A, Mockenhaupt M, Naldi L, Correia O, Schröder W, Roujeau JC; for the SCAR Study Group: Correlations between clinical patterns and causes of erythema multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis. Arch Dermatol 2002;138:1019–1024. 6 Schröder W, Mockenhaupt M, Schlingmann J, Schneck B, Hering O, Schöpf E: Clinical re-classification of severe skin reactions and evaluation of their etiology in a population-based registry; in Victor N, Blettner M, Edler L, Haux R, Knaup-Gregori P, Pritsch M, Wahren-

dorf J, Windeler J, Ziegler S (eds): Medical Informatics, Biostatistics and Epidemiology for Efficient Health Care and Medical Research: 44th Annual Conference of the GMDS. Heidelberg, Urban & Vogel, 1999, pp 107–110. 7 Staughton RC, Payne CM, Harper JI, McMichen H: Toxic pustuloderma – a new entity? J R Soc Med 1984;77(suppl 4):6–8. 8 Sidoroff A, Halevy S, Bouwes Bavinck JN, Vaillant L, Roujeau JC: Acute generalized exanthematous pustulosis (AGEP) – a clinical reaction pattern. J Cutan Pathol 2001;28:113–119. 9 Beylot C, Bioulac P, Doutre MS: Pustuleuses exanthématiques aigues généralisées: a propos de 4 cas. Ann Dermatol Venereol 1980;107:37–48.

29

10 Roujeau JC, Bioulac-Sage P, Bourseau C, Guillaume JC, Bernhard P, Lok C, Plantin P, Claudy A, Delavierre C, Vaillant L, Wechsler J, Danan G, Benichou C, Beylot C: Acute generalized exanthematous pustulosis. Analysis of 63 cases. Arch Dermatol 1991; 127:1333– 1338. 11 Brandenburg VM, Kurts C, Eitner F, Hamilton-Williams E, Heintz B: Acute reversible renal failure in acute generalized exanthematous pustulosis. Nephrol Dial Transplant 2002;17:1857– 1858. 12 Shear NH, Spielberg SP: Anticonvulsant hypersensitivity syndrome: in vitro assessment of risk. J Clin Invest 1988;82:1826–1832. 13 Peyrière H, Dereure O, Breton H, Demoly P, Cociglio M, Blayac JP, HillaireBuys D; Network of the French Pharmacovigilance Centers: Variability in the clinical pattern of cutaneous sideeffects of drugs with systemic symptoms: does a DRESS syndrome really exist? Br J Dermatol 2006;155:422–428. 14 Bocquet H, Bagot M, Roujeau JC: Druginduced pseudolymphoma and drug hypersensitivity syndrome (Drug Rash with Eosinophilia and Systemic Symptoms: DRESS). Semin Cutan Med Surg 1996;15:250–257. 15 Seishima A, Yamanaka S, Fujisawa T, Toyama M, Hashimoto K: Reactivation of human herpesvirus (HHV) family members other than HHV-6 in druginduced hypersensitivity syndrome. Br J Dermatol 2006;155:344–349. 16 Kardaun SH, Sidoroff A, Valeyrie-Allanore, Halevy S, Davidovici BB, Mockenhaupt M, Roujeau JC: Variability in the clinical pattern of cutaneous sideeffects of drugs with systemic symptoms: does a DRESS syndrome really exist? Br J Dermatol 2006; in press. 17 Kano Y, Hiraharas K, Sakuma K, Shiohara T: Several herpesviruses can reactivate in a severe drug-induced multiorgan reaction in the same sequential order as in graft-versus-host disease. Br J Dermatol 2006;155:301–306. 18 Schöpf E, Stühmer A, Rzany B, Victor N, Zentgraf R, Kapp JF: Toxic epidermal necrolysis and Stevens-Johnson syndrome. An epidemiologic study from West-Germany. Arch Dermatol 1991;127:839–842. 19 Roujeau JC, Guillaume JC, Fabre JP, Penso D, Fléchet ML, Girre JP: Toxic epidermal necrolysis (Lyell’s syndrome):

30

20

21

22

23

24

25

26

27

incidence and drug etiology in France, 1981–1985. Arch Dermatol 1990;126: 37–42. Rzany B, Mockenhaupt M, Baur S, Schröder W, Stocker U, Mueller J, Holländer N, Bruppacher R, Schöpf E: Epidemiology of erythema exsudativum multiforme majus (EEMM), StevensJohnson syndrome (SJS) and toxic epidermal necrolysis (TEN) in Germany (1990–1992). Structure and results of a population-based registry. J Clin Epidemiol 1996;49:769–773. Mockenhaupt M, Schröder W, Schneck B, Hering O, Schlingmann J, Schöpf E: Populationsbezogene Erfassung von schweren Hautreaktionen in Deutschland. Allergo J 1998;7:381–384. Kelly JP, Auquier A, Rzany B, Naldi L, Bastuji-Garin S, Correia O, Shapiro S, Kaufman D: An international collaborative case-control study of severe cutaneous adverse reactions (SCAR). Design and methods. J Clin Epidemiol 1995;48:1099–1108. Roujeau JC, Kelly JP, Naldi L, Rzany B, Stern RS, Anderson T, Auquier A, Bastuji-Garin S, Correia O, Locati F, Mockenhaupt M, Paoletti C, Shapiro S, Shear N, Schöpf E, Kaufman DW: Medication use and the risk of Stevens-Johnson syndrome or toxic epidermal necrolysis. N Engl J Med 1995;333:1600–1607. Mockenhaupt M, Viboud C, Dunant A, Naldi L, Halevy S, Bouwes Bavinck JN, Sidoroff A, Schneck J, Roujeau JC, Flahault A: Stevens-Johnson syndrome and toxic epidermal necrolysis: assessment of medication risks with emphasis on recently marketed drugs. The EuroSCAR-study. J Invest Dermatol 2006; in press. Sidoroff A, Flahault A, Bouwes Bavinck JN, Halevy S, Sassolas B, Roujeau JC: The spectrum of drugs that may cause acute generalized exanthematous pusulosis (AGEP) is different from the one for Stevens-Johnson syndrome or toxic epidermal necrolysis (TEN). J Invest Dermatol 2003;12:121. Orpha News Europe: The newsletter of the rare disease task force. www.orphanet.net/actor/EuropaNews/2006/ 060124.html. Chan HL, Stern RS, Arndt KA, Langlois J, Jick SS, Jick H, Walker AM: The incidence of erythema multiforme, Stevens-Johnson syndrome, and toxic epidermal necrolysis: a population-based study with particular reference to reac-

28

29

30

31

32

33

34

35

36

37

38

tions caused by drugs among outpatients. Arch Dermatol 1990;126:43–47. Naldi L, Locati F, Marchesi I, Cainelli T: Incidence of toxic epidermal necrolysis in Italy. Arch Dermatol 1990; 126: 1103–1104. Strom BL, Carson JL, Halpern AC, Schinnar R, Snyder ES, Shaw M, Tilson HH, Joseph M, Dai WS, Chen D, Stern RS, Bergmann U, Stolley PD: A population-based study of Stevens-Johnson syndrome: Incidence and antecedent drug exposures. Arch Dermatol 1991; 127:831–838. Fagot JP, Mockenhaupt M, BouwesBavinck JN, Naldi L, Viboud C, Roujeau JC: Nevirapine and the risk of StevensJohnson syndrome or toxic epidermal necrolysis: preliminary results of a case-control study. AIDS 2001;15:1–6. Schwabe U, Paffrath D: Arzneiverordnungs-Report. Stuttgart, Fischer, 1992– 2006. Mockenhaupt M, Rzany B: The epidemiology of serious cutaneous drug reactions; in Williams HC, Strachan DP (eds): The Challenge of Dermatoepidemiology. Boca Raton, CRC Press, 1997, pp 329–342. Miettinen OS: Estimability and estimation in case-referent studies. Am J Epidemiol 1976;103:226–235. Thomas DG: Exact confidence limits for the odds ratio in a 2x2 table. Appl Stat 1971;20:105–110. Slone D, Shapiro S, Miettinen O: Case control surveillance of serious illnesses attributable to ambulatory use of drug; in Colombo F, Slone D, Shapiro S, Tognoni G (eds): Epidemiological Evaluation of Drugs. Amsterdam, Elsevier North Holland Biomedical Press, 1977. Fournier S, Bastuji-Garin S, Mentec H, Revuz J, Roujeau JC: Toxic epidermal necrolysis associated with Mycoplasma pneumoniae infection. Eur J Clin Microbiol Infect Dis 1995;14:558–559. Kauppinen K: Cutaneous reactions to drugs with special reference to severe bullous mucocutaneous eruptions and sulphonamides. Acta Derm Venereol Suppl (Stockh) 1972;68:1–89. Mockenhaupt M, Kelly JP, Kaufman D, Stern RS: The risk of Stevens-Johnson syndrome and toxic epidermal necrolysis associated with non steroidal antiinflammatory drugs: a multinational perspective. J Rheumatol 2003;30: 2234–2240.

Mockenhaupt

39 Mockenhaupt M, Messenheimer J, Schlingmann J, Tennis P: Risk of Stevens-Johnson syndrome and toxic epidermal necrolysis in new users of antiepileptics. Neurology 2005;64: 1134–1138. 40 Stern RS, Steinberg LA: Epidemiology of adverse cutaneous reactions to drugs. Dermatol Clin 1995; 13:681–688. 41 Barreiro P, Soriano V, Casas E, Estrada V, Tellez M, Hoetelmans R, Gonzalez de Requena D, Jimenez-Nacher I, Gonzalez-Lahoz J: Prevention of Nevirapineassociated exanthema using slow dose escalation and/or corticosteroids. AIDS 2000;14:2153–2157. 42 Blodgett TP, Camisa C, Gay D, Bergfeld WF: Acute generalized exanthematous pustulosis secondary to diltiazem therapy. Cutis 1997;60:45–47.

43 Diffee JJ, Hayes JM, Montesi SC, Greene WL, Milinor P: Chlorpropamideinduced pulmonary infiltration and eosinophilia with multisystem toxicity. J Tenn Med Assoc 1986;79:82–84. 44 Roujeau JC, Huynh TN, Bracq C, Guillaume JC, Revuz J, Touraine R: Genetic susceptibility to toxic epidermal necrolysis. Arch Dermatol 1987;123: 1171–1173. 45 Chung WH, Hung SI, Hong HS, Hsih MS, Yang LC, Ho HC, Wu JY, Chen YT: A marker for Stevens-Johnson syndrome. Nature 2004;428:486. 46 Lonjou C, Thomas L, Borot N, Ledger N, de Toma C, LeLouet H, Graf E, Schumacher M, Hovnanian A, Mockenhaupt M, Roujeau JC: A marker for Stevens-Johnson syndrome …: ethnicity matters. Pharmacogenomics J 2006; 6:265–268.

47 Hung SI, Chung WH, Liou LB, Chu CC, Lin M, Huang HP, Lin YL, Lan JL, Yang LC, Hong HS, Chen MJ, Lai PC, Wu MS, Chu CY, Wang KH, Chen CH, Fann CS, Wu JY, Chen YT: HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci USA 2005;102: 4134–4139. 48 Lonjou C, Sekula P, Ledger N, Borot N, de Toma C, Halevy S, Schumacher M, Roujeau JC, Hovnanian A, Mockenhaupt M: A European genetic study of Stevens-Johnson syndrome and toxic epidermal necrolysis: HLA-B association with specific high risk drugs. J Invest Dermatol 2006; 126:37.

PD Dr. Maja Mockenhaupt Dokumentationszentrum schwerer Hautreaktionen Department of Dermatology University Medical Center Freiburg, Hauptstrasse 7 DE–79104 Freiburg (Germany) Tel. +49 761 270 6723, Fax +49 761 270 6834 E-Mail [email protected]

Epidemiology of SCARs

31

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 34 – 46

Chemical-Induced Contact Hypersensitivity in the Mouse Model Stefan F. Martin Allergy Research Group, Department of Dermatology, University Medical Center Freiburg, Freiburg, Germany

Abstract Hypersensitivity reactions to drugs most commonly manifest in skin. Like chemical- or metal ion-induced allergic contact dermatitis and contact hypersensitivity, most acute adverse drug reactions are mediated by cytotoxic T cells and controlled by immunosuppressive regulatory T cells. In both cases, progress has been made in the identification of the recognition structures for T cells and of the effector and regulatory mechanisms of disease, especially with respect to the adaptive immune response. However, a crucial, yet poorly understood event in these hypersensitivities is the activation of the innate immune system and the subsequent induction and polarization of the adaptive immune response. Recent findings in this field using the mouse contact hypersensitivity model will be discussed here. Future challenges are similar for the management of drug- and chemical-induced hypersensitivities. Thus, great efforts are being made to develop reliable in vitro assay systems for predictive risk assessment. The success of this research in assay development is tightly linked to progress in basic research that will not only lead to a more detailed understanding of the pathomechanisms, but also to the development of new treatments for these clinically and economically increasingly important hypersensitivity reactions.

Drug Hypersensitivity and Allergic Contact Dermatitis

The immune system plays an important role in both drug hypersensitivity and allergic contact dermatitis (ACD). Many drugs and chemicals induce potent T-cell responses that often cause immunopathology in the skin [1–4]. Cytotoxic effector T cells that are specifically activated following the recognition of drugs and chemicals are crucial in both reactions, and regulatory T (Treg) cells [5] are also important players for the control of the immune response to these substances [6, 7]. The absence or dysfunction of Treg cells may determine the susceptibility to hypersensitivity reactions due to the loss of immunologic tolerance, and may explain the rather common occurrence of contact dermatitis and drug hypersensitivity.

Copyright © 2007 S. Karger AG, Basel

Innate Immunity and Contact Dermatitis

ACD is a T-cell-mediated inflammatory skin disease [1, 3, 8]. 5–10% of the population suffers from contact eczema. Besides irritant toxic contact dermatitis, ACD poses severe problems to

elicitation

sensitization epidermis

TC

La n g e rh a n s cell

keratinocyte LC Treg

effector T cell

dermis „lethal hit“

TC

dDC • i n d uc ti o n of infla mm ati on • activation of skin cells • DC migration to local lymph nodes • T cell pr imin g an d in ductio n of skin-specific homing receptors

TC

• in ductio n of inflammation • activation of skin cells • production of cytokines and chemokines • recruitment of skin-specific effector/ memory and regulatory T cells

Fig. 1. Crucial role of the innate immune response in CHS. Contact allergens do not only modify cell surfaces or extracellular proteins to generate antigens for T-cell

recognition but also activate skin cells. This innate immune response results in inflammation and is required both for sensitization and elicitation of CHS.

human health, especially as an occupational disease [9]. It has been most widely studied in the mouse model for ACD, the contact hypersensitivity (CHS) model. Inflammation resulting from the potent activation of the innate immune system by contact allergens is thought to be crucial for both sensitization and elicitation of ACD (fig. 1). The induction of inflammatory mediators, the activation of dendritic cell (DC) migration and the recruitment of effector cells such as cytotoxic T cells into the skin depends on the innate immune response [10]. It is therefore assumed that chemical haptens and metal ions have dual effects: they form immunogenic determinants for T cells, and some for B cells as well, by protein modification and they somehow activate the innate immune system, a property that is designated as the irritant effect of contact allergens because it shares

features of irritant contact dermatitis. Thus, many contact allergens show an irritant effect as well. However, until now the molecular mechanisms by which contact allergens activate the innate immune system remain unclear. It is fair to say that chemicals can induce danger signals that are perceived by the innate immune system. Whether this directly or indirectly involves Tolllike receptors (TLR), a family of innate receptors that recognize pathogen-associated molecular patterns (PAMPs) like lipopolysacchide (LPS), lipopeptides or double-stranded RNA, is unclear (fig. 2). TLR are expressed on keratinocytes and DC in the skin [11]. TLR triggering is seen as a starting point of inflammation and crucially contributes to the activation and polarization of the adaptive immune response (fig. 3). The experimental use of TLR ligands has been shown to influence CHS

Chemical-Induced Contact Hypersensitivity in the Mouse Model

35

Tc1 Th1

Tc 0

Ps ers M iz PA sit I n pe se ty act nt co

DC 0

res type pira II P A t pol ory se MPs len n (PA sitize LM rs s)

IFN- ␥, TNF-␣

inflammation allergic contact dermatitis

DC1 IL-12 +++

Tc0

Tc2 Th2

IL-4, IL-5, IL-13

humoral immunity type I allergies

DC2 I L - 1 2 + /Fig. 2. Polarization of immune responses by different danger signals. Pathogens harbor so-called pathogenassociated molecular patterns (PAMPs) that are recognized by innate immune receptors like TLR. This, in most cases, leads to the polarization of DC to DC1 and the sub-

sequent induction of Th1/Tc1 cells. Contact allergens seem to directly or indirectly exert similar effects on DC, but it is unknown whether TLR are involved. Some, like pollen-derived lipid mediators, seem to booster DC2 development.

responses. Thus, intradermal injection of synthetic TLR9 ligands, CpG-ODN, prior to sensitization resulted in enhanced CHS to 2,4-dinitrofluorobenzene (DNFB) [12]. Systemic application of the hydroxylated imiquimod derivative R848, a TLR7 ligand, during sensitization enhanced CHS to fluorescein isothiocyanate (FITC) and DNFB, while it reduced CHS when injected during the challenge process due to leukocyte retention in peripheral organs and enhanced endothelial adhesiveness [13]. Topical imiquimod treatment before UV exposure of the skin prevented the suppression of CHS to DNFB and tolerance induction in mice [14]. This effect was dependent on IL-12 production since protection was abrogated upon treatment of the mice with anti-IL-12p70 antibodies. These TLR-mediated effects are due to efficient activation of DC emi-

gration from skin and induction of inflammation. The downregulation of CHS by probiotics as shown for Lactobacillus casei and its cell walls is intriguing [15]. Although it has not been formally demonstrated that TLR were involved in the immunoregulatory effects, these and other data suggest that TLR may be a potential target to be exploited for the treatment of CHS and other dermatological diseases [11]. Therefore, one must be able to fine tune the TLR trigger so that suppression rather than stimulation of the CHS response occurs. Timing and dosing of application will have to be carefully tested to prevent adverse effects. Inhibitors for TLR will be developed in the future based on the increasing knowledge of TLR signaling and its negative regulation [16].

36

Martin

immunity to infection

TLR

PAMPS pathogen

IL -12

innate immune response => inflammation

DC1

?

Tc1 Th1 IFN- TNF- cytotoxicity

contact allergen

allergic contact dermatitis (ACD)

Fig. 3. Polarization of immune responses by allergens. Many pathogens can drive differentiation of DC to IL-12 secreting and Th1/Tc1 inducing DC1 by PAMPs. Accumulating evidence supports the theory that allergens also polarize immune responses by influencing DC matura-

tion and differentiation. Thus, contact allergens like type I PAMPs may directly or indirectly induce DC1 and Th1/ Tc1 cells whereas respiratory sensitizers or phytoprostanes from pollen, like type II PAMPs, induce DC2 and Th2 cells.

Up to now, a physiological role for TLR in CHS has not been formally demonstrated. Of note, a recent report described a reduction of CHS in mice deficient for NALP3/CIAS1/cryopyrin or the adapter protein ASC. NALP3 and ASC form a caspase-1 activating ‘inflammasome’ that is involved in the LPS/TLR4-driven and ATP-dependent processing of pro-IL-1
[17]. The efficient secretion of the pro-inflammatory cytokines IL-1/
and IL-18 that are important in CHS require caspase-1, as is the migration of Langerhans cells [18]. Our own data have now clearly demonstrated a role for TLR2 and TLR4 in CHS [S.F. Martin et al., submitted]. These and other findings clearly point at a role for TLR in CHS and suggest that infections may affect the susceptibility to and the potential outcome of ACD or CHS.

Innate immune cells such as granulocytes are rapidly recruited to allergen-sensitized skin. The recruitment of neutrophils in 2,4,6-trinitrochlorobenzene (TNCB)- and oxazolone-induced CHS is dependent on mast cell-derived TNF-  and MIP-2 [19] and mast cell-derived TNF- contributes to DC migration [20]. A further role for mast cells has been demonstrated in IgE–/– mice [21]. These animals have decreased CHS responses to oxazolone and DNFB. It was shown that the skin infiltration of inflammatory cells, emigration of DC and the production of inflammatory mediators were impaired. FcR1+ mast cells were activated by IgE in an antigen-independent fashion. Recently,  T cells, B-1 B cells, and NKT cells have been implicated in CHS reactions [22, 23]. Depletion of  T cells before sensitization or the use of  T-cell-deficient mice resulted in in-

Chemical-Induced Contact Hypersensitivity in the Mouse Model

37

creased CHS responses [24]. In CHS to p-phenylendiamine (PPD),  T-cell depletion revealed a positive contribution to CHS induction in adoptive cell transfer experiments [25]. The role of B-1 B cells and NKT cells has now been intensively studied. It was reported that the initiation of CHS requires the rapid activation of IL-4-producing hepatic V14+ invariant NKT cells (i NKT) which support migration of peritoneal B-1 B cells to lymphoid organs and their production of hapten-specific IgM. This IgM activates complement and an inflammatory cascade involving complement-dependent mast cell activation then helps in the activation of endothelial cells and the recruitment of effector T cells [26, 27]. Further evidence for a role of i NKT cells in the initiation of CHS was provided by Nieuwenhuis et al. [28]. CHS to oxazolone was blocked in NKT cell-deficient CD1d–/– or T-cell receptor (TCR) J28–/– mice or upon administration of non-activating CD1d-binding ligands for the invariant TCR of i NKT cells. New data now have revealed a role for activated i NKT cells in promoting DC migration. Thus, injection of -GalCer, a CD1d ligand recognized by i NKT cells, at the time of DNFB sensitization resulted in an increased CHS response [29]. An interesting finding was reported for the rather Th2-dependent CHS to FITC [30]. In this study a role for MHC class II-independent CD4+ T cells was identified. MHC class II and CD8–/– mice still developed CHS to FITC that was, however, diminished. This implies that CD4+ i NKT cells may be effectors in CHS to FITC. A surprising result was obtained when delayedtype hypersensitivity (DTH) and CHS was studied in T- and B-cell-deficient RAG2–/– mice. These mice mounted a DTH reaction in the bladder that was similar to the reaction in wild-type mice concerning the recruitment of inflammatory cells after sensitization. Moreover, cutaneous CHS to oxazolone or DNFB was efficiently induced in RAG2–/– mice and was shown to be haptenspecific and NK cell-dependent [31]. This response

38

was long-lasting as shown by rechallenge after 4 weeks and could be transferred with Thy1+Ly49C/ Ly49I+ NK cells present as a minor population in spleen and liver of sensitized mice. These NK cells may have been primed in skin-draining lymph nodes. It will be interesting to identify the activating NK receptors and the hapten-modified molecules that are recognized by them.

Polarization of Immune Responses by Contact Allergens

Haptens and metal ions can directly activate DC and induce the activation of MAP kinase- and NF-B-dependent signaling pathways [10]. This leads to the production of pro-inflammatory cytokines and chemokines by DC and the upregulation of surface molecules such as costimulatory molecules associated with DC maturation and differentiation. Thus, it has recently been shown that nickel (Ni) and 2,4-dinitrochlorobenzene (DNCB) induce CCR7 expression on human DC [32]. p38 MAPK and JNK were required in both cases. However, Ni induced CCR7 upregulation directly via these pathways while DNCB required TNF-, the secretion of which is regulated by the kinases. A crucial event in the induction of CHS is the polarization of the innate immune response for Th1/Tc1 induction (fig. 2, 3). This already seems to occur in the skin [33]. DC are the most prominent targets of hapteninduced polarization due to their function as naive T-cell activators. Dearman and colleagues [34, 35] have shown a dominant Th1/Tc1-mediated IFN- response induced by the contact sensitizer DNCB, while the respiratory sensitizer trimellitic anhydride (TMA) biased the immune response to IL-4-producing Th2 cells. Together with the recent findings that plant pollen contains phytoprostanes that directly inhibit IL-12 production by DC and promote differentiation to DC2 and type I allergy [36], these studies reveal

Martin

that allergens are not solely recognition structures for T cells. The type of antigen/allergen directly influences the outcome of innate and adaptive immune responses by driving inflammation and polarization. Interestingly, recent data indicate that the type of polarization of the immune response may segregate with the chemical modification of target proteins. Western blot analysis revealed that type I immune responses as elicited by DNCB and DNFB correlated with hapten modification of cellular proteins while the type II immune response inducers FITC, TMA and 2,4-dinitrobenzenesulfonylchloride bound selectively to serum proteins [37]. Induction of CHS with FITC results in a different type of immune polarization as compared to most contact allergens. Thus, it was shown that FITC induces Th2 cells that produce IL-4 and IL10 [38]. Transfer of these cells resulted in the typical delayed cutaneous response following challenge. Interestingly, the early phase of the CHS response could be reproduced in recipient mice following transfer of serum and is probably dependent on IgE. In line with these findings, Stat6–/– mice with a defect in IL-4- and IL-13mediated signal transduction failed to develop CHS to FITC [30]. However, CHS to TNCB, DNFB and oxazolone was also reduced in Stat6–/– mice. Granulocyte infiltration of the skin was also reduced as was the production of IgE and IgG1 [39, 40]. It is for these reasons that FITC-induced CHS is often referred to as an atopic dermatitis-like skin inflammation. These data support the view that Th1/Tc1-dependent CHS responses have a Th2-type immune component.

Costimulation and CHS

DC are the prototypic antigen-presenting cells of the immune system. They are crucial for the activation of naive T cells and costimulatory mol-

Chemical-Induced Contact Hypersensitivity in the Mouse Model

ecules upregulated on activated DC are required. Upregulation of these molecules can be triggered by TLR ligands, but also by contact allergens [3, 10]. Therefore, a number of studies have analyzed the role of costimulation and its blockade in CHS. It was reported that OX40L–/– mice have impaired CHS responses [41] and that treatment of mice with an agonist anti-CD40 antibody during DNFB sensitization resulted in enhanced CHS with increased Langerhans cell activation and migration and increased CD4+ Th1 and CD8+ Tc1 responses that were IL-12 independent [42]. Recent studies have also shown a role for the negative costimulatory CD28-B7 family members PD-1 on T cells and B7-H1 (PD-1L) on DC and T cells in CHS [43]. Anti-PD-1 treatment of mice at sensitization enhanced and prolonged CHS and correlated with an increase in the number of T cells in draining lymph nodes. Similar effects were observed upon anti-B7-H1 treatment [44]. B7-H1 interaction with PD-1 resulted in the suppression of the proliferation and cytokine production of activated T cells. Furthermore, DNFB treatment resulted in upregulation of B7-H1, and treatment of DNFB-sensitized mice with a DC line that expressed high levels of B7-H1 resulted in prevention of CHS upon challenge. These findings imply a successful therapeutic intervention in ACD upon interference with costimulation.

Langerhans Cells and Their Role in CHS

The recent development of transgenic mice in which CD11c+ DC or Langerhans cells can be deleted by administration of diphtheria toxin has allowed the analysis of the role of Langerhans cells in CHS. Different studies have obtained different results. Unaltered [45], diminished [46] or enhanced CHS responses were found [47]. These heterogeneous results imply that Langerhans cells are dispensable for CHS, at least in this artificial situation. Nevertheless, under physiological conditions, Langerhans cells encounter contact

39

allergens in the epidermis and migrate to the draining lymph nodes where they can efficiently induce skin-homing effector T cells [48, 49]. Dermal DC seem to be capable to do the same as suggested by these new studies. The relative contribution of these skin-specific DC subsets to CHS remains to be determined.

Immune Regulation and Tolerance Induction in CHS

Immune responses are controlled by regulatory cells like CD4+CD25+Foxp3+ Treg cells [5]. These cells also play a role in the control of ACD and CHS [7]. Transfer of Foxp3+ Treg cells generated in vitro by infection of CD4+CD25– T cells with a Foxp3 encoding retrovirus significantly suppresses CHS in DNFB-sensitized mice [50]. Depletion of Treg cells with anti-CD25 antibodies prevents oral tolerance induction to DNFB [51] and application of the drug cyclophosphamide increases CHS to TNCB due to Treg depletion [52]. It is, however, not clear yet if the immune regulation takes place in the skin draining lymph nodes and/or the skin. Previous studies have implicated a role for Treg in the inflamed skin and a delayed appearance after a first wave of infiltrating effector cells following elicitation of CHS [53]. The dependence of Treg cells on exogenous IL-2 has recently been demonstrated and studies in the CHS model suggest that CD8+ CHS effector T cells deliver the IL-2 required by CD4+CD25+ Treg cells which control CHS [54]. Similarly, NKT cell help for Treg has been shown [55, 56]. Whether skin-specific Treg subsets exist is not yet clear, but the existence of a high proportion of CLA+CD4+CD25+FOXP3+ cells in human peripheral blood seems to prove that [57– 59]. It has been demonstrated in the CHS model that feeding of DNFB results in a state of oral tolerance that is systemic. CHS can no longer be induced in the tolerized mice [51]. Tolerance induc-

40

tion was ineffective in Invariant Chain (Ii)-deficient mice or impaired in mice treated with anti-CD25 antibodies. The Ii–/– mice developed chronic CHS. Transfer of naive CD4+ T cells into Ii–/– mice prior to hapten feeding restored tolerization. Transfer of CD4+CD25+ T cells into Ii–/– mice completely prevented CHS and treatment of wild-type mice with anti-CD25 antibodies prior to sensitization impaired oral tolerance. Moreover, oral, intravenous or epicutaneous application of low doses of the contact sensitizer TNCB can induce systemic tolerance. The underlying mechanisms of low zone tolerance have now been understood in greater detail. It was observed that CD8+ Tc2 cells are the suppressors of Tc1mediated CHS. IL-10-producing regulatory CD4+ T cells are obligatory for their induction [60]. Feeding of Ni in drinking water also induces a state of oral tolerance which prevents induction of CHS. In this case a role for i NKT cells was demonstrated. These i NKT cells were required for the induction of Treg cells [61]. The mechanism of Treg induction is dependent on the induction of apoptosis of B cells after they have taken up Ni. Ni treatment of mice resulted in the appearance of preapoptotic B cells with decreased bcl-2 and bcl xl expression and upregulation of the proapoptotic protein bax. FasL-expressing i NKT cells induce this B-cell apoptosis in the spleen. These apoptotic B cells may then be taken up by DC which were shown to mediate Ni-specific tolerance [62]. Hapten application of UV-irradiated skin also induces hapten-specific tolerance due to induction of Treg cells [63]. It is associated with DNA damage that can be reduced by IL-12 or IL-18 treatment [64]. UV-induced Treg can suppress CHS upon intravenous transfer prior to sensitization but not prior to challenge. Intradermal Treg injection, however, prevented elicitation as well, suggesting that intravenously injected Treg do no harm to the skin [65]. In this system a role for the C-type lectin dectin-2 was demonstrated [66]. Injection of soluble dectin-2 prevented UV-

Martin

induced CHS suppression or reverted established tolerance. It was speculated that sDec-2 affected Treg cells. Hapten-specific tolerance can also be induced by treatment of mice with the hapten 2,4-dinitrothiocyanobenzene (DNTB) [67]. Skin painting with DNTB prevents CHS to DNFB and this tolerance can be transferred with cells from sensitized mice. However, at this point the nature of the tolerogenic cells is unclear. It was shown that DNTB treatment causes significantly higher levels of apoptosis of DC in skin and lymph nodes than DNFB. The tolerogenic effect of DNTB was abrogated when IL-12 was injected 24 and 3 h before DNTB application or before sensitization with DNFB. Application of IL-12 and DNTB resulted in efficient sensitization since CHS was induced upon DNFB ear challenge without DNFB sensitization indicating that IL-12 converted the tolerogen to an immunogen. An anti-apoptotic role for IL-12 resulting in protection of APC was also demonstrated in this study. Treg cells can also be induced by some drugs as shown for the immunosuppressive agent leflunomide [68]. In this case, both CD4+ and CD8+ T cells with regulatory function were identified and both populations were required for maximal suppression of CHS. Another recent study demonstrated that oral application of the anti-fungal drug itraconazole before hapten challenge of mice with DNFB or oxazolone significantly suppressed CHS [69]. This correlated with a suppression of IFN- production by T cells. An involvement of Treg cells as shown for leflunomide has not been reported. Interestingly, irritant contact dermatitis was not suppressed. These data show a crucial role of CD4+CD25+ Treg cells and other Treg cell populations including i NKT cells in controlling CHS responses and demonstrate their potential for immunotherapeutic tolerization protocols. Thus, the development of strategies for the induction of Treg cells may be a valid approach in the treatment of ACD (fig. 4).

Chemical-Induced Contact Hypersensitivity in the Mouse Model

Effector Cells and Regulatory Cells in CHS to Weak Allergens

The finding that cytotoxic CD8+ T cells predominate in CHS to strong contact sensitizers [3] has now resulted in parallel studies using common allergens classified as weak sensitizers. It has recently been shown that also in the case of weak allergens commonly used in fragrances, CD8+ T cells are the inducers of CHS. CHS to hexylcinnamaldehyde, hydroxycitronellal or eugenol was only induced following extensive depletion of CD4+ T cells [70]. Similar observations were reported for sesquiterpene lactones (SLs) from Arnica montana [C. Lass et al., submitted]. In the latter case, it was shown that SLs have a dominant anti-inflammatory activity in vivo which gives way to pro-inflammatory effects upon extensive CD4 depletion (fig. 4). Interestingly, as shown in vitro, the anti-inflammatory effects are due to inhibition of NF-B activation at high doses, the pro-inflammatory effects correlated with NF-B activation at low doses of SLs. These findings suggest a general predominance of CD8+ Tc1 effector cells in CHS and, moreover, the prevention of CHS to weak contact sensitizers by CD4+ MHC class II-independent cells with regulatory function (fig. 4). It will be interesting to find out if these cells are CD4+ i NKT cells which have been shown to collaborate with Treg cells [56].

Allergen Testing – Future Challenges for Risk Assessment and Therapies

A major problem with ACD, CHS and drug hypersensitivity is the predictive risk assessment. In vivo and in vitro tests have so far allowed to classify contact sensitizers based on their ability to activate DC [71] or cell proliferation in the local lymph node assay [72, 73]. However, we have not yet managed to develop in vitro assays which al-

41

H

NO2

TNCB O2N

O

sesquiterpene lactones

O O

O

NO2

O

Cl

anti-inflammatory effect

n o CH S

CHS T r eg etc.

Teff

strong contact sensitizer

Teff

T r eg etc.

weak contact sensitizer => tolerance

Fig. 4. CHS – a delicate balance between effector and Treg cells. Strong contact sensitizers shift the balance between immunosuppressive Treg cells such as CD4+CD25+Foxp3+ Treg in favor of the effector T cells (Teff) most likely by influencing the polarization of DC.

In contrast, weak allergens usually fail to do that and tolerance is established. Some contact sensitizers like sesquiterpene lactones from Arnica have potent anti-inflammatory effects that prevent efficient activation of the innate immune response.

low the reliable identification of potential irritants and contact allergens or the differentiation between irritants and contact sensitizers. An even more difficult task is the integration of prohaptens into these studies and to make predictions that will have to incorporate issues of skin metabolism. Among the many problems, the assay systems and readout parameters, the type of solvent, concentration and time of application of the test substances have to be evaluated to allow standardization of the assays. Profiling studies using proteomic or genomic approaches are now used to identify signatures for contact allergens. In a recent study using two-dimensional gel electrophoresis and mass spectrometry, out of 22 prominent Ni-binding proteins identified in human B cells, 9 were heat-shock proteins or

chaperonins [74]. This supports the theory that contact allergens are perceived as danger signals that activate innate immune responses. Moreover, expression profiling of dinitrobenzenesulfonic acid (DNBS) treated versus untreated human CD34+ PBMC-derived DC showed 173 and 1,249 differentially expressed genes at a significance of p ! 0.001 for 1 and 5 mM DNBS, respectively [75]. Of note, a number of the genes that were differentially regulated by allergens are involved in DC maturation. These findings support the theory that contact allergens induce DC maturation. In a follow-up study, a set of 60 selected target genes was evaluated by RT-PCR using 5 skin irritants and 11 contact allergens [76]. This study identified 10 genes that reliably differed in their expression between allergen treated and untreated DC and were altered in their

42

Martin

expression by none, 1 or 2 of the irritants tested. These studies aim at the identification of typical signatures for contact allergens with respect to the relevant hapten-modified target proteins and the differentially regulated genes. It is to be expected that there are hapten-specific differences but the hope is to find a sufficient overlap in the profiles of different haptens as a common denominator for all haptens that act as contact sensitizers, or at least for groups of haptens that may have common chemical reaction mechanisms and are sufficiently different from irritants.

Treatment Options in ACD

The gold standard in the treatment of acute and chronic CHS still is the use of corticosteroids. However, application of non-steroidal immunosuppressants like the calcineurin inhibitors pimecrolimus and tacrolimus has shown positive effects in the treatment of ACD and of CHS. While oral administration of tacrolimus inhibited the sensitization phase in mice, pimecrolimus was ineffective. In contrast, both drugs were able to prevent elicitation due to inhibition of lymphocyte activation [77]. Other treatment strategies try to interfere with the crucial innate im-

mune response by dampening inflammation. NF-B and MAP kinase inhibitors, IL-1R antagonists and other drugs targeting inflammatory pathways may prove to be useful. These drugs will also prevent DC migration and maturation as well as the recruitment of effector cells to the skin, other routes to immune pathology, the blockade of which is also under investigation. If we view ACD as the result of a loss of immunologic tolerance (fig. 4), we should try to restore tolerance by Treg induction using our knowledge of oral or low zone tolerance, UV-induced tolerance, Treg-inducing drugs or tolerogenic allergen analogs such as DNTB. In summary, the recent developments in basic research using the CHS model have pointed out potential new strategies to be followed up for the development of better treatments for ACD and other inflammatory skin diseases. Assay systems for the in vitro identification of contact allergens and predictive risk assessment will have to be developed to improve the safety of consumer products and help to further the cheapest and most effective therapy, that is avoidance of allergens. As our understanding of the pathogenesis will increase, we will be able to identify new drug targets. Given the enormous socio-economic impact of contact dermatitis, this goal must be pursued with great intensity.

References 1 Roujeau JC: Clinical heterogeneity of drug hypersensitivity. Toxicology 2005; 209:123–129. 2 Nassif A, Bensussan A, Boumsell L, Deniaud A, Moslehi H, Wolkenstein P, et al: Toxic epidermal necrolysis: effector cells are drug-specific cytotoxic T cells. J Allergy Clin Immunol 2004;114: 1209–1215. 3 Martin SF: T lymphocyte-mediated immune responses to chemical haptens and metal ions: implications for allergic and autoimmune disease. Int Arch Allergy Immunol 2004;134:186–198.

4 Saint-Mezard P, Rosieres A, Krasteva M, Berard F, Dubois B, Kaiserlian D, et al: Allergic contact dermatitis. Eur J Dermatol 2004;14:284–295. 5 Ziegler SF: FOXP3: of mice and men. Annu Rev Immunol 2006;24:209–226. 6 Azukizawa H, Sano S, Kosaka H, Sumikawa Y, Itami S: Prevention of toxic epidermal necrolysis by regulatory T cells. Eur J Immunol 2005;35:1722– 1730.

Chemical-Induced Contact Hypersensitivity in the Mouse Model

7 Cavani A, Ottaviani C, Nasorri F, Sebastiani S, Girolomoni G: Immunoregulation of hapten- and drug-induced immune reactions. Curr Opin Allergy Clin Immunol 2003;3:243–247. 8 Thierse HJ, Gamerdinger K, Junkes C, Guerreiro N, Weltzien HU: T-cell receptor interaction with haptens: metal ions as non-classical haptens. Toxicology 2005;209:101–107. 9 Belsito DV: Occupational contact dermatitis: etiology, prevalence, and resultant impairment/disability. J Am Acad Dermatol 2005;53:303–313.

43

10 Martin SF, Merfort I, Thierse HJ: Interactions of chemicals and metal ions with proteins and role for immune responses. Mini Rev Med Chem 2006;6: 247–255. 11 McInturff JE, Modlin RL, Kim J: The role of toll-like receptors in the pathogenesis and treatment of dermatological disease. J Invest Dermatol 2005; 125: 1–8. 12 Akiba H, Satoh M, Iwatsuki K, Kaiserlian D, Nicolas JF, Kaneko F: CpG immunostimulatory sequences enhance contact hypersensitivity responses in mice. J Invest Dermatol 2004; 123:488– 493. 13 Gunzer M, Riemann H, Basoglu Y, Hillmer A, Weishaupt C, Balkow S, et al: Systemic administration of a TLR7 ligand leads to transient immune incompetence due to peripheral-blood leukocyte depletion. Blood 2005;106: 2424–2432. 14 Thatcher TH, Luzina I, Fishelevich R, Tomai MA, Miller RL, Gaspari AA: Topical imiquimod treatment prevents UV-light-induced loss of contact hypersensitivity and immune tolerance. J Invest Dermatol 2006; 126:821–831. 15 Chapat L, Chemin K, Dubois B, Bourdet-Sicard R, Kaiserlian D: Lactobacillus casei reduces CD8+ T-cell-mediated skin inflammation. Eur J Immunol 2004;34:2520–2528. 16 Liew FY, Xu D, Brint EK, O’Neill LA: Negative regulation of toll-like receptor-mediated immune responses. Nat Rev Immunol 2005;5:446–458. 17 Sutterwala FS, Ogura Y, Szczepanik M, Lara-Tejero M, Lichtenberger GS, Grant EP, et al: Critical role for NALP3/ CIAS1/cryopyrin in innate and adaptive immunity through its regulation of caspase-1. Immunity 2006;24:317–327. 18 Antonopoulos C, Cumberbatch M, Dearman RJ, Daniel RJ, Kimber I, Groves RW: Functional caspase-1 is required for Langerhans cell migration and optimal contact sensitization in mice. J Immunol 2001;166:3672–3677. 19 Biedermann T, Kneilling M, Mailhammer R, Maier K, Sander CA, Kollias G, et al: Mast cells control neutrophil recruitment during T-cell-mediated delayed-type hypersensitivity reactions through tumor necrosis factor and macrophage inflammatory protein 2. J Exp Med 2000;192:1441–1452.

44

20 Suto H, Nakae S, Kakurai M, Sedgwick JD, Tsai M, Galli SJ: Mast cell-associated TNF promotes dendritic cell migration. J Immunol 2006;176:4102– 4112. 21 Bryce PJ, Miller ML, Miyajima I, Tsai M, Galli SJ, Oettgen HC: Immune sensitization in the skin is enhanced by antigen-independent effects of IgE. Immunity 2004;20:381–392. 22 Askenase PW: Yes T cells, but three different T cells (
,  and NK T cells), and also B-1 cells mediate contact sensitivity. Clin Exp Immunol 2001;125:345–350. 23 Askenase PW, Szczepanik M, Itakura A, Kiener C, Campos RA: Extravascular T-cell recruitment requires initiation begun by V14+ NKT cells and B1 B cells. Trends Immunol 2004;25: 441–449. 24 Guan H, Zu G, Slater M, Elmets C, Xu H:  T cells regulate the development of hapten-specific CD8+ effector T cells in contact hypersensitivity responses. J Invest Dermatol 2002; 119:137–142. 25 Yokozeki H, Watanabe K, Igawa K, Miyazaki Y, Katayama I, Nishioka K:  T cells assist 
T cells in the adoptive transfer of contact hypersensitivity to p-phenylenediamine. Clin Exp Immunol 2001;125:351–359. 26 Askenase PW, Itakura A, Leite-deMoraes MC, Lisbonne M, Roongapinun S, Goldstein DR, et al: TLR-dependent IL-4 production by invariant V14+J18+ NKT cells to initiate contact sensitivity in vivo. J Immunol 2005;175:6390–6401. 27 Campos RA, Szczepanik M, Itakura A, Lisbonne M, Dey N, Leite-de-Moraes MC, et al: Interleukin-4-dependent innate collaboration between iNKT cells and B-1 B cells controls adaptative contact sensitivity. Immunology 2006; 117:536–547. 28 Nieuwenhuis EE, Gillessen S, Scheper RJ, Exley MA, Taniguchi M, Balk SP, et al: CD1d and CD1d-restricted iNKTcells play a pivotal role in contact hypersensitivity. Exp Dermatol 2005; 14: 250–258. 29 Gorbachev AV, Fairchild RL: Activated NKT cells increase dendritic cell migration and enhance CD8+ T-cell responses in the skin. Eur J Immunol 2006;36:2494–4503 (DOI 10.1002/ eji.200636075).

30 Takeshita K, Yamasaki T, Akira S, Gantner F, Bacon KB: Essential role of MHC II-independent CD4+ T cells, IL4 and STAT6 in contact hypersensitivity induced by fluorescein isothiocyanate in the mouse. Int Immunol 2004; 16:685–695. 31 O’Leary JG, Goodarzi M, Drayton DL, von Andrian UH: T-cell- and B-cellindependent adaptive immunity mediated by natural killer cells. Nat Immunol 2006;7:507–516. 32 Boisleve F, Kerdine-Romer S, RougierLarzat N, Pallardy M: Nickel and DNCB induce CCR7 expression on human dendritic cells through different signalling pathways: role of TNF-  and MAPK. J Invest Dermatol 2004; 123: 494–502. 33 Bellinghausen I, Brand U, Enk AH, Knop J, Saloga J: Signals involved in the early TH1/TH2 polarization of an immune response depending on the type of antigen. J Allergy Clin Immunol 1999;103:298–306. 34 Dearman RJ, Humphreys N, Skinner RA, Kimber I: Allergen-induced cytokine phenotypes in mice: role of CD4 and CD8 T-cell populations. Clin Exp Allergy 2005;35:498–505. 35 Kimber I, Dearman RJ: What makes a chemical a respiratory sensitizer? Curr Opin Allergy Clin Immunol 200; 5:119– 124. 36 Traidl-Hoffmann C, Mariani V, Hochrein H, Karg K, Wagner H, Ring J, et al: Pollen-associated phytoprostanes inhibit dendritic cell interleukin-12 production and augment T-helper type 2 cell polarization. J Exp Med 2005;201: 627–636. 37 Hopkins JE, Naisbitt DJ, Kitteringham NR, Dearman RJ, Kimber I, Park BK: Selective haptenation of cellular or extracellular protein by chemical allergens: association with cytokine polarization. Chem Res Toxicol 2005;18: 375–381. 38 Dearman RJ, Kimber I: Role of CD4+ T-helper 2-type cells in cutaneous inflammatory responses induced by fluorescein isothiocyanate. Immunology 2000;101:442–451. 39 Takeshita K, Yamasaki T, Nagao K, Sugimoto H, Shichijo M, Gantner F, et al: CRTH2 is a prominent effector in contact hypersensitivity-induced neutrophil inflammation. Int Immunol 2004;16:947–959.

Martin

40 Yokozeki H, Ghoreishi M, Takagawa S, Takayama K, Satoh T, Katayama I, et al: Signal transducer and activator of transcription 6 is essential in the induction of contact hypersensitivity. J Exp Med 2000;191:995–1004. 41 Chen AI, McAdam AJ, Buhlmann JE, Scott S, Lupher ML Jr, Greenfield EA, et al: Ox40 ligand has a critical costimulatory role in dendritic cell: T-cell interactions. Immunity 1999;11:689–698. 42 Gorbachev AV, Fairchild RL: CD40 engagement enhances antigen-presenting Langerhans cell priming of IFN--producing CD4+ and CD8+ T cells independently of IL-12. J Immunol 2004; 173:2443–2452. 43 Kim HK, Guan H, Zu G, Li H, Wu L, Feng X, et al: High-level expression of B7-H1 molecules by dendritic cells suppresses the function of activated T cells and desensitizes allergen-primed animals. J Leukoc Biol 2006;79:686–695. 44 Tsushima F, Iwai H, Otsuki N, Abe M, Hirose S, Yamazaki T, et al: Preferential contribution of B7-H1 to programmed death-1-mediated regulation of hapten-specific allergic inflammatory responses. Eur J Immunol 2003; 33: 2773–2782. 45 Kissenpfennig A, Henri S, Dubois B, Laplace-Builhe C, Perrin P, Romani N, et al: Dynamics and function of Langerhans cells in vivo: dermal dendritic cells colonize lymph node areas distinct from slower migrating Langerhans cells. Immunity 2005;22:643– 654. 46 Bennett CL, van Rijn E, Jung S, Inaba K, Steinman RM, Kapsenberg ML, et al: Inducible ablation of mouse Langerhans cells diminishes but fails to abrogate contact hypersensitivity. J Cell Biol 2005;169:569–576. 47 Kaplan DH, Jenison MC, Saeland S, Shlomchik WD, Shlomchik MJ: Epidermal Langerhans cell-deficient mice develop enhanced contact hypersensitivity. Immunity 2005;23:611–620. 48 Dudda JC, Simon JC, Martin S: Dendritic cell immunization route determines CD8+ T-cell trafficking to inflamed skin: role for tissue microenvironment and dendritic cells in establishment of T-cell-homing subsets. J Immunol 2004;172:857–863.

49 Dudda JC, Lembo A, Bachtanian E, Huehn J, Siewert C, Hamann A, et al: Dendritic cells govern induction and reprogramming of polarized tissueselective homing receptor patterns of T cells: important roles for soluble factors and tissue microenvironments. Eur J Immunol 2005;35:1056–1065. 50 Loser K, Hansen W, Apelt J, Balkow S, Buer J, Beissert S: In vitro-generated regulatory T cells induced by Foxp3retrovirus infection control murine contact allergy and systemic autoimmunity. Gene Ther 2005;12:1294–1304. 51 Dubois B, Chapat L, Goubier A, Papiernik M, Nicolas JF, Kaiserlian D: Innate CD4+CD25+ regulatory T cells are required for oral tolerance and inhibition of CD8+ T cells mediating skin inflammation. Blood 2003;102:3295– 3301. 52 Ikezawa Y, Nakazawa M, Tamura C, Takahashi K, Minami M, Ikezawa Z: Cyclophosphamide decreases the number, percentage and the function of CD25+ CD4+ regulatory T cells, which suppress induction of contact hypersensitivity. J Dermatol Sci 2005;39:105– 112. 53 Okazaki F, Kanzaki H, Fujii K, Arata J, Akiba H, Tsujii K, et al: Initial recruitment of interferon--producing CD8+ effector cells, followed by infiltration of CD4+ cells in 2,4,6-trinitro-1-chlorobenzene-induced murine contact hypersensitivity reactions. J Dermatol 2002;29:699–708. 54 Kish DD, Gorbachev AV, Fairchild RL: CD8+ T cells produce IL-2, which is required for CD4+CD25+ T-cell regulation of effector CD8+ T-cell development for contact hypersensitivity responses. J Leukoc Biol 2005;78: 725–735. 55 Jiang S, Game DS, Davies D, Lombardi G, Lechler RI: Activated CD1d-restricted natural killer T cells secrete IL-2: innate help for CD4+CD25+ regulatory T cells? Eur J Immunol 2005;35:1193– 1200. 56 La Cava A, Van Kaer L, Fu DS: CD4+CD25+ Tregs and NKT cells: regulators regulating regulators. Trends Immunol 2006;27:322–327. 57 Rensing-Ehl A, Gaus B, Bruckner-Tuderman L, Martin SF: Frequency, function and CLA expression of CD4+CD25+FOXP3+ regulatory T cells in bullous pemphigoid. Exp Dermatol 2007;16:13–21.

Chemical-Induced Contact Hypersensitivity in the Mouse Model

58 Huehn J, Hamann A: Homing to suppress: address codes for Treg migration. Trends Immunol 2005;26:632– 636. 59 Wei S, Kryczek I, Zou W: Regulatory T-cell compartmentalization and trafficking. Blood 2006;108:426–431. 60 Seidel-Guyenot W, Perschon S, Dechant N, Alt R, Knop J, Steinbrink K: Low zone tolerance induced by systemic application of allergens inhibits Tc1mediated skin inflammation. J Allergy Clin Immunol 2006;117:1170–1177. 61 Roelofs-Haarhuis K, Wu X, Gleichmann E: Oral tolerance to nickel requires CD4+ invariant NKT cells for the infectious spread of tolerance and the induction of specific regulatory T cells. J Immunol 2004;173:1043–1050. 62 Nowak M, Kopp F, Roelofs-Haarhuis K, Wu X, Gleichmann E: Oral nickel tolerance: Fas ligand-expressing invariant NK T cells promote tolerance induction by eliciting apoptotic death of antigencarrying, effete B cells. J Immunol 2006;176:4581–4589. 63 Schwarz T: Mechanisms of UV-induced immunosuppression. Keio J Med 2005; 54:165–171. 64 Schwarz A, Maeda A, Stander S, van Steeg H, Schwarz T: IL-18 reduces ultraviolet radiation-induced DNA damage and thereby affects photoimmunosuppression. J Immunol 2006; 176: 2896–2901. 65 Schwarz A, Maeda A, Wild MK, Kernebeck K, Gross N, Aragane Y, et al: Ultraviolet radiation-induced regulatory T cells not only inhibit the induction but can suppress the effector phase of contact hypersensitivity. J Immunol 2004;172:1036–1043. 66 Aragane Y, Maeda A, Schwarz A, Tezuka T, Ariizumi K, Schwarz T: Involvement of dectin-2 in ultraviolet radiation-induced tolerance. J Immunol 2003;171:3801–3807. 67 Riemann H, Loser K, Beissert S, Fujita M, Schwarz A, Schwarz T, et al: IL-12 breaks dinitrothiocyanobenzene-mediated tolerance and converts the tolerogen DNTB into an immunogen. J Immunol 2005;175:5866–5874. 68 Weigmann B, Jarman ER, Sudowe S, Bros M, Knop J, Reske-Kunz AB: Induction of regulatory T cells by leflunomide in a murine model of contact allergen sensitivity. J Invest Dermatol 2006;126:1524–1533.

45

69 Ausaneya U, Kawada A, Aragane Y: Itraconazole suppresses an elicitation phase of a contact hypersensitivity reaction. J Invest Dermatol 2006; 126: 1028–1035. 70 Vocanson M, Hennino A, Cluzel-Tailhardat M, Saint-Mezard P, Benetiere J, Chavagnac C, et al: CD8+ T cells are effector cells of contact dermatitis to common skin allergens in mice. J Invest Dermatol 2006; 126:815–820. 71 Casati S, Aeby P, Basketter DA, Cavani A, Gennari A, Gerberick GF, et al: Dendritic cells as a tool for the predictive identification of skin sensitisation hazard. Altern Lab Anim 2005;33:47–62.

72 Basketter DA, Clapp C, Jefferies D, Safford B, Ryan CA, Gerberick F, et al: Predictive identification of human skin sensitization thresholds. Contact Dermatitis 2005;53:260–267. 73 Kimber I, Dearman RJ, Betts CJ, Gerberick GF, Ryan CA, Kern PS, et al: The local lymph node assay and skin sensitization: a cut-down screen to reduce animal requirements? Contact Dermatitis 2006;54:181–185. 74 Heiss K, Junkes C, Guerreiro N, Swamy M, Camacho-Carvajal MM, Schamel WW, et al: Subproteomic analysis of metal-interacting proteins in human B cells. Proteomics 2005;5:3614–3622. 75 Ryan CA, Gildea LA, Hulette BC, Dearman RJ, Kimber I, Gerberick GF: Gene expression changes in peripheral blood-derived dendritic cells following exposure to a contact allergen. Toxicol Lett 2004;150:301–316.

76 Gildea LA, Ryan CA, Foertsch LM, Kennedy JM, Dearman RJ, Kimber I, et al: Identification of gene expression changes induced by chemical allergens in dendritic cells: opportunities for skin sensitization testing. J Invest Dermatol 2006;126:1813–1822. 77 Bavandi A, Fahrngruber H, Aschauer H, Hartmann B, Meingassner JG, Kalthoff FS: Pimecrolimus and tacrolimus differ in their inhibition of lymphocyte activation during the sensitization phase of contact hypersensitivity. J Dermatol Sci 2006;43:117–126.

Prof. Dr. Stefan F. Martin Allergy Research Group, Department of Dermatology University Medical Center Freiburg Hauptstrasse 7, DE–79104 Freiburg (Germany) Tel. +49 761 270 6738, Fax +49 761 270 6655 E-Mail [email protected]

46

Martin

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 47–54

Lessons from Nickel Hypersensitivity: Structural Findings Hans Ulrich Weltzien Max-Planck-Institut für Immunbiologie, Freiburg, Germany

Abstract Nickel (Ni) represents the most common contact sensitizer in the industrialized world, but its suitability as a general model for contact hypersensitivity remains a matter of dispute. This relates to the fact that unlike chemically reactive hapten allergens, Ni engages in multivalent, but reversible coordinative complexes with proteins. The present contribution summarizes knowledge on the structural basis of Ni recognition by T cells and discusses its relevance and impact on hapten and drug hypersensitivity. Available data indicate that for some T cells, Ni ions may indeed be presented similar to classical haptens, involving coordination bonds to MHC-associated peptides. However, other clones indicate Ni-facilitated cross-linking of the T-cell antigen receptor (TCR) with MHC independent of associated peptides, increasing significantly the chances for TCR-activating matches. A model deduced from these findings suggests that Ni ions may activate T cells by stabilizing low-affinity contacts between TCR and restricting MHC, if those two structures supply sufficient coordination sites. We suggest that similar non-covalent interactions may lead to T-cell activation in hypersensitivity reactions to chemically inert drugs. In both situations the model requires as yet unidentified carrier molecules to concentrate the allergens close to the TCR-MHC interphase. Copyright © 2007 S. Karger AG, Basel

Drugs, Haptens and Their Recognition by T Cells

The understanding of hypersensitivity reactions to drugs is closely related to our knowledge of immunological reactivity to haptens. This dates back to Landsteiner’s finding that haptens via covalent reaction with proteins may produce artificial antigenic determinants, and that these may be the basis for contact hypersensitivity (CHS) to those chemicals in animals or humans [1]. Ever since, CHS reactions against model haptens such as di- or tri-nitrochlorobenzene (DNCB or TNCB), oxazolone, and others provided experimental tools to study T-cell-mediated responses against haptens [2]. Some 30 years ago, Shearer et al. [3] first described that not only immunoglobulins but also T cells reacted specifically to haptens and that this reactivity was clearly MHC restricted. Later, after MHC restriction had been related to MHC-associated peptides [4, 5], we demonstrated that T-cell immunogenic hapten determinants were formed by modification of MHC-binding peptides [6] and that such ‘haptenated’ peptides were capable of inducing hapten-specific CHS in mice [7]. The relevance of

L

TCR

L

L

β




Ni

Hapten

L

L

Peptide

HLA-DR L

Fig. 1. The ‘classical’ hapten model. Based on the findings of Ortmann et al. [6] , haptens are recognized by T cells in the form of covalently modified, MHC-associated peptides.

Fig. 2. Ni coordination complexes. Ni2+ ions may form square planar or octaedric coordination complexes with four or six electron-donating ligands. Ligands may be represented by nitrogen- or oxygen-containing amino acid side chains or even by water molecules. The most preferred amino acid ligand is histidine.

these findings for human drug hypersensitivity was revealed by human, HLA-restricted, penicillin-specific T-cell clones reacting to HLA-binding penicilloyl peptides [8]. Hapten-modified MHC-associating peptides therefore clearly represent a major way of presenting chemically reactive haptens and drugs to T cells (see figure 1 for schematic representation). More recently, however, increasing evidence was found which indicates that this may be only one of several possibilities to trigger hapten-reactive antigen receptors (TCR) on T cells [9].

can form quite stable, though reversible, coordination complexes with proteins, involving 4–6 electron-donating amino acid side chains or water molecules as ligands (fig. 2). These complexes are geometrically highly defined, resembling either a planar square with 4 ligands or a bi-pyramidal arrangement with 6 ligands. Protein-Ni complexes have been best analyzed with model peptides or with serum albumin, a known carrier molecule for Ni or Cu ions in most mammals [11]. Early studies by Sinigaglia et al. [12] and others [13] had shown that the peripheral blood of Ni-allergic patients harbored MHC-restricted T cells with apparent specificity for Ni2+. A first indirect indication that Ni2+ actually might be part of the antigenic determinant for such clones came from an experiment of Romagnoli et al. [14]. They studied a malaria-specific murine T cell, reactive to two peptides, one with, the other without histidine in its sequence. Reactivity to the His peptide, but not to the other one, was inhibitable by Ni2+, suggesting a direct binding of Ni2+ to MHC-associated peptides.

Nickel Allergy

One clue to this latter assumption stems from studies of nickel (Ni)-reactive T cells isolated from patients suffering from CHS to Ni, representing, in fact, the most common form of human contact allergy in the western world [10]. Metal ions do not engage in covalent bonds with proteins and, thus, according to Landsteiner’s definition, do not fulfill the criteria of ‘typical’ haptens. However, Ni2+ among other metal ions

48

Weltzien

Nickel Ions as Haptens?

The existence of Ni-specific CD4 and CD8 T cells in peripheral blood and skin lesions of allergic individuals was confirmed in several laboratories [10, 15], but the structural identification of the allergenic epitopes proved extremely difficult. It was not even clear whether metal ions actually formed part of these determinants [16]. One reason for these difficulties resides in the multivalency of Ni2+ interactions. In some cases, such as human serum albumin (HSA) [11] or bacterial urease [17], a single protein chain may provide enough ligands in appropriate geometrical locations to allow for stable coordination complexes. In many other situations, however, one has to assume that Ni2+ finds only one or few fitting ligands within a single protein chain, requiring a second one (identical or different) to saturate the remaining coordination sites. While this indicates a cross-linking potential for Ni2+, it also implies that in most cases, including the surface of MHC-peptide complexes on antigen-presenting cells (APC), a single protein or peptide will not provide enough ligands for stable binding of Ni2+. Unlike for TNCB, where any accessible lysine will suffice for covalent hapten modification of MHC-associating peptides [18], no Ni2+-binding peptide with adequate affinity for an MHCbinding groove has yet been identified. Consequently, the molecular scenario for Ni2+-dependent T-cell activation involves a multitude of possibilities. The situation simulating the ‘classical’ hapten type, i.e. presentation of a Ni-peptide complex on MHC, if it exists at all, will be a rare exception. On the other hand, data have been published indicating that Ni3+ and other metal ions may affect the processing of selfproteins, resulting in MHC presentation of cryptic and hence immunogenic self-peptides [19]. In that case the apparent ‘Ni-determinant’ would actually not contain any Ni ions. Other possibilities include binding of Ni2+ to a combination of amino acids on MHC and as-

Nickel Hypersensitivity

sociated peptides or on MHC alone. If in these cases the MHC/peptide surface provides enough coordination sites to firmly bind Ni2+, one would expect that APCs could be pulsed with Ni2+, and even after washing would activate corresponding T-cell clones. The complementary TCR would have to supply a fitting geometry, including ligands to saturate free Ni coordination sites. In fact, T-cell clones exhibiting the required characteristics have been isolated and one of them (clone ANi-2.3, restricted to HLA-DR52c, i.e. DRB3*03) has been extensively analyzed [20]. Clone ANi-2.3 was activated by Ni2+-pulsed and aldehyde-fixed APC indicating a determinant produced by coordinative binding of Ni2+ to the APC surface in the absence of cellular processing [21]. This binding depended on the presence first of histidine in position 81 on the surface of the HLA-DR -chain and secondly of a particular, so far not identified, type of B-cell-derived peptide in the MHC groove. These data were interpreted as ‘... representing a general model for Ni2+ recognition in which His81 and two amino acids from the NH2-terminal part of the MHC bound peptide coordinate Ni2+ ...’ [20].

Different TCR Recognize Distinct Types of Ni2+ -Induced Epitopes

Clones like ANi-2.3, though repeatedly isolated [21], are far from representing the majority of Nispecific T cells. Thus, at least 50% of all Ni-reactive clones do not react to pulsed APC, i.e. they require the constant presence of Ni2+ in the medium [22]. Another major proportion of clones requires actively processing APC. Still others exhibit promiscuous HLA restriction to various degrees [21] or are even not restricted by any type of MHC-like protein [23]. The Ni-induced determinants responsible for triggering the TCR of these different types of T cells must therefore exhibit considerable structural variation. In all cases, however, gene transfection experiments have

49

a


b

TCR




TCR

β Ni

Fig. 3. Two models for TCR activation by Ni2+. a Hapten-like binding of Ni2+ to MHC and peptide as shown for the Ni-reactive human T-cell clone ANi-2.3. b Peptide-independent cross-linking of TCR and MHC as exhibited by the T-cell clone SE9.

shown the antigen specificity to be mediated by
TCR, the variability of determinants being mirrored by TCR polyclonality. Nevertheless, Nispecific T-cell responses in peripheral blood [24] or skin lesions [25] of particularly strong sensitized individuals showed in their TCR a preponderance of the V17 element. Within the CDR3 regions of those -chains, a frequent Arg-Asp motif was detected (including the above-mentioned clone ANi-2.3) and shown by mutation to be essential for Ni recognition. However, for functionality and specificity these -chains required pairing with idiotypically defined
chains, excluding a superantigen-like reactivity of Ni2+ with the V17 sequence [26].

A Cross-Linking Model for Nickel Recognition

As mentioned above, Ni2+ is expected to crosslink proteins if they provide appropriate ligand side chains. For one particular clone this model has been experimentally verified. Clone SE9 and the corresponding murine hybridoma line T913 carrying the transfected SE9 TCR were employed in these studies [21, 27]. SE9 expressed V
22 and V17, reacted to Ni2+ on fixed, but in contrast to ANi-2.3 not on Ni-pulsed APC, and was inhibited by antibodies to HLA-DR, but not to DP or DQ [27]. Also in contrast to ANi-2.3, which was

50

HLA-DR

β Ni

HLA-DR

selectively restricted to the DRB3*03 allele, clone SE9 was activated by Ni2+ in the presence of numerous different HLA-DR alleles. Therefore, both clones could be stimulated in the presence of identical APC. The fact that ANi-2.3 but not SE9 reacted to those APC when pulsed with Ni2+ clearly indicated that both clones reacted to different epitopes and that pulsing of APC cannot produce the one for SE9. Luckily, the Ni specificity of SE9 turned out to be almost exclusively defined by the TCR
chain, and mutations of TCR
and DR revealed three of the six possible coordination sites for Ni2+. As for ANi-2.3, His81 in the DR chain was absolutely essential for Ni presentation, but peptides in the MHC groove could be exchanged ad libitum [27]. Thus, two models for (a) hapten-like presentation of Ni2+ to clones such as ANi-2.3 (fig. 3a), and (b) peptide-independent cross-linking of MHC and TCR as in the case of clone SE9 (fig. 3b) have been developed. The independence of SE9 activation of the nature of DR-associated peptides has two important implications. On the one hand, it results in a reduction of the number of coordination sites for Ni2+ on the APC, explaining the failure to pulse APC wit Ni2+ for clone SE9. On the other hand, it tremendously increases the chances for ‘fitting’ Ni2+ coordination complexes since every HLADR molecule on the APC is a potential candidate.

Weltzien

In vivo, Ni-reactive clones sharing this property are therefore likely to react more sensitively to Ni2+ than those requiring the presence of defined MHC-associated peptides. In the TCR
-chain of clone SE9, mutational analyses revealed two amino acids in positions 29 (CDR1) and 94 (CDR3) as essential for Ni recognition, both of them tyrosine [27]. Tyrosine is not usually regarded as a typical ligand for Ni2+, but functional replacement of Tyr29 by histidine strongly argued for their role as coordinative ligands in this TCR [27]. Gamerdinger and Weltzien [unpubl. data] have further shown that also SE9 clone cells could not be pulsed with Ni2+. Thus, neither APC nor T cells are capable of fixing Ni2+ in appropriate locations to facilitate eventual activation of the SE9 TCR. A model to explain HLA-DR-restricted activation of the SE9 TCR therefore had to postulate a close proximity of TCR and MHC prior to the insertion of Ni2+, because only the combination of both structures could provide sufficient coordination sites [27, 28]. In fact, ‘non-stimulating’ contacts of T cells with self-MHC are regarded essential for T-cell survival in the periphery [29].

Delivery of Nickel Ions to TCR/MHC Contact Sites via Carriers

One serious problem in relating in vitro to in vivo situations for T-cell reactivity to Ni arises from the fact that in vitro Ni2+ concentrations of 10 –4 M and more are required for effective clonal activation [22, 27, 30]. These concentrations are already close to the toxic range and are very unlikely to be reached in a lymph node during the period of sensitization. If the above model should apply to reality, we therefore need to postulate a system, e.g. carrier proteins that concentrate Ni2+ in close proximity to TCR/MHC contacts. In humans, one known carrier for Ni2+ is HSA with its amino terminal four amino acids (Asp-

Nickel Hypersensitivity

AlaHisLys) being identified as coordination sites [11]. By producing dialyzed HSA-Ni complexes, we were able to show that these complexes did not withdraw Ni2+ from reactive T cells, but rather allowed their activation at identical molar concentrations as free Ni salts [31]. The presence of metal chelators such as HSA in skin and serum reduces the possibility to find free Ni2+ anywhere in the human body to a minimum. Moreover, effective exchange of Ni2+ between HSA and other proteins or peptides requires at least equal Ni affinity for the acceptor as compared to HSA [31]. This implies that the combination of the SE9TCR with HLA-DR supplies a high-affinity coordination system for Ni2+, schematically depicted in figure 4. In vivo, any effective Ni carrier would thus have to exhibit a sufficient but not too high affinity for Ni2+ to guarantee sufficient transport and effective metal exchange to the target structures. An additional requirement for an effective carrier is the concentration of Ni2+ in the vicinity of TCR/MHC complexes. For this purpose it either would have to exhibit particular affinity to TCR, to MHC or to specialized membrane areas harboring these structures. We do not know at present whether HSA might even fulfill these predictions or whether one should search for other proteins which might be more specialized and efficient in this respect. A recent proteomic approach [32] indeed identified numerous new human Ni-binding proteins. Interestingly, these included a variety of heat-shock proteins known to be involved in the presentation of classical peptide antigens. From these findings it appears not unlikely that Ni carriers can be identified which have a preference for locations involving MHC/peptide complexes, fulfilling the above requirements for selective enrichment of Ni2+ in such locations. HSA, in fact, may even in this scenario serve as an important intermediate transporter.

51

a TCR H

Y

Y

Ni K




A

β

D

H81

HSA

HLA-DR

b

TCR H

A

K

D


Y

Y

HSA

β

Ni

H81 HLA-DR

Fig. 4. a, b Delivery of Ni2+ to TCR-MHC associates by HSA. The combination of the SE9 TCR with most HLA-DR molecules apparently supplies an arrangement of enough coordination sites for Ni2+ to facilitate a transfer of metal ions from carrier molecules such as HSA. Since DR-associated peptides are not involved as ligands, each individual HLA-DR molecule on the surface of an APC may serve as a counterpart for this TCR.

Lessons for Drug Hypersensitivity

The issue whether Ni should be regarded as a ‘typical’ hapten or not is being controversially discussed. However, the fact that, on the one hand, it constitutes the major human contact sensitizer and, on the other hand, it may produce very hapten-like allergenic determinants [20] reduces the matter to a rather academic question. It therefore seems legitimate to draw parallels from what we learned for Ni allergy to general drug hypersensitivity reactions. Clearly, the classical hapten concept envisaging haptenated peptides as the effective allergenic determinants will hold true for many if not most drugs exhibiting hap-

52

tenic or pro-haptenic properties. However, the increasing awareness of the allergenic potential of drugs and anesthetics, which are neither chemically reactive nor appear to require metabolic conversion [33], brings into focus potentially non-covalent drug interactions with MHC and/ or TCR. In this context it is worthwhile to recall that the classical type of antigen recognition by T cells exclusively involves non-covalent interactions between MHC, peptide and TCR. It does not take too much imagination to envisage low molecular non-peptide compounds, other than metal ions, to engage TCR and MHC into noncovalent interactions and thus to trigger the TCRsignaling cascade. A prerequisite for such complexes would be a polyvalence of non-covalent drug interactions, resulting in a molecular scenario largely analogous to the one depicted above for metal ions. Molecular cross-links may be envisaged between TCR and MHC, between TCR, MHC and peptide, or between TCR and peptide. In addition, one might even envisage situations where drug molecules occupy the position of peptides in the MHC-binding groove. The problem in all these cases is, like for Ni, the fact that it is hard to imagine how those chemicals would ever accumulate in vivo at high enough concentrations in close enough vicinity to short-lived TCR-MHC aggregates to stabilize those for full activation of the engaged T cells. In analogy to Ni, one way to circumvent this dilemma is to postulate drug-carrier molecules to concentrate the allergen near the TCR-MHC interphase. Again, one possible candidate might be HSA, which has been identified as a carrier for many different drugs. However, experimental evidence for HSA to bind all those drugs discussed as ‘non-covalent allergens’ is missing, and as mentioned before, we have no evidence as yet for a specific association of HSA complexes with MHC or TCR on the respective cells. As in the case of Ni allergy, it might therefore be worthwhile to search for other carriers which fulfill those requirements.

Weltzien

In conclusion, investigating the structural basis of Ni recognition by T cells from Ni-sensitized individuals has widened our view on the molecular interactions underlying hapten and drug hypersensitivity in general. These studies have demonstrated that even one individual type of chemical allergen, i.e. Ni2+, may induce the activation of specific T cells by a variety of different mechanisms. Why then should all other allergenic chemicals and drugs exclusively be recognized only upon covalent attachment to MHC-bound

peptides? Clearly, the classical hapten concept has to be expanded to accommodate low molecular allergens, which fail to engage in covalent interactions with antigen-presenting molecules. The p-i concept of Gerber and Pichler [33] may be one feasible model to describe such reactions. However, as mentioned above, the questions how such allergens are brought into contact with the TCR/MHC interphase in vivo, and what type of carrier molecules may be involved, remain important topics for future research.

References 1 Landsteiner K, Jacobs JL: Studies on the sensitization of animals with simple chemical compounds. J Exp Med 1935;61:643–656. 2 Gocinski BL, Tigelaar RE: Roles of CD4+ and CD8+ T cells in murine contact sensitivity revealed by in vivo monoclonal antibody depletion. J Immunol 1990;144:4121–4128. 3 Shearer GM, Rehn TG, Garbarino CA: Cell-mediated lympholysis of trinitrophenyl-modified autologous lymphocytes. Effector cell specificity to modified cell surface components controlled by H-2K and H-2D serological regions of the murine major histocompatibility complex. J Exp Med 1975;141:1348– 1364. 4 Buus S, Sette A, Colon SM, Miles C, Grey HM: The relation between major histocompatibility complex restriction and the capacity of Ia to bind immunogenic peptides. Science 1987;235:1353– 1358. 5 Falk K, Rötzschke O, Stefanovic S, Jung G, Rammensee HG: Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules. Nature 1991;351:290–296. 6 Ortmann B, Martin S, von Bonin A, Schiltz E, Hoschutzky H, Weltzien HU: Synthetic peptides anchor T-cell-specific TNP epitopes to MHC antigens. J Immunol 1992;148:1445–1450. 7 Martin S, Lappin MB, Kohler J, Delattre V, Leicht C, Preckel T, Simon JC, Weltzien HU: Peptide immunization indicates that CD8+ T cells are the dominant effector cells in trinitrophe-

Nickel Hypersensitivity

8

9

10

11

12

13

nyl-specific contact hypersensitivity. J Invest Dermatol 2000;115:260–266. Padovan E, Mauri-Hellweg D, Pichler WJ, Weltzien HU: T-cell recognition of penicillin G: structural features determining antigenic specificity. Eur J Immunol 1996;26:42–48. Weltzien HU, Dötze A, Gamerdinger K, Hellwig S, Thierse HJ: Molecular recognition of haptens by T cells: more than one way to tickle the receptor; in Cavani A, Girolomoni G (eds): Immune Mechanisms in Allergic Contact Dermatitis. Georgetown/Tex, Landes Bioscience, 2004, pp 14–27 (online: www. Eurekah.com). Budinger L, Hertl M: Immunologic mechanisms in hypersensitivity reactions to metal ions: an overview. Allergy 2000;55:108–115. Bal W, Christodoulou J, Sadler PJ, Tucker A: Multi-metal binding site of serum albumin. J Inorg Biochem 1998; 70:33–39. Sinigaglia F, Scheidegger D, Garotta G, Scheper R, Pletscher M, Lanzavecchia A: Isolation and characterization of Nispecific T-cell clones from patients with Ni-contact dermatitis. J Immunol 1985;135:3929–3932. Kapsenberg ML, Res P, Bos JD, Schootemijer A, Teunissen MBM, Schooten WV: Nickel-specific T-lymphocyte clones derived from allergic nickel-contact dermatitis lesions in man: heterogeneity based on requirement of dendritic antigen-presenting cell subsets. Eur J Immunol 1987;17: 861–865.

14 Romagnoli P, Labhardt AM, Sinigaglia F: Selective interaction of Ni with an MHC-bound peptide. EMBO J 1991;10: 1303–1306. 15 Cavani A, Ottaviani C, Nasorri F, Sebastiani S, Girolomoni G: Immunoregulation of hapten- and drug-induced immune reactions. Curr Opin Allergy Clin Immunol 2003;3:243–247. 16 Thierse HJ, Gamerdinger K, Junkes C, Guerreiro N, Weltzien HU: T-cell receptor interaction with haptens: metal ions as non-classical haptens. Toxicology 2005, 209:101–107. 17 Jabri E, Carr MB, Hausinger RP, Karplus PA: The crystal structure of urease from Klebsiella aerogenes. Science 1995;268:998–1004. 18 Martin S, Ortmann B, Pflugfelder U, Birsner U, Weltzien HU: Role of hapten-anchoring peptides in defining hapten epitopes for MHC-restricted cytotoxic T cells. Cross-reactive TNP determinants on different peptides. J Immunol 1992;149:2569–2575. 19 Griem P, Vonvultee C, Panthel K, Best SL, Sadler PJ, Shaw CF: T-cell crossreactivity to heavy-metals – identical cryptic peptides may be presented from protein exposed to different metals. Eur J Immunol 1998;28:1941–1947. 20 Lu L, Vollmer J, Moulon C, Weltzien HU, Marrack P, Kappler J: Components of the ligand for a Ni 2+-reactive human T-cell clone. J Exp Med 2003;197:567– 574.

53

21 Vollmer J, Weltzien HU, Gamerdinger K, Lang S, Choleva Y, Moulon C: Antigen contacts by Ni-reactive TCR: typical
 chain cooperation versus
chain-dominated specificity. Int Immunol 2000;12:1723–1731. 22 Moulon C, Vollmer J, Weltzien HU: Characterization of processing requirements and metal cross-reactivities in Tcell clones from patients with allergic contact dermatitis to nickel. Eur J Immunol 1995;25:3308–3315. 23 Moulon C, Choleva Y, Thierse HJ, Wild D, Weltzien HU: T-cell receptor transfection shows non-HLA-restricted recognition of nickel by CD8+ human T cells to be mediated by
 T-cell receptors. J Invest Dermatol 2003; 121:496– 501. 24 Vollmer J, Fritz M, Dormoy A, Weltzien HU, Moulon C: Dominance of the BV17 element in nickel-specific human T-cell receptors relates to severity of contact sensitivity. Eur J Immunol 1997;27: 1865–1874.

25 Budinger L, Neuser N, Totzke U, Merk HF, Hertl M: Preferential usage of TCRV17 by peripheral and cutaneous T cells in nickel-induced contact dermatitis. J Immunol 2001;167:6038–6044. 26 Vollmer J, Weltzien HU, Moulon C: TCR reactivity in human nickel allergy indicates contacts with complementarity-determining region 3 but excludes superantigen-like recognition. J Immunol 1999;163:2723–2731. 27 Gamerdinger K, Moulon C, Karp DR, Van Bergen J, Koning F, Wild D, Pflugfelder U, Weltzien HU: A new type of metal recognition by human T cells: contact residues for peptide-independent bridging of T-cell receptor and major histocompatibility complex by nickel. J Exp Med 2003;197:1345–1353. 28 Weltzien HU, Gamerdinger K, Thierse HJ: Nickel presentation; in Lutz MB, Romani N, Steinkasserer A (eds): Handbook of Dendritic Cells – Biology, Diseases and Therapies. Weinheim, Wiley-VCH, 2006, vol 3, pp 1047–1058. 29 Bender J, Mitchell T, Kappler J, Marrack P: CD4+ T-cell division in irradiated mice requires peptides distinct from those responsible for thymic selection. J Exp Med 1999;190:367–373.

30 Pistoor FHM, Kapsenberg ML, Bos JD, Meinardi MMHM, von Blomberg BME, Scheper RJ: Cross-reactivity of human nickel-reactive T-lymphocyte clones with copper and palladium. J Invest Dermatol 1995;105:92–95. 31 Thierse HJ, Moulon C, Allespach Y, Zimmermann B, Doetze A, Kuppig S, Wild D, Herberg F, Weltzien HU: Metal-protein complex-mediated transport and delivery of Ni 2+ to TCR/MHC contact sites in nickel-specific human Tcell activation. J Immunol 2004;172: 1926–1934. 32 Heiss K, Junkes C, Guerreiro N, Swamy M, Camacho-Carvachal MM, Schamel WWA, Haidl ID, Wild D, Weltzien HU, Thierse HJ: Subproteomic analysis of metal-interacting proteins in human B cells. Proteomics 2005;5:3614–3622. 33 Gerber BO, Pichler WJ: Noncovalent interactions of drugs with immune receptors may mediate drug-induced hypersensitivity reactions. AAPS J 2006; 8:E160–E165.

Prof. Dr. H.U. Weltzien Schillhof 5 DE–79110 Freiburg (Germany) Tel. +49 171 423 2575 E-Mail [email protected] or [email protected]

54

Weltzien

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 55–65

Drugs as Haptens, Antigens, and Immunogens B. Kevin Park
Joseph P. Sanderson
Dean J. Naisbitt Department of Pharmacology and Therapeutics, University of Liverpool, Liverpool, UK

Abstract It is known that drugs and other small molecular weight compounds can activate the immune system. In this review we discuss the known and proposed mechanisms by which such compounds can act as haptens, antigens and immunogens. In particular, we focus on nitrohalobenzenes, nickel, penicillins and sulfamethoxazole in order to draw conclusions about both the differences, but more importantly, the common features that these compounds share in terms of their immunogenicity. Copyright © 2007 S. Karger AG, Basel

Introduction

Immune-mediated adverse drug reactions are a major problem, both clinically and in drug development. Despite extensive research, much is still unknown about the fundamental immunological mechanisms behind such reactions. From an immunotoxicological perspective it is essential that these fundamental mechanisms are elucidated in order to reduce this burden, in order to inform the physician with regard to the safer use of existing treatments and also to inform the medicinal chemist to facilitate the design of safer and better tolerated drugs.

Before considering the possibilities for a drug to activate the immune system, it is instructive to define some of the terms which will be used in this discussion, namely, hapten, antigen, and immunogen [1, 2]. Hapten: A small molecule which can only induce an immunological response when bound to a carrier protein, either exogenous or endogenous. Antigen: Any substance which can be specifically bound by an antibody or T-cell receptor (TCR). Immunogen: A molecule which can stimulate a cellular or humoral immune response. It is important to note that individual compounds may have any of a number of combinations of these properties – while all immunogens are antigens, not all antigens are immunogens; for instance, self-proteins are antigenic, but do not normally stimulate an immune response. One of the most important questions that is still to be answered is how a therapeutic agent activates the immune system. There are two general mechanisms, which can be considered to contribute to the ability of low molecular weight compounds, including drugs, to initiate an immune response.

Table 1. Macromolecular chemical interactions

Type

Energy, kJ/mol

Relationship between strength and distance

Covalent bond Ion-ion Ion-dipole Dipole-dipole Hydrogen Van der Waals forces Hydrophobic interactions

150–600 20–40 8–20 3–20 5–30 0.5–5 3.4 per methylene group

N/A 1/d 1/d2 1/d3 1/d4 1/d5–1/d8

Firstly, several compounds are known to activate the immune system directly by binding to immune-modulating receptors or signaling molecules and thereby hyperstimulating the innate immune system. Good examples of this are imiquimod, a imidazoquinoline derivative which has been found to activate antigen-presenting cells via direct interactions with the TLR-7 receptor [3], and lenalidomide, a thalidomide analog which supplies costimulation to T cells via activation of the B7-CD28 pathway [4]. These drugs have both been associated with cutaneous manifestations [5, 6], although little is known about the mechanisms of these. An additional case which is worth discussing, although it involves a modified antibody rather than a low molecular weight chemical, is that of the TGN-1412 trial at Northwick Park Hospital. This compound was designed to provide costimulatory signals to T cells, but induced a cytokine storm response leading to severe systemic effects [7]. This again illustrates that a compound may activate the immune system without necessarily being either immunogenic or antigenic. These actions are a function of the primary pharmacology of the drug, but they do raise the possibility that drugs from a wide range of pharmacological classes may initiate immune-mediated reactions in this way. The second mechanism by which drugs can activate immune responses is by acting as an antigen. It was originally believed that this occurred almost exclusively via a hapten mechanism, although this is now regarded as being too simplis-

56

tic an assumption. In order to explore this mechanism fully, we will firstly recap the chemical bonds which underlie macromolecular interactions, discuss the mechanisms by which peptidic antigens are recognized by the immune system, followed by an in-depth focus on certain important paradigm compounds, namely, nitrohalobenzenes, nickel, penicillins and sulfamethoxazole. This will illustrate the different possible ways by which the immune system can recognize drugs, and it also offers insights into the common features which they all share.

Chemical Interactions Involved in Macromolecule Binding

Several different interactions are known to be important for drug-receptor interactions and are summarized in table 1. These can also be applied to the interaction between drugs and either antibodies or the TCR. The important determinant controlling overall bond energy, which in turn determines affinity and efficacy, is the summation of all bond energies. Therefore, although individually covalent bonds are markedly stronger than noncovalent interactions, it is certainly possible for multiple noncovalent interactions to form a significant binding force. This is especially likely to be the case when the two binding surfaces fit closely together, and hence allow multiple contacts for Van der Waals and hydrogen bonds to form.

Park ⴢ Sanderson ⴢ Naisbitt

Peptidic Antigen Recognition by T Cells Peptide antigens are specifically recognized by T cells via the TCRs interacting with the peptide bound noncovalently to the major histocompatibility (MHC) molecule. Antigen-presenting cells process proteins into peptides, load them onto nascent MHC molecules and traffic them to the cell surface. The composite MHC-peptide surface then acts as a ligand for peptide-MHC-specific TCRs. Recent work has elucidated many of the structural requirements for the TCR-peptideMHC interaction [for review, see 8]. Peptides fit into a specific groove on the surface of MHC molecules where they are held in place by noncovalent interactions, largely hydrogen bonds and Van der Waals forces, particularly at certain conserved residues, known as anchor residues. MHC-I molecules typically accept nonameric peptides, although some longer peptides have been seen. The N- and C-terminals are fixed in place, and longer peptides are accommodated by bulging of the peptide [9]. In addition to the essential carboxy-terminal anchor position, MHC-I-bound peptides have up to three further anchor residues. MHC-II-bound peptides are generally held less tightly, and are less constrained in terms of peptide length (between 12 and 25 is common) [10]. The interactions between the peptide and MHC molecule are also less constrained, and involve more uniformly dispersed residues. While anchor residues are particularly important for binding to MHC molecules, other residues make up a disproportionate amount of the binding to the TCR, especially for MHC-I-bound peptides. Indeed, in most studied MHC-I-bound 9-amino acid peptides, the residue at position 5 is responsible for over half of all TCR-peptide contacts [11]. As was the case for MHC interactions, peptides bound to MHC-II are less dependent on certain residues for TCR interactions, but they still only contact the TCR through a subset of available residues.

Drugs as Haptens, Antigens, and Immunogens

A striking finding of recent studies of TCRMHC-bound structures is the heterogeneity of the TCR-peptide-MHC interaction, as there appear to be very few common features to all interfaces visualized so far [11]. However, it has been discovered that the overall portion of the total binding area represented by TCR-peptide interactions is significantly lower than that supplied by TCR-MHC interactions. The important conclusions from these studies on TCR recognition of peptides is that although the peptides are relatively large, only a small portion of them is actually involved in binding to the TCR, and that TCR-MHC interactions can overshadow TCR-peptide interactions. Protein Recognition by Antibodies Unlike the TCR, antibodies do not bind to antigens using a structural background such as the MHC. This allows a much wider range of antigens to be recognized, both by individual antibodies and as a population. Therefore, and in further contrast to the TCR, it is well recognized that individual antibodies can bind both to haptenated proteins and to small molecules, which are free in solution. X-ray diffraction studies have characterized the structural features of antibodies bound to protein antigens. This has allowed greater insights into the bonds which underpin the molecular interactions. These studies have allowed the identification of several common features of protein recognition [12]: (1) both the light and heavy chains make contacts with the antigen, although the heavy chain typically dominates; (2) specificity is driven by the variable VH and VL regions; (3) the antibody-contacting residues on the antigen are not sequential, but form a contiguous surface of high complementarity to the antibody; (4) the contacting surfaces form an area of about 600–900 nm2, and (5) the binding is largely mediated by hydrogen bonds and Van der Waals forces, with ionic bonding making up a small proportion of the interactions.

57

Nitrohalobenzenes

The most widely studied classes of model chemical immunogens are the nitrohalobenzenes, particularly 2,4-dinitrochlorobenzene (DNCB), 2,4-dinitrofluorobenzene (DNFB), and 2,4,6-trinitrochlorobenzene. These are highly potent sensitizers, although they are rarely encountered amongst the population at large. However, their relatively simple chemistry and effectiveness for in vivo sensitization has made them essential tools for the investigation of fundamental immunochemistry. Nitrohalobenzenes are powerful electrophiles which readily react with nucleophilic residues on proteins, particularly cysteine and lysine groups, through an SNAr nucleophilic substitution mechanism (fig. 1A). This involves nucleophilic attack at the halogen-bonded carbon, which is stabilized by resonance structures involving the electron-withdrawing nitro groups, followed by preferential ejection of the halide ion as a good leaving group. This has been seen to occur with cysteine and lysine, both in vitro [13, 14] and in vivo [15, 16]. The biological relevance of chemical reactivity with other nucleophilic amino acid residues, such as tyrosine and histidine, is less certain. Some studies have shown that DNFB reacts with these residues [17, 18], whereas a recent analysis found extremely limited reactivity for DNCB to histidine [19]. Whether this reflects genuine differences between the two compounds or simply differences in experimental method is uncertain. It may also reflect the variable reactivity of amino acid side chains in different proteins. Chemical reactivity of individual amino acid residues is a function of their microenvironment within the protein, which, in addition to steric effects, can have a profound influence on pKa [20]. Importantly, all dinitrohalobenzenes generate the same dinitrophenyl (DNP) adduct on proteins, although the residues to which they attach may differ. Trinitrohalobenzenes generate trinitrophenylated (TNP) de-

58

rivatives of proteins, again typically on lysines and cysteines. Early studies focused on the ability of nitrohalobenzenes to induce humoral immune responses to haptenated proteins. In a series of elegant studies, Landsteiner and Jacobs [21, 22] found that the sensitizing potential of a series of nitrohalobenzenes in guinea pigs correlated with their chemical reactivity in vitro. Further work confirmed that in addition to the effects of the parent compounds, nitrophenylated cells and proteins were also potent immunogens [23, 24], although this was cell type and protein dependent. Antibodies generated in response to DNP- and TNPmodified proteins and peptides have been well studied and appear to be heterogenous in specificity and affinity for DNP [25–27], although most react both to the modified amino acids and modified peptides. T-cell responses to both nitrohalobenzenes and nitrophenyl-modified proteins and peptides have also been well characterized. T-cell clones have been generated from hypersensitive individuals which respond to DNCB in a processingand covalent binding-dependent manner [28], suggesting that DNCB must be bound solely to the peptide. This is in keeping with several other studies which have found that TNP-modified peptides are sufficient to activate in vivo primed CD4+ and CD8+ T cells [29, 30]. Additionally, TNP-modified peptides have been shown to be capable of priming TNP-specific T-cell responses in vivo [31]. These studies have found a precise requirement for TNP-modification of a specific residue, albeit on a given peptide. Whether this is a general feature of TNP reactivity is not known at present. An interesting development in our understanding of how contact hypersensitivity to sensitizers such as dinitrohalobenzenes occurs is the finding that they are capable of directly activating dendritic cells and supplying requisite costimulatory signals for primary T-cell activation [32–35]. The mechanism by which these com-

Park ⴢ Sanderson ⴢ Naisbitt

Fig. 1. Formation of protein conjugates by dinitrohalobenzenes (A) where X is any halogen and Nu represents any lone-pair bearing nucleophile; penicillins (B) showing the major penicilloyl determinant and two minor de-

Drugs as Haptens, Antigens, and Immunogens

terminants, penicillamine and penicillenyl, formed from penicilloic and penicillenic acid, respectively, and sulfamethoxazole (C) showing the route of bioactivation to a nitroso derivative via a hydroxylamine metabolite.

59

pounds act is not fully understood, although it appears to depend on the activation of tyrosine kinases [36] and the p38/MAPK signaling pathway [37], but not the NF-B or ERK pathways [38]. There is evidence that binding of DNFB and 2,4,6-trinitrochlorobenzene to thiol groups is particularly important [39], but this requires further work for confirmation. Therefore, it is clear that dinitrohalobenzenes may activate the immune system both by acting as an antigen, in a haptenic fashion, and by direct activation of the innate immune system to drive active T-cell priming. Indeed, this newly discovered function may be the mechanism behind the therapeutic immunomodulatory effects of DNCB as used topically in dermatology [40]. A full consideration of the interactions of this class of compounds will allow an understanding of how it can act as a complete immunogen.

Nickel

Nickel is one of the most commonly encountered contact sensitizers, and a common cause of allergy [41]. Ni2+ ions readily dissolve in sweat and are absorbed through the skin [42], where they bind to carrier proteins via four- and six-coordinate binding to electron donors such as oxygen, nitrogen and free thiols [43]. In biological matrices nickel is typically in equilibrium between binding to serum albumin and in complexes with free histidine [44]. Nickel-specific T cells have been cloned on a number of occasions [45–47], and these have allowed detailed investigations into the nature of the nickel antigen. It appears that for certain clones, nickel acts in a similar fashion to a classical hapten, in that it is not possible to wash the bound nickel away from antigen-presenting cells and eliminate T-cell reactivity [45]. In certain clones that react to nickel in this fashion, it appears that Ni2+ binds to the surface of a preformed peptide-MHC complex via a conserved His81

60

residue on HLA-DR and two residues on the peptide [48], whereas a functional antigen-processing pathway is required for other clones, suggesting that Ni2+ remains bound to the peptide as it is processed [45]. In both of these cases, however, it is clear that the bonds Ni2+ forms to the MHC peptide are of sufficient strength to allow the ion to act as a hapten. Alternatively, certain clones appear to depend on the continual presence of free Ni2+ ions for activation, and in these clones it appears to act in a peptide-independent superantigen-like fashion [49], although the presence of the histidine residue on HLA-DR still appears to be essential. In these cases, it seems likely that Ni2+ does not bind tightly to the MHC peptide, but forms a relatively weak interaction that can be removed by washing. Clinical cross-reactivity between nickel and other metals, particularly copper and palladium, has been observed [50, 51]. There is some evidence that this is due to T-cell cross-reactivity at the molecular level [46], but an interesting alternative possibility was suggested by the finding that heavy metals can produce cryptic epitopes [52], and that different metals can generate the same cryptic epitopes from a single protein [53]. While this mechanism cannot explain the existence of processing-independent nickel-reactive clones, it may well be important for determining the antigen specificity of processing-dependent nickel-reactive T cells. As has been found for nitrohalobenzenes, nickel has been recently associated with dendritic cell activation [32], although this appears to depend on different signaling pathways, particularly the ERK and p38 pathways [38]. The mechanism behind this process is as yet unknown. Nevertheless, an integration of the effects of nickel at both dendritic cell receptors and the TCR is essential to an understanding of when nickel behaves simply as an antigen and under what circumstances it is an effective immunogen.

Park ⴢ Sanderson ⴢ Naisbitt

Penicillin

Penicillin is a directly reactive drug associated with a high incidence of allergic reactions. It has a reactive structure, with the capacity to modify amine, hydroxyl, mercapto and histidine groups on protein. The site and extent of protein modification is influenced by concentration (of penicillin and protein), pH and the presence of certain metal ions [54]. Nucleophilic attack of the -lactam ring by free amino groups of protein leads to a covalently linked penicilloyl determinant, often referred to as a major antigenic determinant for antibodies and T cells. This derivative represents greater than 90% of penicillin-modified protein. Other penicillin-related adducts are formed through the conversion of penicillin in solution to penicillenic or penicilloic acid, both of which are thought to modify cysteine residues in protein. Resultant penicillenyl and penicillamine conjugates, called minor antigenic determinants, are formed in low levels, but may contribute towards the induction of allergic reactions (fig. 1B). T cells from allergic patients are mainly CD4+ and stimulated with penicillin presented in the context of MHC class II [55]. T cells are penicillin-specific (i.e. they are not stimulated with structurally unrelated drugs); however, certain clones accommodate small changes in the penicillin side chain, which are not involved in penicillin binding to protein. The requirement of penicillin modification of MHC-associated peptides and formation of a penicilloyl determinant for T-cell activation has been demonstrated by the stimulation of specific TCRs with designer peptides containing appropriate anchor residues [56]. Importantly, only peptides containing lysine residues in precise positions were able to stimulate TCRs [56]. The ability of penicillenyl or penicillamine conjugates to stimulate T cells has not been investigated. More recently, the involvement of dendritic cells in penicillin allergy has been evaluated.

Drugs as Haptens, Antigens, and Immunogens

Immature dendritic cells from amoxicillin-hypersensitive patients, but not those from controls, were activated in response to amoxicillin, and were able to present processed peptides to autologous peripheral blood mononuclear cells and T cells [57]. However, the mechanism behind this effect is not known at present, and will require further work to elucidate. Thus, while the primary focus on the elucidation of the mechanisms of penicillin hypersensitivity has been on the formation of the penicilloyl hapten, there is still the need to define the chemical basis of the generation of the costimulatory signals involved.

Sulfamethoxazole

Sulfamethoxazole (SMX) is a sulfonamide which was largely phased out of clinical practice because of rare but serious cutaneous hypersensitivity reactions. In recent years it has been more widely prescribed in combination with trimethoprim as treatment for Pneumocystis carinii pneumonia in patients with AIDS. Unfortunately, the incidence of hypersensitivity reactions is greatly increased in this population [58]. Many studies exploring the mechanisms of drug hypersensitivity have focused on SMX, because it is known to cause hypersensitivity and much is known about its disposition in the body (fig. 1C). SMX is metabolized by CYP2C9 in human liver to a proreactive hydroxylamine metabolite (SMX-NHOH) [59–62]. This readily reacts with molecular oxygen to generate nitroso SMX (SMX-NO) [63, 64], which is unstable and reacts both with SMX-NHOH to generate azo and azoxy dimers, and with protein [65, 66] and nonprotein thiols [63]. Further oxidation can also generate nitro SMX [66]. Importantly, reduction of SMXNO can occur either via interaction with nonprotein thiols (e.g., glutathione) and ascorbate, or enzymatically [59, 67]. Thus, the critical balance between metabolic activation and detoxification in

61

a given cell system ultimately determines the level of exposure to the reactive metabolite. Studies with animal models of SMX hypersensitivity have found that SMX-NO is a potent immunogen [66], and is capable of inducing hapten-specific T-cell responses similar to those generated by nitrohalobenzenes, i.e. both covalent binding and processing dependent [68]. Interestingly, a study demonstrated that while administration of SMX was insufficient to induce a T-cell response, despite in vivo generation of SMXNHOH, co-dosing with an adjuvant produced a T-cell response to SMX-NO [69]. This demonstrates that in vivo generated reactive metabolites may form antigens through haptenation, although additional activation of the innate immune system through danger signaling [70] is required to produce immunogenicity. T cells from hypersensitive patients have been isolated and cloned in order to explore the mechanisms by which SMX can act as an antigen in man. In contrast to animal studies, however, no T-cell clones have been found which respond to SMX-NO in a classical hapten-like fashion. Instead, in all patients studied so far, T-cell clones have been found which respond to SMX-NO in a binding-dependent but processing-independent fashion, presumably by binding to the preformed MHC-peptide complex [71]. This represents a vital distinction between immunogenicity in experimental animals, which is consistently observed, and hypersensitivity in a small subgroup of patients. However, the protein milieu in any given in vitro system will differ from that in the in vivo environment at the time of antigen encounter, so it may be premature to rule out the presence of T cells responsive towards metabolite-modified endogenous proteins. Unexpectedly, T-cell clones have been identified which also respond to SMX itself in the absence of covalent binding or antigen processing [72, 73]. This T-cell activation is TCR dependent [74] and MHC restricted, but does not appear to depend on the presence of specific peptides [75],

62

suggesting a reversible interaction between MHC-drug-TCR. Recently, studies with T cells isolated from blister fluid in toxic epidermal necrolysis patients found a similar pattern of antigen specificity [76], suggesting that these T cells may also be important effector cells in drug hypersensitivity. The differences between the nature of the antigen in SMX-immunogenicity between humans and animals are as yet unexplained, but they suggest caution regarding the use of animal models to investigate what is an idiosyncratic reaction in man, and one that is therefore patient-specific. It is important to note that the ability of SMX to function as an antigen with T cells from susceptible individuals does not necessarily prove that SMX per se was the original immunogen in those patients. In addition to the effects of SMX as an antigen, in recent work [77] we have shown the presence of SMX metabolite-modified intracellular proteins when SMX was incubated in vitro with human monocyte-derived dendritic cells. Dendritic cell metabolism of SMX resulted in covalent modification of endogenous protein and subsequent increased expression of the dendritic cell co-stimulatory receptor CD40. Further work has found that the CD40 signaling pathway seems to be important in the development of SMX immunogenicity. Further studies are underway to explore the effects of SMX treatment on dendritic cells from hypersensitive patients to evaluate whether altered immune cell metabolism and dendritic cell activation is associated with individual susceptibility and presentation of the metabolite-modified endogenous proteins to specific T cells.

Conclusions

It is clear that low molecular weight compounds can act as antigens in a variety of ways, from classical hapten mechanisms involving covalent

Park ⴢ Sanderson ⴢ Naisbitt

binding to endogenous protein and antigen processing and presentation, to noncovalent interactions between small molecules, MHC peptides and TCRs. However, much more work is required to be confident that we have fully elucidated the mechanisms at work in the induction and propagation of an immune response. In particular, we need to better understand the role of direct activation of the innate immune system by drugs in the primary induction of an immune response to these compounds, and relate these findings to the binding of the drug or metabolite to the TCR in the overall context of the immunological synapse. In relation to interactions with the TCR, that ability of a drug to function as an antigen will be

determined by the summation of bond energies between receptor and ligand. Therefore, one should not place too much emphasis on the importance of a single bond. Thus, simple arguments about the relative importance of covalent and noncovalent interactions become redundant. For a low molecular weight compound to function as an immunogen there must be an antigenic signal and also a costimulatory signal. There has been much speculation concerning the ability of drugs or metabolites to generate costimulatory (danger) signals. However, at present we have little knowledge of the receptors to which drugs and/or their metabolites interact to activate signaling pathways involved in costimulation.

References 1 Abbas AK, Lichtman AH: Cellular and Molecular Immunology, ed 5. Philadelphia, Saunders, 2003. 2 Park BK, Coleman JW, Kitteringham NR: Drug disposition and drug hypersensitivity. Biochem Pharmacol 1987; 36:581–590. 3 Hemmi H, Kaisho T, Takeuchi O, Sato S, Sanjo H, Hoshino K, Horiuchi T, Tomizawa H, Takeda K, Akira S: Small anti-viral compounds activate immune cells via the TLR7 MyD88-dependent signaling pathway. Nat Immunol 2002; 3:196–200. 4 LeBlanc R, Hideshima T, Catley LP, Shringarpure R, Burger R, Mitsiades N, Mitsiades C, Cheema P, Chauhan D, Richardson PG, Anderson KC, Munshi NC: Immunomodulatory drug costimulates T cells via the B7-CD28 pathway. Blood 2004;103:1787–1790. 5 Hadley G, Derry S, Moore RA: Imiquimod for actinic keratosis: systematic review and meta-analysis. J Invest Dermatol 2006;126:1251–1255. 6 Sharma RA, Steward WP, Daines CA, Knight RD, O’Byrne KJ, Dalgleish AG: Toxicity profile of the immunomodulatory thalidomide analogue, lenalidomide: phase I clinical trial of three dosing schedules in patients with solid malignancies. Eur J Cancer 2006;42: 2318–2325.

7 Suntharalingam G, Perry MR, Ward S, Brett SJ, Castello-Cortes A, Brunner MD, Panoskaltsis N: Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N Engl J Med 2006;355:1018–1028. 8 van der Merwe PA, Davis SJ: Molecular interactions mediating T cell antigen recognition. Annu Rev Immunol 2003; 21:659–684. 9 Speir JA, Stevens J, Joly E, Butcher GW, Wilson IA: Two different, highly exposed, bulged structures for an unusually long peptide bound to rat MHC class I RT1-Aa. Immunity 2001;14:81– 92. 10 Rammensee HG: Chemistry of peptides associated with MHC class I and class II molecules. Curr Opin Immunol 1995; 7:85–96. 11 Rudolph MG, Wilson IA: The specificity of TCR/pMHC interaction. Curr Opin Immunol 2002;14:52–65. 12 Braden BC, Poljak RJ: Structural features of the reactions between antibodies and protein antigens. FASEB J 1995; 9:9–16. 13 Habig WH, Pabst MJ, Jakoby WB: Glutathione S-transferases. The first enzymatic step in mercapturic acid formation. J Biol Chem 1974;249:7130–7139. 14 Tingle MD, Clarke JB, Kitteringham NR, Park BK: Influence of glutathione

Drugs as Haptens, Antigens, and Immunogens

15

16

17

18

19

conjugation on the immunogenicity of dinitrophenyl derivatives in the rat. Int Arch Allergy Appl Immunol 1990;91: 160–165. Kitteringham NR, Maggs JL, Newby S, Park BK: Drug-protein conjugates – VIII. The metabolic fate of the dinitrophenyl hapten conjugated to albumin. Biochem Pharmacol 1985;34:1763– 1771. Maggs JL, Kitteringham NR, Grabowski PS, Park BK: Drug-protein conjugates – X. The role of protein conjugation in the disposition of dinitrofluorobenzene. Biochem Pharmacol 1986;35:505–513. Jackson GE, Young NM: Determination of chemical properties of individual histidine and tyrosine residues of concanavalin A by competitive labeling with 1-fluoro-2,4-dinitrobenzene. Biochemistry 1986;25:1657–1662. Liu LF, Hong JL, Tsai SP, Hsieh JC, Tam MF: Reversible modification of rat liver glutathione S-transferase 3–3 with 1chloro-2,4-dinitrobenzene: specific labelling of Tyr-115. Biochem J 1993; 296:189–197. Gerberick GF, Vassallo JD, Bailey RE, Chaney JG, Morrall SW, Lepoittevin JP: Development of a peptide reactivity assay for screening contact allergens. Toxicol Sci 2004;81:332–343.

63

20 Bello J, Iijima H, Kartha G: A new arylating agent, 2-carboxy-4,6-dinitrochlorobenzene. Reaction with model compounds and bovine pancreatic ribonuclease. Int J Pept Protein Res 1979; 14:199–212. 21 Landsteiner K, Jacobs J: Studies on the sensitisation of animals with simple chemical compounds. J Exp Med 1935; 61:643–656. 22 Landsteiner K, Jacobs J: Studies on the sensitization of animals with simple chemical compounds II. J Exp Med 1936;64:625–639. 23 Soeberg B, Sumerska T, Binns RM, Balfour BM: Contact sensitivity in the pig. III. Induction by intralymphatic infusion of DNP-conjugated cell membranes and soluble proteins, free DNFB and some small molecular weight derivatives. Int Arch Allergy Appl Immunol 1978;57:481–487. 24 Soeberg B, Sumerska T, Binns RM, Balfour BM: Contact sensitivity in the pig. II. Induction by intralymphatic infusion of DNP conjugated cells. Int Arch Allergy Appl Immunol 1978;57:114– 125. 25 Bassolino-Klimas D, Bruccoleri RE, Subramaniam S: Modeling the antigen combining site of an anti-dinitrophenyl antibody, ANO2. Protein Sci 1992;1: 1465–1476. 26 Halsey JF, Cebra JJ, Biltonen RL: Structural and thermodynamic basis of affinity in anti-dinitrophenyl antibody. Biochemistry 1975;14:5221–5224. 27 Knight CG, Green NM: Interaction of dinitrophenyl groups bound to bovine serum albumin with univalent fragments of anti-dinitrophenyl antibody. Biochem J 1979;177:225–236. 28 Pickard C, Smith AM, Cooper H, Strickland I, Jackson J, Healy E, Friedmann PS: Investigation of mechanisms underlying the T-cell response to the hapten 2,4-dinitrochlorobenzene. J Invest Dermatol 2006, in press. 29 Cavani A, Hackett CJ, Wilson KJ, Rothbard JB, Katz SI: Characterization of epitopes recognized by hapten-specific CD4+ T cells. J Immunol 1995;154: 1232–1238. 30 Martin S, von Bonin A, Fessler C, Pflugfelder U, Weltzien HU: Structural complexity of antigenic determinants for class I MHC-restricted, hapten-specific T cells. Two qualitatively differing types of H-2Kb-restricted TNP epitopes. J Immunol 1993;151:678–687.

64

31 Kohler J, Martin S, Pflugfelder U, Ruh H, Vollmer J, Weltzien HU: Cross-reactive trinitrophenylated peptides as antigens for class II major histocompatibility complex-restricted T cells and inducers of contact sensitivity in mice. Limited T cell receptor repertoire. Eur J Immunol 1995;25:92–101. 32 Aiba S, Terunuma A, Manome H, Tagami H: Dendritic cells differently respond to haptens and irritants by their production of cytokines and expression of co-stimulatory molecules. Eur J Immunol 1997;27:3031–3038. 33 Coutant KD, de Fraissinette AB, Cordier A, Ulrich P: Modulation of the activity of human monocyte-derived dendritic cells by chemical haptens, a metal allergen, and a staphylococcal superantigen. Toxicol Sci 1999;52:189–198. 34 Jugde F, Boissier C, Rougier-Larzat N, Corlu A, Chesne C, Semana G, Heresbach D: Regulation by allergens of chemokine receptor expression on in vitro-generated dendritic cells. Toxicology 2005;212:227–238. 35 Manome H, Aiba S, Tagami H: Simple chemicals can induce maturation and apoptosis of dendritic cells. Immunology 1999;98:481–490. 36 Kuhn U, Brand P, Willemsen J, Jonuleit H, Enk AH, van Brandwijk-Petershans R, Saloga J, Knop J, Becker D: Induction of tyrosine phosphorylation in human MHC class II-positive antigenpresenting cells by stimulation with contact sensitizers. J Immunol 1998; 160:667–673. 37 Arrighi JF, Rebsamen M, Rousset F, Kindler V, Hauser C: A critical role for p38 mitogen-activated protein kinase in the maturation of human bloodderived dendritic cells induced by lipopolysaccharide, TNF-alpha, and contact sensitizers. J Immunol 2001;166: 3837–3845. 38 Aiba S, Manome H, Nakagawa S, Mollah ZU, Mizuashi M, Ohtani T, Yoshino Y, Tagami H: p38 Mitogen-activated protein kinase and extracellular signalregulated kinases play distinct roles in the activation of dendritic cells by two representative haptens, NiCl2 and 2,4dinitrochlorobenzene. J Invest Dermatol 2003;120:390–399. 39 Bruchhausen S, Zahn S, Valk E, Knop J, Becker D: Thiol antioxidants block the activation of antigen-presenting cells by contact sensitizers. J Invest Dermatol 2003;121:1039–1044.

40 Holzer AM, Kaplan LL, Levis WR: Haptens as drugs: contact allergens are powerful topical immunomodulators. J Drugs Dermatol 2006;5:410–416. 41 Nielsen NH, Linneberg A, Menne T, Madsen F, Frolund L, Dirksen A, Jorgensen T: Incidence of allergic contact sensitization in Danish adults between 1990 and 1998; the Copenhagen Allergy Study, Denmark. Br J Dermatol 2002;147:487–492. 42 Menne T: Prevention of nickel allergy by regulation of specific exposures. Ann Clin Lab Sci 1996;26:133–138. 43 Zhang Y, Wilcox DE: Thermodynamic and spectroscopic study of Cu(II) and Ni(II) binding to bovine serum albumin. J Biol Inorg Chem 2002;7:327–337. 44 Tabata M, Sarkar B: Specific nickel(II)transfer process between the native sequence peptide representing the nickel(II)-transport site of human serum albumin and L-histidine. J Inorg Biochem 1992;45:93–104. 45 Moulon C, Vollmer J, Weltzien HU: Characterization of processing requirements and metal cross-reactivities in T cell clones from patients with allergic contact dermatitis to nickel. Eur J Immunol 1995;25:3308–3315. 46 Pistoor FH, Kapsenberg ML, Bos JD, Meinardi MM, von Blomberg ME, Scheper RJ: Cross-reactivity of human nickel-reactive T-lymphocyte clones with copper and palladium. J Invest Dermatol 1995;105:92–95. 47 Sinigaglia F, Scheidegger D, Garotta G, Scheper R, Pletscher M, Lanzavecchia A: Isolation and characterization of Nispecific T cell clones from patients with Ni-contact dermatitis. J Immunol 1985; 135:3929–3932. 48 Lu L, Vollmer J, Moulon C, Weltzien HU, Marrack P, Kappler J: Components of the ligand for a Ni++ reactive human T cell clone. J Exp Med 2003;197:567– 574. 49 Gamerdinger K, Moulon C, Karp DR, Van Bergen J, Koning F, Wild D, Pflugfelder U, Weltzien HU: A new type of metal recognition by human T cells: contact residues for peptide-independent bridging of T cell receptor and major histocompatibility complex by nickel. J Exp Med 2003;197:1345–1353. 50 Kanerva L, Kerosuo H, Kullaa A, Kerosuo E: Allergic patch test reactions to palladium chloride in schoolchildren. Contact Dermatitis 1996;34:39–42.

Park ⴢ Sanderson ⴢ Naisbitt

51 Wohrl S, Hemmer W, Focke M, Gotz M, Jarisch R: Copper allergy revisited. J Am Acad Dermatol 2001; 45:863–870. 52 Griem P, Panthel K, Kalbacher H, Gleichmann E: Alteration of a model antigen by Au(III) leads to T cell sensitization to cryptic peptides. Eur J Immunol 1996;26:279–287. 53 Griem P, von Vultee C, Panthel K, Best SL, Sadler PJ, Shaw CF 3rd: T cell crossreactivity to heavy metals: identical cryptic peptides may be presented from protein exposed to different metals. Eur J Immunol 1998;28:1941–1947. 54 Ahlstedt S, Kristofferson A: Immune mechanisms for induction of penicillin allergy. Prog Allergy 1982;30:67–134. 55 Padovan E, Mauri-Hellweg D, Pichler WJ, Weltzien HU: T cell recognition of penicillin G: structural features determining antigenic specificity. Eur J Immunol 1996;26:42–48. 56 Padovan E, Bauer T, Tongio MM, Kalbacher H, Weltzien HU: Penicilloyl peptides are recognized as T cell antigenic determinants in penicillin allergy. Eur J Immunol 1997;27:1303–1307. 57 Rodriguez-Pena R, Lopez S, Mayorga C, Antunez C, Fernandez TD, Torres MJ, Blanca M: Potential involvement of dendritic cells in delayed-type hypersensitivity reactions to beta-lactams. J Allergy Clin Immunol 2006;118:949–956. 58 Pirmohamed M, Park BK: HIV and drug allergy. Curr Opin Allergy Clin Immunol 2001;1:311–316. 59 Cribb AE, Spielberg SP, Griffin GP: N4hydroxylation of sulfamethoxazole by cytochrome P450 of the cytochrome P4502C subfamily and reduction of sulfamethoxazole hydroxylamine in human and rat hepatic microsomes. Drug Metab Dispos 1995;23:406–414. 60 Gill HJ, Maggs JL, Madden S, Pirmohamed M, Park BK: The effect of fluconazole and ketoconazole on the metabolism of sulphamethoxazole. Br J Clin Pharmacol 1996;42:347–353. 61 Mitra AK, Thummel KE, Kalhorn TF, Kharasch ED, Unadkat JD, Slattery JT: Inhibition of sulfamethoxazole hydroxylamine formation by fluconazole in human liver microsomes and healthy

62

63

64

65

66

67

68

69

volunteers. Clin Pharmacol Ther 1996; 59:332–340. van der Ven AJ, Mantel MA, Vree TB, Koopmans PP and van der Meer JW: Formation and elimination of sulphamethoxazole hydroxylamine after oral administration of sulphamethoxazole. Br J Clin Pharmacol 1994;38:147–150. Cribb AE, Miller M, Leeder JS, Hill J, Spielberg SP: Reactions of the nitroso and hydroxylamine metabolites of sulfamethoxazole with reduced glutathione. Implications for idiosyncratic toxicity. Drug Metab Dispos 1991;19: 900–906. Naisbitt DJ, O’Neill PM, Pirmohamed M, Park BK: Synthesis and reactions of nitroso sulphamethoxazole with biological nucleophiles: Implications for immune mediated toxicity. Bioorg Med Chem Lett 1996;6:1511–1516. Manchanda T, Hess D, Dale L, Ferguson SG, Rieder MJ: Haptenation of sulfonamide reactive metabolites to cellular proteins. Mol Pharmacol 2002;62: 1011–1026. Naisbitt DJ, Farrell J, Gordon SF, Maggs JL, Burkhart C, Pichler WJ, Pirmohamed M, Park BK: Covalent binding of the nitroso metabolite of sulfamethoxazole leads to toxicity and major histocompatibility complex-restricted antigen presentation. Mol Pharmacol 2002;62:628–637. Kurian JR, Bajad SU, Miller JL, Chin NA, Trepanier LA: NADH cytochrome b5 reductase and cytochrome b5 catalyze the microsomal reduction of xenobiotic hydroxylamines and amidoximes in humans. J Pharmacol Exp Ther 2004;311:1171–1178. Farrell J, Naisbitt DJ, Drummond NS, Depta JP, Vilar FJ, Pirmohamed M, Park BK: Characterization of sulfamethoxazole and sulfamethoxazole metabolite-specific T-cell responses in animals and humans. J Pharmacol Exp Ther 2003;306:229–237. Naisbitt DJ, Gordon SF, Pirmohamed M, Burkhart C, Cribb AE, Pichler WJ, Park BK: Antigenicity and immunogenicity of sulphamethoxazole: demonstration of metabolism-dependent hap-

70

71

72

73

74

75

76

77

tenation and T-cell proliferation in vivo. Br J Pharmacol 2001;133:295–305. Park BK, Pirmohamed M, Kitteringham NR: Role of drug disposition in drug hypersensitivity: a chemical, molecular, and clinical perspective. Chem Res Toxicol 1998;11:969–988. Schnyder B, Burkhart C, SchnyderFrutig K, von Greyerz S, Naisbitt DJ, Pirmohamed M, Park BK, Pichler WJ: Recognition of sulfamethoxazole and its reactive metabolites by drug-specific CD4+ T cells from allergic individuals. J Immunol 2000;164:6647–6654. Schnyder B, Mauri-Hellweg D, Zanni M, Bettens F, Pichler WJ: Direct, MHCdependent presentation of the drug sulfamethoxazole to human alphabeta T cell clones. J Clin Invest 1997;100: 136–141. Zanni MP, von Greyerz S, Schnyder B, Brander KA, Frutig K, Hari Y, Valitutti S, Pichler WJ: HLA-restricted, processing- and metabolism-independent pathway of drug recognition by human alpha beta T lymphocytes. J Clin Invest 1998;102:1591–1598. Depta JP, Altznauer F, Gamerdinger K, Burkhart C, Weltzien HU, Pichler WJ: Drug interaction with T-cell receptors: T-cell receptor density determines degree of cross-reactivity. J Allergy Clin Immunol 2004;113:519–527. Burkhart C, Britschgi M, Strasser I, Depta JP, von Greyerz S, Barnaba V, Pichler WJ: Non-covalent presentation of sulfamethoxazole to human CD4+ T cells is independent of distinct human leucocyte antigen-bound peptides. Clin Exp Allergy 2002;32:1635–1643. Nassif A, Bensussan A, Boumsell L, Deniaud A, Moslehi H, Wolkenstein P, Bagot M, Roujeau JC: Toxic epidermal necrolysis: effector cells are drug-specific cytotoxic T cells. J Allergy Clin Immunol 2004;114:1209–1215. Sanderson JP, Naisbitt DJ, Farrell J, Ashby CA, Tucker MJ, Rieder MJ, Pirmohamed M, Clarke SE, Park BK: Sulphamethoxazole and its metabolite nitroso sulfamethomazole stimulate denditric cell co-stimulatory signalling. J Immunol, in press.

Dr. Kevin Park Department of Pharmacology and Therapeutics, University of Liverpool Sherrington Building, Ashton Street Liverpool L69 3GE (UK) Tel. +44 151 794 5559, Fax +44 151 794 5540, E-Mail [email protected]

Drugs as Haptens, Antigens, and Immunogens

65

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 66–73

The p-i Concept: Evidence and Implications Basil O. Gerber
Werner J. Pichler Division of Allergology, Clinic and Policlinic for Rheumatology and Clinical Immunology/Allergology, Inselspital, Bern, Switzerland

Abstract For decades, the hapten concept has served as the paradigm which explains how low molecular weight antigens are capable of triggering an immune response. In accordance with this, many drugs have been shown to be haptens and consequently require covalent binding to carrier peptides in order to elicit hypersensitivity reactions. However, a growing body of evidence suggests that a substantial number of drugs employ a different mechanism which involves reversible rather than covalent binding to either major histocompatibility complex proteins or T-cell receptors. We have recently postulated a hypothesis, which we termed ‘p-i concept’, to account for these observations. Drawing on different model systems of drug-mediated and other ‘unconventional’ T-cell activation, we present evidence in support of the p-i concept, and discuss its potential implications for drug hypersensitivity. Copyright © 2007 S. Karger AG, Basel

Introduction: Drugs as Haptens and Prohaptens

For decades, the hapten concept has served as the paradigm to explain the immunogenicity of small pharmacological agents [1], according to which the recognition of low molecular weight compounds by the immune system depends on the

compounds’ ability to covalently bind to a carrier protein. The modified carrier molecules will then be processed by antigen-presenting cells (APC) and presented to T cells, subsequently triggering an immune response. There is abundant evidence for the notion that the hapten concept is applicable to hypersensitivity-inducing drugs, which is reviewed elsewhere in this book. A more recent extension of the hapten concept states that chemically inert drugs can gain reactivity by preceding metabolization; such compounds are classified as prohaptens [2]. While the chain of evidence may not be quite as solid, there is little dissent about the relevance of the prohapten concept for drug hypersensitivity, and recent developments in this field will be presented elsewhere as well. In this chapter, we shall instead focus on the presumable mechanism(s) of action of the increasing number of drugs that evoke hypersensitivity reactions, even though they are neither known to be chemically reactive nor to require metabolic conversion to a reactive intermediate. Further, we shall discuss the implications of these presumable mechanisms on systemic immune activation in drug hypersensitivity.

‘Non-Hapten’ Drugs as Antigens

In essence, there are three lines of evidence that a large number of ‘non-hapten’ drugs are potent antigens that trigger T-cell activation. Specifically, several drugs are antigenic in their parental, non-reactive form, some drugs do not require antigen processing, and the antigen presentation of certain drugs involves reversible and not covalent binding. In many instances, the parental form of a drug has been shown to activate T cells via the T-cell receptor (TCR) in a major histocompatibility complex (MHC)-dependent way. Most of these studies have employed T-cell lines (TCL), T-cell clones (TCC) or TCR transfectants. To date, the list of drugs includes carbamazepine [3, 4], lidocaine [5, 6], lamotrigine [7], mepivacaine [6], norfloxacin [8, 9], p-phenylendiamine [10], and sulfamethoxazole (SMX) [11–13]. It is noteworthy that for most of these drugs, reactive metabolites are known as well. Consequently, carbamazepine [4], p-phenylendiamine [10], and SMX can act as typical (pro)haptens as well. SMX is a particularly instructive example since its hapten metabolite has been characterized extensively. SMX is first bioactivated into its hydroxylamine metabolite (SMX-NHOH) and then oxidized into the highly reactive metabolite SMX-nitroso (SMX-NO) [14, 15]. Nevertheless, a point has been made for the parental drug SMX to be the primary antigen: Glutathione as an antioxidant protects cells from reactive metabolites and prevents the oxidation of SMX-NHOH to SMX-NO [15]. Still, only a small minority of TCC derived from an allergic patient reacted to SMX-NO while most recognized the parental compound [16]. In line with this finding, the addition of glutathione reduced instead of increased T-cell proliferation in response to SMX-NO [17]. Two lines of independent evidence support the notion that antigen processing is not necessary for certain drugs in order to evoke a T-cell response. On one hand, glutaraldehyde-fixed

The p-i Concept: Evidence and Implications

APC, which are incapable of processing after treatment, are as efficient as untreated cells in inducing T-cell proliferation in response to carbamazepine [4], ciprofloxacin [8], lamotrigine [7], moxifloxacin [8], norfloxacin [9], and SMX [11– 13]. On the other hand, several cellular responses are much too fast to allow for prior antigen processing, such as the release of free intracellular calcium in response to SMX [12] (within seconds), TCR downregulation in response to carbamazepine [4] and SMX [12] (within 30 min), or ERK phosphorylation in response to SMX [13] (within 1 min). Pulse-chase experiments with several drugs indicate that the interaction of the drug with specific peptide-MHC (pMHC) molecules and TCR is not covalent since brief washing, usually with the medium used for incubation, was sufficient to abolish T-cell activation. To date, this list includes carbamazepine [4], ciprofloxacin [8], moxifloxacin [8], norfloxacin [9], and SMX [11, 13]. Of note, the same treatment did not have any effect when the hapten SMX-NO was used [11]. In summary, we are left with a comparably extensive list of drugs that clearly are not haptens but frequently elicit clinically relevant drug hypersensitivity nonetheless. In order to provide an explanation for the above observations, we have thus proposed the ‘pharmacological interaction with immune receptors’ concept (p-i concept) [18–20]. Below, we shall present the molecular and the systemic aspects of the p-i concept, review the evidence currently available in its support, and discuss its implications.

Molecular Aspects of the p-i Concept: Drug Interaction with MHC and TCR

The p-i concept states that drugs can interact by reversible (non-covalent) binding with the immune molecules required for the induction of a T-cell response, namely the pMHC complex and the TCR. From the evidence outlined above, it is

67

APC

APC

MHC

MHC Drug

Drug

TCR

TCR

T cell

T cell

APC

APC

APC

MHC MHC Drug

Drug

MHC Drug

TCR

TCR

T cell

T cell

TCR

T cell

APC

APC MHC

MHC

Drug

Drug

TCR TCR

T cell

Fig. 1. T-cell activation by reversible interaction of drug, MHC and TCR. In the top row, the drug first interacts with the MHC of the APC, the complex of which is then presented to the T cell by the APC. In the middle row, lowstrength interactions between MHC and TCR are stabilized by the drug, which bridges the MHC and TCR, and brings the T cell and the APC in close contact to one another. In the bottom row, the drug binds first to the TCR,

68

T cell

and the drug-TCR complex then interacts with the MHC. Note that in this case, drug-TCR interaction by itself might be sufficient to induce T-cell activation. The three models converge in the right column where full T-cell activation results. Light blue arrows designate molecular interactions, and orange arrows indicate transitions between different states during the differing activation processes.

Gerber ⴢ Pichler

clear that both of these molecules are involved: Even though drug-specific TCR are MHC-unrestricted more often than peptide-specific TCR [21], the great majority of drug-specific TCC and TCR transfectants characterized to date nevertheless need the expression of a particular human leukocyte antigen (HLA) allele in order to be activated. Likewise, the expression of a particular TCR determines the specificity and type of response, as recently evidenced for sulfanilamides such as SMX [13] and quinolones such as norfloxacin [9]: In both cases, transfection of TCR derived form drug-specific TCC conferred the same specificity and reactivity to the resulting transfectants. In general, reversible, non-covalent interactions of low molecular weight molecules with proteins are the rule rather than the exception since a plethora of drugs act in this way, regulating the activity of many cellular receptors and enzymes [22]. T-cell activation left aside, covalent interactions are in fact exceedingly rare in comparison, and it is worthwhile to remember that peptides bind to the groove of the MHC exclusively via many weak, reversible and non-covalent interactions. Accordingly, figure 1 depicts the three basic ways of how a trimolecular complex consisting of drug, pMHC and TCR may be formed. Costimulatory signals from the APC to the T cell are omitted for simplicity, but would principally occur in all three models. The first possibility is depicted in the top row. Here, the drug would first bind to the pMHC on the APC, and the ‘drug-loaded’ pMHC would then interact with the TCR. The TCR in turn, supplying a suitable fit for the drug-pMHC complex, would be activated and T-cell activation would result (right column). This scenario is different from normal antigen presentation only in that the antigen is binding from the outside to the pMHC. In fact, such a mechanism is well documented: The Ni-specific T-cell clone ANi-2.3, which has been characterized extensively by Weltzien and co-workers [23], is activated exactly

The p-i Concept: Evidence and Implications

in this way. In this particular case, the surface of the pMHC provided coordination sites for Ni that were strong enough to prevent simple washing from abolishing T-cell activation, documenting that not only covalent but also reversible interactions may be of considerable strength [24]. Of note, such a mechanism would also explain why certain hypersensitivity reactions, which are caused by several drugs that are not known to be haptens, show exceedingly strong associations with particular HLA alleles [25–28]. The second mechanism is depicted in the middle row, according to which the drug would stabilize low-affinity interactions between pMHC and TCR that per se are not sufficient to trigger TCR activation. Once the drug is bound to the pMHC and the TCR, it would bridge the two proteins, thereby bringing the APC and T cell in close enough contact for T-cell activation to ensue. It is important to point out that none of the interactions with either the pMHC or the TCR by themselves would suffice to stably ‘fix’ the drug antigen to either of the two proteins, but that the trimolecular complex would be stable enough due to a comparably large number of non-covalent interactions between the two proteins, and between the drug and either protein. Again, the work by Weltzien et al. furnishes experimental proof for the feasibility of such a mechanism. Ni cross-links the TCR of the TCC SE9 and of the respective TCR transfectant T913 to a conserved histidine side chain of various HLA-DR alleles [29]. TCR activation was processing-independent in this case as well, but it was sensitive to washing. This requirement for the continuous presence of Ni could be accounted for by the reduced number of coordination sites present on the pMHC in comparison to clone Ani-2.3 [23]. The third mode of action is shown in the bottom row. The drug would first engage the TCR, and the resulting drug-TCR complex would then interact with the MHC on the APC. An interesting twist of this notion is that TCR activation may occur not only upon encounter of the T cell with

69

the APC but in an APC-independent way as well, assuming that the drug itself is (somehow) able to activate the TCR by itself. In fact, there is precedence for such an MHC-independent TCR activation from several studies using different haptens. TCC reactive to p-azobenzenearsonate possessed binding sites specific for the antigen and were activated by arsanylated ovalbumin. Non-activating derivatives did not compete for binding, and binding occurred in the absence of MHC proteins as well [30, 31]. In another study, Jurkat cells devoid of endogenous MHC-II were transfected with an MHC-II-restricted TCR specific for fluorescein. These transfectants bound and were activated by fluorescein-coated beads and fluorescein conjugates while fluorescein monomers were inactive under the same conditions, indicating that TCR cross-linking was required [32]. In a similar way, a cytotoxic, trinitrophenyl-specific TCC became MHC-independent at high antigen doses but was MHC-restricted at low doses [33]. Hence, MHC-independent TCR activation has been documented even though the respective hapten antigens employ mechanisms that deviate from the one proposed here. In summary, one can thus imagine three nonconventional ways how T cells may be activated. None of them has been shown to apply for drugspecific T-cell activation, but there is experimental precedence for each of them nonetheless. For strong but non-covalent interactions with the MHC as well as for a superantigen-like crosslinking of MHC and TCR, this evidence is furnished by nickel, which is probably the most prevalent contact sensitizer known and thus a model system of high pathological relevance [34]. Somewhat artificial hapten antigens provide the support for direct TCR interaction and activation, which may even be APC- and pMHC-independent. If already such a limited range of antigens employs all of these unusual mechanisms, it should indeed not come as a surprise if all other allergenic compounds together do not restrict themselves to covalent attachment to the pMHC.

70

Systemic Aspects of the p-i Concept: Cross-Reactivity between Drug and Peptide Antigens

Most drug-induced hypersensitivity reactions that are T-cell-mediated do not occur directly upon initiation of treatment with the drug but upon a considerable delay only, which may stretch from a few days to several weeks. This observation is consistent with the induction of a primary T-cell response to the drug antigen, which requires the activation of the innate branch of the immune system. Covalently binding haptens and prohaptens may be recognized because of ‘altered self’ [2] according to the ‘self–nonself’ hypothesis [35], or because the reactive compound causes cellular damage [36] according to the danger hypothesis [37]. Also, several haptens have been shown to directly exert at least partial activation of APC such as monocytes and monocyte-derived dendritic cells [38–41]. In the frame of these two concepts, it is not directly evident how, at therapeutic and thus non-toxic concentrations, a non-reactive compound could provide this costimulatory signal except if such inert drugs had direct effects on APC as well, which is currently not known. However, there is an alternative possibility, as suggested by the observation that not all drug reactions need several days to manifest themselves. In the case of X-ray contrast media (RCM), T-cell-mediated allergic reactions have been reported to occur within hours of initiation [42]. Symptoms are usually moderate and confined to the skin [43]. Interestingly, reactions to RCM may even manifest themselves upon the first exposure ever to the agent, hence without previous sensitization [44]. Clearly, such reactions are much too fast to be explained by a primary immune response involving T cells, which requires several days to manifest itself. As discussed above, T-cell activation can be independent of antigen processing, and possibly even of antigen presentation. The rapid, delayed-

Gerber ⴢ Pichler

type reactions to RCM may in fact not require a primary immune response but represent the symptoms of a secondary immune response initiated by previously activated memory T cells [43]. These T cells would be specific for RCM but their sensitization and preactivation would have occurred in response to other, unknown peptide antigens. In other words, they would be peptidespecific and cross-reactive for the RCM. A compounding factor for such an activation would be the enormous amounts of RCM that are used, since volumes of up to 200 ml at concentrations of up to 1 M may be used depending on the type of examination. It is imaginable that even T cells with comparably low reactivity might be activated under such circumstances. Thus, preexisting memory T cells, which are known to generally be more reactive to TCR activation by weak antigens [45], might cross-react with a drug to cause hypersensitivity due to ‘accidental’ stimulation. In principle, such cross-reactivity seems feasible, as e.g. suggested by the high incidence of alloreactivity observed for drug-specific TCC [21]. Additionally, it should be taken into account that the activation and expansion of these reacting cells might require a significant amount of time, depending e.g. on drug concentrations and frequency of drug-cross-reactive T cells in the repertoire. Such an immunological response would thus become evident after a certain ‘lag phase’ only. Hence, a delayed onset of a drug hypersensitivity reaction is certainly compatible with the induction of a primary immune response, but not necessarily indicative of it. Three additional observations support this notion. For one, such a mechanism seems all the more possible since drug-specific T cells are even present in individuals who are not allergic and have never been exposed to the drug, as shown recently for SMX and SMX-NO [46]. On the other hand, general stimulation of T cells is known to increase the propensity for adverse drug reactions. IL-2 is a known risk factor in this regard, notably also for the occurrence of hypersensitiv-

The p-i Concept: Evidence and Implications

ity reactions to RCM in particular [42, 43]. Likewise, generalized viral infections, which induce a strong T-cell activation, can greatly increase the incidence of drug allergy, as evidenced e.g. for drugs used for antiviral therapy in HIV-infected subjects [47]. Finally, there are striking parallels between particular forms of drug hypersensitivity reactions such as drug rash with eosinophilia and systemic symptoms and immune stimulation by superantigens in terms of molecular mechanism as well as manifestations. In both cases, the antigen binds from the outside without being processed, and leads to a massive, systemic and longlasting immune activation [48].

Conclusions

In summary, we propose that drugs may evoke hypersensitivity reactions in ways that deviate drastically from the well-characterized hapten and prohapten mechanisms. In fact, they may behave like an allergenic molecule even though they completely circumvent the involvement of a primary response or of the innate branch of the immune system. Two peculiarities of such drug antigens would render this notion feasible: TCR reaction without antigen processing and antigen presentation, and the cross-specific activation of a pool of preexisting T cells. There is experimental evidence for many of these mechanisms although usually not (yet) from work on drug systems themselves. To date, the (pro)hapten concept has remained the prevailing paradigm to explain drug hypersensitivity, and current protocols for e.g. prediction of drug allergenicity focus largely on chemical reactivity and metabolic conversion of known and novel compounds. In fact, drug candidates that prove to be reactive during research and development are usually not pursued further, and functional groups that are reactive are avoided altogether. It remains absolutely undisputed that

71

numerous drugs act as haptens and prohaptens, and that even more will be discovered to do so. Nevertheless, chemical reactivity should no longer be considered the sole defining hallmark of allergy-inducing compounds. Even though the p-i concept deviates considerably from the (pro)hapten, the self–nonself and the danger hypotheses, none of these models are mutually exclusive, and none by itself is at current sufficient to explain all aspects of drug hy-

persensitivity. Clearly, not all drugs or classes of drugs are equally prone to induce an allergic reaction, and not all individuals are alike in their propensity to suffer from drug hypersensitivity. Somewhat unfortunately, the most likely proposition seems to be that the combinations of different immunological, toxicological and genetic factors that contribute to drug hypersensitivity will be as diverse as the chemical components that elicit them.

References 1 Landsteiner K, Jacobs J: Studies on the sensitization of animals with simple chemical compounds. J Exp Med 1935; 61:643–656. 2 Naisbitt DJ, Williams DP, Pirmohamed M, Kitteringham NR, Park BK: Reactive metabolites and their role in drug reactions. Curr Opin Allergy Clin Immunol 2001;1:317–325. 3 Naisbitt DJ, Britschgi M, Wong G, Farrell J, Depta JP, Chadwick DW, Pichler WJ, Pirmohamed M, Park BK: Hypersensitivity reactions to carbamazepine: characterization of the specificity, phenotype, and cytokine profile of drugspecific T cell clones. Mol Pharmacol 2003;63:732–741. 4 Wu Y, Sanderson JP, Farrell J, Drummond NS, Hanson A, Bowkett E, Berry N, Stachulski AV, Clarke SE, Pichler WJ, Pirmohamed M, Park BK, Naisbitt DJ: Activation of T cells by carbamazepine and carbamazepine metabolites. J Allergy Clin Immunol 2006;118:233– 241. 5 Zanni MP, Mauri-Hellweg D, Brander C, Wendland T, Schnyder B, Frei E, von Greyerz S, Bircher A, Pichler WJ: Characterization of lidocaine-specific T cells. J Immunol 1997;158:1139–1148. 6 Zanni MP, von Greyerz S, Hari Y, Schnyder B, Pichler WJ: Recognition of local anesthetics by + T cells. J Invest Dermatol 1999; 112:197–204. 7 Naisbitt DJ, Farrell J, Wong G, Depta JP, Dodd CC, Hopkins JE, Gibney CA, Chadwick DW, Pichler WJ, Pirmohamed M, Park BK: Characterization of drug-specific T cells in lamotrigine hypersensitivity. J Allergy Clin Immunol 2003;111:1393–1403.

72

8 Schmid DA, Depta JP, Pichler WJ: T cell-mediated hypersensitivity to quinolones: mechanisms and cross-reactivity. Clin Exp Allergy 2006;36:59–69. 9 Schmid DA, Depta JP, Luthi M, Pichler WJ: Transfection of drug-specific T-cell receptors into hybridoma cells: tools to monitor drug interaction with T-cell receptors and evaluate cross-reactivity to related compounds. Mol Pharmacol 2006;70:356–365. 10 Sieben S, Kawakubo Y, Al Masaoudi T, Merk HF, Blomeke B: Delayed-type hypersensitivity reaction to p-phenylenediamine is mediated by two different pathways of antigen recognition by specific  human T-cell clones. J Allergy Clin Immunol 2002;109:1005– 1011. 11 Schnyder B, Mauri-Hellweg D, Zanni M, Bettens F, Pichler WJ: Direct, MHCdependent presentation of the drug sulfamethoxazole to human  T cell clones. J Clin Invest 1997;100:136–141. 12 Zanni MP, von Greyerz S, Schnyder B, Brander KA, Frutig K, Hari Y, Valitutti S, Pichler WJ: HLA-restricted, processing- and metabolism-independent pathway of drug recognition by human  T lymphocytes. J Clin Invest 1998; 102:1591–1598. 13 Depta JP, Altznauer F, Gamerdinger K, Burkhart C, Weltzien HU, Pichler WJ: Drug interaction with T-cell receptors: T-cell receptor density determines degree of cross-reactivity. J Allergy Clin Immunol 2004;113:519–527. 14 Cribb AE, Spielberg SP: Sulfamethoxazole is metabolized to the hydroxylamine in humans. Clin Pharmacol Ther 1992;51:522–526.

15 Naisbitt DJ, Vilar FJ, Stalford AC, Wilkins EG, Pirmohamed M, Park BK: Plasma cysteine deficiency and decreased reduction of nitrososulfamethoxazole with HIV infection. AIDS Res Hum Retroviruses 2000;16:1929–1938. 16 Schnyder B, Burkhart C, SchnyderFrutig K, von Greyerz S, Naisbitt DJ, Pirmohamed M, Park BK, Pichler WJ: Recognition of sulfamethoxazole and its reactive metabolites by drug-specific CD4+ T cells from allergic individuals. J Immunol 2000;164:6647–6654. 17 Burkhart C, von Greyerz S, Depta JP, Naisbitt DJ, Britschgi M, Park KB, Pichler WJ: Influence of reduced glutathione on the proliferative response of sulfamethoxazole-specific and sulfamethoxazole-metabolite-specific human CD4+ T-cells. Br J Pharmacol 2001;132:623–630. 18 Pichler WJ: Pharmacological interaction of drugs with antigen-specific immune receptors: the p-i concept. Curr Opin Allergy Clin Immunol 2002;2: 301–305. 19 Gerber BO, Pichler WJ: Cellular mechanisms of T cell mediated drug hypersensitivity. Curr Opin Immunol 2004; 16:732–737. 20 Gerber BO, Pichler WJ: Noncovalent interactions of drugs with immune receptors may mediate drug-induced hypersensitivity reactions. AAPS J 2006; 8:E160–165. 21 Von Greyerz S, Bultemann G, Schnyder K, Burkhart C, Lotti B, Hari Y, Pichler WJ: Degeneracy and additional alloreactivity of drug-specific human (+) T cell clones. Int Immunol 2001;13: 877–885.

Gerber ⴢ Pichler

22 Beeley NR: Can peptides be mimicked? Drug Discov Today 2000;5:354–363. 23 Lu L, Vollmer J, Moulon C, Weltzien HU, Marrack P, Kappler J: Components of the ligand for a Ni++ reactive human T cell clone. J Exp Med 2003;197:567– 574. 24 Vollmer J, Weltzien HU, Gamerdinger K, Lang S, Choleva Y, Moulon C: Antigen contacts by Ni-reactive TCR: typical  chain cooperation versus  chain-dominated specificity. Int Immunol 2000;12:1723–1731. 25 Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, Sayer D, Castley A, Mamotte C, Maxwell D, James I, Christiansen FT: Association between presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet 2002;359:727–732. 26 Martin AM, Nolan D, Gaudieri S, Almeida CA, Nolan R, James I, Carvalho F, Phillips E, Christiansen FT, Purcell AW, McCluskey J, Mallal S: Predisposition to abacavir hypersensitivity conferred by HLA-B*5701 and a haplotypic Hsp70-Hom variant. Proc Natl Acad Sci USA 2004;101:4180–4185. 27 Chung WH, Hung SI, Hong HS, Hsih MS, Yang LC, Ho HC, Wu JY, Chen YT: Medical genetics: a marker for StevensJohnson syndrome. Nature 2004;428: 486. 28 Hung SI, Chung WH, Liou LB, Chu CC, Lin M, Huang HP, Lin YL, Lan JL, Yang LC, Hong HS, Chen MJ, Lai PC, Wu MS, Chu CY, Wang KH, Chen CH, Fann CS, Wu JY, Chen YT: HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci USA 2005;102: 4134–4139. 29 Gamerdinger K, Moulon C, Karp DR, Van Bergen J, Koning F, Wild D, Pflugfelder U, Weltzien HU: A new type of metal recognition by human T cells: contact residues for peptide-independent bridging of T cell receptor and major histocompatibility complex by nickel. J Exp Med 2003;197:1345–1353.

30 Rao A, Ko WW, Faas SJ, Cantor H: Binding of antigen in the absence of histocompatibility proteins by arsonate-reactive T-cell clones. Cell 1984;36: 879–888. 31 Rao A, Faas SJ, Cantor H: Analogs that compete for antigen binding to an arsonate-reactive T-cell clone inhibit the functional response to arsonate. Cell 1984;36:889–895. 32 Diamond DJ, Szalay P, Symer D, Hao P, Shin HS, Dintzis RZ, Dintzis HM, Reinherz EL, Siliciano RF: Major histocompatibility complex independent T cell receptor-antigen interaction: functional analysis using fluorescein derivatives. J Exp Med 1991;174:229–241. 33 Stockl J, Majdic O, Fischer G, Maurer D, Knapp W: Monomorphic molecules function as additional recognition structures on haptenated target cells for HLA-A1-restricted, hapten-specific CTL. J Immunol 2001;167:2724–2733. 34 Thierse HJ, Gamerdinger K, Junkes C, Guerreiro N, Weltzien HU: T cell receptor interaction with haptens: metal ions as non-classical haptens. Toxicology 2005;209:101–107. 35 Janeway CA Jr: The immune system evolved to discriminate infectious nonself from noninfectious self. Immunol Today 1992;13:11–16. 36 Seguin B, Uetrecht J: The danger hypothesis applied to idiosyncratic drug reactions. Curr Opin Allergy Clin Immunol 2003;3:235–242. 37 Matzinger P: Tolerance, danger, and the extended family. Annu Rev Immunol 1994;12:991–1045. 38 Aiba S, Manome H, Nakagawa S, Mollah ZU, Mizuashi M, Ohtani T, Yoshino Y, Tagami H: p38 mitogen-activated protein kinase and extracellular signalregulated kinases play distinct roles in the activation of dendritic cells by two representative haptens, NiCl2 and 2,4dinitrochlorobenzene. J Invest Dermatol 2003;120:390–399.

39 Bruchhausen S, Zahn S, Valk E, Knop J, Becker D: Thiol antioxidants block the activation of antigen-presenting cells by contact sensitizers. J Invest Dermatol 2003;121:1039–1044. 40 Becker D, Valk E, Zahn S, Brand P, Knop J: Coupling of contact sensitizers to thiol groups is a key event for the activation of monocytes and monocyte-derived dendritic cells. J Invest Dermatol 2003;120:233–238. 41 Hulette BC, Ryan CA, Gildea LA, Gerberick GF: Relationship of CD86 surface marker expression and cytotoxicity on dendritic cells exposed to chemical allergen. Toxicol Appl Pharmacol 2005;209:159–166. 42 Christiansen C, Pichler WJ, Skotland T: Delayed allergy-like reactions to X-ray contrast media: mechanistic considerations. Eur Radiol 2000;10: 1965–1975. 43 Christiansen C: Late-onset allergy-like reactions to X-ray contrast media. Curr Opin Allergy Clin Immunol 2002;2: 333–339. 44 Christiansen C: X-ray contrast media – an overview. Toxicology 2005;209:185– 187. 45 Amrani A, Serra P, Yamanouchi J, Trudeau JD, Tan R, Elliott JF, Santamaria P: Expansion of the antigenic repertoire of a single T cell receptor upon T cell activation. J Immunol 2001;167: 655–666. 46 Engler OB, Strasser I, Naisbitt DJ, Cerny A, Pichler WJ: A chemically inert drug can stimulate T cells in vitro by their T cell receptor in non-sensitised individuals. Toxicology 2004;197: 47–56. 47 Pirmohamed M, Park BK: HIV and drug allergy. Curr Opin Allergy Clin Immunol 2001;1:311–316. 48 Lerch M, Pichler WJ: The immunological and clinical spectrum of delayed drug-induced exanthems. Curr Opin Allergy Clin Immunol 2004;4:411–419.

Dr. Basil O. Gerber Division of Allergology, Clinic and Policlinic for Rheumatology and Clinical Immunology/Allergology, Inselspital CH–3010 Bern (Switzerland) Tel. +41 31 632 4233, Fax +41 31 632 3547, E-Mail [email protected]

The p-i Concept: Evidence and Implications

73

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 74–83

Tolerance Mechanisms to Small Molecular Compounds Andrea Cavani
Ornella De Pità Laboratory of Immunology and Allergology, IDI-IRCCS, Rome, Italy

Abstract Unbalanced immune responses to small molecular weight compounds lead to a variety of diseases, including allergic contact dermatitis to skin sensitizers and systemic drug hypersensitivity, both characterized by the uncontrolled expansion of cytotoxic CD8+ and CD4+ T lymphocytes. Evidence exists that specialized subsets of T lymphocytes with regulatory function modulate immune responses to chemicals by preventing the occurrence of the hypersensitivity reactions in healthy, non-allergic individuals. In addition, regulatory mechanisms may have a role in dampening the magnitude of the immune response in allergic subjects to avoid excessive tissue damage. Dendritic cells orchestrate immune responses to chemicals. Depending on their maturative state they can expand effector lymphocytes responsible for hypersensitivity reactions, or interleukin-10-releasing and CD4+CD25+ T lymphocytes, responsible for the establishment of specific tolerance to the potential sensitizer. The understanding of the underlying mechanisms of tolerance induction is a critical step for the development of therapeutical strategies for delayed hypersensitivity reactions. Copyright © 2007 S. Karger AG, Basel

skin and the gut are equipped with efficient defensive mechanisms aimed to protect the organisms from the entry of microorganisms and toxic substances [1]. In most cases, optimized patterns of recognition guarantee protection against the danger, while preventing undesired responses towards innocuous substances [2]. However, in certain circumstances, strong immune responses against non-pathogenic stimuli may develop, causing remarkable systemic or local damage. Maintenance of specific tolerance to innocuous chemicals is a priority for the immune system. The task is guaranteed by multiple mechanisms, including T-cell anergy, release of anti-inflammatory cytokines, and induction of specialized subsets of T lymphocytes with regulatory function [3]. The delicate equilibrium between responsiveness and unresponsiveness involves the fundamental interaction between innate and adaptive immune responses, in particular dendritic cells (DCs).

Introduction

The skin and other boundary organs are frequently penetrated by chemicals either as a consequence of accidental exposure to environmental low molecular weight substances or following therapeutic administration of drugs. Both the

Animal Models of Tolerance to Chemicals

Murine contact hypersensitivity (CH) has become the most valuable model for the study of effector and regulatory mechanisms of immune

responses to chemicals. Consequently, investigations have focused mainly on contact sensitizers, rather than on systemically administered drugs [4]. CH can be induced by the epicutaneous application of the sensitizers onto the shaved abdomen of the mouse and quantified by measuring the ear thickness 24 and 48 h after the ear challenge with the relevant substance. In this scenario, UV-B irradiation at the skin site of chemical exposure not only induces failure of sensitization, but also promotes a specific immune tolerance which prevents subsequent sensitization to the hapten [5]. Similarly, a specific tolerance to chemicals can be induced by orally feeding the animal with the sensitizer prior to skin exposure [6]. UV-B-induced and oral tolerance are clearly antigen-specific and not the consequence of a generalized immune suppression, since the animals fully maintain the responsiveness to unrelated antigens. Unresponsiveness can be adoptively transferred into naive recipients with spleen or lymph node cell suspensions, thus suggesting a role of T cells in the tolerogenic mechanism. These findings demonstrate that the immune system possesses the intrinsic capacity to regulate responses to chemicals, and that tolerance can be boosted by specific experimental procedures. Recent advances indicate that both UVB-induced and oral tolerance depends on the expansion of specialized subsets of T cells with regulatory function belonging to the CD4+ population [7, 8]. Indeed, oral tolerance can be prevented by depleting the animals of CD4+ T lymphocytes, and fails in mice lacking MHC class II molecules, in which the development of mature CD4+ T lymphocytes is greatly impaired [9]. More than one subset of T-regulatory (Treg) cells have been found to have a role in the induction of oral tolerance: TGF--secreting T cells, named Th3 lymphocytes; interleukin (IL)-10/TGF--releasing T regulatory 1 (Tr1) cells, and, more recently, CD4+CD25+ Treg cells [8]. Consensus exists that cytokines, either IL-10 or TGF-, are involved in the mechanism. Oral tolerance re-

Tolerance Mechanisms to Small Molecular Compounds

quires an intact DC recirculation and is impaired in CCR7-deficient mice [10]. CCR7 is a critical chemokine receptor that targets DCs residing in the lamina propria to the mesenteric lymph nodes, where Treg are expanded. An attractive hypothesis suggests that the expansion of Treg in oral tolerance depends on the presence of invariant NK-T cells, a distinct lymphoid lineage expressing an invariant T-cell receptor (TCR) recognizing glycolipids in a CD1d-restricted fashion. Indeed, oral tolerance to nickel fails in invariant NK-T-cell-deficient Ja18–/– mice [11]. Since the mid-1980s, many experimental approaches have clearly demonstrated the capacity of UV-B to impair skin immune responses. Mechanisms involved in UV-B-induced immunosuppression are multiple. UV-B irradiation depletes the skin of Langerhans cells (LCs) by promoting their apoptosis and/or migration to regional lymph nodes [12]. Furthermore, UV-B irradiation impairs the antigen-presenting capacity of LCs as a result of reduced MHC class II, CD80 and CD86 expression. Second, UV-B induces DNA damage in DCs [13, 14]. Several reports have demonstrated that IL-12 reverts UVB-induced tolerance by enhancing DNA repair in skin LCs and reverts the suppressive effect of cis-urocanic acid, which accumulates in the epidermis upon UV-B irradiation and impairs LC functions [13]. IL-18 mimics some of the effects of IL-12 [15]. Finally, IL-10, released by keratinocytes upon UV-B exposure, can further reduce the antigen-presenting capacity of LCs, rendering them tolerogenic. UV-B tolerance can be transferred into naive recipients with CD4+ CD25+ Treg expressing high levels of cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) from UV-B-irradiated syngeneic animals. In this model, tolerance is reverted by blocking the CTLA-4 antigen with monoclonal antibodies [14]. Importantly, although tolerogenic CD4+ CD25+ Treg can prevent induction of CH in naive animals, they fail to impede CH expression in already sensitized mice. This finding may have

75

Fig. 1. Mechanisms of UV-B-induced and oral tolerance. Hapten application onto UV-B-irradiated skin induces DNA damage and apoptosis of DCs, release of cis-urocanic acid and IL-10. Altered maturation of skin DCs migrating in regional lymph nodes is responsible for the expansion of IL-10-releasing CD4+CD25+ Treg. Chemical entry

in the gut is followed by the interaction of the sensitizer with DCs residing in the lamina propria. Chemical-loaded DCs migrate in mesenteric lymph nodes and induce CD4+CD25+ Treg. NK-T cells provide critical signals for Treg expansion, including IL-10.

relevant consequences for the design of immunomodulatory therapy in humans, and could depend on the tissue recirculation of UV-B-induced Treg, which express the L-selectin for the lymph node homing but not the cutaneous lymphocyte-associated antigen (CLA). Although canonic CD4+CD25+ Treg do not release IL-10, Treg induced by UV-B release abundant IL-10, that, together with host-derived IL-10, contributes significantly to the immunosuppressive mechanisms [16] (fig. 1). An additional model of tolerance to chemicals is the low zone tolerance (LZT), which is induced by the repeated application onto the skin of a

non-sensitizing amount of the substance, and is regarded as the most common mechanism of tolerance to hapten in humans. Development of LZT requires IL-10 and appears mediated by CD8+ Treg cells showing a Tc2 pattern of cytokine release, including IL-4 and IL-10 [17].

76

T-Regulatory Cells in Chemical Tolerance: Phenotype and Functions

The hypothesis of suppressor T cells capable of controlling immune responses was formulated about 25 years ago on the basis of experimental

Cavani ⴢ De Pità

evidence. However, the technical difficulty in defining regulatory T-cell populations considerably slowed down the research in the field. Rather, the identification of two dichotomic T-cell subsets, the T-helper (Th)1 and Th2 cells, capable of distinct and often mutually exclusive functions, led to the idea that regulation of immune responses was predominantly due to the balance between Th1 and Th2 cytokines. This approach had some success, and still has estimators, possibly due to the fact that IL-10, a critical regulatory cytokine, is exclusively associated to the Th2 development in the mouse, but not in humans. In the last few years the concept of immunomodulation has been completely revised, according to new advances in the phenotypic and functional characterization of Treg. In the mid-1990s the laboratory of Roncarolo and that of Sagakuchi provided evidence of the existence of at least two distinct Treg subsets, the Tr1 cells and the CD4+CD25+ Treg. Tr1 cells, characterized by high IL-10 production, were generated in vitro through multiple stimulation in the presence of IL-10 from both human blood and mouse spleen cells [18]. Tr1 cells were suppressive in vitro and prevented the development of experimentally induced colitis in the mouse. The suppressive activity of Tr1 mostly depends on IL-10, which is secreted abundantly upon TCR engagement. IL-10 strongly inhibits the antigen-presenting function of macrophages and DCs, as well as the differentiation and IL-12 release of monocyte-derived DCs. IL-10 is a critical regulatory cytokine in immune responses to chemicals. In the mouse, administration of IL-10 prior to hapten challenge strongly impairs the effector phase of CH in the mouse [19, 20]. Conversely, mouse with targeted disruption of the IL-10 gene displayed a strongly enhanced cutaneous immune response to chemicals [21]. In humans, IL-10-releasing nickel-specific Tr1 cells are enriched in the peripheral blood of non-allergic individuals, suggesting a role of Tr1 cells in the control of immune responses to nickel [22]. In addition, nickel-spe-

Tolerance Mechanisms to Small Molecular Compounds

cific Tr1 cells can also be isolated from an ongoing CH reaction to nickel in allergic subjects, indicating that they can be recruited at the site of hapten exposure and could modulate ongoing immune responses to chemicals. This hypothesis is further confirmed by experiments in the mouse showing that blocking IL-10 augments the magnitude and the duration of the CH reaction to skin sensitizers [19]. Although Tr1 cells have been widely characterized phenotypically and functionally, their origin is still debated. Their phenotype resembles that of highly differentiated memory T cells with markers which belong to both the Th1 and Th2 differentiation, such as CD30, LAG, and the IL12R2 chain [22]. Chemokine receptor analysis of Tr1 cells revealed the capacity of these cells to respond to a vast array of chemokines. They share with Th2 cells the expression of CCR8 and the responsiveness to CCL1 [23]. Notably, mRNA of CCL1 appears late during CH reactions, produced by maturing DCs, activated T cells and keratinocytes. Interestingly, effector/memory Th1 cells rendered anergic by incomplete activation have been found to release abundant IL-10 and to possess a suppressive function [24]. Thus, the hypothesis of Tr1 cells as a sort of terminal differentiation of Th1 lymphocytes, rather than a distinct T-cell lineage, could not be excluded. CD4+CD25+ Treg were first identified in the mouse as a distinct T-cell lineage that originates in the thymus after being positively selected by thymic epithelial cells [25]. The major function of naturally occurring CD25+ Treg is to maintain the peripheral tolerance to self-antigens. Both in the mouse and humans, the CD4+CD25+ population represents about 5–10% of circulating CD4+ cells. Among these, Treg appears to be included in the CD25high fraction. Additional markers for CD25+ Treg are the expression of L-selectin, which allows their recirculation through the high endothelial venules in the secondary lymphoid organs, the high levels of CTLA-4, and the expression of the glucocorticoid-induced TNFreceptor family-related gene (GITR). CTLA-4 is a

77

CD28 analogue which binds to CD80 and CD86 and negatively regulates cell activation and IL-2 release. Characteristically, CD25+ Treg express high levels of the transcriptional factor Foxp3, which encodes a transcriptional factor named scurfin, and has been positively correlated with the development of CD25+ T cells, both in the thymus and at the periphery [26]. In the mouse, Foxp3 confers regulatory activity when transduced into T cells. The role of Foxp3 in the development of CD25+ Treg has been confirmed in humans: mutation of Foxp3 induces an autoimmune lymphoproliferative disorder termed the immune dysregulation polyendocrinopathy enteropathy-X-linked (IPEX) syndrome characterized by neonatal autoimmune type 1 diabetes, polyendocrinopathy, autoimmune hemolytic anemia, autoimmune enteropathy and eczema [27]. Clinical manifestations are the consequence of the uncontrolled T-cell proliferation and activation. CD25+ Treg display a broad usage of the TCR V repertoire and have the characteristics of highly differentiated T cells, indicating a repetitive antigen encounter and stimulation. Indeed, the CD25+ Treg population renovates continuously in vitro, but they behave as anergic cells in vitro. This characteristic limits considerably our capacity to determine their antigen specificity. Beside naturally occurring CD25+ Treg involved in the peripheral control of autoreactive T cells that escaped from thymic selection, adaptive CD25+ Treg with a broad antigenic specificity have been demonstrated to expand following an encounter with exogenous antigens. These induced Treg are involved in the control of immune responses against environmental antigens, including bacteria, fungi, protozoa, allergenic molecules and chemicals [28]. In rodents, CD25+ Treg regulate both UV-B-induced and oral tolerance, as well as drug-induced autoimmunity. In this latter model, procainamide-induced antinuclear IgG antibodies synthesis has been shown to be suppressed by CD4+CD25+ Treg [29]. In human beings, it has been shown that CD4+CD25+

78

T cells from peripheral blood of healthy, non-allergic individuals can regulate in vitro the activation of both naive and memory nickel-specific Tcell responses. In comparison, CD4+CD25+ T cells isolated from nickel-allergic subjects showed a limited or absent capacity to regulate CD4+ and CD8+ nickel-specific T cells [30]. CD4+ CD25+ Treg have a pattern of migratory behavior similar to that of Tr1 cells, both expressing high levels of CCR4 and CCR8, receptors for CCL17 and CCL1, indicating that common mechanisms govern the recruitment of distinct Treg subsets in peripheral tissues [31]. In addition, a conspicuous fraction of circulating CD4+CD25+ T cells co-express the CLA, and can be recruited at the site of skin inflammation. According to this hypothesis, skin biopsies at the site of nickel sulfate application in non-allergic individuals displayed a conspicuous T-cell infiltrate, mostly CD4+, which included CD25+ T cells [30]. Once isolated, these infiltrating CD4+CD25+ T cells strongly suppressed in vitro nickel-specific T-cell responses. Thus, CD4+ CD25+ Treg could block T-cell activation to chemicals through multiple pathways: at the site of chemical entry, by dampening the activation of chemical-reactive effector CD8+ T cells, and in secondary lymphoid organs, where they can prevent the expansion of memory-effector CD4+ and CD8+ T cells. In addition, CD4+CD25+ Treg have been shown to inhibit expression of CXCR3 chemokine receptor by differentiating T cells, the latter being crucial for T-cell homing into the skin during chemical hypersensitivity reactions [32]. Finally, it has been shown that CD25+ Treg facilitate the induction of IL-10 producing Tr1-like cells with an anergic/suppressive phenotype [33]. In aggregate, these data indicate that development of exaggerated immune responses to chemicals is the consequence of a defective or altered expansion and/or functional suppressive activity of specific Treg (table 1). Mechanisms of suppression of CD4+CD25+ Treg have not been completely elucidated. In vi-

Cavani ⴢ De Pità

Table 1. Major Treg subsets Treg cell

Identification markers

Mechanism of suppression

Function

Tr1 cells

Characterized by the pattern of cytokine release: IL-10+ IL-5+ IL-4– and IFN-–

IL-10-dependent

Autoimmunity, cancer, infections and contact hypersensitivity reactions

CD4+CD25+

Foxp3, LAG-3, CTLA-4+, GITR+, L-selectin+

Cell-to-cell contact; tryptophan deprivation; perforin/granzymemediated apoptosis

Tolerance to self-antigens, microbial antigens, chemicals, tumor antigens and graft rejection

Th3

No specific markers

TGF--dependent

Autoimmunity

CD8+CD28–

No specific markers

Impair APCs and induce ILT3/4 in DCs

Autoimmunity, graft rejection

CD8 Tc2

No specific markers; IL-10 and IL-4 secretion

IL-10/IL-4-mediated

Low zone tolerance to haptens

NK-T cell

CD3+CD4+CD161+

IL-10, Fas/Fas L-mediated apoptosis

Oral tolerance

tro analysis has determined that CD25+ Treg are anergic cells, unless high IL-2 is provided. CD25+mediated suppression requires activation via TCR, although, once activated, CD25+ Treg can regulate antigen-unrelated CD4+ and CD8+ T lymphocytes. In vitro, the mechanism of suppression appears to be independent of IL-10 and TGF- release, and requires cell-to-cell contact between the Treg and the suppressed T cell. However, some in vivo models stress the role of either IL-10 or TGF- as a major suppressor mechanism of CD25+ Treg. A putative role of CTLA-4 in the suppressive activity of Treg has been suggested by many reports. In this model, CTLA-4 exposed to CD25+ Treg may interact with CD80 and CD86 expressed on activated T cells. CD80 and CD86 ligation may induce an outside-in signal that blocks IL-2 production [34]. In addition, CTLA-4 interaction with CD80 and CD86 expressed by DCs induces indoleamine 2,3-dioxygenase, an enzyme involved in tryptophan catabolism. Following this model, T-cell suppression may be the consequence of tryptophan deprivation [35]. It has been shown that CD25+ Treg express high

Tolerance Mechanisms to Small Molecular Compounds

levels of perforin and granzyme A. Induction of apoptosis of target T cells has been proposed as an additional mechanism of suppression of CD25+ Treg [36]. It is clear that suppression operated by Treg can occur with multiple mechanisms depending on the in vivo or in vitro situations. The relationship between the diverse Treg cell populations remains poorly defined. The current view is that CD25+ Treg and Tr1 cells are distinct cell populations which cooperate to ensure proper regulation of immune responses. The presence to redundant mechanisms indicates that tolerance to self as well as exogenous antigens is a critical task for the immune system. Although most of the Treg cell populations described so far belong to the CD4+ T-cell subset, reports showing similar roles performed by CD8+ T cells are increasing. For example, IL-10/ IL-4-releasing CD8+ T cells have been reported to expand following repeated skin administration of subimmunogenic amounts of chemicals and mediate the development of LZT [37]. CD8+ T cells with a pattern of cytokine expression similar to Tr1 lymphocytes have been shown to ex-

79

pand upon stimulation of naive CD8+ T cells with plasmacytoid DCs [38]. In addition, a subpopulation of CD8+CD28– T suppressor cells has been described to regulate immune responses by inducing the expression of the inhibitory molecules immunoglobulin-like transcript (ILT)-3 and ILT-4 on DCs, which thus acquire tolerogenic activity [39]. Their role in the regulation of chemical hypersensitivity has not been investigated so far.

Breaking Tolerance to Chemicals

The immune system has the intrinsic capacity to develop either protective immune responses or specific tolerance. Most of the individuals exposed to skin sensitizers or to drugs do not develop an immune-mediated adverse reaction, indicating that induction of tolerance to chemicals is the standard response. However, circumstances leading to the failure of the tolerogenic mechanisms may result in uncontrolled hypersensitivity response to the chemical. Both in the case of CH to haptens and drug hypersensitivity, the severity of the disease correlates with the intervention of cytotoxic CD8+ T-cell responses. Since exposed individuals can develop an allergy to chemicals at any time during their lifetime, immune responses to chemicals must be continuously adjusted and modulated. DCs are the major orchestrators of the dynamic interaction between innate and adaptive immune responses, being capable of inducing either protective immune responses or unresponsiveness, depending on their functional status [40]. Immature DCs reside at boundary tissues, where antigen encounter is more frequent. Upon antigen encounter, DCs migrate to regional lymph nodes to present the antigenic determinants to naive T cells [41]. Their maturation and mobilization is strongly promoted by the exposure to a wide spectrum of agents generally named ‘danger signals’, which includes UV-B radiation,

80

bacteria, CpG-DNA, viral products, necrotic cells, and extracellular matrix components, such as heparan sulfate [42]. DCs provide two distinct signals to naive T lymphocytes: (i) the antigenic determinant bound to surface MHC molecules, and (ii) accessory signals, named co-stimulatory molecules, which are poorly expressed on immature DCs and strongly upregulated during DC maturation. The level of DC maturation state and the cytokine milieu at the site of antigen encounter strongly affect the future properties of memory-effector T cells. Differentiation of effector Th1 and Tc1 cells, responsible for CH reactions, is strongly promoted by IL-12, which is released by fully mature and not immature DCs. In contrast, antigen presentation by immature or partially mature DCs, which have migrated in secondary lymphoid organs in steady-state conditions, often leads to unresponsiveness and/or immune tolerance, due to the expansion of T cells with regulatory functions [43]. Thus, the concomitance of danger signals during chemical exposure may be the discriminating factor for allergy versus tolerance [44]. The danger signal hypothesis may explain the high frequency of adverse drug reactions in boundary organs. Indeed, in spite of the fact that the liver is the major detoxification organ, most of the ADR take place in the skin, which is equipped with highly efficient DCs, LCs and dermal DCs. To further confirm this hypothesis, it has been recently shown that T cells activated by DCs whose TLRs were triggered by LPS or CpG are resistant to suppression mediated by CD25+ Treg, with a mechanism at least partially mediated by IL-6 [45]. Thus, it may be hypothesized that concomitant bacterial entry or subinfection may be a relevant cofactor in breaking tolerance to skin-applied chemicals. Together with the increased amount of chemical entry, this may account for the increased risk of sensitization when topical drugs are applied onto chronic inflamed skin, and the increased frequency of nickel allergy in women wearing earrings, but not orthodontic braces [46].

Cavani ⴢ De Pità

Perspectives: Tolerance Induction and Tolerance Restoration

The explosive interest in Treg mostly depends on their supposed clinical application in the prevention and treatment of immune-mediated diseases. Thus, a major goal in the field of chemical hypersensitivity is the definition of strategies aimed at expanding antigen-specific Treg both in vitro and in vivo. In vitro stimulation of naive CD25– T lymphocytes with mature DCs in the presence of chemical antigens, such as TNCB or nickel, induces a small percentage of CD25+ T cells expressing Foxp3 that behave as regulatory cells in in vitro assays [pers. unpubl. observation]. However, the percentage of Foxp3+ Treg obtained from naive T cells greatly increases when the antigen is presented by immature or partially mature DCs [43, 47]. This finding may suggest that although a small percentage of Treg are commonly generated during antigen presentation [48], their differentiation can be augmented when the chemical is presented by DCs modified by cytokines or pharmaceutical compounds, or in the absence of danger signals. Several reports have demonstrated that TGF has a critical role in the induction of CD4+ CD25+ Treg in vitro. Indeed, in the presence of TGF-, naive CD4+CD25– T cells stimulated with immature DCs or with anti-CD3 Ab rapidly convert to a CD4+CD25+ phenotype. TGF--induced Treg express Foxp3 and display potent immunosuppressive activity both in vitro and in vivo [49]. Modulation of DC maturation by cholera toxin [50], 1,25-dihydroxyvitamin D3 [51], mycophenolate mofetil, anti-TNF Ab [52] and dietary antioxidants (-tocopherol and vitamin C) [53] have been reported to increase the yield of Treg upon in vitro stimulation of CD4+CD25– T cells. Whether distinct DC lineages are selectively involved in the generation of Treg is a still debated and attractive hypothesis. Several reports indicate that plasmacytoid DCs, a subset of lym-

Tolerance Mechanisms to Small Molecular Compounds

phoid DC secreting a high amount of type 1 IFN upon exposure to microbial products, preferentially generate either CD8+ or CD4+ IL-10 producing Foxp3+ Treg once matured in the presence of TLR9 ligands or upon CD40 blockade [38, 54, 55]. In addition to signals mediated by DCs, TLR ligands may have a direct effect on CD4+CD25+ T-cell functions. In particular, TLR2 and TLR8 agonists revert the anergic status of Treg and transiently block their suppressive activity [56, 57]. In vivo preliminary results in human beings appear promising. Oral tolerization to hydroxychloroquine was successful in a group of sensitive patients [58] and daily oral administration of nickel in allergic individuals has been reported to decrease significantly the magnitude of the CH [59]. Further experiments are certainly required to elucidate the mechanisms involved in these in vivo desensitization protocols.

Conclusions

Studies aimed at characterizing the mechanisms involved in the regulation of immune responses have advanced since Treg populations have been characterized both in the mouse and humans. Animal studies have clearly demonstrated the participation of Treg in the tolerance mechanisms to chemicals. Similarly, preliminary reports in humans suggest that Treg are also involved in the control of immune responses to environmental chemicals. The reciprocal role and origin of the Treg population still has to be defined, and there is the need of a better definition of markers for their selective isolation. The in vitro or in vivo selective expansion of chemicalspecific Treg may have a huge impact both in the prevention of undesired CH reaction in selected population groups, as well as in the development of a new strategy for the therapy of already sensitized individuals.

81

References 1 Kupper TS, Fuhlbrigge RC: Immune surveillance in the skin: mechanisms and clinical consequences. Nat Rev Immunol 2004;4:211–222. 2 Sabroe I, Read RC, Whyte MK, Dockrell DH, Vogel SN, Dower SK: Toll-like receptors in health and disease: complex questions remain. J Immunol 2003;171:1630–1635. 3 Cavani A, Albanesi C, Traidl C, Sebastiani S, Girolomoni G: Effector and regulatory T cells in allergic contact dermatitis. Trends Immunol 2001;22: 118–120. 4 Cavani A, Ottaviani C, Nasorri F, Sebastiani S, Girolomoni G: Immunoregulation of hapten- and drug-induced immune reactions. Curr Opin Allergy Clin Immunol 2003;3:243–247. 5 Schwarz T: Ultraviolet radiation-induced tolerance. Allergy 1999;54:1252– 1261. 6 Galliaerde V, Desvignes C, Peyron E, Kaiserlian D: Oral tolerance to haptens: intestinal epithelial cells from 2,4-dinitrochlorobenzene-fed mice inhibit hapten-specific T-cell activation in vitro. Eur J Immunol 1995;25:1385–1390. 7 Thorstenson KM, Khoruts A: Generation of anergic and potentially immunoregulatory CD25+CD4 T cells in vivo after induction of peripheral tolerance with intravenous or oral antigen. J Immunol 2001;167:188–195. 8 Dubois B, Chapat L, Goubier A, Papiernik M, Nicolas JF, Kaiserlian D: Innate CD4+CD25+ regulatory T cells are required for oral tolerance and inhibition of CD8+ T cells mediating skin inflammation. Blood 2003;102:3295– 3301. 9 Desvignes C, Bour H, Nicolas JF, Kaiserlian D: Lack of oral tolerance but oral priming for contact sensitivity to dinitrofluorobenzene in major histocompatibility complex class II-deficient mice and in CD4+ T-cell-depleted mice. Eur J Immunol 1996;26:1756– 1761. 10 Worbs T, Bode U, Yan S, Hoffmann MW, Hintzen G, Bernhardt G, Forster R, Pabst O: Oral tolerance originates in the intestinal immune system and relies on antigen carriage by dendritic cells. J Exp Med 2006;203:519–527.

82

11 Roelofs-Haarhuis K, Wu X, Gleichmann E: Oral tolerance to nickel requires CD4+ invariant NKT cells for the infectious spread of tolerance and the induction of specific regulatory T cells. J Immunol 2004; 173:1043–1050. 12 Tang A, Udey MC: Doses of ultraviolet radiation that modulate accessory cell activity and ICAM-1 expression are ultimately cytotoxic for murine epidermal Langerhans cells. J Invest Dermatol 1992;99:71S–73S. 13 Schwarz A, Maeda A, Kernebeck K, van Steeg H, Beissert S, Schwarz T: Prevention of UV radiation-induced immunosuppression by IL-12 is dependent on DNA repair. J Exp Med 2005;201:173– 179. 14 Schwarz A, Beissert S, Grosse-Heitmeyer K, Gunzer M, Bluestone JA, Grabbe S, Schwarz T: Evidence for functional relevance of CTLA-4 in ultraviolet-radiation-induced tolerance. J Immunol 2000;165:1824–1831. 15 Schwarz A, Maeda A, Stander S, van Steeg H, Schwarz T: IL-18 reduces ultraviolet radiation-induced DNA damage and thereby affects photoimmunosuppression. J Immunol 2006;176: 2896–2901. 16 Ghoreishi M, Dutz JP: Tolerance induction by transcutaneous immunization through ultraviolet-irradiated skin is transferable through CD4+CD25+ T regulatory cells and is dependent on host-derived IL-10. J Immunol 2006; 176:2635–2644. 17 Maurer M, Seidel-Guyenot W, Metz M, Knop J, Steinbrink K: Critical role of IL-10 in the induction of low zone tolerance to contact allergens. J Clin Invest 2003;112:432–439. 18 Groux H, O’Garra A, Bigler M, Rouleau M, Antonenko S, de Vries JE, Roncarolo MG: A CD4+ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 1997;389:737–742. 19 Ferguson TA, Dube P, Griffith TS: Regulation of contact hypersensitivity by interleukin-10. J Exp Med 1994;179: 1597–1604. 20 Schwarz A, Grabbe S, Riemann H, Aragane Y, Simon M, Manon S, Andrade S, Luger TA, Zlotnik A, Schwarz T: In vivo effects of interleukin-10 on contact hypersensitivity and delayed-type hypersensitivity reactions. J Invest Dermatol 1994;103:211–216.

21 Berg DJ, Leach MW, Kuhn R, Rajewsky K, Muller W, Davidson NJ, Rennick D: Interleukin-10 but not interleukin-4 is a natural suppressant of cutaneous inflammatory responses. J Exp Med 1995;182:99–108. 22 Cavani A, Nasorri F, Prezzi C, Sebastiani S, Albanesi C, Girolomoni G: Human CD4+ T lymphocytes with remarkable regulatory functions on dendritic cells and nickel-specific Th1 immune responses. J Invest Dermatol 2000;114:295–302. 23 Sebastiani S, Allavena P, Albanesi C, Nasorri F, Bianchi G, Traidl C, Sozzani S, Girolomoni G, Cavani A: Chemokine receptor expression and function in CD4+ T lymphocytes with regulatory activity. J Immunol 2001;166:996–1002. 24 Buer J, Lanoue A, Franzke A, Garcia C, von Boehmer H, Sarukhan A: Interleukin10 secretion and impaired effector function of major histocompatibility complex class II-restricted T cells anergized in vivo. J Exp Med 1998;187:177–183. 25 Sakaguchi S: Naturally arising Foxp3expressing CD25+CD4+ regulatory T cells in immunological tolerance to self and non-self. Nat Immunol 2005;6: 345–352. 26 Kim JM, Rudensky A: The role of the transcription factor Foxp3 in the development of regulatory T cells. Immunol Rev 2006;212:86–98. 27 Le Bras S, Geha RS: IPEX and the role of Foxp3 in the development and function of human Tregs. J Clin Invest 2006;116:1473–1475. 28 Hoglund P: Induced peripheral regulatory T cells: the family grows larger. Eur J Immunol 2006;36:264–266. 29 Layland J, Solaro RJ, Shah AM: Regulation of cardiac contractile function by troponin I phosphorylation. Cardiovasc Res 2005;66:12–21. 30 Cavani A, Nasorri F, Ottaviani C, Sebastiani S, De Pita O, Girolomoni G: Human CD25+ regulatory T cells maintain immune tolerance to nickel in healthy, nonallergic individuals. J Immunol 2003;171:5760–5768. 31 Iellem A, Mariani M, Lang R, Recalde H, Panina-Bordignon P, Sinigaglia F, D’Ambrosio D: Unique chemotactic response profile and specific expression of chemokine receptors CCR4 and CCR8 by CD4+CD25+ regulatory T cells. J Exp Med 2001;194:847–853.

Cavani ⴢ De Pità

32 Sarween N, Chodos A, Raykundalia C, Khan M, Abbas AK, Walker LS: CD4+ CD25+ cells controlling a pathogenic CD4 response inhibit cytokine differentiation, CXCR-3 expression, and tissue invasion. J Immunol 2004; 173: 2942–2951. 33 Dieckmann D, Bruett CH, Ploettner H, Lutz MB, Schuler G: Human CD4+ CD25+ regulatory, contact-dependent T cells induce interleukin-10-producing, contact-independent type 1-like regulatory T cells. J Exp Med 2002;196: 247–253. 34 Paust S, Lu L, McCarty N, Cantor H: Engagement of B7 on effector T cells by regulatory T cells prevents autoimmune disease. Proc Natl Acad Sci USA 2004;101:10398–10403. 35 Fallarino F, Grohmann U, Hwang KW, Orabona C, Vacca C, Bianchi R, Belladonna ML, Fioretti MC, Alegre ML, Puccetti P: Modulation of tryptophan catabolism by regulatory T cells. Nat Immunol 2003;4:1206–1212. 36 Grossman WJ, Verbsky JW, Barchet W, Colonna M, Atkinson JP, Ley TJ: Human T regulatory cells can use the perforin pathway to cause autologous target cell death. Immunity 2004;21: 589–601. 37 Seidel-Guyenot W, Perschon S, Dechant N, Alt R, Knop J, Steinbrink K: Low zone tolerance induced by systemic application of allergens inhibits Tc1mediated skin inflammation. J Allergy Clin Immunol 2006;117:1170–1177. 38 Gilliet M, Liu YJ: Generation of human CD8 T regulatory cells by CD40 ligandactivated plasmacytoid dendritic cells. J Exp Med 2002;195:695–704. 39 Chang CC, Ciubotariu R, Manavalan JS, Yuan J, Colovai AI, Piazza F, Lederman S, Colonna M, Cortesini R, DallaFavera R, Suciu-Foca N: Tolerization of dendritic cells by T(S) cells: the crucial role of inhibitory receptors ILT3 and ILT4. Nat Immunol 2002;3:237–243. 40 Steinman RM, Hawiger D, Nussenzweig MC: Tolerogenic dendritic cells. Annu Rev Immunol 2003;21:685–711.

41 Banchereau J, Paczesny S, Blanco P, Bennett L, Pascual V, Fay J, Palucka AK: Dendritic cells: controllers of the immune system and a new promise for immunotherapy. Ann NY Acad Sci 2003;987:180–187. 42 Gallucci S, Matzinger P: Danger signals: SOS to the immune system. Curr Opin Immunol 2001;13:114–119. 43 Mahnke K, Qian Y, Knop J, Enk AH: Induction of CD4+/CD25+ regulatory T cells by targeting of antigens to immature dendritic cells. Blood 2003;101: 4862–4869. 44 Seguin B, Uetrecht J: The danger hypothesis applied to idiosyncratic drug reactions. Curr Opin Allergy Clin Immunol 2003;3:235–242. 45 Pasare C, Medzhitov R: Toll pathwaydependent blockade of CD4+CD25+ Tcell-mediated suppression by dendritic cells. Science 2003;299:1033–1036. 46 Kalimo K, Mattila L, Kautiainen H: Nickel allergy and orthodontic treatment. J Eur Acad Dermatol Venereol 2004;18:543–545. 47 Jiang S, Camara N, Lombardi G, Lechler RI: Induction of allopeptide-specific human CD4+CD25+ regulatory T cells ex vivo. Blood 2003;102:2180–2186. 48 Yamazaki S, Iyoda T, Tarbell K, Olson K, Velinzon K, Inaba K, Steinman RM: Direct expansion of functional CD25+ CD4+ regulatory T cells by antigenprocessing dendritic cells. J Exp Med 2003;198:235–247. 49 Chen W, Jin W, Hardegen N, Lei KJ, Li L, Marinos N, McGrady G, Wahl SM: Conversion of peripheral CD4+CD25– naive T cells to CD4+CD25+ regulatory T cells by TGF- induction of transcription factor Foxp3. J Exp Med 2003; 198:1875–1886. 50 Lavelle EC, McNeela E, Armstrong ME, Leavy O, Higgins SC, Mills KH: Cholera toxin promotes the induction of regulatory T cells specific for bystander antigens by modulating dendritic cell activation. J Immunol 2003;171:2384– 2392.

51 Adorini L: Tolerogenic dendritic cells induced by vitamin D receptor ligands enhance regulatory T cells inhibiting autoimmune diabetes. Ann NY Acad Sci 2003;987:258–261. 52 Valencia X, Stephens G, GoldbachMansky R, Wilson M, Shevach EM, Lipsky PE: TNF down-modulates the function of human CD4+CD25hi Tregulatory cells. Blood 2006;108:253– 261. 53 Tan PH, Sagoo P, Chan C, Yates JB, Campbell J, Beutelspacher SC, Foxwell BM, Lombardi G, George AJ: Inhibition of NF-B and oxidative pathways in human dendritic cells by antioxidative vitamins generates regulatory T cells. J Immunol 2005;174:7633–7644. 54 Moseman EA, Liang X, Dawson AJ, Panoskaltsis-Mortari A, Krieg AM, Liu YJ, Blazar BR, Chen W: Human plasmacytoid dendritic cells activated by CpG oligodeoxynucleotides induce the generation of CD4+CD25+ regulatory T cells. J Immunol 2004;173:4433– 4442. 55 Ochando JC, Homma C, Yang Y, Hidalgo A, Garin A, Tacke F, Angeli V, Li Y, Boros P, Ding Y, Jessberger R, Trinchieri G, Lira SA, Randolph GJ, Bromberg JS: Alloantigen-presenting plasmacytoid dendritic cells mediate tolerance to vascularized grafts. Nat Immunol 2006;7:652–662. 56 Liu H, Komai-Koma M, Xu D, Liew FY: Toll-like receptor 2 signaling modulates the functions of CD4+CD25+ regulatory T cells. Proc Natl Acad Sci USA 2006;103:7048–7053. 57 Peng G, Guo Z, Kiniwa Y, Voo KS, Peng W, Fu T, Wang DY, Li Y, Wang HY, Wang RF: Toll-like receptor 8-mediated reversal of CD4+ regulatory T-cell function. Science 2005;309:1380–1384. 58 Mates M, Zevin S, Breuer GS, Navon P, Nesher G: Desensitization to hydroxychloroquine – experience of 4 patients. J Rheumatol 2006;33:814–816. 59 Morris DL: Intradermal testing and sublingual desensitization for nickel. Cutis 1998;61:129–132.

Dr. Andrea Cavani Laboratory of Immunology and Allergology, IDI-IRCCS Via dei Monti di Creta 104, IT–00167 Rome (Italy) Tel. +39 06 6646 4776, Fax +39 06 6646 4705 E-Mail [email protected]

Tolerance Mechanisms to Small Molecular Compounds

83

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 84–94

HIV and Drug Hypersensitivity Munir Pirmohamed Department of Pharmacology and Therapeutics, University of Liverpool, Liverpool, UK

Abstract

Introduction

Hypersensitivity reactions are a well-known manifestation of therapy with both anti-HIV drugs, and drugs used for the treatment or prevention of opportunistic reactions. In the course of the history of HIV, the epidemiology of these reactions has changed. Before 1995, hypersensitivity was most frequently seen with the antimicrobials used for opportunistic infections (in particular co-trimoxazole). Since then, with the advent of highly active antiretroviral therapy, the reactions are more commonly seen with the antiretrovirals themselves (in particular abacavir and nevirapine). The manifestations are typical of all hypersensitivity reactions with the skin and liver most commonly affected, and sometimes accompanied by systemic manifestations such as fever and eosinophilia. The reactions still cause a lot of morbidity; however, there has been progress. HIV physicians are now aware of these reactions, monitor for them, and stop drugs or provide symptomatic treatment with the first signs. There is also increasing laboratory evidence that T cells are involved in the pathogenesis of many of these reactions, perhaps providing an opportunity in the future of developing diagnostic tests. Genetic factors predisposing to hypersensitivity reactions have also been identified – the prime example of this is with abacavir, where pre-prescription genotyping for HLA B*5701 has been shown to reduce the incidence of hypersensitivity. Copyright © 2007 S. Karger AG, Basel

HIV is the most researched virus in the world – this intense activity has resulted in a change in the natural history of the disease, from one which was rapidly fatal to one which is now a chronic disease. While we await the development of a vaccine, drugs are the mainstay of treatment. The introduction of highly active antiretroviral therapy (HAART) about a decade ago has led to a reduction in the mortality associated with the disease. The consequent maintenance of immune function has also resulted in a reduction in the frequency of opportunistic infections [1]. The net effect of the change in the natural history of the disease is a change in the pattern of drugs used in HIV, and a consequent change in the adverse drug reaction profile. Drugs used for the treatment of the virus itself, and those used for opportunistic infections and complications associated with the disease, are associated with a wide variety of adverse effects affecting almost every organ system [2]. Drug allergy is one form of adverse reaction that these patients encounter (table 1). Many drugs used to treat both HIV and its associated opportunistic infections cause allergic reactions, although the epidemiology has altered over the years.

Table 1. Factors contributing to the high prevalence of hypersensitivity in HIV-positive patients U

U

U

U

U

U

U

U

Chronic drug therapy (HIV is a chronic disease requiring therapy for years) Multiple drug therapy with unforeseen interactions (both pharmacokinetic and pharmacodynamic) Use of novel drugs often with unknown biological effects (many drugs have off-target effects which are not understood) Use of high mass drug doses (gram doses used with some drugs, e.g. co-trimoxazole) Perturbation of drug metabolism (induction, inhibition, or downregulation) Increased oxidative stress (changes in intra- and extracellular thiols) Immune hyperactivation (increased cytokines, increased HLA expression) Presence of ‘danger’ signals (costimulatory danger signals, some of which are known, many of which are not)

Adapted from Pirmohamed and Park [5].

Epidemiology

The drugs implicated in causing the reactions have changed over the last decade. In the 1980s, the commonest drugs responsible for allergy were the antimicrobials used to treat or protect against opportunistic infections. Co-trimoxazole, used for Pneumocystis carinii pneumonia, received the greatest attention [3]. However, a significant incidence of hypersensitivity was observed with other antimicrobials, including aminopenicillins, isoniazid, rifampicin, thioacetazone, clindamycin, primaquine and dapsone, as well as the other sulfonamides [4]. With the advent of HAART, there is no longer the same need for prophylactic therapy, and thus problems associated with the antimicrobials have decreased, but not disappeared. These have, however, largely been replaced by reactions to the new antiretroviral agents including abacavir, nevirapine and protease inhibitors [5]. A study in 1993 showed that the frequency of drug hypersensitivity in HIV-positive patients

HIV and Drug Allergy

ranges from 3 to 20% [6]. It was estimated that drug rashes were 100 times more frequent in HIV-positive patients than in the general population [7]. More recently, since the advent of HAART, there have been no surveys on the incidence of drug allergy with different antiretrovirals. However, data exist for individual drugs and are covered in the individual sections below. The clinical manifestations are similar for the older drugs and the new compounds. With the antimicrobials used to treat opportunistic infections, most patients presented with erythematous rashes which, in about 0.5% of cases, were blistering in nature [7]. The rashes were accompanied by fever, and in some cases by internal organ involvement. The skin remains the commonest organ to be affected by the antiretrovirals, although severity may vary and there are subtle clinical differences. For example, rechallenge is strictly contraindicated with abacavir [8] where deaths have been reported on re-exposure, while it is deliberately attempted as a desensitization procedure in patients with a history of co-trimoxazole allergy [9]. Nevirapine is one of the commonest causes of blistering skin reactions [10]. Liver injury is also reported frequently with antiretroviral compounds, and may occur in the absence of skin manifestations [11, 12]. Indeed, liver injury has recently resulted in a high profile withdrawal of a new CCR5 antagonist before the drug was licensed [13]. It is important to note that many of the skin and liver manifestations associated with drugs used in HIV disease are assumed to be immunological in origin, usually based on the clinical symptoms and signs, but there is rarely good laboratory evidence of immune system involvement.

Co-Trimoxazole and Hypersensitivity Reactions

Co-trimoxazole is used in the treatment and prophylaxis of P. carinii pneumonia. The use of cotrimoxazole has decreased over the last decade

85

Fig. 1. Structure of drugs particularly associated with hypersensitivity reactions.

since the advent of HAART, and consequent immune reconstitution [14, 15]. Nevertheless, it is still associated with hypersensitivity reactions, and is a good paradigm by which to investigate the mechanisms of hypersensitivity in HIV-positive patients. Co-trimoxazole is associated with hypersensitivity in 1–3% of HIV-negative patients [16]. In HIV-positive patients, the frequency is much higher occurring in 30% of patients at prophylactic doses, and 50% of patients at doses used for treatment [3]. Sulfamethoxazole (SMX; fig. 1) has been suggested to be responsible for 80% of cases, with trimethoprim accounting for about 20% [17]. Although there is anecdotal evidence to suggest that hypersensitivity reactions to SMX are not as frequently seen since the advent of HAART, it is not clear whether this is a true reduction in incidence or merely reflects reduced used of this drug. The mechanism by which SMX causes hypersensitivity is not clear. Two complementary hypotheses, both of which can be supported by experimental evidence, have been put forward. First, the parent drug itself may be responsible through a pathway that is major histocompatibility complex (MHC)-restricted, but processing-

86

and metabolism-independent [18, 19]. This requires labile, reversible binding of SMX to the MHCs on antigen-presenting cells. This has been termed the pharmacological interaction with immune receptors, or the p-i concept. The presence of T-cell clones that proliferate only in response to SMX, and the rapid downregulation in expression of the T-cell receptor upon stimulation are consistent with this mechanism. Further evidence for the p-i concept has been provided recently by transfecting drug-specific T-cell receptors into hybridoma cells, which were then able to recognize their specific drugs in a direct, processing-independent and dose-dependent way, similar to the response obtained with T-cell clones [20]. Second, immune response occurs through the traditional hapten pathway, where bioactivation and haptenation are critical in initiating the immune response [21, 22]. SMX undergoes biotransformation by CYP2C9 to the pro-toxin hydroxylamine metabolite, which is then auto-oxidized to the ultimate toxin, nitroso-SMX [23]. In support of the role of nitroso-SMX in the pathogenesis of hypersensitivity, the following has been shown: (1) The hydroxylamine metabolite can be detected in the urine of HIV-positive patients [24]. (2) Keratinocytes, which are the site of attack in cutaneous eruptions, can bioactivate SMX to the hydroxylamine [25] consistent with the finding that keratinocytes express multiple P450 isoforms and drug transporters [26]. Keratinocytes form adducts within the cell and on the cell surface after exposure to SMX suggesting that they may act as a source of antigen on cell death, but may also be able to present antigen to hapten-specific T lymphocytes [27]. Human dermal fibroblasts are less susceptible to the cytotoxic effects of the hydroxylamine metabolite than keratinocytes, but nevertheless respond to exposure to the hydroxylamine metabolite by forming intracellular adducts, and increased oxidative stress with a consequent reduction in glutathione levels [28].

Pirmohamed

(3) Nitroso-SMX can bind covalently to thiol residues in proteins and on surfaces of lymphocytes [29] and keratinocytes [25]. (4) CD8+ dermal T cells proliferate more vigorously in response to microsome-generated SMX metabolites when compared to the parent compound alone [30]. (5) Nitroso-specific proliferation of splenocytes from mice, rats and rabbits treated with nitroso-SMX has been shown. There was no response on in vitro stimulation with SMX indicating that the reactive metabolites are the source of antigen [31]. It is of course possible, and perhaps likely, that both mechanisms may be important in the overall pathogenesis. For example, the hapten-dependent pathway may be more important for primary immune stimulation (sensitization), while the metabolism-independent pathway may be all that is necessary for secondary stimulation and elicitation of tissue damage. It is also known that nitroso-SMX is directly toxic to cells [29]. However, in HIV-negative patients, cells from hypersensitive patients are more sensitive to the nitroso metabolite than control cells suggestive of a deficiency in cellular defense [32]. A similar finding was also reported in HIVpositive patients [33], although this was not replicated in a more recent study [34]. Interestingly, cytotoxicity with SMX-hydroxylamine was recently correlated with HIV-1 viral tat protein expression, with enhanced sensitivity also being associated with reduced intracellular glutathione concentrations [35]. In this respect, it is interesting to note that HIV-1 tat protein has been shown to modulate the redox status of the cell by lowering the level of reduced and total glutathione in cell lines in vitro and in vivo in a transgenic model [36]. Glutathione, along with other thiols and ascorbic acid, thus seems to be involved in the mechanism of cytotoxicity associated with SMX metabolites. Glutathione can reduce the cytotoxicity of the nitroso metabolite [29], and can also

HIV and Drug Allergy

reduce the lymphocyte proliferative response of T-cell clones that are reactive to nitroso-SMX [31, 37]. Biochemically, glutathione can act as a reducing agent, converting the nitroso metabolite to the hydroxylamine and the parent drug [38]. Interestingly, cysteine reacts more rapidly with the nitroso metabolite than glutathione [29], suggesting that in certain body compartments (it is mainly an extracellular thiol), it may be more important in detoxification of nitroso-SMX than glutathione. The importance of thiols in detoxification of the toxic metabolites of SMX lies in the fact that HIV infection exerts a degree of oxidative stress, and leads to alteration in reduced to oxidized thiol ratios [39]. It has been shown that HIV-positive patients have decreased plasma cysteine levels [40] and a decreased capacity to reduce nitroso-SMX back to its parent compound [40]. Furthermore, ascorbic acid is also reduced in HIV-positive patients, and its deficiency was strongly correlated with an impaired reduction of nitroso-SMX to its hydroxylamine [41]. Taken together, this suggests that there may be a detoxification defect in HIV-positive patients; further evidence for this has been provided in a recent study which showed that endogenous ascorbate and glutathione are important for the intracellular reduction of nitroso-SMX, and that redox cycling of the hydroxylamine and nitroso metabolites may contribute to the cytotoxicity of these metabolites in vitro [42]. Since hypersensitivity only affects a proportion of patients treated with co-trimoxazole, it has been postulated that genetic factors may determine individual predisposition. In HIV-negative patients, a deficiency in N-acetylation (NAT2; at both phenotypic and genotypic levels) has been reported to be a risk factor [43–45]. In contrast, the N-acetylator genotype does not predict individual susceptibility in HIV-positive patients [46]. Furthermore, polymorphisms in CYP2C9, and the glutathione transferase genes (GSTM1, GSTT1, GSTP1) were also not associated with cotrimoxazole hypersensitivity [46]. Analysis of the

87

MHC region on chromosome 6 has also not revealed any immunogenetic polymorphisms that predispose to SMX hypersensitivity [47]. Given that in HIV, the rate of allergic reactions with co-trimoxazole is at least 10 times higher than that in non-HIV patients [2], and no genetic factors have been identified to date, this may represent an example of an idiosyncratic reaction where environmental factors predominate. The explanation for this may well lie in the recently proposed danger hypothesis [48]. According to this hypothesis, the immune system responds not to non-self but to dangerous entities, with the ultimate aim of the immune response being the need to detect and protect against danger. In the context of HIV disease, the increased expression of proinflammatory cytokines may be one source of the danger signal. For instance, IFN-
is elevated in the sera of HIVpositive patients [49], which in turn increases the expression of various proinflammatory cytokines, and contributes to the oxidative stress seen in HIV disease [50]. Another effect of IFN-
overexpression is the upregulation of MHC class II and costimulatory molecules on antigen-presenting cells and other cells including keratinocytes, which overall will lead to enhancement of antigen presentation [51]. However there may also be alternative reasons for the increase in ‘danger’. For example, with SMX, it has been shown that HIV-infected cells may be more sensitive to the cytotoxic effects of nitroso-SMX [52]. Necrosis of cells may further increase proinflammatory cytokine expression and, in addition, release other factors, the nature of which has not been clearly defined, that may act to increase ‘danger’. Theoretically, therefore, nullification of these danger signals may be a method that could be used to reduce the frequency of drug hypersensitivity in HIV disease. Interestingly, nitrosoSMX may not necessarily need to cause cell death to lead to a danger signal. Recent studies have shown that the nitroso metabolite leads to increased CD40 expression on dendritic cells [Nais-

88

bitt, unpublished data], and the increased costimulatory signal produced from this, taken together with the other proinflammatory stimuli in HIV disease, may account for the higher prevalence of hypersensitivity to SMX when compared with the non-HIV population. Clearly, the association of a viral infection with an increased incidence of drug hypersensitivity is not unique to HIV, but is well recognized with Epstein-Barr virus, and more recently, human herpesvirus 6 infections [53].

Abacavir Hypersensitivity

Abacavir (fig. 1), an HIV-1 reverse transcriptase inhibitor, causes hypersensitivity reactions in about 5% of patients [54]. The adverse effect occurs after a mean of 11 (range 3–56) days, and is characterized by (in rank order of frequency) fever, rash, gastrointestinal, mouth/throat, respiratory, and musculoskeletal symptoms, as well as malaise, lymphadenopathy, paresthesia, and fatigue [55]. This reaction needs to be distinguished from rash (without systemic manifestations) which can occur in 15% of patients. Abacavir is metabolized in the body, with alcohol dehydrogenase being one of the enzymes involved [56]. A putative reactive metabolite has been suggested, but whether this is involved in the immune reaction is unclear. In our laboratory we have shown the presence of T cells that proliferated in vitro on exposure to abacavir, while in one case T cells proliferated only in the presence of an ADH-metabolizing system indicating that they were responding to a metabolite [57]. These proliferating T cells were identified as being CD8+ cells by flow cytometry [58]. Analysis of the MHC by Mallal et al. [59] found a strong association between abacavir hypersensitivity and the haplotype comprising HLA-B*5701, HLA-DR7, and HLA-DQ3 with an odds ratio of over 100. Subsequent studies have shown that this haplotype resides on the ances-

Pirmohamed

tral haplotype 57.1, and that the combination of HLA-B*5701 and polymorphism in heat shock protein-Hom has greater predictive accuracy than HLA-B*5701 by itself [60]. This association has now been shown in two other cohorts by GlaxoSmithKline Pharmaceuticals, the manufacturer of the drug, and independently in a cohort of patients from the UK [61–63]. However, the same association has not been shown in an African-American population presumably because of ethnic differences in linkage disequilibrium patterns in the MHC [62 ; see Nolan et al., pp 95–104]. An analysis of the predictive value of prospective HLA-B*5701 genotyping prior to abacavir hypersensitivity based on a meta-analysis of three cohorts showed that to prevent one case of hypersensitivity, eight HLA-B*5701-positive patients would be denied abacavir, and to identify them 48 patients would require testing [63]. The conclusion from this analysis was that genotyping for HLA-B*5701 prior to the prescription of abacavir in Caucasians is a cost-effective strategy. In a prospective study of 260 abacavir-naive patients, genotyping for HLA-B*5701 was able to prevent hypersensitivity reactions, with none of the 148 HLA-B*5701-negative developing hypersensitivity [64]. Similar results have also been observed in a UK cohort [65]. More rapid methods for typing have also been developed [66, 67], which will further enhance translation into clinical practice. A randomized controlled trial to test the utility of pre-prescription genotyping for HLA-B*5701 to prevent abacavir hypersensitivity has just been completed and the results are eagerly awaited.

Non-Nucleoside Reverse Transcriptase Inhibitor

The non-nucleoside reverse transcriptase inhibitors (NNRTIs) include delavirdine, efavirenz and nevirapine. All NNRTIs are associated with rash-

HIV and Drug Allergy

es which are presumed to be a hypersensitivity phenomenon. All grades of rash have been observed in a 10% excess to controls with efavirenz and a 15–17% excess with nevirapine (fig. 1) and delavirdine [68]. Nevirapine-induced hypersensitivity has received most attention because of the high frequency and severity of rash [69]. For example, Stevens-Johnson syndrome occurs in 0.3% of patients on nevirapine [70, 71] but rarely with the other NNRTIs. Slow introduction of nevirapine reduces the risk of rash [72]. Steroids may actually increase the risk of rash [73]. The risks of severe rashes and discontinuation of nevirapine therapy in women are 7- and 3.5-fold higher, respectively, than in men [74]. High CD4+ count also predisposes to hypersensitivity. Nevirapine is also associated with raised transaminases in 7% of patients, with 1% developing hepatitis [68]. A recent study showed that the independent risk factors for severe hepatotoxicity were: body mass index !18.5; female sex; serum albumin level !35 g/l; mean corpuscular volume 185 fl; plasma HIV-1 RNA load !20,000 copies/ml; aspartate aminotransferase level !75 IU/l, and lactate dehydrogenase level !164 IU/l [75]. In some cases the hepatic abnormalities are due to concomitant infection with hepatitis viruses, but undoubtedly in some of the patients hepatitis is a manifestation of the hypersensitivity reaction. There are also reports of severe hepatotoxicity and rash in individuals given nevirapine as part of post-exposure prophylaxis [76]. Evaluation by the FDA of reports received by MedWatch over the last 3 years has revealed that there were 22 cases of serious adverse events associated with post-exposure prophylaxis, including 12 reports of hepatotoxicity, 14 of skin reactions and one of rhabdomyolysis, occurring a median of 21 and 9 days after start of therapy for the hepatic and skin reactions, respectively [77]. The use of nevirapine for post-exposure prophylaxis is therefore no longer recommended.

89

The contribution (if any) of nevirapine serum concentration profile to the risk of hypersensitivity is not known. Humans show considerable variation in the expression and activity of the drug metabolizing enzymes (CYP3A4, CYP2B6 and UGT) involved in nevirapine metabolism [78]. Patients with impaired metabolic capacity, for example those with polymorphisms in CYP2B6, have lower clearance rates [78] and have higher overall drug exposure at standard doses. However, there are contradictory reports as to whether elevated drug levels increase the risk of liver toxicity associated with nevirapine [79, 80]. Interestingly, a recent randomized study has shown that the MDR1 3435 C]T polymorphism was associated with a significantly reduced risk of hepatotoxicity, even though nevirapine is not a substrate for P-glycoprotein [81]. While skin reactions associated with nevirapine seem to have all the clinical features consistent with an immune etiology, it is more difficult to be sure of an immune pathogenesis with liver injury when it is not accompanied by hypersensitivity manifestations. The pattern of hepatic injury with nevirapine varies, some reactions occurring early, while others occur late after prolonged exposure. The latter may be non-immune in nature while the former may have an immune pathogenesis. Consistent with this, we have recently shown the presence of drug-reactive T cells in a patient with such a history [82]. Further evidence of the ability of nevirapine to cause immune-mediated reactions comes from two sources. First, the findings from an animal model using Brown Norway rats developed by Uetrecht [83] and co-workers is consistent with a hypersensitivity phenomenon, although it is not clear whether the immune system is responding to the parent drug or reactive metabolite [see Shenton et al., pp 115–128]. Second, a recent study in an Australian population has shown that HLA-DRB1*0101, together with CD4 counts, acts as a predisposing factor for nevirapine hypersensitivity [84]. A small study in Sardinian patients

90

has shown an association with different alleles (HLA Cw8-B14 haplotype) than those found in the Australian study [85].

Protease Inhibitors

Adverse reactions, such as skin rashes and abnormal liver function tests, presumed to be manifestations of hypersensitivity have been reported with all protease inhibitors. Rash, however, seems to be more common with amprenavir than with the other protease inhibitors [74]. It has been reported in up to 28% of patients: 4% have grade 3 or 4 rashes; 3% discontinue treatment, and 1% can develop Stevens-Johnson syndrome. A sulfa moiety in amprenavir has been implicated in the higher incidence of rashes; however, there is no laboratory evidence to support this. The major problem with protease inhibitors, and with HAART in general, seems to be hepatotoxicity [85]. However, it is clear that the occurrence of hepatotoxicity in such patients is a complex phenomenon, probably representing many different forms of liver injury, with different mechanisms [86]. Immune mechanisms may be important in some cases, but there is a dearth of information on this.

Conclusions

Hypersensitivity reactions to a variety of drugs are common in HIV-positive patients. Before 1995, the main culprit was co-trimoxazole, a drug used for the treatment of opportunistic infections, while after 1995, it has been replaced by the antiretrovirals themselves. The skin and liver are the commonest organs to be affected. There is increasing evidence to implicate T cells in the pathogenesis of these reactions, with individual susceptibility being due to genetic factors residing in the MHC on chromosome 6. However, there is still a need for further research to define

Pirmohamed

the mechanisms of the hypersensitivity reactions, including whether the drugs or metabolites are antigens, and to identify individual genetic and non-genetic predisposing factors. The importance of this research agenda is obvious given the high frequency of adverse reactions in HIV patients, and the fact that drugs are still being withdrawn from development even before they are licensed because of their propensity to cause such reactions.

Acknowledgements The support of the Wellcome Trust, MRC, Department of Health, Pfizer, Astra Zeneca and GSK is acknowledged.

References 1 Palella FJ, Delaney KM, Moorman AC, Loveless MO, Fuhrer J, Satten GA, Aschman DJ, Holmberg SD: Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med 1998;338:853–860. 2 Carr A, Cooper DA: Adverse effects of antiretroviral therapy. Lancet 2000; 356:1423–1430. 3 Carr A, Cooper DA: Pathogenesis and management of HIV-associated drug hypersensitivity. AIDS Clin Rev 1995; 96:65–97. 4 Pirmohamed M, Park BK: Drug reactions in HIV infected patients. Postgrad Doctor 1995;18:438–44. 5 Pirmohamed M, Park BK: HIV and drug allergy. Curr Opin Allergy Clin Immunol 2001;1:311–316. 6 Coopman SA, Johnson RA, Platt R, Stern RS: Cutaneous disease and drug reactions in HIV infection. N Engl J Med 1993;328:1670–1674. 7 Roujeau JC, Stern RS: Severe adverse cutaneous reactions to drugs. N Engl J Med 1994;331:1272–1285. 8 Escaut L, Liotier JY, Albengres E, Cheminot N, Vittecoq D: Abacavir rechallenge has to be avoided in case of hypersensitivity reaction. AIDS 1999; 13:1419–1420. 9 Rich JD, Sullivan T, Greineder D, Kazanjian PH: Trimethoprim/sulfamethoxazole incremental dose regimen in human immunodeficiency virusinfected persons. Ann Allergy Asthma Immunol 1997;79:409–414.

HIV and Drug Allergy

10 Fagot JP, Mockenhaupt M, BouwesBavinck JN, Naldi L, Viboud C, Roujeau JC: Nevirapine and the risk of StevensJohnson syndrome or toxic epidermal necrolysis. AIDS 2001;15:1843–1848. 11 Abrescia N, D’Abbraccio M, Figoni M, Busto A, Maddaloni A, De Marco M: Hepatotoxicity of antiretroviral drugs. Curr Pharm Des 2005;11:3697–3710. 12 Martinez E, Blanco JL, Arnaiz JA, Perez-Cuevas JB, Mocroft A, Cruceta A, Marcos MA, Milinkovic A, GarciaViejo MA, Mallolas J, Carne X, Phillips A, Gatell JM: Hepatotoxicity in HIV-1infected patients receiving nevirapinecontaining antiretroviral therapy. AIDS 2001;15:1261–1268. 13 Crabb C: GlaxoSmithKline ends aplaviroc trials. AIDS 2006;20:641. 14 Lopez Bernaldo de Quiros JC, Miro JM, Pena JM, Podzamczer D, Alberdi JC, Martinez E, Cosin J, Claramonte X, Gonzalez J, Domingo P, Casado JL, Ribera E: A randomized trial of the discontinuation of primary and secondary prophylaxis against Pneumocystis carinii pneumonia after highly active antiretroviral therapy in patients with HIV infection. Grupo de Estudio del SIDA 04/98. N Engl J Med 2001;344: 159–167. 15 Ledergerber B, Mocroft A, Reiss P, Furrer H, Kirk O, Bickel M, Uberti-Foppa C, Pradier C, D’Arminio Monforte A, Schneider MM, Lundgren JD: Discontinuation of secondary prophylaxis against Pneumocystis carinii pneumonia in patients with HIV infection who have a response to antiretroviral therapy. Eight European Study Groups. N Engl J Med 2001;344:168–174.

16 van der Ven AJ, Koopmans PP, Vree TB, van der Meer JW: Adverse reactions to co-trimoxazole in HIV infection. Lancet 1991;ii:431–433. 17 Cribb AE, Lee BL, Trepanier LA, Spielberg SP: Adverse reactions to sulphonamide and sulphonamide-trimethoprim antimicrobials: clinical syndromes and pathogenesis. Adverse Drug React Toxicol Rev 1996;15:9–50. 18 Schnyder B, Mauri-Hellweg D, Zanni M, Bettens F, Pichler WJ: Direct, MHCdependent presentation of the drug sulfamethoxazole to human / T cell clones. J Clin Invest 1997;100:136–141. 19 Schnyder B, Burkhart C, SchnyderFrutig K, von Greyerz S, Naisbitt DJ, Pirmohamed M, Park BK, Pichler WJ: Recognition of sulfamethoxazole and its reactive metabolites by drug-specific CD4+ T cells from allergic individuals. J Immunol 2000;164:6647–6654. 20 Schmid DA, Depta JP, Luthi M, Pichler WJ: Transfection of drug-specific T-cell receptors into hybridoma cells: tools to monitor drug interaction with T-cell receptors and evaluate cross-reactivity to related compounds. Mol Pharmacol 2006;70:356–365. 21 Park BK, Naisbitt DJ, Gordon SF, Kitteringham NR, Pirmohamed M: Metabolic activation in drug allergies. Toxicology 2001;158:11–23. 22 Naisbitt DJ, Gordon SF, Pirmohamed M, Park BK: Immunological principles of adverse drug reactions: the initiation and propagation of immune responses elicited by drug treatment. Drug Saf 2000;23:483–507.

91

23 Cribb AE, Spielberg SP, Griffin GP: N4hydroxylation of sulfamethoxazole by cytochrome P450 of the cytochrome P4502C subfamily and reduction of sulfamethoxazole hydroxylamine in human and rat hepatic microsomes. Drug Metab Dispos 1995;23:406–414. 24 van der Ven AJ, Vree TB, van EwijkBeneken Kolmer EW, Koopmans PP, van der Meer JW: Urinary recovery and kinetics of sulphamethoxazole and its metabolites in HIV-seropositive patients and healthy volunteers after a single oral dose of sulphamethoxazole. Br J Clin Pharmacol 1995;39:621–625. 25 Reilly TP, Lash LH, Doll MA, Hein DW, Woster PM, Svensson CK: A role for bioactivation and covalent binding within epidermal keratinocytes in sulfonamide-induced cutaneous drug reactions. J Invest Dermatol 2000; 114: 1164–1173. 26 Baron JM, Holler D, Schiffer R, Frankenberg S, Neis M, Merk HF, Jugert FK: Expression of multiple cytochrome p450 enzymes and multidrug resistance-associated transport proteins in human skin keratinocytes. J Invest Dermatol 2001;116:541–548. 27 Roychowdhury S, Vyas PM, Reilly TP, Gaspari AA, Svensson CK: Characterization of the formation and localization of sulfamethoxazole and dapsoneassociated drug-protein adducts in human epidermal keratinocytes. J Pharmacol Exp Ther 2005;314:43–52. 28 Bhaiya P, Roychowdhury S, Vyas PM, Doll MA, Hein DW, Svensson CK: Bioactivation, protein haptenation, and toxicity of sulfamethoxazole and dapsone in normal human dermal fibroblasts. Toxicol Appl Pharmacol 2006; 215:158–167. 29 Naisbitt DJ, Hough SJ, Gill HJ, Pirmohamed M, Kitteringham NR, Park BK: Cellular disposition of sulphamethoxazole and its metabolites: implications for hypersensitivity. Br J Pharmacol 1999;126:1393–1407. 30 Hertl M, Jugert F, Merk HF: CD8(+) dermal T-cells from a sulfamethoxazole-induced bullous exanthem proliferate in response to drug-modified liver-microsomes. Br J Dermatol 1995; 132:215–220. 31 Farrell J, Naisbitt DJ, Drummond NS, Depta JPH, Vilar FJ, Pirmohamed M, Park BK: Characterization of sulfa-

92

32

33

34

35

36

37

38

39

methoxazole and sulfamethoxazole metabolite-specific T-cell responses in animals and humans. J Pharmacol Exp Ther 2003;306:229–237. Rieder MJ, Uetrecht JP, Shear NH, Cannon M, Miller M, Spielberg SP: Diagnosis of sulfonamide hypersensitivity reactions by in-vitro ‘rechallenge’ with hydroxylamine metabolites. Ann Intern Med 1989;110:286–289. Carr A, Tindall B, Penny R, Cooper DA: In vitro cytotoxicity as a marker of hypersensitivity to sulphamethoxazole in patients with HIV. Clin Exp Immunol 1993;94:21–25. Reilly TP, MacArthur RD, Farrough MJ, Crane LR, Woster PM, Svensson CK: Is hydroxylamine-induced cytotoxicity a valid marker for hypersensitivity reactions to sulfamethoxazole in human immunodeficiency virus-infected individuals? J Pharmacol Exp Ther 1999;291:1356–1364. Arp J, Rieder MJ, Urquhart B, Freeman D, Tucker MJ, Krizova A, Lehmann D, Dekaban GA: Hypersensitivity of HIV1-infected cells to reactive sulfonamide metabolites correlated to expression of the HIV-1 viral protein tat. J Pharmacol Exp Ther 2005;314:1218–1225. Choi J, Liu RM, Kundu RK, Sangiorgi F, Wu W, Maxson R, Forman HJ: Molecular mechanism of decreased glutathione content in human immunodeficiency virus type 1 tat-transgenic mice. J Biol Chem 2000;275:3693–3698. Burkhart C, von Greyerz S, Depta JP, Naisbitt DJ, Britschgi M, Park KB, Pichler WJ: Influence of reduced glutathione on the proliferative response of sulfamethoxazole-specific and sulfamethoxazole-metabolite-specific human CD4+ T-cells. Br J Pharmacol 2001;132:623–630. Naisbitt DJ, O’Neill PM, Pirmohamed M, Park BK: Synthesis and reactions of nitroso sulfamethoxazole with biological nucleophiles – implications for immune-mediated toxicity. Bioorg Med Chem Lett 1996;6:1511–1516. Aukrust P, Svardal AM, Muller F, Lunden B, Berge RK, Ueland PM, Froland SS: Increased levels of oxidized glutathione in CD4+ lymphocytes associated with disturbed intracellular redox balance in human immunodeficiency virus type 1 infection. Blood 1995;86: 258–267.

40 Naisbitt DJ, Vilar J, Stalford A, Wilkins EGL, Pirmohamed M, Park BK: Plasma cysteine and decreased reduction of nitroso sulphamethoxazole with HIV infection. AIDS Res Hum Retrov 2000; 16:1929–1938. 41 Trepanier LA, Yoder AR, Bajad S, Beckwith MD, Bellehumeur JL, Graziano FM: Plasma ascorbate deficiency is associated with impaired reduction of sulfamethoxazole-nitroso in HIV infection. J Acquir Immune Defic Syndr 2004;36:1041–1050. 42 Lavergne SN, Kurian JR, Bajad SU, Maki JE, Yoder AR, Guzinski MV, Graziano FM, Trepanier LA: Roles of endogenous ascorbate and glutathione in the cellular reduction and cytotoxicity of sulfamethoxazole-nitroso. Toxicology 2006;222:25–36. 43 Zielinska E, Niewiarowski W, Bodalski J, Rebowski G, Skretkowicz J, Mianowska K, Sekulska M: Genotyping of the arylamine N-acetyltransferase polymorphism in the prediction of idiosyncratic reactions to trimethoprim-sulfamethoxazole in infants. Pharm World Sci 1998;20:123–130. 44 Rieder MJ, Shear NH, Kanee A, Tang BK, Spielberg SP: Prominence of slow acetylator phenotype among patients with sulfonamide hypersensitivity reactions. Clin Pharmacol Ther 1991;49: 13–17. 45 Wolkenstein P, Carriere V, Charue D, Bastujigarin S, Revuz J, Roujeau JC, Beaune P, Bagot M: A slow acetylator genotype is a risk factor for sulfonamide-induced toxic epidermal necrolysis and Stevens-Johnson syndrome. Pharmacogenetics 1995;5:255–258. 46 Pirmohamed M, Alfirevic A, Vilar J, Stalford A, Wilkins EGL, Sim E, Park BK: Association analysis of drug metabolizing enzyme gene polymorphisms in HIV-positive patients with co-trimoxazole hypersensitivity. Pharmacogenetics 2000;10:705–713. 47 Pirmohamed M: Genetic factors in the predisposition to drug-induced hypersensitivity reactions. AAPS J 2006;8: E20–E26. 48 Matzinger P: Tolerance, danger, and the extended family. Annu Rev Immunol 1994;12:991–1045. 49 Pantaleo G, Fauci AS: New concepts in the immunopathogenesis of HIV infection. Annu Rev Immunol 1995;13:487– 512.

Pirmohamed

50 Baier-Bitterlich G, Fuchs D, Wachter H: Chronic immune stimulation, oxidative stress, and apoptosis in HIV infection. Biochem Pharmacol 1997;53:755– 763. 51 Schnyder B, Frutig K, Mauri-Hellweg D, Limat A, Yawalkar N, Pichler WJ: Tcell-mediated cytotoxicity against keratinocytes in sulfamethoxazol-induced skin reaction. Clin Exp Allergy 1998;28:1412–1417. 52 Rieder MJ, Krause R, Bird IA, Dekaban GA: Toxicity of sulfonamide-reactive metabolites in HIV-infected, HTLVinfected, and noninfected cells. J Acquir Immune Defic Syndr Hum Retrovirol 1995;8:134–140. 53 Sullivan JR, Shear NH: The drug hypersensitivity syndrome: what is the pathogenesis? Arch Dermatol 2001; 137: 357–364. 54 Hetherington S, McGuirk S, Powell G, Cutrell A, Naderer O, Spreen B, Lafon S, Pearce G, Steel H: Hypersensitivity reactions during therapy with the nucleoside reverse transcriptase inhibitor abacavir. Clin Ther 2001;23:1603–1614. 55 Clay PG, Rathbun RC, Slater LN: Management protocol for abacavir-related hypersensitivity reaction. Ann Pharmacother 2000;34:247–249. 56 Walsh JS, Reese MJ, Thurmond LM: The metabolic activation of abacavir by human liver cytosol and expressed human alcohol dehydrogenase isozymes. Chem Biol Interact 2002;142:135–154. 57 Dodd CC, Naisbitt DJ, Vilar FJ, Gordon SF, Maggs JL, Park BK, Pirmohamed M: Abacavir hypersensitivity in human immunodeficiency virus- positive patients: demonstration of the presence of drug- specific lymphocyte proliferation in vitro. Br J Clin Pharmacol 2003;55: 421–421. 58 Phillips EJ, Wong GA, Kaul R, Shahabi K, Nolan DA, Knowles SR, Martin AM, Mallal SA, Shear NH: Clinical and immunogenetic correlates of abacavir hypersensitivity. AIDS 2005;19:979–981. 59 Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, Sayer D, Castley A, Mamotte C, Maxwell D, James I, Christiansen FT: Association between presence of HLA-B *5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse transcriptase inhibitor abacavir. Lancet 2002;359:727–732. 60 Martin AM, Nolan D, Gaudieri S, Almeida CA, Nolan R, James I, Carvalho F, Phillips E, Christiansen FT, Purcell

HIV and Drug Allergy

61

62

63

64

65

66

67

68

AW, McCluskey J, Mallal S: Predisposition to abacavir hypersensitivity conferred by HLA-B*5701 and a haplotypic Hsp70-Hom variant. Proc Natl Acad Sci USA 2004;101:4180–4185. Hetherington S, Hughes AR, Mosteller M, Shortino D, Baker KL, Spreen W, Lai E, Davies K, Handley A, Dow DJ, Fling ME, Stocum M, Bowman C, Thurmond LM, Roses AD: Genetic variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet 2002;359:1121–1122. Hughes AR, Mosteller M, Bansal AT, Davies K, Haneline SA, Lai EH, Nangle K, Scott T, Spreen WR, Warren LL, Roses AD: Association of genetic variations in HLA-B region with hypersensitivity to abacavir in some, but not all, populations. Pharmacogenomics 2004; 5:203–211. Hughes DA, Vilar FJ, Ward CC, Alfirevic A, Park BK, Pirmohamed M: Cost-effectiveness analysis of HLA B*5701 genotyping in preventing abacavir hypersensitivity. Pharmacogenetics 2004;14:335–342. Rauch A, Nolan D, Martin A, McKinnon E, Almeida C, Mallal S: Prospective genetic screening decreases the incidence of abacavir hypersensitivity reactions in the Western Australian HIV cohort study. Clin Infect Dis 2006; 43:99–102. Reeves I, Fisher M, Churchill D: Clinical utility of HLA-B*5701 testing in a UK clinic cohort. Abstract 667a; Presented at the 13th Conference on Retroviruses and Opportunistic Infections, Denver USA, February 2006. Martin AM, Nolan D, Mallal S: HLAB*5701 typing by sequence-specific amplification: validation and comparison with sequence-based typing. Tissue Antigens 2005;65:571–574. Martin AM, Krueger R, Almeida CA, Nolan D, Phillips E, Mallal S: A sensitive and rapid alternative to HLA typing as a genetic screening test for abacavir hypersensitivity syndrome. Pharmacogenet Genomics 2006;16: 353–357. Moyle G: The emerging roles of nonnucleoside reverse transcriptase inhibitors in antiretroviral therapy. Drugs 2001;61:19–26.

69 Gangar M, Arias G, O’Brien JG, Kemper CA: Frequency of cutaneous reactions on rechallenge with nevirapine and delavirdine. Ann Pharmacother 2000; 34:839–842. 70 Warren KJ, Boxwell DE, Kim NY, Drolet BA: Nevirapine-associated Stevens-Johnson syndrome. Lancet 1998; 351:567. 71 Metry DW, Lahart CJ, Farmer KL, Hebert AA: Stevens-Johnson syndrome caused by the antiretroviral drug nevirapine. J Am Acad Dermatol 2001; 44(suppl):354–357. 72 Barreiro P, Soriano V, Casas E, Estrada V, Tellez MJ, Hoetelmans R, de Requena DG, Jimenez-Nacher I, GonzalezLahoz J: Prevention of nevirapine-associated exanthema using slow dose escalation and/or corticosteroids. AIDS 2000;14:2153–2157. 73 De Luca A, Baracchini A, Baldini F, Zaccarelli M, Cingolani A, De Longis P, Di Giambenedetto S, Tozzi V, Rizzo MG, Murri R, Ippolito G, Ammassari A, Antinori A: Gender, use of corticosteroids and CD4 counts are predictive factors of nevirapine-associated rash. AIDS 2000;14:P182. 74 Max B, Sherer R: Management of the adverse effects of antiretroviral therapy and medication adherence. Clin Infect Dis 2000;30(suppl 2):S96–S116. 75 Sanne I, Mommeja-Marin H, Hinkle J, Bartlett JA, Lederman MM, Maartens G, Wakeford C, Shaw A, Quinn J, Gish RG, Rousseau F: Severe hepatotoxicity associated with nevirapine use in HIVinfected subjects. J Infect Dis 2005;191: 825–829. 76 Benn PD, Mercey DE, Brink N, Scott G, Williams IG: Prophylaxis with a nevirapine-containing triple regimen after exposure to HIV-1. Lancet 2001;357: 687–688. 77 Centers for Disease Control and Prevention (CDC): Serious adverse events attributed to nevirapine regimens for postexposure prophylaxis after HIV exposures – worldwide, 1997–2000. MMWR Morb Mortal Wkly Rep 2001; 49:1153–1156. 78 Rotger M, Colombo S, Furrer H, Bleiber G, Buclin T, Lee BL, Keiser O, Biollaz J, Decosterd L, Telenti A: Influence of CYP2B6 polymorphism on plasma and intracellular concentrations and toxicity of efavirenz and nevirapine in HIVinfected patients. Pharmacogenet Genomics 2005;15:1–5.

93

79 Almond LM, Boffito M, Hoggard PG, Bonora S, Raiteri R, Reynolds HE, Garazzino S, Sinicco A, Khoo SH, Back DJ, Di Perri G: The relationship between nevirapine plasma concentrations and abnormal liver function tests. AIDS Res Hum Retroviruses 2004;20:716–722. 80 Nunez M, Gonzalez-Requena D, Gonzalez-Lahoz J, Soriano V: Short communication: interactions between nevirapine plasma levels, chronic hepatitis C, and the development of liver toxicity in HIV-infected patients. AIDS Res Hum Retroviruses 2003;19: 187–188.

81 Haas DW, Bartlett JA, Andersen JW, Sanne I, Wilkinson GR, Hinkle J, Rousseau F, Ingram CD, Shaw A, Lederman MM, Kim RB: Pharmacogenetics of nevirapine-associated hepatotoxicity: an Adult AIDS Clinical Trials Group collaboration. Clin Infect Dis 2006;43: 783–786. 82 Drummond NS, Vilar FJ, Naisbitt DJ, Hanson A, Woods A, Park BK, Pirmohamed M: Drug-specific T cells in an HIV-positive patient with nevirapineinduced hepatitis. Antivir Ther 2006; 11:393–395. 83 Uetrecht J: Role of animal models in the study of drug-induced hypersensitivity reactions. Aaps J 2005;7:E914– 921.

84 Martin AM, Nolan D, James I, Cameron P, Keller J, Moore C, Phillips E, Christiansen FT, Mallal S: Predisposition to nevirapine hypersensitivity associated with HLA-DRB1*0101 and abrogated by low CD4 T-cell counts. AIDS 2005;19:97–99. 85 Littera R, Carcassi C, Masala A, Piano P, Serra P, Ortu F, Corso N, Casula B, La Nasa G, Contu L, Manconi PE: HLA-dependent hypersensitivity to nevirapine in Sardinian HIV patients. Aids 2006; 20:1621–1626. 86 Nunez M, Soriano V: Hepatotoxicity of antiretrovirals: incidence, mechanisms and management. Drug Saf 2005;28: 53–66.

Prof. Munir Pirmohamed Department of Pharmacology and Therapeutics University of Liverpool Ashton Street Liverpool L69 3GE (UK) Tel. +44 151 794 5549, Fax +44 151 794 5540, E-Mail [email protected]

94

Pirmohamed

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 95–104

Abacavir Hypersensitivity David Nolan
Coral-Ann Almeida
Elizabeth Phillips
Simon Mallal Centre for Clinical Immunology and Biomedical Statistics, Royal Perth Hospital and Murdoch University, Perth, Australia

Abstract Abacavir is a commonly prescribed HIV drug belonging to the nucleoside reverse transcriptase inhibitor (NRTI) class. The major treatment-limiting side effect associated with its use is an early onset multi-system drug hypersensitivity reaction typically including some combination of rash, fever and gastrointestinal symptoms, occurring within 6 weeks of initiating treatment in approximately 5–8% of abacavir recipients. Susceptibility to this drug hypersensitivity syndrome is strongly predicted by the presence of a specific human leukocyte antigen (HLA) allele – HLA-B*5701 – which represents the dominant risk factor for abacavir hypersensitivity among Caucasian and Hispanic populations. The frequency distribution of this genetic marker in different populations is likely to provide a rational basis for racially-defined differences in susceptibility, while the critical role of HLAB*5701 in directing CD8+ T-cell-dependent, HLA-restricted immune responses provides a key role for this genetic variant in the pathogenesis of abacavir-specific immune responses. In this chapter, we review the implications of this genetic association for clinical practice as well as current knowledge of its immunological basis. Copyright © 2007 S. Karger AG, Basel

Abacavir Structure, Antiretroviral Function, and Disposition

Abacavir is a potent nucleoside analogue reverse transcriptase inhibitor (NRTI) drug that has been prescribed for the treatment of HIV infection

since 1999, as a component of combination treatment regimens that typically include two NRTI drugs and a third drug from a different antiretroviral drug class (e.g. HIV protease inhibitors, non-NRTIs). Abacavir is currently available and used either as a single drug (ZiagenTM) or in coformulated drugs that combine two nucleoside analogue drugs (abacavir/lamivudine: Epzicom/ KivexaTM ; abacavir/zidovudine: Combivir TM) or three NRTIs (abacavir/lamivudine/zidovudine: Trizivir TM) [1]. As with all NRTI drugs, antiretroviral activity is conferred by the ability to inhibit HIV reverse transcription performed by the viral RNA-dependent DNA polymerase that allows the creation of a nascent DNA sequence from its own RNA template (fig. 1). Within the NRTI drug class, abacavir has a unique chemical structure comprising a 2, 3 -unsaturated bond in the deoxyribose ring as well as a carbocyclic ring instead of a sugar ring, and is also notable for its relatively complex bioactivation pathway in which abacavir essentially acts as a pro-drug (fig. 1) [2]. Following cellular entry of the parent drug – achieved primarily by passive diffusion – abacavir is phosphorylated to its monophosphate derivative by adenosine phosphotransferase and then deaminated by a novel enzyme to form carbovir monophosphate [2]. Formation of the ac-

N

N H 2N

N

N H 2N O

N

N

HO

HO O

P

a

adenosine phosphotransferase

HO

O

P

O

HO

O

carbovir-MP

abacavir-MP

N

NH

N

N H2 N

N

N

H

O

abacavir

alcohol dehydrogenase

N

N

N

N

O

HO

N

N

NH

N

b

H 2N O

O

NH

N

HN

N

N

HO

abacavir

H 2N

O

NH

NH

H 2N

N

N

O

O

aldehyde intermediate

HO

alcohol dehydrogenase

O

carboxylic acid

Fig. 1. Abacavir metabolism and antiretroviral activity. a Abacavir is shown here, along with its major metabolic derivatives. Following intracellular phosphorylation, NRTI drugs compete with naturally occurring nucleotides for utilisation by HIV reverse transcriptase. Incorporation of the ‘false’ (NRTI) substrate then causes DNA

chain termination, as these nucleotides lack a 3 hydroxyl group (closed arrow) that is critical for chain elongation. b Catabolism of abacavir to form a carboxylic acid derivative via an aldehyde intermediate accounts for
30% of urinary abacavir excretion. Shaded area shows the putatively immunogenic aldehyde derivative.

tive carbocyclic guanosine analogue moiety, carbovir triphosphate, is then accomplished via sequential phosphorylation (fig. 1a). Hence, an efficient bioactivation pathway for abacavir exists within the drug’s target cells (lymphocytes) without a requirement for facilitated transport or extracellular metabolism [2].

Abacavir catabolism pathways may also be pertinent to the development of abacavir hypersensitivity reactions [3]. In particular, the oxidation of abacavir to form an aldehyde intermediate and finally 5-carboxylate, in a process catalysed by cytosolic alcohol dehydrogenase (fig. 1b), appears to be an important determinant of abaca-

96

Nolan
Almeida
Phillips
Mallal

vir’s immunogenicity. Here, the involvement of alcohol dehydrogenase isozymes that are sensitive to 4-methylpyrazole inhibition (i.e., class I and potentially class II or IV isozymes) provides for a widespread capacity for abacavir oxidation in epithelial tissues including the skin [4], as well as in the liver.

The Clinical Syndrome of Abacavir Hypersensitivity

Drug hypersensitivity represents the major treatment-limiting toxicity for abacavir use, occurring in approximately 5–8% of recipients within 6 weeks of commencing therapy [5–7]. Diagnosis of this multi-system inflammatory syndrome has been primarily based on clinical criteria, and is therefore potentially complicated by the use of concurrent drug therapy or the presence of infections or inflammatory syndromes. The clinical classification of abacavir hypersensitivity includes at least two symptoms of fever, rash, nausea, vomiting, headache, respiratory and gastrointestinal symptoms, lethargy, myalgia, or arthralgia occurring less than 6 weeks after exposure, and resolving within 72 h of withdrawal of the drug. More recently, epicutaneous patch testing, involving the application of 1 and 10% concentrations of abacavir to the skin, has proven to be a useful adjunctive method for confirming suspected abacavir hypersensitivity [8]. This diagnostic approach is likely to have the advantage of revealing abacavir-specific immune reactivity with a greater degree of specificity than can be provided clinically [8]. A history of definite abacavir hypersensitivity precludes any further use of abacavir, as rechallenge can evoke severe and potentially life-threatening reactions. As elegantly demonstrated in a recent retrospective analysis of the French Pharmacovigilance Database encompassing a 15-year period [9], some of the clinical and laboratory features of abacavir drug hypersensitivity (n = 68 in this

Abacavir Hypersensitivity

study) are unusual when compared with other idiosyncratic drug reactions. For example, in abacavir hypersensitivity reactions the predilection for gastrointestinal system involvement (diarrhoea, abdominal pain, nausea) and/or upper respiratory symptoms (cough, pharyngitis, tachypnoea) in the majority of cases is notable, whereas these symptoms are relatively rare in other multi-system drug reactions [9]. On the other hand, abacavir hypersensitivity is characterised by an absence of eosinophilia (!10%) and liver function abnormalities (!20%), as well as a relatively short time to onset after initiating treatment (median 7–10 days). The high frequency of fever and generalised non-bullous rash (both 170% in abacavir hypersensitivity) is perhaps the most consistent feature across the spectrum of severe drug reactions noted in this study and in the literature more generally.

Genetically Determined Susceptibility to Abacavir Hypersensitivity Reactions

The involvement of genetic factors in the pathogenesis of abacavir hypersensitivity was initially suggested by clinical reports of a multi-system inflammatory syndrome that affected only a proportion of susceptible abacavir-treated individuals, and only in the earliest phases of treatment. Cases of familial predisposition, and significantly decreased frequency in individuals of African origin were also consistent with this possibility. Subsequent research published from 2002 onwards has revealed a strong predictive association between carriage of HLA-B*5701 and abacavir hypersensitivity reactions in Caucasian and Hispanic ethnic groups [10–14], sufficient to stratify the predicted risk of abacavir hypersensitivity by identifying low-risk (!1%) and high-risk (170%) individuals according to the presence or absence of the HLA-B*5701 allele. Accordingly, in a recent study in which HLA-B*5701 genetic screening was utilized prospectively among aba-

97

HLA residues

97

a

114

116

15

6

HLA-B*5701

Valine

Aspartic acid

Serine

Leucine

HLA-B*5702

Valine

Asparagine

Tyrosine

Arginine

HLA-B*5703

Valine

Asparagine

Tyrosine

Leucine

HLA-B*5801

Arginine

Aspartic acid

Serine

Leucine

T cell receptor 97 P4

P5

P9

P2

116 114 156

116

HLA molecule

b

Fig. 2. HLA-B*5701 and abacavir hypersensitivity. a Alignment of variant amino acid positions between HLA-B*5701 and closely related alleles within the B17 serological family. Position 116 is a serine in the drug hypersensitivity-associated alleles HLA-B*5701, HLA-B*5801 and HLA-B*1502. b Left panel: Representation of class I HLA-peptide interactions, with a putative 9 amino acid epitope occupying the HLA-peptide-binding groove.

Approximate positions of residues 97, 114 and 116 are displayed. Right panel: Cross-section of the HLA-peptide-T-cell receptor interaction, revealing the importance of HLA residues in contact with peptide anchor residues P2 and P9 for HLA peptide binding. The 116 position lies at the base of the HLA F-pocket involved in P9 side-chain interactions.

cavir-naive patients in the Western Australian HIV cohort during a study period from January 2002 until July 2005 (n = 260), there were no cases of drug hypersensitivity among 148 HLAB*5701-negative abacavir recipients. (On the other hand, all 3 HLA-B*5701-positive patients who commenced abacavir in this study developed definite abacavir hypersensitivity [14].) HLA-B*5701 belongs to a group of B17 subtypes defined by serological techniques that also includes HLA-B*58 and at least 9 different HLAB*57 subtypes (HLA-B*5701 to B*5709) identified by high-resolution sequence-based typing (SBT). We have examined the frequency of these HLA genotypes among abacavir-tolerant individuals

in the Western Australian abacavir-exposed cohort (n = 360), demonstrating the presence of HLA-B*5801 (n = 8) and HLA-B*5802 (n = 2) as well as HLA-B*5702 (n = 2) and HLA-B*5703 (n = 1) among individuals who tolerated the drug for more than 6 weeks. These alleles are 190% similar, varying only at amino acid residues 96, 114, 116 and 156 within the peptide-binding domain, and have distinct but overlapping peptide-binding repertoires (fig. 2). Hence, the specificity of the HLA-B*5701 allele for abacavir hypersensitivity points to a specific accommodation of abacavir or its haptenated derivative into the deepseated peptide-binding pocket of HLA-B*5701 (fig. 2a, b).

98

Nolan
Almeida
Phillips
Mallal

CTL TCR

viral protein (or haptenated endogenous peptide)

CD8

DRUG

5

(fixed exposure)

GOLGI

PATHOGEN (adaptable exposure)

PROTEASOME COMPLEX

1 4

TAP

2

ER

3 peptide epitopes

MHC I

p-MHC I

Fig. 3. Antigen processing and HLA class I-restricted immune responses. Cell-mediated immune responses involving the HLA class I (endogenous) pathway typically involve the processing of antigens (pathogen-derived or drug-derived) to form short peptide epitopes that are then presented on specific HLA molecules to CD8+ T lymphocytes. This pathway involves (1) peptide cleavage in the proteasome complex; (2) translocation of pep-

tides into the endoplasmic reticulum (ER), and (3) loading of peptides onto HLA molecules, both involving the transporter of antigen processing (TAP) and other ERresident chaperones that may further optimise peptide folding/editing; (4) transport via the secretory pathway to the plasma membrane, and (5) cell surface presentation of the peptide-HLA complex to T-cell receptors.

It would be anticipated that the serine-116 and aspartic acid-114 positions that define differences between the HLA-B*5701 and HLA-B*5703 molecules (fig. 2a) might be important determinants of the repertoire and composition of peptide epitopes recognised by these molecules, as they lie at the base of the HLA-peptide-binding domain that binds the C-terminal ‘P9 anchor’ residues that dictate the affinity of peptide-MHC binding

(fig. 2b). This already provides something of a conundrum, as high-resolution crystal structures of HLA-B*5703 molecules bound to HIV epitopes have revealed a similar preference for bulky aromatic amino acids (tryptophan, phenylalanine) at the P9 position for epitopes bound to HLA-B*5701, B*5702 or B*5703, with little evidence for involvement of either the 114 or 116 position in these HLA-peptide interactions [15].

Abacavir Hypersensitivity

99

It is also notable that the selection of amino acids at position 116 of HLA-B molecules has been shown to influence interactions between HLA and the transporter of antigen processing (TAP) protein [16] which is involved in translocating peptides into the endoplasmic reticulum and optimising HLA-peptide interactions (fig. 3). In general, the presence of serine at position 116 tends to abrogate the TAP dependence of these HLA molecules [16], which may increase the propensity for the presentation of empty HLA and/ or suboptimal conformations of HLA-peptide complexes. (On the other hand, this lack of TAP dependence among selected HLA alleles may represent a valuable adaptation in response to an array of viral-encoded proteins that actively inhibit antigen processing via TAP, reflecting the context dependence of ‘evolutionary advantage’ [17].) It is at present unknown if the TAP dependence of HLA molecules plays a role in determining which alleles are more likely to present drugderived antigens, although it is intriguing to note that serine-116 is common to (but not exclusive to) the three HLA-B alleles most strongly associated with drug hypersensitivity: HLA-B*5701 (abacavir), HLA-B*5801 (allopurinol) and HLAB*1502 (carbamazepine). The finding that HLA-B*5701 provides an exquisitely fine-tuned restriction factor for abacavir-specific drug reactions has several implications relevant to clinical practice. First, it highlights the importance of high-resolution HLA typing methods to identify the HLA-B*5701 allele within the B17 serological group. This is particularly important in ethnically diverse populations, as HLA-B*58 is among the most common HLA-B alleles among Asian populations (
10%) while HLA-B*5703 is relatively specific to populations of African origin [18]. In both cases these alleles are more frequent than HLA-B*5701 at a population level, making it more likely that a ‘B17’ result for serological HLA typing will represent a false positive than a true positive finding [18]. Highresolution SBT methods for HLA-B allele identi-

100

fication are available in transplantation laboratories, but for the moment this method remains relatively specialised, costly and may be limited by longer turn-around times. To this end, we have developed a high-resolution genetic testing using a sequence-specific primer (SSP) assay for HLA-B*5701 identification which provides a feasible method for use in routine molecular biology laboratories [19]. This multiplexed PCR test can differentiate between highly similar alleles of HLA-B*5701 including HLA-B*5702 and HLAB*5703, and although requiring sample processing (DNA extraction) can be batched to improve cost-effectiveness, and can generally be performed within 3–4 h. Second, these findings also raise the possibility that genetic markers that are closely linked to HLA-B*5701 within an extended MHC haplotype are also involved in conferring susceptibility to abacavir hypersensitivity. It is highly plausible that the specific (and highly penetrant) association between abacavir drug reactions and the HLA-B*5701 allele reflects not only an HLArestricted drug-specific immune response, but also implicates other genetically variable immunoregulatory proteins encoded within the MHC region on the HLA-B*5701-specific extended haplotype (denoted the 57.1 ancestral haplotype). Here, candidate genetic factors that exhibit extremely strong linkage disequilibrium with HLA-B*5701 include cellular chaperones (e.g. heat-shock proteins), pro-inflammatory cytokines (e.g. TNF- ) and proteins involved in the stress response (MHC class I chain-related genes, MIC-A and MIC-B). We therefore examined genetic markers specific for the 57.1 ancestral MHC haplotype, identifying a parsimonious susceptibility region that included HLA-B*5701 and a haplospecific allotype of the complement C4 locus (C4A6) [12]. Using a panel of patients recombinant for the 57.1 ancestral haplotype, the susceptibility region was further narrowed to the genetic region bounded by HLA-B*5701 and a non-synonymous variant within the heat-shock

Nolan
Almeida
Phillips
Mallal

protein 70 (HSP70) gene cluster which was identified in 190% of the abacavir hypersensitivity cases compared with !1% of the abacavir-tolerant control population [12]. The functional significance of the identified HSP70 SNP is yet to be fully understood, although it has been proposed that the encoded methionine-to-threonine substitution at position 493 within the HSP70Hom peptide-binding domain may facilitate HLAB*5701-restricted presentation of abacavir-specific ligands by this cellular chaperone. The observation that multiple genetic determinants within a common haplotype structure may contribute to disease susceptibility may also provide some insight into the lack of association between HLA-B*5701 and abacavir hypersensitivity in non-Caucasian (or Hispanic) populations to date, where it appears that the low frequency HLA-B*5701 allele does not contribute to increased risk [13]. It is notable in this context that populations of African origin do not demonstrate the same MHC haplotype in association with HLA-B*5701 carriage, being more likely to co-express HLA-C *0701 rather than the HLAC *0602 allele which commonly defines the Caucasian 57.1 ancestral haplotype [20].

Immunological Markers of Abacavir Hypersensitivity: Drug-Specific Immune Responses

It would be anticipated in the case of abacavir hypersensitivity (and other drug hypersensitivity reactions with evidence of profound class I HLA restriction) that the effector cell populations for the drug-specific immune response are CD8+ cytotoxic T lymphocytes with a strong bias towards a pro-inflammatory (T1-type) cytokine profile. Accordingly, investigations of the abacavir hypersensitivity response have shown that patients who had an adverse reaction had elevated TNF- and IFN- levels after ex vivo exposure to the drug [12], with additional evidence of dra-

Abacavir Hypersensitivity

matic increases in intracellular TNF levels in CD14+ monocytes following stimulation of peripheral blood mononuclear cells (PBMCs) with abacavir [12]. We have also been able to demonstrate that chronic in vitro exposure of PBMCs from genetically susceptible abacavir-naive patients is sufficient to induce an immune response characterised by elevated IFN- production. Unpublished data has also revealed that inhibiting 4-MP-sensitive alcohol dehydrogenase isoforms significantly abrogates abacavir-specific immune responses, implicating the products of oxidative abacavir and/or carbovir metabolism (such as  unsaturated aldehydes) as antigenic determinants (see fig. 1b). These findings are in keeping with earlier in vitro data from Walsh et al. [3], although in this case the development of specific anti-abacavir antibodies in rabbits was the observed endpoint. As discussed earlier, an additional method of measuring abacavir-specific responses relies on the application of an epicutaneous patch test, a technique that has been shown to be highly specific for HLA-B*5701-associated hypersensitivity reactions in West Australian HIV and Canadian cohorts [8, 12]. Given that the clinical classification of abacavir hypersensitivity is based on the presence of two or more clinical symptoms that may also be present in other drug reactions (e.g. nevirapine, sulphamethoxazole) or in unrelated illnesses (e.g. viral illnesses, immune restoration syndromes), the use of this diagnostic strategy in suspected cases of abacavir hypersensitivity offers evidence of a drug-specific immune response that strongly supports the diagnosis. It must be remembered, however, that a positive epicutaneous patch test result requires the presence of an immune response that has been primed by drug exposure, and (unlike genetic testing for HLAB*5701) is not predictive for abacavir hypersensitivity reactions.

101

Pharmacogenetic Screening Tests: A Move towards Personalised HIV Treatment

HLA-B*5701-Associated Abacavir Hypersensitivity as an Immunological Model

Following the identification of HLA-B*5701 as a clinically useful marker of susceptibility to abacavir hypersensitivity, particular consideration has been given to the fact that HLA typing is generally a highly specialised and often expensive diagnostic technique that is offered in a limited number of laboratories involved in transplantation medicine. Hence, there has been an increased focus on developing accurate, robust and cost-effective methods for identifying the HLAB*5701 allele that may be performed in HIV laboratories in a timely manner. These include the molecular HLA typing methods described earlier, which enable targeted assessment of the HLA-B*5701 allele [19]. Further refinements of these techniques could also provide for use of whole blood samples without requiring DNA extraction (as already demonstrated in the case of HLA-B*27 testing in the diagnosis of ankylosing spondylitis [21]). A serological approach to HLA screening can also be provided by flow cytometry methods, utilising B17 monoclonal antibody (mAb) hybridisation to HLA cell surface antigens on CD45+ leukocytes [22]. The B17 mAb binds to all subtypes within the HLA-B*57 and HLA-B*58 molecule families, so subsequent molecular testing is necessary to exclude non-HLA-B*5701 alleles in the case of a positive serological test result. This technology would allow for rapid screening of individuals, and is particularly valuable as a screening tool that can rapidly identify ‘low-risk’ patients with a negative test result (i.e. 190% of those tested). In addition, flow cytometry laboratories are established in most HIV laboratories and are set up for processing of samples daily, thereby ensuring rapid sample analysis that is practical and economical for most diagnostic centres.

Readers of this text will already be familiar with the proposed models of immune pathogenesis underlying drug hypersensitivity reactions in humans. For the most part, these depend on Tlymphocyte-dependent immunological responses that may be considered in light of classical representations of antigen presentation and recognition (fig. 3). In this context, drug hypersensitivity syndromes with a highly predictable genetic basis are most likely to be associated with specific HLA alleles, as opposed to drug reactions that rely on direct T-cell receptor activation [reviewed in 23] in which the repertoire of drugreactive cells (among an estimated 25 ! 106 different TCR combinations in an adult human) is not genetically determined to the same degree. The more stringent conditions for HLA peptide binding provided by highly polymorphic HLA-B molecules, compared with the more cross-reactive and/or less genetically diverse HLA-A, HLAC and class II HLA molecules (fig. 2b), might also explain the observation that HLA-B associations have revealed the most highly genetically-determined drug hypersensitivity syndromes. In the case of abacavir hypersensitivity, the favoured pathogenic model involves HLA-specific presentation of a peptide epitope that has been rendered antigenic by an abacavir metabolite (i.e. the ‘hapten’ model), within an immune environment that may also be shaped by genetic factors that are coinherited along with HLA-B*5701 on the 57.1 ancestral haplotype. It also seems likely that HLApeptide interactions specifically involve HLA residues located at the base of the peptide-binding groove such as the serine-116 position, again favouring a model of peptide antigen processing and HLA-restricted presentation rather than direct MHC drug binding on the cell surface [23]. Finally, the available experimental evidence suggests that an aldehyde derivative may play an important role in the creation of an antigenic pep-

102

Nolan
Almeida
Phillips
Mallal

tide [3], possibly by facilitating the formation of protein adducts, although at this stage the drugspecific antigen has not been identified. Further insights into the phenomenon of abacavir hypersensitivity will hopefully illuminate some aspects of basic immunology as well as pharmacogenetics, providing (as eloquently described by Amalio Telenti [24] in an early editorial review of this topic) ‘… the first drops before a downpour’. From a clinical management perspective, the antiretroviral efficacy, addition of one-pill fixed-dose combinations and apparent favourable long-term toxicity profile, have contributed to the common use of abacavir in HIV treatment regimens [1]. Given the power of ge-

netic screening to sharply differentiate high versus low risk of the most frequent treatment-limiting drug side effect, it will be interesting to observe to what extent this pharmacogenetic approach will be adopted in the field of HIV medicine in the near future. Prospective, randomised trials to specifically test the utility of genetic screening in abacavir-naive populations are underway which will add further clarity to this issue and its application and generalizability into routine clinical practice. In this instance, a pharmacogenetic approach to risk stratification, although not a substitute for clinical pharmacovigilance, would appear to pave the way for the more intelligent and cost-effective use of abacavir.

References 1 Dando TM, Scott LJ: Abacavir plus lamivudine: a review of their combined use in the management of HIV infection. Drugs 2005;65:285–302. 2 Faletto MB, Miller WH, Garvey EP, St Clair MH, Daluge SM, Good SS: Unique intracellular activation of the potent anti-human immunodeficiency virus agent 1592U89. Antimicrob Agents Chemother 1997;41:1099–1107. 3 Walsh JS, Reese MJ, Thurmond LM: The metabolic activation of abacavir by human liver cytosol and expressed human alcohol dehydrogenase isozymes. Chem Biol Interact 2002;142:135–154. 4 Lockley DL, Howes D, Williams FM: Cutaneous metabolism of glycol ethers. Arch Toxicol 2005;79:160–168. 5 Hetherington S, McGuirk S, Powell G, Cutrell A, Naderer O, Spreen B, Lafon S, Pearce G, Steel H: Hypersensitivity reactions during therapy with the nucleoside reverse transcriptase inhibitor abacavir. Clin Ther 2001;23:1603–1614. 6 Cutrell AG, Hernandez JE, Fleming JW, Edwards MT, Moore MA, Brothers CH, Scott TR: Updated clinical risk factor analysis of suspected hypersensitivity reactions to abacavir. Ann Pharmacother 2004;38:2171–2172.

Abacavir Hypersensitivity

7 Peyriere H, Guillemin V, Lotthe A, et al: Reasons for early abacavir discontinuation in HIV-infected patients. Ann Pharmacother 2003;37:1392–1397. 8 Phillips EJ, Wong GA, Kaul R, Shahabi K, Nolan DA, Knowles SR, Martin AM, Mallal SA, Shear NH: Clinical and immunogenetic correlates of abacavir hypersensitivity. AIDS 2005;19:979–981. 9 Peyriere H, Dereure O, Breton H, Demoly P, Cociglio M, Blayac JP, HillaireBuys D, Network of the French Pharmacovigilance Centers: Variability in the clinical pattern of cutaneous sideeffects of drugs with systemic symptoms: does a DRESS syndrome really exist? Br J Dermatol 2006;155:422–428. 10 Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, Sayer D, Castley A, Mamotte C, Maxwell D, James I, Christiansen FT: Association between presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet 2002;359:727–732. 11 Hetherington S, Hughes AR, Mosteller M, Shortino D, Baker KL, Spreen W, Lai E, Davies K, Handley A, Dow DJ, Fling ME, Stocum M, Bowman C, Thurmond LM, Roses AD: Genetic variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet 2002;359:1121–1122.

12 Martin AM, Nolan D, Gaudieri S, Almeida CA, Nolan R, James I, Carvalho F, Phillips E, Christiansen FT, Purcell AW, McCluskey J, Mallal S: Predisposition to abacavir hypersensitivity conferred by HLA-B*5701 and a haplotypic Hsp70-Hom variant. Proc Natl Acad Sci USA 2004;101:4180–4185. 13 Hughes AR, Mosteller M, Bansal AT, Davies K, Haneline SA, Lai EH, Nangle K, Scott T, Spreen WR, Warren LL, Roses AD, CNA30027 Study Team/ CNA30032 Study Team: Association of genetic variations in HLA-B region with hypersensitivity to abacavir in some, but not all, populations. Pharmacogenomics 2004;5:203–211. 14 Rauch A, Nolan D, Martin A, McKinnon E, Almeida C, Mallal S: Prospective genetic screening decreases the incidence of abacavir hypersensitivity reactions in the Western Australian HIV cohort study. Clin Infect Dis 2006; 43:99–102. 15 Stewart-Jones GBE, Gillespie G, Overton IM, Kaul R, Roche P, McMichael AJ, Rowland-Jones S, Jones EY: Structures of three HIV-1 HLA-B*5703-peptide complexes and identification of related HLAs potentially associated with longterm non-progression. J Immunol 2005;175:2459–2468.

103

16 Prilliman KR, Crawford D, Hickman HD, Jackson KW, Wang J, Hildebrand WH: Alpha-2 domain polymorphism and class I peptide loading. Tissue Antigens 1999;54:450–460. 17 Nolan D, Gaudieri S, Mallal S: Host genetics and viral infections: immunology taught by viruses, virology taught by the immune system. Curr Opin Immunol 2006;18:413–421. 18 Nolan D, Gaudieri S, Mallal S: Pharmacogenetics: a practical role in predicting antiretroviral drug toxicity? J HIV Ther 2003;8:36–41.

19 Martin AM, Nolan D, Mallal S: HLAB*5701 typing by sequence-specific amplification: validation and comparison with sequence based typing. Tissue Antigens 2005;65:571–574. 20 Cao K, Hollenbach J, Shi X, Shi W, Chopek M, Fernandez-Vina MA: Analysis of the frequencies of HLA-A, B and C alleles in the five major ethnic groups of the United States reveals high levels of diversity in these loci and contrasting distribution patterns in these populations. Hum Immunol 2001;62:1009– 1030. 21 Sayer DC, Cassell HS, Christiansen FT: HLA-B*27 typing by sequence-specific amplification without DNA extraction. Mol Pathol 1999;52:300–301.

22 Martin AM, Krueger R, Almeida CA, Nolan D, Phillips E, Mallal S: A sensitive and rapid alternative to HLA typing as a genetic screening test for abacavir hypersensitivity syndrome. Pharmacogenet Genomics 2006;16: 353–357. 23 Gerber BO, Pichler WJ: Cellular mechanisms of T-cell-mediated drug hypersensitivity. Curr Opin Immunol 2004; 16:732–737. 24 Telenti A, Aubert V, Spertini F: Individualising HIV treatment – pharmacogenetics and immunogenetics. Lancet 2002;359:722–723.

Prof. Simon Mallal Centre for Clinical Immunology and Biomedical Statistics 2nd Floor, North Block, Royal Perth Hospital Perth 6000 (Australia) Tel. +61 89 224 2899, Fax +61 89 224 2920 E-Mail [email protected]

104

Nolan
Almeida
Phillips
Mallal

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 105–114

Genetics of Severe Drug Hypersensitivity Reactions in Han Chinese Shuen-Iu Hung a
Wen-Hung Chung a, b
Yuan-Tsong Chen a a

Institute of Biomedical Sciences, Academia Sinica, and b Department of Dermatology, Chang Gung Memorial Hospital, Taipei, Taiwan

Abstract

Introduction

Drug hypersensitivity, an immune-related idiosyncratic adverse reaction, was historically referred to as being unpredictable. However, recent studies in Han Chinese have revealed that several types of severe drug hypersensitivity reactions have a strong genetic predisposition and might be predicted, particularly with the use of the genes coding for major histocompatibility complex (MHC), a key molecule for immune response. The genetic predisposition is drug- and phenotype-specific, HLA-B*1502 is associated with carbamazepine-induced Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN), and HLA-B*5801 with allopurinol-SJS/TEN/hypersensitivity syndrome. These genetic associations are not limited to the Han Chinese, and their significance in other ethnic populations depends on the allele frequency of the populations. The genetic studies in Han Chinese support the concept that MHC-restricted presentation of drug is involved in the pathogenesis of immune-mediated severe drug hypersensitivity reactions. These genetic markers have the potential to be used for further development of tests to identify individuals at risk for these drug-related life-threatening conditions, as well as for an increased understanding of the pathogenesis of these clinical syndromes. Copyright © 2007 S. Karger AG, Basel

Clinically, adverse drug reactions (ADR) are commonly classified into two types: type A and type B [1]. Type A reactions are predictable from the known pharmacology of the drug and often dose-dependent. By contrast, type B reactions are historically referred to as being unpredictable and dose-independent idiosyncratic reactions. Severe drug hypersensitivity reactions, such as Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN) and hypersensitivity syndrome (HSS), belong to the type B ADR. Despite numerous reports on the familial occurrence of severe drug hypersensitivity and cases occurring in identical twins [2–5], identifying genetic factors has yielded very few genes and/or markers that have sufficient sensitivity/specificity and predictive values which can be used clinically to identify individuals at risk for the serious ADRs. A successful disease genetic study is highly dependent on the well-characterized phenotypes and endophenotypes, genetic background of the study population, and the quality and resolution of the genetic markers. Recent advances in genomic and genetic research with a vast increase

in numbers of DNA polymorphic markers offer the opportunity to examine the genetic association of the disease at the whole genome level. Finding the genetic susceptibility to severe drug hypersensitivity should improve our understanding of the pathogenic mechanisms mediating these serious adverse events and has the potential to provide clinically useful pharmacogenetic tests to identify individuals who are at risk for ADRs and thereby preventing drug toxicity. In this chapter, we describe our approach of genetic study of severe drug hypersensitivity reactions with particular emphasis on two drugs: carbamazepine (CBZ), a commonly prescribed anticonvulsant, and allopurinol, a commonly prescribed medication for gout and hyperuricemia. The studies were primarily done in Han Chinese but their implication should extend to other ethnic populations.

Clinical Features of Severe Drug Hypersensitivity Reactions

Most drug hypersensitivity reactions are accompanied by cutaneous manifestations. These skin manifestations may range from mild maculopapular eruption (MPE), urticaria, to fixed drug eruption, and with increasing severity, to HSS, SJS, and TEN (also called Lyell’s syndrome). HSS is characterized by systemic manifestations with multiorgan involvement, in addition to the skin rashes. SJS is characterized by a rapidly developing blistering exanthema of macules and targetlike lesions accompanied by mucosal involvement. TEN has similar presentations as SJS with an even more extensive skin detachment and a higher mortality rate (30–40%) [6]. In Han Chinese residing in Taiwan, the incidence of SJS/TEN is estimated to be 8 cases per million person-years, which is higher than that reported in Western countries with a range of 1– 6 cases per million person-years [7–9]. Variation in incidence of SJS/TEN between countries might

106

be explained by a difference in the drug utilization profile, such as types of drugs and how frequently they are used, and the difference in ethnic genetic background leading to a different genetic susceptibility.

Common Culprit Drugs for HSS/SJS/TEN Vary with Populations

More than 100 drugs have been implicated as causes of HSS/SJS/TEN [6, 10]. However, the commonly offending drugs vary among different ethnic populations (table 1). In Western countries, the most common causes of SJS/TEN are nonsteroidal anti-inflammatory drugs (NSAIDs) and sulfonamides, followed by phenobarbital, aminopenicillins and allopurinol [10]. By contrast, the leading cause of SJS/TEN in Southeast Asian countries (including India, Malaysia, Singapore, Taiwan and Hong Kong) is CBZ (table 1) [11]. It is interesting to note that in Western countries CBZ causes more HSS than SJS/TEN. Allopurinol is also a frequent cause of severe drug hypersensitivity reactions [11, 12]. However, allopurinol-induced severe cutaneous adverse drug reactions (SCAR, including HSS, SJS and TEN) do not appear to have a geographic or ethnic predilection.

Genetic Approach to Discover Genotypes Underlying Clinical Phenotypes

Linkage analysis and association study are two powerful approaches in uncovering genes associated with human diseases. Classical linkage analysis followed by positional cloning remain the methods of choice for identifying rare, high-risk, disease-associated mutations, owing to the clear inheritance patterns they display. Although familial occurrence of drug hypersensitivity reactions has been reported, the inheritance pattern is often not clear, nor are the families sufficiently

Hung ⴢ Chung ⴢ Chen

Table 1. Allele frequencies of HLA-B*1502, B*5801, and common culprit drugs for SJS/TEN among different populations Country/ continent

Ethnic background

HLA-B*1502

HLA-B*5801

The three most common culprit drugs [10, 11]

Taiwan

Han Chinese

2.7–5.9%

8.8–10.9%

1. Carbamazepine 2. NSAIDs 3. Allopurinol

Singapore

Han Chinese Malay Indonesians

11.6% 8.4% 8.2%

10.4% 5% 3.1%

1. Carbamazepine 2. NSAIDs 3. Allopurinol

Malaysia

Malay Indian Han Chinese

8.4% unknown 11.6%

5% unknown 10.4%

1. Carbamazepine 2. Aminopenicillins 3. Sulfonamides/phenytoin

Thailand

Thai

8.5%

7.7%

data not available

India

Indian

1–6%

3–15%

1. Carbamazepine/phenytoin 2. Sulfonamides 3. Phenobarbital

China

South Han North Han

7.1% 1.9%

8.9% 2.9%

data not available

Japan (Central)

Japanese

0.1%

0.4%

data not available

Europe

Caucasian

0%

1~6%

1. NSAIDs 2. Sulfonamides 3. Phenobarbital

Africa

African

0–3%

2–4%

data not available

NSAIDs = Non-steroidal anti-inflammatory drugs; SJS = Stevens-Johnson syndrome; TEN = toxic epidermal necrolysis.

large enough for the linkage study. Furthermore, the hypersensitivity status of many family members may not be available and cannot be reliably determined without challenging the individuals with the culprit drug.

Association Study to Detect Genetic Susceptibility to Severe Drug Hypersensitivity Reactions

An association study requires collection of cases and appropriate controls. In this pharmacogenetic study, we used two control groups. The first

Genetics of Drug Hypersensitivity in Han Chinese

control group, called tolerant control, included the patients who received drugs for an extended duration but without evidence of adverse reactions. The second control group consisted of 93 healthy subjects randomly selected from the Taiwan Cell and Genome Bank under a nationwide population study, in which 3,312 Han Chinese descendants were recruited based on their geographic distribution across Taiwan [13]. There was no self-report of ADR by any of these 93 subjects. All participants were Han Chinese residing in Taiwan. The Han Chinese forms the largest ethnic group in Taiwan, making up roughly 98% of the population and is commonly categorized

107

HLA-DRB

30 20 10 0 29.9

into three major ethnic groups – Minna, Hakka, and mainlanders. There is a high degree of genetic homogeneity in the general Taiwan Han population as no statistical significant population stratification has been detected in the three Han groups [14]. Selection of Candidate Genes We chose two categories of genes as candidate genes to study the genetic susceptibility for severe drug hypersensitivity reactions. The first category included genes involved in the bioactivation or detoxification of the drugs of interest. Specifically for CBZ, we included single-nucleotide polymorphisms (SNPs) in CYP3A4, CYP2B6, CYP2C8, CYP2C9, CYP1A2 and EPHX1 [9, 15]; for allopurinol, we included SNPs in xanthine oxidase, aldehyde oxidase, purine nucleoside phosphorylase, and hypoxanthine phosphoribosyl transferase [12]. In addition to the potential defects in the drug metabolism, immune reaction is also a key feature of drug hypersensitivity. The clinical, histopathological, immunocytological and functional findings support the concept that drug hypersensitivity is a specific drug sensitivity reaction initiated by cytotoxic lymphocytes [16]. Prior in vitro studies suggest that the drug presentation is major histocompatibility complex (MHC) class I re-

108

HLA-C rs3130690 HLA-B MICA rs2848716 rs750332

HLA-A 40

–log10 (p value)

Fig. 1. Association screening of 220 SNPs in the MHC region with CBZ-induced SJS/TEN. On the x-axis, 220 SNPs are ordered by their physical positions (29.9–33.9 Mb) on chromosome 6p21.3. On the y-axis, the –log10 p values were calculated by comparison of the allele frequencies between 60 cases of CBZ-induced SJS/TEN and 144 tolerant controls using the Cochran-Armitage trend test.

30.9

31.9 32.9 Chromosome position (Mb)

33.9

stricted, and there is a clonal expansion of CD8+ cytotoxic T cells, which induce effector cytotoxic responses [16, 17]. Therefore, our second category of the candidate genes were genes encoding immune-related molecules, including 220 SNPs from 4 Mb of the MHC region on chromosome 6p21.3 and 1200 SNPs of apoptosis-related molecules (death-associated proteins, caspases) and immune mediators (complement components, immune cell receptors and cytokines) [12, 15]. Association Screen for Candidate Gene SNPs Of the total of
800 SNPs screened (220 SNPs in the MHC region and
600 SNPs of immune-related genes and drug-metabolizing genes), 37 SNPs in the MHC region were found to have a significant association (p ! 0.001) in patients with the CBZ-induced SJS/TEN as compared to the tolerant group (fig. 1) [15]. Among them, 7 SNPs located between HLA-DRA and HLA-C (physical position 31.3–32.5 Mb, chromosome 6) showed strong associations (p ! 10 –10). The three most important SNPs were rs3130690, rs2848716 and rs750332, located near HLA-B, MICA and BAT3 genes. The rs3130690, an intergenic SNP with 36 kb telomeric to the HLA-B locus, demonstrated the strongest association with CBZ-induced SJS/TEN (Pc = 1.29 ! 10 –39) (fig. 1). A similar finding of a strong association of the

Hung ⴢ Chung ⴢ Chen

SNPs in the MHC region was also seen in the case of allopurinol-induced SCAR [12]. No SNPs in genes involved in the drug metabolism demonstrated a significant association. HLA Allele Frequency Since the most significant association was seen with the SNPs in the MHC region, we genotyped the individual HLA alleles. In the case of CBZinduced SJS/TEN, we found that 59 of 60 patients (98.3%) carried the HLA-B*1502 allele, while only 4.2% (6/144) of the CBZ-tolerant group carried the allele (Pc = 1.6 ! 10 –41, odds ratio (OR) = 1,357 [95% CI, 193.4–8838.3]) (table 2) [15]. The only CBZ-SJS patient who did not have the B*1502 allele had another HLA-B15 allele instead: HLA-B*1558 [15]. The HLA-Cw*0801 and MICA*019 alleles flanking the HLA-B*1502 showed strong linkage disequilibrium with HLAB*1502, and were present in 93.3% (56/60) and 95% (57/60) of SJS/TEN patients, respectively. An extended B*1502 haplotype formed by polymorphic alleles (A*1101-Cw*0801-HLABC-CA*119rs3130690T-B*1502-MICA*019-DRB1*1202) had a strong association with CBZ-induced SJS/TEN [9, 15]. In the case of allopurinol-induced SCAR, alleles HLA-A*3303, B*5801, Cw*0302 and DRB1*0301 occurred at significantly increased frequencies among the allopurinol-SCAR patients compared to the two control groups [12]. In particular, the HLA-B*5801 was present in all 51 (100%) patients with allopurinol-SCAR but in only 15% (20/135) of the allopurinol-tolerant group (Pc = 4.7 ! 10 –24, OR = 580.3 [95% CI, 34.4–9780.9]) (table 2), and 20% (19/93) of the general population (Pc = 8.1 ! 10 –18, OR = 393.5 [95% CI, 23.2–6665.26]) [12]. The HLA-B*5801 extended haplotype was formed by conserved alleles at closely linked loci as HLA-A*3303Cw*0302-B*5801-DRB1*0301. This extended haplotype was present in 21 (41%) of the 51 patients with allopurinol-SCAR, 7% of the tolerant patients, and 10% of the healthy subjects [12].

Genetics of Drug Hypersensitivity in Han Chinese

Are Genes Other than HLA-B Associated with Severe Drug Hypersensitivity Induced by CBZ and Allopurinol?

Fine Mapping To determine whether HLA-B genes themselves or genes in the vicinity of the B locus were involved in the pathogenesis of CBZ-induced SJS/ TEN, we performed fine mapping in the MHC region using 220 SNPs, 20 short tandem repeat polymorphisms, and analyzed HLA alleles [15]. We further found that polymorphisms located between HLA-DRA and HLA-C showed strong association. In particular, the TT or GT genotypes of rs3130690, 36 kb telemetric to the HLAB locus, were present in 98% CBZ-SJS/TEN patients, but in 5% of tolerant controls. MICA*019, 47 kb centromeric to the HLA-B locus, was present in 95% of CBZ-SJS/TEN patients, but in 15% of tolerant controls. The recombinant map of Cw *0801-HLABC-CA*119-rs3130690T-B*1502MICA*019 defined the susceptible region within 86 kb (i.e., between the T allele of rs3130690 and MICA*019) flanking the B*1502 gene in the 4Mb MHC region [15]. Within this 86-kb region, HLA-B is the only known gene. Taken together, the data suggested that one or more alleles in the vicinity of the HLA-B locus, particularly B*1502 itself, participate in the pathogenesis of CBZ-induced SJS/TEN. Genome-Wide Study Notably, although all CBZ-SJS/TEN or allopurinol-SCAR patients carried at least one HLAB*1502/5801 allele, some tolerant patients also carried the allele. This suggests that HLAB*1502/5801 is necessary but not sufficient for causing drug hypersensitivity. There could be other cofactors, such as virus infection or other allelic variants of genes, e.g., T-cell receptor genes, genes related to apoptosis, or genes for costimulatory molecules involved in the interaction between antigen-presenting cells and T cells. We have performed a whole genome scan

109

Fig. 2. Genetic roots of patients with CBZ-induced SJS/TEN who are HLAB*1502-positive. The genetic association of HLA-B*1502 and CBZ-SJS/ TEN has been found in more than 100 Southeast Asian patients or their descendants, despite living in widely scattered geographic areas (Taiwan, Hong Kong, China, USA, Europe and Australia) [9, 15, 20]. Reunion Island and US patients are Chinese descendants.

using an Affymetrix GeneChip Human Mapping 100K set in a case-control association study to identify additional markers/susceptibility genes other than HLA-B molecules that might predispose individuals for these severe adverse events. Using 56 cases/54 controls for CBZ-SJS/ TEN and 60 cases/89 controls for allopurinolSCAR, we confirmed the previous observation that the most significant association was observed in the SNPs of the HLA-B regions on chromosome 6 [unpubl. data]. We did not find additional SNPs that were significantly associated with CBZ-SJS/TEN, however, we found clustering of SNPs located between cysteine and tyrosine-rich 1 (CYYR1) gene and a disintegrin and metalloprotease with thrombospondin motifs-1 (ADAMTS1) gene on chromosome 21, which were strongly associated with allopurinolSCAR, albeit the p value is less than those seen in the HLA-B*5801 [18].

Genetic Association of Drug Hypersensitivity Is Drug-Specific

Both CBZ-SJS/TEN and allopurinol-SCAR are associated with a HLA-B allele. However, the involved B alleles are different and the two alleles

110

United States China Thailand

Taiwan Hong Kong Vietnam Cambodia

Reunion Island

have significant differences in the DNA coding sequences and in the peptide sequences that bind to the alleles. Hypersensitivity to abcabavir, a drug for AIDS, was also reported to be associated with a HLA-B allele, but again with a different one (B*5701) (table 2) [19]. This suggests that the genetic association of drug hypersensitivity is drug-specific. On the other hand, it also suggests the important role of HLA-B alleles in mediating drug-induced hypersensitivity.

Genetic Association of Drug Hypersensitivity Is Ethnicity-Specific

The tight association of HLA-B*1502 and CBZSJS/TEN and HLA-B*5801 and allopurinolSCAR was further confirmed in our extended study group of many newly enrolled patients (
40) from widely separated geographic areas (Taiwan, Hong Kong, China, and USA) (fig. 2); however, all patients enrolled were Chinese or Chinese descendants. A recent European study of 12 patients with CBZ-SJS/TEN revealed the presence of HLA-B*1502 in 4 patients who were originally from China, Cambodia, Vietnam and Reunion Island, all of whom had a parent of Asian origin (fig. 2) [20]. The remaining 8 who were

Hung ⴢ Chung ⴢ Chen

Table 2. Genetic susceptibility to drug hypersensitivity is drug-, phenotype-, and ethnicity-specific Culprit drug (population)

SJS/TEN

HSS

MPE

Carbamazepine (Han Chinese in Taiwan)

HLA-B*1502 (Pc 3.1 ! 10–27 to 1.6 ! 10–41; OR 1,357–2,504)

rs2894342 of motilin gene (Pc 0.0064; OR 7.1)

HLA-A*3101 9, 15 (Pc 2.2 ! 10–3, OR 17.5)

Carbamazepine (Caucasian in Europe)

weak association with B44 no association with HLA-B*1502

TNF-308-DR3-DQ2 haplotypes (Pc 0.02; OR 3.2)

unknown

20, 28, 29

Carbamazepine (Southeast Asians)

HLA-B*1502

unknown

unknown

20

unknown

12

Allopurinol (Han Chinese in Taiwan)

HLA-B*5801 (Pc 4.7 ! 10-24, OR 580.3)

Allopurinol (Han Chinese in Singapore)

HLA-B17/BW58 (Pc 2.9 ! 10–9, RR 46.3)

Ref.

30

Abacavir (Caucasian)

unknown

HLA-B*5701 (Pc 5.2 ! 10-20; OR 960)

19

Abacavir (Hispanics, or African)

unknown

no association with HLA-B*5701

21

SJS = Stevens-Johnson syndrome; TEN = toxic epidermal necrolysis; HSS = hypersensitivity syndrome; MPE = maculopapular eruption; Pc = corrected p value for Bonferroni’s adjustment; OR = odds ratio; RR = relative risk.

negative for the HLA-B*1502 allele are all Caucasians (table 2) [20]. Furthermore, 2 patients, 1 from Australia and 1 from the UK, who were positive for the HLA-B*1502, were of Thai origin [M. Pirmohamed, S. Mallal, pers. communications] (fig. 2). Similar ethnic difference in the genetic association has also been observed in the drug hypersensitivity to abacavir in which HLAB*5701 is strongly associated with abacavir hypersensitivity in Caucasians but not in Hispanics or Africans (table 2) [21].

HLA-B*1502 Allele Frequency Is Positively Correlated with the Prevalence of CBZ-SJS/ TEN in Different Populations

An allele associated with a particular phenotype or disease may be present in different frequencies in different ethnic groups. Therefore, pharmacogenetic studies are more likely to yield a positive

Genetics of Drug Hypersensitivity in Han Chinese

result when conducted in a population with a high frequency of such an allele [22, 23]. It is known that B*1502 is present at a higher allele frequency in countries of Southeast Asia than in countries of Northeast Asia, Europe, Africa, America, and Australia (table 1) [24]. The allele frequency of HLA-B*1502 is
2.7–11.6% in Southeast Asians, but only 0–0.1% in Caucasians. The low frequency of B*1502 in Caucasians may explain the apparent lower incidence of CBZ-SJS in Caucasians. The high frequency of B*1502 in many Southeast Asian countries, including Taiwan, Hong Kong, Singapore, Thailand, Malaysia and India, explains the high prevalence of CBZSJS/TEN in those regions. In contrast, HLA-B*5801 allele is more evenly distributed among different populations (2–4% in Africa, 1–6% in Caucasian, 3–15% in Asian Indian, and 8.8–10.9% Chinese) (table 1) [24], suggesting that the association of HLA-B*5801 and allopurinol-SCAR may also exist in other ethnic

111

groups. Indeed, the positive association was found in patients from the southern part of Japan as well in Caucasian patients (reported at the 7th International Congress on Cutaneous Adverse Drug Reactions, Paris, France, September 2006 [25, 26]).

Genetic Association of CBZ-Induced Drug Hypersensitivity Reactions Is Phenotype-Specific

We also investigated whether HLA-B*1502 is associated with other cutaneous adverse reactions (cADRs) caused by CBZ, such as MPE, lichenoid eruption, exfoliative dermatitis, or HSS. However, none of them showed a significant association with HLA-B*1502, suggesting that the genetic association between HLA-B*1502 and CBZ-SJS/ TEN is specific to SJS and TEN (table 2) [15]. Our recent study showed that MPE was associated with SNPs in the HLA-E region and a nearby allele, HLA-A*3101 (Pc = 2.2 ! 10 –3, OR = 17.5 [95% CI, 4.6–66.5]); HSS with SNPs in the motilin gene (Pc = 0.0064, OR = 7.11 [95% CI, 3.1–16.5]) located terminal to the MHC class II genes (table 2) [15]. These data suggest that genetic susceptibility to CBZ-induced cADRs is phenotype-specific.

Preventing Drug Hypersensitivity with a Gene Test: HLA-B*1502 as a Test to Identify Individuals at Risk for CBZ-Induced SJS/TEN

Although the incidence of SJS/TEN is relatively low compared to common diseases, SJS/TEN is serious and can be deadly. Many surviving patients have long-term complications, some of which result in permanent damage (e.g., ocular complications). Using a CBZ-tolerant group as the control in a test for SJS/TEN, the HLA-B*1502 allele should have 100% sensitivity and 97% specificity. Assuming a 0.25% prevalence rate, the presence of B*1502 has a 7.7% positive predictive

112

value for detecting CBZ-SJS/TEN, whereas its absence has a 100% negative predictive value. The ratio of the odds in test-positive patients to the odds in test-negative patients of having CBZ-SJS/ TEN is 13,200, which far exceeds those of the classical example of B27 and ankylosing spondylitis (OR = 100–200) [27]. Although the test for HLA-B*1502 is highly informative, 3% of patients who are test-positive may never develop the disease and thus may unnecessarily be denied the drug. Given the serious and life-threatening consequences of developing SJS/TEN and the availability of alternative drugs, withholding CBZ from these 3% of patients could be justified. These patients may be treated with other antiepileptic medications, such as phenobarbital, lamotrigine, and valproic acid. Taken together, the pre-prescription use of HLA-B*1502 as a test to screen out individuals at risk for SJS/TEN will be valuable in preventing drug toxicity caused by CBZ in Southeast Asians.

Conclusion

The pathogenesis of drug hypersensitivity reactions is believed to be immune-mediated. Prior in vitro studies suggest that the drug presentation is MHC class I restricted and there is a clonal expansion of CD8+ CTLs and these cells induce effector cytotoxic responses [16, 17]. This concept is now supported by our findings of a strong genetic association between HLA-B alleles and hypersensitivity reactions to certain specific drugs [9, 12, 15]. This strong association suggests a direct functional involvement of HLA-B molecules in the pathogenesis of the disease. A specific HLA-B molecule may serve to present the drug or its metabolite for HLA-restricted T-cell activation. Our data also indicate that the genetic predisposition of drug hypersensitivity reactions is drug-, phenotype- and ethnicity-specific. Although our studies were primarily done in Han

Hung ⴢ Chung ⴢ Chen

Chinese, the significance and its implication go beyond the Chinese. The high sensitivity and specificity of some of these markers in certain ethnic populations provide a plausible basis for the development of such a test to identify individuals at risk for these potentially life-threatening conditions. Application of HLA-B*1502 genotyping as a screening tool before prescribing CBZ should be a valuable tool to prevent CBZSJS/TEN in many Southeast Asian countries.

Acknowledgments This work has been supported by grants from the National Science & Technology Program for Genomic Medicine, National Science Council, Taiwan (National Clinical Core and National Genotyping Core), the Genomics and Proteomics Program, Academia Sinica, and the Foundation for Biomedical Sciences.

References 1 Pirmohamed M, Naisbitt DJ, Gordon F, Park BK: The danger hypothesis – potential role in idiosyncratic drug reactions. Toxicology 2002;181–182:55–63. 2 Fischer PR, Shigeoka AO: Familial occurrence of Stevens-Johnson syndrome. Am J Dis Child 1983;137:914–916. 3 Johnson-Reagan L, Bahna SL: Severe drug rashes in three siblings simultaneously. Allergy 2003;58:445–447. 4 Pritchett JH, Austin AC: Stevens-Johnson syndrome occurring in identical twins with apparent response to terramycin and aureomycin. J Med Assoc Ga 1951;40:374–376. 5 Edwards SG, Hubbard V, Aylett S, Wren D: Concordance of primary generalised epilepsy and carbamazepine hypersensitivity in monozygotic twins. Postgrad Med J 1999;75:680–681. 6 Roujeau JC, Stern RS: Severe adverse cutaneous reactions to drugs. N Engl J Med 1994;331:1272–1285. 7 Rzany B, Mockenhaupt M, Baur S, Schroder W, Stocker U, Mueller J, Hollander N, Bruppacher R, Schopf E: Epidemiology of erythema exsudativum multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis in Germany (1990–1992): structure and results of a population-based registry. J Clin Epidemiol 1996;49:769– 773. 8 Chan HL, Stern RS, Arndt KA, Langlois J, Jick SS, Jick H, Walker AM: The incidence of erythema multiforme, Stevens-Johnson syndrome, and toxic epidermal necrolysis. A population-based study with particular reference to reactions caused by drugs among outpatients. Arch Dermatol 1990;126:43–47.

9 Chung WH, Hung SI, Hong HS, Hsih MS, Yang LC, Ho HC, Wu JY, Chen YT: Medical genetics: a marker for StevensJohnson syndrome. Nature 2004;428: 486. 10 Roujeau JC, Kelly JP, Naldi L, Rzany B, Stern RS, Anderson T, Auquier A, Bastuji-Garin S, Correia O, Locati F, Mockenhaupt M, Paoletti C, Shapiro S, Shear N, Schöpf E, Kaufman DW: Medication use and the risk of Stevens-Johnson syndrome or toxic epidermal necrolysis. N Engl J Med 1995;333:1600–1607. 11 Hung SI, Chung WH, Chen YT: HLA-B genotyping to detect carbamazepineinduced Stevens-Johnson syndrome: implications for personalizing medicine. Personalized Med 2005;2:225– 237. 12 Hung SI, Chung WH, Liou LB, Chu CC, Lin M, Huang HP, Lin YL, Lan JL, Yang LC, Hong HS, Chen MJ, Lai PC, Wu MS, Chu CY, Wang KH, Chen CH, Fann CS, Wu JY, Chen YT: HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci USA 2005;102: 4134–4139. 13 Pan WH, Fann CS, Wu JY, Hung YT, Ho MS, Tai TH, Chen YJ, Liao CJ, Yang ML, Cheng AT, Chen YT: Han Chinese cell and genome bank in Taiwan: purpose, design and ethical considerations. Hum Hered 2006;61:27–30. 14 Yang HC, Lin CH, Hsu CL, Hung SI, Wu JY, Pan WH, Chen YT, Fann CS: A comparison of major histocompatibility complex SNPs in Han Chinese residing in Taiwan and Caucasians. J Biomed Sci 2006;13:489–498.

Genetics of Drug Hypersensitivity in Han Chinese

15 Hung SI, Chung WH, Jee SH, Chen WC, Chang YT, Lee WR, Hu SL, Wu MT, Chen GS, Wong TW, Hsiao PF, Chen WH, Shih HY, Fang WH, Wei CY, Lou YH, Huang YL, Lin JJ, Chen YT: Genetic susceptibility to carbamazepineinduced cutaneous adverse drug reactions. Pharmacogenet Genomics 2006; 16:297–306. 16 Nassif A, Bensussan A, Boumsell L, Deniaud A, Moslehi H, Wolkenstein P, Bagot M, Roujeau JC: Toxic epidermal necrolysis: effector cells are drug-specific cytotoxic T cells. J Allergy Clin Immunol 2004;114:1209–1215. 17 Naisbitt DJ, Britschgi M, Wong G, Farrell J, Depta JP, Chadwick DW, Pichler WJ, Pirmohamed M, Park BK: Hypersensitivity reactions to carbamazepine: characterization of the specificity, phenotype, and cytokine profile of drugspecific T-cell clones. Mol Pharmacol 2003;63:732–741. 18 Hung SI, Chung WH, Fang WH, Ou Yang CW, Chen CH, Fann CS, Wu JY, Chen YT: Genome-wide scan for genetic markers associated with carbamazepine- and allopurinol-induced severe cutaneous adverse reactions. 2nd International Drug Hypersensitivity Meeting, Liverpool, April 18–21, 2006, abstract 53. 19 Martin AM, Nolan D, Gaudieri S, Almeida CA, Nolan R, James I, Carvalho F, Phillips E, Christiansen FT, Purcell AW, McCluskey J, Mallal S: Predisposition to abacavir hypersensitivity conferred by HLA-B*5701 and a haplotypic Hsp70-Hom variant. Proc Natl Acad Sci USA 2004;101:4180–4185.

113

20 Lonjou C, Thomas L, Borot N, Ledger N, de Toma C, LeLouet H, Graf E, Schumacher M, Hovnanian A, Mockenhaupt M, Roujeau JC; RegiSCAR Group: A marker for Stevens-Johnson syndrome...: ethnicity matters. Pharmacogenomics J 2006;6:265–268. 21 Hughes AR, Mosteller M, Bansal AT, Davies K, Haneline SA, Lai EH, Nangle K, Scott T, Spreen WR, Warren LL, Roses AD; CNA30027 Study Team; CNA30032 Study Team: Association of genetic variations in HLA-B region with hypersensitivity to abacavir in some, but not all, populations. Pharmacogenomics 2004;5:203–211. 22 Daar AS, Singer PA: Pharmacogenetics and geographical ancestry: implications for drug development and global health. Nat Rev Genet 2005;6:241–246.

23 Wilson JF, Weale ME, Smith AC, Gratrix F, Fletcher B, Thomas MG, Bradman N, Goldstein DB: Population genetic structure of variable drug response. Nat Genet 2001;29:265–269. 24 Website: Allele Frequencies (http:// www.allelefrequencies.net/). 25 Dainachi T: Common HLA allele presented in the case of Stevens-Johnson syndrome and drug-induced hypersensitivity syndrome by allopurinol. 7th International Congress on Cutaneous Adverse Drug Reactions, Paris, September 6, 2006, abstract 608. 26 Lonjou C: A European genetic study of Stevens-Johnson syndrome and toxic epidermal necrolysis: HLA-B associations with specific high-risk drugs. 7th International Congress on Cutaneous Adverse Drug Reactions, Paris, September 6, 2006, abstract 207.

27 Schlosstein L, Terasaki PI, Bluestone R, Pearson CM: High association of an HL-A antigen, W27, with ankylosing spondylitis. N Engl J Med 1973;288: 704–706. 28 Pirmohamed M, Lin K, Chadwick D, Park BK: TNF promoter region gene polymorphisms in carbamazepinehypersensitive patients. Neurology 2001;56:890–896. 29 Alfirevic A, Jorgensen AL, Williamson PR, Chadwick DW, Park BK, Pirmohamed M: HLA-B locus in Caucasian patients with carbamazepine hypersensitivity. Pharmacogenomics 2006; 7: 813–818. 30 Chan SH, Tan T: HLA and allopurinol drug eruption. Dermatologica 1989; 179:32–33.

Dr. Yuan-Tsong Chen Institute of Biomedical Sciences, Academia Sinica 128 Academia Road, Section 2, Nankang, Taipei (Taiwan) Tel. +886 2 2789 9104, Fax +886 2 2782 5573 E-Mail [email protected]

114

Hung ⴢ Chung ⴢ Chen

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 115–128

Nevirapine Hypersensitivity J.M. Shenton a, b
M. Popovic a, c
J.P. Uetrecht a a Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ont., Canada; b Department of Immunotoxicology, Bristol-Myers Squibb Co., Syracuse, N.Y., USA; c Department of Investigative Toxicology, Novartis Pharma AG, Basel, Switzerland

Abstract Nevirapine (Viramune TM) can cause severe and life-threatening idiosyncratic skin rash. No clear mechanistic understanding exists; thus, it is impossible to predict which patients will suffer nevirapine-induced rash or to design a safer nevirapine analogue. Animal models of idiosyncratic drug reactions are rare, but animal models are arguably the key to mechanistic understanding. Recently, an animal model of nevirapine-induced rash was discovered and characterized; nevirapine causes rash in female Brown Norway rats with characteristics akin to nevirapine-induced rash in humans. The animal model has permitted considerable gains in the mechanistic understanding of nevirapineinduced rash, although it has yet to be determined if the animal model findings reflect the pathogenesis of nevirapine-induced rash in humans. Investigations with the animal model confirmed the essential role of the immune system. Ongoing research should determine the relative importance of parent drug versus reactive metabolite as the root cause; this is a fundamental and unanswered question in the field of idiosyncratic drug reactions. Importantly, the animal model of nevirapine-induced rash does not provide a predictive test of the ability of other drugs to cause idiosyncratic reactions, but may provide a sufficient mechanistic understanding of nevirapine-induced rash to allow the prevention of the disease in patients. Copyright © 2007 S. Karger AG, Basel

Introduction

In 1996, Boehringer-Ingelheim received Food and Drug Administration (FDA) approval for nevirapine (ViramuneTM) indicated in combination

with nucleoside analogues for the treatment of human immunodeficiency virus (HIV)-1 infections. Nevirapine was the first in its class, the non-nucleoside reverse transcriptase inhibitor (NNRTI), to reach the market. In clinical trials, some patients developed idiosyncratic skin rash and/or liver toxicity. Rashes occurred in 16% of patients and varied from mild and maculopapular to severe and life-threatening – prompting permanent discontinuation of nevirapine treatment. The severe and life-threatening rashes, Stevens-Johnson syndrome (SJS) and SJS-toxic epidermal necrolysis (TEN)-transition syndrome were responsible for 0.3% of rashes in early clinical trials [1] and cases of TEN have since been reported [2]. More recent Boehringer-Ingelheim compiled data indicate that nevirapine-induced rash occurs in 8.9% of patients with 16% of these rashes being severe or life-threatening [3]. Liver toxicity is reported both in combination with and in the absence of rash. The reported incidence of liver toxicity is much lower than for rash; Boehringer-Ingelheim reports the incidence to be 2.8% [3]. Thus, many patients are affected by nevirapine-induced rash in the absence of liver toxicity. Several risk factors for developing nevirapine-induced rash and/or hepatotoxicity have been identified (table 1). These reactions can have dire consequences for those affected. Severe rashes such as TEN are

Table 1. Risk factors identified for the development of nevirapine-induced rash and/or liver toxicity Risk factor

Details

Ref.

Sex

Severe rash and/or liver toxicity is/are more common in females with body mass index <18.5

23, 34

Antiallergic drugs

Increased incidence of rash with prophylactic use of glucocorticoids or antihistamines

24

Third trimester in pregnant women

Indications that nevirapine-induced hepatotoxicity may have a higher incidence in women who are given the drug late in pregnancy

25

Previous nevirapineinduced rash

Earlier onset and an increased risk of severe rash on rechallenge with nevirapine if rash occurred on first exposure

26

Nevirapine plasma concentrations

Findings with regard to the role of nevirapine plasma concentrations in patients that develop nevirapine reactions are contradictory: Higher nevirapine plasma levels predict rash and/or liver toxicity

27, 28

Nevirapine plasma levels are not predictors of rash and/or liver toxicity

29

CD4 count

Lower pretreatment CD4 counts in patients are protective: initiation of treatment is contraindicated in females with CD4 counts >250 cells/mm3 and males with CD4 counts >400 cells/mm3

3

HLA genes

Combinations of hepatitis, fever, or rash are associated with a HLA-DRB1*0101 haplotype; no associations were detected for isolated rash

30

Liver enzymes

Increased liver enzymes may warrant permanent discontinuation of nevirapine and should prompt investigation into alternative treatment options depending on the degree of increase

31

not understood and supportive measures are usually the only available treatment. Patients can die or are left with permanent sequelae such as visual impairment or blindness [4]. Nevirapineinduced liver toxicity has also caused deaths or required life-changing interventions including liver transplantation [5]. Rash and liver toxicity also occur in HIV-negative patients, and nevirapine for post-exposure prophylaxis is discouraged [6]. Idiosyncratic reactions restrict the prescription of efficacious drugs to patients by clinicians, villainize the government regulators responsible for the drug approval, and frustrate the pharmaceutical manufacturers who have invested hundreds of millions of dollars in drug development.

116

Despite significant research efforts, the mechanism(s) of idiosyncratic drug reactions remain(s) a ‘black box’. The ultimate goal is to be able to treat, predict, and/or prevent these reactions and other idiosyncratic reactions. To reach this goal, an understanding of the mechanisms of idiosyncratic drug reactions is imperative. Reactions such as nevirapine-induced rash have characteristics incriminating the immune system as a key mechanistic player. Thus, understanding the mechanism of nevirapine-induced rash will not only help in the handling of nevirapine-induced rashes, but may also significantly contribute to an understanding of how the immune system ‘recognizes drugs’ [discussed in the chapter by Park et al., pp 55–65]. Elucidation of the mecha-

Shenton ⴢ Popovic ⴢ Uetrecht

Fig. 1. Basic characteristics of the nevirapine animal model. Female Brown Norway rats administered nevirapine (150 mg/kg/day) in the diet develop red ears and skin lesions between days 7–10 and 10–21, respectively. When nevirapine treatment is stopped the rats recuperate and the skin lesions resolve. If the animals are then rechallenged with nevirapine, red ears develop within

approximately 8 h and skin lesions (albeit fewer than on first exposure) within 4–9 days. The lower number of skin lesions is misleading because the animals actually present with more severe clinical signs than on first exposure such as weight loss, hunched posture, and chromorhinorrhea.

nism of nevirapine-induced rash may also help in the comprehension of other idiosyncratic drug reactions. Important contributions to the understanding of mechanisms are facilitated via animal models. Nevirapine-induced rash is one of the few reactions for which a reasonable animal model is available [7]. This chapter describes the discovery and characterization of the nevirapine animal model, the most important contributions the model has made toward understanding the mechanism of nevirapine-induced rash, and the importance of the model in ongoing mechanistic investigations.

George Lai, in the laboratory of Jack Uetrecht at the University of Toronto. George was administering nevirapine (150 mg/kg/day) to female Sprague-Dawley rats in order to investigate in vivo covalent binding of nevirapine to liver proteins. During this investigation he noted that 2 of 4 rats developed erythema, exhibited excessive scratching around the nose/mouth area, and lost body weight after several weeks of treatment. This study was replicated and expanded to result in a characterization of the response of various sexes and strains of rats (and a mouse strain) to nevirapine [8]. Female Brown Norway rats developed nevirapine-induced rash following a consistent time course (fig. 1) [8] and with a high incidence (100%). These characteristics of the reaction were advantageous in that they facilitated mechanistic studies [7]. Interestingly, none of the rats tested developed liver toxicity; thus, the Brown Norway model represents the portion of patients developing rash in the absence of liver

Discovery and Characterization of an Animal Model of Nevirapine-Induced Skin Rash

The discovery of the animal model of nevirapineinduced rash ultimately stemmed from an observation in the late 1990s by a graduate student,

Nevirapine-Induced Skin Rash in Rats

117

Table 2. Similar characteristics of nevirapine-induced rash in humans and female Brown Norway rats Characteristic

Humans

Rats

Rashes

Vary from mild erythematous maculopapular rashes to blistering skin eruptions [32]

Vary from mild to severe, no blistering skin reactions observed [8]

Nevirapine plasma levels

1–10 g/ml [3]

20–40 g/ml [7]

Time to onset

Highest risk within first 6 weeks of treatment; rash generally develops 1–3 weeks after initiation of nevirapine treatment [1]

Skin lesions occur after 2–3 weeks of nevirapine dosing [8]

Dose response

Incidence increases with dose [1]

Incidence increases with dose [8]

Female sex

Increased susceptibility [23]

Increased susceptibility (likely due to metabolic differences) [7]

Escalating dose regimen

200 mg/day for 2 weeks prior to starting full dose – 400 mg/day – decreases incidence of rash by 50% [1]

40 or 75 mg/kg/day for 2 weeks prior to starting full dose – 150 mg/kg/day – prevents rash [7]

Rechallenge

Immediate onset and increased severity in some patients (especially after a severe rash) [26]

Decreased time to onset and increased severity [8]

Skin histology

Little data – mild perivascular lymphocytic inflammatory infiltrate [33]

Predominantly lymphocytic infiltrate including CD4+ T and CD8+ T cells, but also macrophages [8]

T-cell dependence

Reactions may be dependent on CD4+ T cells – rash with other constitutional findings is more commonly seen in HIV-negative and HIV-positive patients with higher CD4+ T-cell counts [6]

Reaction is dependent on T cells; CD8+ T cells do not appear to mediate the reaction, evidence points to CD4+ T cells [7]

toxicity. Nevirapine-induced rash in rats shares many characteristics with nevirapine-induced rash in humans (table 2), suggesting a common mechanistic pathway. This is the most important attribute of an animal model; indeed, the nevirapine-induced rash animal model meets many of the requirements of a viable animal model of an idiosyncratic drug reaction [for review, see 7]. The nevirapine-induced rash model is arguably the most practical animal model of an idiosyncratic drug reaction currently available for mechanistic studies.

118

Contributions of the Animal Model toward an Understanding of the Mechanism of Nevirapine-Induced Rash

CD4+ T Cells Play a Key Role in the Development of Nevirapine-Induced Rash The immune system is believed to play an important role in the development of many idiosyncratic drug reactions including rash. Many of the characteristics of nevirapine-induced rash in humans and rats implicate the immune system in the mechanism of this reaction. For instance, T cells and macrophages were present in the skin of rats with rash, and rats developed more severe

Shenton ⴢ Popovic ⴢ Uetrecht

Fig. 2. Splenocytes, total T cells, or CD4+ T cells, but not CD8+ T cells, adoptively transfer susceptibility to nevirapine-induced rash. Total T, CD4+ T, or CD8+ T cells were purified from splenocytes obtained on day 9 of second exposure and injected intravenously into naive recipients. The naive rats responded to nevirapine similarly to previously exposed (sensitized) rats after receiving total splenocytes, total T, or CD4+ T cells, and like previously naive rats after receiving CD8+ T cells.

(systemic) symptoms and with a faster onset on second exposure (rechallenge) to nevirapine (fig. 1). Furthermore, the spleens of the rats rechallenged with nevirapine were significantly larger than those of control animals, and sensitivity to the rash could be adoptively transferred with total splenocytes from these rats to naive rats, that is, the naive rats responded to nevirapine similarly to previously exposed (sensitized) rats (fig. 2) [8]. For further studies it would have been useful to elucidate a biomarker of the reac-

Nevirapine-Induced Skin Rash in Rats

tion. Although serum total immunoglobulin E (IgE) concentration appeared promising because it was significantly increased on day 7 of primary nevirapine exposure, no increase in IgE was detectable on rechallenge [9]. Thus, a subjective endpoint was decided upon for further mechanistic studies. Skin rash was not used as the endpoint, because on rechallenge, the rats developed far fewer skin lesions than on primary exposure despite more severe systemic symptoms and a more intense inflammatory infiltrate in the skin (tail, footpad, ear, and torso (ventral and dorsal) skin were all affected). However, female Brown Norway rats developed red ears prior to the onset of skin lesions on both primary and secondary exposure to nevirapine – day 7 of primary exposure and within approximately 8 h after rechallenge (fig. 1). Thus, ‘red ears’ was chosen as albeit a crude endpoint. To determine which splenocyte population was responsible for transferring susceptibility to nevirapine-induced rash, splenic total, CD4+, or CD8+ T cells were purified from rats with severe nevirapine-induced rash (second exposure) and subsequently adoptively transferred into naive recipients [9]. It was evident that total T cells, or purified CD4+ T cells in some cases, could transfer susceptibility (naive rats developed red ears in approximately 8 h), whereas CD8+ T cells could not (fig. 2). Rats partially depleted of CD4+ T cells, or depleted of CD8+ T cells, via intravenous injection with depleting monoclonal antibodies prior to nevirapine treatment, provided corroborating evidence. The development of nevirapineinduced rash was delayed in rats partially depleted of CD4+ T cells, but depletion of CD8+ T cells appeared to have no effect or actually worsen the reaction [9]. The observations that depletion of CD8+ T cells appeared to worsen the reaction, that partial depletion of CD4+ T cells did not completely prevent nevirapine-induced rash, and that CD4+ T cells could not consistently transfer sensitivity to nevirapine-induced rash to naive recipients led us to hypothesize that regulatory T

119

cells (CD4+ and/or CD8+) are also involved in the mechanism of nevirapine-induced rash. To this end, the Uetrecht laboratory has preliminary data demonstrating that CD8+ T cells in rats exposed to nevirapine upregulate FoxP3, a transcription factor expressed by regulatory T cells. An important future experiment would be to illustrate whether depletion of regulatory T cells from a purified population of CD4+ T cells prior to adoptive transfer would generate a consistent response, mimicking prior sensitization to nevirapine, in the naive recipient rats. Importantly, CD4+ T cells have also been implicated in the mechanism of nevirapine-induced skin rash in humans. In fact, Boehringer-Ingelheim has added an additional warning to the product label for nevirapine that states that nevirapine should not be initiated in women with CD4 cell counts 1250 cells/mm3 and men with CD4 counts 1 400 cells/mm3 unless the benefit outweighs the risk [3]. Nevirapine-specific T cells have also been identified ex vivo in an HIV-positive patient with nevirapine-induced hepatitis [10]. The evidence that T cells are involved in human nevirapine reactions lends further credence to the validity of the animal model and the animal model allows one to delve more deeply into the role CD4+ T cells play in the reaction than would be possible in human studies. Inflammatory Infiltrates and Changes in the Expression of Cell Surface Molecules in the Skin over the Course of Nevirapine-Induced Rash Microscopic and immunohistochemical evaluation of skin from rats with rash on first exposure (measured after 21 days) and second exposure (measured after 9 days) illustrated significant inflammatory cell infiltrates in the skin. On second exposure the infiltrate was denser than on primary exposure [8]. T cells and macrophages were the predominant cell types, and cells were observed in the dermis, epidermis, and at the dermal-epidermal junction. Similar expression patterns of  T-cell receptor, CD4, and CD8 were

120

observed, and thus it is believed that the inflammatory infiltrate contained both CD4+ and CD8+ T cells. Lesions of satellite-cell necrosis, i.e., apoptotic cells (likely keratinocytes) adjacent to mononuclear cells, were observed in the epidermis. This lesion may represent immune-mediated keratinocyte cell death [8]. The infiltrates were visible in all the skin samples analyzed, i.e., ear, tail, footpad, thorax (dorsal and ventral) skin; however, ear skin provided the clearest picture. A time course investigation using immunohistochemical evaluation looked at histological changes in ear skin in the days preceding rash [11]. Skin sections were evaluated on days 7, 14, or 21 of first exposure to nevirapine and prior to second exposure or 1 or 9 days after second exposure. On first exposure, CD4+ and CD8+ mononuclear cells were not visible in the skin on day 7 despite the characteristic ear redness that occurs on day 7 of nevirapine exposure. By day 14, skin lesions were erupting and dermal CD4+ and CD8+ cells were visible. On second exposure, a similar picture was observed, that is, CD4+ or CD8+ mononuclear infiltrates did not occur concurrently with red ears (onset occurs within
8 h of rechallenge), but were present by day 9 as reported previously [8]. However, macrophages were present in the dermis at the onset of the ear redness illustrating that macrophages enter the skin before T cells and suggesting an important role for macrophages in the early stages of the immune response; indeed, the dermal macrophage infiltrate grew denser over time. Macrophages were mostly observed throughout the dermis, around the hair follicles, and surrounding sebaceous and apocrine glands. Some macrophages were also detected in the skin muscle layer. Questions yet to be answered include what causes the macrophage infiltration into the skin and if and how the infiltrating macrophages contribute to the pathology. Other early changes, observed in the skin beginning early after the initiation of nevirapine

Shenton ⴢ Popovic ⴢ Uetrecht

Fig. 3. Potential pathway leading to nevirapine-induced rash. (1) Nevirapine is metabolized in the liver by cytochrome P450 enzymes to several hydroxylated metabolites including 12-hydroxynevirapine (2). (3) 12-Hydroxynevirapine would likely circulate through the blood to the skin. In the skin, 12-hydroxynevirapine may be sulfated (4) in skin-resident cells such as fibroblasts [21]. The sulfate is a good leaving group and may dissociate from the molecule leaving a reactive quinone methide (5). The reactive quinone methide may covalently bind to nearby cellular proteins (6). The covalent binding may disrupt cellular homeostasis and result in the production of danger signals [22] , such as release of cytokines, into the surrounding milieu. This may result in the

Nevirapine-Induced Skin Rash in Rats

recruitment of macrophages and the upregulation of MHC-I, MHC-II, and ICAM-1 on local cells such as Langerhans cells (skin-resident antigen-presenting cells) resulting in activation and mobilization of Langerhans cells (7). Nevirapine-specific CD4+ T cells may be activated in the draining lymph nodes by Langerhans cells presenting quinone methide-modified peptides and then home to the skin (8). Adhesion molecules such as ICAM-1 facilitate infiltration of T cells into the skin and antigenpresenting cell-T cell interactions (9). Nevirapine-specific T cells may result in direct attack on quinone-methidemodified skin cells and/or initiate downstream events that result in destruction of skin cells with the final result being an outbreak of nevirapine-induced rash (10).

121

dosing, included increased expression of intracellular adhesion molecule (ICAM)-1 (CD54) and major histocompatibility complex (MHC)-I and II (to a lesser extent). ICAM-1 is an important adhesion molecule, which binds to leukocyte function antigen (LFA)-1 and is expressed by endothelial cells, skin-resident antigen-presenting cells (Langerhans cells), and by activated fibroblasts and macrophages [12]. ICAM-1 expression by endothelial cells supports macrophage and/or T-cell infiltration into the skin, and ICAM-1 expression by antigen-presenting cells facilitates antigen-presenting cell–T-cell interactions. Elevated ICAM-1 expression was mostly observed in the dermis around blood vessels and hair follicles and at the dermal-epidermal junction. Increased MHC expression primarily occurred around dermal blood vessels and hair follicles and by day 14 of the treatment on the infiltrating leukocytes. Because changes that occurred at early time points were primarily evident in the dermis, it may be that the inciting nevirapine antigen is formed in the dermis. Fibroblasts are the most prominent cell type of the dermis and may play a key role in the development of rash, but this remains to be demonstrated. Despite such studies, the sequence of events from nevirapine intake to nevirapine-induced rash remains hypothetical (fig. 3). The relevance of these findings to nevirapine-induced rash in humans cannot be determined because there are no clinical studies aimed at understanding the histological or immunohistochemical appearance of nevirapine-induced rashes. To our knowledge, the only data available in the literature stems from 3 patients in whom a mononuclear infiltrate was described (as reported in table 2). However, one can compare the histological picture of nevirapine-induced rash in rats to rash induced by other causes in humans. The emerging histological picture for nevirapine-induced rash in rats is most akin to the histologic changes described for maculopapular rashes in humans [for review, see 13].

122

Prevention or Treatment of Nevirapine-Induced Rash When a patient is starting nevirapine treatment, dose escalation is recommended. It is clear that 200 mg/day of nevirapine for 2 weeks prior to starting full-dose nevirapine, 400 mg/day, significantly decreases the incidence of rash. Similar dose escalation in the nevirapine animal model completely prevents the rash (table 1). Both corticosteroids and antihistamines have been studied in clinical trials to determine their ability to prevent nevirapine-induced rash. Prophylactic corticosteroids are not recommended because patients suffered increased incidence and severity of rash when treated with prednisone for the first 2 weeks of nevirapine treatment [14]. Prophylactic antihistamine (cetirizine) use was also unsuccessful in mitigating nevirapineinduced rash [15]. Accordingly, in the nevirapine animal model, a prophylactic regimen containing an antihistamine (astemizole), a mast cell stabilizer (cromolyn), and a serotonin antagonist (ketanserin) was unable to prevent nevirapineinduced rash or even red ears [9]. The T-cell immunosuppressant, cyclosporine, has been used clinically to treat severe rashes such as SJS and TEN [discussed in 16]. Because in the nevirapine animal model nevirapine-induced rash was shown to be T-cell-dependent, studies were initiated to determine the effects of cyclosporine and tacrolimus (a similar immunosuppressant) on nevirapine-induced rash [9]. Prophylactic treatment with tacrolimus or cyclosporine on first exposure successfully prevented skin rash, and the rats remained tolerant even after tacrolimus or cyclosporine treatment was stopped (fig. 4a). Complete tolerance only remained with continued nevirapine treatment. Only partial tolerance remained if nevirapine treatment was stopped and restarted (fig. 4b). Unlike on first nevirapine exposure, tacrolimus prophylaxis was unsuccessful at preventing nevirapine-induced rash on second exposure (rechallenge; fig. 4c). Tacrolimus was also able to limit rash and induce

Shenton ⴢ Popovic ⴢ Uetrecht

Fig. 4. Effect of tacrolimus prophylaxis on nevirapineinduced rash. a Prophylactic treatment with tacrolimus (or cyclosporine) can prevent nevirapine-induced rash on first exposure. Tacrolimus treatment (1 mg/kg/day) was started 2 days prior to nevirapine treatment and continued for 5 weeks concurrently with nevirapine. The animals did not develop nevirapine-induced rash even after tacrolimus treatment was discontinued. b Only par-

Nevirapine-Induced Skin Rash in Rats

tial tolerance remains if nevirapine treatment is stopped and restarted. c Prophylactic treatment with tacrolimus cannot prevent nevirapine-induced rash on second exposure. Tacrolimus treatment (1 mg/kg/day) was started 2 days prior to nevirapine rechallenge in previously sensitized animals. The rats developed red ears in approximately 8 h.

123

Fig. 5. Tacrolimus treatment can reverse an ongoing rash and induce tolerance. Tacrolimus treatment (1 mg/kg/day) was initiated at the first sign of rash (red ears). Red ears resolved by day 1 of tacrolimus treatment. 50% of rats developed mild skin lesions that resolved by day 14 and 50% of rats developed no skin lesions. After tacrolimus treatment was stopped the rats remained tolerant to nevirapine.

tolerance to nevirapine if treatment was initiated at the first sign of rash (red ears; fig. 5). The most clinically relevant finding would be if tacrolimus reduced recovery time after a severe nevirapineinduced rash because, when nevirapine-induced rash is life-threatening, it is generally due to infection secondary to de-epithelization caused by SJS or TEN. However, tacrolimus did not reduce the time to recovery after a severe rash in the nevirapine animal model. The ability of a T-cell immunosuppressant to modulate rash further implicates T cells as key players in the development of nevirapine-induced rash in the animal model. Roles of Metabolic Tolerance and Immune Tolerance in Nevirapine-Induced Rash Tolerance to nevirapine could be due to either the induction of metabolism (metabolic tolerance) or the immune system (immunologic tolerance). Indeed, nevirapine induces the P450 enzymes responsible for its metabolism [17] and, in humans, this autoinduction of metabolizing enzymes is thought to be responsible for the decreased risk of nevirapine-induced rash when dose escalation is implemented at the initiation of nevirapine treatment [18]. In the nevirapine animal model, the tolerance induced by dose escalation is also metabolic tolerance. If metabolism is inhibited once

124

the full dose of nevirapine is started, then the animals develop nevirapine-induced rash. On the other hand, the tolerance induced by tacrolimus prophylaxis appears to be immune tolerance. In our hands, tacrolimus reduced the plasma concentrations of nevirapine, and that led to the question of whether the tolerance observed after tacrolimus treatment was metabolic or immunologic in nature. However, rats tolerized by tacrolimus prophylaxis remained tolerant despite inhibition of nevirapine metabolism. Furthermore, these rats remained partially tolerant to nevirapine on second exposure (fig. 6). Thus, both metabolic and immune tolerances play a role in nevirapine-induced rash in the animal model. Evidence that the Reaction Is Induced by the Parent Drug and/or a Metabolite An important unanswered mechanistic question is whether the parent drug or a metabolite is causing nevirapine-induced rash. Substantial evidence points to reactive drug metabolites as the cause of idiosyncratic drug reactions, but there is also evidence that the parent drug may be the culprit [see chapters by Park et al., pp 55–65, and Gerber and Pichler, pp 66–73]. Of course, the truth may lie somewhere in between. The Uetrecht laboratory has hypothesized that a reac-

Shenton ⴢ Popovic ⴢ Uetrecht

Fig. 6. Rats tolerized by tacrolimus treatment remain tolerant to nevirapine in the presence of high nevirapine plasma concentrations. Nevirapine metabolism was inhibited with aminobenzotriazole (ABT) after rats were tolerized by tacrolimus prophylaxis. Despite high plas-

ma concentrations of nevirapine, the rats remained tolerant. Rats also remained partially tolerant (25% not tolerant, 50% mild skin lesions, i.e. partial tolerance, and 25% fully tolerant) on second exposure.

tive quinone methide, formed by dissociation of the sulfate group from sulfated 12-hydroxynevirapine, is the cause of nevirapine-induced rash (fig. 3). The 12-hydroxy metabolite was singled out from other metabolites (primarily 2- and 3-hydroxynevirapine) because inhibition of nevirapine metabolism, with the cytochrome P450 inhibitor aminobenzotriazole, still led to the development of skin rash in the rats, while significantly reducing plasma concentrations of 2- and 3-hydroxynevirapine, but not 12-hydroxynevirapine [unpubl. data]. To investigate this hypothesis, 12-hydroxynevirapine was administered to nevirapine-naive female Brown Norway rats. In support of the hypothesis, 12-hydroxynevirapine was able to induce rash. 12-Hydroxynevirapine was also able to induce a rash that mimicked rash on second nevirapine exposure (rechallenge) in rats previously sensitized with nevirapine. Importantly, 4-chloronevirapine (position 4 methyl

group is replaced by chlorine), which cannot be metabolized to 12-hydroxynevirapine, did not induce rash in rats previously sensitized with nevirapine [unpubl. data]. On the other hand, Drummond et al. [10] recently reported nevirapine-specific T cells in an HIV-positive patient with nevirapine-induced hepatitis. The lymphocyte transformation test was positive with the addition of nevirapine and not with the addition of nevirapine metabolites. This evidence does not support the reactive metabolite theory. Additionally, in vitro experiments using auricular lymph node cells (ear skin draining lymph nodes) from rats with nevirapine-induced rash to investigate the role of the parent drug versus the metabolites have also not supported the in vivo findings with the Brown Norway rat model [unpubl. data] highlighting the importance of in vivo animal models. Studies are ongoing, using the nevirapine animal model as a platform, to conclusively demonstrate the role of parent drug

Nevirapine-Induced Skin Rash in Rats

125

and/or metabolite in the development of nevirapine-induced rash. On first thought, if nevirapine-plasma concentrations paralleled the incidence of nevirapine reactions, the conclusion might be that the parent drug is causing the reaction. However, this in itself does not implicate the parent drug because a specific metabolite might parallel higher parent drug levels. Moreover, in the nevirapine model, 12-hydroxynevirapine plasma concentrations remained significant even in the presence of a metabolic inhibitor that significantly increased the plasma concentrations of nevirapine [unpubl. data]. In fact, in rats, a clear relationship exists between nevirapine plasma concentrations and skin rash. In initial characterization studies, female Sprague-Dawley rats developed rash at a much lower incidence (
20%) than female Brown Norway rats (100%) and male Brown Norway and female Lewis rats did not develop rash [8]. The most susceptible rats, the female Brown Norways, also exhibited the highest nevirapine plasma concentrations. Furthermore, in later studies, when we pushed the dose or inhibited nevirapine metabolism, we did start to see rash at least in male Brown Norway rats and female Lewis rats. The rash in humans is clearly dose-dependent; the incidence is higher at 400 mg/day than at 200 mg/day and a low dose pretreatment decreases the incidence (table 2). In contrast, the correlation between rash and nevirapine plasma concentrations is not clear; data from different studies are conflicting (table 1). The disparity may be explained by differences in the rate of nevirapine metabolism. Plasma nevirapine concentrations are dependent on both dose and rate of metabolism and an increased rate of metabolism might lead to an increased level of reactive metabolite. Taiwo [18] highlights that despite the lack of a clear relationship between nevirapine plasma concentrations and nevirapine rash, rash is more common with 400 mg daily than 200 mg twice daily dose, and a dose of 400 mg daily is pharmacokinetically different from 200 mg twice daily.

126

This difference should be further investigated in relation to rash. A study investigating differential production of nevirapine metabolites in relation to nevirapine reactions may be warranted.

Utility of the Animal Model of NevirapineInduced Rash in Predictive Testing

Many people have asked if the female Brown Norway rat develops rash in response to other drugs known to cause skin rash. We have not tested other NNRTIs or other classes of drugs that have been associated with skin rash in this model. Certainly, female Brown Norway rats do not develop idiosyncratic reactions observed in people to clozapine (agranulocytosis, hepatoxicity, cardiotoxicty) or felbamate (aplastic anemia, hepatoxicity) [7]. The Brown Norway rat is not a panacea for the study of idiosyncratic drug reactions. The importance of the nevirapine-induced rash animal model is to facilitate studies toward an understanding of the mechanism of this reaction. Once the mechanism is understood it may be possible to prevent nevirapine-induced skin rash, for instance by developing ways to predict which patients may develop a severe reaction or by altering the structure of the drug to prevent the reaction. The nevirapine-induced rash animal model may also have a more general use as a model of maculopapular skin rash; we have illustrated the utility of the animal model for following the histological changes in the skin over time and how this data results in new theories and hypotheses for testing. Furthermore, understanding the mechanism of nevirapine-induced rash may shed light on the mechanism of other idiosyncratic drug reactions. However, there are likely many mechanisms of idiosyncratic drug reactions. Investigations with another animal model, penicillamine autoimmunity in the Brown Norway rat, have illustrated this point. For instance, dose escalation in the penicillamine model induces immune tolerance [19], whereas in the ne-

Shenton ⴢ Popovic ⴢ Uetrecht

virapine model, tolerance induced by dose escalation is metabolic, not immunologic, in nature. Also, with the exception of tacrolimus, treatments that decrease the incidence of penicillamine autoimmunity [20] have no effect on the incidence of nevirapine-induced rash [9]. These differences presumably reflect mechanistic differences in the pathogenesis of these idiosyncratic drug reactions. It is important that other useful animal models are developed as it is likely that different, albeit potentially overlapping, mechanisms are involved in the development of various idiosyncratic drug reactions.

Conclusions

Despite intense efforts by a few investigators, over the past 25 years little substantive progress has been made in understanding the mechanisms of idiosyncratic drug reactions. The dearth of animal models is largely to blame because animal models afford significant mechanistic insight. Thus, the animal model of nevirapine-induced

rash is an important contribution to the field of idiosyncratic drug reactions. In the few years since the discovery of this animal model, it has been clearly demonstrated that the immune system is involved in the development of rash. Without the animal model it is doubtful that this conclusion could have been confidently drawn. The animal model also provides a platform on which to investigate the fundamental unanswered question of whether reactive drug metabolite or parent drug elicits the idiosyncratic drug reaction. For instance, in vitro test systems can be validated or refuted by comparing in vitro results to in vivo responses. The next step is to determine if the findings in the animal model reflect the pathogenesis of nevirapine-induced rash in humans. Certainly, the characteristics of the rash in the animal model and in humans are similar, but studies in humans are an important next step in order for the animal model to gain broad acceptance as a tool for the study of the mechanism of nevirapine-induced rash and for the study of idiosyncratic drug reactions in general.

References 1 Pollard RB, Robinson P, Dransfield K: Safety profile of nevirapine, a nonnucleoside reverse transcriptase inhibitor for the treatment of human immunodeficiency virus infection. Clin Ther 1998;20:1071–1092. 2 Warren KJ, Boxwell DE, Kim NY, Drolet BA: Nevirapine-associated Stevens-Johnson syndrome. Lancet 1998; 351:567. 3 Product Insert: Viramune TM (Nevirapine) Tablets and Oral Suspension. Boehringer-Ingelheim. 4 Roujeau JC: The spectrum of StevensJohnson syndrome and toxic epidermal necrolysis: a clinical classification. J Invest Dermatol 1994; 102:28S–30S. 5 Piliero PJ, Purdy B: Nevirapine-induced hepatitis: a case series and review of the literature. AIDS Read 2001; 11:379–382.

Nevirapine-Induced Skin Rash in Rats

6 Anonymous: Serious adverse events attributed to nevirapine regimens for postexposure prophylaxis after HIV exposures – worldwide, 1997–2000. MMWR Morb Mortal Wkly Rep 2001; 49:1153–1156. 7 Shenton JM, Chen J, Uetrecht JP: Animal models of idiosyncratic drug reactions. Chem Biol Interact 2004;150:53–70. 8 Shenton JM, Teranishi M, Abu-Asab MS, Yager JA, Uetrecht JP: Characterization of a potential animal model of an idiosyncratic drug reaction: nevirapine-induced skin rash in the rat. Chem Res Toxicol 2003;16:1078–1089. 9 Shenton JM, Popovic, M, Chen J, Masson MJ, Uetrecht JP: Evidence of an immune-mediated mechanism for an idiosyncratic nevirapine-induced reaction in the female Brown Norway rat. Chem Res Toxicol 2005;18:1799–1813.

10 Drummond NS, Vilar FJ, Naisbitt DJ, Hanson A, Woods A, Park BK, Pirmohamed M: Drug-specific T cells in an HIV-positive patient with nevirapineinduced hepatitis. Antivir Ther 2006; 11:393–395. 11 Popovic M, Caswell J, Mannargudi B, Shenton JM, Uetrecht JP: A study of the sequence of events involved in nevirapine-induced skin rash in Brown Norway rats. Chem Res Toxicol 2006;19: 1205–1214. 12 Ohh M, Takei F: New insights into the regulation of ICAM-1 gene expression. Leuk Lymphoma 1996;20:223–228. 13 Pichler W, Yawalkar N, Schmid S, Helbling A: Pathogenesis of drug-induced exanthems. Allergy 2002;57: 884–893.

127

14 Knobel H, Miro JM, Domingo P, Rivero A, Marquez M, Force L, Gonzalez A, De Miguel V, Sanz J, Boix V, Blanco JL, Locutura J: Failure of a short-term prednisone regimen to prevent nevirapine-associated rash: a double-blind placebo-controlled trial: the GESIDA 09/99 study. J Acquir Immune Defic Syndr 2001;28:14–18. 15 Launay O, Roudiere L, Boukli N, Dupont B, Prevoteau du Clary F, Patey O, David F, Lortholary O, Devidas A, Piketty C, Rey E, Urbinelli R, Allaert FA, Treluyer JM, Caumes E: Assessment of cetirizine, an antihistamine, to prevent cutaneous reactions to nevirapine therapy: results of the viramune-zyrtec double-blind, placebo-controlled trial. Clin Infect Dis 2004;38:e66–72. 16 Chave TA, Mortimer NJ, Sladden MJ, Hall AP, Hutchinson PE: Toxic epidermal necrolysis: current evidence, practical management and future directions. Br J Dermatol 2005; 153:241–253. 17 Bardsley-Elliot A, Perry CM: Nevirapine: a review of its use in the prevention and treatment of paediatric HIV infection. Paediatr Drugs 2000;2:373–407. 18 Taiwo BO: Nevirapine toxicity. Int J STD AIDS 2006;17:364–369; quiz 370. 19 Masson MJ, Uetrecht JP: Tolerance induced by low dose D-penicillamine in the Brown Norway rat model of druginduced autoimmunity is immunemediated. Chem Res Toxicol 2004;17: 82–94. 20 Sayeh E, Uetrecht JP: Factors that modify penicillamine-induced autoimmunity in Brown Norway rats: failure of the Th1/Th2 paradigm. Toxicology 2001;163:195–211.

21 Johnson GA, Baker CA, Knight KA: Minoxidil sulfotransferase, a marker of human keratinocyte differentiation. J Invest Dermatol 1992,98:730–733. 22 Seguin B, Uetrecht JP: The danger hypothesis applied to idiosyncratic drug reactions. Curr Opin Allergy Clin Immunol 2003;3:235–242. 23 Bersoff-Matcha SJ, Miller WC, Aberg JA, van Der Horst C, Hamrick HJ Jr, Powderly WG, Mundy LM: Sex differences in nevirapine rash. Clin Infect Dis 2001;32:124–129. 24 Antinori A, Baldini F, Girardi E, Cingolani A, Zaccarelli M, Di Giambenedetto S, Barracchini A, De Longis P, Murri R, Tozzi V, Ammassari A, Rizzo MG, Ippolito G, De Luca A: Female sex and the use of anti-allergic agents increase the risk of developing cutaneous rash associated with nevirapine therapy. AIDS 2001;15:1579–1581. 25 Joy S, Poi M, Hughes L, Brady MT, Koletar SL, Para MF, Fan-Havard P: Third-trimester maternal toxicity with nevirapine use in pregnancy. Obstet Gynecol 2005;106:1032–1038. 26 Gangar M, Arias G, O’Brien JG, Kemper CA: Frequency of cutaneous reactions on rechallenge with nevirapine and delavirdine. Ann Pharmacother 2000; 34:839–842. 27 De Requena DG, Jimenez-Nacher I, Soriano V: Changes in nevirapine plasma concentrations over time and its relationship with liver enzyme elevations. AIDS Res Hum Retroviruses 2005;21:555–559. 28 De Maat MM, ter Heine R, Mulder JW, Meenhorst PL, Mairuhu AT, van Gorp EC, Huitema AD, Beijnen JH: Incidence and risk factors for nevirapine-associated rash. Eur J Clin Pharmacol 2003; 59:457–462.

29 Dailly E, Billaud E, Reliquet V, Breurec S, Perre P, Leautez S, Jolliet P, Bourin M, Raffi F: No relationship between high nevirapine plasma concentration and hepatotoxicity in HIV-1-infected patients naive of antiretroviral treatment or switched from protease inhibitors. Eur J Clin Pharmacol 2004;60: 343–348. 30 Martin AM, Nolan D, James I, Cameron P, Keller J, Moore C, Phillips E, Christiansen FT, Mallal S: Predisposition to nevirapine hypersensitivity associated with HLA-DRB1*0101 and abrogated by low CD4 T-cell counts. AIDS 2005;19:97–99. 31 Wooltorton E: HIV drug nevirapine (Viramune): risk of severe hepatotoxicity. CMAJ 2004;170:1091. 32 Murphy RL: Defining the toxicity profile of nevirapine and other antiretroviral drugs. J Acquir Immune Defic Syndr 2003;34(suppl 1):S15–S20. 33 Havlir D, Cheeseman SH, McLaughlin M, Murphy R, Erice A, Spector SA, Greenough TC, Sullivan JL, Hall D, Myers M: High-dose nevirapine: safety, pharmacokinetics, and antiviral effect in patients with human immunodeficiency virus infection. J Infect Dis 1995; 171:537–545. 34 Sanne I, Mommeja-Marin H, Hinkle J, Bartlett JA, Lederman MM, Maartens G, Wakeford C, Shaw A, Quinn J, Gish RG, Rousseau F: Severe hepatotoxicity associated with nevirapine use in HIVinfected subjects. J Infect Dis 2005;191: 825–829.

Dr. Jacintha Shenton Department of Immunotoxicology, Bristol-Myers Squibb Co. 6000 Thompson Road East Syracuse, NY 13057 (USA) Tel. +1 315 431 9382, Fax +1 315 432 2295 E-Mail [email protected]

128

Shenton ⴢ Popovic ⴢ Uetrecht

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 129–139

Animal Models of Toxic Epidermal Necrolysis Hiroaki Azukizawa a
Satoshi Itami b Departments of a Dermatology and b Regenerative Dermatology, Osaka University, Graduate School of Medicine, Osaka, Japan

Abstract Toxic epidermal necrolysis (TEN) is a severe acute exfoliative skin disease characterized by detachment of large epidermal sheets as a result of massive keratinocyte apoptosis. Although TEN is mostly caused by drugs, how drug-related antigens are presented to T cells is not clarified. As for other drug hypersensitivities, it is difficult to establish an animal model of TEN using a drug antigen. Currently, TEN can be reproduced by a combination of transgenic mice expressing a model antigen in the epidermis and its specific CD8+ T-cell receptor. The animal models of TEN introduced here not only reproduce the skin symptoms and histopathological findings of TEN in humans, but also support the findings of clinical reports, e.g. epidermal injury by drug-specific cytotoxic T lymphocytes and lymphopenia. Moreover, they provide new aspects to understand the underlying pathomechanisms of TEN, e.g. possible involvements of epidermal antigens and regulatory T cells (Tregs), and link known clinical observations, e.g. the relation of AIDS and TEN. Although the therapeutic effect of Tregs for TEN has not been proven so far in this animal model, Tregs must be a key player as indicated by prophylaxis experiments, and might open the way for new therapies of TEN. Copyright © 2007 S. Karger AG, Basel

Introduction

Toxic epidermal necrolysis (TEN) is a severe adverse drug reaction characterized by detachment of large epidermal sheets as a result of massive

keratinocyte apoptosis. Due to its high mortality, the invention of new therapies for TEN is a challenge to medical research. However, satisfactory treatments for this devastating skin disease have not been established so far. Since the pathogenesis of drug hypersensitivity is still largely unknown, it is difficult to reproduce this severe skin disease in animals using drugs or chemicals. Currently, a mouse skin disease induced by epidermal antigen-specific cytotoxic T lymphocytes (CTLs) is proposed as an animal model of TEN. Although this model is not applicable as a drug hypersensitivity model in general, it provides many hints to the understanding of the pathomechanisms of TEN.

Severe Acute Graft-versus-Host Disease and TEN

TEN is a severe widespread epidermal necrosis caused by drugs such as antibiotics, anticonvulsants, etc. A similar devastating skin change is observed in the severest cases of acute graft-versus-host disease (aGVHD) after receiving bone marrow transplantation or hematopoietic stem cell transplant for the treatment of lymphoproliferative diseases [1]. Although the trigger of

Fig. 1. Effector cells and antigens in acute GVHD, drug hypersensitivity, and autoimmune disease. The effector T cell (Teff) and antigen (Ag) are essential factors for each disease, whereas effector cells of aGVHD and antigen of drug hypersensitivity are exogenous factors. The origin of these factors distinguishes these 3 diseases fundamentally.

aGVHD is different from that of drug-induced TEN, there is a close resemblance between them in clinical symptoms and histopathological findings, e.g. detachment of large epidermal sheets accompanied by a positive Nikolsky’s sign and mucosal involvement. Correia et al. [2] showed that the blister fluid from bullous skin lesions of severe aGVHD and TEN predominantly contains CD8+ T cells, indicating that both diseases are mediated by CTLs. Murine models using allogeneic cell transfer have been well studied for understanding the pathomechanisms of aGVHD. Asagoe et al. [3] demonstrated that the adoptive transfer of an extremely large amount of T-cellenriched spleen cells (5 ! 107 cells/mouse) from C57BL/6 mice expressing major histocompatibility complex (MHC) H-2b into BALB/c nude mice expressing H-2d results in a severe lethal aGVHD developing body weight loss, diarrhea, erythematous skin changes, and erosions with positive Nikolsky’s signs. Thus, the skin disease of the mouse aGVHD model resembles Stevens-Johnson syndrome or TEN observed in drug-hypersensitive patients. However, murine aGVHD models are not suitable for a drug hypersensitivity model in several aspects, since there are essential differences between aGVHD and TEN (fig. 1).

130

For example, donor T cells in aGVHD are activated by single or several host-derived alloantigens, whereas effector T cells in drug hypersensitivity can be activated by a single chemical. Moreover, drug-specific T cells arise from a T-cell repertoire of the same individual, while effector T cells of aGVHD are derived from a different individual.

Why Do Drug-Specific T Cells Attack Keratinocytes?

When we think about the pathogenesis of drug hypersensitivity, the reason why drug-specific T cells preferentially attack epidermal keratinocytes is unclear. The same question can be asked for aGVHD. One of the hypothetical mechanisms is the involvement of epidermal self-antigens during the priming of drug-specific T cells. Selfproteins in the cytoplasm, e.g. keratins in epidermal keratinocytes, are usually processed into peptides within the proteasomes [4, 5]. Self-peptides are transported to the endoplasmic reticulum, bind to MHC class-I molecules, and are presented on the cell surface. Drugs or their metabolites may bind to proteins, peptides, or MHC

Azukizawa ⴢ Itami

Fig. 2. Indirect antigen presentation of epidermal selfantigens and drugs. Keratinocytes undergo apoptosis physiologically. Apoptotic materials of keratinocytes can be phagocytosed by antigen-presenting cells (APCs) such as dendritic cells (DCs). Exogenously engulfed antigen can be cross-presented on MHC class I molecules

like an endogenous antigen. APCs migrate to the skindraining lymph node (LN) and present antigens to T cells. Drugs may bind to peptides or MHC molecules at any point of this indirect epidermal self-antigen presentation pathway.

molecules during antigen presentation. However, epidermal keratinocytes themselves cannot directly prime T cells circulating in blood, lymph, or lymphatic organs, since there are no vessels in the epidermis. Another antigen-presenting mechanism involves the indirect recognition of epidermal self-antigens by T cells. Epidermal keratinocytes undergo apoptosis physiologically, and they are phagocytosed by antigen-presenting cells (APCs) such as dendritic cells (DCs). Engulfed proteins in DCs can also enter the processing pathway for endogenous proteins as described above. When such an engulfed ‘exogenous’ antigen is presented on the MHC class-I molecule

that usually presents an endogenous antigen produced in the cell, this process is called ‘cross-presentation’ [6] (fig. 2). Drugs or their metabolites may also bind to proteins, peptides or MHC molecules in any step of this indirect antigen presentation.

Animal Model of TEN

Contact Hypersensitivity and Drug Hypersensitivity

Haptens are low molecular weight chemicals that usually penetrate the skin through the epidermis. Some of the haptens covalently bind to

131

Fig. 3. Contact hypersensitivity and drug hypersensitivity. The immune response of contact hypersensitivity can be divided into a sensitization phase and an elicitation phase. In the sensitization phase, haptens penetrate into the body through the epidermis, and may covalently bind to proteins or peptides. Haptens reach the skindraining lymph nodes by transport through antigenpresenting cells or lymphatic flow, and activate the adaptive immune system. In the elicitation phase, sensitized T cells infiltrate the epidermis and cause epidermal

injury. On the other hand, culprit drugs in drug hypersensitivity are usually absorbed from the gut by oral medications, or directly infused into the blood stream by intravenous medications. They circulate in the blood stream and penetrate tissues of the whole body. Unlike contact hypersensitivity, it is unclear how drug-specific T cells obtain the ability to migrate into the skin. Also whether drugs are actually binding to target cells like keratinocytes or not is unknown for both the sensitization and elicitation phases.

an epidermal protein or peptide, recognized by the immune system in a MHC-restricted manner, and cause contact hypersensitivity by CTLs [7, 8] (fig. 3). In contrast, culprit drugs in drug hypersensitivity are usually absorbed from the gut by oral medications, or directly infused into

the blood stream by intravenous medications. They circulate in the blood stream and penetrate all tissues of the body. Some are converted to metabolites in the liver or other organs. Although culprit drugs of drug hypersensitivity are not administered transepidermally, they might have

132

Azukizawa ⴢ Itami

a chance to haptenate proteins in peripheral tissue such as skin. Like contact hypersensitivity, the immune response of drug hypersensitivity may also divide into the two phases, sensitization and elicitation. In the sensitization phase, priming of T cells may occur in the skin-draining lymph nodes or other lymphoid organs. Unlike contact hypersensitivity, it is unclear how drug-specific T cells obtain the ability to migrate into the skin, as it is unknown for both the sensitization and elicitation phases whether or not drugs actually bind to target cells such as keratinocytes.

Transfer of Epidermal Antigen-Specific CTLs Is Not Sufficient to Induce TEN

Blanca et al. [9] showed that the expression of cutaneous lymphocyte antigen, a skin-homing receptor, is significantly increased on the peripheral blood T cells of drug-hypersensitive patients. Since drug-specific T cells must be determined to attack the epidermis after priming, a keratinocyte-specific T cell could be a model for TEN. Currently, a combination of transgenic mice is available as a model of keratinocyte-specific antigen-specific T cell. K5-mOVA is a transgenic mouse expressing chicken ovalbumin (OVA) under control of the keratin-5 promoter [10]. The OVA-specific CD8+ T cells from T-cell receptor transgenic mice (OT-I) have been used in models for organ-specific autoimmune diseases such as type-1 diabetes [11]. OT-I cells can also be used as a model of epidermal antigen-specific CTLs in K5-mOVA mice. It is interesting to know whether these two factors are sufficient to induce epidermal damage like TEN or not. Adoptively transferred epidermal antigen-specific CTLs proliferate in the skin-draining lymph node, infiltrate into the epidermis, and induce apoptosis of keratinocytes. The histopathology of the skin from these mice is clinically comparable to mild drug eruption or mild aGVHD, whereas full-

Animal Model of TEN

thickness epidermal necrosis causing widespread erosion is never observed even when an extremely large amount of CTLs is transferred. Thus, a combination of epidermal antigen-specific CTL is not sufficient to induce TEN.

Inductive Model of TEN

It is clear that an epidermal antigen-specific CD8+ T cell can be an effector cell for inducing apoptosis of epidermal keratinocytes, whereas other cells and/or factors must be necessary to reproduce a severe epidermal injury like TEN. Irradiated mice or athymic nude mice provide a lymphopenic environment in vivo with a minimum of residual T cells. When OT-I cells are transferred into sublethally (5 Gy) irradiated K5mOVA mice or athymic K5-mOVA nude mice, recipient mice develop detachment of large epidermal sheets (fig. 4). Widespread erosion with positive Nikolsky signs, mucosal involvement in the eyes, and severely wasted appearance are similar to the clinical findings in TEN patients. Histopathologically, a full thickness of epidermal necrosis, subepidermal blister formation, and relatively few lymphocyte infiltrations are identical to TEN. In addition, OVA-specific CD4+ T cells of OT-II transgenic mice never induce apoptosis of keratinocytes in K5-mOVA mice even if they are transferred into the lymphopenic hosts (unpublished data). Taken together, this inductive model provides three pieces of evidence on the pathogenesis of TEN. First, CD8+ T cells induce apoptosis of keratinocytes in TEN. This fact supports the data by Nassif et al. [12] who suggested that effector cells in human TEN are drug-specific CTLs. Second, mice with normal T-cell repertoires thus having low frequencies of antigenspecific T cells do not develop cytotoxicity, resulting in TEN even when such CTLs accidentally developed in vivo. Finally, a keratinocytederived self-antigen might be involved in CTL priming to facilitate their skin migration.

133

Fig. 4. Inductive model of TEN. TEN can be reproduced by transfer of epidermal antigen-specific CTLs into lymphopenic recipients like irradiated mice or athymic nude mice. a K5-mOVA nude mice develop detachment of large epidermal sheets with positive Nikolsky signs on

day 7 after adoptive transfer of 1 ! 107 OT-I cells (arrows). b Full thickness epidermal necrosis and relatively few lymphoid cell infiltrations in the dermis are identical to the histopathology of TEN patients.

3-Phase Model for TEN: Importance of Lymphopenia

duced tissue damage followed by production of proinflammatory cytokines like TNF-, IL-1, etc. Thus, the 3-phase model for aGVHD could be applied for drug-induced TEN as follows: (1) pre-activation of APCs by drugs or metabolites; (2) antigen recognition by drug-specific CTLs, and (3) tissue damage by CTLs. In aGVHD, the host immunity is severely destroyed by conditioning regimens like total body irradiation or chemotherapies. Such an immunosuppressive state may not completely differentiate an immunological background of TEN from that of aGVHD, since leukopenia is also found in TEN patients in the acute phase [15]. Hence, lymphopenia in phase 1 should be considered for the 3phase model of TEN.

A 3-phase model has been proposed to understand the pathophysiology of aGVHD as follows: (1) pre-activation of APCs by a conditioning regimen; (2) host alloantigen recognition by donor T cells, and (3) tissue damage by CTL [13] (table 1). Danger signals are important for maturing APCs and to facilitate T-cell priming. The significance of the danger signal has also been proposed for drug hypersensitivity as the ‘danger hypothesis’ [14]. Not only pathogens but also drugs or metabolites may directly upregulate the expression of costimulatory molecules on APCs, or may indirectly upregulate them by drug-in-

134

Azukizawa ⴢ Itami

Table 1. A 3-phase model for aGVHD and corresponding immune reactions of TEN in each phase aGVHD

TEN

Phase 1 Effect of conditioning TNF-, IL-1, and danger signal by a conditioning regimen mature host DC Tissue damage leaks LPS

Effect of drug Danger signal by drugs matures DC? Lymphopenia

Phase 2 Donor T-cell activation Th1 cytokines

Drug-specific T-cell activation Th1 cytokines

Phase 3 Cellular and inflammatory effector phase CTL Fas/FasL Perforin/granzyme LPS

Cellular and inflammatory effector phase CTL Fas/FasL Perforin/granzyme

Lymphopenia is not included in the original 3-phase model for aGVHD; however, it might be more important than danger signals in TEN.

Double Transgenic Model

In the inductive model of TEN, the OT-I cells were not selected in a thymus expressing OVA and were adoptively transferred into K5-mOVA mice as in an aGVHD model (fig. 5). However, effector cells of drug hypersensitivity arise from endogenous T-cell repertoires, indicating that adoptive transfer itself might cause an essential difference in the background of effector cells between drug hypersensitivity and aGVHD. On the other hand, drug hypersensitivity is not defined as an autoimmune disease, since an exogenously administered drug antigen triggers drug-specific T-cell immunity (fig. 1). If an epidermal model antigen and its specific T cells are transgenically expressed in the same individual lacking the thymus, the immune reaction between them can be tested without considering any influences of endogenous effector cells. K5-mOVA.OT-I double transgenic (dTg) mice generated in the euthymic mouse background never develop TEN, although

Animal Model of TEN

their isolated effector cells do induce TEN when they are adoptively transferred into lymphopenic hosts [16] (fig. 6). Strikingly, dTg mice generated in an athymic nude mouse background spontaneously develop TEN, indicating that adoptive transfer itself is not a sufficient trigger for TEN. Moreover, even danger signals are not necessary to initiate TEN in athymic mice, as TEN in the athymic mouse occurs at 2–3 months of age without any triggers. Interestingly, inoculation of euthymic dTg mouse cells prevents TEN of athymic dTg mice, indicating that thymus-derived tolerogenic cells regulate effector cells of TEN in euthymic mice (fig. 6). Notably, the inoculation of wild-type cells also prevents TEN in athymic dTg mice, indicating that thymus-derived tolerogenic T cells are not selected by a particular antigen expressed in the thymus. In general, T cells are regulated by two tolerance mechanisms. One is central tolerance by thymic clonal deletion eliminating autoreactive T cells against autoantigen expressed in the thymus [17]. Indeed, the number

135

Fig. 5. Background of effector cells in aGVHD, drug hypersensitivity, and autoimmune disease. In aGVHD or the inductive model of TEN, effector T cells (Teff) arise in different individuals without encountering a target antigen during thymic selection. On the other hand, Teff in drug hypersensitivity arise from an endogenous T-cell

repertoire without encountering drug antigen in the thymus. The euthymic dTg model resembles an autoimmune disease model. The background of Teff in the athymic dTg model has a close resemblance to that of drug hypersensitivity since these cells are not influenced by thymic clonal deletion.

of OT-I cells is dramatically decreased in euthymic dTg mice, as OVA which is also expressed in the thymus induces clonal deletion. The other mechanism is peripheral tolerance-suppressing autoreactive T cells in the periphery. The contribution of central tolerance may be small for preventing drug-induced TEN, as thymic clonal deletion only occurs for autoantigens. Taken together, the peripheral tolerance maintained by thymus-derived tolerogenic cells plays a key role in regulating effector cells of TEN in the dTg model.

Regulatory T Cells and TEN

136

Interestingly, thymus-derived CD4+CD25+ regulatory T cells (Tregs) are critical to transfer tolerance in the dTg model, since transfer of euthymic dTg mouse cell-depleted Tregs do not prevent TEN in athymic dTg mice. Tregs have increasingly been in the focus of tolerance research since the first report by Sakaguchi et al. [18] in 1995. Tregs are thymus derived CD4+ T cells constitutively expressing CD25 (IL-2 receptor  chain) and cytotoxic T-lymphocyte-associ-

Azukizawa ⴢ Itami

Fig. 6. Tregs prevent TEN mediated by CTLs. Euthymic K5-mOVA.OT-I double transgenic (dTg) mice never develop TEN spontaneously. In contrast, athymic dTg mice spontaneously develop TEN at 8–12 weeks of age. In vivo CD4+ T cell depletion of euthymic dTg mice results in

TEN. Adoptive transfer of isolated cytotoxic T cells (CTLs) from euthymic dTg mice induced TEN in athymic K5mOVA. CD4+CD25+ regulatory T cells (Tregs) prevent CTL activity in euthymic dTg mice, whereas peripheral tolerance by Tregs is lacking in athymic dTg mice.

ated antigen (CTLA4). The transcription factor, FOXP3, is a key regulator of Tregs and its mutation results in autoimmune diseases called IPEX [19]. The naturally occurring Tregs usually suppress T-cell immunity caused by CD4+CD25– effector cells and CD8+ T cells; however, little is known about the mechanism of CD8+ T-cell suppression. In euthymic dTg mice, naturally occurring Tregs continuously suppress CTLs causing TEN in the periphery, as depletion of CD4+ T cells including Tregs in vivo breaks peripheral tolerance on CTLs and results in TEN in euthymic dTg mice [16] (fig. 6). This model nicely demonstrates the link between AIDS and TEN. Patients with AIDS are at increased risk of TEN [20]. The number of CD4+ T cells is low in AIDS patients and high CD8+ cell numbers indicate a

risk of drug eruption [21]. Thus, the dTg model suggests that CD4+ T cells prevent TEN induced by CD8+ T cells.

Animal Model of TEN

Death Receptor and TEN

One of the important tasks for animal models is to test the efficacy of therapy. Viard et al. [22] showed that human intravenous immunoglobulin treatment inhibits TEN by blocking Fas (CD95). To date, many groups have applied this treatment for TEN patients, however, the outcome is controversial [23]. Recombinant monoclonal antibody (mAb) treatment is more specific and effective than intravenous immunoglobulin treatment because of the frequency of specific an-

137

tibodies. Recently one group reported the effectiveness of anti-TNF- antibody treatment for a TEN patient by using infliximab [24]. The therapeutic effect of anti-Fas-L mAb or anti-TNF-  mAb was preliminarily tested in the inductive model of TEN. On days 2, 4 and 6 after induction of TEN, 0.5 mg/mouse of either mAb was inoculated intraperitoneally. Although mice inoculated with anti-TNF- mAb had a slightly prolonged survival, a significant difference compared to the control mAb-treated mice was not observed (unpublished data). Further studies are required to conclude the efficacy of these mAbs blocking death receptor interaction in TEN.

Tregs for TEN Treatment

The therapeutic effect of Tregs on TEN is an issue to clarify in animal models, as there are few satisfactory treatments for this life-threatening skin disease. In principle, it is difficult to isolate Tregs for a therapeutic application by in vitro expansion or stimulation during the acute phase of disease, as TEN proceeds rapidly after onset. However, a key molecule or a key cytokine of Tregs suppressing CTLs might open a way for new therapy. The prophylactic or therapeutic effects of purified Tregs on TEN have been tested in the inductive model of TEN [16]. Unexpectedly, the transfer of purified Tregs never prevented TEN induced in irradiated hosts, although Tregs prevent TEN in athymic dTg mice as described above. Interestingly, co-transfer of Tregs and CD11c+ DCs, but not DCs alone, prevents TEN induced in the irradiated host. These findings suggest that Tregs cannot suppress the effector cells of TEN directly or independently of APCs but need antigen presentation by DCs or regulation occurs through DCs [25]. Little is known about the mechanism of CTL suppression by Tregs. Some mechanisms are proposed based on clinical or experimental studies as follows: (1) Tregs mediated killing by the perforin or granzyme pathway; (2) indoleamine 2,3-dioxy-

138

genase expression in APCs by the CD80 or CD86CTLA-4 signaling pathway, and (3) TGF- mediated CTL regulation by Tregs [26]. So far, a therapeutic effect of Tregs during the acute phase of TEN has not been observed in this animal model, although the application of Tregs before induction of TEN shows a completely protective effect. Since the regulatory pathway from Tregs to CTLs might be disrupted during the acute phase of TEN, further detailed studies may lead to the establishment of effective CTL regulation after onset of disease.

Concluding Remarks

In this chapter, the pathomechanisms of TEN have been discussed by comparing them with contact hypersensitivity, aGVHD, and autoimmune disease. Although the animal models of TEN introduced here are not directly applicable for a drug hypersensitivity model in general, they not only reproduce the clinical symptoms and histopathological findings of TEN in humans, but also sustain the findings of clinical reports, e.g. drug-specific CTLs and lymphopenia. Moreover, these animal models underline the importance of Tregs in the pathomechanisms of TEN. Finally, the failure to elicit TEN in immunocompetent mice but not in athymic or irradiated mice underlines the importance of immunodeficiency in this disease and may explain why patients with AIDS have such a high incidence of TEN. So far, the therapeutic effect of Tregs in TEN has not been proven in this animal model, but Tregs are key players in prophylactic experimental settings and might open the way for new therapy of TEN.

Acknowledgement We thank Dr. Hideo Yagita (Juntendo University, Japan) for collaboration by providing anti-Fas-L and anti-TNF mAb. We also thank Dr. Manfred B. Lutz (University of Erlangen, Germany) for useful comments and suggestions on this work.

Azukizawa ⴢ Itami

References 1 Vargas-Diez E, Garcia-Diez A, Marin A, Fernandez-Herrera J: Life-threatening graft-vs-host disease. Clin Dermatol 2005;23:285–300. 2 Correia O, Delgado L, Barbosa IL, Domingues JC, Azevedo R, Vaz CP, Pimentel P: CD8+ lymphocytes in the blister fluid of severe acute cutaneous graft-versus-host disease: further similarities with toxic epidermal necrolysis. Dermatology 2001;203:212–216. 3 Asagoe K, Takahashi K, Yoshino T, Kondo E, Tanaka R, Arata J, Akagi T: Numerical, morphological and phenotypic changes in Langerhans cells in the course of murine graft-versus-host disease. Br J Dermatol 2001; 145:918– 927. 4 York IA, Rock KL: Antigen processing and presentation by the class I major histocompatibility complex. Annu Rev Immunol 1996;14:369–396. 5 Yoneda K, Furukawa T, Zheng YJ, Momoi T, Izawa I, Inagaki M, Manabe M, Inagaki N: An autocrine/paracrine loop linking keratin 14 aggregates to tumor necrosis factor alpha-mediated cytotoxicity in a keratinocyte model of epidermolysis bullosa simplex. J Biol Chem 2004;279:7296–7303. 6 Shen L, Rock KL: Priming of T cells by exogenous antigen cross-presented on MHC class I molecules. Curr Opin Immunol 2006;18:85–91. 7 Weltzien HU, Moulon C, Martin S, Padovan E, Hartmann U, Kohler J: T cell immune responses to haptens. Structural models for allergic and autoimmune reactions. Toxicology 1996;107: 141–151. 8 Akiba H, Kehren J, Ducluzeau MT, Krasteva M, Horand F, Kaiserlian D, Kaneko F, Nicolas JF: Skin inflammation during contact hypersensitivity is mediated by early recruitment of CD8+ T cytotoxic 1 cells inducing keratinocyte apoptosis. J Immunol 2002;168: 3079–3087.

Dr. Hiroaki Azukizawa Department of Dermatology Osaka University, Graduate School of Medicine 2–2 Yamadaoka, Suita Osaka 565-0871 (Japan) Tel. +81 6 6879 3031, Fax +81 6 6789 3039 E-Mail [email protected]

Animal Model of TEN

9 Blanca M, Posadas S, Torres MJ, Leyva L, Mayorga C, Gonzalez L, Juarez C, Fernandez J, Santamaria LF: Expression of the skin-homing receptor in peripheral blood lymphocytes from subjects with nonimmediate cutaneous allergic drug reactions. Allergy 2000; 55:998–1004. 10 Azukizawa H, Kosaka H, Sano S, Heath WR, Takahashi I, Gao XH, Sumikawa Y, Okabe M, Yoshikawa K, Itami S: Induction of T-cell-mediated skin disease specific for antigen transgenically expressed in keratinocytes. Eur J Immunol 2003;33:1879–1888. 11 Kurts C, Heath WR, Carbone FR, Allison J, Miller JF, Kosaka H: Constitutive class I-restricted exogenous presentation of self antigens in vivo. J Exp Med 1996;184:923–930. 12 Nassif A, Bensussan A, Boumsell L, Deniaud A, Moslehi H, Wolkenstein P, Bagot M, Roujeau JC: Toxic epidermal necrolysis: effector cells are drug-specific cytotoxic T cells. J Allergy Clin Immunol 2004;114:1209–1215. 13 Reddy P, Ferrara JL: Immunobiology of acute graft-versus-host disease. Blood Rev 2003;17:187–194. 14 Pirmohamed M, Naisbitt DJ, Gordon F, Park BK: The danger hypothesi – potential role in idiosyncratic drug reactions. Toxicology 2002;181–182:55–63. 15 Roujeau JC, Moritz S, Guillaume JC, Bombal C, Revuz J, Weil B, Touraine R: Lymphopenia and abnormal balance of T-lymphocyte subpopulations in toxic epidermal necrolysis. Arch Dermatol Res 1985;277:24–27. 16 Azukizawa H, Sano S, Kosaka H, Sumikawa Y, Itami S: Prevention of toxic epidermal necrolysis by regulatory T cells. Eur J Immunol 2005;35:1722– 1730. 17 Stockinger B: T lymphocyte tolerance: from thymic deletion to peripheral control mechanisms. Adv Immunol 1999;71:229–265.

18 Sakaguchi S, Sakaguchi N, Asano M, Itoh M, Toda M: Immunologic selftolerance maintained by activated T cells expressing IL-2 receptor alphachains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases. J Immunol 1995;155:1151–1164. 19 Sakaguchi S: Naturally arising Foxp3expressing CD25+CD4+ regulatory T cells in immunological tolerance to self and non-self. Nat Immunol 2005;6: 345–352. 20 Saiag P, Caumes E, Chosidow O, Revuz J, Roujeau JC: Drug-induced toxic epidermal necrolysis (Lyell syndrome) in patients infected with the human immunodeficiency virus. J Am Acad Dermatol 1992;26:567–574. 21 Eliaszewicz M, Flahault A, Roujeau JC, Fillet AM, Challine D, Mansouri S, Wolkenstein P, Aractingi S, Penso-Assathiany D, Maslo C, Bourgault-Villada I, Chosidow O, Caumes E: Prospective evaluation of risk factors of cutaneous drug reactions to sulfonamides in patients with AIDS. J Am Acad Dermatol 2002;47:40–46. 22 Viard I, Wehrli P, Bullani R, Schneider P, Holler N, Salomon D, Hunziker T, Saurat JH, Tschopp J, French LE: Inhibition of toxic epidermal necrolysis by blockade of CD95 with human intravenous immunoglobulin. Science 1998; 282:490–493. 23 Bachot N, Revuz J, Roujeau JC: Intravenous immunoglobulin treatment for Stevens-Johnson syndrome and toxic epidermal necrolysis: a prospective noncomparative study showing no benefit on mortality or progression. Arch Dermatol 2003;139:33–36. 24 Hunger RE, Hunziker T, Buettiker U, Braathen LR, Yawalkar N: Rapid resolution of toxic epidermal necrolysis with anti-TNF-alpha treatment. J Allergy Clin Immunol 2005;116:923–924. 25 Bluestone JA, Tang Q: How do CD4+CD25+ regulatory T cells control autoimmunity? Curr Opin Immunol 2005;17:638–642. 26 Zou W: Regulatory T cells, tumour immunity and immunotherapy. Nat Rev Immunol 2006;6:295–307.

139

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 140–150

Non-Clinical Testing Approaches for Drug Development: Possibilities and Limitations Thomas T. Kawabata a
Joseph R. Piccotti b Drug Safety R&D, Pfizer Global Research and Development, a Groton, Conn. and b Ann Arbor, Mich., USA

Abstract Drug hypersensitivity reactions produced by drugs administered by the oral or parenteral routes are usually observed during later stages of development (clinical trials or commercialization phase). This presents a risk to patients and significant loss of investment for the pharmaceutical industry. Late stage attrition due to hypersensitivity reactions is attributed to the poor predictive values of non-clinical approaches currently being used. Animal models have been developed, however, due to significant limitations of each assay or inadequate validation, they are rarely used during drug development. Future development of better animal models or in vitro approaches will depend on applying information obtained from studies on the mechanism of human hypersensitivity reactions and the identification of patient and environmental susceptibility factors. Copyright © 2007 S. Karger AG, Basel

Introduction

Due to the high attrition rate of candidate compounds during drug development, the pharmaceutical industry has been trying to identify the best candidates and uncover potential toxicities as early as possible to minimize investment and risk during clinical trials. This has been very difficult to achieve for compounds that produce immune-mediated hypersensitivity reactions due to

the lack of predictive non-clinical models. For compounds that produce a relatively high incidence rate of hypersensitivity reactions (11%) in humans, standard non-clinical toxicology studies are unable to detect potential risk. Reactions at this incidence rate are observed during clinical trials and continued development would be based on the clinical indication, severity of the reactions, potential mechanism and other factors. A greater concern is with severe, low incidence hypersensitivity reactions (! 0.1%) that only appear when larger populations are exposed (during marketing). Hepatotoxicity is the most common cause for post-market regulatory actions and the mechanisms for these reactions are unclear. Immune-mediated hypersensitivity mechanisms may be responsible for reactions with some of these drugs. A few marketed compounds have had changes to ‘boxed’ warnings of labels or were withdrawn from market due to severe adverse hypersensitivity reactions of the skin (table 1). Reactions with these marketed drugs are of significant public health concerns since idiosyncratic severe hypersensitivity reactions may only become apparent during the commercialization phase of the drug. Moreover, the total investment made in the development of each com-

Table 1. FDA MedWatch safety alerts of drugs that produce skin hypersensitivity reactions from 1996 to 2006 Drug

Reaction

Action by FDA

Lamotrigine Valdecoxib

Severe life-threatening rash Potentially life-threatening skin reactions

Chlormezanone Abacavir

Low incidence serious reactions (TEN) Fatal hypersensitivity reactions

Letter/Box (1997) Box (2004); request for voluntary withdrawal (2005) Withdrawal letter (1996) Box/Letter (2000)

MedWatch alerts are given to marketed drugs (http://www.fda.gov/medwatch/safety.htm). Drugs with skin hypersensitivity warnings at the time of marketing would not be included in MedWatch. Many drugs had alerts for hepatotoxicity but it is not known which of these reactions were attributed to immune-mediated hypersensitivity reactions. Letter = Dear Healthcare Professional letter; Box = Change in Box warning on label. The action year by the FDA is indicated in parentheses.

pound is significant. Patient-related factors that determine the susceptibility to hypersensitivity reactions are unknown and therefore these reactions are unpredictable in individual patients. These concerns warrant increased research to understand the etiology of drug hypersensitivity reactions and the development of predictive animal models and in vitro approaches [1]. The goal of this review is to provide an overview on the current non-clinical approaches used to minimize the risk for drug hypersensitivity reactions, when and how hypersensitivity testing methods could be used in drug development, and potential animal models that could be developed in the future. This review will focus on low molecular weight (MW) compounds that produce hypersensitivity reactions when administered systemically. Hypersensitivity testing approaches for drugs administered topically or by the inhalation route will not be covered. Although reviews of animal models of hypersensitivity reactions and predictive testing have been discussed in other publications [2–5], this review will take the perspective of the pharmaceutical industry and focus more on how predictive models can be used for the drug development process.

Non-Clinical Testing Approaches for Drug Development

Testing Approaches Currently Used

Predictive animal models used to assess the potential of low MW compounds to produce hypersensitivity reactions or to be immunogenic in humans have not been adequately developed and validated. Therefore, some companies use the following approaches to minimize the risk for hypersensitivity reactions. Total Dose and Structural Alerts In a survey of drugs known to produce hypersensitivity reactions in humans it was found that most of these drugs are administered at doses 110 mg/day [5]. Thus, some companies may focus on efforts to increase potency of the drug to minimize the dose administered. Another approach is to compare the chemical structure of the test compound to drugs known to produce hypersensitivity reactions. For example, avoiding the structural features that would lead to reactive metabolites (e.g., epoxides, acyl glucuronides) is a common practice in the drug industry [4].

141

Discovery Toxicology Reactive Metabolite Screening Up until the mid to late 1990s, standard nonclinical toxicity studies in animals had been the primary focus of non-clinical safety testing for most pharmaceutical companies. Once the discovery groups had identified a potential candidate, a large amount of drug would be synthesized and handed on to toxicology groups for non-clinical toxicity testing. This approach, however, proved to be very costly since many compounds were discontinued due to toxicities observed in the animal studies. Currently, most companies have initiated safety testing during the discovery process to identify potential safety issues earlier to minimize investment. This ‘Discovery Toxicology’ approach uses in vitro screening assays and short-term animal studies to eliminate or rank order compounds with potential safety concerns [6]. This may involve testing for genetic toxicity, cardiovascular toxicity and/or adverse pharmacology. As a part of their discovery toxicology studies, many companies also have incorporated reactive metabolite screening assays into the discovery toxicology battery of tests. Reactive intermediates may be directly cytotoxic via different mechanisms and/or bind to proteins and act as immunogenic haptens. The generation of haptenspecific T cells and/or antibodies may result in hypersensitivity reactions with re-exposure. Thus, a positive in the reactive metabolite assay may trigger additional in vitro and in vivo studies to assess the significance of the reactive metabolites. The findings from these studies may stop further development of the drug candidate or lead to further ranking in comparison to other candidates. Reactive metabolite screening assays may involve generating reactive metabolites using liver microsomes and trapping electrophilic intermediates with GSH. The GSH adducts are then measured by liquid chromatography and mass spectroscopy methods. Additional methods that trap

142

free radicals and iminium ions or that use peroxidases to generate reactive intermediates have been reported. With compounds that form carboxylic acid metabolites, methods that look for acyl glucuronide formation are used [7]. To further minimize the risk of toxicity due to reactive intermediates, some companies also test for covalent binding to proteins in addition to the reactive metabolite screen [8]. This approach requires the preparation of 14C-labeled drugs. In vitro studies with the radiolabeled drug and human or animal liver microsomes or isolated hepatocytes would be conducted. In vivo studies with rats may also be considered. The amount of covalent binding to proteins is then measured. Binding 150 pg/mg of protein will trigger concern and will be taken into consideration along with other factors to determine future development. Since the formation of reactive metabolites and binding to proteins are just two steps in the complex process that may lead to hypersensitivity reactions, additional methods are needed to better understand the clinical risks for hypersensitivity reactions with these in vitro findings. Standard Toxicity Studies Once a candidate compound has been selected, additional safety concerns not identified during discovery toxicology studies are addressed by the standard repeat dose toxicity studies required by regulatory agencies before the initiation of clinical trials. These toxicity studies involve repeated dosing of rodents (usually rats) and non-rodent species (dogs or non-human primates) by the clinical route of administration at doses that result in blood levels greater than the expected therapeutic doses in the planned clinical studies. Clinical signs, body weight, hematological parameters and serum chemistry endpoints are evaluated during the in-life portion of the study. At study termination, selected organ weights and gross and histopathology of an extensive panel of tissues are evaluated. For most toxicities, standard toxicology studies are able to identify poten-

Kawabata ⴢ Piccotti

tial safety issues before clinical studies are initiated [9]. On the other hand, drug hypersensitivity reactions with low MW compounds are rarely observed in early discovery or non-clinical toxicity testing in animals but do occur in clinical studies. Exceptions include -lactam antibiotics that are known to produce immune-mediated hemolytic anemia in rhesus monkeys [10]. The lack of sensitivity of standard toxicity studies further strengthens the need to develop predictive testing models. Hypersensitivity-like type 1 reactions that do occur in non-clinical studies are commonly nonimmune-mediated anaphylactoid or pseudoallergic reactions [11, 12]. These usually occur with the first dose (mainly with intravenous infusion) due to direct activation of mast cells or complement. For example, opiates, smooth muscle relaxants, vancomycin and fluoroquinolones directly activate mast cells to cause histamine release. Other compounds such as radiocontrast dye or vehicles activate complement and form C5a and C3a. C5a is known to stimulate histamine release from mast cells. Methods that differentiate immune-mediated from pseudo-allergic reactions have been reported.

Potential Animal Models for Predictive Testing

The following sections provide an overview of the different types of predictive animal models reported to have the ability to identify drugs with the potential to produce hypersensitivity reactions. Most of these models evaluate the induction phase of the immune response and do not use hypersensitivity endpoints. A discussion regarding the pros and cons of these assays are presented. There are several animal models of hypersensitivity reactions such as the Brown Norway rat model for penicillamine and nevirapine reactions and the guinea pig halothane hepatitis

Non-Clinical Testing Approaches for Drug Development

model. Since these models are limited to specific drugs (e.g., halothane, nevirapine), they cannot be used as a general screen to predict hypersensitivity reactions in humans. However, the information gained from these models could be used to develop better predictive assays. These models are discussed in other chapters in this book and in other reviews [2]. Popliteal Lymph Node Assay The popliteal lymph node assay (PLNA) is one of the better characterized models developed for predictive testing. This assay appears to be capable of identifying drugs known to produce hypersensitivity and autoimmune-like reactions in humans, along with evaluating direct immunostimulation [13, 14]. Like other models for predictive testing, the readout of the PLNA is not a hypersensitivity reaction (e.g., rash) per se. The assay evaluates the ability of a test compound to activate lymphocytes in the popliteal lymph nodes by draining the hind paw, inducing a druginduced graft-versus-host like reaction. The direct (primary) PLNA was first reported by Gleichmann [15], who demonstrated that injecting the anticonvulsant diphenylhydantion into the footpad of mice induced significant increases in lymph node weight. Since this initial report, significant work has been done to further develop and characterize this experimental model [16]. The direct PLNA and its modifications (e.g., secondary, adoptive and reporter antigen PLNA) are discussed below. Direct PLNA The direct PLNA was developed as a rapid, reproducible and inexpensive test, to provide objective evaluation of the immunostimulatory potential of chemicals or the ability to produce hypersensitivity or autoimmune reactions. Test compound is injected subcutaneously (s.c.) into one footpad of a mouse or rat, and the induction of an immune response is measured in the popliteal lymph node by draining the hind limb approxi-

143

mately 7 days after treatment [13]. The contralateral footpad is left untreated, and serves as a control. Although the assay can be conducted in rats, mice are more commonly used. Objective endpoints such as lymph node weight and cell count are routinely measured, although evaluating lymph node cell number appears to be more sensitive than lymph node weight. The results of the assay are typically expressed as lymph node indices, which are the ratios of the treated versus controls. Indices of 62 and 65 are considered positive responses in weight and cellularity, respectively [16]. Other endpoints such as 3H-thymidine incorporation and cytokine production have been evaluated in the direct PLNA, although these parameters do not appear to improve the sensitivity of the assay. A significant limitation of the direct PLNA appears to be a lack of metabolizing capabilities in the footpad, which has resulted in false-negative results with compounds such as sulfamethoxazole, isoniazid, -methyldopa, gold(I) disodium thiomalate, procainamide and propylthiouracil, all which cause hypersensitivity or autoimmunelike reactions in humans [reviewed in 13]. The direct PLNA also is incapable of distinguishing an irritating compound from a sensitizer, which can lead to false-positives. Modifications of the direct PLNA, which focus on measuring secondary immune responses (see below), have been developed to help differentiate a compound’s irritancy and sensitization potential. Another important drawback of the direct PLNA that limits its use in drug development is the relevance of s.c. route since most drugs are taken orally by humans. Although other exposure routes have been investigated in the PLNA (and its modification), the number of compounds tested orally is limited. The limitations of the direct PLNA discussed above question the predictability of this assay for immunotoxicity assessment, and have led to significant refinement of the assay over the years. However, in its basic form the direct PLNA assay

144

could be used as an initial screen to evaluate the ability of a drug to stimulate lymphocytes in the popliteal lymph node as a measure of immune cell activation. A positive finding in the PLNA may trigger further testing to differentiate irritation and direct immunostimulation (e.g., adjuvant effects) from specific T-cell involvement. A negative result would question the requirement of metabolic activation, which may be overcome by preincubating the test compound with S9 fractions, or treating mice with the reactive metabolite [reviewed in 13]. However, the latter approach is not practical since the reactive metabolite would be unknown for the majority of sensitizing compounds, especially new drug candidates. Secondary and Adoptive PLNA As mentioned above, it is possible that the direct PLNA may not distinguish antigen-specific sensitization from an inflammatory response caused by a chemical irritant. Further, the direct assay is incapable of differentiating allergenicity from immunostimulation. In the secondary PLNA [14], mice are treated s.c. with the test compound as in the direct PLNA. Following 4–6 weeks, animals are given a low non-sensitizing dose of the drug into the same footpad. Popliteal lymph node weight and cellularity are measured within 4–6 days after challenge, which is faster than a primary immune response due to the presence of memory T cells. Thus, the secondary PLNA was developed as a challenge assay to test for drugspecific T-cell sensitization. In the adoptive PLNA [13, 16], a donor mouse is given the test chemical by a relevant human route. Splenocytes from the donor mouse are irradiated, and then injected s.c. into the footpad of a syngeneic recipient. A non-sensitizing dose of test chemical is then injected s.c. into the same footpad the following day, and the popliteal lymph nodes are harvested 4–6 days later. Since the concentration of the test chemical used is too low to induce a significant primary lymph node response alone,

Kawabata ⴢ Piccotti

increases in lymph node weight in the recipient result from the transfer of memory T cells from the donor mouse. Advantages of both the secondary and adoptive PLNA are their ability to investigate drugspecific priming of the immune system. A limitation of the adoptive PLNA, as with the direct method, is the lack of metabolizing activity. With compounds such as gold(I) disodium thiomalate, procainamide and propylthiouracil, the reactive metabolites are needed to generate a positive response in recipients [reviewed in 13]. Another drawback of the adoptive PLNA is a significant increase in the numbers of animals needed compared to the direct assay. Further, both assays are time consuming, and thus have limited use as initial screens. However, the secondary and adoptive PLNA may be used on a case-by-case basis to distinguish potential antigen-specific immune responses from irritancy and direct immunostimulation. Reporter Antigen PLNA Another modification of the direct PLNA is the reporter antigen assay, which was first reported by Albers and Pieters [17] in 1997. In this assay [reviewed in 14, 17], test compound is injected into the footpad together with either the reporter antigen TNP-ovalbumin or TNP-Ficoll. Serum is collected approximately 7 days later to measure anti-TNP antibodies responses. TNP-ovalbumin is recognized by T cells, which can provide help to B cells to produce anti-TNP antibodies. However, the concentration of TNP-ovalbumin used in the assay is too low to induce an immune response by itself. In order to make anti-TNP antibodies, adjuvant-like or costimulatory signals are needed. In contrast to TNP-ovalbumin, TNP-Ficoll is a T-cell-independent antigen that is not recognized by T cells. Treatment of mice with TNP-Ficoll can stimulate the production of immunoglobulin (Ig)M against TNP; however, Tcell help (i.e., the production of cytokines) is needed for the class switch to anti-TNP IgG. This

Non-Clinical Testing Approaches for Drug Development

assay allows for the differentiation of compounds that are capable of forming neo-antigens and stimulating T cells (positive IgG response to TNP-Ficoll and TNP-ovalbumin), chemical irritants (positive IgG to TNP-ovalbumin but not TNP-Ficoll), and innocuous compounds (no IgG produced to either reporter antigen). The reporter PLNA appears to be more sensitive than measuring lymph node weight and determining cell number. In addition to this advantage, the reporter PLNA provides information on the possible mechanism(s) of drug-induced immune reactions. A disadvantage of the reporter PLNA, as with the direct assay, is the use of s.c. exposures. Nierkens et al. [18] recently published a study that evaluated the hypersensitivity potential of D -penicillamine, diclofenac, and nevirapine in the reporter antigen PLNA following oral and intraperitoneal (i.p.) administration. The authors showed that all three compounds were capable of stimulating TNP-specific antibodies to TNP-ovalbumin following oral and i.p. injection of the drugs, demonstrating an adjuvant-like response. Further, mice pretreated with a single oral dose of D -penicillamine or diclofenac, and then given suboptimal doses of these compounds s.c. in the footpad 3 weeks later, had enhanced immune responses (IgG and interferon-) to TNP-Ficoll. These findings are consistent with results reported earlier by Gutting et al. [19], which demonstrated increases in footpad swelling, and popliteal lymph node weight and cellularity following a s.c. injection of diclofenac into the footpad of mice pretreated orally with the drug. Collectively, these results demonstrated the priming of drug-specific T cells following oral administration of compounds known to cause systemic drug reactions in humans. This approach shows promise in the prediction of systemic hypersensitivity reactions; however, additional compounds need to be tested, including sensitizing and non-sensitizing drugs. Future work also is needed to evaluate the metabolizing

145

capabilities of the assay. Ideally, a standardized experimental design (e.g., mouse strain, route of exposure, dosing protocol) is needed before the assay will be suitable for routine use in determining the sensitization potential of new drug candidates. Lymph Node Proliferation Assay The lymph node proliferation assay (LNPA) is a recent modification of the murine local lymph node assay (LLNA). In the standard LLNA [20], which is accepted as a stand-alone assay to evaluate the contact sensitization potential of compounds, mice are treated topically with a test compound to the dorsum of both ears on 3 consecutive days, followed by 2 days of rest. The proliferative responses of lymphocytes in the auricular lymph nodes draining the ears are measured by injecting 3H-thymidine into a tail vein approximately 5 h prior to lymph node harvest. In the LNPA [21], this experimental design is mimicked except that that mice are given a s.c. injection of the test compound on the top of the head between the ears and the cervical lymph nodes are harvested. This approach was developed to test its ability to predict the systemic hypersensitivity potential of compounds known to cause immune-mediated drug hypersensitivity reactions. A positive response required a SI 63, and a dose response. Four of the 8 drugs known to cause systemic drug reactions in humans (nevirapine, lamotrigine, abacavir and ofloxacin) were positive or gave mixed responses in this assay. As in the PLNA, 2 of the false-negatives, sulfamethoxazole and procainamide, appear to require metabolic activation in order to induce positive responses in rodents. The doses of clonidine and zomepirac, the other 2 false-negative drugs, were limited by toxicity, which may have prevented accurate evaluation of the sensitization potential of these compounds. Some of the insoluble compounds were administered as a suspension and therefore there may be concern about nonspecific irritation at the injection site. Until ad-

146

ditional compounds are tested in LNPA, the utility of this approach to evaluate drug-induced immune reactions is unclear. Contact Hypersensitivity Testing and Other Approaches The general mechanism for the induction of systemic and contact hypersensitivity reactions is very similar. Both involve the test compound becoming immunogenic by binding to cellular proteins and acting as a hapten. Thus, the potential use of contact sensitization methods such as guinea maximization test, guinea pig Buehler patch test and mouse LLNA to identify potential systemic sensitizers have been evaluated [22, 23]. Several positives in the contact hypersensitivity assays were found with drugs known to rarely produce hypersensitivity reactions in humans. These false-positives may be attributed to induction of immunological tolerance with oral administration. There were also compounds known to produce hypersensitivity reactions in humans which were negative with the contact hypersensitivity animal models. These false-negatives could be attributed to inadequate metabolic activation by the skin or penetration through the epidermis. Therefore, it can be concluded at this time that contact sensitization should not be used as a predictor for systemic sensitizers. Methods to assess immunogenicity (antigenicity) of systemically administered drugs have also been developed for the guinea pig. In these assays, test compound may be administered s.c. with adjuvant or by the clinical route of exposure. Specific antibodies against the drug (IgG1 and IgE) are measured by the active systemic anaphylaxis and/or passive cutaneous anaphylaxis assays in guinea pigs. Although these assays may be useful for -lactam and protein biologicals, their general use as a predictive model for systemic hypersensitivity testing of all low MW compounds is not recommended [1].

Kawabata ⴢ Piccotti

Testing Approaches Needed to Address Key Issues

The drug discovery and development process involves gathering efficacy, pharmacokinetics, pharmaceutics and safety data at different stages to make key decisions to move forward, delay or stop. Data that raise potential concerns for hypersensitivity reactions may be generated at different stages of drug development. The following four scenarios illustrate when and how predictive animal models or in vitro screens could be useful if these tests were available. Scenario 1 A lead candidate in discovery with good efficacy in animal models and pharmacokinetics profile was found to be positive in a reactive metabolite screen. Should this compound be selected as our development candidate? Is there a model we could use to rank our lead against other candidates? Is there a model that we can use to better understand the potential risk for hypersensitivity reactions? Since there are drugs known to produce reactive intermediates but have a low risk for hypersensitivity reactions (e.g., acetaminophen), additional in vitro methods or animal models are needed to further understand the potential risk with compounds found to be positive in the reactive metabolite assay. Without such approaches, compounds are moved forward with the risk in mind or discontinued prematurely. Scenario 2 Hypersensitivity reactions are observed during non-clinical or clinical studies. The severity and/ or incidence of reactions are at such a level that studies with the lead are stopped. This may trigger the initiation of a discovery program to find a backup candidate. For these efforts, an animal model that could be used to screen for a candidate that will not produce hypersensitivity reactions as the original lead compound would be needed. Models may also help to better under-

Non-Clinical Testing Approaches for Drug Development

stand why the compound produces hypersensitivity reactions and to find approaches to avoid this liability with other compounds. Mechanistic information may also help to determine if the hypersensitivity reaction is immune-mediated or pseudoallergic. This may involve measuring for specific IgE by immunoassays or using the in vitro lymphocyte transformation test (LTT) to detect drug-specific T cells. This differentiation may be important since pseudoallergic reactions may be easier to manage due to the dose-response relationship with these reactions in comparison to immune-mediated reactions in which a small challenge dose may trigger a significant response. Scenario 3 During phase 1 or 2 clinical trials, minor skin rashes that are treatment-related are observed. There is concern that these reactions may progress to more severe reactions with longer exposure and/or when larger populations of patients are exposed. A similar concern occurred with the development of gemifloxacin, a fluoroquinolone antibiotic [24]. A separate clinical rash safety study was then conducted and demonstrated that the rashes were benign exanthems. Rather than conducting a special clinical study, an animal model to specifically provide guidance to the manufacturer and regulatory agency would have saved a significant amount of time and resources. Scenario 4 There are general concerns about all safety issues that may arise during the later stages of clinical development or during commercialization. Thus, a method to screen out compounds that may produce low incidence (! 0.1%) severe hypersensitivity reactions in either late discovery or early development would be needed. Ideally, this assay would not be limited by the issues with the PLNA or LPNA assays and could be broadly used against a variety of compounds.

147

Future Considerations – The Possibilities

The type of animal model or in vitro assay developed for hypersensitivity testing will depend on the question being addressed. Models to address questions raised for scenarios 1 and 2 will be different from models for scenarios 3 and 4. This could range from a simple ‘fit-for-purpose’ assay used in discovery toxicology studies to rank compounds of a particular chemical class (scenarios 1 and 2) to a model which can more broadly screen out potential sensitizers (scenario 4). For the former question, the PLNA, LNPA or a modification could be used as an initial screen to evaluate reactivity of the parent compound. This type of assay could be useful if only a small amount of drug was available, if adequate drug solubility can be achieved for an injectable formulation, and if metabolism, irritation or direct immunostimulation are not expected. An assay that could be used to more broadly screen different types of compounds without the caveats of the PLNA or LNPA would be ideal. The endpoints of this type of model will not necessarily require hypersensitivity reactions (e.g., skin rash) since it may be too difficult to develop or too focused for the needs of screening. Thus, a model that simply measures the ability of a compound to induce a specific immune response against itself would be very useful. Conceptually, this model would use rats or mice and the same dosing route and formulation as used for the standard toxicity studies for the induction or sensitization phase. This would be followed by a challenge phase by the same route of administration and the measurement of sensitive endpoint of drug-specific responses. Alternatively ex vivo challenge may be used and lymphocyte responses measured (e.g., proliferation, cytokine production, expression of activation markers). In general, this assay would be a modification of the secondary PLNA and oral PLNA models described above.

148

A key hurdle in developing this model is being able to sensitize the animals to the compound using the clinical route of administration. It has been shown that high doses of drugs known to produce hypersensitivity reactions in humans do not produce immune responses to drugs in outbred rats (most commonly used for toxicity testing). For example, rats were administered sulfamethoxazole [25] or benzylpenicillin [26] at high doses for 4 weeks, which did not result in the production of antibodies to these drugs. Therefore, special modifications are needed, such as transgenic animals that express different cytochrome P450 isoenzymes, to generate the reactive intermediates and/or co-treatments that enhance costimulatory factors (e.g., toll-like receptor agonists) and down-modulate T-regulatory cells involved in tolerance. A model that predicts hypersensitivity responses that occur at a higher incidence (as with scenario 2) may be different from that used to predict compounds that produce the rare reaction (scenario 4). A better understanding of the patient and environmental factors that play a role in these rare idiosyncratic hypersensitivity reactions are needed. A model that addresses the issue in scenario 3 would be difficult to develop since it is not known why certain compounds produce more severe reactions while others just produce mild skin rashes. However, since the type of immune response produced (T-helper vs. cytotoxic T cells) could determine the severity of the reaction, this could be evaluated in a model. Before these assays are used by industry or recommended by regulatory agencies, they must be validated for predictivity and performance. This will involve evaluating a wide panel of compounds known to produce hypersensitivity reactions in humans (positive controls) as well as compounds that are not known to produce positives (negative controls). Weaver et al. [21] identified positive control compounds for the development of the LNPA. This was based on identifying drugs with adverse reaction reports that could be

Kawabata ⴢ Piccotti

Discovery Non-clinical

Concept

Lead

-SAR -Increase potency -Decrease reactive metabolites -Evaluate protein binding

In vitro T-cell and/or dendritic cell screen

Standard toxicity testing

Rank order & select optimal candidate(s )

Clinical

Diagnosis with LTT

Backup discover y

Confirm findings with non-clinical model

Fig. 1. Once in vitro and in vivo methods for hypersensitivity testing are developed, the following scheme may be used at different stages of drug development to minimize the risk for hypersensitivity reactions. During the concept stage of a discovery program, structure-activity relationship studies, efforts to increase potency, reactive metabolite and/or covalent binding assays may be conducted. If needed, in vitro assays for the potential to produce hypersensitivity reactions (e.g., dendritic cell acti-

vation) may be used to initially rank compounds. This may be followed up by in vivo animal models for hypersensitivity testing. If hypersensitivity reactions are observed, the lymphocyte transformation test or another in vitro diagnostic assay may be used to determine the causative drug and/or if the reaction is immune-mediated. A program may be initiated to identify backup compounds with a lower risk for hypersensitivity reactions.

attributed to a hypersensitivity reaction (e.g., rash, purities, urticaria, edema, allergic reaction, asthma). Metformin and phenobarbital were identified as negative controls in the review of the adverse events reported. For future validation studies, more negative controls will be needed. Validation of methods that predict drug-induced liver injury by immune-mediated hypersensitivity reactions may be more difficult since the mechanisms of liver injury/hepatitis for many drugs may be mediated by non-immune-mediated mechanisms (not involving drug-specific T cells or antibodies). Figure 1 illustrates when and how these different non-clinical methods may be used during the drug development process. During discovery

toxicology studies, screens such as the reactive metabolite screen and other in vitro methods could be initially used. Lead compounds may be further evaluated in predictive animal models. In conclusion, in order to develop better models, more information about mechanisms of hypersensitivity reactions observed in human or in animal models is needed. Fit-for-purpose models that address a specific drug development question for a specific type or class of compounds may be more practical at this time. In the future, however, models that more broadly predict hypersensitivity reactions that occur at very low incidence rates may be feasible with the incorporation of environmental and patient-related risk factors.

Non-Clinical Testing Approaches for Drug Development

149

References 1 Adkinson NF, Essayan D, Gruchalla R, Haggerty H, Kawabata T, Sandler JD, Updyke L, Shear NH, Wierda D; The Health and Environmental Science Institute Task Force: Task force report: future research needs for the prevention and management of immune-mediated drug hypersensitivity reactions. J Allergy Clin Immunol 2002;109: S461–S478. 2 Shenton JM, Chen J, Uetrecht JP: Animal models of idiosyncratic drug reactions. Chem Biol Interact 2004;150:53– 70. 3 Van Wijk F, Nierkens S: Assessment of drug-induced immunotoxicity in animal models. Drug Discov Today Technol 2006;3:103–109. 4 Uetrecht J: New concepts in immunology relevant to idiosyncratic drug reactions: the ‘danger hypothesis’ and innate immune systems. Chem Res Toxicol 1999;12:387–395. 5 Uetrecht J: Screening for the potential of a drug candidate to cause idiosyncratic drug reactions. Drug Discov Today 2003;8:832–837. 6 Mayne JT, Ku WW, Kennedy SP: Informed toxicity assessment in drug discovery: systems-based toxicology. Curr Opin Drug Discov Dev 2006;9: 75–83. 7 Nassar AE, Lopez-Anaya A: Strategies for dealing with reactive intermediates in drug discovery and development. Curr Opin Drug Discov Dev 2004;7: 126–136. 8 Evans DC, Watt, AP, Nicoll-Griffith DA, Baillie TA: Drug-protein adducts: an industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem Res Toxicol 2004;17:3–16. 9 Olson H, Betton G, Stritar J, Robinson D, Thomas K, Monro A, Kolaja G, Lilly P, Sanders J, Sipes G, Bracken W, Dorato M, van Duen K, Smith P, Berger B, Heller A: Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul Toxicol Pharmacol 2000; 32:56–67.

10 Lankas GR, Coleman JB, Klein HJ, Bailly Y: Species specificity of 2-aryl carbapenem-induced immune-mediated hemolytic anemia in primates. Toxicology 1996;108:207–215. 11 Szebeni J: Complement activationrelated pseudoallergy: a new class of drug-induced acute immune toxicity. Toxicology 2005;216:106–121. 12 Watkins J: Markers and mechanism of anaphylactoid reactions; in Assem ESK (ed): Allergic Reactions to Anesthetics. Clinical and Basic Aspects. Monogr Allergy. Basel, Karger, 1992, vol 30, pp 108–129. 13 Bloksma N, Kubicka-Muranyi M, Schuppe HC, Gleichmann E, Gleichmann H: Predictive immunotoxicological test systems: suitability of the popliteal lymph nodes assay in mice and rats. Crit Rev Toxicol 1995;25:369–396. 14 Pieters R: The popliteal lymph node assay: a tool for predicting drug allergies. Toxicology 2001;158:65–69. 15 Gleichmann H: Studies on the mechanism of drug sensitization: T-cell-dependent popliteal lymph node reaction to diphenylhydantoin. Clin Immunol Immunopathol 1981;18:203–211. 16 Ravel G, Descotes: Popliteal lymph node assay: facts and perspectives. J Appl Toxicol 2005;25:451–458. 17 Albers R, Broeders A, van der Pijil A, Seinen W, Pieters R: The use of reporter antigens in the popliteal lymph node assay to asses immunomodulatory chemicals. Toxicol Appl Pharmacol 1997;143:102–109. 18 Nierkens S, Aalbers M, Bol M, van Wijk F, Hassing I, Pieters R: Development of an oral exposure mouse model to predict drug-induced hypersensitivity reactions by using reporter antigens. Toxicol Sci 2005;83:273–281. 19 Gutting BW, Updyke LW, Amacher DE: BALB/c mice orally pretreated with diclofenac have augmented and accelerated PLNA responses to diclofenac. Toxicol 2002;172:217–230.

20 National Institute of Environmental Health Sciences: The Murine Local Lymph Node Assay: A Test Method for Assessing the Allergic Contact Dermatitis Potential of Chemicals/ Compounds. NIH Publ 99-4494. NIH, Research Triangle Park, NC, 1999. 21 Weaver JL, Chapdelaine JM, Descotes J, Germolec D, Holsapple M, House R, Lebrec H, Meade J, Pieters R, Hasting KL, Dean JH: Evaluation of a lymph node proliferation assay for its ability to detect pharmaceuticals with potential to cause immune-mediated drug reactions. J Immunotoxicol 2005;2:11– 20. 22 Weaver JL, Staten D, Swann J, Armstrong G, Bates M, Hastings KL: Detection of systemic hypersensitivity to drugs using standard guinea pig assays. Toxicology 2003;193:203–217. 23 Warbick EV, Dearman DJ, Kimber I: Prediction of drug allergenicity. Possible use of the local lymph node assay. Curr Opin Drug Discov Dev 2001;4:60– 65. 24 Shear NH: Gemifloxacin – a part of the slide presentation at the Factive (Gemifloxacin) FDA Advisory Committee Meeting, March 4, 2003. Slides 101– 158; http://www.fda.gov/ohrms/dockets/AC/03/slides/3931S1_04_LFLife% 20Sciences-Factive.pdf#search=%22 gemifloxacin%20cutaneous%20manifestations%22 25 Gill H, Hough SJ, Naisbitt DJ, Maggs JL, Kitteringham NR, Pirmohamed M, Park BK: The relationship between the disposition and immunogenicity of sulfamethoxazole in the rat. J Pharmacol Exp Ther 1997;282:795–801. 26 Kitteringham NR, Christie G, Colemen JW, Yeung JH, Park BK: Drug-protein conjugates. XII. A study of the disposition, irreversible binding and immunogenicity of penicillin in the rat. Biochem Pharmacol 1987;36:601–608.

Dr. Thomas T. Kawabata Pfizer Global Research and Development 1 Eastern Point Road, MS 8274-1206, Groton, CT 06340 (USA) Tel. +1 860 441 0527, Fax +1 860 715 9246 E-Mail [email protected]

150

Kawabata ⴢ Piccotti

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 151–165

Adverse Side Effects to Biological Agents Werner J. Pichler a  Paolo Campi b a

Department for Rheumatology and Clinical Immunology/Allergology, Inselspital, University of Bern, Bern, Switzerland; b Nuovo Ospedale San Giovanni di Dio, Florence, Italy

Abstract During the last years biological agents like cytokines, monoclonal antibodies and fusion proteins have been introduced in anti-inflammatory and tumor therapy. They are highly efficient, but can cause a great variety of adverse side effects. The peculiar features of biological agents require a novel approach to these side effects, which appear to be distinct from adverse side effects to xenobiotics. This novel approach is based on the often immunological activity of biological agents and differentiates five distinct types, namely clinical reactions due to high cytokine levels (type
), hypersensitivity due to an immune reaction against the biological agent (), immune or cytokine imbalance syndromes (), symptoms due to cross-reactivity (), and symptoms not directly affecting the immune system (). This approach and classification could help to better deal with the clinical features of these side effects, to identify possible individual and general risk factors, and to direct research in this novel area of medicine. Copyright © 2007 S. Karger AG, Basel

Introduction

Many new biological immune modulators (‘biological agents’) have entered the market during the last decade. They represent new therapeutic principles and include proteins such as cytokines, monoclonal antibodies and fusion pro-

teins (solubilized receptors) (fig. 1). Many of these biological agents have proven to be highly efficient tools in various inflammatory and oncologic diseases as their direct and focused effect makes them superior to immunosuppressive or cytotoxic drugs, whose use is often limited by severe generalized and unwanted side effects. The progress in this field was based on a better understanding of the immunological basis of many diseases, the identification of relevant molecules in inflammation as well as on tumor cells and the application of biotechnological techniques, which allowed to produce recombinant proteins like cytokines as well as humanized antibodies at a large scale [1]. The appropriate use of biological agents requires a special knowledge and familiarity with the disease to be treated. Moreover, not only the function of these compounds has to be understood, but also the underlying immunology – which is often rather complex. These biological agents are expensive medicines and force the treating doctors to consider economic aspects as well. Last but not least, there is some concern regarding the side effects of these biological agents, which are proteins used like drugs [2].

...ximab

...zumab

...mumab

TNF-
R

TNF-
Fab

Fc

Fc

b Monoclonal murine Ab

Chimeric Ab

Humanized Ab

Human Ab

a Development and non-human components of therapeutic antibodies

Fig. 1. Schematic representation of types of antibodies and fusion proteins. a Monoclonal antibodies are normally produced in mice and were originally used as mouse antibodies, causing a substantial anti-mouse immune response. Subsequently, only Fab (chimeric), with ca. 25–32% mouse origin, later only the antigen-binding site (humanized), with 5–10% of mouse origin, were pro-

duced, while the rest was human and thus not immunogenic. The newest generation of antibodies is completely human – in theory one can only react to the antigen binding site by developing anti-idiotypic antibodies. b Fusion protein between e.g. a TNF-
R and Fc portion of human IgG1.

Adverse side effects to drugs are clinically very heterogeneous. One approach was to differentiate so-called type A reactions, which correspond to the pharmacological activity of the drug, and are thus predictable [3] from type B reactions, which are not related to the pharmacological activity of the drug and are non-predictable (table 1). About 14–16% of side effects after drug treatment are type B reactions [4]. They are mainly immune-mediated side effects like hypersensitivity reactions. Clinically, these immunemediated side effects are very heterogeneous and can be subdivided according to different pathomechanisms [5, 6]. Biological agents differ from most drugs as they are not small chemical compounds (xenobiotics) but are proteins produced in a way to make them as similar to human proteins as possible (table 2, fig. 1). They are not metabolized like drugs but are processed like other proteins, and therefore need to be applied parentally, to avoid digestion in the gastrointestinal tract. Quite a few of

them are actually naturally occurring proteins (e.g. cytokines) or humanized antibodies able to neutralize natural proteins. These distinct biological and chemical features explain already that adverse reactions to biological agents differ from those elicited by drugs. The clear difference of xenobiotics and biological agents with regard to mode of action, chemistry, metabolism and immunogenicity (table 2) suggests that a somewhat different approach to their side effects is needed. Here some general aspects of these adverse side effects are outlined based more on the action of these substances than on their immunogenic potential. This results in a new classification for adverse effects of biological agents, which is based on mechanistic considerations. This subclassification might help to better understand and treat the patient who experiences them. Moreover, it might provide some help to avoid them in the future by defining risk factors and give directions for future research in this novel area.

152

Pichler ⴢ Campi

Table 1. Classification of adverse drug reactions vs. adverse reactions to biologicals (after Naisbitt et al. [3]) Adverse drug reactions

Adverse reactions to biologicals

Type A (augmented) reactions Predicted from the known pharmacology of the drug. These reactions are dose-dependent: examples are bleeding with anticoagulants

Type
: high cytokine and cytokine release syndrome Side effects might be connected to the systematic application of cytokines in relatively high doses or to high concentrations of cytokines released into the circulation [9]

Type B (bizarre) reactions Reactions are not predicted from the known pharmacology of the drug. They appear (but actually are not) relatively dose independent, as very small doses might already elicit symptoms. They include immune-mediated side effects like maculopapular exanthema, but also other hypersensitivity reactions, like aspirin induced asthma

Type : hypersensitivity The second group of reactions can be termed as ‘hypersensitivity’. Thereby basically three forms of allergies can be differentiated: IgE-, IgG-mediated and T-cell-mediated reactions

Type C (chemical) reactions1 – which are related to the chemical structure and its metabolism, e.g. paracetamol hepatotoxicity

Type : immune (cytokine) imbalance syndromes A major group of side effects have immunological features, but cannot be explained by high cytokine levels or typical hypersensitivity reactions. As shown in figure 2, these reactions can be further subdivided in impaired functions, and unmasking or causing an immune imbalance leading to either autoimmune or auto-inflammatory reactions

Type D (delayed) reactions1 – which appear after many years of treatment, e.g. bladder carcinoma after treatment with cyclophosphamide

Type : cross-reactivity Another cause for side effects might be that antibodies, e.g. generated to an antigen expressed on tumor cells might also cross-react with normal cells, which express this structure as well, albeit to a lower degree [40]

Type E (end of treatment) reactions1 – occur after drug withdrawal, e.g. seizures after stopping phenytoin

Type : non-immunological side effects Quite a few of the biological agents may elicit symptoms not directly related to the immune system, sometimes revealing unknown functions of the biological agents given or targeted

1

Types C–E are rarely used.

Table 2. Biological agents and drugs: important differences related to adverse side effects Biological agent

Drug

Structurally similar to autologous proteins

Synthesized chemicals (xenobiotics)

Are digested and processed, but not metabolized

Metabolized, reactive intermediates with potential immunogenicity (haptens)

Parental application required

Oral or parental

Immune-mediated effects are inherent in their activity, but hypersensitivities are rare and mainly due to immunoglobulins (IgE, IgG)

Immune-mediated side effects are unexpected, differ from the normal action of the drug and are often T-cell-mediated Drug interactions, organ toxicity

Adverse Side Effects to Biological Agents

153

Table 3. Types of biologicals (examples) Cytokines IFN-
, IFN-, IL-2, etc. Antibodies To soluble proteins like cytokines: anti-TNF-
(infliximab or adalimumab), anti-IL-2 (daclizumab) To cell surface molecules: anti-CD20 (rituximab), anti-IL-2 receptor (basiliximab), anti-LFA-1 (efalizumab) To IgE (omalizumab) To tumor antigens (e.g. epidermal growth factor receptor (EGFR), cetuximab, anti-HER2, trastuzumab) Fusion proteins (soluble receptors for cytokines or soluble cellular ligands) TNF-
RII (etanercept), a soluble TNF-
receptor CTLA4-Ig (abatacept) blocking CD28-CD80/CD86 interaction IL-1-receptor antagonist (anakinra)1 1

Not a fusion protein, but acting in a similar way.

Biological Agents

The biological agents can be subdivided into the following classes (table 3, fig. 1). Cytokines Cytokines, like for example interferon (IFN)-
, IFN-, interleukin-2 (IL-2) etc., are widely used biological agents. Some of these cytokines have been modified to prolong their in vivo half-life (containing polyethylene glycol, which reduces degradation, e.g. peg-IFNs). Their amino acid sequence is identical to human proteins but their glycosylation might differ. Antibodies While the original monoclonal antibodies used for therapeutic purposes were of mouse origin, the progress of molecular biological techniques allowed to modify these. Consequently, the majority of antibodies in use are in the meantime chimeric, humanized or fully human antibodies (fig. 1a). The chimeric antibodies like infliximab are characterized by ‘ximab’, while humanized antibodies like daclizumab or omalizumab carry a ‘zumab’ and fully human antibodies like adalimumab a ‘mumab’.

154

They may be directed to cytokines, in order to neutralize their action, to cell surface molecules, for example to inhibit migration or to inactivate or deplete cells (table 3). Fusion Proteins Natural receptors have often a very high affinity for their ligands and are thus as potent as highaffinity antibodies. To solubilize and increase the half-life of these normally cell-bound molecules, they are fused with the Fc part (CH2, CH3) of human IgG1. A special case is the naturally occurring soluble IL-1 receptor antagonists (anakinra) [7]. Soluble cytokine receptors are named using the ending -cept, like in etanercept (the p55, soluble TNF-
receptor II (TNF-
RII), see fig. 1b). Soluble cell ligands interfere with the cell-tocell communication. To block this interaction, either antibodies to the ligand or a soluble form of the ligand itself can be used to interfere. Thereby, co-stimulation of cells or their migration can be blocked.

Pichler ⴢ Campi

General Principle of Adverse Effects of Biological Agents

In a recent review of adverse reactions of biological agents, Lee and Kavanaugh differentiated between target-related or agent-related adverse side effects [10]. Indeed, target-related side effects are common with biological agents, as for example a biological agent may alter the composition and functional integrity of the normal immune response, and thereby predispose the patient to certain side effects, while the agent itself is rather ‘harmless’: Consequently, to understand the great variety of adverse side effects of biological agents, one has to be aware not only of the activity of the biological agent for the particular disease, but also of its influence on the normal immune balance. (a) Many biological agents have a well-defined range of physiological actions [8], some of which may already explain some adverse ‘side’ effects. For example, it is expectable that high concentrations of a proinflammatory cytokine like IFN-
can cause symptoms which are also observed during an immune reaction with high IFN-
levels (e.g. flu-like syndrome). (b) Biological agents often affect T and B cells or their products, as well as the different effector cells leading to various forms of inflammation. Side effects of biological agents affecting the immune response are to a certain degree predictable and consequently cannot be classified as unpredictable ‘type B’ reactions [3]. A question in the same line is: Can a hypersensitivity reaction to an injected protein, which contains parts of a foreign protein, be classified as an unpredictable type B reaction, or is it not actually a predictable reaction, even if it does not occur in each individual? (c) Immunological reactions during therapy with small molecular compounds (drugs) are mainly classified as hypersensitivity reactions [5, 6]. Hypersensitivity reactions are immune responses against the substance applied, which

Adverse Side Effects to Biological Agents

surely does not explain most of the side effects seen with these biological agents. The many distinct functions of these biological agents make it impossible to subclassify their adverse side effects based on clinical symptoms. More appropriate is a subclassification based on mechanisms of action and structure, as proposed in figure 2. To distinguish it from the classification of side effects to chemicals/drugs (type A–E reactions; see table 1), the Greek alphabet is used for the five types –
, , , ,  (table 1, fig. 2). This classification considers the well-accepted classification of side effects to drugs (table 1), as the first two types are similar (type A/
are both dose-dependent and related to the function of the drug or biological agent, type B/ comprise hypersensitivity).

Classification of Adverse Effects of Biological Agents

Type
– High Cytokine and Cytokine Release Syndrome Most cytokines (as well as chemokines) are produced locally and have a predominant local activity: their action is directed to the neighboring cell (paracrine) or has even an autocrine function [1]. Some cytokines like TNF-
or IL-5 have also a systemic activity, which comes into play if the immune reaction is strong and a systemic reaction of the immune system is required [1]. Thus, for most cytokines only the local concentration is relatively high, while the systemic concentrations are rather low and affect often bone marrow-derived progenitor cells. If the cytokine is applied therapeutically, the situation is different: comparatively high systemic concentrations are applied to achieve a sufficiently high concentration locally. Such high systemic concentrations can sometimes cause severe, non-tolerable side effects, limiting the use of cytokines (fever, myalgia, headache, etc.). Very high cytokine lev-

155

Type
High cytokine and cytokine release syndrome (anti-CD3) Immediate (IgE)

Type  Hypersensitivity Delayed (IgG+C', T cells)

Impaired function (immunodeficiency)

Type  Immune imbalance syndromes

Autoimmunity Immune or cytokine imbalance syndromes Auto inflammatory diseases

Type  Cross-reactivity

Type  Non-immunological side effects

Fig. 2. Classification of adverse effects of biological agents.

els can lead to severe diseases with multi-organ defects, as illustrated with the TGN1412 incident (see insert). One of the first monoclonal antibodies on the market was directed against CD3 (muromunab), which is the signal transmitting complex associated with the specific T-cell receptor for antigen. Cross-linking these T-cell receptor-associated molecules leads to activation of T cells and rapid release of different cytokines into the circulation with generalized symptoms like flash, arthralgia, capillary leak syndrome with pulmonary edema, encephalopathy, aseptic meningitis, pyrexia, gastrointestinal symptoms like severe vomiting or diarrhea, called cytokine release syndrome [9].

156

Type  – Hypersensitivity Reactions to Biological Agents Different factors determine the immunogenicity of the biological agent, and the type of clinical symptoms due to real hypersensitivity: Degree of Humanization Allergic reactions to biological agents are directed against the protein itself. The frequency of such reactions depends on the degree of humanization of the applied protein, which is often an antibody. The allergic immune response can be directed to the constant or the variable part. For example, while mouse antibodies (almost not used any more) as well as chimeric antibodies have at least some xenogenic determinants on its constant part, which can elicit quite rapidly an

Pichler ⴢ Campi

The TGN 1412 Event A dramatic illustration of the potency of certain biologicals gave the phase I study with TGN1412 [10]. This fully humanized antibody is directed to a certain epitope of CD28 on T cells, which is an important co-stimulatory molecule on T cells. It is called a superagonistic antibody, as it alone can stimulate T cells to proliferate and to release cytokines. Some experiments suggested also a preferential activation of T-regulatory cells, but injecting 0.1 mg/kg of this antibody in the frame of phase I study into healthy volunteers resulted in a massive generalized T-cell stimulation with rapid release of cytokines into the circulation: surprising was (a) the speed of this T-cell-mediated reaction (already starting after 30 min), (b) the type and severity of symptoms (massive swelling, mainly of the face, then of lymph nodes with rapid multiple organ failures and development of necrotic vasculitis, leading to loss of fingers), and (c) that previous in vitro and in vivo studies using also cynomolgus and rhesus monkeys had failed to predict the strengths of the reactions (only 2 animals showed a moderate lymph node swelling). As there was no mistake in handling the drug, as the preparation was not contaminated and as all 6 volunteers obtaining the drug were affected while two placebo-treated volunteers had no symptoms, it was surely not an allergic reaction: This effect of TGN1412 was interpreted as cytokine storm – and illustrates the catastrophic effect of extremely high cytokine concentrations.

immune response, humanized or fully human antibodies have a low immunogenicity as immunological tolerance exists to the constant part of the immunoglobulin. Nevertheless, the antigenbinding site of the monoclonal antibody can still elicit an immune response (anti-idiotypic) [13]. Cofactors Another important aspect for the immunogenicity of a biological agent is its content of adjuvants: For the cases of pure red cell aplasia observed in 2000 and 2001 outside the USA and related to erythropoietin injections, differences in rubber stoppers used for the vials containing the erythropoietin is thought to have contributed to the immunogenicity, as certain stoppers allowed the

Adverse Side Effects to Biological Agents

There is a lot to learn from this incident: For people who are familiar with the clinical symptoms of an acute DRESS (drug rash with eosinophilia and systemic symptoms), the similarity of this ‘cytokine storm’ to DRESS is striking, as facial and general swellings, multiorgan failure, systemic exanthema can be found in DRESS as well (but the lympho-, monocytopenia in the TGN1412 incident is different), see table 3 in Pichler, pp 168–189. Massive elevations of cytokines like IFN- and IL-5 can be detected in DRESS, albeit not as high as with TGN1412. Since the pathomechanism of T-cell stimulation in the TGN1412 incident is rather well explained by the action of TGN1412, one can infer that the DRESS symptoms may have a similar pathomechanism: this xenobiotic-induced reaction may be due to a superantigen-like activity of drugs on the T-cell receptor [11]. The even more acute and dramatic course with TGN1412 may be due to the intravenous injection and the high affinity of the TGN1412 antibody for CD28. Together, the clinical picture and the known activity of TGN1412 was due to a massive cytokine release syndrome through T-cell stimulation (type
of the classification presented). It was an overdose and misconception of the antibody activity, which happened because the preclinical in vitro or in vivo tests were not informative or not correctly interpreted [12] and because the original concept of a superagonistic antibody, which stimulates all CD28+ T cells, was neglected in view of the possible effect on T-regulatory cells.

diffusion of some organic compounds with adjuvant activity inside the vial, which was enough to cause immunogenicity of the erythropoietin [14]. The way of application (s.c. versus i.v.), the IgG isotype of the biological agent, and in particular the amount of immunosuppressive co-treatment may also have an influence. For example, the sensitization and antibody formation to infliximab, a chimeric anti-TNF-
antibody, can be reduced by the co-medication with methotrexate [15, 16]. Type of Allergic Reaction IgE-mediated reactions can cause a local whealand-flare reaction at the injection site if applied subcutaneously, but it may also cause urticaria and anaphylaxis. Such a reaction appears rather

157

rapidly, that means within 20 min after the injections. One has to differentiate it from an unspecific irritation induced by the solvent, which might also lead to local redness and a typical wheal. Irritative responses are often diminished at subsequent applications – but this is not a strict criterion to differentiate it from real allergy, as tolerance might develop in IgE-mediated reactions as well. The majority of these allergic reactions are mild, but severe IgE-mediated anaphylaxis has also been described [17]. In table 4 the hypersensitivity reactions of immediate type described for IFN-
, IFN-, IL-2, infliximab, adalimumab, daclizumab, rituximab, basiliximab, efalizumab, omalizumab, cetuximab, trastuzumab, etanercept, abatacept and anakinra over the last 5 years are summarized, based on a Medline search. Allergological workups (in vivo or in vitro) were performed in only 7 out of 17 reports. In these altogether 11 cases, skin tests were positive in 4 patients (1 without controls), and serum IgE was detected in 1 case. In conclusion, true, IgE-mediated immediate reactions to biologicals appear to be infrequent. Acute infusion reactions are mostly not IgEmediated. They occur in 3–5% of patients treated with chimeric antibodies, often already during the (first) infusion and can be reduced by slowing the infusion rate [18]. Their pathomechanism is unclear, but may be related to activation of cells (by Fc-IgG receptors) or of the complement system via immune complexes, as they appear more frequently when antibodies are detectable [15, 18, 19]. Symptoms are chills, nausea, dyspnea, headache and fever [18]. Delayed reactions appear 16 h after the application. They can be subdivided in immunoglobulin- and T-cell-mediated reactions. The normal physiological immune response to a foreign, soluble protein is immunoglobulin-mediated. Thus, the development of IgG antibodies directed to the biological agent is by far the most frequent reaction. Formation of IgG antibodies against the biological agent may occur rather frequently, if

158

the biological agent is immunogenic and if no immunosuppression-like methotrexate accompanies the treatment [15, 19]. In a study with infliximab, up to 68% of the treated patients developed antibodies to this chimeric antibody [15, 16]. These antibodies are not necessarily associated with symptoms. The most frequent effect is inactivation of the biological agent: the half-time of an injected cytokine or antibody is reduced and the patient needs more of the biological agent or an alternative to achieve the same effect. However, if the substance injected is unique for a certain function, the inactivation may have severe consequences: this has been shown for antierythropoietin antibodies, which lead to pure red cell aplasia [14]. Formation of antibodies to the biological agent may also result in activation of the complement cascade via immune complex formation as well as by Fc-IgG receptor-mediated activation of the neutrophils and may thus cause immune complex diseases like serum sickness, vasculitis and nephritis. Some symptoms appear after 3–12 days, and are classified as delayed infusion reactions, characterized by myalgia, arthralgia, fever, ‘rash’, pruritus, facial and lip edema, dysphagia and urticaria [18]. Another immunoglobulin-associated side effect may be thrombocytopenia, if immune complexes are formed that bind to FcIgG receptors on thrombocytes, which are then removed from the circulation by the phagocytic system in liver and spleen [19; also see chapter of Aster, pp 306–320]. In these immunoglobulin-dependent reactions, T cells are probably also involved, but mainly as regulators of the humoral immune response. In contrast to hypersensitivity reactions to small molecular weight compounds (chemicals/drugs), where T-cell-mediated reactions cause different forms of exanthems or hepatitis, etc. [6], biological agents seem to elicit such reactions quite rarely. However, immunohistological examinations of delayed appearing and persisting injection site reactions to etanercept (soluble TNF-
R) re-

Pichler ⴢ Campi

Table 4. Immediate hypersensitivity reactions to some biologicals (type  hypersensitivity) Drug

Cytokines IFN-1b

Disease treated

Symptoms

Onset of the reaction

Allergological studies

Desensitization

Multiple sclerosis

Urticaria

after 9 months

I.d. positive; negative with IFN-1a

ND

47

Facial erythema, flushing, dyspnea, shock

after 6 months

ND

ND

48

Flush, chest tightness, dyspnea, nausea, hypotension

3rd dose

ND

Successful

49

Nausea, dizziness, dysphagia, urticaria dyspnea

2nd dose

Scratch negative

Unsuccessful (reaction to 1.3 mg)

Etanercept

50

Dyspnea

1st dose

ND

ND

Infliximab

51

Diffuse erythema, mild dyspnea

2nd dose

ND

ND

Infliximab

51

Diffuse erythema, facial swelling, dyspnea

2nd dose

ND

ND

51

Flushing, urticaria, chest pain, dyspnea, hypotension, nausea, vomiting

2nd dose

ND

ND

52

Flushing, urticaria, dyspnea, dysphagia, chest pain, hypotension

2nd dose

ND

ND

53

Flushing, urticaria, dyspnea, dysphagia, chest pain, hypotension

3rd dose

ND

ND

53

Flushing, tachycardia, dyspnea (‘red man syndrome’)

2nd dose

ND

ND

1st dose

ND

ND

2nd dose

ND

ND

3rd dose

ND

ND

Antibodies Crohn’s Infliximab disease (chimeric mAb to TNF-
)

Flushing, tachycardia, dyspnea, vomiting

vealed infiltration of T cells [20], suggesting that T-cell reactions themselves may cause clinical symptoms. If a hypersensitivity reaction is suspected, one can confirm it by skin tests (prick, intradermal) with the biological agent: if specific IgE to the biological agent is present, a local wheal-and-flare reaction might appear after a prick test or intradermal test (preferentially with 1: 10–1: 1,000 in PBS-diluted material); if T cells are involved, induration and vesicle formation can be seen after

Adverse Side Effects to Biological Agents

Drug tolerated afterwards

Ref.

Infliximab (starting slowly) Infliximab (starting slowly) Infliximab (starting slowly)

54

Adalimumab

55

24–72 h. Alternatively, ELISA assays detecting newly formed antibodies to the biological agent (human anti-chimeric or anti-mouse-Ig antibodies, etc.) can confirm the presence of antibodies. Type  – Immune/Cytokine Imbalance Syndromes Quite a few side effects to biological agents cannot be explained by high concentrations or by an immune response directed to the biological agent and are thus not hypersensitivities. Tests to detect

159

54 54

Table 4 (continued) Drug

Disease treated

Symptoms

Onset of the reaction

Allergological studies

Desensitization

Drug tolerated afterwards

Ref.

Dyspnea, lightheadedness, hypotension, flushing, fever or rash (5 patients)

3rd–11th dose

ND

ND

Adalimumab (4 out of 5)

56

Dyspnea, angioedema

6th dose

I.d. negative

Successful

57

Dysphagia, facial erythema, chest pain, tachycardia

7th dose

I.d. negative

Successful

57

Dyspnea, tachycardia, angio-edema

2nd dose

I.d. negative

Successful

57

Skin and mouth erythema, itching, dyspnea, chest pain

5th dose

I.d. negative

Successful

57

Dyspnea, urticaria, hypotension Adult onset. Still disease

3rd dose

I.d. positive (controls?)

ND

58

Psoriatic arthritis

Dyspnea, laryngeal spasm, vomiting, hypotension, hypoxemia

2nd dose

ND

ND

59

Basiliximab (chimeric mAb to IL-2 receptor)

Renal transplant

Dyspnea, angioedema, rash

2nd dose

I.d. positive (daclizumab negative)

ND

Bronchospasm, hypotension

2nd dose

Serum IgE (ELISA) and CAST positive; both negative for daclizumab

ND

Trastuzumab (humanized mAb to HER2)

Breast cancer

Crohn’s Antibodies disease Infliximab (chimeric mAb to TNF-
)

60

61

Rash, angioedema, nausea, vomiting

4th dose

I.d. positive

Successful

62

Itching, throat tightening

after 9 months

I.d. negative

Successful

62

At 20 min: severe headache, back pain, general fatigue; at 90 min: hypotension, hypoxia (collapsed lung)

1st dose

ND

ND

63

hypersensitivities, like skin tests with the biological agent as well as in vitro determinations of antibodies to the substrate, are negative. Some side effects might be explained by the potent and unique activity of the biological agent in certain types of the normal immune response or by the elimination of certain cytokine activity by an injected antibody. Other effects are often not explainable as easily. They may reveal a new or hitherto neglected activity of the biological agent given or eliminated. Thereby, the broader the

160

Daclizumab

physiological role of the biological agent, the more heterogeneous effects can be seen. For example, recombinant erythropoietin or omalizumab, an antibody directed against IgE, have a limited pattern of adverse side effects, as it replaces or reduces a certain effector molecule with a limited function in the immune system. In contrast, targeting essential, broadly active cytokines like IFN-
or TNF-
are associated with a wide variety of quite different side effects, due to the very broad activity of these cytokines (table 5) [2, 21–23].

Pichler ⴢ Campi

Table 5. Subclassifying side effects of INF-
and anti-TNF-


Type
High dose

Type  Hypersensitivity

INF-


Anti-TNF-
(infliximab)

Flu-like symptoms – Myalgia – Arthralgia – Fever

–

Local and generalized urticaria, local dermatitis

Local and systemic urticaria, erythema, serum sickness, loss of efficiency, acute and delayed infusion reactions, local dermatitis

Type  Cytokine or immune imbalance syndromes Immunodeficiency –

Autoimmune/ autoinflammatory disorders

Type  Cross-reactivity Type  Non-immunological side effects

Tuberculosis, listeriosis Other granulomatous infectious diseases

Thrombocytopenia, hemolytic anemia, IgA nephropathy, dermatitis herpetiformis, SLE, vasculitis, thyroid disease, pernicious anemia sarcoidosis, psoriasis, vitiligo

Interstitial pneumopathy, acute fibrosis, systemic sclerosis, SLE, demyelinating diseases, pancytopenia, lichenoid skin reaction, neutrophilic (psoriasis) or eosinophilic skin diseases (atopic dermatitis)

–

– (?)

Neurological symptoms like Bell‘s palsy, hearing loss, depression, dystonia, restless legs

Heart insufficiency

Impaired Function (Immunodeficiency) Quite a few of the biological agents are actually used in inflammatory disorders or transplantation and one aim of the treatment is to dim the inflammation or the immune response to the transplanted organ [13, 24]. The best understood and to a certain extent expected adverse side effect of certain biological agents is the impaired function of the immune system resulting in a certain immunodeficiency. Actually, one could classify ‘impaired function’ also as a predictable, type
reaction. However, type A/
reactions are mainly due to a high dose, which may or may not lead to immunodeficiency and therefore impaired function is classified within imbalance syndromes. Typical examples would be efalizumab, an antibody to LFA1 (CD11a), the ligand for CD18 on

Adverse Side Effects to Biological Agents

neutrophils and T cells. It inhibits the migration of these cells into the affected tissue [24]. While this may be beneficial for example in psoriasis, it may be contraproductive for the optimal and rapid control of infections. TNF-
is another example: one main obstacle in the use of anti-TNF-
therapy is the danger that an underlying disease like tuberculosis or listeriosis escape the control of the immune response and disseminate, as TNF-
is essential for the control of these intracellular infections by stimulating macrophage function [2, 25]. Unmasking a Preexisting Imbalance or Causing an Imbalance The immune system is well balanced, and central and peripheral tolerance mechanism, regulatory T cells, certain cytokines like TGF- and IL-10,

161

as well as the Th1/Th2 balance are involved [8, 21, 23]. A disturbance of this balance can occur by eliminating or injecting certain cytokines, which have an immunoregulatory function. It can result in autoimmunity (with autoantibodies, e.g. systemic lupus erythematodes) and autoinflammatory disorders (e.g. T-cell-regulated eosinophilic or neutrophilic inflammations without autoantibodies, e.g. psoriasis) if the immunological tolerance to autoantigens is altered. Or it might lead to the appearance of other immunological reactions, which are normally suppressed, like for example an immune response to a harmless exogenous antigen (atopic disorders). All these patterns have been described for anti-TNF
, IFN-
, anti-CTLA4 antibodies and others [21, 22, 26–29] (table 5). Autoimmunity and Autoinflammatory Disorders. Both reactions are well documented in the frame of anti-TNF-
or IFN-
treatment, and also possible with other treatments with biologicals. TNF-
neutralization leads rather frequently to autoimmune phenomena and rarely even to autoimmune diseases (with autoantibodies). Antinuclear antibodies can be found in up to 11% of patients treated with etanercept, a soluble TNF
R [30], and in up to 68% in patients treated with infliximab [15, 16], but the development of a clinical symptom of lupus is a rather rare event (
0.5%) [2, 28]. The development of demyelinating diseases has also been observed under anti-TNF-
treatment [30] and treatment of patients with multiple sclerosis with lenercept (the soluble p75 form of the TNF-
RI) had to be stopped as the disease became more severe [2]. Not so unusual are skin diseases under antiTNF-
treatments: these could be neutrophilic or eosinophilic. Even a pustular psoriasis can develop under anti-TNF-
treatment, which is quite astonishing, as psoriasis itself responds rather well to anti-TNF-
treatment [31]. The reason for autoimmune reactions by this presumably im-

162

munosuppressive treatment is unclear, but deregulated TNF-
production has been associated with autoimmunity [21, 23], and TNF-
is expressed in normal eccrine sweat duct cells [31]. Autoinflammatory diseases may also arise if a shift of the Th1-Th2 balance occurs. Th1 cells booster macrophage function, the production of complement fixing antibodies and cellular immune responses, while Th2 cells enhance the production of IgE/IgG4 and eosinophilic inflammations. These T-cell subsets do not only regulate the type of immune response but both T-cell subsets do also control each other, as for example the Th2 cytokine IL-4 down-regulates Th1-driven macrophage functions, but boosters IgE responses, while the Th1 cytokine IFN- stimulates macrophages but can suppress IgE [1]. Biological agents can interfere with this balance, for example a Th1-driven autoinflammatory process might be enhanced by IFN-
or be dimmed by reducing the high activity of for example TNF-
. This shift may uncover a hitherto controlled readiness to generate a Th2 response [27, 32–34]. As a consequence, the development of an eosinophilic inflammation due to suppression of TNF
may occur under anti-TNF-
treatments, with the clinical features similar to an atopic dermatitis. It is unknown whether auto- or exogenous antigens are driving this reaction [32, 33]. Interestingly, the use of an immunostimulatory cytokine like IFN-
treatment may also induce autoimmune and autoinflammatory diseases, as lupus-like syndrome, systemic sclerosis, Guillain-Barré syndrome, autoimmune thyroid disease, idiopathic thrombocytopenic purpura, vitiligo and psoriasis have been described [19, 29, 35–38] (table 5). The underlying pathomechanism is not yet understood: beside the discussed shift in cytokine balance, also the immunostimulatory effects of IFN-
leading to the appearance of hidden antigens, an increased expression of co-stimulatory molecules or an enhanced signaling in activated B cells secreting autoantibodies may be responsible [39].

Pichler ⴢ Campi

Abrogating the suppressive function of activated CTLA4+ T cells, which have immunoregulatory properties, by anti-CTLA4 antibodies (MDX-010) may also lead to skin symptoms like an eosinophilic dermatitis similar to drug hypersensitivity [27]. The immune imbalance syndromes are clinically very heterogeneous, dependent on the effect of eliminating or bolstering a crucial cytokine or function/expansion of a cell. They occur only in a minority, suggesting that either the individual predisposition (e.g. psoriasis) or an individual comorbidity may be important in order that the treatment results in a clinical symptom. These immune or cytokine imbalance syndromes are complex diseases and surely need to be better defined for each biological agent, which may reveal interesting, neglected aspects of the target molecule and pave the way to identify individuals at risk. Type  – Cross-Reactivity Cross-reactivity can be due to expression of the same antigen on different tissue cells or to the reaction of the antibody with a similar structure. Tumor antigens are often ‘normal’ proteins, which are overexpressed on tumor cells. Antibodies to these antigens may also react with these structures on normal cells: for example, epidermal growth factor receptor (EGFR) is strongly expressed on a variety of carcinomas of different origin and is thought to be partly associated with tumor progression [40]. In addition, EGFR plays a major role in the homeostasis of the epidermis and epidermal appendages. Antibodies to these EGFR (e.g. cetuximab) are used in the treatment of various tumors. Interestingly, acneiform eruptions appear very frequently in the frame of these anti-EGFR treatments – possibly due to cross-reactivity with EGFR on skin cells [40]. Actually, the appearance of such side effects correlates to the efficacy of treatment. It also cannot be ruled out that some of the antibodies used do in addition react with structurally similar proteins – and thus cause unexpected side effects.

Adverse Side Effects to Biological Agents

Type  – Non-Immunological Side Effects Many molecules originally detected in the immune system and inflammatory response may also be involved in other physiological functions. Actually, the in vivo use of a biological agent in humans may reveal these ‘new’ functions. Examples are blocking CD40-CD40 ligand interactions (important for immunoglobulin class switch in B cells), where both soluble CD40-Ig or anti-CD40L antibodies precipitated the appearance of thrombosis and subsequently the detection of the CD40 and CD40L on thrombocytes [41]; or the role of TNF-
in heart failure, where high TNF-
levels were detected, but neutralization of TNF-
leads to aggravation of the disease [42]. Also the rather frequent neuropsychiatric adverse effects of IFN-
(acute confusional state, depression), as well as various retinopathies observed during IFN-
treatments may represent such type  reactions [42, 43]. Manifestations of such non-immunological side effects might actually be quite frequent. Some of these type  reactions may be due to cross-reactivity (type  reactions) if antibodies are involved. On the other hand, such unexpected side effects of biological agents provide a chance to detect new functions of molecules which were originally detected in the immune response, but play a role outside of it as well. A further aspect to be considered in the evaluation of side effects is the combined use of biological agents and drugs. An example would be the treatment of hepatitis C infection, where IFN
is often provided in combination with ribavirin: if an anemia develops, it might be related to ribavirin, while the development of autoimmunity is likely due to IFN-
itself [44, 45]. And in oncology, where many biological agents are in use, attempts are made to increase the efficacy of the treatment by coupling cytotoxic or radioactive compounds to biological agents, which of course can also be responsible for adverse side effects [46].

163

Conclusion

Biological agents are often highly potent tools in the treatment of various severe diseases. They can cause a variety of adverse side effects which, at the present state of research, often appear puzzling, but may be very instructive as well. Biological induced side effects should be differentiated from xenobiotic-induced side effects due to distinct chemical and biological features. A subclas-

sification is proposed which is based on the pathomechanism and biological properties of the biological. It follows certain rules, but the clinical features might differ widely. This classification is at first sight rather complex, however this complexity is inherent in the action of various biologicals. It may, nevertheless, be a practical guideline to better approach, treat and advise patients with side effects due to biologicals appropriately and prevent these side effects in the future.

References 1 Abbas AK, Lichtman AH: Cellular and molecular immunology, ed 5. Philadelphia, Saunders, 2003. 2 Weber RW: Adverse reactions to biological modifiers. Curr Opin Allergy Clin Immunol 2004;4:277–283. 3 Naisbitt DJ, Gordon SF, Pirmohamed M, Park BK: Immunological principles of adverse drug reactions: the initiation and propagation of immune responses elicited by drug treatment. Drug Saf 2000;23:483–507. 4 Hoigne R, Schlumberger HP, Vervloet D, Zoppi M: Epidemiology of allergic drug reactions. Monogr Allergy 1993; 31:147–170. 5 Coombs PRA GP: Classification of allergic reactions responsible for clinical hypersensitivity and disease; in Gell RRA (ed): Clinical Aspects of Immunology. Oxford, Oxford University Press, 1968, pp 575–596. 6 Pichler WJ: Delayed drug hypersensitivity reactions. Ann Intern Med 2003; 139:683–693. 7 Waugh J, Perry CM: Anakinra: a review of its use in the management of rheumatoid arthritis. BioDrugs 2005; 19: 189–202. 8 Abbas AK, Murphy KM, Sher A: Functional diversity of helper T lymphocytes. Nature 1996;383:787–793. 9 Vasquez EM, Fabrega AJ, Pollak R: OKT3-induced cytokine-release syndrome: occurrence beyond the second dose and association with rejection severity. Transplant Proc 1995;27:873–874. 10 Suntharalingam G, Perry MR, Ward S, et al: Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N Engl J Med 2006;355: 1018–1028.

164

11 Pichler WJ: Pharmacological interaction of drugs with antigen-specific immune receptors: the p-i concept. Curr Opin Allergy Clin Immunol 2002;2: 301–305. 12 Schrodt BJ, Kulp-Shorten CL, Callen JP: Necrotizing vasculitis of the skin and uterine cervix associated with minocycline therapy for acne vulgaris. South Med J 1999;92:502–504. 13 Chatenoud L: Immunologic monitoring during OKT3 therapy. Clin Transplant 1993;7:422–430. 14 Boven K, Knight J, Bader F, Rossert J, Eckardt KU, Casadevall N: Epoetinassociated pure red cell aplasia in patients with chronic kidney disease: solving the mystery. Nephrol Dial Transplant 2005;20(suppl 3):33–40. 15 Baert F, Noman M, Vermeire S, et al: Influence of immunogenicity on the long-term efficacy of infliximab in Crohn’s disease. N Engl J Med 2003; 348:601–608. 16 Lipsky PE, van der Heijde DM, St Clair EW, et al: Infliximab and methotrexate in the treatment of rheumatoid arthritis. Anti-Tumor Necrosis Factor Trial in Rheumatoid Arthritis with Concomitant Therapy Study Group. N Engl J Med 2000;343:1594–1602. 17 Abramowicz D, Crusiaux A, Goldman M: Anaphylactic shock after retreatment with OKT3 monoclonal antibody. N Engl J Med 1992;327:736. 18 Han PD, Cohen RD: Managing immunogenic responses to infliximab: treatment implications for patients with Crohn’s disease. Drugs 2004;64:1767–1777. 19 Arimura K, Arima N, Ohtsubo H, et al: Severe autoimmune thrombocytopenic purpura during interferon-
therapy

20

21

22

23

24

25 26

27

for chronic myelogenous leukemia. Acta Haematol 2004;112:217–218. Werth VP, Levinson AI: Etanerceptinduced injection site reactions: mechanistic insights from clinical findings and immunohistochemistry. Arch Dermatol 2001;137:953–955. Kassiotis G, Kollias G: TNF and receptors in organ-specific autoimmune disease: multi-layered functioning mirrored in animal models. J Clin Invest 2001;107:1507–1508. Vermeire S, Noman M, Van Assche G, et al: Autoimmunity associated with antitumor necrosis factor-
treatment in Crohn’s disease: a prospective cohort study. Gastroenterology 2003;125:32–39. Banchereau J, Pascual V, Palucka AK: Autoimmunity through cytokine-induced dendritic cell activation. Immunity 2004;20:539–550. Hershberger RE, Starling RC, Eisen HJ, et al: Daclizumab to prevent rejection after cardiac transplantation. N Engl J Med 2005;352:2705–2713. Wellington K, Perry CM: Efalizumab. Am J Clin Dermatol 2005; 6:113–120. Gomez-Reino JJ, Carmona L, Valverde VR, Mola EM, Montero MD: Treatment of rheumatoid arthritis with tumor necrosis factor inhibitors may predispose to significant increase in tuberculosis risk: a multicenter active-surveillance report. Arthritis Rheum 2003;48: 2122–2127. Phan GQ, Yang JC, Sherry RM, et al: Cancer regression and autoimmunity induced by cytotoxic T-lymphocyteassociated antigen 4 blockade in patients with metastatic melanoma. Proc Natl Acad Sci USA 2003;100:8372– 8377.

Pichler ⴢ Campi

28 Debandt M, Vittecoq O, Descamps V, Le Loet X, Meyer O: Anti-TNF-
-induced systemic lupus syndrome. Clin Rheumatol 2003;22:56–61. 29 Seckin D, Durusoy C, Sahin S: Concomitant vitiligo and psoriasis in a patient treated with interferon-alfa-2a for chronic hepatitis B infection. Pediatr Dermatol 2004;21:577–579. 30 Day R: Adverse reactions to TNF-
inhibitors in rheumatoid arthritis. Lancet 2002;359:540–541. 31 Michaelsson G, Kajermo U, Michaelsson A, Hagforsen E: Infliximab can precipitate as well as worsen palmoplantar pustulosis: possible linkage to the expression of tumour necrosis factor-
in the normal palmar eccrine sweat duct? Br J Dermatol 2005; 153: 1243–1244. 32 Devos SA, Van Den Bossche N, De Vos M, Naeyaert JM: Adverse skin reactions to anti-TNF-
monoclonal antibody therapy. Dermatology 2003;206:388–390. 33 Chan JL, Davis-Reed L, Kimball AB: Counter-regulatory balance: atopic dermatitis in patients undergoing infliximab infusion therapy. J Drugs Dermatol 2004;3:315–318. 34 Menon Y, Cucurull E, Reisin E, Espinoza LR: Interferon-
-associated sarcoidosis responsive to infliximab therapy. Am J Med Sci 2004;328:173–175. 35 Solans R, Bosch JA, Esteban I, Vilardell M: Systemic sclerosis developing in association with the use of interferon-
therapy for chronic viral hepatitis. Clin Exp Rheumatol 2004;22:625–628. 36 Niewold TB, Swedler WI: Systemic lupus erythematosus arising during interferon-
therapy for cryoglobulinemic vasculitis associated with hepatitis C. Clin Rheumatol 2005;24: 178–181. 37 Boz C, Ozmenoglu M, Aktoz G, Velioglu S, Alioglu Z: Guillain-Barré syndrome during treatment with interferon-
for hepatitis B. J Clin Neurosci 2004;11:523–525. 38 Doi F, Kakizaki S, Takagi H, et al: Long-term outcome of interferon-
induced autoimmune thyroid disorders in chronic hepatitis C. Liver Int 2005; 25:242–246.

39 Rifkin IR, Leadbetter EA, Busconi L, Viglianti G, Marshak-Rothstein A: Tolllike receptors, endogenous ligands, and systemic autoimmune disease. Immunol Rev 2005;204:27–42. 40 Perez-Soler R, Saltz L: Cutaneous adverse effects with HER1/EGFR-targeted agents: Is there a silver lining? J Clin Oncol 2005;23:5235–5246. 41 Danese S, Fiocchi C: Platelet activation and the CD40/CD40 ligand pathway: mechanisms and implications for human disease. Crit Rev Immunol 2005; 25:103–121. 42 Kwon HJ, Cote TR, Cuffe MS, Kramer JM, Braun MM: Case reports of heart failure after therapy with a tumor necrosis factor antagonist. Ann Intern Med 2003;138:807–811. 43 Kasahara A, Hiraide A, Tomita N, et al: Vogt-Koyanagi-Harada disease occurring during interferon-
therapy for chronic hepatitis C. J Gastroenterol 2004;39:1106–1109. 44 Bagheri H, Fouladi A, Barange K, et al: Follow-up of adverse drug reactions from peginterferon-alfa-2b-ribavirin therapy. Pharmacotherapy 2004;24: 1546–1553. 45 Chamberlain AJ, Poon E: Cutaneous reactions to interferon and ribavirin. Intern Med J 2004;34:519. 46 Panwar P, Iznaga-Escobar N, Mishra P, et al: Radiolabeling and biological evaluation of DOTA-Ph-Al derivative conjugated to anti-EGFR antibody or egf/r3 for targeted tumor imaging and therapy. Cancer Biol Ther 2005;4. 47 Brown DL, Login IS, Borish L, Powers PL: An urticarial IgE-mediated reaction to interferon-1b. Neurology 2001; 56:1416–1417. 48 Corona T, Leon C, Ostrosky-Zeichner L: Severe anaphylaxis with recombinant interferon-. Neurology 1999;52: 425. 49 Puchner TC, Kugathasan S, Kelly KJ, Binion DG: Successful desensitization and therapeutic use of infliximab in adult and pediatric Crohn’s disease patients with prior anaphylactic reaction. Inflamm Bowel Dis 2001;7:34–37. 50 O’Connor M, Buchman A, Marshall G: Anaphylaxis-like reaction to inflix-

51

52

53

54 55

56

57

58 59

60

61

62 Prof. Dr. Werner J. Pichler Department for Rheumatology and Clinical Immunology/Allergology Inselspital, University of Bern CH–3010 Bern (Switzerland) Tel. +41 31 632 2264 or 3174, Fax +41 31 632 2747 E-Mail [email protected]

Adverse Side Effects to Biological Agents

63

imab in a patient with Crohn’s disease. Dig Dis Sci 2002;47:1323–1325. Hyams JS, Markowitz J, Wyllie R: Use of infliximab in the treatment of Crohn’s disease in children and adolescents. J Pediatr 2000;137:192–196. Soykan I, Ertan C, Ozden A: Severe anaphylactic reaction to infliximab: report of a case. Am J Gastroenterol 2000;95:2395–2396. Diamanti A, Castro M, Papadatou B, Ferretti F, Gambarara M: Severe anaphylactic reaction to infliximab in pediatric patients with Crohn’s disease. J Pediatr 2002;140:636–637. Lobel EZ, Korelitz BI, Warman JI: Red man syndrome and infliximab. J Clin Gastroenterol 2003;36:186. Stallmach A, Giese T, Schmidt C, Meuer SC, Zeuzem SS: Severe anaphylactic reaction to infliximab: successful treatment with adalimumab – report of a case. Eur J Gastroenterol Hepatol 2004;16:627–630. Youdim A, Vasiliauskas EA, Targan SR, et al: A pilot study of adalimumab in infliximab-allergic patients. Inflamm Bowel Dis 2004;10:333–338. Lelong J, Duburque C, Fournier C, et al: Desensitisation to infliximab in patients with Crohn’s disease (in French). Rev Mal Respir 2005;22:239–246. Domm A: A patient’s reaction to infliximab. Ann Allergy Asthma Immunol 2003;90:298–301. Chavez-Lopez MA, Delgado-Villafana J, Gallaga A, Huerta-Yanez G: Severe anaphylactic reaction during the second infusion of infliximab in a patient with psoriatic arthritis. Allergol Immunopathol (Madr) 2005; 33:291–292. Leonard PA, Woodside KJ, Gugliuzza KK, Sur S, Daller JA: Safe administration of a humanized murine antibody after anaphylaxis to a chimeric murine antibody. Transplantation 2002;74: 1697–1700. Baudouin V, Crusiaux A, Haddad E, et al: Anaphylactic shock caused by immunoglobulin E sensitization after retreatment with the chimeric anti-interleukin-2 receptor monoclonal antibody basiliximab. Transplantation 2003;76: 459–463. Melamed J, Stahlman JE: Rapid desensitization and rush immunotherapy to trastuzumab (Herceptin). J Allergy Clin Immunol 2002;110:813–814. Tada K, Ito Y, Hatake K, et al: Severe infusion reaction induced by trastuzumab: a case report. Breast Cancer 2003;10:167–169.

165

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 168–189

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation Werner J. Pichler Division of Allergology, Department for Rheumatology and Clinical Immunology/Allergology, Inselspital, University of Bern, Bern, Switzerland

Abstract The clinical characteristics of drug hypersensitivity reactions are very heterogeneous as drugs can actually elicit all types of immune reactions. The majority of allergic reactions involve either drug-specific IgE or T cells. Their stimulation leads to quite distinct immune responses, which are classified according to Gell and Coombs. Here, an extension of this subclassification, which considers the distinct T-cell functions and immunopathologies, is presented. These subclassifications are clinically useful, as they require different treatment and diagnostic steps. Copyright © 2007 S. Karger AG, Basel

Introduction

Drug-induced adverse reactions are common and normally classified as type A reactions, which represent predictable side effects due a pharmacological action of the drug or type B reactions, which are not predictable and comprise idiosyncratic reactions due to some individual predisposition (e.g. an enzyme defect), and hypersensitivity reactions [1]. Drug hypersensitivity reactions account for about one sixth of all adverse drug reactions. They comprise allergic and so-called pseudoallergic reactions. The latter is characterized by having the features of an aller-

gic reaction without detectable reactions of the adaptive immune system. Drug hypersensitivity reactions can become manifest in a great variety of clinical symptoms and diseases, some of which are quite severe and even fatal [2, 3]. The most common allergic reactions occur in the skin and are observed in about 2–3% of hospitalized patients [4, 5]. Any drug is assumed to be able to elicit hypersensitivity reactions. Antibiotics and antiepileptics are the drugs most frequently causing them. The risk of sensitization and the severity of clinical symptoms depend on the state of immune activation of the individual, the dose and duration of treatment, female sex and the immunogenetic predisposition (in particular HLA-B alleles), while a pharmacogenetic predisposition has rarely been detected [see chapters of Hung et al., pp 105–114, and Nolan et al., pp 95–104]. Epicutaneous application of a drug clearly increases the risk of a sensitization compared to oral or parental treatments. It may be due to the high density of dendritic cells in the skin. Atopy – defined as the genetic predisposition to mount an IgE response to inhaled or ingested innocuous proteins – is normally not associated with a higher risk of drug hypersensitivity in

general. However, an atopic predisposition may prolong the persistence of drug-specific IgE in the serum [6], and an ongoing IgE-mediated allergic inflammation may aggravate the symptoms of an IgE-mediated drug hypersensitivity reaction.

Classification of Drug Hypersensitivity Reactions

Drug hypersensitivity reactions can cause many different diseases. To account for this heterogeneity and to better explain the various clinical pictures, Gell and Coombs [7] have classified drug hypersensitivity as well as other immune reactions in four categories termed type I–IV reactions: This classification relies on formation of IgE antibodies, which bind to high-affinity IgE receptors on mast cells and basophilic leukocytes, on complement-fixing antibodies and on T-cell reactions, which orchestrate different forms of inflammations. One has to be aware that these reactions are tightly connected, as for example the maturation of B cells to IgE- or IgGproducing plasma cells depends on the help of T cells. Moreover, this classification was developped in the 1960s, before any functional heterogeneity of T cells was known. In the meantime it has become clear that the immune system is not only specific and has a kind of memory, but is also well adapted to the type of challenge it faces, as for example the defense against intracellular pathogens requires a different approach than the defense against extracellular bacteria, etc. This discrimination seems to be regulated by different types of T cells [8]. To better take into account this heterogeneity of T-cell functions, which are important to understand different forms of diseases, the classification of Gell and Coombs has recently been revised [8] as discussed below. This modified and extended Gell and Coombs classification has an impact on classifying disease severity, treatment, level of cross-reactivity

with other, structurally-related drugs, natural course and prognosis. It requires knowledge of underlying immune mechanisms, and of how these mechanisms result in different forms of clinical disease. On the other hand, one has to be aware that a classification is a simplification of complex events occurring in vivo. The immune system often combines different approaches to defend against a real or – as is the case with allergies – putative pathogen. On the other hand, in many hypersensitivity diseases a certain type of immune reaction dominates the clinical picture, even if the immune response is rather complex. In drug hypersensitivity, an additional level of complexity derives from the p-i concept (pharmacological interaction with immune receptors), as drugs may act more like superantigens [discussed in the chapter by Gerber and Pichler, pp 66–73]: • The p-i concept postulates a bypassing of the development of a normal immune response, as a direct, pharmacological stimulation of memory and effector T cells is implied; • Therefore, it does not follow the normal rules of an immune response, which may already explain some as ‘bizarre’ classified clinical features; • It could appear at the first encounter with the drug, as no sensitization is required; • Those T cells which happen to be stimulated by drugs are assumed to actually have a peptide specificity (to which they were primed); however, which peptides are recognized is unknown. It is unclear (a) whether the drug-induced stimulation of peptide-specific T cells results in a different clinical picture, if the peptides, which are recognized by the drug-stimulated T cells, are present in the body; (b) whether the function of the stimulated T cells remains the same, if the T cell is stimulated by the drug or by its natural ligand, a certain MHC-associated peptide, and (c) whether the constant presence of the presumed natural ligand, the peptide, may contribute to the

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

169

persistence of drug-allergic reactions over years in spite of strict avoidance of the drug. These questions are open and explain why the naming of type B reactions as ‘bizarre’ reactions is actually not so far fetched.

Antibody-Mediated Drug Hypersensitivity Reactions The hapten-like features of a drug allow the modification of soluble and cell-bound proteins. For example, a penicillin molecule could bind covalently to a lysine within a serum protein, but also to cell-bound proteins [see figure 2 in chapter of Torres et al., pp 190–203]. The ‘normal’ reaction of the immune system to such modified proteins is the development of a humoral immune response, consisting of many distinct antibodies with hapten specificity. Consequently, if a coordinated humoral immune response develops (based on T-cell help), one may conclude that the eliciting drug has hapten-like features forming hapten-carrier complexes or is itself a protein bearing ‘foreign’ determinants [e.g. a chimeric antibody, see figure 1 in chapter of Pichler and Campi, pp 151–165]. Indeed, as shown in table 1, the majority of drugs able to elicit IgE-mediated allergies are known to be haptens, or they contain foreign antigenic structures (fig. 1) [9]. Type I (IgE-Mediated) Allergies The IgE system is geared to react to small amounts of antigens. It achieves this extraordinary sensitivity by the ubiquitous presence of mast cells armed with high-affinity Fc-IgE receptors (FcIgE-RI), to which allergen/drug-specific IgE is bound. Very small amounts of a drug are apparently sufficient to interact and stimulate these receptor-bound IgE molecules, as occasionally even skin tests with drugs can elicit systemic reactions [see contribution of Barbaud, pp 366–379]. Upon cross-linking the Fc-IgE-RI, various mediators (histamine, tryptase, leukotrienes, prostaglan-

170

Table 1. IgE-mediated drug allergies Up to 50% of patients with IgE-induced anaphylaxis to certain drugs have no history of previous drug exposure! Anaphylaxis occurs rapidly (<20 min), seldom later Asphyxia is probably the main cause of lethal anaphylaxis Cardiac arrest can be the sole symptom of an anaphylaxis (in particular in perioperative anaphylaxis) In certain cases, desensitization procedures are possible and may allow reuse of the drug

dins, TNF-
, etc.) are released, which cause the immediate symptoms and may start and facilitate late allergic reactions. IgE-mediated reactions to drugs are usually thought to depend on the prior development of an immune response to a hapten/carrier complex: B cells need to mature into IgE-secreting plasma cells, and T cells help in this process by interacting with B cells (i.e. CD40-CD40L interaction) and by releasing IL-4/IL-13, which are switch factors for IgE synthesis. This sensitization phase is asymptomatic and may have occurred during an earlier drug treatment. Upon renewed contact with the drug, a haptencarrier complex is formed again, which then cross-links preformed drug-specific IgE on mast cells. The drug itself is normally too small to cross-link two adjacent IgE molecules, and needs to bind to proteins to cross-link [see chapter of Torres et al., pp 190–203]. However, it is to be noted that in ca. 50% of patients with immediate reactions, no prior contact with the drug can be elucidated [see contribution of Christiansen, pp 233–241] and that 80% of the patients with lethal anaphylaxis had no prior contact with the drug [10]. This indicates that either (a) a silent sensitization to a cross-reactive compound had occurred, or (b) preformed IgE existed which happened to react with the small

Pichler

O

Hapten

C H2

Penicillin G H H N S b CH3 N

1

CH3

Oa

2

ONa O

3

Processing

Hapten (penicillin G) Binding to 1: soluble proteins or 2: membrane-bound proteins or 3: the MHC-peptide complexes (I und II) directly binding (a) via -lactam ring-forming penicilloyl (PPL-PLL) or (b) via thiaolidin structure

Clinic: ‘Everything’ 1 and 2: binding to cell-bound and soluble proteins r IgE or IgG to hapten-protein: anaphylaxis, hemolytic anemia, thrombocytopenia 3: MHC class I and II modification: T-cell reaction with exanthem, hepatitis, interstitial lung disease, contact dermatitis, AGEP, TEN...

Metabolism-dependent hapten formation (e.g. sulfamethoxazole, SMX) Uptake of the non-hapten drug SMX in cells able to metabolize it, generation of a hapten (SMX-NO), which can bind to intracellular proteins: presentation of processed modified peptides and binding to extracellular soluble proteins (r both T- and B-cell responses might develop): the metabolism may also induce co-stimulatory molecules on antigen-presenting cells

Clinic: ‘Everything’ Potentially immunogenic for B and T cells; Immunogenicity and clinical manifestation might be restricted to the liver (hepatitis!) or kidney (interstitial nephritis!), where metabolism occurs

p-i concept The drug happens to fit into some TCR (1) with sufficient affinity to cause a signal. This drug-TCR interaction is supplemented by MHC interaction (2); the T cells react and proliferate. No metabolism of drugs required. The reacting T cells are probably preactivated and have an additional peptide specificty

Clinic: Only T cells An exclusive T-cell response might develop with exanthems, hepatitis, etc. Whether B cells (by drug binding to Ig) can similarily be stimulated, remains unclear

3

a Prohapten NH2or NO

SMX (-NO)

R NH

R NH

R NH

SO

SO

SO

NHOH

NO

N-S H

O S O NH

N O H3C

b p-i concept

SMX

NH2

1 MH C c l a ss I

SMX O S O N O

c

NH

T cell MH C class II

2

H3C

Fig. 1. Hapten and prohapten concept and the non-covalent drug presentation to T cells. a Haptens: Drugs are haptens if they can bind covalently to molecules, be they soluble or cell bound (e.g. penicillin G). They can even bind directly to the immunogenic major histocompatibility complex (MHC)/peptide complex on antigen-presenting cells (APC), either to the embedded peptide or to the MHC molecule itself. Thus, the chemical reactivity of haptens leads to the formation of many distinct antigenic epitopes, which can elicit both humoral and cellular immune responses. Some examples of a B- or T-cellmediated immune response are listed on the right side. b Prohaptens: Other drugs are ‘prohaptens’, requiring metabolic activation to become haptens (= chemically reactive). The metabolism leads to the formation of a

chemically reactive compound (e.g. from sulfamethoxazole (SMX) to the chemically reactive form SMX-NO). The resulting intake may lead to modification of cell-bound or soluble proteins by the chemically reactive metabolite, similar to a real hapten. c The p-i concept (pharmacological interaction with immune receptors): Drugs are often designed to fit into certain proteins/enzymes to block their function. Some drugs may happen to bind also into some of the available T-cell receptors. Under certain conditions (see text), this drug–T-cell receptor interaction may lead to an immune response of the T cell with a ‘fitting’ T-cell receptor. For a full T-cell stimulation by such an inert drug, an interaction of the T-cell receptor with the MHC molecule is required. This type of drug stimulation results in an exclusive T-cell stimulation.

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

171

molecular compound, or (c) that the reaction was pseudoallergic (see below). These reactions were erroneously considered to be dose-independent, as sometimes very small amounts can already cause severe reactions. But further diminishing the dose – as is done in desensitization procedures – illustrates that also these reactions are clearly dose-dependent. IgE-mediated reactions can cause mild to very severe, even lethal, diseases. In sensitized individuals the reaction can start within seconds after contact with the parentally applied drug, and minutes after oral drug uptake. Symptoms reach from simple local itch, local wheal-andflare reaction upon parenteral drug application, to acute bronchospasm and generalized urticaria and edema, preferentially periorbital, perioral or genital. More severe and complex reactions are called anaphylaxis, whereby in most cases with anaphylaxis some circulatory events with collapse and (transient) unconsciousness are observed together with a generalized redness, itch or urticaria. Anaphylactic shock occurs often within 15 min, and asphyxia due to laryngeal edema often between 15 and 60 min. Starting symptoms may be a palmar, plantar, genital or axillar itch, which should be seen as an alarm sign, as it often heralds a possibly severe, anaphylactic reaction, following rapidly within minutes (fig. 2): the skin becomes red (diffuse erythema), often first at the trunk, later over the whole body. In the next
30–60 min an urticaria may appear, together with swelling of the periorbital, perioral and sometimes genital area. Asphyxia may account for 60% of anapylaxis-related deaths [11]: laryngeal swelling may be suspected if the voice becomes hoarse, the patients have difficulty speaking and swallowing as the tongue is swollen. Patients may also complain about chest tightness and dyspnea – signs of acute bronchospasm. Some patients develop gastrointestinal symptoms (nausea, cramps, vomiting and fecal incontinence). The blood pressure may collapse – ei-

172

ther just due to a shift of volume into the extravascular space or – and more severely – if a cardiac arrhythmia develops. The full syndrome is anaphylactic shock, which is lethal in ca. 1–2% of all anaphylaxis cases and the more dangerous the more rapid it appears! Risk factors for a severe course are high dose, preexisting (undertreated) asthma and older age, as myocardial infarction or cerebral hypoxia and brain damage can lead to death days after the acute event. In perioperative anaphylaxis, the symptoms may affect initially only one organ system (e.g. the cardiovascular system with arrhythmia), and skin symptoms appear later. Anaphylaxis is a severe event, and survivors do not rarely have some cognitive or intellectual impairment. Table 2 summarizes the main drugs causing anaphylaxis. Most IgE-mediated reactions to drugs are less severe – and often only an urticaria, angioedema or a local wheal may develop. However, any IgEmediated drug allergy can be potentially lifethreatening, as the mild symptoms might be due to a relatively low dose, and each treatment might also boost the drug-specific IgE response. Pseudoallergy (Non-Immune-Mediated Hypersensitivity) An unsolved problem are so-called ‘pseudoallergic’ reactions (non-immune-mediated hypersensitivities), which in fact are as frequent as true IgE-mediated reactions. Detection of a specific immune mechanism is negative. The majority of these reactions imitate the clinical features of milder immediate reactions (erythema, urticaria), but some reactions cause anaphylaxis and can be lethal. For non-steroidal anti-inflammatory drug (NSAID)-induced pseudoallergic reactions, it seems that they tend to arise less rapidly (often 115 min) than true IgE-mediated allergies and they may require higher drug doses than for true IgE-mediated reactions. Elevated tryptase levels in the acute stage underline the role of mast cell degranulation, at least in some of these reactions.

Pichler

Multiorgan involvement + Respiratory tract

Skin

+ GI tract

Anaphylaxis

Death

Arrhythmia, cerebral hypoxia, bronchospasm, laryngeal edema, generalized urticaria, collapse

Skin, respiratory and GI symptoms and collapse

Skin symptoms and collapse

+ Laryngeal edema

+ Bronchospasm

(Often facial) local edema

Generalized urticaria

Generalized itch

Symptoms

Local wheal-and-flare

Severity

Anaphylactic shock

Fig. 2. Anaphylaxis is a complex and severe allergic reaction affecting multiple organs. It can be seen as the summary of different, severe allergic manifestations in various organs. In many, but not all cases, the cardiovascular system is involved either as result of relative hypo-

volemia, due to a shift of intravascular volume to extravascular space (less dangerous), or due to arrhythmia. Its most severe form is anaphylactic shock, which can result in death in 1–2% of affected persons (for details, see text).

‘Pseudoallergic’ reactions can be elicited by many drugs, but some drugs seem to elicit them more often (table 2). Some drugs might elicit both pseudoallergic and also presumably real allergic reactions, as positive prick skin tests can be detected in very severe reactions (contrast media, neuromuscular-blocking agents). In vitro, these drugs do not release mediators from basophils. A certain group of people seems to show a higher susceptibility to react in this way, as they develop similar, mostly mild symptoms to a quite heterogeneous panel of drugs, with clearly dis-

tinct chemical and pharmacological features. Neither IgE nor T-cell reactions can be demonstrated, the reactions are recurrent, but provocation tests are often negative, suggesting that additional cofactors might be needed to develop clinical symptoms. A few patients may have constantly elevated tryptase levels as a sign of a mastocytosis. Such patients sometimes report about multiple anaphylactic reactions to various triggers (food, drugs, hymenoptera stings, etc.). Some milder reactions can be suppressed by pretreatment with antihistamines, but it is unclear wheth-

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

173

Table 2. Drugs causing IgE-mediated or pseudoallergic reactions Drugs involved in IgE-mediated allergies1

Drugs causing ‘pseudoallergic’ reactions1

Foreign proteins (chimeric antibodies) Immunglobulin preparations (IgE anti-IgA) -Lactam antibiotics Penicillin Cephalosporin Pyrazolones Quinolones Neuromuscular-blocking agents2

(Radio)contrast media Plasma expanders NSAIDs: acetylsalicylic acid, diclofenac, mefanamic acid, ibuprofen, … Pyrazolones Quinolones Neuromuscular-blocking agents

1

Not complete; only main groups mentioned. The main causes of perioperative anaphylaxis are neuromuscular-blocking agents, followed by antibiotics (mainly i.v. cephalosporins) and latex. Skin tests may be positive even if the drug was given the first time.

2

er pretreatment with antihistamines and corticosteroids can prevent more severe reactions, e.g. to contrast media [12]. The most common form of such reactions is related to NSAIDs and is discussed in detail in the contributon by Sczecklik et al. [pp 340–349]. IgG-Mediated Reactions (Cytotoxic Mechanism, Type II) Type II and III reactions rely on the formation of complement-fixing IgG antibodies (IgG1, IgG3). Occasionally, IgM is involved. They are similar, as both depend on the formation of immune complexes and interaction with complement and Fc-IgG receptor (Fc-IgGI, IIa & IIIa) bearing cells (on macrophages, NK cells, granulocytes, platelets), but the target structures and physiological consequences are different: In type II reactions, the antibody can be directed to cell structures on the membrane (rarely) or immune complex activation occurs on the cell surface: both events can lead to cell destruction or sequestration. Affected target cells include erythrocytes, leukocytes, platelets and probably hematopoietic precursor cells in the bone marrow [see also chapter of Aster, pp 306– 320]. The mechanism of type II reaction is not as

174

clear as originally thought, since a clear haptenspecific immune reaction can often not be documented [13, 14]. One differentiates between: (a) Development of complement-fixing antibodies (IgG1, IgG3, rarely IgM) to the haptencarrier complex, mostly after longer duration of high-dose treatment. It is rather rare and best documented for high-dose penicillin and cephalosporin treatments. Some antibody reactivity may be directed to the carrier molecule itself (= autoantibodies). This autoimmune form is less abrupt, but longer lasting (weeks instead of days) after cessation of the drug. (b) Non-specific adherence with autoantibody inductions can occur when a drug or metabolite is somehow adsorbed to the erythrocyte or thrombocyte membrane, creating a new antigenic complex in combination with the cell membrane. For example, quinine-induced immune thrombocytopenia is caused by a remarkable class of IgG and/or IgM immunoglobulins that react with selected epitopes on platelet membrane glycoproteins, usually GPIIb/IIIa (fibrinogen receptor) or GPIb/IX (von Willebrand factor receptor) only when the drug is present in its soluble form [13]. Well-documented cases are due to quinine, quinidine or sulfonamide antibiotics. The

Pichler

antibodies are clearly not hapten-specific, and it remains enigmatic how a soluble drug can promote binding of an otherwise innocuous antibody to a membrane glycoprotein to cause platelet destruction. The antibody-coated cells will be sequestrated to the reticuloendothelial system in liver and spleen by Fc or complement receptor binding. More rarely, intravascular destruction may occur by complement-mediated lysis. Hemolytic anemia has been attributed to penicillin and its derivatives, cephalosporins, levodopa, methyldopa, quinidine and some antiinflammatory drugs. Today, cephalosporins are the main cause. The clinical symptoms of hemolytic anemia are insidious and may be restricted to symptoms of anemia (fatigue, paleness, shortness of breath, tachycardia) and jaundice with dark urine. The laboratory investigation may reveal reduced erythrocyte and hemoglobin levels, increased reticulocytes, a positive direct and – if the drug is present during the test – indirect Coombs test. Indirect bilirubin levels are elevated, haptoglobulin decreased. In urine, hemoglobin and hemosiderin are increased. Thrombocytopenia is a relatively common side effect of drug treatment: Acute, sometimes severe and life-threatening thrombocytopenia is a recognized complication of treatment with quinine, quinidine, sulfonamide antibiotics and many other medications. It is a not so rare complication of treatment with biologicals. Drug-induced immune thrombocytopenia usually develops after 5–8 days of exposure to the sensitizing medication, or after a single exposure in a patient exposed previously to the same drug. Patients with this condition often present with widespread petechial hemorrhages in the skin and buccal mucosa, sometimes accompanied by urinary tract or gastrointestinal bleeding. Intracranial hemorrhage is rare, but examples have been reported. After discontinuation of the provocative medication, platelet counts usually return to normal within 3–5 days.

A special, intermediate form between type II and III reactions is heparin-induced thrombocytopenia: Platelets contain low-affinity (Fc-IgGRIIa) Fc receptors that are capable of binding immune complexes, leading to platelet activation [15]. Heparin is a high molecular weight, sulfated, linear polysaccharide that inhibits blood coagulation by activating regulatory proteins such as antithrombin III. About 50% of patients anticoagulated with heparin for at least 7 days produce antibodies that recognize complexes consisting of heparin and platelet factor 4, a CXC chemokine normally stored in platelet
-granules. When a patient with such an antibody is given heparin, heparin/PF4 complexes are formed that react with antibodies to form immune complexes, which bind to the platelet Fc-IgG-RIIa receptors, leading to platelet activation, additional PF4 release and eventually, platelet destruction. Thrombocytopenia occurs in about 5% of patients given heparin and is rarely severe enough to cause bleeding. However, about 10% of the affected patients experience paradoxical thrombosis which can be life-threatening. These thrombopenic reactions due to heparin have to be differentiated from other heparin- or heparinoid-induced immune reactions. They may be due to a T-cell-mediated allergy with exanthematous or eczematous skin reactions, and very rarely anaphylaxis due to heparin-specific IgE has been described [16]. In delayed reactions due to subcutaneous applications, intravenous administration of heparin or subcutaneous applications of structurally different anticoagulants like for example fondaparinux of hirudin may be tolerated. IgG-Mediated Reactions (Immune Complex Deposition, Type III) Formation of immune complexes is a common event in the frame of a normal immune response and does not normally cause symptoms. Immune complexes may also be formed during drug treatment: either if the drug forms a hapten-carrier

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

175

complex and thus gives rise to an immune reaction, or if the drug is a (partly) foreign protein, which elicits an immune reaction itself (e.g. a chimeric antibody). Such immune complexes will normally be rapidly cleared, either by Fc-IgG-RI or CR1 binding on reticuloendothelial cells. No symptoms arise, but the efficiency of treatment decreases. Why under certain circumstances an immune complex disease develops is not clear: very high immune complex levels, a relative deficiency of some complement components, and thus lower capacity to eliminate immune complexes or an aberrant Fc-IgG-R function might be considered. Recently, a low copy number of Fc-IgG-RIII genes could be associated with another immune complex disease, namely glomerulonephritis [17]. Thus, reduced removal of immune complexes may lead to inappropriate deposition of immune complexes and recruitment of inflammatory cells, in particular PMNs due to immune complex binding to Fc-IgG-R on PMN. In addition, anaphylatoxins C3a and C5a, generated due to local complement activation, may attract PMNs. The clinical symptoms of a type III reaction may be hypersensitivity, small vessel vasculitis and/or serum sickness. Serum sickness was first described with the use of heterologous or foreign serum for passive immunizations: Antibodies are generated within 4–10 days, which react with the antigen, forming soluble circulating immune complexes. Complement (C1q)-containing immune complexes are deposited in the postcapillary venules and are attracting neutrophilic leukocytes by interacting with their Fc-IgG-RIII [18], which thereby release proteolytic enzymes that can mediate tissue damage. Currently, non-protein drugs are the most common cause of serum sickness. Hypersensitivity vasculitis reportedly has an incidence of 10– 30 cases per million people per year. Most reports concern cefaclor, followed by trimethoprim-sulfamethoxazole, cephalexin, amoxicillin, NSAIDs and diuretics.

176

Vasculitis may be localized mainly to the skin as ‘palpable purpura’ – purplish, red spots, usually found on the legs [see figure 5 on hypersensitivity vasculitis in chapter of Bircher, pp 352– 365]. In children, it is often referred to as HenochSchönlein purpura, sometimes appearing together with an arthritis. Lesions may coalesce to form plaques and they may ulcerate in some instances. The internal organs most commonly affected are the gastrointestinal tract, kidneys and joints. The prognosis is good when no internal involvement is present. The disorder may be acute or chronic. Histology can reveal IgA-containing immune complexes, and the histology of kidney lesions is in fact identical to IgA nephropathy, a main cause of chronic renal failure.

T-Cell-Mediated, Delayed Drug Hypersensitivity Reactions Subclassification of Type IV Reactions The old Gell and Coombs classification was established before a detailed analysis of T-cell subsets and functions was available. In the meantime, immunological research has revealed that the three antibody-dependent types of reactions require an involvement of helper T cells. Moreover, T cells can orchestrate different forms of inflammations. Therefore, T-cell-meditated type IV reactions were further subclassified in IVaIVd reactions as proposed in figure 3 [8]. This subclassification considers the distinct cytokine production by T cells and thus incorporates the well-accepted Th1/Th2 distinction of T cells; it includes the cytotoxic activity of both CD4 and CD8 T cells (IVc), and it emphasizes the participation of different effector cells like monocytes (IVa), eosinophils (IVb), or neutrophils (IVd), which are the cells causing the inflammation and tissue damage (fig. 4): Type IVa reactions correspond to Th1-type immune reactions: Th1-type T cells activate macrophages by secreting large amounts of interfer-

Pichler

Fig. 3. Urticaria with itching wheals (pseudoallergic reaction).

on (IFN)-, drive the production of complementfixing antibody isotypes involved in type II and III reactions (IgG1, IgG3), and are co-stimulatory for pro-inflammatory responses (TNF, IL-12) and CD8+ T-cell responses. The T cells promote these reactions by IFN- secretion and possibly other cytokines (TNF-
, IL-18, etc.). An in vivo correlate would on the one hand be a monocyte activation, for example in skin tests to tuberculin or even granuloma formation as seen in sarcoidosis. On the other hand, these Th1 cells are known to help in the activation of CD8 cells, which might explain the common combination of IVa and IVc reactions (e.g. in contact dermatitis). Type IVb corresponds to the Th2-type immune response. Th2 T cells secrete the cytokines IL-4, IL-13 and IL-5, which promote B-cell production of IgE and IgG4, macrophage deactivation and mast cell and eosinophil responses. The high production of the Th2 cytokine IL-5 leads to an eosinophilic inflammation, which is the characteristic inflammatory cell type in many drug hypersensitivity reactions [8]. In addition, there is a link to type I reactions, as Th2 cells boost IgE production by IL-4/IL-13 secretion. An in vivo

correlate might be an eosinophil-rich maculopapular exanthem, but also infestations with nematodes, or an allergic inflammation of the bronchi or nasal mucosa (asthma and rhinitis). Type IVc: T cells can also act as effector cells themselves: they emigrate to the tissue and can kill tissue cells like hepatocytes or keratinocytes in a perforin/granzyme B- and FasL-dependent manner (fig. 5) [19, 20]. Such reactions occur in most drug-induced delayed hypersensitivity reactions, mostly together with other type IV reactions (monocyte, eosinophil or PMN recruitment and activation). Cytotoxic T cells thus play a role in maculopapular or bullous skin diseases as well as in neutrophilic inflammations (acute generalized exanthematous pustulosis – AGEP), and in contact dermatitis. Type IVc reactions appear to be dominant in bullous skin reactions, where activated CD8+ T cells kill keratinocytes [8, 19, 20], but may also be the dominant cell type in hepatitis or nephritis. Type IVd: Rather neglected was the possibility that T cells could coordinate (sterile) neutrophilic inflammations as well. Typical examples would be sterile neutrophilic inflammations of the skin, in particular AGEP. In this disease, CXCL8- and GM-CSF-producing T cells recruit neutrophilic leukocytes via CXCL8 release and prevent their apoptosis via GM-CSF release [21]. Besides AGEP, such T-cell reactions are also found in Behçet’s disease and pustular psoriasis [22]. Tolerance mechanism: Most patients can take drugs without developing immune-mediated side effects. One could argue that they lack precursor cells able to interact with the drug, but the great heterogeneity of the immune response to drugs [8], a high precursor frequency in sensitized patients [23] and the finding that 2–4% of the normal population, but that 30 to 150% of HIV-infected patients may react with sulfamethoxazole, suggests that not a lack of precursor cells but other factors like the underlying immune status (preactivation of memory T cells) and ‘regulatory’ mechanisms may be important.

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

177

Antibody (I–III) and T- cell orchestrated hypersensitivity reactions (IV a–d)

Ty p e I

I mm un e reactant

IgE

I gG

IgG

Antigen

Soluble antigen

Cell- or matrixas s o c ia t e d antigen

Soluble antigen

Mast cell activation

FcR+ cells (phagocytes, NK cells)

FcR+ cells Complement

Effector

Type IVa

Ty pe III

Typ e II

Type IVb

Type IVc

Type IVd

IL-5, IL-4/IL-13 (TH2 cells)

Perforin/ granzyme B (CTL)

CXCL8, G M -C S F (T cells)

Antigen presented by cells o r d i rect T - c el l stimulation

Antigen presented by cells o r d i re ct T - c e l l stimulation

Cell-associated antigen or direct Tcell stimulation

Soluble antigen presented by cells o r d i r ec t T - c e l l stimulation

Macrophage activation

Eosinophils

T cells

N e utr o p h i ls

IIFN-␥, TNF-␣ (TH1 cells)

Immune complex B lood vessel IFN-γ

Platelets

TH2 TH1

Ag

IL-4 IL-5

Eotaxin

CTL

CXCL8 GM-CSF

PMN

Eosinophil

Chemokines, cytokines, Cytokines, inflammatory mediators cytotoxins

Example of hypersensitivity reaction

Allergic rhinitis, asthma, systemic anaphylaxis

H em o l y t i c a n em i a, t hromb ocytopenia (e.g., penicillin)

Serum sickness, Arthus reaction

Fig. 4. Revised Gell and Coombs classification of drug reactions. Drugs can elicit all types of immune reactions. In fact, all reactions are T-cell regulated, but the effector function rely mainly on antibody-mediated effector functions (type I–III) or more T-cell/cytokine-dependent functions (type IVa–IVd) [8]. Type I reactions are IgE-mediated. Cross-linking IgE molecules on high-affinity IgE receptors (Fc-IgE-RI) on mast cells and basophilic leukocytes lead to degranulation and release of mediators, which cause a variety of symptoms (vasodilatation, increased permeability, bronchoconstriction, itch, etc.). Type II reactions are IgG-mediated, and cause cell destruction due to complement activation or interaction with Fc-IgG receptor-bearing killer cells. Type III reactions are also IgG-mediated: complement deposition and activation in small vessels and recruitment of neutrophilic granulocytes via Fc-IgG receptor interaction

178

Tuberculin reaction, contact dermatitis (w ith IVc )

Chronic asthma, chronic allergic rhinitis Maculopapular exanthema w ith eosinophilia

Cytokines, inflammatory mediators

Contact dermatitis Maculopapular and bullous exanthema Hepatitis

AGEP Behçet's disease

leads to a local vascular inflammation. Type IVa corresponds to Th1 reactions with high IFN- /TNF-
secretion and involves monocyte/macrophage activation. Often, one can also see a CD8 cell recruitment (type IVc reaction). Type IVb reactions correspond to eosinophilic inflammation and to a Th2 response with high IL-4/IL-5/ IL-13 secretion; they are often associated with an IgE-mediated type I reaction. Type IVc: the cytotoxic reactions (both by CD4 and CD8 cells) rely on cytotoxic T cells themselves as effector cells (type IVc). They seem to occur in all drug-related delayed hypersensitivity reactions. Type IVd correspond to a T-cell-dependent, sterile neutrophilic inflammatory reaction. It is clearly distinct from the rapid influx of PMN in bacterial infections. It seems to be related to high CXCL8/GM-CSF production by T cells (and tissue cells).

Pichler

Keratinocyte

Fig. 5. Cytotoxic mechanism. a Typical histology of a maculopapular exanthem: note the focal keratinocyte necrosis, often in close apposition of T cells (which have cytotoxic potential and are perforin-positive). b Schematic representation of CD4mediated killing through perforin/ granzyme B release. Activated keratinocytes, which express MHC class II and ICAM1 molecules, are killed which results in spongiosis, vesicle formation and even bullae.

ICD541 MHC II

Perforin granzyme B TCR LFA-1

a

Thereby ‘regulation’ may occur on different levels and may be different for drugs stimulating the immune system via hapten and the p-i concept (see table 5). Clinical Symptoms The most frequent manifestations of drug allergies are delayed appearing cutaneous reactions, so-called ‘rashes’. They comprise a broad spectrum of clinical and distinct histopathological features which appear 16 h to 10 days after drug intake (fig. 6). It is not clear why the skin is affected so often during drug hypersensitivity. On one hand, one has to be aware that the skin is an important defense region, and anything penetrating it may be considered to be harmful. Indeed, the skin contains a surprisingly high proportion of dendritic cells and lymphocytes, some of which have a sentinel function [24]. Thus, the immune system of the skin is constantly ready to react to possible intruders. This is in contrast to the gastrointestinal tract, where many foreign proteins, which are taken up, are not dangerous and a defense mechanism against them would be contraproductive. Therefore, potent tolerance systems prevail in the gastrointestinal tract.

b

The activation threshold of the skin ‘resident’ lymphocytes (CXCR8+, responsive to CCL1) might be rather low: this might explain on the one hand the higher rate of sensitization if a drug is applied topically to the skin then via the oral or parental route. The same argument is also relevant for the frequent manifestations of drug hypersensitivity in the skin, if the drug reaches the dermis and epidermis via the bloodstream, as is the case for most drugs (fig. 6). On the other hand, albeit drug hypersensitivity reactions are observed often on the skin, they are clearly systemic reactions, and other organs like the liver may actually be affected quite frequently [25], but most of these systemic reactions occur only transiently, are mostly of mild nature, asymptomatic, and overlooked if one does not screen for them. In all forms of delayed drug-induced reactions like exanthems, nephritis and probably hepatitis, cytotoxic mechanisms seem to play an important role. The clinical picture is determined by the strengths of cytotoxicity (amount of drug-specific cytotoxic T cells and tissue destruction), the generation of cytotoxic CD8 versus CD4 cells, and the type of associated effector mechanism

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

179

Hapten/prohapten

p-i concept

From outside (localized contact dermatitis)

From circulation (e.g. generalized maculopapular exanthema)

Metabolism (oxidation) to reactive compounds in skin by keratinocytes (KC) and dendritic cells (DC)

Ubiquitous distribution of drug including the skin

→ Stimulation of DC by reactive compounds (CD40↑, TNF- α ↑, etc.) and consequently activation of KC = local danger signals

Interaction of drug with resident Tsentinel cells (CCR8+)

→ Formation of hapten-protein complexes

→ Activation of locally residing sentinel T cells by drugs, which secrete TNF-α, IFN-γ, CXCL8, etc. leading to KC activation, integrin upregulation on endothelial cells

Emigration of activated DC with hapten-protein complexes to local lymph nodes (LN), processing and presentation to naive T cells

Direct drug-mediated activation of Teff in lymph nodes

Expansion of drug/hapten-specific T cells in LN, induction of Teff

Emigration of Teff (CLA+, CCR6+) to circulation and homing to skin – where these Teff cells organize local inflammation (cytotoxicity, IFN- etc.) leading to exanthem

Fig. 6. Possible role of skin ‘resident’ sentinel T cells in the initiation of drug hypersensitivity manifestations in the skin. If the drug is a hapten or prohapten – as is often the case for contact sensitizers – the formation of a new antigenic compound occurs first, with stimulation of dendritic cells (left-side panel). If the drug is inert and comes to the skin from the circulation, mainly preactivated, memory T cells residing in the skin as sentinel cells may be stimulated via the p-i-mechanism. Drugs are rapidly distributed throughout the body and are found

180

within 30–45 min in the skin in rather high concentrations. There they may bind to the TCR of sentinel T cells in the skin, which are ready to be activated and thus drug binding to their TCR is a sufficient signal to stimulate them. This results in cytokine release and activation of local keratinocytes and dendritic cells. Circulating T cells, previously also stimulated by the drug in the lymph nodes, are recruited to the skin (for details, see text), as a danger signal is present there through the local activation of keratinocytes.

Pichler

45 40

Symptoms

35 30 25 20 15 10 5 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

Days

Fig. 7. Kinetic of T-cell-mediated reactions. Appearance of exanthema in gemifloxacin-treated healthy volunteers: the biggest study ever to investigate drug-induced side effects was performed with gemifloxacin, a quinolone. Initial data revealed that rashes appeared more commonly in women !40 years of age and in those patients submitted to longer treatment (17 days). A prospective study was performed in 987 healthy women aged 18–40 years: 790 (80%) were treated with gemi-

floxacin for 10 days while the remaining 197 (20%) received ciprofloxacin for 10 days. A mostly very mild exanthem appeared in 260/790 (31.7%) gemifloxacintreated women but only in 7/197 (4.3%) ciprofloxacintreated women, with a clear peak on days 8–10 after treatment onset (see presentation of Shear N in http:// www.fda.gov/ohrms/dockets/ac/03/slides/3931S1_ 04_LFLife%20Sciences-Factive.pdf). Aminopenicillin-induced exanthems have a similar time kinetic.

(monocyte or activation of eosinophils and neutrophilic granulocytes). Maculopapular exanthem (MPE) is the most frequent drug hypersensitivity reaction, affecting 2–8% of hospitalized patients, especially after treatment with -lactams, sulfamethoxazole, quinolones, diuretics, and many more [4, 5] (see fig. 8). It appears mainly 8–11 days after starting treatment (fig. 7) – sometimes even 1–2 days after stopping treatment. Its clinical features are described in detail in the contribution of Yawalkar [pp 242–250]. The biggest study ever to investigate the MPE was performed with gemifloxacin, a quinolone: initial data with this drug have revealed that ‘rashes’ appeared more commonly in women !40 years of age and in those patients submitted to longer treatment (17 days). A prospective study was performed in 987 healthy women aged 18–

40: 790 (80%) were treated with gemifloxacin for 10 days while the remaining 197 (20%) received ciprofloxacin for 10 days. A mostly very mild exanthem appeared in 260/790 (31.7%) gemifloxacin-treated women, with a clear peak on days 8– 10 after treatment onset (fig. 7). This time kinetic is very similar to aminopenicillin-induced exanthems. The structurally closely related drug ciprofloxacin induced an exanthem in only 7/197 (4.3%) ciprofloxacin-treated women. The study has not been published, but the most relevant data can be accessed under http://www.fda.gov/ ohrms/dockets/ac/03/slides/3931S1_04_LF Life%20Sciences-Factive.pdf, and are reviewed in Schmid et al. [26]. What can be learned from this study? (a) The reaction was clearly dose-dependent. A higher dose, achieved by longer treatment, induced the reaction.

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

181

(b) Females in the childbearing age had a higher risk of exanthem, suggesting an influence of estrogens on the clinical manifestation. (c) Symptoms were rather mild, and no eosinophilia or liver damage was observed in these reactions. (d) Other data with ciprofloxacin allergy suggest a wide cross-reactivity with other quinolones [27]. However, when the women with exanthem were re-exposed to ciprofloxacin (or placebo) to evaluate cross-reactivity between gemifloxacin and ciprofloxacin, only 10.4% with previous reactions to gemifloxacin reacted to a re-exposure to ciprofloxacin, suggesting that a mild exanthema is often not sufficient to induce a strong memory response with broad cross-reactivity. Crossreactivity to a related compound may thus depend on the strength of the original reaction. (e) Rather striking was the dramatic difference of MPE induction between gemifloxacin and the structurally closely related ciprofloxacin. Thus, moderate modification of the molecule can dramatically affect its ‘immunogenic’ potential – causing a rather frequent, but moderate reaction (e.g. gemifloxacin), or, as illustrated with other quinolones, a rare, but sometimes deadly reaction: temafloxacin, trovafloxacin and tosufloxacin, which all have a difluorophenyl substitution at the 1-position, were withdrawn, suspended or associated with warnings because they caused severe, presumably immune-mediated side effects in selected organs (hemolytic uremic syndrome, hepatotoxic reactions and nephritis, respectively) [26]. The MPE often heals with desquamation within 2–10 days after stopping the incriminated drug. Indeed, desquamation is a typical sign of T-cell infiltration into dermis and epidermis. MPEs – particularly if caused by -lactams or gemifloxacin – are rather mild diseases, and treatment with an emollient cream, and possibly topical corticosteroids, or systemic antihistamines against the pruritus is sufficient. Some patients can be treated through without aggrava-

182

tion. That a severe bullous exanthem develops from a gemifloxacin- or aminopenicillin-induced MPE is rather unlikely, as the cells involved are different (CD4 in MPE, CD8 in SJS/TEN) [compare chapter of Yawalkar, pp 242–250]. On the other hand, certain drugs may induce an MPE with a more mixed CD4/CD8 cell activation (table 3). In such cases a prolonged treatment may lead to the appearance of danger signs [see chapter of Bircher, pp 352–365] with confluence of the papules, the patient may complain about malaise and fever, and liver function tests indicate a hepatitis (ALAT/ASAT 13-fold increased). In the blood an eosinophilia (1 0.5 g/l) and activated CD8 cells are found [25]. This type of reaction might be similar or is already a drug rash with eosinophilia and systemic symptoms (DRESS/DiHS, discussed below). It illustrates that MPE can also be the manifestation of a more severe systemic drug hypersensitivity disease, and that the danger associated with a MPE may be related to the drug used. Acute Generalized Exanthematous Pustulosis AGEP is a rare disease (ca. 1:100,000 treatments) with an estimated incidence equal to severe bullous skin diseases combined [21, 28]. It is caused by drugs in 190% of the cases (table 3). Its clinical hallmark is the presence of myriads of disseminated, sterile pustules in the skin (fig. 8), which appear rather rapidly, namely 3–5 days after starting treatment. The patients have fever and massive leukocytosis in the blood, sometimes with eosinophilia. Epicutaneous patch test reactions can cause a similar pustular reaction locally. Details are found in the contribution of Allanore and Roujeau [pp 267–277]. Immunohistology of the acute lesion reveals intraepidermal pustules, which are filled with neutrophilic granulocytes and are surrounded by activated, HLA-DR-expressing CD4+ and CD8+ T cells. Keratinocytes show an elevated expression of the neutrophil attracting chemokine IL-8 (CXCL8), and even the T cells migrat-

Pichler

Table 3. Similarities of DiHS/DRESS- and TGN1412-induced cytokine storm Symptoms

DiHS/DRESS1

TGN1412

Fever Eruption Lymphadenopathy Liver (ASAT/ALATd) Pneumopathy with eosinophilia Kidney involvement (acute renal failure, proteinuria, interstitial nephritis) Myalgia Neurologic findings (headache, hypotension) Back pain Eosinophilsd Lymphocytosis Lymphoblasts Neutropenia/thrombopenia/hemolytic anemia

+ + + + +

+ + + + +

(+)2 (+) (+) (+)3 + + + (+)

+ + + + ? – – +

Note that almost all symptoms are similar, with the exception of lymphocytosis and lymphoblasts, which are only found in drug-induced hypersensitivity syndrome/drug rash with eosinophilia and systemic symptoms (DiHS/DRESS). Compiled according to references 35 and 36. + = Present in the majority of patients; (+) = only seen with certain drugs or rarely. 1 Clinical picture quite heterogeneous with different drugs, see text and table 5 in chapter of Shiohara [pp 251–266]. 2 Allopurinol. 3 Abacavir.

Table 4. Drugs eliciting severe cutaneous or systemic reactions1 Acute generalized exanthematous pustulosis (AGEP)

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN)

Drug-induced hypersensitivity syndrome/drug rash with eosinophilia and systemic symptoms (DiHS/DRESS)

Aminopenicillins1 Cephalosporins Pristinamycin Celecoxib Quinolone Diltiazem Terbinafin

Nevirapine Allopurinol Phenytoin2 Carbamazepin2 Lamotrigine Cotrimoxazole Barbiturate NSAID (oxicams)

Carbamazepin2 Phenytoin2 Lamotrigine3 Minocyclin Allopurinol Dapsone Sulfasalazine Cotrimoxazole Abacavir3

1

List incomplete, the most frequent elicitors are italicized. The type of reaction might be determined by the presence of a certain HLA-B phenotype. 3 Lamotrigine- and abacavir-induced systemic reactions often lack eosinophilia, and abacavirinduced hypersensitivity affects preferentially the respiratory and gastrointestinal tracts. 2

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

183

Fig. 8. Clinical pictures of (maculo)papular drug eruption with some bullae and a pustular drug eruption (AGEP). a Confluent, papular drug eruption with strong T-cell infiltration which leads already to some blister formation: it illustrates the danger signs confluency and blisters formation. b Pustular drug eruption (AGEP): note the multiple, intraepidermal, non-follicular pustules (for details, see text).

ing into the epidermis express CXCL8 and GMCSF. Analysis of sequential patch test reactions at 48–96 h suggests that drug-specific cytotoxic T cells emigrate first, causing formation of vesicles by killing keratinocytes. Subsequently, T cells and keratinocytes release GM-CSF and CXCL8, which recruits granulocytes into the vesicles, which are thereby transformed into pustules [21]. Some pustules coalesce together and may form bullae. The lethality is about 2–4%, and occurs particularly in older people. This disease and the underlying T-cell reaction seems to be a model for sterile neutrophilic inflammations (type IVd) like pustular psoriasis and Behçet’s disease [22]. Bullous Exanthem, Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis The most severe forms of drug-induced skin reactions go along with formation of bullae. The

184

most severe bullous skin diseases are StevensJohnson syndrome (SJS) and toxic epidermal necrolysis (TEN) [for details, see chapter of Allanore and Roujeau, pp 267–277]. TEN and SJS are rare (ca. 1: 1,000,000 for TEN, ca. 1: 100,000 for SJS). Nowadays they are considered to be the milder and more severe form of the same disease (SJS !10%, TEN 130% skin detachment). They are graded according to SCORTEN, whereby age, underlying disease and the amount of maximal skin detachment are the most important prognostic factors. According to the European study group of severe cutaneous drug reactions, SJS has a mortality of ca. 13%, and TEN of ca. 39%! The intermediate form with 10–30% skin detachment is called SJS/TEN overlap syndrome and has a lethality of ca. 21% [see chapter of Mockenhaupt, pp 18–31]. SJS/TEN are clearly different from erythema exsudativum multiforme, which is mainly caused

Pichler

by viral infections [3], is often recurrent and affects younger persons (mean age 24 years). In ca. 6% no drug treatment was given the week before SJS/TEN started and an infectious origin (Mycoplasma pneumoniae, Klebsiella pneumoniae) is suspected. It can also be due to a graft-versus-host disease. Most reactions start within the first 8 weeks of treatment (mean of first symptoms at ca. 17 days), with some differences according to the causing drug (e.g. with sulfonamide antibiotics, it may appear later). The disease might develop quite rapidly: initially often a macular, purple-red exanthem can be observed, which can become painful – an ominous sign. Within 12–24 h bullae might already be seen, and the Nikolsky sign is positive. Stopping drug treatment at this time point may not prevent SJS any longer, but might prevent an even more severe reaction (TEN). Mucous membranes (mouth, genitalia) are involved with blister formation, and a purulent keratoconjunctivitis with formation of synechiae may result in permanent eye damage. The main causes for SJS/TEN are drugs (table 3), which might differ in frequency in various regions due to genetic and ethnic background. Important risk factors are HIV infection (low CD4, high CD8 counts), renal diseases, and active systemic autoimmune diseases like SLE, Still’s syndrome, Sjögren’s disease and rheumatoid arthritis, which are all associated with immune stimulations. In histology of TEN, many dead keratinocytes are found, but cell infiltration is only scarcely seen. However, the bullae may be filled by cytotoxic CD8+ T cells, expressing CD56+ and
– T-cell receptors, which kill via perforin/granzyme B but not via the Fas-mediated pathway at this stage of the disease [19]. On the other hand, the massive cell death of keratinocytes is hard to reconcile with a cell contact-dependent killing process. It has been proposed that the apoptosis of keratinocytes is due to FasL, a soluble molecule of the TNF family, which binds to keratinocytes

via Fas and functions as a so-called death receptor [29]. Since blocking anti-Fas antibodies are found in immunoglobulin preparations, it has been proposed to treat patients with TEN with immunoglobulins. However, the efficiency is controversial [30]. Extensive research did not reveal a pharmocogenetic predisposition or low GSH levels in the affected persons. However, recent data have revealed striking HLA associations of severe, CD8mediated drug hypersensitivity reactions [31– 33]. Three factors play a role: (a) the type of drug; (b) the type of reaction – as it was found for cytotoxic CD8 reactions only, and (c) possibly race, as some reactions were mainly described in Han Chinese. The HLA-B allele, which is the most polymorphic HLA allele, seems to be involved: in carbamazepine-induced SJS/TEN it is HLAB*1502, for phenytoin HLA-B*5601 and for abacavir HLA-B*5701 together with hsp70 [see chapters of Hung et al., pp 105–114, and Nolan et al., pp 95–104]. It seems that in these CD8+ T-cell reactions the HLA-B alleles favor the presentation of certain peptides which are able to optimally present the drug/hapten. Alternatively, if the stimulation occurred via the p-i concept, only certain MHC class B alleles might supplement the T-cell stimulation by drugs, while in the absence of this allele the T cell fails to react to the drug binding to its T-cell receptor. The extremely high association of certain HLA-B alleles and hypersensitivity reactions to certain drugs may be used to prevent such side effects in the future as HLA typing may identify patients at risk. Studies to determine the benefit of such predictive testing are under way and are promising for abacavir [see chapter of Nolan et al., pp 95–104]. Systemic Drug Reactions – Severe Drug Hypersensitivity Syndromes Some drugs are known to cause a severe systemic disease with fever, lymph node swelling, massive hepatitis and various forms of exanthems (ta-

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

185

ble 3). Few patients develop colitis, pancreatitis or interstitial lung disease [34]. Over 70% of these cases have a marked eosinophilia (often 11 G/l), and activated lymphocytes are often found in the circulation, similar to acute HIV or generalized herpesvirus infections. This syndrome has many names, the most frequently used are drug (induced) hypersensitivity syndrome (DHS or DiHS) or drug rash with eosinophilia and systemic symptoms (DRESS) [for details, see chapter of Shiohara et al., pp 251–266]. Importantly, the symptoms can start up to 12 weeks after the start of drug treatment, often after increasing the dose, and may also persist and recur for many weeks even after cessation of drug treatment. The clinical picture resembles a generalized viral infection, for example an acute Epstein-Barr virus infection, but it is distinguished by the prominent eosinophilia in most, but not all reactions. Actually, a recent study revealed quite dramatic clinical and laboratory differences of this syndrome dependent on the drug used [see table 3 in the chapter of Shiohara et al., pp 251–266, and 35], as only some drugs elicited often a kidney disease (e.g. allopurinol) while others frequently lacked eosinophilia (e.g. lamotrigine). Many patients have a facial swelling, and some have signs of a capillary leak syndrome – like patients having a cytokine release syndrome [36]. Indeed, the clinical features are actually similar to the symptoms described with the TGN1412 incident, which was due to an anti-CD28 antibody eliciting a ‘cytokine storm’ (table 4) [see chapter of Pichler and Campi, pp 151–165, and 37]. Various cytokines are also massively increased in the serum of patients with DiHS/DRESS. As the clinical picture is quite dramatic and as the disease tends to persist in spite of stopping drug treatments, many patients are not diagnosed timely and correctly. However, any doctor using anticonvulsants should know this syndrome, as it might occur in 1:3,000 treated patients! The mortality is ca. 10%, and some patients only survive if they receive an acute liver transplantation.

186

Patients with DiHS/DRESS have many activated T cells in the circulation. These drug-specific T cells are stimulated by the parent compound (p-i concept) and they secrete high amounts of IL-5 and IFN- [38]. A peculiar feature of this syndrome is its long-lasting clinical course despite withdrawal of the causative drug. There may also be persistent intolerance to other chemically distinct drugs, leading to flare-up reactions to a rather innocuous drug (e.g. acetaminophen) weeks to months after stopping the initial drug therapy, further adding to the diagnostic confusion (see multiple drug allergies, and figure 9c). Treatment often includes high doses of corticosteroids, in particular if the hepatitis is severe. Recently, it has been shown that in many patients with this syndrome, human herpesvirus 6 DNA can be found during the 3rd or 4th week of the disease (but not before), followed by a rise of antibodies to human herpesvirus 6 [see chapter of Shiohara et al., pp 251–266, and 39]. Other reports document a reactivation of a cytomegalovirus infection. Thus, similar to HIV, where Tcell activation can also enhance virus production, the drug-induced massive immune stimulation may somehow reactivate these latent lymphotropic herpesviruses, which subsequently replicate and possibly contribute to the chronic course and persistent drug intolerance in the affected patients. Other authors attribute a more causal role to the viral infections in this syndrome altogether [see chapter of Shiohara et al., pp 251–266]. Many drugs can induce an isolated hepatitis [see contribution of Cerny and Bertoli, pp 278– 294 and some drugs (penicillins, proton pump inhibitors, quinolones, disulfiram, etc.) an (interstitial) nephritis [see contribution of Keller et al., pp 295–305]. Recent data of drug-induced interstitial nephritis have illustrated the important role of drug-specific T cells in this disease [40]. Actually, as in cutaneous drug hypersensitivity reactions, the nephritic lesions also show either

Pichler

Symptoms

Table 5. Tolerance mechanism to drugs drug x

drug y x Skin and LTT ++ to x and y tI

Symptoms

a

drug x drug y

Skin and LTT a) ++ to x only b) ++ to both drugs t

Symptoms

b

c

drug x

drug x l

Skin and LTT ++ to x and x l t

Fig. 9. Immune-mediated multiple drug hypersensitivities can be attributed to different mechanisms. a Two structurally distinct drugs elicit a separate immune response and sometimes also different clinical symptoms. The sensitization may occur separately, or during the same time span and simultaneous treatments. b Socalled flare-up reactions are due to the massive immune stimulation during a drug allergy: this facilitates a reaction to a second drug, which may be too weak to generate a persistent immune response (alternatively, the immune response to the second drug may be strong enough to elicit also a memory to the second reaction, then we have the situation described under a). c Structurally-related drugs can cause reappearance of symptoms due to cross-reactivity. The second reaction occurs more rapidly.

more granulomatous, eosinophil-rich or neutrophilic inflammations. In drug-induced interstitial nephritis, eosinophils can sometimes be detected in the urine (even in the absence of eosinophilia in the blood), and might help to indicate that a drug allergy is occurring. Even rarer are interstitial lung diseases (a frequent cause was Furantoin), pancreatitis, isolated fever or eosinophilia as the only symptom of a drug allergy.

Ignorance: Even if drugs bind to immune receptors, the interaction (affinity, surface contact) is too weak to elicit a significant reaction Lack of danger signals and preactivation: The haptencarrier complexes does not sufficiently stimulate the innate immunity, which is necessary to develop a primary immune response. Or, in case of p-i concept, the TCR stimulation by the drug is not sufficient to induce a cytokine production and proliferation, as the T cell is not sufficiently preactivated to react to this signal Regulatory T cells (T-reg as well as IL-10/TGF-producing Tr1 cells) may be insufficiently active in some patients, making them prone to react to small chemical compounds. These patients may have multiple drug allergies and may also suffer from autoimmunity The liver as a tolerogenic organ: The generation of reactive metabolites in the liver may induce tolerance to it, which might even prevent the development of an immune reaction to the drug in the periphery

Multiple Drug Hypersensitivity Syndrome The term multiple drug hypersensitivity is widely used for different forms of side effects to various drugs. Some use it to characterize patients with multiple drug intolerance (‘pseudoallergy’ to various, structurally distinct NSAIDs, etc.) while others reserve this term for well-documented repeated immune-mediated reactions to structurally not related drugs [41]. Cross-reactivity due to structural similarity is not included. According to our experience, about 10% of patients with well-documented drug hypersensitivity (skin and/or lymphocyte transformation test positive) have multiple drug allergies [42], for example, they may have reacted to an injected lidocaine with a massive angioedema; years later the same patient develops an allergy to corticosteroids. Both drugs are positive in skin and lymphocyte transformation tests. Alternatively, a patient reacts to amoxicillin, phenytoin and sulfamethoxazole within a few months, but with different symptoms (MPE, DiHS/DRESS, eryth-

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

187

rodermia). Most patients with multiple drug hypersensitivity have had rather severe reactions to at least one drug. An IgE-mediated reaction might be followed by a T-cell-mediated reaction. The pathomechanism of this syndrome is unknown, but it is tempting to speculate that the tolerance mechanism to small molecular compounds fails in these patients (table 5). An immune reaction to a drug, be it via a hapten or p-i mechanism, can be seen as a failure of tolerance, and the same patient might not only be prone to develop other drug allergies as well, but also autoimmunity. Preliminary data suggest indeed that a previous drug allergy might be a risk factor for the development of a delayed hypersensitivity reaction to contrast media. Multiple drug hypersensitivity should be differentiated from flare-up reactions (fig. 9), which

occur when – due to an existing drug allergy – therapy is changed to a second drug, which is then given into a still highly activated immune system. Such flare-up reactions are common in severe reactions like DiHS/DRESS (see above DRESS/DiHS, and Sullivan and Shear [36]), as in patients with systemic drug allergies, the Tcell immune system is massively activated, similar to acute viral infections. As in the latter, these patients seem to have a higher tendency to react to a new drug and might show a flare up of their exanthema to a new antibiotic. Later, in remission, the second drug may remain negative in skin tests and might also be well tolerated, as the co-stimulatory conditions do not exist any more. However, there are also descriptions that a second sensitization persists, which occurred during the DRESS/DiHS [43].

References 1 Naisbitt DJ, Gordon SF, Pirmohamed M, Park BK: Immunological principles of adverse drug reactions: the initiation and propagation of immune responses elicited by drug treatment. Drug Saf 2000;23:483–507. 2 Lazarou J, Pomeranz BH, Corey PN: Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 1998;279: 1200–1205. 3 Roujeau JC, Stern RS: Severe adverse cutaneous reactions to drugs. N Engl J Med 1994;331:1272–1285. 4 Bigby M, Jick S, Jick H, Arndt K: Druginduced cutaneous reactions. A report from the Boston Collaborative Drug Surveillance Program on 15,438 consecutive inpatients, 1975 to 1982. JAMA 1986;256:3358–3363. 5 Hunziker T, Kunzi UP, Braunschweig S, Zehnder D, Hoigne R: Comprehensive hospital drug monitoring: adverse skin reactions, a 20-year survey. Allergy 1997;52:388–393. 6 Manfredi M, Severino M, Testi S, Macchia D, Ermini G, Pichler WJ, et al: Detection of specific IgE to quinolones. J Allergy Clin Immunol 2004;113:155– 160.

188

7 Coombs PR, Gell PG: Classification of allergic reactions responsible for clinical hypersensitivity and disease; in Gell RR (ed): Clinical Aspects of Immunology. Oxford, Oxford University Press, 1968, pp 575–596. 8 Pichler WJ: Delayed drug hypersensitivity reactions. Ann Intern Med 2003; 139:683–693. 9 Martin S, Weltzien HU: T cell recognition of haptens, a molecular view. Int Arch Allergy Immunol 1994;104:10– 16. 10 Pumphrey R: Anaphylaxis: can we tell who is at risk of a fatal reaction? Curr Opin Allergy Clin Immunol 2004;4: 285–290. 11 Pumphrey RS: Fatal anaphylaxis in the UK, 1992–2001. Novartis Found Symp 2004;257:116–128; discussion 128–132, 157–160, 276–285. 12 Tramer MR, von Elm E, Loubeyre P, Hauser C: Pharmacological prevention of serious anaphylactic reactions due to iodinated contrast media: systematic review. BMJ 2006;333:675. 13 Aster RH: Drug-induced immune thrombocytopenia: an overview of pathogenesis. Semin Hematol 1999;36: 2–6.

14 Arndt PA, Garratty G: The changing spectrum of drug-induced immune hemolytic anemia. Semin Hematol 2005;42:137–144. 15 Greinacher A, Warkentin TE: Recognition, treatment, and prevention of heparin-induced thrombocytopenia: review and update. Thromb Res 2006; 118:165–176. 16 Bircher AJ, Harr T, Hohenstein L, Tsakiris DA: Hypersensitivity reactions to anticoagulant drugs: diagnosis and management options. Allergy 2006;61: 1432–1440. 17 Aitman TJ, Dong R, Vyse TJ, Norsworthy PJ, Johnson MD, Smith J, et al: Copy number polymorphism in Fcgr3 predisposes to glomerulonephritis in rats and humans. Nature 2006;439: 851–855. 18 Stokol T, O’Donnell P, Xiao L, Knight S, Stavrakis G, Botto M, et al: C1q governs deposition of circulating immune complexes and leukocyte Fc receptors mediate subsequent neutrophil recruitment. J Exp Med 2004;200:835– 846.

Pichler

19 Nassif A, Bensussan A, Dorothee G, Mami-Chouaib F, Bachot N, Bagot M, et al: Drug-specific cytotoxic T cells in the skin lesions of a patient with toxic epidermal necrolysis. J Invest Dermatol 2002;118:728–733. 20 Schnyder B, Frutig K, Mauri-Hellweg D, Limat A, Yawalkar N, Pichler WJ: Tcell-mediated cytotoxicity against keratinocytes in sulfamethoxazole-induced skin reaction. Clin Exp Allergy 1998;28:1412–1417. 21 Britschgi M, Steiner UC, Schmid S, Depta JP, Senti G, Bircher A, et al: T-cell involvement in drug-induced acute generalized exanthematous pustulosis. J Clin Invest 2001;107:1433–1441. 22 Keller M, Spanou Z, Schaerli P, Britschgi M, Yawalkar N, Seitz M, et al: T cellregulated neutrophilic inflammation in autoinflammatory diseases. J Immunol 2005;175:7678–7686. 23 Beeler A, Engler O, Gerber BO, Pichler WJ: Long-lasting reactivity and high frequency of drug-specific T cells after severe systemic drug hypersensitivity reactions. J Allergy Clin Immunol 2006;117:455–462. 24 Schaerli P, Ebert L, Willimann K, Blaser A, Roos RS, Loetscher P, et al: A skin-selective homing mechanism for human immune surveillance T cells. J Exp Med 2004;199:1265–1275. 25 Hari Y, Frutig-Schnyder K, Hurni M, Yawalkar N, Zanni MP, Schnyder B, et al: T cell involvement in cutaneous drug eruptions. Clin Exp Allergy 2001; 31:1398–1408. 26 Schmid DA, Campi P, Pichler WJ: Hypersensitivity reactions to quinolones. Curr Pharm Des 2006;12:3313–3326.

27 Schmid DA, Depta JP, Pichler WJ: Tcell-mediated hypersensitivity to quinolones: mechanisms and cross-reactivity. Clin Exp Allergy 2006;36:59–69. 28 Roujeau J, Bioulac-Sage P, Bourseau C: Acute generalized exanthematous pustulosis: analysis of 63 cases. Arch Dermatol 1991;127:1333–1338. 29 Viard I, Wehrli P, Bullani R, Schneider P, Holler N, Salomon D, et al: Inhibition of toxic epidermal necrolysis by blockade of CD95 with human intravenous immunoglobulin. Science 1998;282: 490–493. 30 Bachot N, Roujeau JC: Intravenous immunoglobulins in the treatment of severe drug eruptions. Curr Opin Allergy Clin Immunol 2003;3:269–274. 31 Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, et al: Association between presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet 2002; 359:727–732. 32 Chung WH, Hung SI, Hong HS, Hsih MS, Yang LC, Ho HC, et al: Medical genetics: a marker for Stevens-Johnson syndrome. Nature 2004;428:486. 33 Hung SI, Chung WH, Liou LB, Chu CC, Lin M, Huang HP, et al: HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci USA 2005;102:4134–4139. 34 Knowles SR, Shapiro LE, Shear NH: Anticonvulsant hypersensitivity syndrome: incidence, prevention and management. Drug Saf 1999;21:489–501. 35 Peyriere H, Dereure O, Breton H, Demoly P, Cociglio M, Blayac JP, et al: Variability in the clinical pattern of cutaneous side effects of drugs with systemic symptoms: Does a DRESS syndrome really exist? Br J Dermatol 2006;155:422–428.

36 Sullivan JR, Shear NH: The drug hypersensitivity syndrome: What is the pathogenesis? Arch Dermatol 2001;137: 357–364. 37 Suntharalingam G, Perry MR, Ward S, Brett SJ, Castello-Cortes A, Brunner MD, et al: Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N Engl J Med 2006;355: 1018–1028. 38 Naisbitt DJ, Farrell J, Wong G, Depta JP, Dodd CC, Hopkins JE, et al: Characterization of drug-specific T cells in lamotrigine hypersensitivity. J Allergy Clin Immunol 2003;111:1393–1403. 39 Hashimoto K, Yasukawa M, Tohyama M: Human herpesvirus 6 and drug allergy. Curr Opin Allergy Clin Immunol 2003;3:255–60. 40 Spanou Z, Keller M, Britschgi M, Yawalkar N, Fehr T, Neuweiler J, et al: Involvement of drug-specific T cells in acute drug-induced interstitial nephritis. J Am Soc Nephrol 2006;17:2919– 2927. 41 Sullivan T: Studies of the multiple drug allergy syndrome. J Allergy Clin Immunol 1989;83:270. 42 Gex-Collet C, Helbling A, Pichler WJ: Multiple drug hypersensitivity – proof of multiple drug hypersensitivity by patch and lymphocyte transformation tests. J Investig Allergol Clin Immunol 2005;15:293–296. 43 Gaig P, Garcia-Ortega P, Baltasar M, Bartra J: Drug neosensitization during anticonvulsant hypersensitivity syndrome. J Investig Allergol Clin Immunol 2006;16:321–326.

Prof. Dr. Werner J. Pichler Division of Allergology Department for Rheumatology and Clinical Immunology/Allergology Inselspital, University of Bern CH–3010 Bern (Switzerland) Tel. +41 31 632 2264 or 3174, Fax +41 31 632 2747 E-Mail [email protected]

Drug Hypersensitivity Reactions: Classification and Relationship to T-Cell Activation

189

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 190–203

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins) Maria José Torres ⭈ Cristobalina Mayorga ⭈ Miguel Blanca Allergy Service and Research Unit for Allergic Diseases, Carlos Haya Hospital, Malaga, Spain

Abstract Anaphylaxis and urticaria are the main clinical entities associated with betalactam allergy. They usually occur within 1 h of drug exposure, although urticaria may occasionally appear later. The shorter the interval to the reaction, the greater the possibility of this being IgE mediated, although in some cases of late-onset urticaria a T-cell mechanism has to be considered. Initially the haptens identified were benzylpenicilloyl and minor determinant mixtures of benzylpenicillin. Because of the changing patterns of consumption over the years, other betalactams, such as amoxycillin and cephalosporins, are now the most common drugs inducing reactions with the involvement of side-chainspecific determinants. These points, among others, are essential for performing in vivo and in vitro tests and understanding cross-reactivity. Copyright © 2007 S. Karger AG, Basel

Definition

Betalactams (BL) are the drugs most frequently involved in specific immunological drug reactions, inducing several clinical entities such as urticaria, anaphylaxis, exanthema, fixed drug eruption, toxic epidermal necrolysis, StevensJohnson syndrome, acute exanthematic pustulosis and drug hypersensitivity syndrome. Organspecific reactions, such as hepatitis, have also been reported [1, 2].

From a clinical viewpoint, the reactions can be classified into two groups: immediate reactions, appearing no more than 1 h after drug intake, mediated by specific IgE antibodies, and non-immediate reactions, occurring with a variable interval from 1 h to a few days (usually 24– 48 h) and mediated by sensitized T cells [3]. This chapter deals with immediate reactions, and their two most common entities, anaphylaxis and immediate urticaria. The concept of anaphylaxis applies to symptoms occurring immediately after drug intake and affecting more than one organ, whether or not accompanied by hypotension (anaphylactic shock). In these circumstances, urticaria is considered when lesions are limited to the skin with variable-sized wheals, usually 1 4–5 cm in diameter, and which can be accompanied by angioedema. Whether these are two independent clinical entities or part of the same process, with different degrees of severity, is currently not clear [1]. Allergic reactions were reported immediately after penicillin was first administered, when penicillin G was the first antibiotic to treat bacterial infections. Since then, the original family of drugs has increased with the appearance of new BL, such as cephalosporins, carbapenems, monobactams and clavams. Over many years the expo-

sure of the population to BL compounds has varied; indeed, it still varies between countries and also within the same country over time. The most relevant change over the last two decades has been the progressive replacement of benzylpenicillin (BP) as the sensitizing agent by other BL, such as semisynthetic penicillins or cephalosporins. Older persons have therefore had the possibility of being exposed to a wider variety of BL, including the former impure penicillins, whereas most young people have been exposed mainly to amoxicillin (AX) and, to a lesser extent, cephalosporins.

O O R C NH R C NH

S N

S

N

O

COOH Penicillins

CH2 R2

O COOH Cephalosporins O

R' S

N O

C

R

R C NH N

O

OH Carbapenems

O

SO3H

Monobactams

Drugs Involved

BL are formed by a large number of chemical structures that share the BL nuclear ring and pharmacological activity. The chemical structure of the four major groups of BL is shown in figure 1. In the case of penicillins, the BL nucleus is attached to a thiazolidine ring. The side chain contributes to the specific name of the penicillin, which is relevant not only for its pharmacological and antibacterial properties but also for its immunological specificity, because it contributes to the structure of the epitope [2, 3] (table 1). Cephalosporins are BL that contain a dihydrothiazine in place of the thiazolidine ring, with two different side chains at the R1 and R 2 positions [4]. Compared to penicillins, the number of cephalosporins is higher and they are divided into four generations, with different activity against Gram-positive and Gram-negative bacteria (table 2). Carbapenems differ from penicillins in that they are unsaturated and contain a carbon atom instead of sulfur in the thiazolidine ring. The two compounds used are imipenem and meropenem, with similar therapeutic equivalence and clinical toxicity. Aztreonam (AZT) is a monocyclic BL isolated from Chromobacterium violaceum which interacts with penicillin-binding proteins

Fig. 1. Chemical structure of different betalactams.

of susceptible microorganisms; it is resistant to many betalactamases. Finally, there is the group of betalactamase inhibitors produced by Grampositive and Gram-negative bacteria, whose most relevant compound is clavulanic acid produced by Streptomyces clavuligerus. These compounds are shown in table 3.

Pathophysiology

BL molecules have the capacity, by the spontaneous opening of the BL ring, to bind to serum and cell membrane proteins, forming stable covalent drug-protein adducts, also known as haptencarrier conjugates. Antigen-presenting cells process and present these conjugates to T cells, which in immediate reactions interact with B cells inducing the production of IgE antibodies [3]. As in all type I allergic reactions, when bi- or multivalent hapten-carrier conjugates bridge two or more IgE antibodies located at the mast cell/basophil surface, they trigger an intracellular activation cascade that induces the release of

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins)

191

Table 1. Side chain chemical structures of penicillin antibiotics Non-proprietary name

Side chain

Penicillin G

2-Phenylacetic acid

0)

0

0 /

0 / )

4 ) 0

Penicillin V

0)

2-Phenoxylacetic acid

0

0 /

0 / )

4 ) 0

Methicillin

0)

2,6-Dimethoxybenzoic acid

0 0 / / )

4

0 0

)

Oxacillin

0

2-(3-Phenyl-4-isoxazolyl)acetic acid

0) 0 / 4

0

Cloxacillin

/

/ )

)

0

0

0)

3-(2)-Chlorophenyl-5-methyl-4isoxazolecarboxilic acid

0 / 4

0

/ )

)

0 / $* 0

0)

Dicloxacillin

3-(2,6-Dichlorophenyl)-5-methyl-4isoxazolecarboxilic acid

0 / 4

/ )

)

0

$* 0 / $*

Nafcillin

0

2-Ethoxy-1-naphthoic acid

0) 0 / 0

/ )

4

0

) 0

Ampicillin

0)

2-Amino-2-phenylacetic acid

0 /

) /

4

/ ) )

)

0

Amoxicillin

0

0)

2-Amino-2-(4-hydroxyphenyl)acetic acid

0 ) /

/

0)

4

Carbenicillin

/ )

0 /B

0o

0

2-Phenylmalonic acid

0 0) /

0

4

Ticarcillin

/ )

0

0

0)

2-(3-Thienyl)malonic acid

0 4

/ 4

0)

) )/ 0

192

0

Torres ⴢ Mayorga ⴢ Blanca

Table 1 (continued) Non-proprietary name

Side chain

Mezclocillin 0

4

2-{[(Methylsulfonyl)-2-oxo-1-imidazolidinyl]carbonyl}amino-2-phenylacetic acid

0

/ 0

0

0) 0

/ 0

/

)/

4

/ )

)

Piperacillin

0 0

/ 0

0)

0 0

/ 0

/ 4 )

2-{[(4-Ethyl-1,2,3-dioxopiperazino) carbonyl]amino}-2-phenylacetic acid

)/ / )

inflammatory mediators. Although no precise studies about the kinetics of this haptenization process exist, the in vivo binding is sufficiently rapid to induce symptoms within 1 h of drug intake. In vitro studies have shown that optimal binding with serum proteins occurs within 1 h whereas binding with cell proteins occurs in about 8 h, although symptoms can sometimes occur after more than 1 h, the so-called accelerated reactions. Whether these are due to a slower rate of haptenization or to a T-cell response is not known at present, although the longer the time before the appearance of urticaria, the greater the possibility that a non-IgE mechanism may be involved [5, 6]. The induction of drug-carrier conjugates endows them with immunogenic properties, able to induce T cells or antibodies of different isotypes, which in turn leads to different pathological responses. Memory T cells or specific antibodies that recognize penicillin haptens can be found in most individuals with good tolerance to BL. However, reactions occur only in some subjects exposed to the drug; the reasons for this remain unknown, although a genetic basis may well be involved. The structure formed after the binding of the opened BL ring of BP to an amino group of the protein is known as benzylpenicilloyl (BPO), also

0

called the major determinant [3]. This determinant is the best characterized, both in vivo and in vitro [2, 3]. From the outset, researchers reported that these structures were recognized by all isotype antibodies, both in animals and humans, including monoclonal antibodies [7, 8]. In vivo, this BPO group is mainly formed within the human serum albumin (HSA). When this carrier is fully haptenated, only surface epitopes are accessible to IgE antibodies. This can be avoided by conjugating BP to a linear synthetic carrier such as polylysine (PLL). This conjugate, known as PPL, induces a wheal-and-flare response in a large proportion of cases when applied to the skin, and is able to bind IgE antibodies when used in vitro tests [9]. Nevertheless, it was soon detected that the application of BP itself to the skin induced a positive response, and it was seen that BP generated other determinants that were the products of its metabolism. This structure was called the minor determinant mixture (MDM), because only a limited proportion of sensitized patients responded. At that time, several chemical structures generated by penicillin metabolism were also considered as candidate minor haptens, although because of their instability, immunochemical studies had difficulties proving them to be relevant haptens in humans [10]. To date, the concept of MDM is maintained

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins)

193

Table 2. Side chain chemical structures of cephalosporin antibiotics Generation

Non-proprietary name

First

Cephalothin

Side chain 0o

0

2-(2-Thienyl)acetic acid

/B 0

0

0 /

0

4

/

Cefazolin

4

/ )

) /

2-(1H-1,2,3,4-Tetraazol-1yl)acetic acid

/ /

/ )

) /

4

4

0

/

/

4

0 )0

Cefalexin

0

0

2-Amino-2-phenylacetic acid

0) 0 /

0

4

/ )

)

/) 0

Cefadroxil

0)

2-Amino-2-(4-hydroxyphenyl)acetic acid

0 0) /

0

4

/ )

)

/)

Second

Cefamandole

0

2-Hydroxy-2-phenylacetic acid

0) 0

4

)0

/

/ 4

/

/ )

) /

0

/

0

Cefoxitin

0)

2-(2-Thienyl)acetic acid

0 4

/

0

/)

0

4 )

/ )

0

4

Cefaclor

0

2-Amino-2-phenylacetic acid

0) 0

$* /

/) / )

4

0

) 0

Cefuroxime

0)

2-(2-Furyl)-2(methoxyimino)acetic acid

0 0

0

/

0

/)

/

4

0

/ )

)

0 0

Loracarbef

0)

2-Amino-2-phenylacetic acid

0 $* /

0

4 )

/ )

0

0)

/)

Cefonicid

4

2-Hydroxy-2-phenylacetic acid

0

)0 0

4

0

/

/

4

/ /

)

)0

/ )

0

/

Cefotetan

0

0

0)

0

0 4

)/

/

0)

4

2-(4-Carboxy-1,3-dithietan-2-yliden)malonic acid

4 /

/

4 )

/

194

/

/ 0 )

0

Torres ⴢ Mayorga ⴢ Blanca

Table 2 (continued) Generation

Non-proprietary name

Side chain 0

Ceforanide 0

0

4

4

0

4

/ /

2-[2-(Aminomethyl)phenyl]acetic acid

)/

/

/

Third

0)

)0

/ )

)

/

0

0

Cefotaxime

0) 0

0

0

2-(4-Amino1,3-thiazol-2-yl)-2-(methoxyimino)acetic acid

0 /

/

4

4

/

/ )

)

/)

0

0

Cefpodoxime

0) 0

0

4

4

/

/ )

)

2-(4-Amino1,3-thiazol-2-yl)-2-(methoxyimino)acetic acid

0 /

/

/)

0

0

Cefixime

0) 0

4

2-(5-Imino-4,5-dihydro-1H-pyrrol-2-yl)2(methoxyimno)acetic acid

0 ) /

/

/

/)

/ )

)

0 0

Ceftriaxone

0

0

2-(2-Amino-1,3-thiazol-4-yl)-2-(methoxyimino)acetic acid

0)

) 0

0

)/ /

4

/

/

4 4

/ )

)

/ /)

0

Cefoperazone

/

0

0

0 /

4

4

/

/ )

)

2-{[(4-Ethyl-2,3-dioxopiperazino)carbonyl]amino}-2-(4-hydroxyphenyl)acetic acid

0

/

/)

0

0

Ceftazidime

0) 0

/

/

4

0

4

/) /

/ )

) )0

2-({[(2-Amino-1,3-thiazol-4-yl)(carboxy)methylidene]amino}oxyl)-2-methylpropionic acid

/ 0 0

0

Fourth

Cefepime

2-(4-Amino1,3-thiazol-2-yl)-2-(methoxyimino)acetic acid

N 0 0

0)

/ 0

4

0 )/

/

0 / /

/ /

4 )

/ )

for BP and benzylpenicilloic acid, though a number of questions need to be answered regarding the chemical identities of all these proposed compounds [6]. Studies using polyclonal and monoclonal antibodies have provided clues about the contribution of different parts of the BL molecule to the antibody-binding site. The experiments of de Haan et al. [7] showed that three different epitopes could be identified: the side chain of BP; the

0

part formed by the binding of the carbonyl of the BL to the amino group of the protein, the socalled new antigenic determinant, and the thiazolidine ring. Other studies have shown that the side chain and the nuclear part of the BL are more important. However, later studies indicated that recognition is not limited to the side chain and/or to the nuclear part of the BL, but rather that the whole structure is required for optimal recognition, although the side-chain structure is often

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins)

195

Table 3. Chemical structures of carbapenems, monobactams and ␤-lactamase inhibitors Compound

Non-proprietary name

Chemical structure

Carbapenems

Imipenem

)

) / 0

)

/

)

) 4

) / 0

0 0

)

) OH ) OH

Meropenem 0

) OH

/ )

)

/

) 0 )

4 /

0 0 0 )

Monobactams

Aztreonam 0 0 )0

0)

0

4

/

/

0

4 / )

)

␤-Lactamase inhibitors

Clavulanic acid

0

/ )

0o

0

,

/

0 )

/ 0

0)

essential to differentiate specificities [8]. Studies with polyclonal IgE antibodies obtained from allergic patients support the concept that either the side chain or the nuclear part are important when considering antibody specificity [11]. In penicillins, the chemical structure formed from the fusion of the BL with the thiazolidine ring results in a high degree of tension, which de-

196

)

termines a greater chemical reactivity. In cephalosporins, however, the heterocycles are formed by four and six carbons, which result in less tension than for the BL ring. Haptenization of proteins by cephalosporins is thus slower and less efficient than with penicillins. The R 2 chemical structure can modulate chemical reactivity, but because it is lost after the opening of the BL ring,

Torres ⴢ Mayorga ⴢ Blanca

only the R1 and part of the BL ring remain bound to the carrier, contributing to the formation of the epitope [6]. However, our knowledge concerning the immunological response to cephalosporins is more incomplete. Initially, during the development of first-generation cephalosporins, cephalothin was shown to have high cross-reactivity with BP. In some other instances, similarities between the side chain of cephalothin or cephalexin, thiopene, and the side chain of BP, phenylacetic acid, seemed to contribute to cross-reactivity. Another explanation was attributed to molecular impurities, although cross-allergenicity seems to exceed the role of cross-contamination. The most relevant structure for IgE hypersensitivity reactions is the chemical structure of the R1 side chain [12]. Although selective responses to cephalosporins were reported, their number has increased since the 1990s, including responses to frequently prescribed compounds such as cefuroxime, ceftriaxone, cefotaxime, cefepime and ceftazidime.

Clinical Course

In BL hypersensitivity, symptoms usually appear immediately after drug intake, within 1 h. They may occasionally occur after more than 1 h, the so-called accelerated reactions, but in general it is accepted that the longer the interval between drug intake and the appearance of symptoms, the less likelihood of it being induced by an IgE mechanism [13]. The shorter interval is associated with anaphylaxis, and symptoms often follow the order palmo-plantar pruritus extending to many other parts of the body, generalized erythema, angioedema and/or maculopapulae, frank urticaria, difficulty of breathing, tachycardia, abdominal cramps, nausea, vomits, diarrhea, hypotension and loss of consciousness. Urticaria may arise during the wide clinical expression of anaphylaxis, though it can be limited to the skin and

be the sole manifestation of immediate hypersensitivity. Clinical observations suggest that urticaria is more frequently detected in subjects positive to PPL but anaphylaxis only in those positive to MDM [5]. Furthermore, because of the increase in responses to side-chain-specific determinants that are negative to PPL, the number of cases of anaphylaxis has increased and this entity now predominates over urticaria. Early studies suggested that PPL-positive patients developed milder responses because they also had blocking IgG anti-BPO antibodies. However, this has not been confirmed [5]. Patients with immediate allergic reactions to BL can be classified into at least two groups: selective responders and cross-reactors. In the former group, the patient is diagnosed as having a reaction to one BL, by skin test or a positive drug provocation test, but with good tolerance to BP [6]. The advantage of identifying selective responders is because in a number of situations, such as cystic fibrosis, bone marrow transplantation, severe infections with specific microorganisms, syphilis and AIDS, the existence of an allergic reaction to just one BL does not exclude the use of other drugs from this family, including those of the same group such as BP, penicillin V, or cloxacillin in the case of selective responders to AX. Avoiding the whole group may imply more risks than benefits because of the potential adverse consequences resulting from the use of other antibiotic drugs that could be either more toxic, more expensive or more likely to induce bacterial resistance than BL. Finally, in those patients who have cross-reactions, the whole group of BL has to be avoided.

Diagnosis

The diagnostic approach for patients with IgE hypersensitivity to BL is based on the clinical history, although this is not always reliable because

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins)

197

very often patients do not provide enough information about a previous episode, healthcare personnel do not register properly the minimum information about past episodes, or the information has been registered by poorly trained personnel. The general guidelines for the evaluation of subjects with immediate hypersensitivity reactions to BL are described in detail in the position paper of the European Academy of Allergy and Immunology and are based on the performance of skin testing, in vitro testing and, if indicated, a drug provocation test (DPT) [14]. Skin Testing The usual procedure is to perform skin testing with PPL and MDM. In subjects with anaphylactic reactions to AX with negative skin tests to the penicillin determinants mentioned above, application of free AX determinants is required to diagnose an immediate response. This may also be useful even in subjects positive to PPL or MDM because AX is currently the drug that most often elicits an immediate allergic response. In 1990 a study showed that up to one third of all cases responded only to AX and that the role of AX in inducing positive skin responses by either prick or intradermal testing was more important than the role of MDM [15]. Further studies in a larger well-defined positive group showed that in populations where AX is the most important drug eliciting the reaction, the use of only PPL and MDM leaves over 60% of subjects undiagnosed [16]. This tendency is increasing and results from several European countries support this observation. Although AX is the offending drug in many studies, cloxacillin, ampicillin (AMP), different generation cephalosporins and others may also be involved. In all cases the recommended approach is to begin the study with PPL and MDM and, if positive, the diagnosis can be made, although as stated above the use of the culprit drug provides additional information about severity and cross-reactivity.

198

Some subjects may have a negative skin test though the clinical history is consistent with an immediate reaction, and a few subjects have had repeated episodes. Several reasons exist for negative skin tests in patients with a clear positive history. These include the use of an incomplete battery of determinants (including if necessary the culprit drug), the use of antihistamines, poorly trained personnel and the evaluation of patients with a long time between the drug reaction and the performance of the study [17]. In contrast to this recommendation is the statement that subjects who are skin-test negative to PPL and/or MDM can safely be given the drug, even if they have a clear positive history [1]. This is in disagreement with an increasing number of studies and daily clinical practice that point out that skin-test sensitivity is not optimal. The proposals for this practice are based on studies performed more than 20 years ago and even included many cases with very long intervals between the reaction and the study. Although skin testing is generally more sensitive than in vitro testing, mainly when dealing with protein allergens, skin testing in the case of haptens, such as BL, can be negative but specific IgE antibodies can be detected in vitro; thus, in vivo and in vitro evaluation can be complementary. Because in some instances, such as severe anaphylaxis, where skin testing with maximal concentrations is not exempt from risks, in vitro testing can be the first choice [18]. Furthermore, this technique can be considered in countries where the use of MDM and/or other penicillin determinants are not approved and in vitro testing enables the use of any available determinant providing that the laboratory methodology is reproducible and consistent. Unfortunately, sensitivity is not as good as with the in vivo test. In vitro Methods These are based on the determination of specific IgE antibodies and can detect serum IgE antibodies by immunoassay [19], IgE antibodies

Torres ⴢ Mayorga ⴢ Blanca

bound to the surface of basophils by flow cytometry [20, 21] or a functional assay measuring the basophil mediator release upon IgE-mediated activation [22]. Several of these techniques can be combined to improve sensitivity. Nevertheless, because of the cost and/or the lack of accessibility, they are not all performed in the diagnostic procedure. Immunoassay Immunoassay is the most widely used method, and was soon available after its first description. Throughout the years different techniques have existed: (a) radioimmunoassay, also known as the radioallergosorbent test (RAST); (b) enzymelinked immunosorbent assay (ELISA), and (c) fluoroenzyme immunoassay (CAP-FEIA). The latter is widely used and has the advantage that no radioactivity is used, and it can be performed with automated equipment, which facilitates its use. The RAST method is based on the use of activated cellulose discs as the solid phase, to which the hapten-carrier conjugate is bound. Critical factors for this test include the carrier used in the solid phase and the density of the hapten. The first approach was the conjugation of BP to HSA, but it is now recognized that PLL or aliphatic spacers that bind covalently to the penicillin are more appropriate [19]. Dendrimers (DH), to which BP, AX, cephalosporins or different determinants have been bound, have been used recently to improve the density of the hapten. Preliminary results using BPO and AXO determinants are promising [19]. Other RAST solid phases have often been used for the in vitro quantitation of IgE antibodies. These solid phases consist of Sepharose beads activated with cyanogen bromide, to which penicillin-HSA or penicillin-DH is bound [23]. Other methodologies use epoxy resin-activated Sepharose but these techniques are less feasible and do not seem to represent an advantage. Hydrophobic interactions, non-specific IgE binding and false-positive cases have been claimed to be major pitfalls of these techniques. However, because

no comparative studies with this technique between different groups have been carried out so far, a final evaluation regarding its use is still pending. Unless collaborative studies and position statements describing the comparison of these approaches are made, the answer to this question will remain open. Basophil Activation Test (BAT) This test is based on the capacity of basophils to be activated in the presence of the culprit hapten and afterwards express the CD63 marker on their membrane. The quantification of this marker plus the IgE on the surface of the basophil enables the identification in peripheral blood of these cells by flow cytometry. The sensitivity of BAT is 48.6%, with a specificity of 93%, but in cases with a negative skin test and RAST, the sensitivity is lower [20]. It is important to note that the sensitivity of CAP-FEIA does not match that of BAT, but the combination of both tests increases sensitivity [20, 21]. Basophil Mediator Release This test is based on the ability of basophils from peripheral blood to release sulfidopeptide leukotrienes (LTC4, LTD4, LTE4) after in vitro stimulation with allergens. This method has also been proposed for the diagnosis of IgE-mediated reactions to BL [22]. Further studies are required to establish its value. Competition Inhibition Assays Although not a first-line in vitro diagnostic technique, this test has proved useful to establish the final specificity of the antibody detected [11]. Most studies undertaken with BL use immunoassays, to which the hapten carrier is bound to the solid phase, and free or monomeric hapten-carrier conjugates are used in the fluid phase to compete for the specific IgE. The results should express the hapten concentration in mol/l producing 50% inhibition. The reason for using monomeric conjugates is because of the ease of determining

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins)

199

the hapten conjugated to the protein carrier, which is not precise when we use protein as the carrier. These assays enable study of the specificity and also cross-reactivity, as well as the relevant role of the different parts of the penicillin molecule in forming the epitope. Two main patterns of recognition have been established in IgE antibodies to penicillins: those that specifically recognize the side chain and those that recognize the nuclear part of the BL corresponding to a subject who cross-reacts to both drugs [11]. Further studies with cephalosporins have been reported using Sepharose as the solid phase [24]. In general, this may have the limitation that we attributed to this technique earlier in this chapter. One of the major problems is that it is not known how the hapten binds to the solid phase and how it is exposed to the antibody-binding site, and the different cephalosporins used in the inhibition were not conjugated to monomeric structures. Furthermore, no clinical information concerning skin testing or tolerance to different BL was provided in these studies. Drug Provocation Tests As stated above, skin testing does not have an optimal sensitivity compared to protein allergens. Contrary to what is recommended in many texts and clinical guidelines, an increasing proportion of subjects cannot be diagnosed, even if we use AX and the culprit drug in addition to PPL and MDM. The main reason for this is that BP is not the classical inducer and that other drugs probably generate determinants, at present unknown, and they are therefore not used for skin testing. Even the combination with in vitro tests will fail to rescue a sufficient number to give the combination of both techniques an acceptable sensitivity. Thus, the situation is such, that either we perform a DPT in certain cases or we may leave an important number of cases undiagnosed. This applies mainly to many persons with an unreliable history but who claim to be allergic to BL. Using a well-validated protocol [14], 330 persons

200

were evaluated because of a history of penicillin allergy and DPT was performed in those cases that were still skin-test and in-vitro negative after completion of the workup. A total of 89 cases (77% of the whole group) required DPT to establish the diagnosis [16]. Our group prefers to use the term ‘controlled administration’, which applies to those cases with no consistent history but who still require a complete study for diagnosis.

Natural Evolution and Resensitization

Immediate allergic reactions to BL do not last forever and IgE antibodies may disappear progressively over time. The natural course of this process first involves decreased serum levels of specific IgE antibodies followed by subsequent negativization of skin testing. However, the time between these two events can vary greatly from case to case. Unpublished data indicate that serum-specific IgE antibodies may disappear quite quickly but the skin tests remain positive for longer, although subjects change from positive to negative according to the specificity of IgE recognition. Few studies have been carried out to analyze this concept and most are based on retrospective data, with different sources of bias [25]. Prospective data from a group of subjects allergic to penicillin classified as cross-reactive or selective to AX showed that sensitivity decreased faster in the latter group, with all the subjects becoming skin test negative to AX at 5 years, but in the cross-reactors (positive to either PPL, MDM or both), 50% still remained positive at 5 years [17]. Accordingly, the selectivity of the response influences the rate of negativization and it is reasonable to assume that this also occurs with cephalosporins, though no studies have yet been reported with these compounds. It is important to mention that although this study has provided relevant information about the natural course of skin sensitivity, it does not imply that subjects can tolerate the drug.

Torres ⴢ Mayorga ⴢ Blanca

Less certainty exists regarding resensitization. Several studies have reported figures ranging from 1 to 16%, although in most studies the percentage is low or not significant. In one study, where repeated courses of penicillins were given to subjects suspected of being allergic to BL, the figures were very low. In another study where skin testing was done with no interference from simultaneous drug administration, a 2.5% rate of resensitization appeared after repeated skin testing [26]. In a further attempt to examine these observations in more detail, a prospective study was undertaken and a group of subjects with selective responses was followed up over several years with periodic administration of BP and penicillin V (PV). This group was compared to a positive control group to whom BP or PV was administered. In all cases a decline in IgE antibodies to AX occurred with no differences in the controls and no influence on the maintenance of IgE antibodies was observed. This implies that selective responders do not become cross-reactors. On the other hand, if these subjects have inadvertent contact with AX, they may occasionally become resensitized to this drug alone [27].

Cross-Reactivity

In clinical terms cross-reactivity implies that a subject allergic to one compound can respond to another with a similar structure. This is important in the area of BL antibiotics because they all share a common BL structure, although differences or similarities may exist in their side chains. Evaluation of cross-reactivity differs depending on whether we are dealing with immediate or non-immediate reactions; in addition, it varies according to the chemical structure, as mentioned above. For immediate reactions, the greater the difference in the chemical structure, the lesser the possibility of developing a cross-reactive response [10].

Cross-reactivity is not a bilateral phenomenon, with subjects primarily sensitized to penicillins equally recognizing cephalosporins or subjects first sensitized to cephalosporins recognizing penicillins with the same equivalence. Since the demonstration of side-chain-specific IgE antibodies, either cross-reactivity or selective responses must be considered within the penicillin group, within the cephalosporin group, between penicillins and cephalosporins, and between more distant groups, such as cephalosporins and monobactams (e.g., AZT and ceftazidime) [10]. In general, excluding selective responders to AX, subjects primarily sensitized to BP will also respond to AMP, AX, cloxacillin and the other compounds in this group. Crossreactivity between penicillins and cephalosporins has mainly been studied in patients with immediate allergic reactions to penicillins. Based on side-chain similarities, cases with a selective response to AX show a greater cross-reactivity when a cephalosporin with the same side chain is administered than with a non-side-chain-related cephalosporin. Thirty-eight percent of subjects with a selective response to AX developed cross-reactivity with cefadroxil, a cephalosporin that has an identical side chain [28]. Patients who are evaluated after an immediate allergic reaction to cephalosporins can show any of three different patterns of response: patients who respond to cephalosporin determinants and are also skin test positive to penicillin determinants, and thus may have a clinical reaction; patients who respond to cephalosporin determinants but are negative to penicillin determinants and tolerate BP, and patients who respond exclusively to the cephalosporin that was involved in the reaction [29, 30]. The explanation for these differences requires analysis of how the epitopes are formed, though this is beyond the scope of this chapter. Cross-reactivity between cephalosporins must be considered in terms of the similarity of the chemical structure of the R1 side chain. In this context, important cross-reactivity has been de-

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins)

201

tected between cefuroxime and ceftazidime, compounds that have an identical side chain at the R1 position, as well as between cefuroxime, ceftazidime and cefepime. Nevertheless, some patients have cross-reactivity between cephalosporins with different side chains, probably because the specific antibodies are directed to the common chemical structure shared by all of them. Regarding other BL, it is considered potentially harmful to administer imipenem to patients with IgE-mediated hypersensitivity to penicillins, although a recent study has indicated that very low cross-reactivity exist [31]. AZT shares the same side chain as ceftazidime and monoclonal antibodies enabled three patterns of recognition to be established, two recognizing both AZT and ceftazidime, indicating that the same acyl side chain is a relevant part of the epitope, and a third recognizing a new antigenic determinant and displaying broad cross-reactivity with several BL. Selective responses to AZT have been reported, although we are unaware of any evidence of specific IgE antibodies; on the other hand, cross-reactivity with ceftazidime has been observed. However, in all patients with penicillin allergy, good tolerance to AZT has been observed [32]. Selective hypersensitivity reactions to clavulanic acid have been reported, although no studies have yet shown the immunochemical characteristics of the epitope.

Treatment

Subjects who develop an allergic response to BL derivatives are recommended to avoid BL in the future. As recent studies indicate that third- and fourth-generation cephalosporins are well tolerated in patients who are primarily sensitized to PPL, and that BP can be administered to patients who are primarily sensitized to AX or cephalosporins, recommendations are not so restrictive. The question of safe tolerance within the penicillin group is more difficult to answer. At present, we know that selective responders will tolerate BP but no data are available concerning whether, in the long term, subsequent treatment will be tolerated. Regarding other BL, a number of studies indicate that good tolerance to monobactams exists in patients who are primarily sensitized to penicillin determinants. The only restriction concerns when a patient who was originally sensitized to ceftazidime is given AZT. In some cases the only alternative to giving a BL is to induce desensitization, consisting of the progressive administration of the drug, usually parentally, although the oral route has occasionally been used. Different protocols have been published elsewhere with varying success rates; while some patients developed good tolerance with no or only mild reactions, others had reactions and the protocol had to be stopped.

References 1 Weis ME, Adkinson NF: Immediate hypersensitivity reactions to penicillin and related antibiotics. Clin Allergy 1988;18:515–540. 2 Batchelor FR, Dewdney JM, Gazzard D: Penicillin allergy. The formation of the penicilloyl determinant. Nature 1965; 206:362–364. 3 Levine BB, Ovary Z: Studies of the mechanism of the formation of the penicillin antigen III. The N-(D-(benzyl penicilloyl)) group as an antigenic determinant responsible for hypersensitivity to penicillin G. J Exp Med 1961;114:875.

202

4 Batchelor FR, Dewdney JM, Weston RD, Wheeler AW: The immunogenicity of cephalosporin derivatives and their cross-reaction with penicillin. Immunology 1966;10:21–33. 5 Torres MJ, Mayorga C, Pamies R, Rodriguez JL, Juarez C, Romano A, Blanca M: Immunologic response to different determinants of BP, amoxicillin and ampicillin. Comparison between urticaria and anaphylactic shock. Allergy 1999;54:936–943. 6 Blanca M, Torres MJ, Mayorga C, Perez-Inestrosa E, Suau R, Montañez MI,

Juarez C: Immediate allergic reactions to betalatacms. Facts and contorversies. Curr Opin Allergy Clin Immunol 2004;4:261–266. 7 De Haan P, Jonge AJ, Verbrugge E, Boorsma DM: Three epitope-specific monoclonal antibodies against the hapten penicillin. Int Arch Allergy Appl Immunol 1985;76:42–46. 8 Mayorga C, Obispo T, Jimeno L, Blanca M, Moscoso del Prado J, Carreira J, Garcia JJ, Juarez C: Epitope mapping of betalactam antibiotics with the use of monoclonal antibodies. Toxicology 1995;97:225–234.

Torres ⴢ Mayorga ⴢ Blanca

9 Solley GO, Gleich GJ, Van Dellen RG: Penicillin allergy: clinical experience with a battery of skin test reagents. J Allergy Clin Immunol 1982;69:238– 244. 10 Dewdney JM: Immunology of the antibiotics; in Sela M (ed): The Antigens. New York. Academic Press, 1977, vol IV, pp 114–122. 11 Moreno F, Blanca M, Mayorga C, Terrados S, Moya C, Pérez E et al: Studies of the specificities of IgE antibodies found in sera from subjects with allergic reactions to penicillins. Int Arch Allergy Immunol 1995;108:74–81. 12 Sanchez-Sancho F, Perez-Inestrosa E, Suau R et al: Synthesis, characterization and immunochemical evaluation of cephalosporins antigenic determinants. J Mol Recognit 2003;16:148–156. 13 Terrados S, Blanca M, Garcia J, Vega JM, Torres MJ, Carmona MJ et al: Nonimmediate reactions to betalactams: prevalence and role of the different penicillins. Allergy 1995;50:563–567. 14 Torres MJ, Blanca M, de Weck A, Fernandez J, Demoly P, Romano A, Abereer W and ENDA and the EAACI Interest Group on Drug Hypersensitivity: Diagnosis of immediate allergic reactions to betalactam antibiotics. Allergy 2003;58:854–863. 15 Blanca M, Vega JM, Garcia J, Carmona MJ, Terrados S, Avila MJ: Allergy to amoxicillin with good tolerance to other penicillins. Study of the incidence in patients allergic to betalactams. Clin Exp Allergy 1990;20:475–481. 16 Torres MJ, Romano A, Mayorga C, Moya MC, Guzman AE, Reche M et al: Diagnostic evaluation of a large group of patients with immediate allergy to penicillins: the role of skin testing. Allergy 2001;56:850–856.

17 Blanca M, Torres MJ, Garcia JJ, Romano A, Mayorga C, de Ramon E et al: Natural evolution of skin test sensitivity in patients allergic to betalactam antibiotics. J Allergy Clin Immunol 1999;103:918–924. 18 Co Minh HB, Bousquet PJ, Fontaine C, Kvedariene V, Demoly P: Systemic reactions during skin tests with beta-lactams: a risk factor analysis. J Allergy Clin Immunol 2006;117:466–468. 19 Blanca M, Mayorga C, Perez E, Suau R, Juarez C, Vega JM et al: Determination of IgE antibodies to the benzylpenicilloyl determinant. A comparison between poly-L-lysine and human serum albumin as carriers. J Immunol Methods 1992;153:99–105. 20 Torres MJ, Padial A, Mayorga C, Fernandez T, Sanchez-Sabaté E, CornejoGarcia JA et al: The diagnostic interpretation of basophil activation test in immediate allergic reactions to betalactams. Clin Exp Allergy 2004;34: 1768–1755. 21 Sanz ML, Gamboa PM, Antepara I, Uasuf C, Vila L, Garcia-Aviles C et al: Flow cytometric basophil activation test by detection of CD63 expression in patients with immediate-type reactions to beta-lactam antibiotics. Clin Exp Allergy 2002;32:277–286. 22 Garcia-Aviles C, Sanz ML, Gamboa PM, Urrutia I, Anteapra I, Jauregui I, De Weck AL: Antigen-specific quantification of sulfidopeptideleukotrienes in patients allergic to betalactam antibiotics. J Invest Allergol Clin Immunol 2005;15:37–45. 23 Pham NH, Baldo BA: Beta-lactam drug allergens: fine structural recognition patterns of cephalosporin-reactive IgE antibodies. J Mol Recognit 1996;9:287– 296.

24 Romano A, Gueant-Rodriguez RM, Viola M, Pettinato R, Gueant JL: Crossreactivity and tolerability of cephalosporins in patients with immediate hypersensitivity to penicillins. Ann Intern Med 2004;141:16–22. 25 Sullivan TJ, Wedener HJ, Shatz GS, Yecies LD, Parker CW: Skin testing to detect penicillin allergy. J Allergy Clin Immunol 1981;68:171–180. 26 Nuguent JS, Quin JM, McGrath CM, Hrneir D et al: Determination of the incidence of sensitisation after skin testing. Ann Allergy Asthma Immunol 2003;90:398–403. 27 Blanca M, García J, Vega JM, Miranda A, Carmona MJ, Mayorga C et al: Anaphylaxis to penicillins after non-therapeutic exposure: an immunological investigation. Clin Exp Allergy 1996; 26:335–340. 28 Miranda A, Blanca M, Vega JM, Moreno F, Carmona MJ, Garcia JJ et al: Cross-reactivity between penicillin and a cephalosporin with the same side chain. J Allergy Clin Immunol 1996;98: 671–677. 29 Romano A, Mayorga C, Torres MJ, Artesani MC, Suau R, Pérez E et al: Immediate allergic reactions to cephalosporins: cross-reactivity and selective responses. J Allergy Clin Immunol 2000;106:1177–1183. 30 Antunez C, Blanca-Lopez N, Torres MJ, Mayorga C, Perez-Inestrosa E, Montanez MI et al: Immediate allergic reactions to cephalosporins. Evaluation of cross-reactivity with a panel of penicillins and cephalosporins. J Allergy Clin Immunol 2006;117:404–410. 31 Romano A, Viola M, Gueant-Rodriguez RM, Gaeta F, Pettinato R, Gueant JL: Imipenem in patients with immediate hypersensitivity to penicillins. N Engl J Med 2006;354:2835–2837. 32 Vega JM, Blanca M, García JJ, Miranda A, Carmona MJ, García A et al: Tolerance to aztreonam in patients allergic to beta-lactam antibiotics. Allergy 1991;46:196–202.

Dr. María José Torres Allergy Service and Research Unit for Allergic Diseases Carlos Haya Hospital, Plaza del Hospital Civil s/n ES–29009 Malaga (Spain) Tel. +34 95 103 0346, Fax +34 95 103 0302 E-Mail [email protected]

Urticaria and Anaphylaxis due to Betalactams (Penicillins and Cephalosporins)

203

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 204–215

Perioperative Anaphylaxis Pascale Dewachter Department of Anesthesiology, College of Physicians and Surgeons, Columbia University, New York, N.Y., USA

Abstract Anaphylaxis, occurring during the perioperative period is a clinical syndrome involving multiple organ systems, most commonly the skin, the cardiovascular and respiratory systems. The overall incidence is estimated at 1 in 10,000– 20,000 anesthetic procedures and 1 in 6,500 administrations of neuromuscular blocking agents (NMBAs). The most common drugs or substances involved are NMBAs, latex and antibiotics. Patients at risk have been identified. Some comorbidity factors may worsen the reaction. Prompt recognition of signs and symptoms is crucial because anaphylaxis may progress rapidly and is potentially life-threatening. The severe reactions may involve only one system, most commonly the cardiovascular system. Epinephrine is considered as the first-line treatment in guidelines on the perioperative management of anaphylaxis and, in most cases, an appropriate treatment restores the cardiovascular homeostasis. An allergological assessment is required to prove the immune mechanism, to identify the culprit drug or substance, and the cross-reactive molecules, especially for NMBAs, allowing preventive measures for future anesthetic procedures. Copyright © 2007 S. Karger AG, Basel

Definition

Anaphylaxis occurring during the perioperative period is a potential life-threatening condition, when multiple drugs or substances are used for anesthesia or surgical/radiological procedures. It

is a clinical syndrome involving multiple organ systems. The clinical manifestations are the consequence of the immediate release of preformed mediators from mast cells and basophils. In 2001 the European Academy of Allergology and Clinical Immunology (EAACI) Task Force [1] proposed a revised nomenclature for allergic and related reactions based on the present knowledge of the mechanisms which initiate and mediate allergic reactions. This nomenclature, published as the official EAACI Position Statement, was updated in 2003 by the Committee of the World Allergy Organization [2]. According to this nomenclature, immediate reactions with signs consistent with allergy are called immediate hypersensitivity reactions and are subdivided into allergic hypersensitivity reactions in which a specific immune mechanism can be demonstrated, and non-allergic hypersensitivity reactions in which an immune mechanism is ruled out. Hypersensitivity causes reproducible symptoms or signs initiated by exposure to a defined stimulus at a dose tolerated by normal subjects. Because the term anaphylaxis has been applied to different clinical entities and immunologic mechanisms, a clinical definition was proposed to define this nosologic entity as ‘a severe life-threatening generalized or systemic hy-

persensitivity reaction’ [1, 2]. But, hypotension and severe bronchospasm do not have to be present for a reaction to be classified as anaphylaxis [1]. Nevertheless, the former definition of Gell and Coombs of, ‘an immune mechanism reaction due to specific immunoglobulin E’, must be kept in mind.

in Australia [6] and 5% in Japan [7]. More recently, anaphylaxis was considered to be involved in 3% of the deaths totally or partially related to anesthesia in France [8], whereas 10% of anesthesiarelated immediate hypersensitivity reactions reported to the UK Medicines Control Agency were fatal [9].

Epidemiology

Which Drugs Cause It

A better knowledge of perioperative immediate hypersensitivity reactions, including allergic and non-allergic mechanisms, has been obtained through epidemiologic studies conducted during the last 25 years in different countries, such as Australia, France, New Zealand, the United Kingdom, and more recently in Scandinavia. These surveys enabled estimation of the incidence, understanding of the pathophysiological mechanisms, and identification of the responsible drugs or substances as well as the patients at risk of reactions. About 60% of the immediate hypersensitivity reactions occurring during anesthesia are IgE-mediated. The incidence of anaphylaxis occurring during anesthesia has been estimated at between 1 in 10,000 and 1 in 20,000 in Australia [3] and at 1 in 13,000 in France [4]. When the studied population was restricted to patients administered neuromuscular blocking agents (NMBAs), the estimated incidence was 1 in 6,500 in France [4] and 1 in 5,200 in Norway [5]. Nevertheless, the exact incidence is difficult to establish because not all cases are diagnosed, explored, published, reported to the Drug Safety Monitoring Authorities, or included in a national register. Actually, all reporting systems are voluntary and underreporting is thought to be common. In France, underreporting is estimated to be greater than 30% of cases. In other respects, the prevalence of these reactions is not known and nor is the morbidity, which is excluded from these epidemiologic studies. At the beginning of the 1990s, the mortality rate was estimated at 3%

Drugs or Substances Most Commonly Involved NMBAs, latex and antibiotics are the molecules most often found to be responsible for anaphylaxis during the perioperative period (table 1).

Perioperative Period and Anaphylaxis

Neuromuscular Blocking Agents NMBAs are the most common cause of anaphylaxis during anesthesia, with a range of 50–70% [4]. NMBAs are classified as depolarizing agents (succinylcholine, decamethonium) and nondepolarizing agents: benzylisoquinolinium compounds such as alcuronium, atracurium, cisatracurium, D -tubocurarine, gallamine and mivacurium, and amino steroids such as vecuronium, pancuronium and rocuronium. All NMBAs have been reported to cause anaphylaxis. In France [10] and Australia [3], in 12 or !50% of the cases, respectively, anaphylaxis to NMBAs occurred in patients with no previous exposure to any NMBA. In France, some authors proposed classifying the different NMBAs according to their risk of sensitization as high (succinylcholine, rocuronium), medium (pancuronium, vecuronium), and low (atracurium, cisatracurium) [10]. In Australia, Rose et al. [11] suggested that NMBAs with a high risk of causing anaphylaxis in the relaxant-sensitive population were alcuronium, succinylcholine, D -tubocurarine and decamethonium, intermediate-risk agents were atracurium, cisatracurium, rocuronium, mivacurium and gallamine, and low-risk agents were vecuronium and pancuronium. There is general agreement that the use of succinylcholine induces

205

Table 1. Frequent and less frequent elicitors of perioperative anaphylaxis Agents frequently involved NMBAs Depolarizing agents Succinylcholine Decamethonium Non-depolarizing agents Benzylisoquinolinium Alcuronium Atracurium Cis-atracurium D-Tubocurarine Gallamine Mivacurium Aminosteroids Pancuronium Rocuronium Vecuronium

the greatest risk of anaphylaxis. Nevertheless, the potential for an actual increased incidence of anaphylaxis to rocuronium is currently controversial. Reports from France [10] and Norway [12] suggest an increased frequency, leading the Norwegian Medicine Agency to publish an alert recommending withdrawal of rocuronium from routine practice and reserving its use for urgent intubations [12]. This trend was not observed in Australia [11] or in the USA [13], a discrepancy attributed to: (1) false-positive testing; (2) a rapid increase in the market share of the drug, and (3) biased reporting of adverse effects of new drugs [14]. Further efforts are therefore needed among anesthesiologists in the reporting of adverse drug effects.

Latex Antibiotics Penicillins, cephalosporins Agents less frequently involved Other anesthetic agents Hypnotics Etomidate Ketamine Propofol Thiopental Local anesthetics Amides Esters Opioids Other non-anesthetic agents Colloids Albumin Dextran Gelatin Hetastarch Aprotinin, protamine Antiseptics Chlorhexidine Povidone-iodine Dyes Isosulfan blue Patent blue Methylene blue Iodinated contrast agents Ionic Non-ionic

206

Latex Latex products are manufactured from a milky fluid derived from the rubber tree, Hevea brasiliensis. Latex is a component of most sterile and nonsterile gloves, and it is also contained in other products as face masks, tourniquets, urinal catheters, etc. Sensitization occurs through upper airways via latex particle aeroallergens carried by cornstarch glove powder or through cutaneous contact. Latex sensitization is frequent in the latex industry and healthcare workers, and in children with spina bifida and multiple operations. Atopic patients are at a higher risk of latex sensitization with an increased risk of latex allergy (36%) if compared to nonatopic patients (9.4%) [4]. Latex-induced anaphylaxis occurs in sensitized patients following exposure to latex gloves or devices. The incidence of anaphylaxis to latex has increased in the last few years. In the last prospective French report, latex was responsible for 22.3% of perioperative reactions [15] compared to 16.7% in the preceding report [10]. Thirty percent of these patients had a history of symptoms suggestive of latex sensitization which could have been detected before the reaction [15].

Dewachter

Antibiotics Antibiotics are commonly injected during the perioperative period as prophylaxis or treatment and are responsible of 15% of the anaphylactic reactions [15]. Penicillins and cephalosporins are mainly involved in France [10, 15] and cephalosporins in Australia [3]. Other Anesthetic Drugs Involved Hypnotics Hypnotics commonly used during the perioperative period are thiopental, propofol, etomidate and ketamine. The incidence of anaphylaxis to thiopental was estimated at 1 in 30,000 in the UK in 1982. Recently, thiopental was involved in less than 1% of allergic reactions in France [10]. Propofol was originally formulated with Cremophor EL which induced frequent anaphylactic reactions. This finding prompted a change in the formulation to a lipid vehicle. A few cases of anaphylaxis have been reported since then, representing less than 2.5% of the anaphylactic reactions in France [10, 15]. Anaphylactic reactions to etomidate and ketamine remain extremely rare. Local Anesthetics Local anesthetics include amides (bupivacaine, levobupivacaine, lidocaine, mepivacaine, prilocaine, ropivacaine) and esters (chloroprocaine, procaine, tetracaine, etc.). Anaphylactic reactions to local anesthetics are very rare when compared to their wide use. Less than 1% of reactions to local anesthetics proceed through an immunological mechanism. Amides have been involved in 0.6% of the perioperative reactions [15]. Anaphylactic reactions to esters are due to their common metabolite, p-aminobenzoic acid, leading to the contraindication of all the related drugs in case of documented anaphylaxis to one. There is no cross-reactivity between esters and amides. Allergic reactions to the preservatives (methylparaben, paraben) of local anesthetic solutions remain very rare.

Perioperative Period and Anaphylaxis

Opioids Anaphylactic reactions to opioids have been reported but remain very rare. Inhaled Anesthetics No anaphylactic reactions have been reported with inhaled agents as sevoflurane, isoflurane and desflurane, despite their wide use. Other Non-Anesthetic Drugs Involved Colloids Colloids (albumin, dextran, gelatin, hetastarch) are plasma expanders used to restore intravascular fluid volume. The overall frequency of anaphylaxis was estimated to be 0.033 and 0.22%, respectively, in Germany [16] and in France [4]. Gelatins (0.34%) and dextrans (0.27%) were more frequently involved than albumin (0.1%) or hetastarch (0.06%) [4]. In the last French report, hetastarches were responsible for 2.8% of the perioperative reactions [15]. Aprotinin Aprotinin is a polypeptide serine protease inhibitor used intravenously to decrease blood loss and blood transfusions or as a component of biological sealants. The frequency of anaphylactic reactions is estimated at 0.5–5.8% when used intravenously during cardiac surgery procedures, and at 5 for 100,000 applications when used as a biological sealant [4, 17]. Protamine Protamine sulfate is used to reverse the anticoagulant effects of heparin. A few cases of anaphylaxis have been reported. Antiseptics Life-threatening immediate hypersensitivity reactions have been described after insertion of central catheters impregnated with chlorhexidine or after intraurethral use or topical application of chlorhexidine. The incidence is not known. Only 7 documented cases of anaphylaxis have

207

been reported due to topical use of povidone-iodine during the perioperative period. The incidence is not known. Dyes Isosulfan blue is the 2,5-disulfonated isomer of patent blue dye and is used for lymphangiography and during sentinel lymph node biopsies. Isosulfan blue is used in the USA and patent blue in Europe. Anaphylactic reactions have been reported with patent blue dye, whereas the number of documented allergic reactions to isosulfan blue is more important. The incidence of allergic reactions to isosulfan blue ranges from 1 to 2% [17]. Patent blue was involved in less than 1% of perioperative anaphylactic reactions [15]. Only 1 documented case of anaphylaxis due to methylene blue dye, structurally less related to isosulfan/patent blue, has been reported. Iodinated Contrast Agents Iodinated contrast agents are widely used and may be injected during a surgical procedure. Furthermore, interventional radiologic procedures are in constant progression and a growing number of these is now performed under general anesthesia. Anaphylaxis to iodinated contrast agents has been reported. Nevertheless, its contribution to the incidence of perioperative anaphylaxis remains unknown [Christiansen, pp 233–241].

Pathophysiology

Allergic hypersensitivity reactions are due to a massive release of mediators from sensitized mast cells and basophils, as an inappropriate response to an allergen. Allergic hypersensitivity is initiated by an immune mechanism due in most cases to specific IgE antibodies. On initial exposure to an allergen in susceptible individuals, IgEs are produced and bound to high-affinity IgE receptors (Fc
RI) located in the membrane of tissue

208

mast cells and blood basophils, whereas lymphocytes, eosinophils and platelets bind them via low-affinity receptors (Fc
RII). On re-exposure, the multimeric allergen cross-links specifically two corresponding IgE receptors, creating a bridge between two IgEs. This initiates a signal transduction cascade, which increases the intracellular calcium concentration, and the immediate release of preformed mediators such as histamine, proteases (e.g. tryptase) and proteoglycans (e.g. heparin) in tissues and blood. Phospholipid metabolism then leads to the synthesis of newly synthesized mediators such as leukotrienes (LTC4, LTD4, LTE4), prostaglandins (PGD2), thromboxane (TXA2) and platelet-activating factor released soon after, and different cytokines released later on. A very small amount of allergen is sufficient for the cells to react. The molecules stored in the cell granules (histamine, tryptase, etc.) are immediately released in tissues and blood. Histamine is a potent vasoactive mediator with a wide variety of actions through its receptors H1, H2, H3 and H4 located in the skin, gastrointestinal tract, heart, vascular bed, bronchial smooth muscle: vasodilation and increase in capillary permeability, smooth muscle constriction, direct cardiac effect, and nerve end stimulation. The corresponding signs are erythema, edema, arterial hypotension, bronchial and gastrointestinal constriction, tachycardia and pruritus. Protease prevents local coagulation and degrades bronchodilating peptides and therefore is a potential bronchoconstrictor. PGD2 and LTC4 receptors are present in the skin, bronchial smooth muscle and vascular bed and cause bronchoconstriction and increased vascular permeability whereas platelet-activating factor acts as a vasoactive mediator. These different signs, associated or not, belong to the immediate allergic hypersensitivity syndrome. Nevertheless, the mechanisms of sensitization in patients who experienced anaphylactic reactions at the first encounter, especially with NMBAs, remain unknown.

Dewachter

Clinical Signs and Symptoms

Anaphylaxis is generally an unanticipated reaction, often explosive in onset. Prompt recognition of signs and symptoms is crucial because anaphylaxis may progress in a few minutes and is potentially life-threatening. Reactions may occur at any time during the perioperative period but are mostly observed soon after induction [4]. Clinical symptoms usually appear within minutes after intravenous injections of anesthetic drugs or antibiotics. Latex-induced reactions are usually delayed 30–60 min after beginning the surgical procedure but might occur immediately [18]. The clinical syndrome associated with an anaphylactic reaction is variable, both in its features and severity. The patterns of target organ involvement commonly concern the skin, the cardiovascular and respiratory systems, the gastrointestinal tract. During anesthesia, the reported clinical features, whether or not associated, are cardiovascular symptoms (bradycardia, tachycardia, hypotension, cardiovascular collapse, cardiac arrest), bronchospasm and cutaneous-mucous signs (erythema, urticaria, angioedema) [10]. The most commonly reported initial clinical features are lack of pulse and/or lung insufflation difficulty due to severe bronchospasm and/or desaturation [3]. Cardiovascular signs are present in 79% of the cases, respiratory signs in 40%, and cutaneous signs in 66% [15]. The most severe reactions may involve only one system, most commonly the cardiovascular system with cardiovascular collapse or cardiac arrest [3, 10]. Potentially life-threatening recurrences of anaphylactic reactions with a delay of up to 8 h have been observed in 28% of patients who responded to initial therapy, indicating that patients should be followed carefully after apparent remission of anaphylaxis [19]. The literature provides three predictive criteria of the severity of the ongoing reaction: (1) The more rapidly the anaphylaxis occurs after exposure to the allergen, the more likely the reaction is to be severe and potentially

Perioperative Period and Anaphylaxis

Table 2. Clinical severity scale of immediate reactions according to Ring and Messmer [16] Grade

Clinical signs

I

Cutaneous-mucous signs

II

Cutaneous-mucous signs Cardiovascular signs (tachycardia, hypotension) Respiratory signs

III

Cardiovascular collapse Bronchospasm

IV

Cardiac arrest

life-threatening [18]. (2) Urticaria and angioedema are the most common manifestations of anaphylaxis but might be delayed or absent in rapidly progressive anaphylaxis [18], appearing secondarily when blood pressure is restored by therapy. The subcutaneous vascular bed is susceptible to vasoconstrictive influences especially during anaphylaxis, the vascular bed being the first compromised when circulatory homeostasis is threatened. (3) Sinus bradycardia is seen with major hypovolemia and in this case bradycardia may be considered as a mechanism that allows the ventricles to fill adequately before they start to contract again. Additionally, anaphylaxis to NMBAs or antibiotics seems to be more severe than with latex [10]. The presumptive diagnosis of anaphylaxis is initially based on a detailed description of the clinical episode. A clinical scale was proposed by Ring et al. [16] to quantify the severity of anaphylactoid reactions due to colloids (table 2). It is presently used in France during immediate hypersensitivity reactions occurring during anesthesia [10]. This classification does not take into account the pathophysiological mechanism but is appropriate for immediate reactions occurring during anesthesia. Typically, this classification details four grades of increased severity. Generally, grades I and II are not life-threat-

209

ening conditions whereas grades III and IV correspond to emergency situations urging prompt treatment. The clinical diagnosis of anaphylaxis is not always easy, as cutaneous signs are not always detected because patients are under drapes or in the darkness of the operating room sometimes needed for the surgical procedure. Therefore, bronchospasm and cardiovascular collapse may be the first apparent clinical signs, although cardiovascular collapse may be recognized only with a delay in absence of an invasive arterial pressure monitoring. In addition, when clinical features occur as an isolated clinical symptom (e.g. cardiovascular collapse, bronchospasm), anaphylaxis may be misdiagnosed because many other clinical conditions can present identical manifestations occurring during anesthesia. Under such circumstances, the lack of proper diagnosis and appropriate allergological assessment could lead to fatal re-exposure [4].

chestnut, buckwheat, etc., because of the high rate of cross-reactivity with latex. In contrast, patients who are atopic or those who are allergic to a drug or product that is not likely to be used during the perioperative period are not to be considered at risk for anaphylaxis during anesthesia.

Comorbidity Factors

Asthma There are no epidemiologic studies indicating that anaphylaxis occurs more frequently in asthmatic patients. Nevertheless, asthma especially, if not optimally controlled, is an important risk factor for death from anaphylaxis [21]. Cardiovascular Disease Concurrent underlying cardiovascular disease such as ischemic cardiomyopathy or cardiac insufficiency might be aggravated by the occurrence of anaphylactic reactions.

Risk Factors

Patients at risk of perioperative anaphylaxis have been identified [20] and are as follows: • Patients who are allergic to one of the drugs or products likely to be administered or used during anesthesia and for which the diagnosis was established by a previous allergy investigation; • Patients who have shown clinical signs suggesting an allergic reaction during a previous anesthesia; • Patients who have experienced clinical manifestations of allergy when exposed to latex, whatever the circumstances in which this occurred; • Children who have multiple operations, especially those with spina bifida, because of the high incidence of sensitization to latex and of the high incidence of anaphylactic shock caused by latex in such patients ; • Patients who have experienced clinical manifestations of allergy to avocado, kiwi, banana,

210

-Blockers and Other Drugs There are no epidemiologic studies indicating that anaphylaxis occurs more frequently in patients receiving -blockers [18]. But epinephrine administered to patients taking -blockers may be ineffective. Under this circumstance, glucagon could be necessary. Concurrent administration of angiotensin-converting enzyme and to a lesser extent angiotensin II blockers, might also interfere with the hemodynamic response to the treatment [21].

Diagnosis

The initial diagnosis of an anaphylactic reaction occurring during anesthesia is only presumptive and based on the clinical signs according to the clinical severity scale of Ring and Messmer [16] and the treatment of the reaction. Therefore this

Dewachter

diagnosis should be confirmed by an allergological assessment including biochemical and cutaneous tests. Typical Laboratory Changes Biochemical tests are available either in vivo or in vitro. Early increased plasma histamine confirms in vivo histamine release. Histamine concentrations in plasma can be reliably measured by immunoassay. During allergic reactions to drugs, histamine concentrations are immediately maximal and decrease thereafter with a half-life of about 20 min [22]. Tryptase concentrations, measured in serum by immunoassay, reach a peak between 15 min and 1 h and decrease slowly, with a half-life of 90 min [22]. An increase in histamine concentrations indicates activation of mast cells and/or basophils, whereas an increase in tryptase is specific for mast cell activation, as basophils have low tryptase levels. Histamine and tryptase measurements have been recommended for the diagnosis of allergic reactions during anesthesia [20]. In vitro tests available in clinical practice detect the presence of IgE antibodies by the radioallergosorbent test (RAST). The presence of NMBAspecific IgEs in serum was initially demonstrated using NMBA coupled to epoxy sepharose (CAPRAST). The sensitivity of this test, however, is limited and therefore was improved by coupling an analog of choline to sepharose, or p-aminophenylphosphoryl-choline on agarose [23]. These in vitro tests may help to confirm the diagnosis in patients experiencing an anaphylactic reaction to a NMBA, or to identify the culprit drug in patients who cannot be skin-tested. Placing the tested serum in the presence of the soluble form of the culprit molecule controls the inhibition of specific IgE binding [20]. The concordance rate with skin tests is about 80% in most cases [4]. Specific IgEs (radioimmunoassay) against thiopental, propofol, morphine and latex have also been identified in the serum of patients experiencing adverse reactions during anesthesia. The

Perioperative Period and Anaphylaxis

RAST available for latex is less sensitive than the skin-prick test with latex, being positive only in 50–70% of the cases [18]. Actually, the indications for specific IgE assays are restricted to the diagnosis of anaphylaxis to NMBAs, thiopental and latex [20]. Several other tests have been proposed for indirect detection of specific IgE to anesthetic drugs. The basophile histamine release test is reliable for NMBAs and its sensitivity is about 60–70%. This technique might be required when cross-reactivity among NMBAs is investigated for NMBA reintroduction in patients with documented anaphylaxis to NMBAs [20]. The study of basophil activation in the presence of the allergen is assessed by the expression of CD63 on the plasma membrane using flow cytometry. Its sensitivity and specificity are estimated to be 64 and 93%, respectively. Nevertheless, further investigations are needed to assess the value of this method in the diagnosis of anaphylaxis to NMBAs [4]. Skin Testing to Identify the Responsible Drugs Skin tests remain the gold standard for the detection of IgE-mediated reactions. Skin tests should be performed preferentially after a delay of 6 weeks after the reaction because of mast-cell depletion. When they are carried out earlier, a risk of false-negative results exists and therefore only positive skin tests should be considered relevant. Immediate hypersensitivity reactions should be investigated by immediate reading of skin tests. Skin test reactivity should first be verified by negative and positive controls. Skin tests must include all drugs injected just before the reaction, as well as latex, in order to respect the chronology of the event. Investigation of anesthetic agents is performed by prick tests and/or intradermal tests using commercial solutions, undiluted or diluted in aqueous phenol. In France, the different concentrations for anesthetic agents which are normally non-reactive in the practice of skin tests have been recommended (table 3) [20]. In France, investigation of latex is performed by prick test

211

Table 3. Concentrations of anaesthetic agents normally non-reactive during skin tests according to the Société Française d’Anesthésie et de Réanimation [20] (reproduced with permission) Drugs

Prick tests mg
ml–1

INN

C,

Atracurium Cis-atracurium Mivacurium Pancuronium Rocuronium Suxamethonium Vecuronium Etomidate Midazolam Propofol Thiopental Alfentanil Fentanyl Morphine Remifentanil Sufentanil Bupivacaine Lidocaine Mepivacaine Ropivacaine

10 2 2 2 10 50 4 2 5 10 25 0.5 0.05 10 0.05 0.005 2.5 10 10 2

Intradermal tests mg
ml--1

dilution

MC,

1/10 undiluted 1/10 undiluted undiluted 1/5 undiluted undiluted undiluted undiluted undiluted undiluted 1/10 undiluted undiluted undiluted undiluted undiluted undiluted undiluted

1 2 0.2 2 10 10 4 2 5 10 25 0.5 0.05 1 0.05 0.005 2.5 10 10 2

dilution

MC, g
ml--1

1/1,000 1/100 1/1,000 1/10 1/100 1/500 1/10 1/10 1/10 1/10 1/10 1/10 1/10 1/1,000 1/10 1/10 1/10 1/10 1/10 1/10

10 20 2 200 100 100 400 200 500 1,000 2,500 50 5 10 5 0.5 250 1,000 1,000 200

INN = International Nonproprietary Name; C = concentration; MC = maximal concentration.

using 2 different commercial extracts [20] which are not available in the USA [17]. With the aid of prick tests, intradermal tests are first realized with a 1/1,000 dilution of the agent, except with NMBAs and morphine for which a 1/10,000 dilution is recommended. When the first intradermal test is negative, then successive dilutions are used with 15- to 20-min intervals between each intradermal test without exceeding the maximal concentration. The sensitivity of skin tests for NMBAs is approximately 94% [20]. The sensitivity of skin tests with latex is excellent. The sensitivity of skin tests with penicillins is approximately 97% [18]. The sensitivity of intradermal test with morphine is good, while specificity is not because morphine induces histamine release. False-positive skin tests (intradermal test) are de-

212

scribed when insufficiently diluted morphine is tested (a 1/10,000 dilution is recommended and should not be exceeded). Finally, skin tests should be performed by trained physicians in a setting with adequate resuscitative equipment. A consultation performed by a team including 2 physicians with expertise in anesthesia and in allergy would be highly desirable in order to optimize the proper character of the diagnosis. Additionally, patients must be informed of the investigations and should possess a detailed report before eventual future anesthesias. Problems of Cross-Reactivity When the prick test and/or the intradermal test with a NMBA is positive, investigation for cross-

Dewachter

reactivity with the other commercialized NMBAs should be done by intradermal tests. The maximal concentration recommended should not be exceeded (table 3) [20]. Therefore a NMBA-negative intradermal test might be proposed before further re-introduction. Nevertheless, this recommendation does not guarantee absolute prevention of a further adverse reaction. Challenge Tests Challenge tests are limited to local anesthetics and latex in the case of negative skin tests. The tests are considered negative if no adverse reaction occurs within 30 min after subcutaneously injecting the local anesthetic or wearing a latex glove [17, 20]. In conclusion, when the clinical presentation is suggestive of an immediate allergic reaction, skin tests performed with the alleged drug or substance confirm the diagnosis. However, a negative IgE test (RAST) with NMBAs does not rule out the involvement of a NMBA when the clinical history and the skin tests are contributive to the diagnosis. In contrast, when the observed data remain inconclusive (negative skin tests and biology), it is mandatory to contact the anesthesiologist in order to analyze the clinical events and consider all data that may eventually lead to an alternate diagnosis.

Clinical Conclusions: What Can Be Done to Avoid These Side Effects and What Can Be Done to Recognize Them Earlier?

There is no role for systematic pre-operative screening in the general population, one does not know the predictive value of skin tests for the occurrence of perioperative anaphylaxis. Additionally, a pre-medication with an anti-H1-receptor agonist alone, or with an anti-H2-receptor agonist alone or with a corticosteroid alone or an association of two or more of these drugs has not proven to be able to prevent an anaphylactic reac-

Perioperative Period and Anaphylaxis

tion [18, 20]. Therefore, the best pre-medication consists actually of performing an allergological assessment in patients who experienced an immediate reaction during the perioperative period, in order to identify the culprit drug/substance and to avoid it definitely [20]. A reduction in the incidence of sensitization to latex can be achieved by using non-powdered latex gloves, whereas a latex-free environment is effective for the prevention of an anaphylactic reaction in patients allergic to latex [17, 18, 20]. Finally, anesthesia crisis resource management is a concept that has been used to train anesthetists in full-scale simulators in an attempt to improve human performance during anaphylaxis occurring during anesthesia.

Special Therapy

Management of perioperative anaphylaxis consists of withdrawing the culprit drug when possible, immediate discontinuation (when possible) of anesthetic drugs when it occurs during the anesthetic induction, or abbreviating the surgical procedure when it occurs during surgery. Since the different drugs are delivered by the intravenous route, reactions generally appear very rapidly. Treatment should not be started by intramuscular epinephrine, but by intravenous use according to the different guidelines concerning the management of perioperative anaphylaxis [9, 20, 24] and since an intravenous line is immediately available. In contrast, the officially recommended treatment by allergologist guidelines is epinephrine (0.3–0.5 mg) injected intramuscularly. Early administration of epinephrine by the intravenous route together with vascular expansion initiated with crystalloids are the key points of the treatment [9, 18, 20, 24]. Epinephrine is considered the first line of treatment in most guidelines on perioperative management of anaphylaxis [9, 18, 20, 24], and there is no absolute contraindication to its administration during

213

anaphylaxis [18]. Its beneficial effects in anaphylactic shock are mediated by the 1-adrenergic effects which increase the left ventricular preload by reducing venous capacitance, while -adrenergic effects reverse bronchoconstriction and increase cardiac inotropy and chronotropy. Evidence in the literature suggests that a poor outcome during anaphylactic shock is associated with late administration of epinephrine, but also when epinephrine has been given inadequately or in excessive doses [20]. Epinephrine should not be injected during grade-I reactions whereas a titrated bolus (10–20 g) might be necessary under certain circumstances during grade-II reactions. A titrated intravenous bolus administration of epinephrine (100–200 g) is required in gradeIII reactions, which should be renewed every 1 or 2 min if necessary according to the hemodynamic response and followed by a continuous infusion (1–4 g
min–1) to prevent the need of repeat injections [24]. Grade-IV reactions require cardiopulmonary resuscitation, volume administration and high-dose epinephrine [20, 24]. A commonly used sequence is 1–3 mg intravenously (3 min), then 3–5 mg intravenously (3 min) followed by a continuous infusion (4–10 g
min–1) [24]. In most cases, appropriate treatment restores the cardiovascular homeostasis. But, epinephrine has been reported to be ineffective despite an appropriate dosage in some cases of anaphylactic shock, even in previously healthy patients. In some anaphylactic shocks refractory to catecholamines, the use of -agonists such as metaraminol [9] or methoxamine produced rapid return of spontaneous circulation. Recently, experimental work has provided arguments for the possible use of arginine vasopressin (AVP) as therapy for anaphylactic shock [25]. In addition, the most recent American guidelines on anaphylaxis [24] and several case reports suggest that adding AVP to standard therapy might be considered a potential therapeutic approach to anaphylactic shock when catecholamines are not ef-

214

ficient. Nevertheless, further investigations are needed to recommend AVP under such clinical circumstances. Bronchospasm is treated with inhaled 2-agonists (salbutamol). In case of refractory bronchospasm or immediate increased severity, intravenous injection of a bolus dose of salbutamol (100– 200 g) is recommended and continuous infusion (5–25 g
min–1) should be considered [9, 20, 24]. Glucocorticoids should also be considered and injected in case of bronchospasm. But when cardiovascular collapse and bronchospasm occur together, epinephrine should be injected as first-line therapy to rapidly correct the cardiovascular collapse. Corticosteroids are also recommended in the management of anaphylaxis [9, 24], nevertheless their effects have never been evaluated in placebo-controlled trials [18]. In a retrospective study analyzing recurrent episodes of anaphylaxis, corticosteroids did not prevent biphasic anaphylaxis [19].

Conclusion

Perioperative anaphylaxis is a potentially lifethreatening clinical condition which remains underestimated because it is underreported. NMBAs, latex and antibiotics are the most common drugs involved but other drugs used for anesthesia or the surgical procedure might be involved. In order to prevent further reactions, an allergological assessment is required to prove the immune mechanism, to identify the culprit drug/ substance, and to avoid it definitely.

Dewachter

References 1 Johansson SGO, Hourihane J, Bousquet J, Bruijnzeel-Koomen C, Dreborg S, Haahtela T, Kowalski ML, Mygind N, Ring J, van Cauwenberge P, van HageHamsten M, Wüthrich B: A revised nomenclature for allergy. An EAACI position statement from the EAACI nomenclature task force. Allergy 2001; 56:813–824. 2 Johansson SGO, Bieber T, Dahl R, Friedmann PS, Lanier BQ, Lockey RF, Motala C, Martell JAO, Platt-Mills TAE, Ring J, Thien F, van Cauwenberge P, Williams HC: Revised nomenclature for allergy for global use: report of the Nomenclature Review Committee of the World Allergy Organization, October 2003. J Allergy Clin Immunol 2004; 113:832–836. 3 Whittington T, Fisher M: Anaphylactic and anaphylactoid reactions. Baillieres Clin Anaesthesiol 1998;12:301–323. 4 Mertes PM, Laxenaire MC: Allergic reactions occurring during anaesthesia. Eur J Anaesthesiol 2002;19:240– 262. 5 Harboe T, Guttormsen AB, Irgens A, Dybendal T, Florvaag E: Anaphylaxis during anesthesia in Norway: a 6-year single-center follow-up study. Anesthesiology 2005;102:897–903. 6 Currie M, Webb R, Williamson J, Russell W, Mackay P: The Australian Incident Monitoring Study. Clinical anaphylaxis: an analysis of 2000 incident reports. Anaesth Intensive Care 1993; 21:621–625. 7 Mitsuhata H, Matsumoto S, Hasegawa J: The epidemiology and clinical features of anaphylactic and anaphylactoid reactions in the perioperative period in Japan. Masui 1992;41: 1664–1669.

8 Lienhart A, Auroy Y, Pequignot F, Benhamou D, Warszawski J, Bovet M, Jougla E: Survey of anaesthesia-related mortality in France. Anesthesiology 2006; 105:1085–1097. 9 http://www.aagbi.org/pdf/Anaphylaxis.pdf 10 Mertes PM, Laxenaire MC, Alla F; Groupe d’Etudes des Réactions Anaphylactoïdes Peranesthésiques: Anaphylactic and anaphylactoid reactions occurring during anesthesia in France in 1999–2000. Anesthesiology 2003;99: 536–545. 11 Rose M, Fisher M: Rocuronium: high risk for anaphylaxis? Br J Anaesth 2001;86:678–682. 12 Guttormsen AB: Allergic reactions during anaesthesia – increased attention to the problem in Denmark and Norway. Acta Anaesthesiol Scand 2001;45:1189– 1190. 13 Bhanaker SM, O’Donnell JT, Salemi JR, Bishop MJ: The risk of anaphylactic reactions to rocuronium in the United States is comparable to that of vecuronium: an analysis of food and drug administration reporting of adverse events. Anesth Analg 2005;101:819–822. 14 Laake JH, Rottingen JA: Rocuronium and anaphylaxis – a statistical challenge. Acta Anaesthesiol Scand 2001; 45:1196–1203. 15 Mertes PM, Laxenaire MC; Membres du GERAP: Epidémiologie des réactions anaphylactiques et anaphylactoïdes peranesthésiques en France. Septième enquête multicentrique (Janvier 2001–Décembre 2002). Ann Fr Anesth Reanim 2004;23:1133–1143. 16 Ring J, Messmer K: Incidence and severity of anaphylactoid reactions to colloid volume substitutes. Lancet 1977;1:466–469. 17 Hepner DL, Castells MC: Anaphylaxis during the perioperative period. Anesth Analg 2003;97:1381–1395.

18 Joint Task Force on Practice Parameters; American Academy of Allergy, Asthma and Immunology; American College of Allergy, Asthma and Immunology; Joint Council of Allergy, Asthma and Immunology: The diagnosis and management of anaphylaxis: an updated practice parameter. J Allergy Clin Immunol 2005; 115(suppl 2):S483–S523. 19 Stark BJ, Sullivan TJ: Biphasic and protracted anaphylaxis. J Allergy Clin Immunol 1986;78:76–83. 20 Société Française d’Anesthésie et de Réanimation: Reducing the risk of anaphylaxis during anaesthesia. Abbreviated text. Ann Fr Anesth Reanim 2002; 21(suppl 1):7s–23s. 21 Simons FER: Anaphylaxis, killer allergy: Long-term management in the community. J Allergy Clin Immunol 2006;117:367–377. 22 Laroche D, Vergnaud MC, Sillard B, Soufarapis H, Bricard H: Biochemical markers of anaphylactoid reactions to drugs. Comparison of plasma histamine and tryptase. Anesthesiology 1991;75:945–949. 23 Guilloux L, Ricard-Blum S, Ville G, Motin J: A new radioimmunoassay using a commercially available solid support for the detection of IgE antibodies against muscle relaxants. J Allergy Clin Immunol 1992;90:153–159. 24 American Heart Association: Guidelines for cardiopulmonary resuscitation. Part 10.6: anaphylaxis. Circulation 2005;112:IV143–IV145. 25 Dewachter P, Jouan-Hureaux V, Lartaud I, Bello G, de Talancé N, Longrois D, Mertes PM: Comparison of arginine vasopressin, terlipressin or epinephrine to correct hypotension in a model of anaphylactic shock in anesthetized Brown Norway rats. Anesthesiology 2006;104:734–741.

Dr. Pascale Dewachter Department of Anesthesiology, College of Physicians and Surgeons Columbia University, 168th Street, Black Building, Room 713 New York, NY 10032 (USA) Tel. +1 212 305 6372, Fax +1 212 543 8832 E-Mail [email protected]

Perioperative Period and Anaphylaxis

215

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 216–232

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics Paolo Campi a  Mariangela Manfredi b  Maurizio Severino a a Allergy Clinic and b Allergy Laboratory, Azienda Sanitaria di Firenze, Nuovo Ospedale San Giovanni di Dio, Florence, Italy

Abstract IgE-mediated allergic reactions to pyrazolones, quinolones and other non-
-lactam antibiotics are reviewed based on the literature and personal data. Evidence of immediatetype hypersensitivity comes from prick tests and intradermal tests upon immediate reading with positive results, provided that a non-irritating concentration has been used. Skin tests in control subjects are necessary to evaluate the non-irritating concentration. Serum-specific IgE to these molecules has also been demonstrated and the different in vitro methods are discussed. One of the most sensitive solid phase assays is to couple the drug to epoxy-activated Sepharose TM 6B (Amersham Biosciences AB, Uppsala). Management of these reactions through the various steps of history, skin tests, in vitro tests, provocation tests and desensitization is discussed, also following the indications of the European consensus papers published by the European Network for Drug Allergy. Copyright © 2007 S. Karger AG, Basel

Introduction

Most allergic-type hypersensitivity reactions to drugs, e.g. those sustained by an immunological mechanism, are IgE-mediated (type I according to Gell and Coombs) or cell-mediated (type IV). The IgE reactions present with one or more of the following symptoms: itching, urticaria, angio-

edema, rhinitis, conjunctivitis, bronchospasm, abdominal cramps and shock. They are defined as ‘immediate’ because they often appear within a few minutes of taking the drug, usually within 1 h. They can also appear later during the course of therapy, mainly in cases of oral consumption.

Non-
-Lactam Antibiotics

Incidence The Boston Collaborative Surveillance Program [1] showed that 300 of 14,111 inpatients (2.12%) had adverse cutaneous reactions during antibiotic therapy; with non-
-lactam antibiotics 49 reactions were observed in 5,769 patients (0.85%). In another Boston Collaborative Surveillance Program study [2], in the 3,968 patients treated there were 173 adverse reactions to antibiotics (4.36%), and 134 seemed immune mediated. Non-
-lactam antibiotics have provoked apparently allergic reactions in 39 of 1,675 patients treated (2.33%) and therefore with a greater frequency than
-lactams (table 1). From our outpatient case histories of adverse reactions to drugs at Nuovo Ospedale San Gio-

Table 1. Incidence of cutaneous adverse reactions (Boston Collaborative Surveillance Program)

Study

Number adverse reactions

Number patients treated

Incidence %

Systemic antibiotics [1] Non-
-lactam antibiotics [1] Systemic antibiotics [2] Non-
-lactam antibiotics [2]

300 49 134 39

14,111 5,769 3,968 1,674

2.1 0.8 3.4 2.3

Table 2. Adverse reactions to drugs in 6,500 patients studied from 1988 to 2004 at Nuovo Ospedale San Giovanni di Dio in Florence Class
-Lactams ASA Other NSAIDs Pyrazolones Non-
-lactam antibiotics Co-trimoxazole Sulfonamides Local and general anesthetics Radiographic contrast media Insulin Other drugs Total

Reactions n

Incidence %

2,387 2,058 1,549 1,368 1,135 615 42 488 63 4 859

22.5 19.4 14.6 12.9 10.7 5.8 0.4 4.6 0.6 0.03 8.1

10,610

100

vanni di Dio in Florence, of 6,500 patients examined between 1988 and 2004, 1,792 had taken non-
-lactam antibiotics (43% of the antibiotics; table 2). The incidence of adverse reactions is not only dependent on the allergenic potential of the single molecule, but also on the consumption: this correlation becomes evident also from a comparison of the adverse reactions reported by our patients and the sale of antibiotics (table 3). Non-
lactam antibiotics constitute 36% of the systemic antibiotics sold in Italy. Even in this case the rate of adverse reactions is higher than that for
-lactams if it is correlated with consumption.The ex-

Table 3. Comparison between sale of antibiotics in Italy in 2003 and adverse reactions reported by 6,500 patients from 1988 to 2004 at Nuovo Ospedale San Giovanni di Dio in Florence Antibiotics

Sales in 2003, %

Adverse reactions reported, %


-Lactam Non-
-lactam

64 36

57 43

100

100

Total

amination of 2,224 cutaneous adverse reactions, gathered in 1996 and 1997 through a system of spontaneous observation in four Italian regions (Friuli Venezia Giulia, Lombardy, Sicily and Veneto) with a population of 20 million people, has correlated the incidence of reactions with the consumption of drugs [3]: for antibiotics the relationship between the incidence and consumption was highest for co-trimoxazole (n = 78), followed by cephalosporins (n = 68), fluoroquinolones (n = 65), penicillins (n = 45) and macrolides (n = 35). The adverse reactions to non-
-lactam antibiotics therefore constitute a problem of significant frequency. Skin Tests Reports of skin tests with positive outcome were reported only for some non-
-lactam antibiotics. There are few reports for each drug, with the ex-

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

217

ception of widely used molecules such as erythromycin and streptomycin, or drugs used topically, such as rifamycin SV and bacitracin. It is well known that this method of administration carries greater risk of sensitization, and for this reason
-lactams have not been used for topical application for many years. These data should lead to caution in the topical use of these drugs, especially polypeptide antibiotics such as bacitracin, which can provoke near-fatal reactions by cutaneous application, particularly to a wound. The bibliography shown in table 4 refers to cases of immediate adverse reactions in which the immunological mechanism type I has been proven by skin tests (prick and/or intradermal reactions with immediate reading), provided that the authors have defined the non-irritating concentrations [see chapter of Solensky, pp 404–412, table 1] in control subjects. Serum-Specific IgE Table 5 shows the non-
-lactam antibiotics for which the literature has shown serum-specific IgE: the greater sensitivity of sepharose compared to nitrocellulose as the solid phase for binding drugs with a low molecular level is to be noted. Already since 1983, Harle et al. [4] and Baldo and Harle [5] have devised a method to identify specific IgE for low molecular weight drugs using the epoxy-activated SepharoseTM 6B (Amersham Biosciences AB, Uppsala) as a solid phase with a spacer arm. This method was initially used to demonstrate serum IgE directed towards trimethoprim and sulfamethoxazole and later even towards other antibiotics such as quinolones [6– 8] and erythromycin [9]. It has been used by our group with excellent results for those
-lactam and non-
-lactam antibiotics for which trustworthy methodology is not available commercially, such as the Phadia CAP
test. We have been able to demonstrate serum IgE directed towards numerous cephalosporins and also rifamycin and vancomycin (table 5).

218

Quinolones

Recently our group has shown the presence of serum-specific IgE towards these antibiotics [6–8]. In recent years our attention has been drawn to various cases of immediate-type reactions to quinolones. The consumption of these antibiotics is increasing because the new fluoroquinolones have a broad spectrum of action, as opposed to the first quinolones which were used principally for infections of the urinary tract. The literature has described numerous adverse reactions of a presumable immunological nature and among these the most numerous are those of the immediate type. With regard to the incidence of adverse reactions of the anaphylactic type, it is interesting to note the case histories reported by Burke and Burne [10]: of 3,200 university students who took a single dose of 500 mg ciprofloxacin as prophylaxis against meningococcal disease during an epidemic, 3 cases of an immediate anaphylactoid reaction were reported, a rate of 1:1,000 which is much higher than the 1:100,000 quoted by Davis et al. [11] who reported 12 cases in a population of 972,000. Some authors reported positive results with quinolone skin tests (prick and intradermal reactions with immediate reading), but others found that skin tests are not reliable, being positive even in the controls, or they obtained negative skin tests and positive provocation tests (table 6). These molecules have a high capacity to release histamine from mastocytes in a nonspecific manner, so that the skin test is irritating unless a very high dilution is used. In table 7 the results are reported of intradermal tests done in control subjects by different researchers in order to find the nonirritating concentration [12]. Only a few quinolones are available in vials to perform the intradermal test. The wide range of concentrations reported depends only in part on the different techniques used to perform the intradermal test. For ciprofloxacin (table 7), in particular Empedrad et al. [13] were not able to

Campi  Manfredi  Severino

Table 4. Evidence of cutaneous specific IgE by means of prick test or intradermal test upon immediate reading for non-
-lactam antibiotics and antimycotics Class and drug Macrolides Erythromycin Spiramycin Roxithromycin Fosfomycin Telithromycin Aminoglycosides Streptomycin Tobramycin Framycetin (neomycin B) Gentamicin Rifamycins Rifampicin Rifamycin SV Glycopeptides Vancomycin

References van Ketel WJ 1976, Lopez Serrano CL 1993, Pascual C 1995, Jorro G 1996, Goossens C 1997, Barbaud A 1998, Alvarez Fernandez JA 1998, Lammintausta K 2005 Halpern B 1967, Davies RJ 1975, Malet A 1992, Barbaud A 1998, Benahmed S 2004 Kruppa A 1998, Gurvinder SK 2002 Rosales MJ 1998 Pérez CI 2004 Chakravarty S 1961, Zelger J 1969, Maucher OM 1972, Tinkelman DG 1984, Fernandez E 1987, Chandler MJ 1992, Abeck D 1994, Perez R 1996, Iikura M 2002, Romano A 2002 Earl HS 1987, Schretlen-Doherty JS 1995 Agathos M 1980, Proebstle TM 1995 Schulze S 2003 Nessi R 1973, Cnudde F 1994, Dutau H 2000, Campi P and Severino M 2000 (personal data), Strauss RM 2001, Buergin S 2006 Grob JJ 1987, Mancuso G 1992, Cardot E 1995, Laxenaire MC 1996, Erel F 1998, Garcìa F 1999, Magnan A 1999, Scala E 2001, Ebo DG 2006

Teicoplanin

Lin FL 1990, Knudsen JD 1992, Anne S 1994, Campi P and Severino M 2000 (personal data), Hwu JJ 2005 Knudsen JD 1992

Tetracyclines Oxytetracycline Tetracycline

Fellner MJ 1965, Zelger J 1969 Menon MP 1977

Polypeptides Bacitracin

Polymyxin B

Comaish JS 1967, Roupe G 1969, Schechter JF 1984, Goh CL 1986, Palungwachira P 1991, Dick ED 1997, Lin RY 1998, Saryan JA 1998, Gall R 1999, Blas M 2000, Carver ED 2000, Antevil JL 2003, Namazy JA 2003, Freiler JF 2005 Lakin JD 1975

Imidazoles Ketoconazole Fluconazole Amphotericin B

Gonzales-Delgado P 1994 Neuhaus G 1991 Kemp SF 1995

Nitrofurans Nitrofurantoin

Tykal P 1972

Trimethoprim

Harle DG 1987, Alonso MD 1992, Cabanas R 1996, Bijl AMH 1998, Lammintausta K 2005

Sulfonamides Sulfamethoxazole

Gruchalla RS 1991

Co-trimoxazole

White MV 1989, Barbaud A 1998

Isoniazid

Asai K 1987, Fujiwara K 1987, Puente Crespo Y 2003

Pentamidine

Belsito DV 1993, Pradalier A 1993

Streptogramins (synergistins) Pristinamycin Barbaud A 2004

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

219

Table 5. Evidence of serum IgE to non-
-lactam antibiotics Class and drug

Solid phase nitrocellulose

Quinolones Macrolides Erythromycin

Neg Neg

References sepharose

other

Pos

Campi P 2000, Campi P 2003, Manfredi M 2004

Pos

Pascual C 1995 Jorro G 1996 Manfredi M and Campi P 1997 and 2006 (personal data) Wälti M 1986 Harle DG 1987, Smal MA 1988, Pham NH and Manfredi M 1996 Cabanas R 1996 Alonso MD 1992

Pos Trimethoprim

Pos Pos Pos Neg

Sulfonamides Sulfamethoxazole

Pos

Pos

Carrington DM 1987, O’Neil CE 1991, Gruchalla RS 1991, Gutt L 1998 Harle DG 1988 Manfredi M and Campi P 1994 (personal data)

Pos Pos Rifamycins Rifamycin SV Rifampicin

Pos

Glycopeptides Vancomycin

Pos

Pos CAP


Isoniazid

Magnan 1999 Manfredi M and Campi P 2000 (cross-reactivity with rifamycin SV; personal data) Manfredi M and Campi P 2000 (no crossreactivity with teicoplanin; personal data)

Pos BSApolystyrene

Fujiwara K 1987

Table 6. Adverse reactions to quinolones: in vivo tests

220

Number of cases

Skin tests

8 1 1 1 33 3 9 1 1 1 1

Positive (5 of 8) Positive Negative Negative Not reliable (positive in controls) Not reliable (positive in controls) Not reliable (positive in controls) Not reliable (positive in controls) Negative (prick test) Negative (prick test) Positive

Provocation test

Positive Positive Positive Positive Positive Positive Positive

References

Arboit F 1997 Munoz-Pereira M 1995 Valdivieso R 1988 Reaňo M 1995 Vieluf D 1991 Davila I 1993 Colàs C 1993 Ronnau AC 1997 Gonzales-Mancebo E 2002 Alemàn AM 2002 Lopes da Silva S 2003

Campi  Manfredi  Severino

Table 7. Recommended nonirritating intradermal test concentrations for quinolones Drug

Nonirritating concentration (mg/ml)

Reference

Ciprofloxacin Ciprofloxacin Ciprofloxacin Ciprofloxacin Ciprofloxacin Ciprofloxacin Ofloxacin Levofloxacin Levofloxacin Levofloxacin

0.004 0.02 0.000001 Not determinable 0.002 0.0002 0.5 0.025 0.025 0.025

Erdem G 1999 Barbaud A 2001 Empedrad RB 2000 Empedrad RB 2003 Scherer K 2005 Campi P and Severino M 2005 (personal data) Barbaud A 2001 Empedrad RB 2000 Empedrad RB 2003 Campi P and Severino M 2005 (personal data)

Modified from Scherer and Bircher [12].

find the nonirritating concentration because it varied by 1,000-fold in several subjects on separate testing dates! The nonirritating concentration is so low that also the sensitivity of the skin test becomes very low. For this reason we preferred to study these patients using serum-specific IgE. The literature reports only 2 cases in which serum-specific IgE has been researched using nitrocellulose for the solid phase [14, 15]: the test gave a negative result, notwithstanding a positive provocation test. The test of histamine release from the basophils [14, 16] was negative too, even though the provocation test was positive [see chapter of Sanz et al., pp 391–402]. We have, therefore, used sepharose as the solid phase. We examined 55 patients who reported 69 immediate-type adverse reactions to quinolones. In 30 of these the serological test was positive for a quinolone-specific IgE. Most of the patients were positive for more than one drug, therefore demonstrating a large cross-reactivity which was also confirmed by numerous bibliographic reports of adverse reactions with more than one quinolone. Inhibition tests were carried out with the sera of some patients: the binding of IgE to the solid-phase sepharose/drug was inhibited by

adsorption with the same drug at different concentrations (1, 10, 100 and 1,000 nmol according to the method of Harle et al. [4]), confirming the specificity of the binding of the quinolone-specific IgE. The test was carried out on 4 patients with lomefloxacin, ciprofloxacin, rufloxacin and ofloxacin and revealed a maximal inhibition of 95, 85, 72 and 70%, respectively. No inhibition was seen with other drugs (penicillin G, amoxicillin, cefuroxim, 4-aminoantipyrine and rifampicin) [8].

Pyrazolones

Pyrazolones are old drugs. Antipyrine was synthesized in 1883 and aminopyrine in 1897 and then withdrawn from sale in the 1970s because of the tendency to form nitrosamines. Dipyrone has been used since 1922; in some countries it has been withdrawn because of the risk of agranulocytosis and in other countries it is among the most used analgesics. Pyrazolones possess good analgesic and antipyretic action while the antiinflammatory action is low, as they rather poorly inhibit the synthesis of prostaglandins compared to other NSAIDs. Dipyrone, for example, acts at

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

221

Table 8. Synonyms and commercial names of pyrazolones Molecule

Synonyms

Commercial names

Aminopyrine

Aminophenazone, amidazophen, dipyrine, dimethylaminophenazone, amidopyrine

Amidopyrazoline, Amidofebrin, Anafebrina, Baukal, Clormetadone, Dimapyrin, Dipirin, Brufaneuxol, Farmidone, Febrinina, Flexipyrin, Ftalazone, Fugantil, Itamidone, Mamallet A, Malivan, Novamidon, Netsusarin, Optalidon suppositories, Piridol, Polinalin, Pyradone, Pyramidon,Virdex

Antipyrine

Phenazone

Analgesine, Anodynine, Optrex, Otalgan, Parodyne, Phenylone, Sedatine

Apazone

Azapropazone

Cinnamin, Prolixan, Sinnamin

Bumadizon

Eumotol

Chlormezanone (anxiolytic, muscle relaxant)

Dichloromethazanone, chlor-methazanone

Alinam, Banabin-Syntial, Fenarol, Lobak, Mio-Sed, Rexan, Rilansyl, Rilaquil, Rilassol, Supotran, Tanafol, Trancopal, Trancote, Transanate

Dipyrone

Metamizol, noramidopyrine, analgin, sulpyrin

Alginodia, Algocalmin, Baralgina, Bonpyrin, Buscopan Compositum, Conmel, Divarine, Dolazon, D-Pron, Dya-Tron, Eblimon, Espyre, Farmolisina, Feverall, Fevonil, Keypyrone, Metilon, Minalgin, Narone, Nartate, Nevralgina, Nolotil, Novacid, Novaldin, Novalgin, Novemina, Novil, Paralgin, Pyralgin, Pyril, Pyrilgin, Pyrojec, Tega-Pyrone, Unagen, etc.

Feprazone

Phenyl-prenazone, prenazone

Analud, Fepramol, Methrazone, Zepelin, Zerinol

Nifenazone

Nicotinoyl-aminoantipyrine

Algotrex, Dolongan, Neopiran, Nicopyron, Nikofezon, Niprazina, Phenicazone, Reumatosil, Reupiron, Supermidone, Thylin

Oxyphenbutazone

Hydroxy-phenylbutazone

Californit, Crovaril, Flogitolo, Flogoril, Frabel, Neo-Farmadol, Oxalid, Rapostan, Tandacote, Tandearil, Tanderil, Visubutina

Phenylbutazone

Flexazone, diphebuzol, fenibutazona

Ambene, Artrizin, Artrowas, Azolid, Bizolin, Butacote, Butadion, Butapirazol, Butadiona, Butatron, Butoz, Butazolidin, Buzon, Ecobutazone, Equipalazone, Exrheudon N, Fenibutol, Intrabutazone, Intrazone, Mepha-Butazon, Phenyzene, Robizone-V, Tevcodyne, Ticinil, Uzone

Propyphenazone

Isopropylphenazone

Antalgil, Budirol, Caffalgina, Causyth, Cibalgina, Eufibron, Flexidone, Mindol, Isoprochin P, Optalidon dragees, Ribelfan, Saridon, Uniplus, Veramon

Sulfinpyrazone (uricosuric, antithrombotic)

Sulfoxy-phenylpyrazolidine

Anturan, Anturane, Anturano, Enturen

Suxibuzone

Hydroxy-methylbutazolidine

Calibène, Danilon, Flogos, Solurol

the level of the central nervous system inhibiting prostanoids. In table 8 the synonyms and commercial names of pyrazolones are given based on the Merck Index [17].

222

Pyrazolones can provoke adverse reactions through the inhibition of cyclooxygenase in patients with the ASA triad (respiratory reactions) and in patients with chronic urticaria (skin reactions) [Szczeklik et al., pp 340–349]. On the other

Campi  Manfredi  Severino

hand, a real ‘immunological’ mechanism is also possible and highly probable, as for many years it has been known that a number of patients present hypersensitivity reactions only with pyrazolones but tolerate all the other classes of NSAIDs. We are dealing with reactions mainly of the immediate type, i.e. urticaria-angioedema and anaphylaxis, but also of delayed type, like fixed drug eruption, the Stevens-Johnson syndrome and toxic epidermal necrolysis. Indeed, the first accurate description of the entire and mainly allergic adverse skin reactions relates to antipyrine and dates back to 1898 [18]. To our knowledge, in 1958 Halpern and Holtzer [19] gave the first demonstration that hypersensitivity reactions to pyrazolones may be mediated by ‘reagins’. They studied in depth a patient with several episodes of severe anaphylactic shock with pyrazolones, initially taken for epileptic seizures. Within 8 min, a provocation test with a dose of 1 g aminopyrine produced anaphylactic shock, and an EEG excluded an epileptic seizure. The scratch test with aminopyrine gave a strongly positive result in the patient. The Prausnitz-Künstner test in 15 subjects was highly positive in 12, slightly positive in 1 and doubtful in 2. They were able to obtain a local reaction in the subjects at the site of injection of the serum but only when aminopyrine was given orally or intramuscularly (symptoms appeared after 5– 30 min). No reaction was observed when aminopyrine was inoculated intradermally at the same injection site. The interpretation of this discrepancy is difficult even today. Incidence In the 1970s and 1980s, mainly in Europe pyrazolones headed the list of drugs classified as provoking hypersensitivity reactions, particularly of the anaphylactic type; while in Anglo-Saxon countries most pyrazolones were not on sale. Currently they are used less and are in fourth place after
-lactams, ASA and other NSAIDs (table 2).

Skin Tests There are numerous reports of immediate positive prick and intradermal reactions obtained with pyrazolones, mainly dipyrone (table 9). Some authors also report systemic reactions during intradermal testing. Various research groups have also used scratch and patch tests with immediate readings (30 min). Some groups have used pyrazolones conjugated with a polylysine as a carrier and found a higher positive number compared to the use of the molecule alone [20]. However, other groups [21] found a positive skin test with a conjugate salicylate-polylysine even in 50% of subjects with positive challenge with ASA, which would indicate a high percentage of truly IgE-positive reactions to ASA in NSAID intolerance – but the specificity of this reagent has not been definitely proven [21]. Another reagent used for skin tests is phenylbutazone conjugated to bovine albumin (table 9) which according to Girard and Zawodnik [22] presents considerably greater sensitivity compared to the drug as such. Nevertheless, such a simple reagent has never been used routinely, nor has it been commercialized like penicilloyl-polylysine. Serum-Specific IgE The studies which have tried to demonstrate the presence of serum-specific IgE are reported in table 9. The tests using pyrazolones bound directly to nitrocellulose activated with cyanogen bromide have given very different results. In some studies positivity ranged from 14 to 17% in the subjects studied [20, 23], and others reported nearly 90%, that is 17/19 patients [24]. Other studies found no positivity with a commercial ELISA method [25] and also false-positives in the controls using an RIA method [26]. Recently serum-specific IgE towards propyphenazone was detected by means of an ELISA test using a conjugate of propyphenazone with human serum albumin (HSA) [27]. It was positive in 31 of 53 patients with adverse reactions of the immediate type. In this group 44 of 53 sub-

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

223

Table 9. Evidence of IgE to pyrazolones by means of positive skin tests and results of in vitro tests Molecule

Skin tests (number of patients with positive results)

ANT, AMI OXY DIP DIP, AMI DIP, AMI PHE-BSA AMI, DIP AMI, DIP, PRO, PHE SUX, AMI ANT, PRO, DIP, PHE Polylysine-pyrazolone Pyrazolones Polylysine-pyrazolone AMI, PRO DIP, PRO, ANT DIP

Scratch, P-K (1) Intradermal (1) Scratch, intradermal (23), P-K (5) Scratch, intradermal (14) Scratch (19), intradermal (4) ? (11) Intradermal (118) Scratch (2), patch at 30 min (10) Not done Scratch (9) Intradermal (14) Prick, intradermal (31) Prick, intradermal (46) Prick (1) Intradermal (6) Prick (8), intradermal (11)

DIP DIP

Intradermal (1) Intradermal (12)

PRO

Prick, intradermal (44)

DIP DIP

Prick, intradermal (13) Prick (1)

Serum IgE, HR, LTT, BAT, number of patients

HR and LTT positive (1)

RAST positive (25) RAST positive (7)

LTT positive (1) RAST positive (17) ELISA (commercial) negative (No?); RAST positive in controls

References

Halpern B 1958 Littlejohns DW 1973 Patriarca G 1973 Szczeklik A 1977 Czerniawska-Mysik G 1981 Girard JP 1983 Cirstea M 1985 Maucher OM 1986 Wälti M 1986 Wüthrich B 1986 Wüthrich B 1986 Mandallaz M 1987 Mandallaz M 1987 Brattig NW 1988 Zhu D 1992 Rubio M 1994

Bellegrandi S 1999 ELISA (commercial) Kowalski ML 1999 negative (28) ELISA positive Himly M 2003 (propyphenazone-HSA) (31) BAT positive (11) Sanz ML 2003 Molto L 2004

AMI = Aminopyrine; ANT = antipyrine; DIP = dipyrone; OXY = oxyphenbutazone; PHE = phenylbutazone; PRO = propyphenazone; SUX = suxibuzone; HR = leukocyte histamine release; LTT = lymphocyte transformation test; BAT = basophil activation test (CD63); P-K = Prausnitz-Künstner.

jects presented positive skin tests [27]. To circumvent steric hindrance of antibody-propyphenazone-HSA interaction, propyphenazone was linked to a spacer molecule with 4 carbon atoms via the nitrogen in position 1 of the pyrazolone moiety. For all four studies cited [20, 23, 24, 27], several important methodological criticisms remain: • The high percentage of positive results found on testing could be due to the low cutoff value (1.5 or 3 standard deviations above the mean of the negative controls) when compared, for example,

224

with the cutoff value used in sepharose-based tests (three times the mean of the negative controls). For example: if the mean of the negative controls is 1% (percentage of bound radioactivity) and the standard deviation is 0.3%, the cutoff value is 1.45 or 1.9% in the first case, with sepharose the cutoff would be 3%; • In the studies in the literature the results for single patients are not reported. Therefore it is not possible to know how high the positive results were; • The results have not been confirmed by later studies carried out with the same methodology;

Campi  Manfredi  Severino

Table 10. Results of assay for serum IgE to 4-aminoantipyrine using EA-Sepharose 6B as solid phase and 125I-labelled anti-human IgE Patient No.

Drug

Bound radioactivity, %

Total IgE kU/l

Inhibition with dipyrone, %

6900 5696 6914 7257 6591 7190 Cord blood Normal blood

Propyphenazone Propyphenazone Propyphenazone Dipyrone Dipyrone Feprazone

2.1 4.6 3.0 5.3 3.2 4.4 0.5 0.6

33 123 84 n.d. 100 n.d.

n.d. 73 54 47 62 73

Manfredi M, Zerboni R, Campi P, Severino M 1994 (personal data). n.d. = Not done.

• With the same methodology [24, 28] there have been reports of serum IgE towards salicylates in a large percentage of subjects: 27/28 [28]. The relevance of this is controversial; • In the work of Wälti et al. [23] there are reports of IgE specific for salicylates with a greater frequency than with pyrazolones. Moreover, the patients were not selected on the basis of skin tests and the symptoms of adverse reactions are not reported. Mention is made of a RAST inhibition test but without giving the number of subjects or reporting the results. The activation test of the basophils by means of flow cytometry with CD63 has given a positive outcome in 11 of 26 patients with adverse reactions of the immediate type; the skin tests were positive in 13 of 26 patients [29]. This test, whilst promising, requires further confirmation by other centers. In 1994 our group studied 25 patients with adverse reactions of the immediate type which had occurred during the previous 3 years following the intake of pyrazolones. Subsequently, all of them tolerated other classes of NSAIDs. We applied epoxy-activated Sepharose 6B for the research of serum IgE against 4-aminoantipyrine and 4-hydroxyphenylbutazone [Beeler and Pich-

ler, pp 380–390]. These 2 molecules are structural analogs of the pyrazolones [5]; on the one hand they are capable of binding themselves to the solid phase and, on the other, of exposing the possible epitopes of pyrazolones to IgE. Only 6/25 patients had a positive result with 4-aminoantipirine, with three times greater values obtained with cord serum and a pool of control subjects (table 10). All 6 subjects experienced the reaction within 1 h of taking the drug: propyphenazone in 3 cases, dipyrone in 2 cases, and feprazone in 1 case. Atopy was present in 3 of 6 cases. The inhibition test with dipyrone carried out in 5 cases showed a maximum inhibition from 47 to 73%, thus confirming the specificity of the bond. Dipyrone has demonstrated a greater inhibition capacity than 4-aminoantiprine. We repeated the dosage of specific and total IgE 6 and 12 months after the adverse reaction and thereafter every year for 4 years. All the patients have shown a fall in IgE, both specific and total, with different kinetics from patient to patient. In 3 patients IgE disappeared 12, 13 and 36 months, respectively, after the adverse reaction. In the other 3 patients it was still present 7, 24 and 48 months after the adverse reaction (personal data).

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

225

In conclusion, there is no doubt that the pyrazolones are capable of provoking even serious IgE-mediated reactions, with a rather significant frequency and in any case more than that of all other classes of NSAIDs. This has been demonstrated both with skin and provocation tests in numerous studies. In vitro tests for pyrazolonespecific IgE appear to have a low sensitivity, with a positive result in only a small percentage of patients, even if they were well chosen by means of history, skin and provocation tests. Moreover, the serum-specific IgE values were always low compared to those found against
-lactams and quinolones by means of sepharose-based tests. The best results were obtained with an ELISA test using pyrazolones coupled to HSA [27], but these results need to be confirmed. The low reactivity in otherwise well-documented IgE-mediated drug hypersensitivity reactions illustrate well the difficulty of these assays in general. For the pyrazolones, they can be summarized as follows. • The pyrazolone molecules do not bind easily to the solid phase, lacking suitable reactive groups; • The modifications introduced in the molecule to bind them to the carrier molecule probably altered or blocked their epitopes; • Perhaps epitopes are in part made up of the drug and a carrier protein which is not identical in the test compared to the in vivo situation. On the other hand, it is not likely that the metabolites are responsible for the immediate reactions, given that the reactions have a very short latent period and the skin tests, which are read immediately, are often positive. Cross-Reactivity In the group of patients with suspected pyrazolone hypersensitivity who tolerate ASA and other NSAIDs, the immune reaction seems sufficiently specific for a single molecule, even if not in all cases. In 26 patients who reacted to propyphenazone or dipyrone a challenge was carried out using both drugs, and a positive outcome was found

226

in only 4 cases (15%) [30]. Czerniawska-Mysik and Szczeklik [31] carried out the tolerance test with phenylbutazone and sulfinpyrazone in 32 patients with adverse reactions of the immediate type with dipyrone or aminopyrine, with a negative outcome. Vice versa, in our records on 35 patients with reactions to pyrazolones who tolerated ASA and other NSAIDs, 74% reported reactions with more than one pyrazolone (personal data). Genetic Predisposition Based on their clinical history and skin tests, Kowalski et al. [32] selected 26 patients with allergic reactions of the immediate type. The classII HLA-DQ7 and DR11 occurred in these subjects with a statistically higher frequency than in the general population. Management Whenever a patient reports a hypersensitivity reaction exclusively to a pyrazolone, a skin test should be performed. It is important to use very low initial drug concentrations for patients who had anaphylactic shock. A positive result in the skin tests indicates an immunological mechanism and excludes cross-reactivity with other classes of NSAIDs, but not with other pyrazolones. In these cases one can proceed using a tolerance test with ASA or other NSAIDs which the patient may need. If the skin tests are negative, it is more prudent to carry out a tolerance test first with an ‘alternative’ for classical NSAIDs, such as nimesulide, acetaminophen and selective inhibitors of COX-2, such as etoricoxib.

Management of Patients with Adverse Reactions to Drugs (Except
-Lactams)

The management of patients with adverse reactions to drugs requires, as a general guide: (1) history; (2) skin tests; (3) in vitro tests; (4) graded

Campi  Manfredi  Severino

challenge/provocation tests [Solensky, pp 404– 412], and (5) desensitization. History In order to collect clinical histories, it is useful to employ the questionnaire [33] edited and published by the European Network for Drug Allergy (ENDA), the study group on allergy to drugs of the European Academy of Allergy and Clinical Immunology. A precise history is fundamental as none of the available tests can overrule a convincing history; in general, tests have a limited specificity. Nevertheless, even a detailed history may not be sufficient to attribute the cause of the symptoms to a certain drug. In fact, it has been demonstrated that during drug treatments, with whichever molecule, one can frequently observe symptoms which imitate an allergic reaction but which may not be provoked by the drug itself. Indeed, in a study by Messaad et al. [34], only 241 of 1,372 provocation tests performed with the suspected drug yielded a positive result (17.6%) in a series of 898 consecutive patients with a history suggesting an immediate reaction. The same group applied a provocation test with the suspected drug to 107 patients with adverse reactions to macrolides with a positive outcome in only 8 cases (7.5%), 7 cases with spiramycin and 1 case with roxithromycin [35]. Similar results were obtained in children in whom provocation tests had always given a negative result [36]; viral infections are common causes of rash, often together with an antibiotic treatment. It is therefore necessary to extend the history and ultimately the diagnostic tests to other possible and concomitant causes of the reported symptoms: other drugs taken contemporarily, food allergies, infective agents, physical exercise, and so on. Skin Tests Recently ENDA has edited and published a document on the general considerations concerning the procedures for skin tests, which should be re-

ferred to for details [37]. The allergological skin tests are standardized and predictive only for a few drugs, i.e.
-lactams and drugs of high molecular weight (polypeptides). Most drugs have a low molecular weight, lower than 1,000 daltons; for these prick tests and intradermal reactions with immediate readings are not standardized and validated with regard to sensitivity and specificity. The skin tests should be carried out not sooner than 3 weeks after the reaction, but not later than a few months, by experts trained in allergy to drugs. As side effects might appear even during skin testing, one should be prepared to treat eventual adverse reactions including anaphylaxis [37]. In relation to the type of adverse reaction, the most common indications [37] to carry out skin tests are reported in table 11. In delayed-type reactions, it is recommended to start with the patch test and, in case of negativity, follow with the prick and intradermal reaction. The intradermal test with delayed reading seems to be more sensitive than the patch test, but less specific. In immediate-type reactions, one should begin with the prick test, which is less sensitive but safer compared with the intradermal test. If this is negative, carry out the intradermal test. In other adverse reactions to drugs with a suspected immunological pathogenesis, skin tests do not help diagnosis: the relevance of skin tests has never been demonstrated in hematological manifestations (anemia, thrombocytopenia, leukocytopenia), renal (glomerulonephritis) or hepatic (hepatitis) and in autoimmune diseases such as lupus erythematosus, bullous pemphigoid, pemphigus vulgaris and interstitial pneumopathies. On the other hand, there are clearly positive skin tests in patients who only had druginduced hepatitis. In these cases, a positive reaction should be controlled by testing other individuals. In general, skin tests are not carried out in pregnant patients.

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

227

Table 11. Common clinical indications for skin testing in the diagnosis of drug hypersensitivity [37] Patch tests can be used as first line of investigation

Skin prick tests and intradermal tests

Acute generalized exanthematous pustulosis Contact dermatitis Erythema multiforme Exanthematous drug eruption Fixed drug eruption Photoallergic reactions Purpura/leukocytoclastic vasculitis Stevens-Johnson syndrome Toxic epidermal necrolysis

Anaphylaxis Bronchospasm Conjunctivitis Rhinitis Urticaria/angioedema

In the last decade many centers have identified concentrations above which a skin test is irritating for most patients. In particular, Empedrad et al. [13] studied 25–30 healthy subjects (male and female, 18–50 years old) and published the nonirritating concentrations following intradermal tests of 5 cephalosporins and 10 other antibiotics in common use [Solensky, table 1, p 406]. Our group has extended this list by studying 20 control subjects for clarithromycin reactivity. We found 0.05 mg/ml to be not irritating in the intradermal test. A positive test is an indication for avoiding the drug (if the test is not irritating). As these tests cannot be controlled by provocation tests, it is possible that a certain number of patients, while presenting positive skin tests, can tolerate the administration of therapeutic doses of the drug. But this is the same with other IgE-mediated skin tests to foods, hymenoptera venoms, etc. A negative result is not a guarantee that the drug is tolerated, because the immune response might have been directed towards metabolites of the drug, or may not be IgE-mediated, or too long a period of time could have passed since the adverse reaction. To be sure, a provocation test is needed or, as it is more practical, a graded challenge test is done [Solensky, pp 404–412].

228

In vitro Tests No commercial kits are available for detecting serum-specific IgE to the drugs discussed above. Some specialist laboratories, like ours, have developed epoxy-activated Sepharose TM 6B to be used as the solid phase for some molecules in an RIA test using 125I-labelled anti-human IgE [4, 5]. Epoxy-activated Sepharose 6B has a high capacity to bind drugs. It is superior to nitrocellulose, as most drugs have a low molecular weight and do not bind easily to nitrocellulose, except through a protein such as albumin. The epoxy-activated Sepharose 6B possesses a long hydrophilic spacer arm which inserts itself between the matrix and the immobilizing molecule facilitating the bond with the antibody and therefore allowing a better presentation of the antigen. The functional groups which are necessary to realize the covalent bond with the spacer from the solid phase are nucleophilic groups such as the hydroxyl, carboxyl, thiol and amino groups. If the molecule does not possess one of these groups it is necessary to synthesize structural analogs of the drug, i.e. new molecules which possess both a functional group for the bond to the solid phase and all the possible epitopes: for example 4-aminoantipyrine and 4phenylbutazone for pyrazolones.

Campi  Manfredi  Severino

A drawback of these assays is the possibility that the protein- or spacer-binding site of the drug is the antigenic epitope itself and is not free to be recognized by the IgE antibody. In this regard, the basophil-based tests, which use free drug antigens, seem to have an advantage. It is to be noted that serum-specific IgE has less persistence in time compared with cutaneous IgE, and moreover they persist for less time in non-atopic compared with atopic subjects [8]. With regard to the lymphocyte transformation test and the flow cast, refer to the relative chapters by Beeler and Pichler [pp 380–390] and Sanz et al. [pp 391–402]. Graded Challenge and Provocation Test The provocation test is the controlled administration of a drug for the diagnosis of hypersensitivity. In provocation tests one would like to prove an allergy and may expect symptoms to prove an association with drug intake. Even the provocation test does not represent the ‘gold standard’, since it does not take into account the cofactors which are present in the daily use of the drug, such as an ongoing infection, physical exercise, food allergies and perhaps other unknown factors. ENDA has edited and published a document [38] which should be consulted for details concerning the procedures. These tests can be carried out after carefully evaluating the risks/benefits for the patient and in a clinical setting prepared to handle anaphylaxis. Whenever possible, one should carry out skin tests and in vitro tests in advance, and patients with positive tests should not be provoked if the test is considered reliable and controlled (see above). If a drug is indispensable, a desensitization program can be undertaken. Often, a ‘graded challenge’ test is done because one does not expect hypersensitivity as the history is inconsistent and skin tests are negative. One wants to prove tolerance of the drug, but for safety reasons one starts with a smaller dose. Practically it means that the drug is started with

a rather low dose (ca. 1/100 of the usual dose), followed by a stepwise increase (5- to 10-fold concentrations, reaching the maximal dose in 3– 4 h). The initial low dose already reduces the risk of a severe reaction substantially, but the scheme does not induce a desensitization. The principal general indications for a graded challenge and certain provocation tests are [38]: • To exclude hypersensitivity in a nonsuggestive history of drug hypersensitivity and in patients with nonspecific symptoms, such as vagal symptoms under local anesthesia; • To provide safe pharmacologically and/or structurally nonrelated drugs in proven hypersensitivity such as other antibiotics in
-lactamallergic patients. This may also be helpful for anxious people who would refuse to take the recommended drug without proof of tolerance; • To exclude cross-reactivity of related drugs in proven hypersensitivity, for example a cephalosporin in a penicillin-allergic subject or an alternative NSAID in an ASA-sensitive asthmatic; • To establish a firm diagnosis in suggestive history of drug hypersensitivity with negative, nonconclusive or nonavailable allergologic tests, for example a maculopapular eruption during aminopenicillin treatment with negative allergological tests. Contraindications are [38]: • Pregnant patients; • Concomitant illnesses such as acute infections or cardiac, hepatic, renal or other pathologies in which the provocation test could lead to an uncontrollable situation, with the exception of cases in which the drug is indispensable, such as penicillin in neurosyphilis; • Previous severe life-threatening reactions, such as severe anaphylaxis, immunocytotoxic reactions, specific organ manifestations such as hepatitis, nephritis, pneumonitis, systemic vasculitis, exfoliative dermatitis, erythema multiforme major/Stevens-Johnson syndrome, drug reaction with eosinophilia and systemic symptoms, toxic epidermal necrolysis, fixed drug

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

229

Table 12. Desensitization protocols Class and drugs

References

Aminoglycosides Streptomycin Tobramycin

Vervloet D 1999 Earl HS 1987, Schretlen-Doherty JS 1995

Rifamycins Rifampicin Glycopeptides Vancomycin

Holland CL 1990, Matz J 1994 (rifampicin + ethambutol), Alonso MD 1996, Alvarez Cuesta JA 1997, Drira I 1997, Portu Zapirain J 1997, Arrizabalaga J 1998, Pech O 2001, Buergin S 2006 Lerner A 1984, Lin RY 1990, Anne S 1994, Wong JT 1994, Villavicencio AT 1997, Sorensen SJ 1998, Chopra N 2000, Wazny LD 2001, Kitazawa T 2006

Imidazoles Fluconazole Itraconazole Metronidazole Amphotericin B

Craig TJ 1996, Witkowska R 2003 Bittleman DB 1994, Douglas R 1997 Kurohara ML 1991, Pearlman MD 1996 Kemp SF 1995

Sulfonamides Sulfadiazine

De la Hoz Caballer B 1991, Tenant-Flowers M 1991, Moreno JN 1995

Co-trimoxazole

Finegold I 1986, Hughes TE 1986, Greenberger PA 1987, MacLean Smith R 1987, White MV 1989, Kletzel M 1991, Absar N 1994, Gluckstein D 1995, Nguyen M 1995, Belchì-Hernandez J 1996, Kalanadhabhatta V 1996, Palusci VJ 1996, Piketty C 1996, Caumes E 1997, Marsh BJ 1997, Rich JD 1997, Demoly P 1998, Theodore CM 1998, Ryan C 1998, Bonfanti P 2000, Yoshizawa S 2000, Wacker P 2000

Isoniazid

Holland CL 1990 (isoniazid + rifampicin)

Clindamycin

Martin JA 1992, Marcos S 1995

Pentamidine

Greenberger PA 1987, Baum CG 1992

eruption of bullous type, acute generalized exanthematous pustulosis, autoimmune diseases such as systemic lupus erythematosus, pemphigus vulgaris, bullous pemphigoid. Desensitization Apart from penicillin and other
-lactams, the literature describes desensitization procedures carried out successfully with various drugs, in particular in patients affected by opportunistic infections while suffering from AIDS or affected by cystic fibrosis (table 12). The term ‘desensitization’ is used classically to indicate the induction of an immunological

230

tolerance towards an antigen which has caused an IgE response. The term ‘desensitization’ is also used when an IgE-mediated mechanism was not demonstrated, as with ASA and the sulfonamides. In the case of rapid desensitization, the responsible drug is administered within several hours, beginning with tiny doses and progressively increasing them until a state is reached which allows the administration of the drug without an adverse reaction or with a minimal reaction. The drug must therefore be taken daily; if it is interrupted for 48 h it is necessary to repeat the procedure from the beginning. Desensitiza-

Campi  Manfredi  Severino

tion does involve a modification of the immunological state of the patient with regard to the given drug: it has, in fact, been demonstrated that skin tests with penicillin become negative during the course of desensitization to penicillin. However, the cutaneous sensitivity to histamine, to compound 48/80 (which induces degranulation of cutaneous mastocytes) and to aeroallergens is preserved. This has also been demonstrated for vancomycin [39, 40]. This is a procedure which carries considerable risk for the patient, therefore the risk/benefit factors must be evaluated very carefully. Desensitization must be distinct from a provocation test which has a diagnostic scope: if the skin test with the suspected drug is positive using a nonirritat-

ing concentration [13], desensitization might be considered – but not provocation or graded challenge. If, on the other hand, the skin test is negative, and no reaction is expected, graded challenge or provocation may be envisaged. For example, if a patient reports a rash with itching (without any other symptoms) during the first day of administration of a sulfonamide 20 years ago, it would be reasonable to carry out a provocation test. If, on the other hand, the patient has a recent history of anaphylaxis, it would be necessary to carry out desensitization [41]. For particulars of the procedure, indications and contraindications, refer to the chapter on desensitization by Solensky [pp 404–412].

References 1 Arndt KA, Jick H: Rates of cutaneous reactions to drugs. A report from the Boston Collaborative Drug Surveillance Program. JAMA 1976;235:918–922. 2 Caldwell JR, Cluff LE: Adverse reactions to antimicrobial agents. JAMA 1974;230:77–80. 3 Naldi L, Conforti A, Venegoni M, Troncon MG, Caputi A, Ghiotto E, et al: Cutaneous reactions to drugs. An analysis of spontaneous reports in four Italian regions. J Clin Pharmacol 1999;48: 839–846. 4 Harle DG, Baldo BA, Smal MA, Van Nunen SA: An immunoassay for the detection of IgE antibodies to trimethoprim in the sera of allergic patients. Clin Allergy 1987;17:209–216. 5 Baldo BA, Harle DG: Drug allergenic determinants. Monogr Allergy 1990; 28:11–51. 6 Campi P, Manfredi M, Severino MG, Zerboni R: IgE specifiche verso chinoloni. G Ital Allergol Immunol Clin 2000;11:63–67. 7 Campi P, Pichler WJ: Quinolone hypersensitivity. Curr Opin Allergy Clin Immunol 2003;3:275–281. 8 Manfredi M, Severino M, Testi S, Macchia D, Ermini G, Pichler WJ, Campi P: Detection of specific IgE to quinolones. J Allergy Clin Immunol 2004;113:155–160.

9 Pascual C, Crespo JF, Quiralte J, Lopez C, Wheeler G, Martin-Esteban M: In vitro detection of specific IgE antibodies to erythromycin. J Allergy Clin Immunol 1995;95:668–671. 10 Burke P, Burne SR: Allergy associated with ciprofloxacin. BMJ 2000;320: 679. 11 Davis H, McGoodwin E, Greene Reed T: Anaphylactoid reactions reported after treatment with ciprofloxacin. Ann Intern Med 1989;111:1041–2043. 12 Scherer K, Bircher AJ: Hypersensitivity reactions to fluoroquinolones. Curr Allergy Asthma Rep 2005;5:15–21. 13 Empedrad R, Carter AL, Earl HS, Gruchalla RS: Nonirritating intradermal skin test concentrations for commonly prescribed antibiotics. J Allergy Clin Immunol 2003;112:629–630. 14 Davila I, Diez ML, Quirce S, Fraj J, De La Hoz B, Lazaro M: Cross-reactivity between quinolones – report of three cases. Allergy 1993;48:388–390. 15 Colàs C, Pola J, Zapata C, Dominguez A, Gimeno A, Duce F: Quinolone hypersensitivity: clinical and immunological aspects (abstract). J Allergy Clin Immunol 1993;91:365. 16 Valdivieso R, Pola J, Losada E, Subiza J, Armentia A, Zapata C: Severe anaphylactoid reaction to nalidixic acid. Allergy 1988;43:71–73.

IgE-Mediated Allergy to Pyrazolones, Quinolones and Other Non-
-Lactam Antibiotics

17 The Merck Index. Whitehouse Station, Merck Research Laboratories, 1996. 18 Apolant H: Die Antipyrinexantheme. Arch Derm Syph 1898;46:345–398. 19 Halpern B, Holtzer A, Liacopoulust P, Meyer J: Allergy to pyrazolone derivates (aminopyrine) with evidence of reaginic type antibody. J Allergy 1958; 29:1–12. 20 Wüthrich B, Fabro L, Wälti M: ‘Soforttyp’-Reaktionen auf Pyrazolderivate: Ergebnisse von Hauttesten und Antikörperbestimmungen. Dermatologica 1986;173:24–28. 21 Mandallaz M, de Weck AL, Frosch P, Haustein U, Maucher OM, Patriarca G, et al: A prospective survey of allergy/ intolerance to pyrazolones, salicylates and betalactams and skin-testing with polylysine conjugates. Int Arch Allergy Appl Immunol 1987;83:S1. 22 Girard JP, Zawodnik S: Diagnostic procedures in drug allergy; in de Weck AL, Bundgaard H (eds): Allergic Reactions to Drugs. Berlin, Springer, 1983, pp 207–225. 23 Wälti M, Neftel KA, Cohen M, Winkler P, de Weck AL: Radioimmunologische Erfassung von IgE- und IgG-Antikörpern gegen Medikamente. Schweiz Med Wochenschr 1986;116:303–305.

231

24 Zhu D, Becker WM, Schulz KH, Schubeler K, Schlaak M: Detection of IgE antibodies specific for 1-phenyl2,3-dimethyl-3-pyrazoline-5-one by RAST: a serological diagnostic method for sensitivity to pyrazoline drugs. Asian Pacific J Allergy Immunol 1992; 10:95–101. 25 Kowalski ML, Bienkiewicz B, Woszczek G, Iwaszkiewicz J, Poniatowska M: Diagnosis of pyrazolone drug sensitivity: clinical history versus skin testing and in vitro testing. Allergy Asthma Proc 1999;20:347–352. 26 Rubio M, Herrero MT, de Barrio M, Tornero P, Rodriguez V, Aranzàbal A, et al: Alergia a pirazolonas. Allergol Immunopathol 1994;22:104–106. 27 Himly M, Jahn-Schmid B, Pittertschatscher K, Bohle B, Grubmayr K, Ferreira F, et al: IgE-mediated immediate-type hypersensitivity to the pyrazolone drug propyphenazone. J Allergy Clin Immunol 2003;111:882–888. 28 Zhu D, Becker WM, Schulz KH, Schubeler K, Schlaak M: The presence of specific IgE to salicyloyl and o-methylsalicyloyl in aspirin-sensitive patients. Asian Pacific J Allergy Immunol 1992;10:25–32.

29 Sanz ML, Garcìa MC, Caballero MR, Diéguez I, Gamboa PM: Test de activaciòn de basòfilos en el diagnòstico de alergia a medicamentos. An Sis Sanit Navar 2003;26(suppl 2):39–47. 30 Przybilla B, Ring J, Harle R, Galosi A: Oraler Expositionstest mit Schmerzmittel-Inhaltsstoffen bei Patienten mit anaphylaktoiden Unverträglichkeitsreaktionen auf ‘leichte’ Analgetika. Allergologie 1986;9:315–318. 31 Czerniawska-Mysik G, Szczeklik A: Idiosyncrasy to pyrazolone drugs. Allergy 1981;36:381–384. 32 Kowalski ML, Woszczek G, Bienkiewicz B, Mis M: Association of pyrazolone drug hypersensitivity with HLADQ and DR antigens. Clin Exp Allergy 1998;28:1153–1158. 33 Demoly P, Kropf R, Bircher A, Pichler W: Drug hypersensitivity questionnaire. Allergy 1999;54:999–1003. 34 Messaad D, Sahla H, Benahmed S, Godard P, Bousquet J, Demoly P: Drug provocation tests in patients with a history suggesting an immediate drug hypersensitivity reaction. Ann Intern Med 2004;140:1001–1006. 35 Benahmed S, Scaramuzza C, Messaad D, Sahla H, Demoly P: The accuracy of the diagnosis of suspected macrolide antibiotic hypersensitivity: results of a single-blinded trial. Allergy 2004;59: 1130–1133.

36 Ponvert C, Scheinmann P: Antibiotic allergy and pseudo-allergy in child. Rev Fr Allergol Immunol Clin 2003; 43: 385–392. 37 Brockow K, Romano A, Blanca M, Ring J, Pichler W, Demoly P, et al; ENDA (European Network for Drug Allergy): General considerations for skin test procedures in the diagnosis of drug hypersensitivity. Allergy 2002;57:45– 51. 38 Aberer W, Bircher A, Romano A, Blanca M, Campi P, Fernandez J, Brockow K, Pichler WJ, Demoly P; ENDA Group: Drug provocation testing in the diagnosis of drug hypersensitivity reactions: general considerations. Allergy 2003;58:854–863. 39 Anne S, Middleton E, Reisman RE: Vancomycin anaphylaxis and successful desensitization. Ann Allergy 1994; 73:402–404. 40 Lin RY: Desensitization in the management of vancomycin hypersensitivity. Arch Intern Med 1990;150:2197–2198. 41 Solensky R: Drug desensitization. Immunol Allergy Clin North Am 2004;24: 425–443.

Dr. Paolo Campi Allergy Clinic, Azienda Sanitaria di Firenze Nuovo Ospedale San Giovanni di Dio, Via di Torregalli 3 IT–50143 Florence (Italy) Tel. +39 05 5719 2303, Fax +39 05 5719 2542 E-Mail [email protected]

232

Campi  Manfredi  Severino

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 233–241

Hypersensitivity Reactions to Iodinated Contrast Media: An Update Cathrine Christiansen Research and Development, GE Healthcare, Oslo, Norway

Abstract

Introduction

Iodinated contrast media are diagnostic agents used in millions of X-ray procedures yearly. Although current agents are usually well tolerated, both immediate (^1 h) and late (11 h) hypersensitivity reactions can occur in susceptible individuals. New data indicate that many of the immediate reactions, which range from mild urticaria to severe anaphylaxis, are caused by an IgE-mediated activation of mast cells in the organs affected. The late reactions are mainly non-serious skin eruptions, induced by contrast medium-reactive T cells. Previous reactors should not be re-exposed to the implicated product, as there is a possibility for a more severe repeat reaction. However, because immediate and late reactions are caused by different mechanisms, a patient with a previous late hypersensitivity reaction is not at increased risk of an immediate hypersensitivity reaction and vice versa. Current skin test data indicate that patients with previous immediate reactions may benefit from a change of product, as cross-reactivity between different products seems to be rare. In previous late reactors, extensive cross-reactivity between different products is common, and a change of product is therefore no guarantee against a new late skin reaction. Premedication with corticosteroids/antihistamines has no proven efficacy in preventing either serious immediate reactions or generalized late skin eruptions. Radiological examination should therefore be carried out with means of resuscitation at hand. Oral administration of corticosteroids both before and for 3 days after the procedure may be tried in an attempt to avoid late skin reactions. Copyright © 2007 S. Karger AG, Basel

Iodinated contrast media are diagnostic agents used in more than 75 million X-ray procedures yearly. Currently available products are tri-iodinated benzene derivatives, commonly divided into four classes based on osmolality and chemical structure: high-osmolality ionic monomers; lowosmolality non-ionic monomers; low-osmolality ionic dimers, and finally iso-osmolality non-ionic dimers. As can be seen from figure 1a and b, there is a considerable difference in chemical structure between the ionic and the newer generation of non-ionic monomers. While the ionic monomers contain a carboxylate group and two relatively short side chains in the remaining ring positions, the non-ionic monomers have long side chains rich in hydroxyl groups in all three positions. Two of these side chains are usually identical. It is also worth noting that most non-ionic monomers have one or two side chains in common (fig. 1b). As the name implies, the dimers contain two tri-iodinated benzene derivatives linked together (fig. 1c). In the ionic dimer, ioxaglate, these two benzene derivatives are different, while in the two non-ionic dimers, iotrolan and iodixanol, they are identical. Iodixanol is derived from iohexol, and the four identical side chains have the same composition as two of the side

chains in iohexol. Each of the four side chains in iotrolan have an additional CH2OH group. All contrast media for intravascular use are chemically inert substances that are commonly administered in quantities of 60–120 g within a short interval in order to enhance contrast in Xray examinations. Subsequently, these low molecular weight molecules mix with the circulating plasma volume, and simultaneously pass freely across the blood vessel walls into the extracellular fluid. Most of the injected dose is recovered unmetabolized in urine within the next 24 h in patients with normal kidney function. While the ionic contrast media show some protein binding due to their negative charge, the non-ionic contrast media are devoid of any measurable plasma protein binding. Considering the enormous dose given, the currently available products are remarkably safe. However, hypersensitivity reactions, which sometimes occur in susceptible patients, remain a problem both for the radiologists and the patients involved. The aim of this review is to provide an update on recent data related to the clinical presentation, prevalence, pathophysiology, treatment and prevention of hypersensitivity reactions induced by iodinated contrast media. As for other drugs, these reactions are divided into immediate (^1 h) and non-immediate or late (11 h) reactions.

Immediate Hypersensitivity Reactions

Clinical Manifestations and Prevalence Immediate hypersensitivity reactions to contrast media range from mild skin reactions to severe anaphylaxis. Itching and limited urticaria are reported to occur in 6% of patients after exposure to ionic contrast media and in 0.9% of patients exposed to non-ionic contrast media [1]. In more severe reactions, the respiratory and/or cardiovascular systems are involved. The frequency of severe immediate anaphylactic reactions was re-

234

ported by The International Collaborative Study of Severe Anaphylaxis to be 0.071% for patients exposed to ionic contrast media versus 0.035% for patients exposed to non-ionic contrast media [2]. Of the 34 cases of anaphylaxis identified for non-ionic products, 30 had reacted to iopromide, 4 to iohexol and 1 to iodixanol. If iopromide is excluded from the calculation, the frequency of severe anaphylaxis to non-ionic products is as low as 0.005%. The same low frequency has been reported by Cochran et al. [3]. From 1996 to 1999, they observed 10 severe reactions and 1 death after use of non-ionic contrast media in 41,060 examinations. Only 2 (0.005%) of these reactions were considered to be hypersensitivity reactions. Although anaphylaxis to non-ionic contrast media is rare, many patients will suffer from such reactions yearly due to the vast number of contrast medium-enhanced X-ray examinations conducted. Most of the immediate symptoms become apparent within the first 5–15 min of contrast medium administration [1] and disappear within 30–60 min, although reactions lasting from 6 to 24 h or sometimes longer have been reported [4]. Time to onset of severe contrast medium-induced anaphylaxis has varied from immediately during injection to 30 min after contrast medium administration [5]. Pathophysiology There is convincing evidence that the immediate hypersensitivity reactions to contrast media are mediated by histamine and possibly other vasoactive substances released from contrast mediumstimulated mast cells present in the organs affect-

Fig. 1. Chemical structure of commercially available Xray contrast media: ionic monomers (a), non-ionic monomers (b), ionic and non-ionic dimers (c). For each contrast medium the generic name is listed together with the most frequently used trade name in parentheses. Common structural elements among the different products are highlighted in red.

Christiansen

O

NH

O

NH

O COOH

I

I

O

O

N H

COOH

N H I

COOH

N H I

Iothalamate (Conray)

I

Ioxithalamate (Telebrix)

OH

O

NH

I

I

I

I

Amidotrizoate (Radioselectan)

a

OH

HO

N

O

N

H

I

OH

H

N

N OH HO

HO

HO

b

HO

OH

N H I H N OH HO

HO

OH

Iomeprol (Iomeron)

O

I O

HO

OH HO

I O

I

OH

CH3 CH3

OH

Iopromide (Ultravist)

OH

N

OH OH

HO

I O

N

H

I

H

OH

Iobitridol (Xenetix)

O I

N

I

O

OH

HO

Ioversol (Optiray)

O

HN

I

N

O

OH HN

I N H CH3

OH

OH HO

OH

O N

O

N H I H N

N

H

HO

O I

O

I

O I

I O

O

Iopentol (Imagopaque)

O HN

H

OH

Iohexol (Omnipaque)

I

O

OH N

I

I

I

HO

O

N

O

I O

O

OH

N

O I

I

HO

OH

O

O

N

O N H I H N

OH

HO

OH

OH

HO

Ioxitol (Ioxilan)

Iopamidol (Iopamiro)

O R1

R1 I

I

I

N

I I

R2

R4

R2

I

I

2 O

I

HO HO

c

Contrast Media Hypersensitivity

2 N

I

O

I

H

OH

I

N

H

HO

Iotrolan (Isovist)

O

I

HN

O NH

I

O

OH

O O

N H I H N

OH HO

O I

I

O N

O

OH

CH2 N

I

H N

OH

HO

Iodixanol (Visipaque)

OH

COOH N H I OH Ioxaglate

(Hexabrix)

235

ed by the hypersensitivity reaction. Laroche et al. [6] demonstrated that patients with moderate to severe immediate reactions had increased plasma levels of both histamine and tryptase (a specific marker of mast cell activation) after the imaging procedure. The levels of these compounds correlated with the severity of the reaction. Other investigators have reported high levels of tryptase in connection with severe or fatal contrast mediuminduced anaphylactic reactions [5]. The immediate hypersensitivity reactions to contrast media have generally been considered as non-allergic since the patients can react on first exposure, the reaction does not always recur, and contrast medium-specific IgE antibodies have seldom been detected. However, new results from France have challenged this dogma. The French Society of Radiology, through its working group on contrast agents (CIRTACI) in 2001 set up a network involving radiologists, allergologists and biochemists in order to study the mechanism of the immediate hypersensitivity reactions. The results from the first 3 years of study were presented at the International Contrast Media Research Symposium in Evian in October 2005 [7]. In summary, 26 patients with immediate reactions to contrast media were skin tested. Nineteen (73 %) had a positive intradermal test to the implicated contrast medium when read after 20 min. Seven of these 19 patients had experienced a life-threatening reaction. Cross-reactivity between different contrast media was low (9.3%). This unexpected low incidence of cross-reactivity between different products has been confirmed in the ongoing multicenter study of the contrast medium task force of the European Network of Drug Allergy (ENDA) [Brockow et al., in preparation]. Based on their skin test data, the CIRTACI group concluded that allergic, IgE-mediated reactions seem to be the most frequent mechanism involved in immediate hypersensitivity reactions to contrast media. The structural elements that constitute the IgE-binding epitopes are presently unknown.

236

Since severe immediate hypersensitivity reactions can occur in previously unexposed patients, Brasch [8] proposed some 25 years ago that these patients could have been sensitized with halogenated benzene derivatives which are used as food additives, pesticides and herbicides. No studies have been conducted to verify or disprove this hypothesis. Nevertheless, the situation is reminiscent of the anaphylactic reactions to neuromuscular blocking agents [Dewachter, pp 204– 215] where IgE to neuromuscular blocking agents can be detected in spite of a lack of previous exposure. It should also be added that the lack of reaction on re-exposure to the contrast medium that caused the previous reaction is not an argument against an IgE-mediated mechanism. It is well known from studies of immediate hypersensitivity reactions to
-lactams that patients lose sensitivity over time. Treatment and Prevention The unpredictable nature of severe anaphylactic reactions can be illustrated by the unpublished data collected in the ongoing ENDA multicenter study [Brockow et al., in preparation]. Twentytwo patients had experienced a severe immediate hypersensitivity reaction after contrast medium exposure. Only 9 of these had previously been exposed to an iodinated contrast medium, and only 5 had experienced a previous contrast mediuminduced immediate reaction. Most fatal reactions are also unpredictable. All of the 10 contrast medium-related anaphylactic deaths registered in the UK in a 10-year period from 1992 were firsttime reactions, although most patients had been exposed to contrast media previously without reaction [9]. The vast majority of patients with severe anaphylaxis recover if they are treated quickly and appropriately. The first-line drugs and equipment should be readily available, and proper training of personnel is essential. The patients should never be left alone for at least 30 min after

Christiansen

contrast medium administration, and the venous access should be left in place. Simple guidelines for first-line treatment of serious immediate reactions were recently published by the European Society of Urogenital Radiology [10]. The most important risk factor for a severe immediate hypersensitivity reaction to contrast media is a previous well-documented contrast medium-induced immediate hypersensitivity reaction [1]. If such patients need new contrast medium-enhanced imaging procedures, they should not be exposed to the product that caused the previous reaction. The apparent low degree of cross-reactivity between different contrast media means that these patients may tolerate another product. However, it is important to bear in mind that cross-reactivity can occur, and that the negative predictive value of skin testing has not yet been established. As recommended by Brown et al. [11], patients who experience a severe anaphylactic reaction to contrast media should be given a MedicAlert bracelet and should be entered into hospital and health care network clinical information systems in an attempt to reduce the risk of a future severe reaction. The prophylactic use of corticosteroids and antihistamines in these patients remains controversial, and premedication practice varies from country to country. There are no convincing clinical data available to show that premedication prevents serious repeat reactions to non-ionic contrast media [12]. Severe anaphylactic reactions have been reported despite optimal pretreatment [for review, see 13]. The role of skin testing in preventing severe repeat reactions remains to be established. The safe re-administration of a skin test-negative contrast medium in patients with previous lifethreatening anaphylactic reactions has so far only been described for 5 patients [14–16]. There is no documented evidence that allergy, including seafood allergy, predisposes to severe immediate hypersensitivity reactions to non-ion-

Contrast Media Hypersensitivity

ic products. It is, however, well established that poorly controlled asthma is a risk factor for immediate bronchoconstriction, and that both cardiovascular disorders and the use of
-blockers may aggravate the consequences of a contrast medium-induced anaphylaxis.

Late Hypersensitivity Reactions

Clinical Manifestations and Prevalence Late hypersensitivity reactions to contrast media are mainly various types of non-serious skin eruptions such as macular, maculopapular and urticaria-like eruptions that usually resolve within 1–7 days [17, 18]. As for other drugs, maculopapular exanthema is the most frequent presentation, occurring in more than 50% of affected patients [18]. The skin eruptions can be accompanied by angioedema, mild fever and eosinophilia, and desquamation may occur on healing [19]. Whereas immediate urticaria can be the first sign of an anaphylactic shock, late occurring urticaria-like eruptions have never been associated with life-threatening reactions. Large observational studies have shown that severe contrast medium-induced skin eruptions are rare. While Munechika et al. [20] detected no severe reactions among their 6,764 investigated inpatients, 2 of 11,121 patients (0.02%) in the study of Hosoya et al. [18] experienced severe skin reactions after exposure to iohexol. The types of eruption were not specified. Occasionally, case reports describing bullous fixed-drug eruption, acute generalized exanthematous pustulosis, bullous exanthema, Stevens-Johnson syndrome and toxic epidermal necrolysis have appeared in the literature [for review, see 21]. The reported frequency of non-serious late skin eruptions varies greatly due to data collection by means of questionnaires, which results in over-reporting of minor symptoms not necessarily related to the contrast medium itself. This is well illustrated in the study of Yasuda and Mune-

237

chika [22]. The frequency of late adverse reactions experienced by inpatients decreased from 9.6 to 2.5% in the group receiving plain CT and from 10.8 to 3.8% in the group receiving contrast medium-enhanced CT when the questionnaires were completed by radiologists rather than by the patients themselves. This study also demonstrated that late hypersensitivity symptoms such as itching and limited skin eruptions are also common after unenhanced CT. It has been postulated that exposure to X-rays during CT scanning may itself contribute to the development of cutaneous symptoms [20]. In the two largest prospective surveys so far conducted, the frequency of skin eruptions occurring from 1 h to 7 days after contrast medium exposure to non-ionic monomers was found to be 0.9 and 2.9%, respectively [18, 20]. Again, it is worth noting that the frequency was about threefold lower for carefully observed inpatients [20] than for outpatients asked to fill in a questionnaire themselves [18]. Hosoya et al. [18] in their survey of some 11,000 patients reported that more than half of the reactions (itching, fever and rash) occurred within 24 h and most within 3 days. The ENDA task force observed a similar time course in their retrospective survey of 84 patients with various types of late skin eruptions. No difference in time to onset of reaction could be detected for patients with or without previous contrast medium exposure [Brockow et al., in preparation]. Pathophysiology Most non-serious as well as severe drug-induced eruptions appear to be caused by drug-reactive T cells which, upon recognition of the drug or a drug metabolite, proliferate and subsequently orchestrate an inflammatory reaction in the skin [23]. Contrast media are no exception [17]. The evidence for this is as follows: (a) patients with late skin eruptions are frequently skin test-positive when challenged with the contrast medium that caused the reaction, either by patch or de-

238

layed intradermal tests [18, 19]; (b) the eruptions have the same appearance as other drug eruptions known to be T-cell-mediated; (c) infiltrates of T cells are present in affected skin and in positive skin test sites [19]; (d) contrast media are able to stimulate the proliferation of peripheral T cells from previous reactors in vitro [19, 24], and (e) contrast medium-specific CD4+ and CD8+ Tcell clones have been generated from such proliferating cultures [25]. The diversity of skin symptoms can be explained by the diversity of inflammatory cytokines secreted by the contrast medium-stimulated T cells and by the subset of T cells activated. While non-serious skin eruptions are caused mainly by cytotoxic CD4+ T cells, bullous skin eruptions are caused by massive activation of the more dangerous cytotoxic CD8+ T cells [26]. The frequently observed development of skin eruptions 1–2 days after contrast medium administration in previously unexposed individuals indicates that contrast media do not induce a primary immune response but instead stimulate easily activated memory T cells with fitting T-cell receptors raised against some peptide antigens [27]. The time to onset of skin eruptions will depend on the patient’s pool of such T cells. The contrast medium-induced skin eruptions are mainly mild because the patients are not repeatedly exposed to the agent over a prolonged period. However, a previous reactor is at risk of a more serious skin reaction if re-exposed to the eliciting contrast medium over a short time interval, due to the expansion of contrast medium-reactive T cells during the previous exposure(s). This is in line with the observation that repeat reactions often develop faster and are more widespread than the previous reaction. It is also well known that for a number of patients, the reaction will not recur on re-exposure to the contrast medium that caused the first reaction. This might be due to the lack of necessary costimulation during re-exposure, or simply a depletion of the circulating pool of contrast me-

Christiansen

dium-reactive T cells since the last exposure. In some patients, the pool of contrast medium-reactive T cells may be maintained for years [19], whereas in others these cells may decline rapidly and may no longer be detectable after some months [24]. A third possibility is initial misdiagnosis. In contrast to patients with immediate reactions, patients with late skin eruptions are frequently skin test-positive to a range of different products [19]. This means that a patient with a previous late skin reaction to one contrast medium may also react when given a different product. Recent in vitro cross-reactivity studies with contrast medium-specific T-cell clones have shown that the clinically observed cross-reactivity between different contrast media is due to the presence of T cells with broad cross-reactivity patterns [25]. However, since patients may harbor T cells with different cross-reactivity, it is impossible at present to predict which contrast medium the patient will tolerate based on structural considerations alone. Vernassiere et al. [28] postulated that the late skin reactions after contrast medium exposure could be due to a reaction to iodine itself, since 4 of their 15 patients were patch test-positive to iodine-containing disinfectants. However, if iodine liberated from the iodinated contrast medium molecules was the cause of these reactions, one would have expected that all iodinated contrast media should cross-react. However, only 1 of the 4 patients with a positive patch test to iodine-containing disinfectants was skin test-positive to all 7 contrast media used for testing. The 3 other patients reacted to 0/7, 1/7 and 2/7 of the tested contrast media, respectively. Studies with contrast medium-specific T-cell clones have shown that each clone has a distinct cross-reactivity pattern which in no case involves all contrast media tested [25]. When potassium iodide was added to cultures of such clones, no proliferation was observed [Lerch et al., in preparation]. It is therefore highly unlikely that either iodide (I–) or iodine

Contrast Media Hypersensitivity

(I2) is the cause of the late skin eruptions. The authors of a literature survey on iodine allergy published last year, concluded that asking a patient if he/she is ‘allergic to iodine’ is a question that should be avoided because its significance is null [29]. Treatment and Prevention Late hypersensitivity reactions are mainly mild to moderate skin eruptions which can easily be treated with oral or topical corticosteroids. If vesicles develop, the patient should be advised to consult a dermatologist. In such cases laboratory analyses should be performed to look for the involvement of internal organs such as liver and kidneys. Patients with previous contrast medium-induced late skin reactions are at risk of a new late skin reaction if re-exposed to the contrast medium that caused the previous reaction(s). Because of the often extensive cross-reactivity between different products, a repeat reaction may not be prevented by the simple change of contrast medium. For patients with generalized exanthemas, patch and intradermal tests with delayed reading on days 1–3, may be attempted in search for a contrast medium that will be tolerated. Undiluted products must be used for testing as 1/10 diluted products have resulted in false-negative results [28]. Even if skin test negative, the product that caused the previous reaction should not be re-injected, since the sensitivity and negative predictive value of these skin tests remains unknown. A validated treatment protocol for prevention of late skin eruptions is currently not available. It has, however, been suggested that for patients with previous mild to moderate skin eruptions, oral medication with 50 mg prednisone the day before and 25 mg prednisone daily for 3 days after contrast medium exposure may be useful. Since the reaction is delayed, continued treatment is necessary. If symptoms appear, the dose

239

should be increased to 50 mg. Since it is unknown how effective this treatment is in preventing severe reactions, patients with previous severe skin eruptions should not be re-exposed to any iodinated contrast medium. This advice should not pose a significant problem since the number of such patients is very low. Apart from previous late reactors, patients with well-documented exanthema to other drugs and those with contact dermatitis seem to be at risk of late skin reactions [19, 28]. This has also been observed for other drugs and may result from an inability of these patients to develop a tolerogenic immune response to xenobiotics [30]. Viral infection is likely to be another risk factor due to the massive stimulation of the immune system during such infections [27]. Severe eruptions have been reported in patients medicated with hydralazine and in patients with systemic lupus erythematosus, bone marrow transplantation and lymphoma [for review, see 21]. It is frequently stated that late skin eruptions occur more frequently in Japan than elsewhere with reference to clinical studies with iotrolan in Europe and Japan. However, the reported incidence of late reactions to iotrolan in Japan was no higher than the incidence in the US [31]. There is therefore a need to stop perpetuating the false statement that the Japanese are at special risk of contrast media-induced late skin reactions.

It should also be pointed out that patients with previous late hypersensitivity reactions are not at increased risk of an immediate hypersensitivity reaction and vice versa, since these reactions are caused by different mechanisms involving different types of cells. This has been confirmed in a large prospective survey [18].

Conclusion

It appears that most late skin eruptions are T-cellmediated while new data indicate that many of the immediate hypersensitivity reactions are IgEmediated. In accordance with good allergy practice, the following advice can therefore be given to radiologists and cardiologists: • Do not re-expose previous reactors to the contrast medium that caused the first reaction; • Use a skin test-negative product, if known; • Do not rely on pretreatment with corticosteroids/antihistamines to prevent anaphylaxis; • Make certain the radiological examination is carried out near resuscitation equipment and drugs; • Be aware of risk factors for late skin eruptions and advise patients at risk; • Avoid re-administration of iodinated contrast media in patients with previous severe late skin eruptions.

References 1 Katayama H, Yamaguchi K, Kozuka T, Takashima T, Seez P, Matsuura K: Adverse reactions to ionic and nonionic contrast media. A report from the Japanese Committee on the Safety of Contrast Media. Radiology 1990;175:621– 628. 2 The International Collaborative Study of Severe Anaphylaxis: Risk of anaphylaxis in a hospital population in relation to the use of various drugs: an international study. Pharmacoepidemiol Drug Saf 2003;12:195–202.

240

3 Cochran ST, Bomyea K, Sayre JW: Trends in adverse events after IV administration of contrast media. Am J Roentgenol 2001;176:1385–1388. 4 Shehadi WH: Contrast media adverse reactions: Occurrence, recurrence, and distribution patterns. Radiology 1982; 143:11–17. 5 Dewachter P, Mouton-Faivre C, Felden F: Allergy and contrast media. Allergy 2001;56:250–251.

6 Laroche D, Aimone-Gastin I, Dubois F, Huet H, Gérard P, Vergnaud MC, Mouton-Faivre C, Guéant JL, Laxenaire MC, Bricard H: Mechanisms of severe, immediate reactions to iodinated contrast material. Radiology 1998;209:183–190. 7 Laroche D, Dewachter P, MoutonFaivre C, Clément O: Immediate reactions following ICM injection: results of a 3-year prospective multicenter survey. Contrast Med Mol Imaging 2006;1: 81.

Christiansen

8 Brasch RC: Allergic reactions to contrast media: accumulated evidence. Am J Roentgenol 1980;134:797–801. 9 Pumphrey R: Anaphylaxis: can we tell who is at risk of a fatal reaction. Curr Opin Allergy Clin Immunol 2004;4: 285–290. 10 Thomsen HS, Morcos SK: Management of acute adverse reactions to contrast media. Eur Radiol 2004;14:476–481. 11 Brown SGA, Mullins RJ, Gold MS: 2. Anaphylaxis: diagnosis and management. Med J Aust 2006;185:283–289. 12 Tramèr MR, von Elm E, Loubeyre P, Hauser C: Pharmacological prevention of serious anaphylactic reactions due to iodinated contrast media: systematic review. BMJ 2006;333:675–680. 13 Brockow K, Christiansen C, Kanny G, Clément O, Barbaud A, Bircher A, Dewachter P, Guéand JL, Rodriguez Guéant RM, Mouton-Faivre C, Ring J, Romano A, Sainte-Laudy J, Demoly P, Pichler WJ: Management of hypersensitivity reactions to iodinated contrast media. Allergy 2005;60:150–158. 14 Kanny G, Maria Y, Mentre B, MoneretVautrin DA: Case report: recurrent anaphylactic shock to radiographic contrast media. Evidence supporting an exceptional IgE-mediated reaction. Allerg Immunol (Paris) 1993;25:425– 430. 15 Valfrey J, Newinger G, Arbogast R, Pauli G: Anaphylactic shock with ioxaglate (Hexabrix 320) during coronary angiography: Two cases. Rev Fr Allergol Immunol Clin 2002;42:157–162.

16 Dewachter P, Mouton-Faivre C: Anaphylactoid reactions to radio-opaque contrast media. Br J Anaesth 2002;88: 879. 17 Christiansen C, Pichler WJ, Skotland T: Delayed allergy-like reactions to X-ray contrast media: mechanistic considerations. Eur Radiol 2000;10:1965–1975. 18 Hosoya T, Yamaguchi K, Akutsu T, Mitsuhashi Y, Kondo S, Sugai Y, Adachi M: Delayed adverse reactions to iodinated contrast media and their risk factors. Radiat Med 2000;18:39–45. 19 Kanny G, Pichler W, Morisset M, Franck P, Marie B, Kohler C, Renaudin JM, Beaudouin E, Sainte Laudy J, Moneret-Vautrin DA: T cell-mediated reactions to iodinated contrast media: Evaluation by skin and lymphocyte activation tests. J Allergy Clin Immunol 2005;115:179–185. 20 Munechika H, Hiramatsu Y, Kudo S, Sugimura K, Hamada C, Yamaguchi K, Katayama H: A prospective survey of delayed adverse reactions to iohexol in urography and computed tomography. Eur Radiol 2003;13:185–194. 21 Böhn I, Shield HH: A practical guide to diagnose lesser-known immediate and delayed contrast media-induced adverse cutaneous reactions. Eur Radiol 2006;16:1570–1579. 22 Yasuda R, Munechika H: Delayed adverse reactions to nonionic monomeric contrast-enhanced media. Invest Radiol 1998;33:1–5. 23 Roujeau JC: Immune mechanisms in drug allergy. Allergol Int 2006;55:27– 33.

24 Pichler WJ, Tilch J: The lymphocyte transformation test in the diagnosis of drug hypersensitivity. Allergy 2004;59: 809–820. 25 Lerch M, Keller M, Britschgi M, Kanny G, Tâche V, Schmid D, Gerber BO, Bircher AJ, Christiansen C, Pichler WJ: Immunological mechanisms of delayed-type hypersensitivity reactions to iodinated contrast media. Contrast Med Mol Imaging 2006;1:81–82. 26 Pichler WJ: Immune mechanism of drug hypersensitivity. Immunol Allergy Clin North Am 2004;24:373–397. 27 Pichler WJ: Direct T-cell stimulations by drugs – bypassing the innate immune system. Toxicology 2005;209:95– 100. 28 Vernassiere C, Trechot P, Commun N, Schmutz JL, Barbaud A: Low negative predictive value of skin tests in investigating delayed reactions to radio-contrast media. Contact Dermatitis 2004; 50:359–366. 29 Dewachter P, Tréchot P, Mouton-Faivre C: ‘Iodine allergy’: point of view. Ann Fr Anesth Reanim 2005;24:40–52. 30 Gex-Collet C, Helbling A, Pichler WJ: Multiple drug hypersensitivity – proof of multiple drug hypersensitivity by patch and lymphocyte transformation tests. J Invest Allergol Clin Immunol 2005;15:293–296. 31 Niendorf HP: Isovist-280: clinical data and delayed allergy-like reactions. Eur Radiol 1996;6(insert to issue 5: Delayed allergy-like reactions to X-ray contrast media:4–8).

Dr. Cathrine Christiansen GE Healthcare Nycoveien 2, Postbox 4220 Nydalen NO–0401 Oslo (Norway) Tel. +47 2318 5671, Fax +47 2318 6008 E-Mail [email protected]

Contrast Media Hypersensitivity

241

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 242–250

Maculopapular Drug Eruptions Nikhil Yawalkar Department of Dermatology, University Hospital, Inselspital, Bern, Switzerland

Abstract Maculopapular (exanthematous) reactions are the most common adverse drug eruptions affecting the skin. Several studies indicate that immunological mechanisms including cytotoxic T cells (CD4+ > CD8+), both type 1 (e.g. IFN-
) and type 2 (e.g. IL-5) cytokines and various chemokines are critically involved in the pathogenesis of these eruptions. While maculopapular exanthems can virtually be elicited by any drug, antimicrobials (e.g. -lactam antibiotic, sulfonamides), anticonvulsants, allopurinol, and NSAIDs are most frequently involved. Clinical manifestations are variable and range from faint macules to widespread erythematous and maculopapular lesions, which usually begin on the trunk, neck and upper extremities and subsequently spread downwards in a symmetrical fashion. Although the clinical course is often relatively mild, these exanthems may sometimes progress to erythroderma or represent the beginning of even more severe drug reactions like Stevens-Johnson syndrome, toxic epidermal necrolysis or a drug rash with eosinophilia and systemic symptoms. In most cases, management includes early withdrawal of the offending drug and usually supportive treatment with emollients, topical corticosteroids and systemic antihistamines depending on the severity of the eruption. Allergological work-up is recommended to provide the patient with appropriate information about the causative drug and possible alternatives for future use. Copyright © 2007 S. Karger AG, Basel

Introduction

Cutaneous drug eruptions account for a large proportion of all adverse drug reactions, affecting about 2–3% of hospitalized patients [1, 2]. They are known to be the great imitators of diseases

and may encompass a large variety of clinical patterns, e.g. urticaria, macular or maculopapular, papulosquamous, lichenoid or pustular skin eruptions, fixed drug eruptions, photosensitive reactions, vesicular/bullous reactions, erythema exsudativum multiforme, Stevens-Johnson syndrome (SJS) or toxic epidermal necrolysis (TEN) [3]. Although some reactions are rather mild and mainly confined to the skin, others may be lifethreatening and involve further organs such as liver, kidney and bone marrow. Therefore, early diagnosis and, particularly, recognition of signs which differentiate less severe from severe reactions together with prompt withdrawal of the offending drug are essential to minimize morbidity. This chapter focuses on the etiopathogenesis, clinical and histopathological features as well as treatment of maculopapular exanthems (MPE).

Epidemiology

Maculopapular (often also referred to as exanthematous) reactions followed by urticaria are the most common adverse drug eruptions affecting the skin. Previous studies have revealed that MPE account for 31–95% of all drug-induced cutaneous eruptions [4, 5]. Albeit these different percentages, MPE were generally found to be the predominant cutaneous manifestation to drugs in these studies.

Causative Drugs and Cofactors

While virtually any drug can give rise to an MPE, antimicrobials (e.g. -lactam antibiotic, sulfonamides), anticonvulsants, allopurinol and NSAIDs are most frequently involved [6, 7]. Other common agents include benzodiazepines, captopril, fluroquinolonos, gold, lithium, oral antihyperglycemic agents, phenothiazines, quinidine and thiazide diuretics. Various cofactors which are related to concomitant diseases, the host, and treatment regimen may have an important role in drug hypersensitivity reactions including drug-induced MPE [1, 5–7]. In particular, viral infections (e.g. HIV, EBV, CMV) are well-recognized cofactors in the elicitation of MPE. Preexisting disease (e.g. connective tissue diseases) can lead to immune perturbance and enhance the risk of cutaneous drug eruptions. Various studies have shown that cutaneous reactions are more likely in women than men. Some reports also indicate that these reactions are more frequent in elderly patients. Host genetic factors linked to the major histocompatibility complex (e.g. HLA-B*5701 in abacavir hypersensitivity) or polymorphisms that alter immune responses (e.g. TNF-  promoter polymorphisms in abacavir and carbamazepin hypersensitivity) or possibly drug metabolism are associated with drug hypersensitivity reactions. With regard to treatment regimen usage of multiple drugs, higher dosage as well as intermittent and repeated administration may also represent further risk factors.

Pathophysiology

Drug-induced MPE are included in type B (‘bizarre’) reactions, which are considered to be unpredictable and occur in susceptible individuals. Immunological mechanisms play a central part in these drug-induced MPE, which may now be classified as a subtype of type IV (Coombs and

Maculopapular Exanthems

Gell) delayed cell-mediated hypersensitivity reactions [8]. Growing evidence obtained from immunohistological studies of acute skin lesions as well as from T cells derived from peripheral blood and skin of drug-allergic patients indicates that drug-specific, cytotoxic T cells (CD4+ > CD8+) are the dominant effector cells in these cutaneous drug eruptions [9–12]. Immunohistochemical studies have shown that the cell infiltrate in druginduced MPE is largely composed of CD3+ T cells (40–70%). There is a predominance of CD4+ T cells, which are mainly located in the perivascular dermis (fig. 1). Together with some CD8+ T cells, these CD4+ T cells are also found along the dermoepidermal junction zone and in the epidermis [11]. The infiltrating T cells appear highly activated, and express CD25 (-chain of the IL-2 receptor), human leukocyte antigen (HLA)-DR and adhesion molecules like the leukocyte function-related antigen 1, and L-selectin. An increase of circulating cutaneous lymphocyte-associated antigen-positive T cells is found in these reactions compared to healthy controls [12]. In addition to T cells, other leukocytes including different dendritic cell subpopulations, macrophages and eosinophils are also observed in the cell infiltrate. Interestingly, CD56+ NK cells, which have been found to be increased in severe blistering exanthems like TEN, are barely present in MPE. Among the resident cells, endothelial cells are activated and express various adhesion molecules such as E- and P-selectin, platelet endothelial cell adhesion molecule 1, and intercellular adhesion molecule (ICAM)-1 [11, 13]. Major histocompatibility complex (MHC) class II molecules and ICAM-1 are also found on keratinocytes and dendritic cells, possibly rendering these cells capable of drug presentation to T cells. Up to 20% of the infiltrating T cells in drug-induced MPE express perforin and granzyme B. These cytotoxic granule proteins trigger cell death by forming pores in the target cell membrane and inducing degradation of their DNA. Interestingly, perforin and granzyme B are observed in both CD4+ and

243

Fig. 1. Immunohistochemical studies of the cell infiltrate in drug-induced MPE show a predominance of CD4+ T cells, upregulation of MHC class II molecules activation and adhesion molecules as well as cytotoxic proteins like perforin and granzyme B. Alkaline phosphatase antialkaline phosphatase (APAAP) method for granzyme B and avidinbiotin-complex/alkaline phosphatase (ABC/AP) method for the other markers. Orig. magnif. !250.

CD8+ T cells and are found at the dermo-epidermal junction zone or in the epidermis, along with signs of keratinocyte cell damage [11]. Posadas et al. [14] also demonstrated an increased expression of perforin and granzyme B, but not of Fas ligand in MPE. Studies on T cells isolated from peripheral blood and skin of drug-allergic patients have demonstrated that most drug-specific T cells are also CD4+. They can contain perforin and are capable of killing autologous keratinocytes, which are pretreated with IFN-
to upregulate MHC class II molecules in a perforin-dependent manner [15]. Taken together, studies of drug-specific T cells eluted from the skin and

244

blood as well as immunohistochemical stainings strongly suggest a crucial role for cytotoxic T cells (CD4+ > CD8+) in the elicitation of MPE [11, 15–17]. Cytotoxic CD4+ T cells infiltrating into the epidermis may damage MHC class II and ICAM-1 expressing keratinocytes and thereby be responsible for some of the typical morphological changes seen in drug-induced MPE, namely the vacuolar interface dermatitis. Cytotoxic CD8+ T cells, which recognize drugs/drug metabolites in a MHC class I-dependent manner may also collaborate in inducing keratinocyte cell damage. However, their preferential activation may lead to more severe reactions like bullous exanthema

Yawalkar

Fig. 2. Clinical manifestations of maculopapular (exanthematous) eruptions are variable and range from faint macules to widespread erythematous and maculopapular lesions, which usually begin on the trunk, neck and upper extremities and subsequently spread downwards in a symmetrical fashion.

[18]. In addition to their cytotoxic function, T cells as well as other leukocytes and resident cells may also orchestrate skin inflammation through the release and induction of different cytokines and chemokines in MPE. A heterogeneous cytokine profile including both type 1 (e.g. IFN-
, TNF-) and type 2 (e.g. IL-5) cytokines as well as various chemokines (e.g. CCL11/eotaxin, CCL5/ RANTES, CCL27/CTACK) has been found in these drug eruptions [11, 19–21]. The increased IFN-
production provides an explanation for the upregulation of MHC class II on keratinocytes, which subsequently enables drug presentation to CD4+ T cells [15]. CCL27-CCR10 interactions have been reported to play a role in the recruitment of CD4+ and CD8+ T cells in MPE and particularly in severe exanthems like SJS and TEN [21]. IL-5 together with chemokines like eotaxin are involved in the recruitment, growth and differentiation of eosinophils. Although the presence of eosinophils in the infiltrate is suggestive of drug eruption, it is important to note that they may also be absent.

Maculopapular Exanthems

Clinical Features

Typically, the lesions consist of faint, pink or red to salmon-colored macules which may further develop into a maculopapular rash and progressively become confluent (fig. 2). However, the clinical features are often variable and lesions may be morbilliform, rubelliform or scarlatiniform resembling viral or bacterial exanthems. Furthermore, a polymorphous aspect including some annular or polycyclic, partly urticarial eruptions as well as purpuric lesions (especially on the limbs) may be found (fig. 3). The rash usually begins on the trunk, neck and upper extremities and subsequently spreads downwards in a symmetrical fashion. Lesions may sometimes also develop in the intertriginiuos areas, but palms, feet and mucous membranes are usually spared. Patients may also experience moderate to severe pruritus and fever. The eruption usually occurs between 4 to 14 days (peak 9–10th day) after initiation of a new therapy, but may begin later or even a few days after the drug has been

245

Fig. 3. Eruption can be polymorphous with some partly urticarial, purpuric, annular or polycyclic lesions especially on the limbs.

discontinued. In sensitized individuals who are re-exposed to the causative drug, skin lesions usually develop within 1–3 days. The reaction usually fades with desquamation within 1 or 2 weeks after withdrawal of the causative agent. However, in some cases these exanthems may progress to erythroderma or represent the beginning of even more severe drug reactions like SJS, TEN or a drug rash with eosinophilia and systemic symptoms (DRESS). In this context, rapidly evolving painful, dusky red or purpuric, partly infiltrated macules may represent an important initial sign for the development of TEN. In DRESS, skin lesions are usually more infiltrated and rather have the appearance of a maculopapular rash often with a follicular accentuation. Notably, such severe cutaneous eruptions may particularly develop when drugs are involved which are known to elicit these reactions like aromatic anticonvulsants or sulfonamides in DRESS. Therefore, all patients with drug-induced maculopapular or erythematous reactions, and especially those taking drugs which are known to elicit dangerous eruptions, should be carefully examined for markers of more severe reactions. Taken together, these particularly include skin pain or burning, widespread eruptions

246

(i.e. confluent erythema) affecting more than 60% of the body surface area, dusky red or purpuric macules, atypical target lesions, blisters or epidermal detachment, positive Nikolsky’s sign (epidermal detachment induced through lateral pressure, e.g. by the thumb), involvement of mucus membranes, facial edema, lymphadenopathy, arthralgia and high fever (i.e. 1 40 ° C). Laboratory results indicating a more serious reaction include eosinophilia, atypical lymphocytes and abnormal liver-function tests [22].

Differential Diagnosis

The differential diagnosis of drug-induced MPE should include acute viral and bacterial infections, collagen vascular disease, serum sicknesslike reactions, and acute graft-versus-host reactions depending on the patient’s history and associated clinical features. Particularly, viral exanthems (caused by e.g. paramyxovirus [measles], togavirus [rubella], Epstein-Barr virus, enteroviruses, adenovirus, HIV, HHV-6, parvovirus B19, CMV) may resemble drug-induced MPE. The polymorphous aspect and the confluence of lesions in drug-induced MPE as well as the pres-

Yawalkar

Fig. 4. Differential diagnosis should particularly include viral (e.g. rubella (a)) and bacterial (e.g. syphilis (b)) infections, which may closely resemble drug-induced MPE.

ence of peripheral blood eosinophilia favor a drug reaction. Furthermore, drug eruptions are more commonly observed in elderly persons whereas a viral cause is favored in the pediatric population. Bacterial infections, e.g. scarlet fever and syphilis (roseola) may mimic drug-induced MPE. Palmoplantar sites are rather uncommonly involved in drug-induced MPE and should alert the clinician to rule out secondary syphilis (fig. 4).

Histopathology

Fig. 5. Histological features of MPE consist of an interface dermatitis with a vacuolar alteration (synonym: hydropic degeneration) of the basal cell layer (arrowhead), individual dyskeratotic (apoptotic) keratinocytes (arrow) and a superficial perivascular mononuclear cell infiltrate at times with few eosinophils (inset). HE. Orig. magnif. !250.

Maculopapular Exanthems

MPE typically show a mild interface dermatitis with a vacuolar alteration (synonym: hydropic degeneration) of the basal cell layer [23]. In addition, mild spongiosis of the lower epidermis may be present along with scattered individual dyskeratotic keratinocytes as well as some infiltrating lymphocytes (fig. 5). These findings correlate with observations made by electron microscopy, where the primary changes of MPE have been reported to be intercellular and intracellular edema

247

as well as disruption of epidermal basal cells with some pyknotic nuclei [24]. The papillary dermis may show mild edema and a superficial, mainly perivascular, mild to moderate mononuclear cell infiltrate with some eosinophils. It should be noted that there are no absolute histological criteria for the diagnosis of drug-induced MPE. In addition, even if the observed histological changes are compatible with a drug-induced eruption, alternative causes cannot be definitely excluded since there is a considerable overlap with features seen in other entities. Moreover, epidermal changes may be minimal or even absent in some reactions like mild scarlatiniform eruptions. Thus, in cases with no severe signs or symptoms and a clear temporal relationship, history and the morphologic pattern of skin lesions are often sufficient for diagnosis and a skin biopsy is not mandatory. However, in complex and severe cases or when the precise morphology is unclear, histopathological findings may provide some clues and assist in reaching a correct diagnosis. Histopathologic findings that favor a drug eruption over a viral exanthema are the presence of tissue eosinophilia and the lack of dermal hemorrhage. In contrast to drug-induced maculopapular reactions, histopathologic findings of secondary syphilitic lesions demonstrate a plasma cell-rich mononuclear infiltrate. A connective tissue disease would be favored if epidermal atrophy, focal parakeratosis (= retention of nuclei in the stratum corneum), thickening of the basement membrane zone and mucinosis accompany the interface dermatitis. Epidermal atrophy, focal or confluent parakeratosis, hypergranulosis and necrotic keratinocytes may represent clues for an acute graft-versus-host disease.

Laboratory and Diagnostic Tests

confirming or ruling out clinical diagnoses. Therefore, depending on the clinical features, differential diagnosis and severity of the eruption, various laboratory tests including complete blood count (eosinophils, atypical lymphocytes), liver and renal function tests, anti-nuclear antibody, bacterial and viral serology can be performed. Different diagnostic methods, namely epicutaneous tests (patch or scratch-patch tests), lymphocyte transformation test (LTT) and occasionally provocation tests (rechallenge) can be performed after recovery in order to gain further information about the causative drug. In general, epicutaneous tests should be performed within 6 months after recovery. In maculopapular (exanthematous) reactions these tests have been reported to be especially useful with certain drugs including antimicrobial agents (e.g. -lactam antibiotics, sulfonamides, macrolides, imidazoles, gentamicin, isoniazid), antiepileptic agents (e.g. carbamazepine, phenytoin), cardiovascular agents (e.g. angiotensin-converting enzyme inhibitors, diltiazem) and allupurinol, among others [25–28]. These tests rarely induce adverse reactions and should particularly be considered when the drug is important. They might be negative in normal skin, but positive in previously affected skin areas. However, epicutaneous tests to many drugs are still rather poorly standardized and false-negative or false-positive results are possible. In addition to epicutaneous tests, LTT may also be helpful in determining the causative drug [28]. As with epicutaneous tests, validation of LTT is lacking for many drugs and this in vitro test is not readily available. Finally, provocation tests (rechallenge) may be considered in some cases. However, these tests require rather prolonged treatment or high doses to become positive, are time-consuming and hold the risk of inducing a more severe reaction.

Until today, there are no laboratory markers to specifically identify drug-induced MPE during the acute phase. However, blood tests may aid in

248

Yawalkar

Treatment

Treatment of MPE is generally supportive. First of all, withdrawal of the offending agent should be done as soon as possible. However, in many patients, several drugs may be incriminated making identification of the causative drug difficult. Therefore, withdrawal of all drugs that are not essential and particularly those newly commenced in the last weeks may be necessary. For mild cases, discontinuation of the offending agent and use of emollients may prove sufficient, with the rash fading over several days. For more serious reactions, systemic antihistamines and topical corticosteroids are given; and in severe, widespread reactions a short course of systemic corticosteroids may be added. In milder cases, when the causative drug is of paramount importance for treating a severe disease, it may be possible to ‘treat through’ the eruption without discontinuing the drug. However, some patients may progressively worsen and develop severe cutaneous eruptions like SJS or TEN. Therefore, the benefit/risk ratio of this therapeutic strategy must be carefully evaluated and patients must be strictly monitored for markers of more severe reactions. While desensitization procedures are mainly used and successful with immediate reactions [see chapters of Solensky, pp 404–412, and Castells, pp 413–425], desensitization may also

be considered in some cases with MPE where the drug is mandatory and no alternative is available. This strategy has particularly been described in patients with HIV infection with previous cutaneous reactions to sulfonamides [29].

Conclusion

The severity of MPE is variable and ranges from mild macular to widespread maculopapular and erythematous eruptions. Diagnosis is mainly based on a detailed clinical history suggesting that the onset of the eruption correlates temporally with drug ingestion as well as morphology and distribution of the skin lesions. Since some exanthems may progress into more severe drug reactions like SJS, TEN or DRESS, a high index of suspicion is required for recognition of markers of severity. This is especially important when drugs are involved which may particularly induce such reactions, e.g. aromatic anticonvulsants with regard to DRESS. Finally, after recovery, patients should receive clear information (i.e. allergy pass) regarding name of the causative drug, manner of identification (e.g. history, epicutaneous tests, LTT), potentially cross-reacting drugs and those which can be taken safely in the future.

References 1 Bigby M, Jick S, Jick H, Arndt K: Druginduced cutaneous reactions. A report from the Boston Collaborative Drug Surveillance Program on 15,438 consecutive inpatients, 1975–1982. JAMA 1986;256:3358–3363. 2 Hunziker T, Kunzi UP, Braunschweig S, Zehnder D, Hoigne R: Comprehensive hospital drug monitoring: adverse skin reactions, a 20-year survey. Allergy 1997;52:388–393.

Maculopapular Exanthems

3 Crowson AN, Brown TJ, Magro CM: Progress in the understanding of the pathology and pathogenesis of cutaneous drug eruptions: implications for management. Am J Clin Dermatol 2003;4:407–428. 4 Apaydin R, Bilen N, Dokmeci S, Bayramgurler D, Yildirim G: Drug eruptions: a study including all inpatients and outpatients at a dermatology clinic of a university hospital. J Eur Acad Dermatol Venereol 2000;14:518–520.

5 Bigby M: Rates of cutaneous reactions to drugs. Arch Dermatol 2001; 137:765– 770. 6 Nigen S, Knowles SR, Shear NH: Drug eruptions: approaching the diagnosis of drug-induced skin diseases. J Drugs Dermatol 2003;2:278–299. 7 Gomes ER, Demoly P: Epidemiology of hypersensitivity drug reactions. Curr Opin Allergy Clin Immunol 2005;5: 309–316.

249

8 Lerch M, Pichler WJ: The immunological and clinical spectrum of delayed drug-induced exanthems. Curr Opin Allergy Clin Immunol 2004;4:411–419. 9 Hertl M, Merk HF: Lymphocyte activation in cutaneous drug reactions. J Invest Dermatol 1995; 105:95–98. 10 Yawalkar N, Pichler WJ: Immunohistology of drug-induced exanthema: clues to pathogenesis. Curr Opin Allergy Clin Immunol 2001;1:299–303. 11 Yawalkar N, Egli F, Hari Y, Nievergelt H, Braathen LR, Pichler WJ: Infiltration of cytotoxic T cells in drug-induced cutaneous eruptions. Clin Exp Allergy 2000,30:847–855. 12 Blanca M, Posadas S, Torres MJ, Leyva L, Mayorga C, Gonzalez L, Juarez C, Fernandez J, Santamaria LF: Expression of the skin-homing receptor in peripheral blood lymphocytes from subjects with non-immediate cutaneous allergic drug reactions. Allergy 2000;55:998–1004. 13 Barbaud AM, Bene MC, Schmutz JL, Ehlinger A, Weber M, Faure GC: Role of delayed cellular hypersensitivity and adhesion molecules in amoxicillininduced morbilliform rashes. Arch Dermatol 1997;133:481–486. 14 Posadas SJ, Torres MJ, Mayorga C, Juarez C, Blanca M: Gene expression levels of cytokine profile and cytotoxic markers in non-immediate reactions to drugs. Blood Cells Mol Dis 2002;29: 179–189.

15 Schnyder B, Frutig K, Mauri-Hellweg D, Limat A, Yawalkar N, Pichler WJ: Tcell-mediated cytotoxicity against keratinocytes in sulfamethoxazole-induced skin reaction. Clin Exp Allergy 1998;28:1412–1417. 16 Yawalkar N: Drug-induced examthems. Toxicology 2005;209:131–134. 17 Pichler W, Yawalkar N, Schmid S, Helbling A: Pathogenesis of drug-induced exanthems. Allergy 2002;57: 884–893. 18 Hertl M, Bohlen H, Jugert F, Boecker C, Knaup R, Merk, HF: Predominance of epidermal CD8+ T lymphocytes in bullous cutaneous reactions caused by lactam antibiotics. J Invest Dermatol 1993;101:794–799. 19 Yawalkar N, Shrikhande M, Hari Y, Nievergelt H, Braathen LR, Pichler WJ: Evidence for a role for IL-5 and eotaxin in activating and recruiting eosinophils in drug-induced cutaneous eruptions. J Allergy Clin Immunol 2000;106:1171– 1176. 20 Posadas SJ, Leyva L, Torres MJ, Rodriguez JL, Bravo I, Rosal M, Fernandez J, Juarez C, Blanca M: Subjects with allergic reactions to drugs show in vivo polarized patterns of cytokine expression depending on the chronology of the clinical reaction. J Allergy Clin Immunol 2000;106:769–776. 21 Tapia B, Padial A, Sanchez-Sabate E, Alvarez-Ferreira J, Morel E, Blanca M, Bellon T: Involvement of CCL27-CCR10 interactions in drug-induced cutaneous reactions. J Allergy Clin Immunol 2004;114:335–340.

22 Roujeau JC, Stern RS: Severe adverse cutaneous reactions to drugs. N Engl J Med 1994 10;331:1272–1285. 23 Bronnimann M, Yawalkar N: Histopathology of drug-induced exanthems: Is there a role in diagnosis of drug allergy? Curr Opin Allergy Clin Immunol 2005;5:317–321. 24 Fellner MJ, Prutkin L: Morbilliform eruptions caused by penicillin. A study by electron microscopy and immunologic tests. J Invest Dermatol 1970; 55: 390–395. 25 Barbaud A: Drug patch testing in systemic cutaneous drug allergy. Toxicology 2005;209:209–216. 26 Hug K, Yawalkar N, Helbling A, Pichler WJ: Scratch-patch and patch testing in drug allergy – an assessment of specificity. J Investig Allergol Clin Immunol 2003;13:12–19. 27 Lammintausta K, KortekangasSavolainen O: The usefulness of skin tests to prove drug hypersensitivity. Br J Dermatol 2005;152:968–974. 28 Pichler WJ, Tilch J: The lymphocyte transformation test in the diagnosis of drug hypersensitivity. Allergy 2004;59: 809–820. 29 Belchi-Hernandez J, Espinosa-Parra FJ: Management of adverse reactions to prophylactic trimethoprim-sulfamethoxazole in patients with human immunodeficiency virus infection. Ann Allergy Asthma Immunol 1996; 76:355–358.

Dr. Nikhil Yawalkar Department of Dermatology, Inselspital CH–3010 Bern (Switzerland) Tel. +41 31 632 8042, Fax +41 31 381 5815 E-Mail [email protected]

250

Yawalkar

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 251–266

Drug-Induced Hypersensitivity Syndrome and Viral Reactivation Tetsuo Shiohara ⭈ Ryo Takahashi ⭈ Yoko Kano Department of Dermatology, Kyorin University School of Medicine, Tokyo, Japan

Abstract Drug-induced hypersensitivity syndrome (DIHS), often referred to as drug reaction with eosinophilia and systemic symptoms (DRESS), is a life-threatening multi-organ system reaction characterized by rash, fever, lymphadenopathy, hepatitis, and leukocytosis with eosinophilia. DIHS has a worldwide distribution but is undoubtedly underdiagnosed in many countries probably due to a lack of awareness. This syndrome has several unique features that cannot be solely explained by drug etiology: they include delayed onset, a paradoxical worsening of clinical symptoms after withdrawal of the causative drugs and unexplained crossreactivity to multiple drugs with different structures. These unique features suggest the additional role of viral infections. Indeed, we demonstrated that not only human herpesvirus 6 (HHV-6) but also other herpesviruses (EpsteinBarr virus, cytomegalovirus and HHV-7) can be reactivated in a sequential manner during the course of DIHS, as demonstrated in graft-versus-host disease. In this review, we will discuss recent work including our own, focusing on the interactions among what we considered the main players of inflammatory responses in DIHS: cross-reactive memory T cells and various herpesviruses. Copyright © 2007 S. Karger AG, Basel

In 1938, Merritt and Putnam [1] reviewed the toxic symptoms caused by therapy with phenytoin and noted that the symptoms could be divided into two cutaneous reactions: the first one being a mild, morbilliform eruption that healed when phenytoin was withdrawn and often did not re-

cur after readministration, and the other being a severe exfoliative dermatitis with fever and eosinophilia. During the next 15 years, it became clear that the second is also associated with lymphadenopathy and multivisceral involvement such as hepatitis, and Chaiken et al. [2] were the first to describe the systemic implications of this reaction. Since then there have been many case reports describing similar reactions to other anticonvulsants, and the reactions induced by relatively long-term therapy with therapeutic doses of anticonvulsants, such as phenytoin, phenobarbital and carbamazepine, were diagnosed using various terms. In 1988, Shear and Spielberg [3] coined the term ‘anticonvulsant hypersensitivity syndrome’ to refer to these diverse entities, and suggested that this reaction may be secondary to a genetic defect of drug metabolism. Bocquet et al. [4] introduced the term ‘drug reaction with eosinophilia and systemic symptoms’ (DRESS) for this syndrome to distinguish it from other severe drug reactions which are not associated with eosinophilia. Subsequently, during the past 10 years, the diagnostic criteria and clinical spectrum of this syndrome were defined: there have been no significant differences in the clinical and laboratory findings of the reported cases despite their different names. Although the terms DRESS and anticonvulsant hypersensi-

tivity syndrome are still used occasionally to describe this syndrome, we propose that the term ‘drug-induced hypersensitivity syndrome’ (DIHS) be used to replace these names to avoid confusion due to the lack of consensus in the literature about terminology [5, 6]. Although there are conflicting views on the pathogenesis of this syndrome in different parts of the world, recent studies [7, 8] including ours [3, 6, 9, 10] suggest an intimate relationship between human herpesvirus 6 (HHV-6) and the development of DIHS. In this review, therefore, much emphasis has been placed on not only the clinical symptoms of DIHS but also the possible etiologic role of herpesviruses including HHV-6 in the development of DIHS.

Epidemiology

The incidence of DIHS is estimated to be between 1 in 1,000 and 1 in 10,000 exposures to phenytoin [11]. DIHS has a worldwide distribution but it is undoubtedly underdiagnosed in many countries. Although this syndrome is much more common than Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN), its true incidence remains unknown because its variable presentation and diverse clinical features and laboratory abnormalities have resulted in inaccurate reporting [12]. Since this syndrome in its florid form has become widely recognized in Japan, the reported incidence has risen sharply. Although the incidence of DIHS has been suggested to be highest in elderly black men [13], it remains undetermined whether a racial predilection could exist. Generally, DIHS has no age or sex predilection. There were no seasonal variations. Our series with this syndrome showed no increased incidence of a personal or family history of atopy and drug eruption. About half of the patients has had an infection within the previous 6 weeks, most commonly a flu-like illness.

252

Clinical Features

The syndrome typically occurs with fever or cutaneous lesions 3 weeks to 3 months after starting therapy with a limited number of drugs (fig. 1), mainly anticonvulsants (table 1). Significant differences exist among these drugs with regard to the potential to cause DIHS depending on race: in Japan minocycline rarely causes DIHS while mexiletine frequently does, and this is precisely the opposite of what happens in Caucasians. Although these drugs appear to have no common epitopes, they have in common the capacity to inhibit B-cell differentiation into immunoglobulin (Ig)-producing cells [9]. The delayed onset in relation to the introduction of the causative drug is one of the important features of this syndrome that can be distinguished from other types of drug eruptions which usually start 1–2 weeks after starting therapy. Sometimes there might be a prodrome with upper airway infection. More severe and rapid reactions occur in the setting of rechallenge in previously sensitized individuals or with continued use of the offending drugs [14– 16]. However, we have never seen patients who developed DIHS within 2 weeks after starting therapy. DIHS has even been reported in patients who took anticonvulsants for up to 40 years [17]. The cutaneous eruption usually begins as patchy erythematous macules, which may be slightly pruritic and can become confluent. Fever usually precedes the rash by several days and temperature ranges from 38 to 40 ° C with spikes that usually generate concern regarding an underlying infection [12]. The fever often persists even for weeks despite discontinuation of the causative drugs. The face, upper trunk and upper extremities are initially affected and followed by involvement of the lower extremities [12]. Periorbital and facial edema with pinhead-sized pustules, reminiscent of acute generalized exanthematous pustulosis, is one of the characteristic features of early cutaneous lesions in DIHS. Palms and soles are usually spared, but can occasionally show a few

Shiohara ⴢ Takahashi ⴢ Kano

Serum Ig level B cell NK cell

Drug 3 weeks <

Eruption

Fever Lymphadenopathy Hepatitis Eosinophilia

HHV-6 DNA

HHV-6 IgG

2–3 weeks after onset

Fig. 1. Typical clinical course of DIHS. Patients develop rashes and other clinical symptoms more than 3 weeks after starting therapy with a limited number of drugs as demonstrated in table 1. At onset, a decrease in serum Ig levels, B and NK cell numbers can be seen in the majority

of patients. After withdrawal of the causative drug, a dramatic paradoxical deterioration of clinical symptoms can often be observed, coincident with a rapid increase in CD4+ T-cell number. Modified from Shiohara et al. [6].

Table 1. The main causative drugs of DIHS [6]

lesions. The eruption often generalizes into severe exfoliative dermatitis or erythroderma, which usually occurs with continued treatment with the causative drug after this syndrome has developed. Tender lymphadenopathy can be seen in 170% of patients early in the illness, predominantly affecting the cervical nodes or can be generalized [12, 16, 17]. Follicular accentuation of the erythematous papules is a characteristic finding of DIHS [18]. Mucosal surfaces are not usually affected which helps to distinguish this syndrome from other forms of severe drug eruptions such as SJS and TEN. Nevertheless, mucosal surfaces, particularly lips and oral mucous membranes, may sometimes show a few lesions so that their involvement does not necessarily exclude the diagnosis of DIHS. Bilateral swelling of the salivary glands with severe xerostomia has frequently

Carbamazepine Phenytoin Phenobarbital Zonisamide Lamotrigine Allopurinol Dapsone Salazosulfapyridine Mexiletine Minocycline Abacavir1 Nevirapine1 1

Both drugs cause systemic hypersensitivity reactions with some peculiar symptoms. See text, table 2 and Peyrière et al. [19].

DIHS and Virus

253

been seen early in the illness, suggesting that reactivation of mumps virus may occur before onset of this syndrome. Some patients may also complain of dryness of the mouth which makes swallowing of food and even talking difficult. This dryness is probably due to severe xerostomia. The severity of the clinical symptoms at onset provides only a guide to prognosis and is not absolute, because the more severe reactions often occur 3–4 days after withdrawal of the causative drug. This paradoxical worsening of clinical symptoms after withdrawal of the causative drug is also characteristic of this syndrome and, in those with a febrile illness, may at first be mistaken for severe infectious diseases by the doctor of first contact who may not have seen such a case. Due to this high level of suspicion of infection, unnecessary empirical antibiotic therapy may be started, which increases the risk of developing additional drug rashes. Indeed, patients with DIHS often show unexplained cross-reactivity to multiple drugs with different structures, including those used after onset of symptoms. Hepatic involvement is common. Hepatomegaly accompanied by splenomegaly is a common finding. Thus, the onset of the symptoms is highly variable; usually patients develop two or three symptomatic features followed by stepwise development of other symptoms (fig. 1). In most cases, withdrawal of the causative drug at an early stage is not followed by a rapid clearing of symptoms, unlike in other mild forms of drug eruptions. Many patients may continue to deteriorate or even show dramatic paradoxical worsening of clinical symptoms even for weeks and months after withdrawal of the causative drugs. With resolution, desquamation occurs, but this may often be followed by flare-ups. Recently, a survey of the reported clinical patterns of DIHS has revealed some clear distinctions attributable to the drug used [19]. In this survey of cases from France and in the general literature, the incidence of eosinophilia was rather low in lamotrigine-induced DIHS, and other

254

clinical features were different as well. Allopurinol rather frequently induced kidney disease, and minocycline often induced massive lymphadenopathy (table 2) and thrombocytopenia in a quarter of the cases reported. Indeed, if one includes the severe hypersensitivity reactions elicited by nevirapine and abacavir in HIV-positive patients as well in this syndrome, the heterogeneity of symptoms and the relationship to the use of a certain drug becomes even more striking [19].

Laboratory Findings

Leukocytosis with atypical lymphocytes and eosinophilia of various degrees are prominent features of this syndrome. Nevertheless, leukopenia or lymphopenia have also been reported [20] and this occasionally precedes leukocytosis. A leukemoid reaction may occur in some patients. The lymphocytosis is primarily due to an increase in either CD4 or CD8 T-cell counts. Eosinophilia can be seen in 60–70% of the patients and may have a delayed onset of up to 1–2 weeks [12], often occurring after elevations in liver enzymes have returned to baseline. Liver abnormalities occur in up to 70% of patients and are characterized by a marked increase in the serum alanine aminotransferase value, while the hepatitis is usually anicteric. The degree of hepatitis may be related to the interval between the onset of this syndrome and withdrawal of the causative drug [21], although this relation cannot be confirmed in our series of patients. The development of severe hepatitis with jaundice increases the risk of reported mortality [22, 23]. Various forms of renal involvement have also been reported [24, 25], ranging from tubulointerstitial nephritis to granulomatous necrotizing angiitis. The mortality from DIHS can be approximately 20% [26] and has been correlated with the degree of renal involvement rather than hepatic involvement. Other features of DIHS include a dramatic decrease in serum IgG, IgA, and IgM levels: the de-

Shiohara ⴢ Takahashi ⴢ Kano

Table 2. Main clinical and biological data (%) of cases reported in the French Pharmacovigilance Database (FPD) and in the literature Drug

Carbamazepine Phenytoin Phenobarbitone Lamotrigine Allopurinol Minocycline Nevirapine Abacavir

Fever

FPD Literature FPD Literature FPD Literature FPD Literature FPD Literature FPD Literature FPD Literature FPD Literature

65 84 43 97 80 100 100 100 23 95 94 100 59 83 93 66

Eruption

Adenopathy

100 84 100 97 80 100 100 100 95 94 94 94 73 100 78 100

22 53 14 21 20 25 28 10 10 0 78 83 4 16 4 0

Liver1

73 66 57 73 80 75 28 65 62 88 72 78 64 100 18 0

Eosinophilia

Lung2

Kidney3

92 53 100 58 100 75 0 21 62 60 72 61 73 83 9 0

11 19 0 18 0 12 14 0 0 0 11 33 0 33 10 66

8 9 0 18 0 0 14 27 43 84 17 17 4 0 3 33

From Peyrière et al. [19]. 1 Liver involvement, i.e. increase in liver enzymes, especially aspartate aminotransferase and alanine aminotransferase. 2 Lung involvment, i.e. pneumopathy with eosinophilia, cough, dyspnea. 3 Kidney involvement, i.e. acute renal failure, proteinuria, interstitial nephritis.

crease is typically observed at onset and the lowest levels are usually detected several days or a week after withdrawal of the causative drug [6, 9]. After the nadir in the decrease, the overshoot in IgG levels can transiently be observed within 1–2 weeks and their levels eventually return to normal upon full recovery (fig. 1).

Pathological Findings

The histologic picture of DIHS is superficial perivascular lymphocytic infiltrates composed mainly of lymphocytes with some extravasated erythrocytes or eosinophils but is not diagnostic. Depending on biopsied lesions, variable degrees of focal spongiosis associated with a lichenoid infiltrate can be seen, but severe epider-

DIHS and Virus

mal damage as frequently seen in SJS and TEN is never detectable. Occasionally dense perivascular lymphocytic infiltrates consisting of T cells can be marked, mimicking mycosis fungoides. Most of the lymphocytes are CD4+ and CD8+ T cells. In our first report on HHV-6 reactivation in DIHS, expression of HHV-6 DNA and viral antigens was documented in the tissue specimens obtained from cutaneous lesions in earlier phases of disease [5], suggesting that this organism plays a part in causing cutaneous lesions in DIHS. Nevertheless, the presence of viral DNA in the tissue cannot be taken as proof of causation of the lesions, because the possibility that viral DNA might enter the cutaneous lesions as a part of the systemic infection or by seeding from a site elsewhere in the body cannot be excluded. However,

255

Table 3. Diagnostic criteria for DIHS established by a Japanese consensus group [6] 1

Maculopapular rash developing >3 weeks after starting with a limited number of drugs

2

Prolonged clinical symptoms after discontinuation of the causative drug

3

Fever (>38° C)

4

Liver abnormalities (ALT >100 U/l)1

5

Leukocyte abnormalities (at least one present) a Leukocytosis (>11 ! 109/l) b Atypical lymphocytosis (>5%) c Eosinophilia (>1.5 ! 109/l)

6

Lymphadenopathy

7

HHV-6 reactivation2

The diagnosis is confirmed by the presence of the seven criteria given above (typical DIHS) or five (1–5) of the seven criteria (atypical DIHS). 1 This can be replaced by other organ involvement, such as renal involvement. 2 Reactivation is detected in the 2nd to 3rd week after the start of symptoms.

HHV-6 DNA was not detected in the blood obtained simultaneously with the cutaneous lesions.

Diagnosis

The diagnosis of DIHS itself is usually not difficult for the physician who is familiar with this syndrome, if it is seriously considered in every patient on anticonvulsant therapy who presents with fever, skin rashes, lymphadenopathy, and the symptoms described above. The clinical picture of this syndrome, however, is not so distinctive that the diagnosis of this syndrome can be made at first glance. The clinical findings alone can be difficult to differentiate from other viral exanthemas, such as infectious mononucleosis

256

(IM) which is caused by Epstein-Barr virus (EBV) which occurs most commonly in teenagers and young adults. The incidence of skin rashes in IM increases with the concomitant use of antibiotics such as ampicillin [27, 28]. In these patients, the eruption usually occurs 2–10 days after starting therapy with antibiotics which rarely cause DIHS, while DIHS typically occurs 3 weeks to 3 months after ingestion of an anticonvulsant. Thus, the interval between the introduction of the causative drug and the onset of clinical symptoms is of diagnostic value. DIHS is often missed in the differential diagnosis of patients presenting with fever, rash and lymphadenopathy, probably due to a lack of awareness. A Japanese consensus group has recently established a set of criteria for diagnosis (table 3) [6]. The diagnosis of DIHS is based on seven criteria: maculopapular rashes (MRs), prolonged clinical symptoms, fever, leukocytosis, elevation of liver enzymes (most commonly alanine aminotransferase), lymphadenopathy, and – albeit delayed – an increase in HHV-6 IgG titers. A probable diagnosis requires the presence of five of these seven criteria, and a definite diagnosis requires all seven. In most patients with obvious findings, little disagreement exists about the diagnosis of DIHS among dermatologists without examining HHV-6 IgG titers. In patients with milder clinical symptoms, however, diagnosis can be difficult without the data on HHV-6. Although the presence of lymphadenopathy is characteristic of DIHS, it is also a feature of Kikuchi-Fujimoto disease (KFD) which is an idiopathic, selflimiting lymphadenopathy associated with high fever and occurring most often in young women [29, 30]. Interestingly, many infectious agents have also been identified in association with KFD including parvovirus B19, HHV-6, and EBV [30, 31], as demonstrated in DIHS, and adverse drug reactions have been reported to frequently occur in this setting. Thus, the clinical picture of KFD appears to share similarities with that of DIHS. Failure to make the correct diagnosis often leads

Shiohara ⴢ Takahashi ⴢ Kano

VZV CMV HHV-7 EBV HHV-6

DIHS

HSV VZV CMV EBV HHV-6 1

Fig. 2. Sequential reactivations of herpesviruses in DIHS and GVHD.

to unnecessary investigations and therapeutic trials with antibiotics, which may increase the risk of developing additional drug allergy and may aggravate the clinical course of the disease.

Reactivation of Herpesviruses

A range of viruses have long been implicated as possible triggers for DIHS due to close clinical similarities between DIHS and IM. However, the idea that viruses can trigger this syndrome seemed heretical because no viral DNA was detected in association with the development of DIHS. This view began to change 8 years ago when we and others [5, 6] demonstrated that HHV-6 can be specifically reactivated in patients with DIHS. In addition to the increase in serologic IgG titers, HHV-6 DNA has been detected in the blood and skin of a patient with DIHS by PCR and in situ hybridization [5]. Because in the earlier studies HHV-6 was the only virus that was found to be reactivated in patients, reactivation of HHV-6 has been specifically implicated in the pathogenesis of DIHS. Sporadic reports, however, demonstrated that not only HHV-6 but also other herpesviruses, such as HHV-7 [5], EBV [32]

DIHS and Virus

2

After bone marrow transplantation 3

4

5 6 Time, months

12

24

and cytomegalovirus (CMV) [33], can be reactivated in the course of DIHS. In order to determine which herpesviruses can initially reactivate, we performed longitudinal real-time PCR analyses of viral loads in blood samples sequentially obtained from patients with DIHS. Our results of PCR analyses showed that various herpesviruses sequentially reactivate during the course of DIHS in most patients: the cascade of reactivation events initiated by EBV or HHV-6 extends, with some delay, to HHV-7 as well, and eventually to CMV [10] (fig. 2). In some patients, clinical manifestations of this syndrome occurring despite discontinuation of the causative drug coincided with herpesvirus reactivations, while reactivation events were not necessarily associated with overt symptoms in other patients. This indicates that the timing of reactivations is often unknown and may pass unnoticed because the ‘illness’ itself caused by the reactivation may be subclinical in many cases. Therefore, it is quite difficult to demonstrate full cascades of reactivation events by conventional weekly sampling of blood, because the change in viral load in the blood of patients with DIHS would be so rapid that we could not identify the true peak viral load.

257

These data show that reactivations of various herpesviruses are frequent events in DIHS and that they may contribute to the perpetuation and sometimes even the aggravation of symptoms after cessation of the drug allergy-causing therapy. What remains unclear is the role of herpesviruses for the start of DIHS: two possibilities exist. • DIHS is started by an ‘allergic’ immune reaction to the incriminated drugs (table 1) which seem to have a substantial ability to stimulate T cells [34, 35]. In the frame of this massive T-cell stimulation herpesviruses, harbored in T cells, are reactivated, as T-cell stimulation may reactivate the viral genome within the cells. This reactivation leads to activation of the herpesvirusspecific immune response, mediated by antibodies and T cells. This would explain why many different herpesviruses are activated and why in another massive immune process, namely graftversus-host disease (GVHD), a similar reactivation can be seen. • An alternative view, which we prefer, would attribute a more causal role to herpesviruses. Reactivation may occur, which is initially clinically unapparent. However, the virus-stimulated T cells show a substantial cross-reactivity with certain drugs, and the provision of the drugs leads to an expansion of these drug (and virus)-specific T cells, which even continues after cessation of the drug due to the persistence of the viral antigens. The simultaneous appearance of multiple concomitant reactivations could be explained by the ability of HHV-6 and HHV-7 to reactivate heterologous viruses [36]. These herpesviruses might be functionally linked in vivo, the reactivation of one leading to the reactivation of the other, thus explaining the common detection. These considerations raise the possibility that the clinical and biological manifestations originally attributed to HHV-6 or EBV can be ascribed to HHV-7 or CMV, and vice versa. If the symptoms of DIHS were indeed mediated by either gene products of various herpesviruses or the immune response to viral replication, then

258

frequent deterioration or several flare-ups occurring despite discontinuation of the causative drug could be explained, at least in part, by sequential reactivation of these herpesviruses.

The Role of Antiviral T Cells in the Pathogenesis of DIHS

Importantly, the sequential occurrence of herpesvirus reactivation has also been reported in the setting of GVHD [37–40] in the same sequential order as described here for DIHS (fig. 2). Interestingly, recent studies have provided strong evidence suggesting a role for viral infections in the emergence of alloreactive T cells and the development of GVHD [37–39], although the literature contains conflicting evidence regarding the role. Indeed, clinical symptoms of GVHD are often indistinguishable from viral eruptions. Postulated mechanisms for virus-induced GVHD include activation of donor-derived antiviral T cells cross-reactive with alloantigens secondary to an encounter with the alloantigens or viral antigens in the recipients, and bystander activation of the antiviral T cells by massive cytokine production during viral infection [40, 41]. It is difficult, however, to determine whether antiviral T cells are responding as a result of bystander mechanisms, or recognition of the relevant viral antigens or cross-reactive stimulation with alloantigens. Among the various viruses, the herpesvirus family is the most likely candidate responsible for the development of GVHD because herpesviruses can induce and maintain a potent memory Tcell response due to their common properties of ubiquitous prevalence in human populations and the capacity to grow in lymphoid cells [42, 43]. Indeed, longitudinal studies of bone marrow transplant (BMT) patients demonstrated that increased levels of the herpesvirus genome can be detected in serum, PBMC, and skin biopsy specimens from BMT patients, coincident with clinical symptoms of GVHD [37–40]. In this regard,

Shiohara ⴢ Takahashi ⴢ Kano

it is of note that reactivations of herpesviruses sequentially observed after BMT cannot necessarily be reflected in the increase in serum antivirus IgG titers, because previous studies on the recovery of immune cells following BMT showed that IgG production was more delayed than the appearance of B cells in the periphery and first became normal 7–9 months following BMT [44]. This result clearly indicates that the sequential occurrence of herpesvirus reactivations cannot be detected by serologic tests in the setting of GVHD and the acute stage of DIHS, at which time reversible hypogammaglobulinemia is specifically detected. Thus, real-time PCR assays constitute a useful method for precisely quantifying viral loads in the blood sequentially obtained from patients with DIHS. Nevertheless, the presence of viral DNA in blood during the course of disease cannot be taken as proof of causation of the disease. Irrelevant viruses might enter the blood from a site elsewhere in the body. How, then, can an etiological role be confirmed? The existence of a virus-specific immune response, particularly if it is selective for the affected tissue or relevant to the known features of disease pathogenesis, constitutes strong circumstantial evidence. In recent years, the role of antiviral T cells in allograft rejection and GVHD has received increasing attention, although similar attention has not been drawn to the role in DIHS. In support of this theory, several reports provide evidence that cross-reactivity can be observed between certain viruses and alloantigens: Burrows et al. [45–47] demonstrated that human clones specific for an EBV peptide presented in the context of HLA-B8 also recognize three common allogeneic HLA molecules (B14, B44, or B35). Further evidence of the importance of cross-reactive memory T cells has also been provided: in healthy seropositive donors, estimated precursor frequencies of cytotoxic T cells against EBV and CMV are roughly equivalent to those estimated against individual allogeneic HLAs [46, 48, 49]. Many aspects of DIHS, including

DIHS and Virus

clinical manifestations, hypogammaglobulinemia and the sequential occurrence of herpesvirus reactivations, further suggest close similarities between GVHD and DIHS. Keeping these considerations in mind, we suggest that virus-specific memory T cells that can cross-react with drug antigens could play a role in DIHS. Given that HLA-B*1502 was identified as an ‘immunogenic’ for carbamazepine-induced SJS and DIHS on the basis of a strong association between them in a large registry analysis [50], it might be expected that memory T cells specific for a viral peptide presented in the context of certain HLAB could be involved in the pathogenesis of DIHS due to their cross-reactivity to drug antigens. If so, an individual’s T-cell repertoire capable of reacting with drug antigens could be shaped by an individual’s past history of viral infections and their persistence in the host. Thus, even after the virus is cleared, the severity of clinical symptoms of DIHS could be increased by expansions of such cross-reactive T cells initially induced by past viral infections and subsequent exposure to certain drugs. Expansions of such cross-reactive memory T cells would occur upon reactivations of various herpesviruses before, at the time of, or subsequent to the onset of this syndrome. Inversely, these cells could be retained as a long-term drugspecific memory T-cell population by stimulation with herpesviruses in patients who are no longer exposed to the relevant drug antigens. This hypothesis would explain why the majority of patients with DIHS maintain massive proliferative responses to the causative drug reaction without stimulation by the drug antigens, as described in the following section.

In vitro Detection of Drug-Specific T Cells in DIHS

Solely on clinical grounds, it is difficult to determine which drug(s) elicits severe adverse reactions in patients with DIHS. The most important

259

step for the diagnosis is to obtain an accurate medical history, including information on dates of administration and discontinuation, dosage, and any previous drug exposure history. Because drugspecific T cells are generally thought to play a central role in mediating these reactions [5–10], patch tests and the lymphocyte transformation tests (LTTs) have often been used for the diagnostic assessment of drug-specific T-cell responses in clinical settings [51]. Laboratory-based in vitro technology such as LTTs offers numerous advantages over patch tests, including absolute safety, simultaneous assessment of T-cell responses to multiple drugs, and a lack of risk of developing additional drug allergies [51]. Nevertheless, a number of practical concerns related to the time of testing and the influence of therapy continue to limit their frequent application in clinical diagnosis. We therefore sequentially performed LTTs during the evolution of the disease in patients with DIHS and those with SJS/TEN and MPs. LTT was also repeated several months to 1 year later. In the vast majority of patients with SJS/ TEN and MPs, positive LTT reactions were obtained when the test was performed within 1 week after onset of rashes. Although at that time all of the patients with SJS/TEN were on therapy with oral prednisone (30–60 mg/day), the proliferation of lymphocytes appeared not to be influenced by prednisone. In contrast, negative results were constantly observed in patients with DIHS, when the test was performed within 1 week after onset, regardless of whether patients were on therapy with prednisone [Kano Y et al., in preparation]. Even when the tests were repeated at 2 weeks after onset, only 1 of 11 patients with DIHS showed a positive result, although the LTT reactions were somewhat enhanced in some patients. At the 5- to 7-week measurement, most of the patients with SJS/TEN and MPs did not show positive reactions while, with the exception of 1 patient, the majority of patients with DIHS exhibited positive reactions. In many of the patients with DIHS, a positive result was noted for the first time

260

5–7 weeks after onset and positive reactions continued to increase after that. Even at the 1-year measurement, the levels of LTT remained unchanged in patients with DIHS. Similar results were observed in patients who had not been treated with oral prednisone, confirming that treatment modalities did not affect the results. To determine whether the loss of LTT reactions to the causative drug during the acute stage of DIHS was associated with a global loss of responsiveness of mitogenic stimulation, we also examined the capacity of PBMC from the same patients who respond to stimulation by the mitogen PHA. Responses to PHA were strong in all patients regardless of whether PBMC were obtained at either the acute or convalescent stage. These results indicate the importance of the timing of the LTT performed during the course of the disease, depending on the type of drug reactions: the utility of the LTT in the diagnosis of drug allergy is much more promising than previously believed, in so far as the test is performed at the right time. One of the most important questions raised by these findings is why positive LTT reactions in patients with DIHS could be observed upon recovery, but not in the acute stage, and persist for months to a year while in other types of drug reactions LTT reactions became negative 5–7 weeks after onset. Considering that the sequential occurrence of herpesvirus reactivations was specifically observed in DIHS but not in SJS/TEN, expansions of cross-reactive T cells could be maintained for a long time after recovery by repeated stimulation with reactivation of several herpesviruses which may occur subclinically even during the convalescence stage. However, if this is the case, it is unclear why patients with DIHS, unlike those with other types of drug reactions, consistently showed negative LTT reactions during the acute phase. These results could be explained by assuming that regulatory T cells (Treg) that can suppress proliferation of memory T cells are expanded in the acute stage of DIHS and that at the end of the immune response the majority of the

Shiohara ⴢ Takahashi ⴢ Kano

expanded Treg population would be removed by apoptosis, which leads to the persistence of memory T-cell populations. Indeed, our preliminary results show that a Treg population with a naive phenotype (CD4+CD25brightFoxP3+) was expanded during the acute stage and decreased upon clinical recovery [Takahashi R et al., in preparation]. An increase in Treg frequency correlated with the disease state. In contrast, the frequency of Treg was relatively low in patients with SJS even in the acute stage and was not different from that in healthy individuals. These results suggest that the expansion of circulating Treg is specific enough to distinguish DIHS from other types of drug reactions.

DIHS as Another Manifestation of Immune Reconstitution Syndrome

The immune reconstitution syndrome (IRS) has been mainly described in HIV-infected patients [52–54]. IRS occurring in this setting is characterized by a paradoxical deterioration of clinical symptoms or laboratory parameters attributable to restoration of the host immune system following the introduction of highly active antiretroviral therapy [52–54]. The onset of this syndrome is usually associated with increased CD4+ T-cell counts and a reduction in HIV viral loads [55]. Reconstitution of a valid immune response against previously unrecognized viruses would reduce viral loads on the one hand, but cause tissue damage at sites of subclinical infection that had not been clinically recognized before starting antiretroviral therapy on the other. Given that paradoxical worsening of clinical symptoms associated with a reduction in viral loads typically observed in IRS is also the hallmark of DIHS, it is attractive to suppose that DIHS is a manifestation of the newly observed IRS. Despite the usefulness of recognizing DIHS as another manifestation of IRS, little attention has been focused on the similarity between DIHS and IRS.

DIHS and Virus

In order to expand the spectrum of IRS to include DIHS, it is important to know to what extent DIHS resembles IRS. To this end, we investigated whether a dramatic paradoxical deterioration of the clinical symptoms and laboratory parameters could be specifically observed after withdrawal of the causative drug, along with an increase in CD4+ T-cell numbers in 12 patients with DIHS. Our results show that dramatic paradoxical deterioration of clinical symptoms associated with no detection of virus genome was typically observed in the majority of patients with DIHS 3–4 days after withdrawal of the causative drug. In many patients, the initial increase in CD4+ Tcell number was seen within 2 weeks after onset, followed by the subsequent decrease along with improvements in clinical status (fig. 3). In some patients, CD8+ T-cell numbers initially increased and then gradually declined, reaching normal values by 2 months after onset. These alterations in lymphocyte subsets during the observation period were not related to the use of oral prednisone for the treatment of DIHS, because similar alterations were observed in patients who never received prednisone. The degree of the increase in CD4+ T-cell numbers is best correlated with the severity of clinical symptoms such as the extent of skin lesions at the acute stage. Our virologic results in 12 patients with DIHS showed that no herpesvirus (EBV, CMV, HHV-6, HHV-7, herpes simplex virus, varicella zoster virus) DNA was detected at their first presentation by real-time PCR assays [Kano et al., in preparation]. In some patients, either HHV-6 or EBV DNA was initially detected in the blood around 1–2 weeks after onset. Because this detection of viral DNA in the blood was followed by the significant rise in herpesvirus-specific IgG titers but not IgM, the detection of herpesviruses can be interpreted as reactivations of latent herpesviruses. Regardless of the severity of the initial clinical presentation and treatment, HHV-DNA was detected 2–3 weeks after the onset of rashes in the vast majority of patients (fig. 1), as described pre-

261

Causative drugs with immunosuppressive function

Unrecognized viruses

Clinical symptoms

Immune recovery (CD4+ T cells)

Viral load

Fig. 3. IHS as another manifestation of IRS. During the prodromal period, reactivation of a variety of latent viruses would be clinically unrecognized due to the immunosuppressive state induced by the protracted use of the causative drugs with immunosuppressive properties. Upon withdrawal of the drug, antiviral immune re-

sponses would be rapidly restored, thereby causing immunopathology and reducing viral loads. These events may be clinically recognized as the onset of drug eruption or viral exanthema, if physicians are neither familiar with this syndrome nor pay special attention to the drug patients receive. Modified from Shiohara et al. [6].

viously [5–10]. Detectable plasma or leukocyte CMV load was found in 5 of 12 patients at later stages of DIHS. Of these 5 patients, 4 were treated with oral prednisone, with mean and median doses of 15 mg/day for 6–8 weeks. In contrast, detectable plasma or leukocyte CMV load was never found up to 7 months after onset in 3 patients not receiving oral prednisone during the observation period. However, doses and the duration of prednisone therapy were not necessarily dependent on the detection of CMV load in plasma or leukocytes and the subsequent occurrence of overt CMV disease because 1 patient developed CMV ulcers 2 weeks after initiation of oral prednisone (50 mg/day). Rather, the development of CMV ulcers was temporarily associated with a reduction in oral prednisone from 50 to 40 mg/day within 1 week; detectable plasma and leukocyte CMV load

coincided with a reduction in oral prednisone rather than secondary to long-term oral prednisone. There was a significant difference in age between patients with or without detectable CMV load: aged patients more than 60 years old were at a greater risk of higher CMV burden. Manifestations of IRS include worsening or reactivation of opportunistic infections: for example, Mycobacterium avium complex, Mycobacterium tuberculosis, CMV, crytococcus, hepatitis B or C, or herpes zoster [52], as shown in table 4. As the characterization of IRS grows, it has become necessary to broaden the concept of IRS to include autoimmune diseases that occur following institution of highly active antiretroviral therapy. In view of the common immunosuppressive properties of the causative drugs responsible for DIHS, the clinical and laboratory

262

Shiohara ⴢ Takahashi ⴢ Kano

Table 4. Pathogens capable of inducing IRS in HIV-negative patients after withdrawal of immunosuppressive therapy Mycobacterium avium complex Mycobacterium tuberculosis Cytomegalovirus Crypotococcus Hepatitis B or C Varicella zoster virus

findings of our cases fit in the spectrum of immunopathological damage associated with the reversal of immunosuppressive processes (fig. 3). Indeed, in many if not all patients with DIHS, a dramatic paradoxical deterioration of the clinical symptoms associated with an increase in CD4+ T-cell numbers and no detection of virus genome was typically observed 3–4 days after withdrawal of the causative drug. This scenario explains why one of our patients developed overt CMV ulcers 2 weeks after the beginning of oral prednisone; such short-term use of prednisone is unlikely to cause overt CMV disease, because recent studies have shown no significant effect of short-term (8 weeks) prednisone therapy on HIV RNA load even in the setting of advanced HIV infection [56]. When considering that the immune reconstitution phenomenon generally occurs within the first 2–3 months of prednisone therapy coincident with the tapering or discontinuation of the drug in HIV-negative patients, discontinuation or rapid tapering of prednisone therapy may be a risk factor for the development of IRS which may be subclinical at the onset of DIHS.

Prognosis and General Treatment

From the many case series, the most consistent finding has been that the prognosis is worse in elderly patients while recovery is more rapid and more likely to be complete in children. Although various clinical symptoms develop at various

DIHS and Virus

time points after onset despite withdrawal of the causative drugs and depending on the sites and severity of organ damage, with treatment most patients with DIHS eventually recovered after an undefined period (months) of clinical illness except for the occasional onset of CMV inclusion diseases. In our experience, patients with a previous CMV-related disease had a more severe disease and a delayed recovery than had other patients. Thus, the presence of CMV reactivation during the course of this disease predicts delayed recovery and a worse prognosis. Although EBV reactivation is generally associated with milder forms of the disease at the acute stage, there may be a greater risk of the eventual development of some autoimmune diseases such as type-1 diabetes mellitus [57], Graves disease [6], and systemic sclerosis-like illnesses. Thus, outcome measures in DIHS are important and have been the subject of recent investigations. For many years, oral corticosteroids have been the main therapeutic options for treating DIHS. Rapid resolution of rashes and fever occurs within several days after starting systemic corticosteroids: the usual dosage is 1–1.5 mg prednisone/ kg/day. Regular monitoring including measurement of viral loads in blood and plasma is essential. All patients with severe disease should be carefully monitored for possible reactivation of CMV. Careful consideration should be given to the balance between the need for relieving clinical symptoms by corticosteroids and the need for reducing the amplitude and duration of viral load by antiviral medication. Other complications requiring careful management include limbic encephalitis, thyroid disease, renal failure, and the syndrome of inappropriate secretion of antidiuretic hormone. The questions of whether all DIHS patients should initially receive systemic corticosteroids at what dose and for how long remain to be determined. Although systemic corticosteroids can give promising results in terms of ameliorating vigorous restoration of immune responses to pathogens, the drug dose should be

263

reduced slowly once systemic corticosteroids have been started, even upon rapid resolution of clinical manifestations in DIHS. This is because patients with DIHS are at a greater risk of subsequently developing the wide spectrum of IRS ranging from CMV disease to autoimmune diseases, and the use of systemic corticosteroids represents an important factor that increases the risk of disease progression to the full manifestation of IRS upon withdrawal or reduction of systemic corticosteroids. If symptoms deteriorate despite systemic corticosteroids, other options used successfully in a small series of patients include pulsed intravenous IgG and plasma exchange, or a combination of these [6]. A randomized controlled trial is needed to provide more support for these observations. The outcome of studies on patients with DIHS treated with intravenous IgG suggests that intravenous IgG alters the underlying virus reactivation process in ways not presently understood, and that its administration goes beyond the simple reduction of viral load from the blood and tissues.

Conclusion

Despite big advances in our understanding of the pathogenesis of DIHS and improved therapeutic options, our struggle with drugs and viruses in the setting of DIHS is still in its infancy. We, as clinician-investigators, should aim at maintaining a delicate balance between protective immunity and pathogens, but not ultimately eradicate them. Due to its strong association with the sequential occurrence of herpesvirus reactivation, this disease offers a unique opportunity to develop new immunosuppressive regimens that can maintain the balance between protective immunity and immunopathology.

Acknowledgements This work was supported in part by grants from the Ministry of Education, Sports, Science and Culture of Japan and the Ministry of Health, Labor, and Welfare of Japan to T.S. We would like to thank Werner J. Pichler for critical review of the manuscript and his careful editorial assistance.

References 1 Merritt HH, Putnam TJ: Sodium diphenylhydantoinate in the treatment of convulsive disorders. JAMA 1938;111: 1068–1073. 2 Chaiken BH, Goldberg BI, Segal JP: Dilantin hypersensitivity: report of a case of hepatitis with jaundice, pyrexia, and exfoliative dermatitis. N Engl J Med 1950;242:897–898. 3 Shear NH, Spielberg SP: Anticonvulsant hypersensitivity syndrome. In vitro assessment of risk. J Clin Invest 1988;82:1826–1832. 4 Bocquet H, Bagot M, Roujeau JC: Druginduced pseudolymphoma and drug hypersensitivity syndrome (Drug Rash with Eosinophilia and Systemic Symptoms: DRESS). Semin Cutan Med Surg 1996;15:250–257.

264

5 Suzuki Y, Inagi R, Aono T, Yamanishi K, Shiohara T: Human herpesvirus 6 infection as a risk factor for the development of severe drug-induced hypersensitivity syndrome. Arch Dermatol 1998;134:1108–1112. 6 Shiohara T, Inaoka M, Kano Y: Druginduced hypersensitivity syndrome (DIHS): a reaction induced by a complex interplay among herpesviruses and antiviral and antidrug immune responses. Allergol Int 2006;55:1–8. 7 Tohyama H, Yahata Y, Yasukawa M, Inagi R, Urano Y, Yamanishi K, Hashimoto K: Severe hypersensitivity syndrome due to sulfasalazine associated with reactivation of human herpesvirus 6. Arch Dermatol 1998;134:1113– 1117.

8 Hashimoto K, Yasukawa M, Tohyama M: Human herpesvirus 6 and drug allergy. Curr Opin Allergy Clin Immunol 2003;3:255–260. 9 Kano Y, Inaoka M, Shiohara T: Association between anticonvulsant hypersensitivity syndrome and human herpesvirus 6 reactivation and hypogammaglobulinemia. Arch Dermatol 2004; 140:183–188. 10 Kano Y, Hirahara K, Sakuma K, Shiohara T: Several herpesviruses can reactivate in a severe drug-induced multiorgan reaction in the same sequential order as in graft-versus-host disease. Br J Dermatol 2006;155:301–306. 11 Gennis MA, Vemuri R, Burns EA, Hill JV, Miller MA, Spielberg SP: Familiar occurrence of hypersensitivity to phenytoin. Am J Med 1991;91:631–634.

Shiohara ⴢ Takahashi ⴢ Kano

12 Vittorio CC, Muglia JJ: Anticonvulsant hypersensitivity syndrome. Arch Intern Med 1995;155:2285–2290. 13 Gaffey CM, Chun B, Harvey JC, Manz HJ: Phenytoin-induced systemic granulomatous vasculitis. Arch Pathol Lab Med 1986;110:131–135. 14 Taylor JW, Stein MN, Murphy MJ, Mitros FA: Cholestatic liver dysfunction after long-term phenytoin therapy. Arch Neurol 1984;41:500–501. 15 Pezzimenti JF, Hahn AL: Anicteric hepatitis induced by diphenylhydantoin. Arch Intern Med 1970;125:118–120. 16 Silverman AK, Fairley J, Wong RC: Cutaneous and immunologic reactions to phenytoin. J Am Acad Dermatol 1988; 18:721–741. 17 Chan HL, Stern RS, Arndt KA, Langlois J, Jick SS, Jick H, Walker AM: The incidence of erythema multiforme, Stevens-Johnson syndrome, and toxic epidermal necrolysis. A population-based study with particular reference to reactions caused by drugs among outpatients. Arch Dermatol 1990;126:43–47. 18 Stanley J, Fallon-Pellicci V: Phenytoin hypersensitivity reaction. Arch Dermatol 1978;114:1350–1353. 19 Peyrière H, Dereure O, Breton H, Demoly P, Cociglio M, Blayac J-P, HillaireBuys D; Network of the French Pharmacovigilance Centers: Variability in the clinical pattern of cutaneous sideeffects of drugs with systemic symptoms: does a DRESS syndrome really exist? Br J Dermatol 2006;155:422–428. 20 Horneff G, Lenard HG, Wahn V: Severe adverse reaction to carbamazepine: significance of humoral and cellular reactions to the drug. Neuropediatrics 1992;23:272–275. 21 Gropper AL: Diphenylhydantoin sensitivity: report of fatal case with hepatitis and exfoliative dermatitis. N Engl J Med 1956;254:522–523. 22 Handfield-Jones SE, Jenkins RE, Whittaker SJ, Besse CP, McGibbon DH: The anticonvulsant hypersensitivity syndrome. Br J Dermatol 1993; 129:175– 177. 23 Kleier RS, Breneman DL, Boiko S: Generalized pustulation as a manifestation of the anticonvulsant hypersensitivity syndrome. Arch Dermatol 1991; 127: 1361–1364. 24 Hogg RJ, Sawyer M, Hecox K, Eigenbrodt E: Carbamazepine-induced acute tubulointestinal nephritis. J Pediatr 1981;83:830–832.

DIHS and Virus

25 Imai H, Nakamoto Y, Hirokawa M, Akihama T, Miura A: Carbamazepineinduced granulomatous necrotizing angiitis with acute renal failure. Nephron 1989;51:405–408. 26 Alldredge BK, Knutsen AP, Ferriero D: Antiepileptic drug hypersensitivity syndrome: in vitro and clinical observations. Pediatr Neurol 1994;10:169– 171. 27 Puller H, Wright N, Murdoch JM: Hypersensitivity reactions to antibacterial drugs in infectious mononucleosis. Lancet 1967;2:1176–1178. 28 Patel M: Skin rash with infectious mononucleosis and ampicillin. Pediatrics 1967;40:910–911. 29 Yasukawa K, Matsumura T, Sato-Matsumura KC, Takahashi T, Fujioka Y, Kobayashi H, Shimizu H: Kikuchi’s disease and the skin: case report and review of the literature. Br J Dermatol 2001;144:885–889. 30 Dominguez DC, Torres ML, Antony S: Is human herpesvirus 6 linked to Kikuchi-Fujimoto disease? The importance of consistent molecular and serologic analysis. South Med J 2003;96:226–233. 31 Hudnall SD: Kikuchi-Fujimoto disease. Is Epstein-Barr virus the culprit? Am J Clin Pathol 2000;1113:761–764. 32 Descamps V, Mahe E, Houhou N, Abramowitz L, Rozenberg F, RangerRogez S, Crickx B: Drug-induced hypersensitivity syndrome associated with Epstein-Barr virus infection. Br J Dermatol 2003;148:1032–1034. 33 Aihara M, Sugita Y, Takahashi S, Nagatani T, Arata S, Takeuchi K, Ikezawa Z: Anticonvulsant hypersensitivity syndrome associated with reactivation of cytomegalovirus. Br J Dermatol 2001;144:1231–1234. 34 Wu Y, Sanderson JP, Farrell J, Drummond NS, Hanson A, Bowkett E, Berry N, Stachulski AV, Clarke SE, Pichler WJ, Pirmohamed M, Park BK, Naisbitt DJ: Activation of T cells by carbamazepine and carbamazepine metabolites. J Allergy Clin Immunol 2006;118:233– 241. 35 Naisbitt DJ, Farrell J, Wong G, Depta JP, Dodd CC, Hopkins JE, Gibney CA, Chadwick DW, Pichler WJ, Pirmohamed M, Park BK: Characterization of drug-specific T cells in lamotrigine hypersensitivity. J Allergy Clin Immunol 2003;111:1393–1403.

36 Katsafanas GC, Schirmer EC, Wyatt LS, Frenkel N: In vitro activation of human herpesviruses 6 and 7 from latency. Proc Natl Acad Sci USA 1996;93:9788– 9792. 37 Maeda Y, Teshima T, Yamada M, Shinagawa K, Nakao S, Ohno Y, Kojima K, Hara M, Nagafuji K, Hayashi S, Fukuda S, Sawada H, Matsue K, Takenaka K, Isimura F, Ikeda K, Niiya K, Harada M: Monitoring of human herpesviruses after allogeneic peripheral blood stem cell transplantation and bone marrow transplantation. Br J Haematol 1999; 105:295–302. 38 Maeda Y, Teshima T, Yamada M, Harada M: Reactivation of human herpesviruses after allogeneic peripheral blood stem cell transplantation and bone marrow transplantation. Leuk Lymphoma 2000;39:229–239. 39 Dockrell DH, Paya CV: Human herpesvirus-6 and -7 in transplantation. Rev Med Virol 2001;11:23–36. 40 Appleton AL, Sviland L: Pathogenesis of GVHD: role of herpes viruses. Bone Marrow Transplant 1993;11:349–355. 41 Ferrara JLM, Levy R, Chao NJ: Pathophysiologic mechanisms of acute graftvs.-host disease. Biol Blood Marrow Transplant 1999;5:347–356. 42 Singh N, Carrigan DR: Human herpesvirus-6 in transplantation: an emerging pathogen. Ann Intern Med 1996; 124:1065–1071. 43 Levy JA: Three new human herpesviruses (HHV-6, 7, and 8). Lancet 1997; 349:558–562. 44 Keever CA, Small TN, Flomenberg N, Heller G, Pekle K, Black P, Pecora A, Gillio A, Kernan NA, O’Reilly RJ: Immune reconstitution following bone marrow transplantation: Comparison of recipients of T-cell depleted marrow with recipients of conventional marrow grafts. Blood 1989;73:1340–1350. 45 Burrows SR, Khanna R, Burrows JM, Moss PJ: An alloresponse in humans is dominated by cytotoxic T lymphocytes (CTL) cross-reactive with a single Epstein-Barr virus CTL epitope: implication for graft-versus-host disease. J Exp Med 1994;179:1155–1161. 46 Burrows SR, Silins SL, Moss DJ, Khanna R, Misko IS, Argaet VP: T cell receptor repertoire for a viral epitope in humans is diversified by tolerance to a background major histocompatibility complex antigens. J Exp Med 1995;182: 1703–1715.

265

47 Burrows SR, Silins SL, Khanna R, Burrow JM, Rischmueller M, Mc Cluskey J, Moss DJ: Cross-reactive memory T cells for Epstein-Barr virus augment the alloresponse to common human leukocyte antigens: degenerate recognition of major histocompatibility complexbound peptide by T cells and its role in alloreactivity. Eur J Immunol 1997;27: 1726–1736. 48 Sharrock CE, Kaminski E, Man S: Limiting dilution analysis of human T cells: a useful clinical tool. Immunol Today 1990;11:281–286. 49 Borysiewicz LK, Graham S, Hickling JK, Mason PD, Sissons JG: Human cytomegalovirus-specific cytotoxic T cells: their precursor frequency and stage specificity. Eur J Immunol 1988; 18:269–275.

50 Chung WH, Hung SI, Hong HS, Hsih MS, Yang LC, Ho HC, Wu JY, Chen YT: Medical genetics: a marker for StevensJohnson syndrome. Nature 2004;428: 486. 51 Pichler WJ, Tilch J: The lymphocyte transformation test in the diagnosis of drug hypersensitivity. Allergy 2004;59: 809–820. 52 Hirsch HH, Kaufmann G, Sendi P, Battegy M: Immune reconstitution in HIVinfected patients. Clin Infect Dis 2004; 38:1158–1166. 53 Cooney EL: Clinical indication of immune restoration following highly active antiretroviral therapy. Clin Infect Dis 2002;34:224–233. 54 Rodriguez-Rosado R, Soriano V, Donna C, Gonzales-Lahoz J: Opportunistic infections shortly after beginning highly active antiretroviral therapy. Antivir Ther 1998;3:229–231.

55 Tubiana TSLi, Katalama C, Calvez V, Mohand Hait, Autran B: Long-lasting recovery in CD4 T-cell function and viral-load reduction after highly active antiretroviral therapy in advanced HIV-1 disease. Lancet 1998;351:1682– 1686. 56 McComsey GA, Whalen CC, Mawhorter SD, Asaad R, Valdeg H, Patki AH, Klaumunzner J, Gopalakrishna KV, Calabrese LH, Lederman MM: Placebocontrolled trial of prednisone in advanced HIV infection. AIDS 2001;15: 321–327. 57 Chiou CC, Chung WH, Hung SL, Yang LC, Hong HS: Fulminant type 1 diabetes mellitus caused by drug hypersensitivity syndrome with human herpesvirus 6 infection. J Am Acad Dermatol 2006;54(suppl):S14–S17.

Dr. Tetsuo Shiohara Department of Dermatology, Kyorin University School of Medicine 6-20-2 Shinkawa, Mitaka Tokyo 181-8611 (Japan) Tel. +81 422 47 5511, Fax +81 422 41 4741, E-Mail [email protected]

266

Shiohara ⴢ Takahashi ⴢ Kano

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 267–277

Clinic and Pathogenesis of Severe Bullous Skin Reactions: Stevens-Johnson Syndrome, Toxic Epidermal Necrolysis Laurence Allanore
Jean-Claude Roujeau Department of Dermatology, Hôpital Henri Mondor, Université Paris XII, Créteil, France

Abstract Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are the most severe forms of hypersensitivity reactions affecting the skin. They are best considered severity variants of a single disease on the basis of similar pathology (epidermal necrolysis), risk factors, causes and frequent progression from SJS to TEN in a few days. Most cases occur in patients with normal metabolic pathways, after taking for a mean of 2 weeks a normal dosage of a medication for the first time. A few ‘high-risk’ drugs account for half of the cases of SJS as well as of TEN. Drug-specific cytotoxic T lymphocytes (hypersensitivity type IVc) induce the diffuse apoptosis of epidermal cells, probably in conjunction with amplification mechanisms not yet deciphered. Animal models point to the resemblance of epidermal necrolysis with acute graft-versus-host disease. Since mortality and morbidity of epidermal necrolysis are high, patients should be referred to specialized wards, burn units for the most severe cases. Symptomatic treatment is of tremendous importance in the absence of any intervention proven yet as capable of stopping the disease. Recent observations suggest that corticosteroids, more than highdose immunoglobulins, may deserve a formal study. Copyright © 2007 S. Karger AG, Basel

Definitions

Drug hypersensitivity may result in skin blisters through three different mechanisms:

(1) Rarely, medications usually taken for months can promote the production of autoantibodies that react with adhesion proteins within the skin and induce the same damages as ‘spontaneous’ autoimmune blistering diseases. Antibodies recognizing proteins at the junction between the epidermis and dermis lead to a subepidermal blister (bullous pemphigoid, linear IgA bullous disease). Antibodies directed against desmogleins induce intraepidermal blister by disrupting the adhesion between keratinocytes (pemphigus). Drug-induced cases are a rare proportion (!5%) of autoimmune blistering diseases. (2) Vasculitis and/or thrombosis of skin vessels may lead to skin infarction and hemorrhage. The main skin symptom is purpura, sometimes accompanied by hemorrhagic blisters, skin nodules and livedo. Whether the vasculitis is restricted to the skin or also involves internal organs, such as the kidney, peripheral nerves, gastrointestinal tract or joints, it is much more often caused by infection or by a ‘collagen vascular disease’ than by a drug reaction. Blisters arising on the skin are very rarely the prominent symptom. (3) The third type of reaction is designated by the name of ‘epidermal necrolysis’ a neologism

proposed by Lyell [1] to indicate the ‘necrosis’ and detachment of the epidermis. Blisters are nothing more than exudates that accumulate under the necrotic epidermis. We know at present that the phenomenon of necrolysis results from massive apoptosis of epidermal cells together with degradation of adhesion molecules between basal cells and the basement membrane of the epidermis. Apoptosis of keratinocytes is a frequent event in many common inflammatory dermatoses, e.g. mild sunburn or eczema. What provides specificity to epidermal necrolysis is the acuteness, dissemination of apoptosis on large areas and on the full thickness of the epidermis, together with the mildness of the infiltrate of inflammatory cells. With this definition in mind, ‘epidermal necrolysis’ is the hallmark of a few diseases. Two are related to allogeneic or autoimmune attack of the epidermis: grade 4 acute graft-versus-host disease (GVHD) and some very rare cases of hyperacute cutaneous lupus erythematosus. Three are most often drug-induced and therefore within the scope of this chapter: blistering forms of fixed drug eruptions, Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN).

Drug-Related Epidermal Necrolysis

Fixed drug eruption (FDE) is usually considered an original disease, because of many specific characteristics. Not all episodes comprise blisters but often only round erythematous and edematous well-limited plaques. Blisters, when present, are large and occur as plaques. The characteristic features of FDE lesions are occurrence on a few areas of the skin, their recurrences at the same sites within a few hours in case of rechallenge with the responsible medication and a remaining pigmentation. Mucous membranes are either spared or involved in a single site. Constitutional symptoms are mild or absent. A substantial proportion of the body surface can be involved by the blisters, usually after several prior episodes.

268

Even in these cases of generalized blistering FDE, life is not endangered the way it is in SJS or TEN. On the other hand, the skin pathology is undistinguishable of what is observed in SJS or TEN. Waiting for a better understanding of mechanisms of FDE and TEN, we suggest considering FDE as a separate disease, because of a very original clinical presentation, good prognosis and because drug causes only partially overlap with those of SJS and TEN. SJS and TEN are best considered severity variants of a single disease on the basis of similar pathology (epidermal necrolysis), similar risk factors, causes and frequent progression from initial SJS to TEN a few days later. That concept was validated in a prospective study in which cases had been included and classified blindly for risk factors and causes. Cases classified as SJS, TEN or overlap SJS/TEN only differed by the extent of necrolysis and prognosis, but not by any other criteria. On the other hand, all three groups mentioned above significantly differed from erythema multiforme major by demographic criteria, risk factors and causality [2]. The pathology of erythema multiforme also differs from that of SJS by more infiltrates and much less apoptosis of keratinocytes [3]. Most cases of erythema multiforme are related to recurrent herpes virus infections (HSV1 more often than HSV2 [4]). Current terminology is based on the maximal extent of epidermal detachment. SJS indicates cases with epidermal necrolysis involving less than 10% of the body surface area, TEN with more than 30% and overlap SJS-TEN in-between. These boundaries are rather artificial and a more simple way would be to report all the cases as epidermal necrolysis, with mention of the extent of necrolysis. For reasons that are not clear, the denomination of SJS is more popular in some countries than in others. It is occasionally used as a synonym for TEN, a minor problem when the extent is described, but also to report eruptions that do not comprise any necrolysis either on the skin or

Allanore ⴢ Roujeau

on mucous membranes [5]. The latter usage should be strongly discouraged as a source of confusion on drug causality, prognosis and effect of treatments. Pathomechanisms In early descriptions of (toxic) epidermal necrolysis, the contrast between the extensive destruction of the epidermis, the mild infiltrate of cells, the absence of alteration of blood vessels and the absence of immune deposits was considered as indicating that a very original model of cell death was responsible for the lesions [1]. Since the relationship to drugs was already recognized, drug toxicity was suspected. Fifty years later only a few cases of epidermal necrolysis have been reported after massive overdose of cytotoxic drugs, e.g. colchicine [6] or methotrexate [7]. But the overwhelming majority of cases occur in patients with normal metabolic pathways, taking a normal therapeutic dosage of a medication to which they are exposed for the first time. The reaction begins after a delay of 4–30 days (mean 12–14 days) after the beginning of treatment. In the very few well-documented cases of recurrences the onset was always within 2 days. Such an accelerated onset is not compatible with direct drug toxicity, but is characteristic of ‘memory’, a hallmark of an immune response. Drug-Specific Immune Response All searches for antibodies or complement activation have been unsuccessful. Accumulated data pointed to the presence of T lymphocytes within the epidermis and at the junction between the dermis and epidermis. When skin samples were analyzed using immunolabelling with monoclonal antibodies it appeared that there were more infiltrating T cells in early lesions than suspected from standard staining, with CD8 T cells predominating within the epidermis. Furthermore, the blister fluid contained a large number of CD8 T cells with a phenotype of activated cytotoxic cells. These cells were actu-

Clinic and Pathogenesis of Severe Bullous Skin Reactions

ally shown to kill autologous lymphocytes and keratinocytes after non-specific triggering of their TCR. In most cases, blister fluid cells also killed the same target cells after specific stimulation by a physiologic concentration of the medication that has been clinically suspected [8]. In contrast with earlier hypotheses, the blister fluid cells all had the phenotypic and functional characteristics of classical cytotoxic T lymphocytes (CTL) that are for example the effector cells in graft rejection. From these findings, SJS and TEN fit with the type IVc mechanism of hypersensitivity [9]. Nevertheless, it remains difficult to believe that these drug-specific CTL are solely responsible for the massive apoptosis for at least two reasons: first, similar CTL were found within the epidermis of patients with mild ‘maculopapular’ drug eruptions, in the absence of necrolysis [10], and second, the killing through perforin/granzyme requires cell-to-cell contact and the relative paucity of infiltrating cells suggests that there is a need for amplification of the death message. An important step was the demonstration that the epidermal cells in the lesions of TEN overexpressed Fas ligand (CD95 ligand) and were therefore capable of killing Jurkat cells, i.e. cells from a lymphocyte line highly sensitive to Fas (CD95)-mediated apoptosis. Elevated concentrations of Fas-L were present in the blood and blister fluid of patients with SJS or TEN. Since normal keratinocytes were known to express Fas it was suggested that the co-expression of the death receptor and its ligand led to some ‘collective suicide’ of epidermal cells [11]. Even though direct proof that keratinocytes expressing Fas-L were able to kill other keratinocytes was missing, the hypothesis was generally accepted. It provided a rationale for treating SJS or TEN with high-dose normal human immunoglobulins, which had been shown to have some Fas antagonist activity [11]. Over the years this hypothesis was challenged by several findings. First, while several teams

269

confirmed the strong expression of Fas-L by keratinocytes in lesional and perilesional skin, Fas was not always detectable on epidermal cells [12]. Second, in an animal model described below, anti-Fas monoclonal antibodies provided no protection against epidermal necrolysis. Further investigations are therefore needed to understand which death pathways are involved in epidermal necrolysis and what are the respective roles of a drug-specific immune response and of non-drug-specific amplification mechanisms. The former probably have some tissue specificity, which explains why the apoptosis is restricted to epithelial cells of the skin and mucous membrane, but also occasionally of the gut, bronchi and bile ducts. Animal Models To our knowledge only two animal models exist, and both are related to acute GVHD resulting in epidermal necrolysis. It may therefore be helpful to decipher the terminal mechanism of keratinocyte death. On the other hand, they provide little information on the main question: Why and how can the immune response to a small molecule like a drug lead to an acute GVHD-like reaction? The first model was developed in hamsters 40 years ago [13]. It was a complicated model of local GVHD resulting from intracutaneous injection of reactive parental cells in a first-generation hybrid. The hybrid animal expressing both parental histocompatibility antigens was unable to reject parental cells, while parental cells reacted towards histocompatibility antigens originating from the other parent. A local inflammatory reaction appeared at the injection site within 48– 72 h as expected in such a cutaneous GVHD reaction. But in addition, between 7 and 10 days after cell injection, a widespread erythema developed with epidermal detachment. Animals died in the following days. Since the inbred hamster lines used in these experiments were no longer available, the above experiments were never repro-

270

duced and the immunologic mechanisms never explored in depth. The second model was elaborated recently by a Japanese team [14, 15; for details, see chapter by Azukizawa and Itami, pp 129–139]. Mice were transfected with a transgene construct of ovalbumin gene associated to the promoter of keratin 5. Transgenic mice expressed ovalbumin at the surface of keratinocytes and of the cells of Hassall corpuscules in the thymus. These OVA+ mice were tolerant to ovalbumin. Non-transgenic mice from the same line were immunized against ovalbumin. Ovalbumin-specific CD8+ CTL were isolated and transferred to OVA+ mice. Even though ovalbumin-specific CTL were detected in peripheral lymph nodes and to a small extent in the dermis of recipient animals, there was no detectable alteration to the skin. On the other hand, when the recipient animals were immunosuppressed, either by transfection with OVA in a context of nude mice, or by irradiation of recipients, a clinical and histological phenotype of epidermal necrolysis was present. The absence of such a reaction in the immunocompetent mice was shown to depend on both CD4CD25 regulatory T cells and on CD11 dendritic cells. Administration of Fas-blocking or antiTNF- monoclonal antibodies to susceptible recipients did not confer any protection against epidermal necrolysis. One may object that this model is more relevant to epidermal necrolysis in a context of acute GVHD than to drug-induced epidermal necrolysis, ovalbumin being a classical T-cell antigen that differs in many respects from medications inducing SJS or TEN. The latter are mostly small non-reactive molecules, while exogenous proteins used in humans (hormones, vaccines, monoclonal antibodies) may induce a serum sickness type of hypersensitivity but not epidermal necrolysis. The model is of great interest anyhow to analyze the mechanisms of keratinocyte apoptosis in epidermal necrolysis.

Allanore ⴢ Roujeau

Genetics As discussed elsewhere [see chapter by Hung et al., pp 105–114] an important recent advance was provided by Taiwanese researchers who discovered a very strong association of HLA-B genes with SJS and TEN provoked by carbamazepine and allopurinol, respectively. These results were the more surprising since prior studies of HLA phenotypes in SJS or TEN in France had found only weak associations of borderline significance [16]. Further analyses of HLA-B among SJS and TEN cases enrolled in the European RegiSCAR study [17, unpubl. data] confirmed the Taiwanese findings for allopurinol-related cases, but also suggested that the associations were more complex in European populations. In other words, relevant HLA-B associations appear to depend both on the responsible drug and on the ethnic background of the patient. In addition, some are linked to the phenotype of the reaction and others not: HLA-B*1502 was associated in Taiwan to carbamazepine-induced epidermal necrolysis, but not to carbamazepine-induced DRESS. On the other hand, HLA-B*5801 was associated with both types of reactions to allopurinol. In the present stage of knowledge, we can conclude that several HLA-B genotypes can be necessary, but not sufficient, for permitting the development of SJS or TEN in relation to a specific medication. A Role for Reactive Metabolites? The dogma that most medications can only induce an immune response after transformation in reactive metabolites (‘hapten hypothesis’ or ‘danger hypothesis’) is discussed elsewhere [see chapter by Pirmohamed, pp 84–94]. Concerning SJS and TEN, data are sparse. Two studies of 18 and 15 patients with SJS or TEN related to sulfonamides found that most had a ‘slow acetylation genotype’ [18, 19]. These studies used PCR primers that detected a small subset of the alleles, a method that was later shown to sometimes result in misclassification [20]. Further-

Clinic and Pathogenesis of Severe Bullous Skin Reactions

more, the hypothesis of slow acetylation as a risk factor for reaction to sulfonamides in AIDS patients was commonly proposed but finally discarded by several prospective studies in a large number of patients. The possible role of polymorphisms in drug-metabolizing enzymes in SJS or TEN should therefore be revisited in a larger number of cases. A Role for Virus Infection? The hypothesis that viral infection and medication can interact in the occurrence of drug reactions remains very popular based on several indirect observations, including the very high rate of rashes induced by antibiotics and especially aminopenicillins during infectious mononucleosis, the increased risk of drug reactions in HIV-infected persons, and the more recent emphasis on the high prevalence of herpes virus reactivation in the HSS/DIHS/DRESS type of reactions, and the fact that erythema multiforme major, which has been confounded with SJS for decades, had herpes as a major etiologic factor. While this hypothesis is attractive in various types of drug reactions, it was never supported by clear evidence concerning SJS and TEN. Furthermore, in a case-control study of 379 cases and 1,505 controls [EuroSCAR study, unpubl. data] we found no evidence that viral infection was a risk factor by itself or by increasing the risk of medications.

Clinic

Early Symptoms and Signs It is commonly stated in medical textbooks that SJS and TEN begin with mucous membrane lesions and fever, followed by skin eruption. Actually it can begin as well with the rash, or with any combination of rash, fever, mucosal symptoms or signs. When the first symptoms are throat pain and fever, antibiotics, antipyretics or analgesics are often administered. In such cases the

271

Fig. 1. SJS of limited extent, characterized by widespread small crusts on areas with prior blisters. Mouth and genitalia were involved.

Fig. 2. Typical pattern of TEN in early stage. Small and large flaccid blisters arose on macules. Large sheets of epidermis are sloughing over the lower part of the back.

rapid development of the disease in the next hours is always attributed, by the patient, relatives, physicians … and lawyers, to the medication(s) that were actually prescribed because of the initial symptoms of the reaction. The most characteristic pattern is the rapid progression and succession of new signs and symptoms. Early skin lesions often resemble a viral eruption and especially varicella (chickenpox), but there are too many lesions involving mucous membrane, too much pain and fever for a trivial viral rash. Early oral lesions resemble aphthae, but again with too much fever, pain and skin lesions. Rapid progression of symptoms and signs, pain and anxiety must alert the physician that the reaction may be more severe than suspected on examination.

Skin lesions comprise a mixture of erythematous spots, patches, blisters and erosions (fig. 1). Erythematous spots have a characteristic grey to violet color, which may resemble purpura. Because their roof is made of necrotic epidermis, vesicles and blisters are flaccid and easily breached, especially on pressure areas (fig. 2). That leads to dark red, oozing erosions. Even when covered by a seemingly normal epidermis, erythematous patches may be prone to a positive Nikolsky sign, i.e. mild friction with a finger results in detachment of a sheet of epidermis. It is what we refer to as ‘detachable areas’, on the grounds that, first, dead epidermis no longer provides a barrier function and, second, that the status of ‘detachable’ versus detached depends critically on careful handling of the patient. We therefore advocate that the extent of the skin lesions should be evaluated as the sum of detachable and detached areas and not only as the latter. The lesions progress for several days, 1 week as a mean. Again, transformation of detachable to detached epidermis does not mean an evolution of the process, and the progression should be judged on the sum of detachable and detached areas.

Later Stage Rapidly (3 days as a mean with extremes of a few hours to 11 week) the diagnosis becomes obvious with the association of severe constitutional symptoms with typical mucous membrane and skin lesions. Patients look acutely ill; they have high fever, shivers and pain.

272

Allanore ⴢ Roujeau

Mucous membranes are involved in about 95% of cases, in at least two sites in 90%. To avoid misinterpretation, physicians should not forget first that the external parts of the lips, eyelids, scrotum, penis and labia major are covered by skin and not mucosa; second that if involvement of mouth and eyes is patent, lesions of the pharynx, larynx, genitalia and anus are easily missed and occasionally revealed by late synechiae that could have been prevented. Mouth involvement predominates on the anterior part: inner lips, gums, tongue and inner cheeks. Erosions are covered by fibrin and so painful as to impair eating, drinking, intake of oral medications and also deglutition of saliva. Redness of eyes is a nearly constant symptom. Examination by an ophthalmologist is important for recognizing mild conjunctivitis from more severe lesions: pseudomembranous conjunctivitis and corneal ulceration. Ulceration of the cornea is the most severe event that may happen during the acute stage, and may result in perforation. Attention should also be directed to genital lesions, including a speculum examination of the vagina.

hypovolemia and sepsis carries the risk of shock and multiple organ failure. But in addition to the above non-specific complications, some patients also suffer from the destruction or alteration of the epithelium of other organs such as the bronchi, gut, bile ducts and renal tubules. These lesions may remain asymptomatic; they often result only in mild alterations of biological markers (e.g., enzymuria and tubular proteinuria indicating kidney tubule alterations) but may sometimes express as organ failure in the acute phase or as sequelae. One of the most severe complications is the involvement of the trachea and bronchial tree. It may occur in up to 20% of cases, without strict correlation to the severity of skin and mucous membrane lesions. Early symptoms are cough and more importantly an increased respiratory rate. Since increased ventilation sometimes allows the maintenance of near-normal oxygenation, pulse oximetry values may be just subnormal. Arterial blood gases should be measured on any suspicion of pulmonary dysfunction. Marked hypoxemia, hypocapnia and metabolic alkalosis indicate a high probability that the patient will need artificial ventilation in the next few hours. In such cases the risk of dying is significantly increased.

Course and Complications

Growth of a new epithelium is rapid. It begins on early erosions on the face, sometimes when the disease is still progressing on lower parts of the body. Skin lesions are completely re-epidermized after a mean of 3 weeks. Cicatrization of the mucous membranes usually requires more time, especially in the tongue. Until cicatrization is complete, patients are exposed to many possible complications. As with extensive burns the loss of the barrier function leads to abundant losses of fluids, electrolytes, proteins and calories. It also permits the colonization of the skin by infectious agents and their penetration into the body. The combination of

Clinic and Pathogenesis of Severe Bullous Skin Reactions

Prognosis

It has been known for years that age and extent of detachment were the two factors that had a major impact on the prognosis. A formal statistical analysis led to the addition of five other criteria in a prognosis score specific for SJS and TEN [21]. After having been validated as highly predictive of the issue of hospitalization in another group of patients, this SCORTEN was generally used. It consists in attributing 1 point to each of the following: age 1 40, detachment superior to 10% of the body surface area, recent cancer, heart rate 1120, plasma bicarbonate !20 mmol/l, glycemia

273

114 mmol/l, urea 110 mmol/l. When the total is !2 the risk of dying is about 3%, when it reaches 64 the risk of death is at least 60%.

Sequelae

As long as the upper dermis has not been involved by infection or trauma, the regenerated epidermis is normal, without atrophic or hypertrophic scars. Only the color of the skin is altered with areas of depigmentation and hyperpigmentation. These will progressively fade with years, the earlier the more attention the patient will give to protection against sun exposure. There may be shedding of hairs and nails. The former is transient with normal regrowth, the latter may result in persistent nail abnormalities. Mucosal alterations can induce taste alteration, phimosis, dyspareunia, dysphagia, intestinal stenosis. Bronchopulmonary sequelae are rare. They consist in chronic bronchitis, bronchiectasis, bronchiolitis obliterans, and bronchiolitis obliterans with organizing pneumonia. Cases reported in the medical literature had a poor prognosis and several required lung transplantation. Mild forms may be less rare and overlooked. The most frequent and severe sequelae affect the eyes. About half of the surviving persons suffer from ocular dryness and/or other problems including palpebral slant with distichiasis and trichiasis, subconjunctival fibrous scarring, and symblepharon. Keratitis is frequent and may be associated with unilateral or bilateral limbic insufficiency leading to superficial corneal neovascularization. These eye lesions may worsen for a few months. They not only impair the vision but are often highly painful and strongly affect the quality of life. After such a severe acute disease with frequent disabilities, it is not surprising that many patients have psychological manifestations that could be the expression of post-traumatic stress disease.

274

Diagnostic Procedures

Even for experienced physicians, the clinical diagnosis of SJS or TEN is not always straightforward. There are several possible pitfalls. The Nikolsky sign is observed in pemphigus and other autoimmune diseases. It can also be present in diseases with superficial cleavage of the epidermis at the level of stratum corneum such as impetigo, staphylococcal scalded skin syndrome or AGEP. Even though very rare, cases of SJS or TEN are of tremendous importance for the evaluation of the risk of medications. Therefore, every case should be as properly documented as possible before being reported to pharmacovigilance institutions. The two most important procedures are to obtain clinical photographs and to perform a skin biopsy. Cameras are nowadays integrated in most mobile phones, allowing transmission of digital images to clinical experts within minutes. This may provide an emergency diagnosis and will permit future confirmation by experts. A skin biopsy, including a frozen section for direct immunofluorescence study is also important to definitely exclude alternative diagnoses. In many hospitals the skin biopsy can be processed in emergency, for example on frozen sections, and provide important diagnostic clues in a few hours. Biologic examinations of blood and urine, radiography of the lungs, and arterial blood gases are important for evaluation of prognosis, management and survey of these acutely ill patients but do not provide diagnosis help.

Risk Factors

As developed in another chapter on epidemiology [see chapter by Mockenhaupt, pp 18–31], many risk factors have been identified. These include several chronic disorders. Some conditions appear to increase the risk only because their treat-

Allanore ⴢ Roujeau

ment requires the utilization of ‘high-risk’ medications. The best example is epilepsy. Others most probably have some impact on the regulation of immune responsiveness to drugs, like collagen vascular diseases, in particular systemic lupus erythematosus and radiotherapy for cancer. HIV infection strongly increases the risk of SJS or TEN, probably by combining the two above factors: abnormal immune response and also frequent exposure to ‘high-risk’ medications such as anti-infectious sulfonamides and nevirapine. Two large-scale multinational case-control studies focused mainly on medication risks. Their results are presented and discussed in another chapter [see chapter by Mockenhaupt, pp 18–31]. The main findings can be summarized as follows: The fraction of clearly drug-related cases is the same for SJS and TEN: around 66%. About one half of the cases are related to a small number of ‘high-risk’ drugs, which are the same for SJS and TEN: antibacterial sulfonamides, allopurinol, carbamazepine, phenytoin, phenobarbital, lamotrigine, oxicam NSAIDs and nevirapine. Several other medications listed in table 1 were withdrawn from the market because of epidermal necrolysis [22]. Most cases occur 2 weeks (range 4–30 days) after the first exposure to the suspected medication. With long-term use the risk disappears. As a conclusion, the frequent statement that SJS can be either related to drugs or to infections while medications are the only definite cause of TEN should be considered obsolete. In both situations drugs are the most frequent cause, but do not explain all cases. In all studies a low percentage of patients (about 5%) has not been exposed to any medication in the week preceding the onset of the reaction. In addition, 5–10% had no medication recently added to a long-term treatment with drugs that were not found to be associated with a detectable risk (vitamins, herbals, contraceptive pill…). These repeated observations strongly suggest that 5–15% of cases are due

Clinic and Pathogenesis of Severe Bullous Skin Reactions

to other causes than a drug allergy. This occurs probably more often in children than in adults. A few cases have been attributed to Mycoplasma pneumoniae, Klebsiella pneumoniae or hepatitis viruses infections. These infections were too rarely documented to explain a substantial proportion of idiopathic cases.

Treatment

SJS and TEN necessitate the following interventions: drug withdrawal, symptomatic management, and measures to avoid recurrences. The possible interest of ‘specific treatments’ aimed at halting disease progression is still debated. In a large retrospective series of patients with SJS or TEN in our institution, we observed that the death rate was significantly lower in patients who stopped a causative drug with short elimination half-life as soon as blisters or erosions occurred in comparison with those who discontinued the suspect drug later [23]. Withdrawal of the suspect drug(s) should be done as soon as possible. Management of patients with extensive lesions must be undertaken in specialized intensive care units or in burn units. Medical transport requires particular attention to the skin. The principles of symptomatic therapy are the same as for burns and include: fluid replacement, anti-infectious therapy, aggressive nutritional support, warming of environmental temperature, skin care with appropriate dressings and correction of any organ failure. In patients with hypoxemia, mechanical ventilation is often necessary. At present there is no evidence-based specific treatment for SJS or TEN. Series claiming benefits from treatments are difficult to interpret because of heterogeneous diagnosis criteria, frequent overestimation of skin detachment, frequent utilization of several consecutive treatments with attribution of the ‘success’, i.e. arrest of disease progression, to the last medication

275

Table 1. Drugs withdrawn because of SJS or TEN [26]

Sulfadimethoxine Sulfathiazole Benoxaprofen1 Tenoxicam Chlormezanone Valdecoxib1 1

1966 1970 1982 1987 1996 2005

Withdrawal also based on other safety concerns.

used. Keeping in mind that extension lasts a mean of 7 days and that patients are hospitalized 2–3 days after the onset, it is usually impossible to decide whether the arrest was spontaneous or due to one treatment. Together with the rarity of the disease, the difficult evaluation of the results of any therapy contributes to the persistence of controversies on the treatment of SJS or TEN. A few publications reported that plasmapheresis induced a halt in the progression of TEN. We stopped using plasmapheresis after seeing no alteration in the progression of skin lesions or in the general status of 6 consecutive patients with early TEN [unpubl. observations]. Cyclophosphamide in high intravenous doses and oral cyclosporin have been claimed to be beneficial. Since they were often administered following an ineffective treatment with corticosteroids, their usefulness remains doubtful. With the rationale of inhibiting TNF- production, we tested thalidomide in a double-blind placebo-controlled trial. It had to be interrupted after inclusion of 25 patients because of an unexpectedly high mortality significantly related to thalidomide [24]. Recently, enthusiasm arose about the use of high-dose intravenous human immunoglobulins (IVIG) [11]. Accumulation of series with contradictory results does not allow a definite conclusion, but certainly does not support using IVIG as a standard treatment [25]. High-dose corticosteroids have been the standard therapy in some countries and considered

276

anti-infectious sulfonamide anti-infectious sulfonamide NSAID NSAID myorelaxant NSAID

harmful in others. After strongly advocating against corticosteroid use for years [26], we reconsidered our position after the results from the EuroSCAR study [submitted]. These results looked at mortality in a large cohort of patients in relation to the treatment they received. Compared with supportive treatment alone, the death rate was moderately and not significantly higher in patients who received IVIG, but importantly (though still not significantly) reduced in those who were administered corticosteroids. Awaiting further evaluation of corticosteroids in a randomized control trial, these results, that we consider the best available evidence, do not support the use of any ‘specific’ treatment outside of a trial. Patients should be advised to avoid exposure to the suspect drug(s). Published cases of recurrences were all attributed to the same generic drug or to compounds chemically closely related (e.g. aromatic anticonvulsants). Therefore there is no rationale for restricting the use of all classes of ‘high-risk drugs’. Since a few familial cases have been reported, and because of HLA associations, first-degree relatives should avoid using the same drug(s). Cases should be notified to regulatory agencies and/or to the relevant pharmaceutical company.

Allanore ⴢ Roujeau

References 1 Lyell A: Toxic epidermal necrolysis: an eruption resembling scalding of the skin. Br J Dermatol 1956; 68:355–361. 2 Auquier-Dunant A, Mockenhaupt M, Naldi L, Correia O, Schroder W, Roujeau JC: Correlations between clinical patterns and causes of erythema multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis: results of an international prospective study. Arch Dermatol 2002; 138:1019– 1024. 3 Cote B, Wechsler J, Bastuji-Garin S, Assier H, Revuz J, Roujeau JC: Clinicopathologic correlations in erythema multiforme and Stevens-Johnson syndrome. Arch Dermatol 1995; 131:1268– 1272. 4 Aurelian L, Ono F, Burnett J: Herpes simplex virus-associated erythema multiforme: a viral disease with an autoimmune component. Dermatol Online Journal 2003;9:1. 5 Cheriyan S, Patterson R, Greenberger PA, Grammer LC, Latall J: The outcome of Stevens-Johnson syndrome treated with corticosteroids. Allergy Proc 1995; 16:151–155. 6 Arroyo MP, Sanders S, Yee H, Schwartz D, Kamino H, Strober BE: Toxic epidermal necrolysis-like reaction secondary to colchicine overdose. Br J Dermatol 2004;150:581–588. 7 Orion E, Matz H, Wolf R: The lifethreatening complications of dermatologic therapies. Clin Dermatol 2005; 23: 182–192. 8 Nassif A, Bensussan A, Boumsell L, Deniaud A, Moslehi H, Wolkenstein P, Bagot M, Roujeau JC: Toxic epidermal necrolysis: effector cells are drug-specific cytotoxic T cells. J Allergy Clin Immunol 2004;114:1209–1215. 9 Pichler WJ: Delayed drug hypersensitivity reactions. Ann Intern Med 2003; 139:683–693.

10 Yawalkar N, Hari Y, Frutig K, Egli F, Wendland T, Braathen LR, Pichler WJ: T cells isolated from positive epicutaneous test reactions to amoxicillin and ceftriaxone are drug specific and cytotoxic. J Invest Dermatol 2000; 115:647– 652. 11 Viard I, Wehrli P, Bullani R, Schneider P, Holler N, Salomon D, Hunziker T, Saurat JH, Tschopp J, French LE: Inhibition of toxic epidermal necrolysis by blockade of CD95 with human intravenous immunoglobulin. Science 1998; 282:490–493. 12 Paquet P, Pierard GE: Keratinocyte injury in toxic epidermal necrolysis: simultaneous but distinct topographic expression of CD95R and calprotectin. Int J Mol Med 2002;10:145–147. 13 Billingham RE, Streilein JW: Toxic epidermal necrolysis and homologous disease in hamsters. Arch Dermatol 1968; 98:528–539. 14 Azukizawa H, Kosaka H, Sano S, Heath WR, Takahashi I, Gao XH, Sumikawa Y, Okabe M, Yoshikawa K, Itami S: Induction of T-cell-mediated skin disease specific for antigen transgenically expressed in keratinocytes. Eur J Immunol 2003;33:1879–1888. 15 Azukizawa H, Sano S, Kosaka H, Sumikawa Y, Itami S: Prevention of toxic epidermal necrolysis by regulatory T cells. Eur J Immunol 2005;35:1722– 1730. 16 Roujeau JC, Bracq C, Huyn NT, Chaussalet E, Raffin C, Duedari N: HLA phenotypes and bullous cutaneous reactions to drugs. Tissue Antigens 1986; 28:251–254. 17 Lonjou C, Thomas L, Borot N, Ledger N, de Toma C, LeLouet H, Graf E, Schumacher M, Hovnanian A, Mockenhaupt M, Roujeau JC; RegiSCAR Group: A marker for Stevens-Johnson syndrome...: ethnicity matters. Pharmacogenomics J 2006;6:265–268.

18 Wolkenstein P, Carriere V, Charue D, Bastuji-Garin S, Revuz J, Roujeau JC, Beaune P, Bagot M: A slow acetylation genotype is a risk factor for sulphonamide-induced toxic epidermal necrolysis and Stevens-Johnson syndrome. Pharmacogenetics 1995;5:255–258. 19 Dietrich A, Kawakubo Y, Rzany B, Mockenhaupt M, Simon JC, Schopf E: Low N-acetylating capacity in patients with Stevens-Johnson syndrome and toxic epidermal necrolysis. Exp Dermatol 1995;4:313–316. 20 Hein DW, Doll MA, Fretland AJ, Leff MA, Webb SJ, Xiao GH, Devanaboyina US, Nangju NA, Feng Y: Molecular genetics and epidemiology of the NAT1 and NAT2 acetylation polymorphisms. Cancer Epidemiol Biomarkers Prev 2000;9:29–42. 21 Bastuji-Garin S, Fouchard N, Bertocchi M, Roujeau JC, Revuz J, Wolkenstein P: SCORTEN: a severity-of-illness score for toxic epidermal necrolysis. J Invest Dermatol 2000;115:149–153. 22 Wysowski DK, Swartz L: Adverse drug event surveillance and drug withdrawals in the United States, 1969–2002. Arch Intern Med 2005;165:1363–1369. 23 Garcia-Doval I, LeCleach L, Bocquet H, Otero XL, Roujeau JC: Toxic epidermal necrolysis and Stevens-Johnson syndrome: does early withdrawal of causative drugs decrease the risk of death? Arch Dermatol 2000;136:323–327. 24 Wolkenstein P, Latarjet J, Roujeau JC, Duguet C, Boudeau S, Vaillant L, Maignan M, Schuhmacher MH, Milpied B, Pilorget A, Bocquet H, BrunBuisson C, Revuz J: Randomised comparison of thalidomide versus placebo in toxic epidermal necrolysis. Lancet 1998;352:1586–1589. 25 Faye O, Roujeau JC: Treatment of epidermal necrolysis with high-dose intravenous immunoglobulins (IV Ig): clinical experience to date. Drugs 2005; 65:2085–2090. 26 Roujeau JC, Stern RS: Severe adverse cutaneous reactions to drugs. N Engl J Med 1994;331:1272–1285.

Prof. Jean-Claude Roujeau Service de Dermatologie, Hôpital Henri Mondor, Université Paris XII 51, Avenue Mal Lattre Tassigny FR–94010 Créteil Cedex (France) Tel. +33 1 4981 2510, Fax +33 1 4981 4504 E-Mail [email protected]

Clinic and Pathogenesis of Severe Bullous Skin Reactions

277

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 278–294

Drug Allergic Liver Injury Andreas Cerny a, b
Raffaela Bertoli a a

Pharmacovigilance Center and b Department of Medicine, Ospedale Regionale di Lugano, Lugano, Switzerland

Abstract Drug-induced liver injury (DILI) is clinically often silent but may progress to severe liver injury resulting in the need for liver transplantation or death. In the last decade, DILI has been the single most frequent reason for drug alerts and withdrawal of newly marketed drugs. The mechanisms of these so-called idiosyncratic reactions depend more on the susceptibility of the host and have traditionally been divided into those that have either a metabolic or a immunoallergic basis. This review aims to comprehensively analyze the reported evidence related to the exposure to a series of drugs that are known to cause DILI. A series of lessons emerge: the majority of DILI continue to be recognized only after marketing of the drug, underscoring the limitations of the preclinical and clinical development process. DILI occurring in the context of certain drugs have typical ‘signatures’ that may help to identify them early. The classical hapten-carrier model is a useful model that will have to accommodate newer findings which suggest that the adaptive immune system is also capable of reacting against self-structures that interact with drugs in a non-covalent fashion. Fever, rash and eosinophilia are considered to be evidence of an immunoallergic mechanism, but their absence does not prove its absence. Copyright © 2007 S. Karger AG, Basel

Introduction

In clinical practice the occurrence of drug-induced liver injury (DILI) remains difficult to recognize and laboratory tests, which would help the

clinician to attribute observed liver injury to a particular drug, are lacking outside of specialized research laboratory facilities. The diagnosis of DILI thus relies mostly on the exclusion of other diagnosis and the improvement of the abnormalities after suspension of the suspected drug. Some types of DILI result from dose-dependent toxic mechanisms, others, however, are so-called ‘idiosyncratic’ drug reactions that occur in individual patients who present a unique susceptibility with conventional therapeutic doses. The types of injuries caused by drugs observed at the level of laboratory tests and histopathology can mimic all known types of liver diseases caused by a variety of pathogenic mechanisms in individual patients. Liver-related adverse events have been the main reasons that led to the withdrawal of drugs from the market in the last decade. In spite of important advances in the process of the drug development process, idiosyncratic liver injury occurring at low frequency remains hard to detect due to statistical reasons and also due to the complex nature of its pathogenesis. The following review aims to put more recent observations in perspective in order to provide a clinicallybased approach to immunologically-mediated DILI.

Definition

Different types of definitions and classifications of DILI are used. DILI may be divided into predictable (type A reaction) and unpredictable or idiosyncratic (type B reaction). Type A reactions produce liver injury in a predictable dose-dependent manner, which can be reproduced in experimental animals. Those reactions are typically known to cause liver necrosis that effects one particular area of the liver lobule predominantly. Acetaminophen typically causes centrilobular necrosis. Most type A hepatotoxins produce liver injury through the effects of the reactive metabolites after phase I biotransformation. This metabolite, which includes free radicals or reactive oxygen, can bind covalently to cellular molecules such as proteins, lipids and nucleic acids, disrupting their function. This in turn leads to inactivation of key enzymes and formation of protein drug complexes that are potential targets for immune-mediated liver injury. Type B reactions on the other hand produce liver injury in an unpredictable manner, appear to be less dependent and are not reproducible in animal models and produce a more diffuse form of liver injury. They are due to an individual host susceptibility. They can be divided into two categories: the first is termed immune idiosyncrasy and typically displays a shortened delay on rechallenge, often eosinophilia, rash and fever, and is sometimes accompanied by the appearance of antibodies against drug-modified proteins. An example is halothane-induced liver injury. The other type is metabolic idiosyncrasy and shows no decrease in delay on rechallenge, there is no evidence of an immune response and the pathogenesis involves differences in metabolism of the implicated drug. An example that is often cited is isoniazid, although more recent data suggest that immune idiosyncrasy may be involved at least in part of the patients.

Allergic Hepatitis

The classification has been extended to include other types such as type C (chronic) reactions which are due to chronic dose- and timedependent exposure, type D (delayed reactions) and type E which are due to withdrawal reactions. A novel concept was introduced by Aronson and Ferner [1], the so-called DoTS (dose-time susceptibility) concept which combines the concepts of exposure (dose, timing) with susceptibility. Another commonly used definition of druginduced liver disorders is mainly based on liver test abnormalities [2]. Liver injury is defined by an increase over two times of the upper limit of the normal range of serum alanine aminotransferase (ALT) or conjugated bilirubin or a combined increase of aspartate aminotransferase (AST), alkaline phosphatase (AP) and total bilirubin provided one of them is above two times the upper limit of the normal range. Acute hepatocellular injury is defined by ALT above the upper limit of normal (ULN) or ALT/ AP 65. Acute hepatocellular injury is clinically indistinguishable from acute viral hepatitis. Acute hepatocellular injury may occur due to overdosage as well as to an idiosyncratic drug reaction or due to drug hypersensitivity or autoimmune manifestations. It is the most common form of hepatic damage caused by drugs. Many compounds can produce acute hepatocellular injury including herbal medicines as well as drugs of abuse, such as cocaine and amphetamine. Discontinuation of the treatment usually results in recovery but sometimes subfulminant or fulminant hepatitis may occur. The appearance of jaundice (total bilirubin level of 12 and elevation of ALT 13 times the ULN) is an ominous sign and should promote immediate discontinuation of the treatment. Acute cholestatic liver injury is characterized by any isolated increase of serum AP above two times normal or by an ALT/AP ratio of !2. There are two types that can be distinguished:

279

(1) Acute cholestatic hepatitis associates cholestasis with clinical features such as pain and hepatic tenderness, that frequently resembles acute biliary obstruction and cholangitis. Hypersensitivity manifestations are often present. The usual course is full recovery within 3 months after drug removal. Rarely, chronic cholestasis may follow particularly when cholangitis is present and destruction of numerous small bile ducts has occurred (vanishing bile ducts). Acute cholestatic hepatitis is the second most common type of DILI and involves several hundred drugs. (2) Pure cholestasis, which is characterized by pruritus and jaundice. Transaminases remain generally normal or only slightly increased. Usually they are no systemic hypersensitivity manifestations. Such reactions are mainly seen with sex steroid hormone derivatives, cytarabine and azathioprine. Mixed hepatocellular and cholestatic acute liver injury is characterized by an ALT/ AP ratio between 2 and 5. This is a frequent form of DILI; it occurs frequently in association with immunoallergic-systemic manifestations and corresponds at the histopathological level to a mixture of hepatocellular and cholestatic hepatitis, sometimes associated with the formation of granuloma. Other types of liver injury induced by drugs, such as fibrosis due to activation of stellate cells by methotrexate or excess of vitamin A and its derivatives, steatosis due to mitochondrial toxicity or oncogenesis promoted by oral contraceptives will not be discussed further since their pathogenesis does not primarily include immunoallergic mechanisms.

Epidemiology

Epidemiology of DILI reveals many limitations especially due to underreporting, difficulties in detection (lack of diagnostic tests, most events are clinically silent) or diagnosis (diagnosis is

280

mostly by exclusion of other causes of liver injury, confounding factors such as pre-existing liver disease and multiplicity of drug exposure), differences in case definition and incomplete observation of exposed persons [3]. For many drugs the occurrence ranges from 1 in 10,000 to 1 in 100,000 patients exposed [4]. Partly because of this low incidence, hepatotoxicity of many drugs fails to be detected during the clinical development of a new drug, which is usually limited to observe drug exposure in a few thousand participants. To reliably determine the frequency of such rare events, studies with large population groups are required. It has been estimated that in order to identify DILI occurring in 1 of 10,000 with a confidence interval of 95%, the number of persons required to be studied in a clinical trial would be about 30,000 patients [5]. Since most new drug applications contain data on smaller numbers of exposed patients, DILI occurring at low frequency are more likely to be detected in the post-marketing period. Hence, the period after a drug is put on the market is most important to identify toxicity on the liver on the basis of spontaneous reports; after approval, large numbers of patients are exposed to a drug and rare toxic events may emerge [6]. For currently used agents DILI is a rare event, although there are some exceptions. The reported rate of drug reactions is a crude indicator of risk because of the inherent inaccuracies of case definition and because case recognition depends on the skill and motivation of observers. The study of Björnsson and Olsson [7] determined which drugs were most frequently reported to the WHO Collaborating Centre for International Drug Monitoring with a suspicion of druginduced fatal liver reaction. The top 10 drugs associated with liver reaction with fatal outcome are shown in figure 1; the five most common drugs were acetaminophen, troglitazone, valproate, stavudine and halothane (fig. 1).

Cerny ⴢ Bertoli

Acetaminophen Troglitazone Valproate Stavudine Halothane Lamivudine Didanosine Amiodarone Nevirapine Sulfa/trimethoprim 0

50

100

150

200

250

300

350

Fig. 1. The most common drugs associated with fatal liver injury reported to the WHO Collaborating Centre for International Drug Monitoring (WHO databases 1968–2003) [reprinted from 7, with permission].

Cofactors and Genetic Disposition for DILI

Individual susceptibility to DILI may be affected by many factors. Drug response is a multifactorial and multigenic process that depends on an interaction between multiple gene products and the environment [8]. One can distinguish between acquired and genetic factors. Adults are generally more susceptible than children and women commonly more affected than men. Nutritional status is also an important factor; obesity is associated with the risk of halothane hepatitis, conversely fasting and alcohol use predispose to paracetamol hepatotoxicity [9]. Nutritional factors can also promote isoniazid hepatotoxicity [10]. Pregnancy, a history of drug reactions and concomitant medications can also increase susceptibility. Heavy alcohol intake has also been identified as a risk factor for hepato-

Allergic Hepatitis

toxocity, for example in patients taking antiretrovirals [11, 12]. Pre-existing liver disease and coexisting illnesses may have a greater effect on the ability of a patient to recover from a liver damage than on the likelihood that it will develop [13]. However, some liver diseases can contribute to the hepatotoxicity of some specific drugs. Patients with chronic viral hepatitis and possibly those with HIV infection have a heightened risk of liver injury during antituberculosis [14] or HAART [15] therapy. Diabetes, renal failure and non-alcoholic steatohepatitis predispose to methotrexate-induced hepatic fibrosis. Veno-occlusive disease induced by anticancer drugs seems more common after bone marrow transplantation [16]. Hyperthyroidism promotes halothane hepatitis and HIV increases the risk of hepatotoxicity to cotrimoxazole.

281

Apart from acquired risk factors, possibly the most important is genetic susceptibility for DILI [17]. Genetic factors determine the activity of the antioxidant- and drug-activating pathway, encode pathways of canicular bile secretion, and modulate the immune system and cell death pathway. The predisposition to the development of DILI are polymorphisms and defects in drugmetabolizing enzymes and in the immune system. For example, polymorphism of the N-acetyltranseferase-2 gene differentiates between fast and slow acetylators, whereby the slow one has increased susceptibility to isoniazid toxicity [18]. Another example in the genetic factor in enzymes responsible for the metabolism of some drugs is deficiency in cytochrome 2C19; troglitazone hepatotoxicity appears to be associated with this deficiency in cytochrome 2C19 [19]. In some members of a family with DILI, a deficiency in detoxification of reactive metabolism suggests a genetic defect. Such observations have been made for example with halothane, phenytoin, and sulfonamides; which defect is involved in these mechanisms is unknown. Variation in the immune system could be involved in DILI; there are very strong associations between some specific human leukocyte antigen (HLA) haplotypes and some drugs [20 ; also see chapters of Nolan et al., pp 95–104, and Hung et al., pp 105–114], and possibly also for amoxicillin-clavulanic acid [21]. Understanding the mechanism of DILI is important for its early recognition, the development of improved preclinical screening procedures, finding specific markers and for the design of effective clinical trials to evaluate a potential hepatotoxic effect of a specific drug before its release onto the market. Moreover, improvements in definition, classification and methods for the diagnosis of DILI are an important challenge and prospective studies of the incidence of hepatic adverse drug reactions are required in order to further elucidate the role of the genetic and environmental factors that contribute to individual

282

susceptibility. Major prospective surveys applying strict criteria also for causality assessment for drug-induced liver damage are needed to record the frequency of liver injury and the potential true hepatotoxicity of a drug.

Principles of Drug Metabolism

The particular susceptibility of the liver to injury by drugs is a consequence of its primary role in the metabolism and elimination of foreign substances. Many drugs are excreted in bile or filtered by the kidney and both pathways require that drugs are water-soluble. The majority of drugs absorbed from the gastrointestinal tract are lipophilic and water-insoluble, and therefore have to be converted into water-soluble metabolites suitable for the elimination in urine and bile. They are rendered water-soluble by the hepatic metabolism. The enzyme systems responsible for the biotransformation are in the hepatocyte, in the smooth endoplasmatic reticulum, and consist in monooxygenase, cytochrome c reductase and cytochrome P450. The enzymatic processes are divided into phase I and phase II. Phase I prepares the compound for conjugation adding polar groups to the lipophilic parent compounds. In phase II the resulting metabolite is conjugated with glucuronic acid, sulfate, acetate, glycine, glutathione, or a methyl group, this enhances its solubility in water (fig 2). The reactions in phase I are catalyzed predominantly by cytochrome P450 (CYP). There are numerous isoforms of CYP and they are grouped, according to the similarities of their amino-acid sequences, into families (e.g. CYP2, CYP21) and subfamilies (e.g. CYP3A4, CYP2D). The most important families for hepatic metabolism of drugs and toxins are CYP1, CYP2 and CYP3. These enzymes are composed of an apoprotein and a heme group working in conjunction with NADPH. A given drug will be metabo-

Cerny ⴢ Bertoli

Phase I

Drug

Fig. 2. Schematic representation of drug biotransformation; enzymatic processes phase I and phase II that convert lipophilic drugs in watersoluble metabolites for the elimination in urine and bile.

First metabolite

Water-soluble conjugate

Elimination (urine/bile) Lipophil

lized following a specific pathway of metabolism that involves one or more enzyme systems. Genetic polymorphism affecting the CYP enzymes may determine diminished or enhanced activity of the affected CYP leading to marked differences between patients in their ability to perform phase I metabolism of some drugs. These genetic alterations lead to diminished or enhanced activity of the affected CYP. Genetic variability may explain some of the hypersensitivity reactions to a specific drug. Most drugs are metabolized in the liver with the production of non-toxic metabolites that can be eliminated; however, some drugs become active only after bioactivation or others are metabolized in reactive toxic metabolites. These reactive metabolites may induce adverse reactions directly causing cell damage, or activating the immune system leading to hypersensitivity reactions. Hepatic metabolism is the main route of activation of drugs that have been linked to hypersensitivity such as sulfamethoxazole, phenytoin, carbamazepine, halothane and abacavir [22]. Factors altering the activity of these enzymes can increase the toxicity of a compound (by reducing the conversion to non-toxic metabolites or by increasing its biotransformation to toxic metabolites) or decrease its therapeutic effectiveness (for example by increasing the rate of transformation of the active drug). The liver receives much greater exposure from reactive metabolites than other tissues; however,

Allergic Hepatitis

Phase II

Hydrophil

despite increased exposure, this organ is rarely the main target for antigen-specific T cells. This discrepancy is likely to be due to two main reasons: first, the liver is protected from toxic insult by high glutathione and N-acetylcysteine levels, and second, the liver is an immunologically privileged organ, activation of T cells by Kupffer can lead to tolerance rather than a pathogenic response. A hypersensitivity reaction can then occur when this tolerance is impaired [23].

Immune System of the Liver

The innate immune system provides a rapid first line of defense against infection, but reducing the microbial is usually not sufficient to eliminate the infectious agent, it needs antigens as a second line of defense. The adaptive immune system (T and B lymphocytes) provide a more targeted means of defense and possess ‘memory’, which is the ability to respond more vigorously to subsequent exposure to a given infectious agent. The liver as a first filter of portal blood flow is continuously exposed to pathogens, toxins, xenobiotics and dietary antigens. The liver has a range of local immune mechanisms to cope with these challenges, it contains large numbers of both innate and adaptive immune cells, including a large number of tissue macrophages (Kupffer cells, KC), natural killer (NK) cells, and natural killer T (NKT) cells. The

283

healthy liver contains a combination of intrahepatic lymphocytes, including CD4+ and CD8+ T cells as well CD4– CD8– T cells. Both the innate and adaptive immune cells work to provide the liver with the capacity to remove pathogenic microorganisms, particles and soluble molecules from the circulation, to delete activated T cells, and to induce and maintain tolerance to food antigens. A special role is provided by KC which phagocytose pathogens entering the liver via portal blood. Activated KC produce various cytokines and other mediators, including prostanoids, nitric oxide, and reactive oxygen intermediates which promote and orchestrate liver inflammation. This also occurs through modulation of the phenotype of liver NK and NKT cells. KC are the predominant type of antigenpresenting cells. They can activate T cells in vitro, but they do so less efficiently than antigen-presenting cells in other parts of the body, suggesting that KC play an important role in the delicate balance between the induction of immunity and tolerance within the liver [24]. The adaptive immune system of the liver is unique in that it favors induction of immunological tolerance rather than immunity. This has been demonstrated in a number of ways, the most compelling being the fact that allogeneic liver organ transplants are accepted across major histocompatibility complex (MHC) barriers. In addition, dietary antigens delivered to the liver tolerize the immune system. Several mechanisms have been suggested to explain this phenomenon, including apoptosis of activated T cells (the liver as a graveyard for activated T cells), immune deviation, and active suppression. Immune deviation was shown to occur when adoptively transferred Th1 and Th2 cells are recovered from the liver that display preferential maintenance of Th2 cytokine production. Active suppression was observed in the case of liver sinusoidal endothelial cells which are capable of selectively suppressing IFN--producing Th1 cells while concurrently promoting the outgrowth of IL-4-expressing Th2

284

cells. Taken together, evidence suggests that liver sinusoidal endothelial cells and KC as well as hepatic dendritic cells are important in the induction of tolerance, rather than the activation of Tcell responses [25].

Pathogenesis of Drug Allergic Liver Injury

The anatomy, cellular composition, and microenvironment of the liver favor tolerance rather than pathogenic immunity as described above. Drug-allergic liver injury can thus be postulated to occur as a failure of physiological tolerance mechanism in susceptible individuals. One line of thought requires so-called ‘danger signals’ to be the key trigger of immune reactivity. The concept developed by Matzinger [26] can also be applied to drug-allergic mechanisms. Such a danger signal may be provided by drug-induced stress and/or damage of hepatocytes. This first ‘danger signal’ may pave the way for activation of the innate immune system within the liver. Work in mice done with acetaminophen provides evidence for this, suggesting that after the initial damage to the hepatocyte a cascade of events implicating the innate immune system amplifies the damage leading to hepatic infiltration with inflammatory cells. Those activated cells of the innate immune system produce a range of inflammatory mediators, including cytokines, chemokines, and reactive oxygen and nitrogen species that contribute to liver injury. Work done in animal models suggest that for acetaminophen poisoning, IFN-, Fas, or Fas ligand are directly involved in causing liver damage, while IL-10, IL6, or COX-2 attenuate liver injury. The adaptive immune system comes into play in various, still incompletely understood ways. The classical hapten model is supported by the detection of antibodies that recognize drug-modified hepatic proteins in the sera of DILI patients. In some cases, drugs which by themselves cannot form hapten-host protein complexes have to be

Cerny ⴢ Bertoli

modified by the host to become reactive haptens. Such substances are termed prohaptens. Halothane- or tienilic acid-induced hepatitis is thought to be caused by a hapten mechanism as discussed below. They elicit drug-specific antibodies or autoantibodies that recognize native liver proteins that have also been found in patients with liver injury caused by other drugs, such as dihydralazine and diclofenac. Modification of host proteins can also lead to the sensitization of MHC-restricted drug-reactive T cells that can mediate a number of reactions including DILI. In patients who developed drug-induced systemic reactions of the liver and other organs, drug-specific T cells have been detected, and in some cases, T-cell clones were generated [27]. Those types of reactions can be replicated in vitro incubating patient-derived lymphocytes, implicated drug and suitable antigen-presenting cells in lymphocyte transformation tests. Reactive cells can be isolated and are further induced to more closely determine their spectrum of reactivity. Following such an approach, Pichler and co-workers [28] were able to identify T cells reacting with compounds that were not covalently bound to host proteins. This so-called p-i concept (pharmacological interaction of drugs with antigen-specific immune receptors) suggests that certain drugs can bind to T-cell receptors, mimicking a ligand and its receptor interaction, and cause T-cell activation in an MHC-dependent fashion. Interestingly, both in lamotrigine- as well as carbamazepine-induced systemic hypersensitivity reactions including hepatitis, the majority of reactive T-cells reacted with the parent compound, not requiring metabolism and covalent binding [29]. Despite the detection of drug-specific antibodies and T cells, it has been difficult to directly prove the pathogenic role of the adaptive immune system in DILI, in part because of the lack of animal models. An important reason for the difficulty in developing animal models is the tendency of the liver to respond to antigens with immunological

Allergic Hepatitis

tolerance. This tolerogenic response could also explain the low occurrence of this type of DILI in humans.

Drugs Causing DILI

Halothane The anesthetic halothane is well known to cause DILI which may present in one of two patterns. A mild elevation in serum ALT level was seen in up to 20% of the patients, probably as a result of direct toxicity by the drug metabolites. Generally, there are no symptoms and the changes in liver test resolve rapidly without treatment. A second form is a dose-independent liver injury which is much less common but more severe and characterized by massive hepatocyte necrosis and acute liver failure. The risk ranges between 1 patient per 35,000 to 1 patient per 10,000 following a single exposure and may increase to 1 in 3,700 following multiple exposures [30]. Beyond repeated exposure to the agent, several risk factors have been identified, these may include female gender, advanced age, hypoxia, hyperthyroidism and obesity. Overweight is possibly related to increased storage of the anesthetic or induction of the enzyme CYP2E1 involved in halothane metabolism. There is additionally data in support of a genetic predisposition to halothane-induced liver injury [31]. Halothane leads to hepatocellular damage, biotransformation appears necessary for halothane hepatotoxicity and it seems to be an immunologically-mediated damage. Hepatitis is in fact associated with suggestive immunological features, including rash, arthralgia, eosinophilia, and antibodies against hepatocytes, which have been seen in patients with severe forms of hepatitis. Toxicity may result from metabolism to reactive metabolites which occur both during the reductive and the oxidative pathway. The reductive mechanism produces free radicals, this may be

285

the mechanism responsible for the mild form of hepatitis [32]. In contrast, the oxidative pathway, mediated by CYP2E1 [31], leads to a very reactive acyl chloride that reacts with lysine to form a highly reactive metabolite trifluoroacetylchloride (TFA), which may bind covalently (prohapten concept, haptenizaition) to hepatocyte macromolecules and phospholipids to form TFA-protein adducts [33, 34]. These TFA-protein adducts may be presented on the plasma membrane of hepatocytes, recognized as foreign to the body and act as neoantigens evoking an immune reaction by producing antibodies. In the serum of patients with halothane-induced hepatitis, several different autoantibodies and TFA-protein adducts can be detected [35]. Neoantigens, including CYP2E1, may be expressed on the hepatocyte surface and recognized by these autoantibodies. However, such adducts are also found in non-hypersensitive patients exposed to halothane, which suggests that the major determinant of response may be idiosyncrasies at the level of the immune system [22]. Njoku et al. [36] found that patients with halothane-induced idiosyncratic hepatitis had significantly elevated levels of CYP2E1-specific IgG4 autoantibodies, while healthy persons exposed to anesthetic had elevated levels of CYP2E1-specific IgG1 autoantibodies. In contrast to IgG1 autoantibodies which form detectable circulating immune complexes cleared by the classical pathway activation of complement system, IgG4 autoantibodies form non-precipitating small immune complexes that may escape clearance and induce liver injury. Symptoms of hepatotoxicity generally occur within 2 weeks after exposure. They may occur within 2–3 days in those previously exposed. Fever is common and precedes jaundice in about 75% of patients. Patients have systemic features and complain of nausea, myalgias, anorexia and rash, followed by manifestation of severe hepatic disease. Eosinophilia, suggesting an immunoallergic reaction, occurs in about 40% of reported cases. Jaundice and hepatomegaly are common.

286

The liver biopsy shows different degrees of necrosis, including zone 3 necrosis that can progress also to massive necrosis [37]. Inflammatory response is usually less marked than that seen in viral hepatitis. Liver tests indicate severe hepatocellular necrosis and are similar to those of severe acute viral hepatitis. ALAT typically are increased, reaching levels of 11,000 U/l and may also be much higher. Bilirubin is increased according to severity of disease, and prothrombin time is prolonged; the degree of prolongation is a prognostic value. Enfluorane and rarely also isofluorane can cause DILI similar to halothane and have to be considered if hepatitis occurs postoperatively [38]. The treatment of halothane hepatitis is supportive and immunosuppressive agents do not seem to alter the outcome. Severe progressive cases may require liver transplantation. Tienilic Acid Tienilic acid is a uricosuric diuretic that was withdrawn from the market following recognition of severe hepatitis that developed in 0.1–0.7% of treated patients [39]. An equal number of males and females were affected. CYP2C9 is the major tienilic acid-metabolizing enzyme [40] that transforms tienilic acid into 5-OH tienilic acid and into an intermediate reactive metabolite, maybe sulfoxide. These reactive metabolites are able to bind covalently to the CYP2C9 enzyme which generated them, inactivating it [41–43]. A neoantigen is formed which triggers an immune response, characterized by the presence of antibodies against liver and kidney microsomes (LKM-2 autoantibodies) [44]. The LKM-2 autoantibody recognizes CYP2C9 [45]. The appearance of these modified proteins on the hepatocyte may be the target for antibodies which may then lead to injury of hepatocytes. It is unclear whether autoantibody production is directly responsible for the liver injury or merely a secondary phenomenon that occurs after liver injury mediated by other mechanisms.

Cerny ⴢ Bertoli

There is a latency between the beginning of treatment and manifestation of hepatitis ranging between 2 and 35 weeks [39]. The latency period upon rechallenge is shorter. 85% of patients with hepatitis associated to tienilic acid developed jaundice and around 10% of icteric patients died. In some cases signs of an immune-mediated reaction have been seen such as fever, rash or eosinophilia. The biopsies showed centrolobular to mid-zonal necrosis and also the presence of an inflammatory infiltrate. In very few cases a fulminant hepatitis, chronic hepatitis or cirrhosis followed the initial liver disease. Severity of injury is independent of the amount of diuretic administered and discontinuation of treatment results in recovery in most cases. Sulfamethoxazole Severe hypersensitivity reactions with fever and eosinophilia affecting liver, kidneys, skin and bone marrow are well known to occur after use of sulfonamides including cotrimoxazole. DILI may be the predominant presentation of this drug hypersensitivity syndrome. However, hepatitis following cotrimoxazole is a rare event; its frequency has been estimated at 0.3–1 per 10,000 cases [46]. The often protracted use of cotrimoxazole in HIV patients has been associated with an increased frequency of hypersensitivity reactions, including DILI, reported to occur between 40 and 80% [47–49], compared with a frequency of only 3% in HIV-negative patients [50, 51]. Slow acetylators are at increased risk for hypersensitivity in the HIV-negative population; interestingly, neither acetylator phenotype nor genotype was shown to predispose HIV-positive patients to sulfamethoxazole allergy [52, 53]. Other genetic associations have not been convincingly demonstrated with other enzymes involved in the metabolism of sulfamethoxazole. The reason for the increase in frequency of sulfamethoxazole hypersensitivity in HIV patients remains unclear [8].

Allergic Hepatitis

Sulfamethoxazole is metabolized by CYP2C9 in the liver to a proreactive hydroxylamine metabolite, which is converted spontaneously and also outside the liver to nitrososulfamethoxazole, which is unstable and can react with hydroxylamine to produce azo/azoxy dimers. Thus, we have the situation where metabolism leads to a compound, which then acts as a hapten also outside of the liver, explaining the generalized nature of the hypersensitivity reaction well. Reduction to nitrososulfamethoxazole occurs enzymatically or via interaction with glutathione and ascorbate. The balance between metabolic activation and detoxification may be crucial for the level of exposure to nitrososulfamethoxazole. Interestingly, patients with HIV often display deficiencies in ascorbate and glutathione, which has been argued to lead to a decreased capacity to reduce the nitroso compound accompanied by an increased risk of hypersensitivity reactions [54, 55]. However, also concurrent viral infections can have immunological effects that influence hypersensitivity and are unrelated to the metabolic detoxification pathway [56]. Nitrososulfamethoxazole can interact directly with cellular macromolecules and act as a cytotoxic agent or it can bind covalently to cellular proteins mainly with cysteine, forming a hapten that is recognized as foreign and targeted by the immune system to elicit an immune response. Even though known metabolites of sulfamethoxazole act covalently as haptens, the unmetabolized form of this drug is also able to stimulate T cells via the T-cell receptor in an MHC-dependent way. This involves non-covalent labile binding to TCR or/and MHC molecules (p-i concept) [27, 57 58]. In fact, only a minority of T-cell clones derived from sulfamethoxazole patients react with the reactive metabolite [59]. Figure 3 summarizes sulfamethoxazole metabolism and the mechanisms of hypersensitivity. In about 70% of the cases, cotrimoxazole-induced injury is predominantly cholestatic. Other manifestations of hypersensitivity are commonly

287

Dimerization

Nitro SMX

NH2

NHOH

CYP2C9/MPO/COX

(Auto)Oxidation

SO2

SO2

NH

NH N O CH

N O

CH3 SMX

NO

N

S NH

SO2

SO2

NH N O CH CH3 CH3 3 Nitroso-SMX Semimercaptal Sulfinamide NH

N O

3

Reduction

Fig. 3. Under current understanding, sulfamethoxazole can induce a hypersensitivity reaction by acting as an antigen, either as the parent drug or as a protein conjugate. Additionally, the reactive metabolite can cause cell damage, leading to danger signalling and activation of costimulatory pathways. CYP = Cytochrome P450; MPO = myeloperoxidase; COX = cyclooxygenase; SMX = sulfamethoxazole; TCR = Tcell receptor; MHC = major histocompatibility complex [reprinted from 22, with permission].

Antigen formation

present in patients experiencing cotrimoxazoleinduced hepatotoxicity, such as fever, rash and eosinophilia, jaundice and pruritus. They may precede the onset of DILI. Symptoms usually develop within 10 days of drug exposure. After reexposure, reactions occur faster and may be more severe requiring aggressive treatment, and deaths have occurred. However, the majority of patients recover when the drug is discontinued. Liver biopsy may show cholestasis with or without necrosis, mild focal necrosis, massive necrosis, or chronic hepatitis. Inflammatory infiltrates are lymphoid, sometimes with eosinophils or granulomas. There are also reports of bile duct injury with cotrimoxazole. Isoniazid-INH Isoniazid has been associated with two types of liver injury: mild isoniazid hepatotoxicity and

288

Dendritic cell

Signal 1: TCR-MHC interaction

O

Protein

SO2

SMX hydroxylamine

Protein S Protein

HO

NH

N O

SMX adducts

Danger signalling

Signal 2: Costimulation

T-cell

Hypersensitivity

isoniazid hepatitis. Approximately 10–20% of patients taking isoniazid have some increase in aminotransferase levels and are asymptomatic. The prognosis is excellent and isoniazid can be continued with careful clinical and laboratory monitoring. Also, the risk of hepatitis induced by isoniazid is much lower; analysis of isoniazid-induced hepatitis in the public health setting have found a risk ranging from 0.1 to 4%, depending most likely on the demography of the populations studied and on the case definitions used for hepatotoxicity [60–62]. There is a relationship between age and susceptibility to liver injury [63]. Regular alcohol intake is another risk factor, as is female sex particularly in non-Whites, malnutrition, concurrent liver disease such as viral chronic hepatitis and concurrent use of medications (e.g. phenytoin, rifampin) that induce CYP450. Slow acetylators of isoniazid may also be at increased risk.

Cerny ⴢ Bertoli

The mechanism of liver damage is incompletely understood and still a matter of discussion [64]. Direct idiosyncratic toxicity of the drug or a metabolite is supposed to be responsible for the injury. The latter is supported by findings in rats which suggest that hydrazine, an INH metabolite, and its effect on CYP2E1 induction is involved in isoniazid liver injury [65]. Isoniazid is metabolized in the human liver by acetylation via N-acetyltransferase and hydrolysis; it is the product of this metabolism, acetylhydrazine, that is thought to be the metabolite most consistently implicated in the pathogenesis of hepatitis [66]. Clinically, the symptoms and signs are those of an acute viral hepatitis. 50–70% of patients have anorexia, nausea, vomiting and fatigue. Fever is observed in 10% of patients, while rash and eosinophilia are uncommon. There may be no systemic complaints and jaundice may be the first signal. There is a rise in serum transaminase level and histologic changes resemble those of an acute viral hepatitis, with bridging or focal necrosis, submassive necrosis and chronic hepatitis. The mortality rate of clinically jaundiced patients is 10%. The interval between the beginning of therapy and the appearance of liver injury varies considerably; for the majority of patients symptoms begin within the first 3 months of treatment, but cases have been described to occur as late as after 14 months. Re-exposure may be associated with an accelerated onset, but gradual reintroduction can be achieved in many cases after resolution of hepatitis [10]. Elderly persons and those receiving isoniazid for 12 months appear to have a worse prognosis. No single monitoring plan can predict early serious disease; however, clinical and biochemical monitoring are recommended. Most deaths from isoniazid hepatitis could be prevented if patients reported early symptoms and isoniazid was discontinued. Effective prevention depends on awareness of early symptoms, no matter how unspecific they are. In the past a hypersensitivity reaction was considered unlikely because of the delayed onset,

Allergic Hepatitis

the clinical manifestation and the fact that rash, eosinophilia and fever were infrequently observed [67]; however, histological findings such as areas of focal necrosis with surrounding inflammatory infiltrates including eosinophils are features that can indicate a drug allergy as the cause of hepatic damage. In fact, TH1 responses may not be accompanied by rash and eosinophilia. A positive in vitro lymphocyte transformation test (LTT) with isoniazid could also support the assumption that this drug may cause a hypersensitivity reaction. In one study [68] lymphocyte involvement in the reaction to isoniazid could be found only in a fraction of the patients. Amoxicillin and Clavulanic Acid The semisynthetic penicillins, ampicillin and amoxicillin have only rarely been linked to hepatotoxicity when used alone. However, when amoxicillin is used in combination with clavulanic acid, the risk of cholestasis is approximately 1/100,000 users. Symptoms such as jaundice occur usually within 1–2 weeks after the initiation of the treatment, but can be delayed for up to 6 weeks following suspension of the drug. Histopathology shows cholestasis and hepatitis sometimes also with granulomas suggesting that the mechanism of injury is probably related to immunological reactions. In some cases DILI was accompanied by fever and eosinophilia. While amoxicillin-reactive T cells have been found in the skin of allergic patients, we currently lack similar direct evidence for the presence of clavulanic acid-reactive T cells in the liver [69]. The fact that there is an association between some HLA alleles and DILI in the context of amoxicillin-clavulanic acid use is another argument in favor of an immunoallergic origin [21, 70]. Oral Vitamin K Antagonists Liver reactions due to coumarin administration were described as early as 1963 [71]. Subsequently, single cases or clusters of cases were described with a clinical picture of hepatitis in some cases

289

accompanied by jaundice both to phenprocumon and acenocumarol. In some cases re-exposure resulted in reappearance of hepatitis. In many cases there was a long interval between initiation of the anticoagulant treatment and the occurrence of hepatitis ranging from approximately 2 to 7 months [72, 73]. After discontinuation of the vitamin K, antagonist recovery followed in most of the patients within 1–5 months. Some cases were treated with prednisone [74]. Dorn-Beineke et al. [75] described a case of phenprocumon-induced hepatitis in which they could demonstrate in vitro cross-reactivity also against acenocumarol and warfarin sodium involving drug-reactive CD4positive, CD45-RO-positive, memory CD4 T cells. Such in vitro cross-reactivity may predict in vivo cross-reactivity and should warrant careful evaluation before switching patients to another oral vitamin K antagonist. Lamotrigine Lamotrigine is a more recently introduced anticonvulsant drug that is effective for a broad range of seizures in adults and children known to cause anticonvulsant hypersensitivity syndrome as well as Stevens-Johnson syndrome. Several adult and pediatric cases of DILI including liver failure have been reported. DILI which can in severe cases be associated with multiorgan failure is in some cases associated with fever and eosinophilia suggesting an immunoallergic mechanism [76]. The drug should be suspended in case of the development of a rash. The experience of immunosuppressive treatment for DILI that occurs in the context of anticonvulsant hypersensitivity syndrome is limited to single case reports. Troglitazone Troglitazone, the first PPAR- agonist, was used for diabetes type 2 and approved in the USA in January 1997 and removed from the market in March 2000 after 94 cases of liver failure had been reported. The ‘signature’ of troglitazone-in-

290

duced liver injury was acute hepatocellular damage with high peak serum transaminases and delayed increase of bilirubin. In spite of stopping the drug, transaminases continued to rise for days and often weeks. Resolution was often protracted. The mechanism is still unclear and direct toxic effects which have been observed under in vitro conditions, oxidative stress, and mitochondrial toxicity have been postulated [77]. Careful analysis of the cases rather suggests an idiosyncratic type B reaction either of the metabolic type or in some cases possibly also of the immunoallergic type as suggested by the presence of autoantibodies [78]. Ximelagatran Ximelagatran is a new oral anticoagulant that acts by direct and reversible inhibition of thrombin and has the potential to replace oral vitamin K antagonists. About 8% of persons treated with the drug developed ALT elevations 13 times the ULN within the first 6 months. Interestingly, normalization of transiently elevated transaminases occurred in most patients that continued the treatment. In the clinical trials, 1.9% of patients developed ALT levels 110 times the ULN and 0.5% developed hyperbilirubinemia (12-fold the norm), but only a few cases of liver failure were reported [79]. The injury is mostly hepatocellular with subsequent hyperbilirubinemia and thought to correspond to a type B (idiosyncratic) injury either of the metabolic or immunoallergic type. These events prompted the company to stop further development of the drug.

Diagnosis of DILI

As outlined above, diagnosis of DILI is based on a complex appraisal of a series of data including the history of drug exposure, clinical features, liver function tests, exclusion of alternative causes of liver injury and occasionally liver biopsy. The latter is done rather rarely, but would of course be

Cerny ⴢ Bertoli

of eminent importance to better understand the pathomechanism. Several algorithms have been proposed to standardize this process such as the Roussel Uclaf Causality Assessment method [80] or a more clinically adapted diagnostic scale proposed by Maria and Victorino [81]. The latter takes the following elements into account: temporal relationship between drug intake and the onset of the event, the exclusion of other causes, the presence of extrahepatic manifestations, information about re-exposure and the relevant literature. As mentioned above, in the quest for more drug-specific tests it was found that some drugs seem to induce particular autoantibodies. Patch tests are occasionally positive, in particular (but not exclusively!) with DILI in association with exanthema (drug hypersensitivity syndrome, DiHS or drug rash with eosinophilia and systemic symptoms DRESS; see chapter of Shiohara et al., pp 251–266], but they might even be positive if only hepatitis occurred. The sensitivity and specificity of this skin test in DILI is unknown. The possibility to mirror the in vivo immune reaction in vitro is of course a very attractive one. The LTT and more recent modifications of this technique would seem ideal diagnostic tools. The group in Lisbon [82] has extensively published on the LTT in DILI. They found positive tests in a modified classical lymphoycyte transformation assay format in 56% of a series of 95 patients suspected of having DILI. However, a negative LTT does not rule out a specific T-cell reaction, which might be directed to certain metabolites. The test thus provides high specificity but lacks sensitivity. More extensive search for drug-specific T-cell reactions may help to better characterize immune response.

Conclusions

DILI has become the ‘Achilles’ heel’ of drug developers, regulators and last but not least of pa-

Allergic Hepatitis

tients exposed to new drugs. The weaknesses of preclinical models, the low incidence especially of idiosyncratic liver reactions as well as the lack of sensitive and specific biomarkers explain the fact that most of the recent problems with newly marketed drugs emerged only after exposure of large numbers of patients after marketing. We herein review the role of immunoallergic mechanisms responsible for idiosyncratic reactions and conclude as follows: (1) As more evidence on comprehensively studied drug-related organ damage in patients emerges, more data argue for an important role of the innate as well as the adaptive immune system including DILI. (2) Novel immunological techniques using more sensitive assays will shed more light on the role of the innate and adaptive immune system in DILI. There is a need to standardize assays and to study liver-infiltrating inflammatory cells. (3) The binomial distinction between allergic and metabolic idiosyncrasy is probably too simple and does not take into account that a metabolic injury may provide the trigger for activation of the innate and/or the adaptive immune system. Also, as more data emerge, older concepts such as the hapten-carrier model need to accommodate new mechanisms such as those of the p-i concept. (4) Another need exists for preclinical models: in silico models, toxigenomics, proteomics and metabonomics applied to new and better animal models as well as new human in vitro tests. (5) ALT and the other liver tests are poor biomarkers of immune-mediated liver injury. There is a need to develop biomarkers eventually in combination that are capable of early detection of liver injury due to activation of the innate and adaptive immune system. (6) Treatment of DILI is mostly still limited to suspension of the offending drug and supportive care. New innovative treatments need to be explored within collaborative networks.

291

References 1 Aronson JK, Ferner RE: Joining the DoTS: new approach to classifying adverse drug reactions. BMJ 2003;327: 1222–1225. 2 Benichou C: Criteria of drug-induced liver disorders. Report of an international consensus meeting. J Hepatol 1990;11:272–276. 3 Larrey D: Epidemiology and individual susceptibility to adverse drug reactions affecting the liver. Semin Liver Dis 2002;22:145–155. 4 Larrey D: Drug-induced liver diseases. J Hepatol 2000;32(1 suppl):77–88. 5 Lee WM: Drug-induced hepatotoxicity. N Engl J Med 2003;349:474–485. 6 Navarro VJ, Senior JR: Drug-related hepatotoxicity. N Engl J Med 2006;354: 731–739. 7 Bjornsson E, Olsson R: Suspected druginduced liver fatalities reported to the WHO database. Dig Liver Dis 2006;38: 33–38. 8 Pirmohamed M: Genetic factors in the predisposition to drug-induced hypersensitivity reactions. AAPS J 2006;8: E20–E26. 9 Whitcomb DC, Block GD: Association of acetaminophen hepatotoxicity with fasting and ethanol use. JAMA 1994; 272:1845–1850. 10 Singh J, Garg PK, Tandon RK: Hepatotoxicity due to antituberculosis therapy. Clinical profile and reintroduction of therapy. J Clin Gastroenterol 1996; 22:211–214. 11 Nunez M, Lana R, Mendoza JL, Martin-Carbonero L, Soriano V: Risk factors for severe hepatic injury after introduction of highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2001;27:426–431. 12 Martin-Carbonero L, Nunez M, Gonzalez-Lahoz J, Soriano V: Incidence of liver injury after beginning antiretroviral therapy with efavirenz or nevirapine. HIV Clin Trials 2003;4:115–120. 13 Russo MW, Watkins PB: Are patients with elevated liver tests at increased risk of drug-induced liver injury? Gastroenterology 2004;126:1477–1480. 14 Wong WM, Wu PC, Yuen MF, et al: Antituberculosis drug-related liver dysfunction in chronic hepatitis B infection. Hepatology 2000; 31:201–206.

292

15 Sulkowski MS, Thomas DL, Chaisson RE, Moore RD: Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of hepatitis C or B virus infection. JAMA 2000; 283:74–80. 16 McDonald GB, Sharma P, Matthews DE, Shulman HM, Thomas ED: Venocclusive disease of the liver after bone marrow transplantation: diagnosis, incidence, and predisposing factors. Hepatology 1984;4:116–122. 17 Larrey D, Pageaux GP: Genetic predisposition to drug-induced hepatotoxicity. J Hepatol 1997;26(suppl 2):12–21. 18 Huang YS, Chern HD, Su WJ, et al: Polymorphism of the N-acetyltransferase-2 gene as a susceptibility risk factor for antituberculosis drug-induced hepatitis. Hepatology 2002; 35:883–889. 19 Kumashiro R, Kubota T, Koga Y, et al: Association of troglitazone-induced liver injury with mutation of the cytochrome P450 2C19 gene. Hepatol Res 2003;26:337–342. 20 Chung WH, Hung SI, Hong HS, et al: Medical genetics: a marker for StevensJohnson syndrome. Nature 2004;428: 486. 21 O’Donohue J, Oien KA, Donaldson P, et al: Co-amoxiclav jaundice: clinical and histological features and HLA class II association. Gut 2000;47:717–720. 22 Sanderson JP, Naisbitt DJ, Park BK: Role of bioactivation in drug-induced hypersensitivity reactions. AAPS J 2006;8:E55–E64. 23 Crispe IN: Hepatic T cells and liver tolerance. Nat Rev Immunol 2003;3:51– 62. 24 Li Z, Diehl AM: Innate immunity in the liver. Curr Opin Gastroenterol 2003;19: 565–571. 25 Crispe IN, Dao T, Klugewitz K, Mehal WZ, Metz DP: The liver as a site of Tcell apoptosis: graveyard, or killing field? Immunol Rev 2000;174:47–62. 26 Matzinger P: Tolerance, danger, and the extended family. Annu Rev Immunol 1994;12:991–1045. 27 Schnyder B, Mauri-Hellweg D, Zanni M, Bettens F, Pichler WJ: Direct, MHCdependent presentation of the drug sulfamethoxazole to human  T cell clones. J Clin Invest 1997;100:136–141.

28 Pichler WJ: Pharmacological interaction of drugs with antigen-specific immune receptors: the p-i concept. Curr Opin Allergy Clin Immunol 2002;2: 301–305. 29 Wu Y, Sanderson JP, Farrell J, et al: Activation of T cells by carbamazepine and carbamazepine metabolites. J Allergy Clin Immunol 2006;118:233–241. 30 Neuberger JM: Halothane and hepatitis. Incidence, predisposing factors and exposure guidelines. Drug Saf 1990;5: 28–38. 31 Spracklin DK, Hankins DC, Fisher JM, Thummel KE, Kharasch ED: Cytochrome P450 2E1 is the principal catalyst of human oxidative halothane metabolism in vitro. J Pharmacol Exp Ther 1997;281:400–411. 32 Plummer JL, Beckwith AL, Bastin FN, Adams JF, Cousins MJ, Hall P: Free radical formation in vivo and hepatotoxicity due to anesthesia with halothane. Anesthesiology 1982;57:160– 166. 33 Park BK, Pirmohamed M, Kitteringham NR: Role of drug disposition in drug hypersensitivity: a chemical, molecular, and clinical perspective. Chem Res Toxicol 1998;11:969–988. 34 Kharasch ED, Hankins D, Mautz D, Thummel KE: Identification of the enzyme responsible for oxidative halothane metabolism: implications for prevention of halothane hepatitis. Lancet 1996;347:1367–1371. 35 Kenna JG, Satoh H, Christ DD, Pohl LR: Metabolic basis for a drug hypersensitivity: antibodies in sera from patients with halothane hepatitis recognize liver neoantigens that contain the trifluoroacetyl group derived from halothane. J Pharmacol Exp Ther 1988;245:1103– 1109. 36 Njoku DB, Mellerson JL, Talor MV, et al: Role of CYP2E1 immunoglobulin G4 subclass antibodies and complement in pathogenesis of idiosyncratic druginduced hepatitis. Clin Vaccine Immunol 2006;13:258–265. 37 Benjamin SB, Goodman ZD, Ishak KG, Zimmerman HJ, Irey NS: The morphologic spectrum of halothane-induced hepatic injury: analysis of 77 cases. Hepatology 1985;5:1163–1171.

Cerny ⴢ Bertoli

38 Gelven PL, Cina SJ, Lee JD, Nichols CA: Massive hepatic necrosis and death following repeated isoflurane exposure: case report and review of the literature. Am J Forensic Med Pathol 1996;17:61–64. 39 Zimmerman HJ, Lewis JH, Ishak KG, Maddrey WC: Ticrynafen-associated hepatic injury: analysis of 340 cases. Hepatology 1984;4:315–323. 40 Lopez Garcia MP, Dansette PM, Valadon P, et al: Human-liver cytochromes P-450 expressed in yeast as tools for reactive-metabolite formation studies. Oxidative activation of tienilic acid by cytochromes P-450 2C9 and 2C10. Eur J Biochem 1993;213:223–232. 41 Lecoeur S, Bonierbale E, Challine D, et al: Specificity of in vitro covalent binding of tienilic acid metabolites to human liver microsomes in relationship to the type of hepatotoxicity: comparison with two directly hepatotoxic drugs. Chem Res Toxicol 1994;7:434– 442. 42 Dansette PM, Amar C, Smith C, Pons C, Mansuy D: Oxidative activation of the thiophene ring by hepatic enzymes. Hydroxylation and formation of electrophilic metabolites during metabolism of tienilic acid and its isomer by rat liver microsomes. Biochem Pharmacol 1990;39:911–918. 43 Lopez-Garcia MP, Dansette PM, Mansuy D: Thiophene derivatives as new mechanism-based inhibitors of cytochromes P-450: inactivation of yeastexpressed human liver cytochrome P450 2C9 by tienilic acid. Biochemistry 1994;33:166–175. 44 Homberg JC, Andre C, Abuaf N: A new anti-liver-kidney microsome antibody (anti-LKM2) in tienilic acid-induced hepatitis. Clin Exp Immunol 1984;55: 561–570. 45 Beaune PH, Lecoeur S, Bourdi M, et al: Anti-cytochrome P450 autoantibodies in drug-induced disease. Eur J Haematol Suppl 1996;60:89–92. 46 Vial T, Biour M, Descotes J, Trepo C: Antibiotic-associated hepatitis: update from 1990. Ann Pharmacother 1997;31: 204–220. 47 Gordin FM, Simon GL, Wofsy CB, Mills J: Adverse reactions to trimethoprimsulfamethoxazole in patients with the acquired immunodeficiency syndrome. Ann Intern Med 1984;100:495–499.

Allergic Hepatitis

48 Jaffe HS, Abrams DI, Ammann AJ, Lewis BJ, Golden JA: Complications of cotrimoxazole in treatment of AIDSassociated Pneumocystis carinii pneumonia in homosexual men. Lancet 1983;ii:1109–1111. 49 Van der Ven AJ, Koopmans PP, Vree TB, van der Meer JW: Drug intolerance in HIV disease. J Antimicrob Chemother 1994;34:1–5. 50 Koch-Weser J, Sidel VW, Dexter M, Parish C, Finer DC, Kanarek P: Adverse reactions to sulfosoxazole, sulfamethoxazole, and nitrofurantoin. Manifestations and specific reaction rates during 2,118 courses of therapy. Arch Intern Med 1971;128:399–404. 51 Bigby M, Jick S, Jick H, Arndt K: Druginduced cutaneous reactions. A report from the Boston Collaborative Drug Surveillance Program on 15,438 consecutive inpatients, 1975–1982. JAMA 1986;256:3358–3363. 52 Pirmohamed M, Alfirevic A, Vilar J, et al: Association analysis of drug metabolizing enzyme gene polymorphisms in HIV-positive patients with cotrimoxazole hypersensitivity. Pharmacogenetics 2000;10:705–713. 53 Alfirevic A, Stalford AC, Vilar FJ, Wilkins EG, Park BK, Pirmohamed M: Slow acetylator phenotype and genotype in HIV-positive patients with sulphamethoxazole hypersensitivity. Br J Clin Pharmacol 2003;55:158–165. 54 Trepanier LA, Yoder AR, Bajad S, Beckwith MD, Bellehumeur JL, Graziano FM: Plasma ascorbate deficiency is associated with impaired reduction of sulfamethoxazole-nitroso in HIV infection. J Acquir Immune Defic Syndr 2004;36:1041–1050. 55 Naisbitt DJ, Vilar FJ, Stalford AC, Wilkins EG, Pirmohamed M, Park BK: Plasma cysteine deficiency and decreased reduction of nitrososulfamethoxazole with HIV infection. AIDS Res Hum Retroviruses 2000;16:1929–1938. 56 Eggena MP, Barugahare B, Jones N, et al: Depletion of regulatory T cells in HIV infection is associated with immune activation. J Immunol 2005;174: 4407–4414. 57 Gerber BO, Pichler WJ: Noncovalent interactions of drugs with immune receptors may mediate drug-induced hypersensitivity reactions. AAPS J 2006; 8:E160–E165.

58 Zanni MP, von Greyerz S, Schnyder B, et al: HLA-restricted, processing- and metabolism-independent pathway of drug recognition by human  T lymphocytes. J Clin Invest 1998;102:1591– 1598. 59 Schnyder B, Burkhart C, SchnyderFrutig K, et al: Recognition of sulfamethoxazole and its reactive metabolites by drug-specific CD4+ T cells from allergic individuals. J Immunol 2000;164:6647–6654. 60 LoBue PA, Moser KS: Use of isoniazid for latent tuberculosis infection in a public health clinic. Am J Respir Crit Care Med 2003;168:443–447. 61 Jasmer RM, Saukkonen JJ, Blumberg HM, et al: Short-course rifampin and pyrazinamide compared with isoniazid for latent tuberculosis infection: a multicenter clinical trial. Ann Intern Med 2002;137:640–647. 62 McNeill L, Allen M, Estrada C, Cook P: Pyrazinamide and rifampin vs. isoniazid for the treatment of latent tuberculosis: improved completion rates but more hepatotoxicity. Chest 2003;123: 102–106. 63 Fountain FF, Tolley E, Chrisman CR, Self TH: Isoniazid hepatotoxicity associated with treatment of latent tuberculosis infection: a 7-year evaluation from a public health tuberculosis clinic. Chest 2005;128:116–123. 64 Schwab CE, Tuschl H: In vitro studies on the toxicity of isoniazid in different cell lines. Hum Exp Toxicol 2003;22: 607–615. 65 Yue J, Peng RX, Yang J, Kong R, Liu J: CYP2E1 mediated isoniazid-induced hepatotoxicity in rats. Acta Pharmacol Sin 2004;25:699–704. 66 Peretti E, Karlaganis G, Lauterburg BH: Acetylation of acetylhydrazine, the toxic metabolite of isoniazid, in humans. Inhibition by concomitant administration of isoniazid. J Pharmacol Exp Ther 1987;243:686–689. 67 Stuart RL, Grayson ML: A review of isoniazid-related hepatotoxicity during chemoprophylaxis. Aust NZ J Med 1999;29:362–367. 68 Schreiber J, Zissel G, Greinert U, Schlaak M, Muller-Quernheim J: Lymphocyte transformation test for the evaluation of adverse effects of antituberculous drugs. Eur J Med Res 1999;4: 67–71.

293

69 Yawalkar N, Hari Y, Frutig K, et al: T cells isolated from positive epicutaneous test reactions to amoxicillin and ceftriaxone are drug specific and cytotoxic. J Invest Dermatol 2000; 115:647– 652. 70 Berson A, Freneaux E, Larrey D, et al: Possible role of HLA in hepatotoxicity. An exploratory study in 71 patients with drug-induced idiosyncratic hepatitis. J Hepatol 1994;20:336–342. 71 Renschler HE, Schmidt FW, Manmmen EF: Research on the effects of longterm anticoagulant therapy on the liver (in German). Dtsch Arch Klin Med 1963;208:524–532. 72 Slagboom G, Loeliger EA: Coumarinassociated hepatitis. Report of two cases. Arch Intern Med 1980;140:1028– 1029. 73 Jones DB, Makepeace MC, Smith PM: Jaundice following warfarin therapy. Postgrad Med J 1980;56:671.

74 Ciorciaro C, Hartmann K, Stoller R, Kuhn M: Liver injury caused by coumarin anticoagulants: experience of the IKS (Intercanton Monitoring Station) and the SANZ (Swiss Center for Drug Monitoring) (in German). Schweiz Med Wochenschr 1996;126: 2109–2113. 75 Dorn-Beineke A, Hassan AW, Brautigam J, Hildenbrand R, Dempfle CE, Schwab M: Phenprocoumon-induced hepatitis as an immunologically mediated drug allergic complication of antithrombotic therapy. Thromb Haemost 2003;90:1210–1213. 76 Naisbitt DJ, Farrell J, Wong G, et al: Characterization of drug-specific T cells in lamotrigine hypersensitivity. J Allergy Clin Immunol 2003;111:1393– 1403. 77 Chojkier M: Troglitazone and liver injury: in search of answers. Hepatology 2005;41:237–246.

78 Maniratanachote R, Shibata A, Kaneko S, et al: Detection of autoantibody to aldolase B in sera from patients with troglitazone-induced liver dysfunction. Toxicology 2005;216:15–23. 79 Lee WM, Larrey D, Olsson R, et al: Hepatic findings in long-term clinical trials of ximelagatran. Drug Saf 2005;28: 351–370. 80 Danan G, Benichou C: Causality assessment of adverse reactions to drugs. I. A novel method based on the conclusions of international consensus meetings: application to drug-induced liver injuries. J Clin Epidemiol 1993;46:1323– 1330. 81 Maria VA, Victorino RM: Development and validation of a clinical scale for the diagnosis of drug-induced hepatitis. Hepatology 1997;26:664–669. 82 Maria VA, Victorino RM: Diagnostic value of specific T cell reactivity to drugs in 95 cases of drug induced liver injury. Gut 1997;41:534–540.

Prof. Dr. med. Andreas Cerny Ospedale Regionale Lugano, Sede Civico Via Tesserete 46, CH–6903 Lugano (Switzerland) Tel. +41 91 805 6046, Fax +41 91 805 6045 E-Mail [email protected]

294

Cerny ⴢ Bertoli

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 295–305

Drug-Induced Interstitial Nephritis Monika Keller
Zoi Spanou
Werner J. Pichler Division of Allergology, Clinic of Rheumatology and Clinical Immunology/Allergology, Inselspital, Bern, Switzerland

Abstract Drug-induced interstitial nephritis (DIN) is characterized by a sudden impairment of renal function and is mainly a result of an immune-mediated reaction after intake of a drug. Many different drugs, such as antibiotics, anticonvulsants, diuretics, proton pump inhibitors, non-steroidal anti-inflammatory drugs and many others, are known to cause DIN. The clinical manifestations are characterized by arthralgias, macular or maculopapular exanthema and fever, together with mild proteinuria, sterile pyuria, and eosinophilia. In many cases the only sign is an asymptomatic increase in serum creatinine. Histopathological analysis shows inflammatory infiltrates in the interstitium, the composition of which differs as a function of different forms of T-cell activation and cytokine production. Consequently, the infiltrate shows mostly T cells, and, dependent on the T-cell function, either a monocyte-rich granulomatous reaction, an eosinophilic or neutrophilic inflammation. Often, plasma cells are present, while glomeruli and vessels are spared. The diagnosis of DIN is confirmed with certainty only by biopsy. The lymphocyte transformation test can demonstrate sensitization to a certain drug, but it is often negative – perhaps because the relevant antigen is a metabolite produced in the endothelial cells of the tubuli. The mainstay of treatment is drug discontinuation; the role of steroids is controversial. Copyright © 2007 S. Karger AG, Basel

and represents the most common form of acute interstitial nephritis. Its etiology is related to the drug and an immune-mediated mechanism is strongly suggested because of its clinical manifestations [1–5]. Only recently a drug-specific immune mechanism could be demonstrated in humans with DIN [6]. This reaction is in agreement with other drug hypersensitivity manifestations [7–10] and supports earlier data in mouse models, which relied, however, on spontaneous immune responses to tubular antigens [11] or on immune responses after immunizations with renal tubular antigens [12–14]. DIN is a relatively rare disease with peculiar clinical manifestations showing quite a heterogeneous profile. The clinical symptoms of DIN arise 2–3 weeks or later after intake of a drug, which is normally well tolerated by most persons. In histology of renal biopsies, an inflammatory infiltrate and edema are observed in the affected interstitial lesions [1, 4, 15].

Epidemiology Definition

Drug-induced interstitial nephritis (DIN) is an interstitial kidney disease related to drug intake. It leads to a sudden impairment in renal function

The precise incidence of DIN is difficult to determine. Important drugs able to cause DIN are listed in table 1. The frequency of occurrence surely relates to the use of these drugs.

Table 1. Drugs responsible for acute DIN1 Antimicrobial agents Abacavir Acyclovir Ampicillin Amoxicillin Atazanavir Aztreonam Azithromycin Carbenicillin Cefaclor Cefamandole Cefazolin Cefixitin Cefotetan Cefoperazone Cefotaxime Cephalexin Cephaloridine Cephalothin Cephapirin Cephradine Chloramphenicol Ciprofloxacin Cloxacillin Colistin Cotrimoxazole Erythromycin Ethambutol Fluoroquinolone Flurithromycin Foscarnet Gentamicin Griseofulvin Indinavir Isoniazid Kanamycin Levofloxacin Lincomycin Methicillin Mezlocillin Minocycline Moxifloxacin Nafcillin Nitrofuantoin Norfloxacin

Oxacillin Pencillin G Piperacillin Piromidic acid Polymyxin acid Quinine Rifampicin Spiramycine Sulfonamides Teicoplanin Telithromycin Tetracycline Vancomycin NSAIDs Alclofenac Aminophenazone Azapropazone Aspirin Benoxaprofen Celecoxib Diclofenac Diflunisal Fenclofenac Fenoprofen Flurbiprofen Ibuprofen Indometacin Ketoprofen Mefenamic acid Meloxicam Mesalazine Naproxen Niflumic acid Phenazone Phenylbutazone Piroxicam Pirprofen Rofecoxib Sulfasalazine Sulfinpyrazone Sulindac Suprofen Tolemetin Zomepirac

Anticonvulsants Carbamazepine Diazepam Lamotrigine Oxcarbazepine Phenobarbital Phenytoin Valproate sodium Anticoagulants Fluindione Phenindione Warfarin Diuretics Bumetanide Chlorthalidone Ethacrynic acid Furosemide Hydrochlorothiazide Indapamide Tienilic acid Triametrene Torsemide PPIs Esomprazole Lansoprazole Omeprazole Pantoprazole Rabeprazole Miscellaneous Allopurinol -Methyldopa Amlodipine Antrafenin Azathioprine Bethanidine Bisphos. alendronate Bismuth salts Captopril Carbimazole Chlorpropamide Chlorprothixene

Cimetidine Clofibrate Clometacin Clozapine Cocaine Colistin Cyamethazine Cyclosporine D-Penicillamine Disulfiram Doxepin Famotidine Fenofibrate Flocatfenin Gold salts Glafenin Interferon Interleukin-2 Leflunomide Loratadine Metamizol Nicergoline Ranitidine Rosiglitazone Pamidronate Paracetamol Phenindione Phenothiazine Phenylpropanolamine Piperazine hydrate Pranlukast Probenecid Propranolol Propylthiouracil Ranitidine Simvastatin Streptokinase Sulfinpyrazone Warfarin Zopiclone

NSAIDs = Non-steroidal anti-inflammatory drugs; PPIs = proton pump inhibitors. Based on data from Rossert [1] and Bennett et al. [31], and others.

1

296

Keller ⴢ Spanou ⴢ Pichler

In several studies, the incidence of acute interstitial nephritis was 1–4% of all native renal biopsies and 10–15% of biopsies performed in the setting of acute renal failure [3, 15]. Rather frequent is proton pump inhibitor (PPI)-induced DIN, which may be a pharmacological class effect as described for different PPIs. In recent series of DIN reported from New Zealand, acute disease occurred at 8 per 100,000 patient years (95% confidence level 2.6–18.7 per 100,000 patient years), which made PPIs the most commonly identified cause of acute DIN [16].

5]. In the majority of reported cases, DIN is caused by antibiotics (especially -lactams, macrolides, quinolones and aminoglycosides), nonsteroidal anti-inflammatory drugs (NSAIDs), anticonvulsants and, recently and rather importantly, by PPIs [16, 19]. Table 1 provides a more detailed list of drugs that are known to be responsible for DIN. Some of these drugs are haptens (-lactams) or may become haptens after metabolism, e.g. in the tubuloendothelial cells.

Cofactors Histopathology

The interstitial infiltrates distinguish DIN from other types of acute renal failure. Histopathological analysis of the affected renal specimens reveals inflammatory infiltrates consisting mostly of lymphocytes (CD4+ and CD8+ T cells) and monocytes/macrophages with a variable number of eosinophils, plasma cells and neutrophils. Typical is also interstitial edema as well as disruption of the tubular basal membrane (TBM). In some cases of primary interstitial nephritis, the mononuclear cell infiltrate is accompanied by granulomas. On the other hand, glomeruli and vessels are spared and intact [15, 17, 18]. Based on additional immunohistochemical studies subtle but clear distinctions could be deciphered, suggesting that DIN can be differentiated into granulomatous (IVa), eosinophilic (IVb), and neutrophilic inflammations (IVd) and reactions where cytotoxic T cells are predominant (IVc) [10]. Details are discussed below.

Drugs Causing DIN

A plethora of drugs are considered to be responsible for the induction of the disease but the frequency of pathogenic effects varies widely [1, 2,

Drug-Induced Interstitial Nephritis

Although DIN may appear in any age group, elderly patients (165 years) with reduced renal clearance seem to have an increased risk of disease development [3]. The exact reasons that predispose this patient group to DIN remains unknown. Other possible cofactors could be genetic factors [5], e.g. altered metabolizing potential or modified drug transport by tubuloendothelial cells. HIV is a very important cofactor for druginduced skin diseases – both with regard to the frequency as well as severity. In this context it is interesting to note that protease inhibitors, such as atazanavir [20] and indinavir [21] that are widely used in HIV+ patients, are also reported to be responsible for DIN. Whether this is a feature of the drug or related to the HIVinduced T-cell activation and deficiency is unclear. Important cofactors may also be related to the pharmacological activity, e.g. reduced renal perfusion (for example for NSAIDs) and prerenal azotemia, or the ability of the drugs to form crystals (as described for antiviral drugs such as acyclovir and indinavir [21]), and direct toxic effects on renal tubuloepithelial cells, which may affect their function and consequently the detoxification of drugs.

297

Pathogenesis

Methicillin is the best-studied drug in DIN, but because of its high potential risk of inducing interstitial nephritis, it is no longer used. Border et al. [22] have proposed that a humoral response may be relevant, as methicillin acts as a hapten, binding to the TBM and finally inducing the formation of anti-TBM antibodies. However, this presumed humoral pathogenesis seems to be an exception. DIN is mainly the result of a T-cell-mediated drug hypersensitivity reaction focusing on the kidney. It is important to know that the same drug can induce allergic reactions in various organs in different patients. It may be localized only to the kidney because the antigen is presented there (and then the diagnosis is often delayed and difficult), or it may appear together with generalized symptoms if the antigen (drug or hapten) is distributed throughout the body. Involvement of T Cells T-cell involvement can be inferred from the following: • The association with a generalized allergic reaction (most often an exanthema), which has a well-known T-cell-mediated pathomechanism [10]; • The clinical course – with its delayed appearance, until the responding T cells are expanded; • The demonstration of a T-cell sensitization against the drug – at least in some patients. In our study, the lymphocyte transformation test (LTT) was found positive in 3/12 patients with DIN [6]; • The cloning of drug-specific T cells from patients with DIN [6]; • The identification of drug-specific T cells which express a peculiar TCR-V phenotype in vitro after drug stimulation, and which are also found in vivo – namely in the interstitial infiltrate of the kidney [6]. The following pathway seems to be feasible: drugs might be metabolized locally (for example

298

in epithelial cells of the tubuli), thereby gaining antigenicity and immunogenicity. They form drug-carrier complexes which might be transported to the lymph nodes, but also may be presented locally. The T cells that are able to react with drug-modified peptide complexes are activated and expanded in the lymph nodes, circulate and home to the kidney where they are restimulated by local drug presentation. If a drug is able to form the same antigenic determinants also outside the kidney and probably also during the in vitro culture of the LTT, specific T cells can be detected. However, if the drug, which presumably caused the DIN, requires metabolism in the kidney to gain its hapten-like features and immunogenicity, the antigenic determinant might not be formed during cell culture and a LTT would remain negative. Such a selective metabolism of the relevant antigenic determinant in the kidney may also be an explanation for the localization of the hypersensitivity reaction to the kidney as the antigenic determinant would only be expressed there. As interstitial nephritis is a rather rare event, it may occur only if certain cofactors facilitate an immune response to the drug. This could be a high reabsorption of the drug in Henle’s loop and/or kidney damage by simultaneously applied aminoglycosides, or unknown genetic factors [see chapter of Park et al., pp 55–65]. Classification of Type-IV Hypersensitivity Reactions The inflammatory reaction in the kidney shows some similarity to other drug-hypersensitivity reactions. Drug-specific T cells may orchestrate an inflammatory reaction by secreting various cytokines, which lead to recruitment and activation of various effector cells, like monocytes, eosinophils and neutrophils. In addition, cytotoxic T cells may also be involved [Keller et al., in preparation]. Thus, the recently described subclassification of type-IV reactions, which is well established in cutaneous drug-hypersensitivity re-

Keller ⴢ Spanou ⴢ Pichler

actions [Pichler, pp 168–189], is also applicable to DIN (fig. 1). Granuloma Formation (Type IVa, fig. 1c) This is characterized by the augmented presence of monocytes and macrophages (CD68+ cells) which can frequently be found with drug-hypersensitivity reactions, and need not implicate sarcoidosis or tuberculosis. In contrast to sarcoidosis, no fibrosis can be observed in DIN. In the granulomatous interstitial areas an increased presence of IL-12 (produced by the monocytes) is detectable [6]. Type-IVa reactions in kidney biopsy specimens were confirmed with the detection of activated monocytes/macrophages with immunohistological staining of CD68 and IL-12. Eosinophilic Inflammations (Type IVb, fig. 1d) Augmented IL-5 production, a cytokine secreted from T cells and playing an important role in the maturation and chemoattraction of eosinophils, may be detectable in the interstitial infiltrations indicating the crucial role of the T cells in the pathogenesis of DIN [6]. In such lesions, eosinophils may be the predominant inflammatory cell type. Type-IVb reactions do not necessarily lead to eosinophilia in the blood, but eosinophils might be detected in the urine. T-Cell-Mediated Cytotoxicity (Type IVc, fig. 1e) Cytotoxic CD8+ and CD4+ T cells play a dominant role in most forms of drug-induced exanthemas [23, 24]. However, this mechanism may be less important in the kidney as the presence of typical cytotoxicity markers such as perforin or Fas/FasL is rather low but still detectable in some biopsies [Keller et al., in preparation]. These findings suggest that cytotoxicity directed, e.g., against tubuloepithelial cells or other cellular antigens may be important.

Drug-Induced Interstitial Nephritis

Neutrophilic Inflammation (Type IVd, fig. 1f) In a fourth group of DIN, neutrophilic leukocytes may represent the major inflammatory cell type. Indeed, infiltration of PMNs into the kidney is often observed despite the absence of an infection. This could be due to the presence of T cells secreting high amounts of CXCL8 and GMCSF, which are also responsible for the sterile inflammations in patients with pustular neutrophilic skin diseases like acute generalized exanthematous pustulosis, pustular psoriasis and Behçet’s disease [25, 26]. The combined evaluation of drug-specific T cells in vitro and the phenotypic analysis of the cellular infiltrate in the kidneys has revealed many similarities between DIN and drug-induced cutaneous hypersensitivity reactions. DIN is a group of heterogeneous T-cell-mediated drug hypersensitivity reactions, where drug-specific T cells can be detected and can elicit various forms of local inflammations dependent on the preferential cytokine produced.

Clinical Presentation

DIN becomes clinically evident 2–3 weeks or later after starting the medication. The clinical signs may arise more rapidly if the patient was already sensitized to the drug previously, but they may appear even later, e.g. 1 year after starting therapy, as demonstrated for NSAIDs [5, 18]. The symptoms of DIN are often unspecific and range from asymptomatic decrease of creatinine clearance to generalized hypersensitivity reactions. The clinical manifestations are listed in table 2. As about 11–12% of the patients are asymptomatic, many patients with DIN may not be diagnosed as described for PPI-induced DIN [19, 27]. Interestingly, DIN induced by NSAIDs and COX-2 inhibitors differs from other forms of DIN. In comparison, it typically develops over

299

a

b CD4+ T cells

c

CD8+ T cells

d Monocytes/macrophages

e

Eosinophils

f Fas

Neutrophils

Fig. 1. The inflammatory reaction in the kidney shows some similarity to other drug hypersensitivity reactions, and effector cells of all classes of type-IV reactions can be observed in the kidney biopsy specimens of patients with DIN. CD4+ (a) and CD8+ (b) T cells are present in all biopsies. An augmented presence of monocytes and macrophages indicates a type-IVa reaction (c, CD68+ cells), a lot of eosinophils are observed in type-IVb reactions (d, hematoxylin-eosin staining), cytotoxic reactions belong to type IVc (e, staining for FAS), and neutrophils are the effector cells in type-IVc reactions (f, neutrophil-elastase staining).

300

Keller ⴢ Spanou ⴢ Pichler

Table 2. Laboratory findings and clinical presentation of DIN and frequency of symptoms due to various drugs, NSAIDs and PPIs Symptoms

Laboratory findings Anemia Edema Eosinophilia Eosinophiluria Hematuria Hypertension Leukocyturia Oliguria Proteinuria Pyuria Clinical findings Fever Flank pain Rash Arthralgias Fatigue/malaise Nausea/vomiting Anorexia/weight loss Asymptomatic patients

Occurrence, % various drugs1

NSAIDs1

PPIs2

n.s. 15 39 n.s. 53 20 n.s. 41 58 49

n.s. 74 41 n.s. 38 17 n.s. n.s. n.s. 41

39–89 n.s. 36–39 11–20 56–61 n.s. n.s. n.s. 56–67 84

46 46 42 12 n.s. n.s. n.s. n.s.

n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.

44 n.s. 8–11 n.s. 16–44 28–39 22–50 11–12

NSAIDs = Non-steroidal anti-inflammatory drugs; PPIs = proton pump inhibitors; n.s. = not specified. 1 Based on Johnson and Freehally [32]. 2 Based on Geevasinga et al. [19] and Myers et al. [27].

a longer period of exposure (months or years after the continuous drug intake), the patients are usually older, and extrarenal signs are uncommon. In the biopsy, a lesser degree of inflammation, tubulitis and eosinophilic infiltration is present compared to DIN caused by other drugs. It seems that NSAID-induced DIN is associated with a milder hypersensitivity reaction. A possible explanation for this difference is the anti-inflammatory effect of the NSAIDs [5, 15, 28]. PPIs are presently one of the main causes of DIN [19]. The frequency of symptoms in general and especially in NSAID- and in PPI-induced DIN is summarized in table 2.

Drug-Induced Interstitial Nephritis

Extrarenal Symptoms The extrarenal clinical features may include fever, malaise, flank pain, rash/maculopapular exanthema, and arthralgias. Only methicillin-induced DIN showed a quite monomorphic clinical picture with fever and eosinophilia in 80%, generalized cutaneous rash in 25%, and occasionally arthralgias [1]. DIN that is induced by other drugs often shows incomplete and less suggestive signs (table 2). Each of the typical symptoms such as low-grade fever, maculopapular rash, mild arthralgias, and eosinophilia is present in fewer than 50% of the patients, and all of them together are present in fewer than 5% of patients [1]. Description of the frequency of these symptoms is

301

very diverse in different studies (fever in 70–100%, and eosinophilia in 50–79%) [1, 2, 18]. Bilateral or also unilateral lumbar pain is described as well and is probably due to dilatation of the renal capsule. Eosinophilia, a typical hallmark of many allergic reactions, is present in some of the patients. It is often only transiently present but can be helpful in the presumptive diagnosis of DIN. Renal Symptoms The renal symptoms are characterized by signs of acute renal failure: namely hematuria, oliguria, leukocyturia, hypertension, a nephrotic-range proteinuria, leading to renal insufficiency. Hematuria is mainly microscopic and mild to moderate proteinuria is found in most of the patients. A severe degree of proteinuria usually does not indicate DIN except in some cases of NSAID-induced interstitial nephritis, where severe proteinuria and the development of a nephrotic syndrome was reported [1, 15]. One of the typical clinical characteristics of the disease is the sterile eosinophiluria, which is not always accompanied by eosinophilia in the peripheral blood [15]. Regarding the limited number of diseases with eosinophilia, the detection of eosinophils in the urine could be a useful diagnostic tool.

Diagnosis

Laboratory Changes One of the characteristics of DIN is the heterogeneous profile and the variety of clinical signs and symptoms. In contrast to other diseases there are neither specific nor unique pathological characteristics that could lead to the diagnosis, apart from positive findings on renal biopsy. Urine analysis in patients with DIN typically shows leukocytes, leukocyte casts, and occasionally red blood cells. A reduction in creatinine clearance indicating acute renal failure is present in all cases. This symptom is not typical for DIN but correlates to renal impairment. It should be pointed

302

out that about 70% of the nephrons have to be damaged in order to detect an increase in creatinine. Eosinophilia in peripheral blood can be present. The absence of eosinophilia does not rule out DIN. Actually, the affected interstitial lesions often do contain some eosinophilic infiltrates which may contribute to pyuria. Therefore, sterile pyuria should raise the suspicion of a possible DIN and further analysis of the urine sample for the presence of eosinophils should be performed. The confirmation of eosinophiluria is rather typical for some forms of DIN. Hansel’s stain, which is the most sensitive, and the Wright’s test are both based on an eosin-methylene blue combination and are widely used to identify eosinophils in urine [1]. When more than 1% of white cells are stained, eosinophiluria is considered to be present. Analysis of patients with DIN due to PPIs reported an elevated erythrocyte sedimentation rate and elevated C-reactive protein in all patients with mean values of 85 mm/h and 81 mg/l, respectively [16]. Imaging Studies The performance of renal ultrasonography as a noninvasive diagnostic procedure has clear limitations. There are no specific signs that could indicate an existing DIN, as the echogenicity is usually normal and no dilated pyelocaliceal cavities are observed. There are no other reliable findings that could confirm or exclude DIN, and no features can be used to distinguish DIN from other forms of acute renal failure [2, 15]. Gallium-67 scanning is another test which has been proposed but has also limited reliability. Specificity is controversial as non-renal disorders (e.g. iron overload, severe liver disease) may give false-positive results, too [2, 15]. Renal Biopsy The gold standard for diagnosis and the test which can confirm the presence of DIN with certainty in the acute phase is only a renal biopsy. Because of the polymorphic clinical presentation,

Keller ⴢ Spanou ⴢ Pichler

it is often essential for diagnosis. However, when the diagnosis seems likely, when a probable precipitating drug can be easily withdrawn or when the disease improves readily after withdrawal of a potentially offending drug, supportive management can proceed safely without the need of renal biopsy. If the clinical signs do not improve after drug discontinuation or if the diagnosis remains unclear, renal biopsy is recommended [2, 15]. If DIN is suspected or diagnosed, the drug(s) should be stopped immediately, as a prolonged exposure to the drug antigen may lead to nonreversible renal damage. In vitro and Epicutaneous Skin Tests A further aim of the diagnosis is to identify the relevant drug. In the acute phase of DIN there is no secure method available to identify the drug responsible for the hypersensitivity reaction. Thus, the history, namely the development of nephritis in relation to drug intake, is the basis to decide for or against drug withdrawal. There are two tests available to identify the relevant drug: (a) the lymphocyte transformation test [29], and (b) the epicutaneous skin test. Both tests are usually performed after remission, and can be done years after the event as the circulating drug-specific T cells may persist for years [30]. Both tests have a low sensitivity, thus a positive test is helpful to identify the drug responsible for the hypersensitivity reaction, especially in patients who were exposed to more than one drug. A negative test does not rule out an allergic drug reaction. The presumption of a positive test result is probably higher if systemic symptoms, in particular exanthemas, were present as well. A positive result is very specific for sensitization, but not for the type of hypersensitivity reaction [29]. Various reasons might be responsible for the lack of sensitivity, as discussed elsewhere [Beeler and Pichler, pp 380–390]. One main reason could be that DIN may have been caused by a drug metabolite which is not available for testing, as it was only generated locally.

Drug-Induced Interstitial Nephritis

Therapy

The most important step is the withdrawal of the responsible medication. After drug discontinuation, renal function spontaneously improves in most of the patients and full or partial recovery is observed after a few weeks [1–3, 15]. If multiple, potentially sensitizing drugs were used by the patient, it is reasonable to substitute other medications for as many of these as possible and to withdraw the most likely etiologic agent among medications that cannot be substituted [2]. An indirect parameter showing improvement in renal function is the increase in and normalization of creatinine clearance. It may take weeks until creatinine returns to the baseline value. If TBM antibodies are present, plasmapheresis can also be considered as a therapeutic option [15]. The role of other therapies in the treatment of DIN is controversial. There is no clear evidence for the beneficial role of steroids in the improvement of renal impairment. In a retrospective study it was shown that there is no difference in outcome or prognosis between patients on conservative or steroid therapy (methylprednisolone, 500 mg/day, i.v., for 2–4 days, followed by oral prednisone, 0.75 mg/kg/day for 3–6 weeks), based on the monitoring of creatinine levels [3]. Other investigators proposed methylprednisolone or prednisone therapy (1 mg/kg/day) for 2–3 weeks only [1, 2], because some clinical studies have demonstrated rapid diuresis, clinical improvement and a return to normal renal function within 72 h after starting steroid treatment [2]. Some clinicians recommend administering this short course of predniso(lo)ne only in patients whose renal function failed to improve within 1 week of stopping the culprit drug [1]. This is in contrast with other reports of NSAID-induced DIN, where no improvement could be detected with corticosteroid therapy. Other immunosuppressive drugs could also be beneficial in the treatment of the disease. For example cyclophosphamide or cyclosporine A are

303

effective in experimental animal models with anti-TBM disease [17]. Another experimental protocol suggests the combination of cyclophosphamide and cyclosporine. However, the nephrotoxic effect of cyclosporine A could even lead to further aggravation of renal function. Some recommend therapy with cyclophosphamide (1–2 mg/kg/day) in humans if there is no improvement in serum creatinine after therapy with steroids [15]. One may assume that this drug would be more efficient when given as first-line therapy, but there are no reports available in humans.

Prognosis

Full recovery from the histopathological changes has been observed in most cases, however, only after discontinuing the drug. Reduced renal function may persist and is often observed in PPI-induced DIN [16, 27]. A number of studies have tried to identify the clinical and renal biopsy features that could provide prognostic information in terms of the recovery of renal function. In general, patients who discontinue the offending mediations within 1 or 2 weeks of the onset of DIN (increased serum creatinine) are more likely to recover to nearly baseline renal function. In contrast, prolonged drug intake (3 or more weeks) can cause atrophy of the parenchyma and necrosis of the tubuli leading to irreversible changes in the renal structures and

prolonged reduced renal function. Other prognostic factors for a bad recovery from DIN have not been clearly identified but may include diffuse (versus patchy) inflammation, an excess number of neutrophilic infiltrations (1–6%), and the extent or severity of interstitial fibrosis [1, 2, 15].

Conclusions

DIN is characterized by a heterogeneous clinical presentation, rather limited laboratory findings, but well-defined histopathological changes. A plethora of drugs can elicit DIN. The restricted tissue accessibility, together with the nonspecific and delayed appearance of clinical symptoms, explains the difficulty in diagnosing DIN. The abrupt reduction in creatinine clearance in association with signs of acute renal failure by patients receiving various medications should raise the suspicions of DIN. In particular sterile leukocyturia accompanied by eosinophiluria should raise the suspicion of DIN. In most cases only a kidney biopsy can prove the diagnosis. Recent studies were able to demonstrate the involvement of drug-specific T cells in DIN and proved the existence of a drug hypersensitivity reaction. The most important step in therapy is the withdrawal of the culprit drug. In general, if DIN is detected early and the drug is promptly discontinued, the long-term outcome seems to be good.

References 1 Rossert J: Drug-induced acute interstitial nephritis. Kidney Int 2001;60:804– 817. 2 Kodner CM, Kudrimoti A: Diagnosis and management of acute interstitial nephritis. Am Fam Physician 2003;67: 2527–2534. 3 Clarkson MR, Giblin L, O’Connell FP, O’Kelly P, Walshe JJ, Conlon P, O’Meara Y, Dormon A, Campbell E, Donohoe J:

304

Acute interstitial nephritis: clinical features and response to corticosteroid therapy. Nephrol Dial Transplant 2004; 19:2778–2783. 4 Schetz M, Dasta J, Goldstein S, Golper T: Drug-induced acute kidney injury. Curr Opin Crit Care 2005;11:555–565. 5 Perazella MA: Drug-induced nephropathy: an update. Expert Opin Drug Saf 2005;4:689–706.

6 Spanou Z, Keller M, Britschgi M, Yawalkar N, Fehr T, Neuweiler J, Gugger M, Mohaupt M, Pichler WJ: Involvement of drug-specific T cells in acute drug-induced interstitial nephritis. J Am Soc Nephrol 2006;17:2919– 2927. 7 Pichler WJ, Schnyder B, Zanni MP, Hari Y, von Greyerz S: Role of T cells in drug allergies. Allergy 1998;53:225–232.

Keller ⴢ Spanou ⴢ Pichler

8 Pichler WJ, Yawalkar N, Britschgi M, Depta J, Strasser I, Schmid S, Kuechler P, Naisbitt D: Cellular and molecular pathophysiology of cutaneous drug reactions. Am J Clin Dermatol 2002; 3: 229–238. 9 Pichler WJ: Immune mechanism of drug hypersensitivity. Immunol Allergy Clin North Am 2004;24:373–397, v–vi. 10 Pichler WJ: Delayed drug hypersensitivity reactions. Ann Intern Med 2003; 139:683–693. 11 Kelly CJ, Korngold R, Mann R, Clayman M, Haverty T, Neilson EG: Spontaneous interstitial nephritis in kdkd mice. II. Characterization of a tubular antigen-specific, H-2K-restricted Lyt2+ effector T cell that mediates destructive tubulointerstitial injury. J Immunol 1986;136:526–531. 12 Zakheim B, McCafferty E, Phillips SM, Clayman M, Neilson EG: Murine interstitial nephritis. II. The adoptive transfer of disease with immune T lymphocytes produces a phenotypically complex interstitial lesion. J Immunol 1984;133:234–239. 13 Neilson EG, McCafferty E, Mann R, Michaud L, Clayman M: Murine interstitial nephritis. III. The selection of phenotypic (Lyt and L3T4) and idiotypic (RE-Id) T cell preferences by genes in Igh-1 and H-2K characterizes the cell-mediated potential for disease expression: susceptible mice provide a unique effector T cell repertoire in response to tubular antigen. J Immunol 1985;134:2375–2382. 14 Mann R, Zakheim B, Clayman M, McCafferty E, Michaud L, Neilson EG: Murine interstitial nephritis. IV. Longterm cultured L3T4+ T cell lines transfer delayed expression of disease as I-Arestricted inducers of the effector T cell repertoire. J Immunol 1985;135:286– 293.

15 Michel DM, Kelly CJ: Acute interstitial nephritis. J Am Soc Nephrol 1998;9: 506–515. 16 Simpson IJ, Marshall MR, Pilmore H, Manley P, Williams L, Thein H, Voss D: Proton pump inhibitors and acute interstitial nephritis: report and analysis of 15 cases. Nephrology (Carlton) 2006; 11:381–385. 17 Neilson EG: Pathogenesis and therapy of interstitial nephritis. Kidney Int 1989;35:1257–1270. 18 Markowitz GS, Perazella MA: Druginduced renal failure: a focus on tubulointerstitial disease. Clin Chim Acta 2005;351:31–47. 19 Geevasinga N, Coleman PL, Webster AC, Roger SD: Proton pump inhibitors and acute interstitial nephritis. Clin Gastroenterol Hepatol 2006;4:597–604. 20 Brewster UC, Perazella MA: Acute interstitial nephritis associated with atazanavir, a new protease inhibitor. Am J Kidney Dis 2004;44:e81–e84. 21 Kopp JB, Falloon J, Filie A, Abati A, King C, Hortin GL, Mican JM, Vaughan E, Miller KD: Indinavir-associated interstitial nephritis and urothelial inflammation: clinical and cytologic findings. Clin Infect Dis 2002;34:1122– 1128. 22 Border WA, Lehman DH, Egan JD, Sass HJ, Glode JE, Wilson CB: Antitubular basement-membrane antibodies in methicillin-associated interstitial nephritis. N Engl J Med 1974;291:381– 384. 23 Kuechler PC, Britschgi M, Schmid S, Hari Y, Grabscheid B, Pichler WJ: Cytotoxic mechanisms in different forms of T-cell-mediated drug allergies. Allergy 2004;59:613–622.

24 Schnyder B, Pichler WJ: Skin and laboratory tests in amoxicillin- and penicillin-induced morbilliform skin eruption. Clin Exp Allergy 2000;30:590– 595. 25 Britschgi M, Pichler WJ: Acute generalized exanthematous pustulosis, a clue to neutrophil-mediated inflammatory processes orchestrated by T cells. Curr Opin Allergy Clin Immunol 2002;2: 325–331. 26 Keller M, Spanou Z, Schaerli P, Britschgi M, Yawalkar N, Seitz M, Villiger PM, Pichler WJ: T cell-regulated neutrophilic inflammation in autoinflammatory diseases. J Immunol 2005; 175: 7678–7686. 27 Myers RP, McLaughlin K, Hollomby DJ: Acute interstitial nephritis due to omeprazole. Am J Gastroenterol 2001; 96:3428–3431. 28 Szalat A, Krasilnikov I, Bloch A, Meir K, Rubinger D, Mevorach D: Acute renal failure and interstitial nephritis in a patient treated with rofecoxib: case report and review of the literature. Arthritis Rheum 2004;51:670–673. 29 Pichler WJ, Tilch J: The lymphocyte transformation test in the diagnosis of drug hypersensitivity. Allergy 2004;59: 809–820. 30 Beeler A, Engler O, Gerber BO, Pichler WJ: Long-lasting reactivity and high frequency of drug-specific T cells after severe systemic drug hypersensitivity reactions. J Allergy Clin Immunol 2006;117:455–462. 31 Bennett WM, Elzinga LW, Porter GA: Tubulointestinal disease and toxic nephropathy; in Brenner BM, Rector FC (eds): The Kidney. Philadelphia, Saunders, 1991, pp 1430–1496. 32 Johnson RJ, Feehally J: Comprehensive Clinical Nephrology. St. Louis, Mosby, 2000.

Dr. Monika Keller Division of Allergology Clinic of Rheumatology and Clinical Immunology/Allergology Sahlihaus 1, Inselspital CH–3010 Bern (Switzerland) Tel. +41 31 632 2245, Fax +41 31 632 3547 E-Mail [email protected]

Drug-Induced Interstitial Nephritis

305

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 306–320

Blood Dyscrasias Caused by Hypersensitivity to Drugs Richard H. Aster Blood Research Institute, Blood Center of Wisconsin, and Medical College of Wisconsin, Milwaukee, Wisc., USA

Abstract Idiosyncratic drug sensitivity can affect blood platelets, red cells, neutrophils or general hematopoiesis through immune and non-immune mechanisms. Here, we consider the epidemiology, pathogenesis, clinical presentation, diagnosis and treatment of blood disorders associated with drug sensitivity, placing emphasis on those thought to have an immune etiology. A considerable body of information is available concerning the pathogenesis of drug-induced immune disorders involving platelets and red cells, and diagnostic tools are available that permit the responsible drug to be identified in many cases. Unfortunately, the same is not true of drug-induced neutropenia/agranulocytosis and bone marrow failure (aplastic anemia) and evidence that certain drugs are prone to trigger those conditions has come mainly from large-scale epidemiologic studies. It is likely, however, that future studies will reveal pathologic mechanisms common to each of these disorders. Copyright © 2007 S. Karger AG, Basel

immune and non-immune mechanisms. In keeping with the immunologic emphasis of this book, we will focus here on drug-induced immune disorders, which for the most part are mediated by drug-induced antibodies capable of causing thrombocytopenia, immune hemolytic anemia, or neutropenia. Occasionally, there is general suppression of hematopoiesis, leading to the condition designated ‘aplastic anemia.’ Although these conditions are uncommon, they are important to consider because they can be lethal. In this chapter, these disorders will be discussed in terms of epidemiology, clinical presentation, pathogenesis, diagnosis, and treatment. Because of page limitations, it will be necessary to provide mere basic descriptions and refer readers to relevant literature for greater detail.

Introduction

Drug-Induced Immune Thrombocytopenia

A ‘blood dyscrasia’ is ‘any abnormal condition of the blood.’ Drugs and other xenobiotics can affect the function of blood cells, suppress or abolish their production or cause their destruction by

Description and Classification For unknown reasons, platelets are targeted more often than red cells or neutrophils for drug-induced immune destruction. In keeping with the well-established role of platelets in hemostasis, patients with drug-induced immune thrombocytopenia (DITP) present with acute, often severe

Supported in part by Grant HL-13629 from the National Heart Lung and Blood Institute, National Institutes of Health.

Table 1. Pathogenesis of drug-induced immune thrombocytopenia [modified from 1] Designation

Mechanism

Examples

Hapten-dependent antibody

Drug (hapten) links covalently to membrane protein and induces a drug-specific immune response

Penicillin, cephalosporin antibiotics?

‘Quinine-type’

Drug induces antibody that binds to membrane protein in the presence of soluble drug

Quinine, sulfonamide antibiotics, non-steroidal anti-inflammatory drugs

Autoantibody induction

Drug induces antibody reactive with autologous platelets in the absence of drug

Gold salts, procainamide

Fiban-induced thrombocytopenia

Drug reacts with glycoprotein IIb/IIIa to induce a conformational change (‘neoepitope’) recognized by antibody?

Tirofiban, epitifibatide

Drug-specific antibody

Antibody recognizes the murine component of a chimeric Fab fragment specific for membrane glycoprotein IIIa

Abciximab

Immune complex

Drug binds to platelet factor 4 to produce a complex for which antibody is specific; the resulting immune complex activates platelets via Fc receptors

Heparin

bleeding symptoms. Others have milder disease characterized by punctate hemorrhages on the skin and mucosal surfaces (petechiae) or no symptoms. DITP is mediated by at least six different mechanisms (table 1). Since the resulting disorders differ in respect to presentation, mechanisms, implicated drugs, diagnosis and treatment, each will be discussed separately. Hapten-Dependent Antibodies It is a basic tenet of immunology that small molecules like drugs induce a specific immune response only when they are covalently linked to a larger carrier molecule such as a protein. Accordingly, when drug-induced, antibody-mediated platelet destruction was first described, it was generally assumed that the sensitizing drugs or their active metabolites became immunogenic by being covalently linked to platelet membrane glycoproteins and that this adduct then triggered the formation of drug-specific antibodies. When an individual who had produced such an antibody ingested the drug, it supposedly became co-

Blood Dyscrasias Caused by Hypersensitivity to Drugs

valently linked to one or more membrane proteins, creating a target recognized by antibody, leading to platelet destruction. Subsequent investigations have shown that this mechanism is operative in, at most, only a small fraction of patients with DITP [1]. Possible exceptions are rare individuals who develop acute thrombocytopenia after treatment with penicillin or penicillin derivatives, which are capable of forming covalent linkages to membrane proteins by virtue of possessing a
-lactam structural element. However, experimental evidence to support this mechanism as a cause of thrombocytopenia is not yet available. Accordingly, it is now thought that immune thrombocytopenia caused by classical hapten-dependent antibodies is extremely rare, if it occurs at all. ‘Quinine-Type’ Immune Thrombocytopenia Epidemiology At least 100 different medications are known to be capable of inducing acute, often severe thrombocytopenia in patients sensitized by prior expo-

307

Table 2. Drugs definitely implicated as causes of ‘quinine-type’ immune thrombocytopenia1 Class of drug

Examples

Cinchona alkaloids

Quinine, quinidine

Sulfonamide antibiotics

Sulfamethoxazole, sulfisoxazole

Other antibiotics

Vancomycin, ceftriaxone, rifampicin, levofloxacin, cefotetin, teicoplanin, trimethoprim, methicillin

Non-steroidal antiinflammatory drugs

Ibuprofen, diclofenac, naproxen

Anticonvulsants

Phenytoin, carbamazepine

Chemotherapeutics/ immunosuppressives

Oxaliplatin, irinotecan, cyclophosphamide, cyclosporine

Others

Acetaminophen, amiodarone, diazepam, ranitidine

1

For a more extensive listing, see references 1 and 3.

sure to these drugs for 6–8 days or intermittently over a longer period of time [1–3]. Because quinine, still widely used for the treatment and prevention of nocturnal leg cramps, is perhaps the most common trigger for this disorder, we will refer to it as ‘quinine-type’ thrombocytopenia. However, various antibiotics, non-steroidal antiinflammatory drugs, anticonvulsants, and many other medications can induce a similar clinical picture [1, 2, 4] (table 2). The incidence of ‘quinine-type’ immune thrombocytopenia has not been well defined. A survey carried out in the Eastern United States estimated that sulfamethoxazole/trimethoprim and quinine/quinidine cause acute DITP in 38 and 26 of every million users per week, respectively [5]. Since these drugs are generally thought to be associated with relatively high risks of DITP, the incidence of DITP in patients taking other implicated drugs is probably lower.

308

Pathogenesis Patients with ‘quinine-type’ immune thrombocytopenia have antibodies that bind to specific glycoproteins on normal platelets when the sensitizing drug is present at pharmacologic concentrations [1, 2]. The favored targets are the platelet receptors for fibrinogen (GPIIb/IIIa, IIb/
3 integrin) and von Willebrand factor (the GPIb/V/ IX complex). In contrast to the behavior expected of a classical hapten-dependent antibody, there is no need for covalent linkage of the drug to the target molecule for this interaction to take place. Other differences are that the reaction of antibody with its target is not inhibited by drug at high concentrations and it recognizes human or primate platelets, but not platelets of other species. These facts led early workers to conclude that the drug might react first with antibody to form a sort of ‘immune complex’ that somehow reacted non-specifically with platelet membrane glycoproteins. Since the platelet itself was not the definitive antigen, it was suggested that, according to this mechanism, platelets would be destroyed as ‘innocent bystanders’. However, the postulated immune complexes were never demonstrated experimentally and subsequent studies showed that drug-dependent antibodies (DDAbs) react with their targets via Fab domains, rather than via Fc domains as would be expected of immune complexes [6]. This finding led to the suggestion that the sensitizing drug reacts first with the target glycoprotein molecule to produce a compound epitope, consisting partly of drug and partly of adjacent peptide sequences, to produce the epitope for which antibody is specific. Alternatively, upon binding to the glycoprotein, the drug might induce a conformational change, that creates an epitope elsewhere in the molecule that is recognized by the DDAb (fig. 1). Recent observations suggest a possible unifying concept in which one structural element of the drug binds to antibody and a second binds to the target glycoprotein [7]. If the affinity of drug for antibody is greater than its affinity for the

Aster

Drug Aby

a

b

Fig. 1. Two mechanisms by which drugs might promote binding of drug-dependent antibodies to platelet membranes. a Drug binds to glycoprotein (vertical dark bar), possibly through hydrophobic interaction. This induces a conformational change that creates an epitope elsewhere in the glycoprotein for which antibody (Aby) is specific. b Drug binds to protein, creating a combinatorial epitope (part drug, part protein) for which antibody is specific. From Aster [1].

protein, the drug reacts first with a complementarity-determining region (CDR) on the antibody to create a conformation favorable for binding to an epitope on the protein. If the drug binds more tightly to protein than to antibody, that reaction takes place first, creating an epitope recognized by antibody. In either case, the drug converts a low-affinity antibody-antigen interaction to one with an affinity high enough to permit significant numbers of antibody molecules to react with the target at concentrations of antibody that might reasonably occur in the course of a vigorous immune response, e.g., 10 –8 M (fig. 2). The experimental and theoretical frameworks supporting this hypothesis are summarized in reference 7. It is apparent that under this model it is unimportant whether drug reacts first with antibody to produce an ‘immune complex’ or first with the target glycoprotein to create a ‘neoepitope’ recognized by antibody. Virtually nothing is known about how drugs stimulate the formation of quinine-type DDAbs.

Drug-dependent antibody CDR H

Drug-dependent antibody CDR H

+

–

H

+

H

– –

+

–

+

+

+

–

–

Platelet antigen

Platelet antigen

Low-affinity fit

High-affinity fit

Fig. 2. A model for drug-dependent antibody binding to an epitope on a platelet glycoprotein. Left: Antibodies capable of causing drug-dependent thrombocytopenia react weakly with an epitope on a target glycoprotein. The K A for this interaction is too low to allow significant numbers of antibody molecules to bind in the absence of drug. Right: In the example shown, the drug contains structural elements complementary to a negatively

Blood Dyscrasias Caused by Hypersensitivity to Drugs

charged site on the glycoprotein (below) and a hydrophobic site (H) on the antibody complementarity-determining region (CDR) (above). The drug interacts with these sites to improve the ‘fit’ between the two proteins, increasing the K A to a value that permits binding to occur at levels of antibody, antigen and drug achieved in the circulation following ingestion of drug. From Bougie et al. [7].

309

However, T-cell lines have been isolated from patients with other forms of drug sensitivity (e.g., those with dermatologic reactions) that carry Tcell receptors (TCR) responsive to drug in the context of an appropriate major histocompatibility complex (MHC) molecule carried on an antigen-presenting cell [8]. It has been shown that drug-dependent T-cell stimulation in this experimental system takes place without any requirement for drug to be linked covalently to either the TCR or the MHC molecule [8]. The analogy between drug-dependent TCR-MHC interaction and the model depicted in figure 2 for drug-dependent binding of a DDAb to its target is apparent, and it could have implications for drug-dependent activation of B cells leading to production of ‘quinine-type’ antibodies that cause platelet destruction in DITP [7]. Presentation Once sensitization to a particular drug has occurred, repeat exposure often leads to systemic symptoms such as headache, dizziness, nausea, and chills followed by fever. Bleeding symptoms, characterized in severe cases by hemorrhage into the skin and buccal mucosa and bleeding from the gastrointestinal or urinary tracts, usually appear 6–24 h later. Rarely, intracranial or intrapulmonary hemorrhage ensues, leading to a fatal outcome. If the association between thrombocytopenia and drug exposure is recognized and the provocative drug is discontinued, platelet levels gradually return to normal, usually within 4– 5 days. Rarely, thrombocytopenia persists for 1– 2 weeks, even when the sensitizing drug is one that is rapidly excreted. Diagnosis The most important element in diagnosis is a high index of suspicion and a careful medical history. Specific inquiries should be made about quinine (usually taken for nocturnal muscle cramps), sulfonamide antibiotics, anticonvulsants, and nonsteroidal anti-inflammatory drugs. Remarkably,

310

a surprising number of patients fail to recall drug exposures even when direct inquiries are made, an oversight that can lead to repeated, unnecessary hospitalizations for ‘acute idiopathic thrombocytopenia’ [9]. As noted, DDAbs bind to their targets in the presence of soluble drug. Accordingly, various assays have been developed to measure drug-dependent binding of antibody to intact platelets or isolated platelet glycoproteins. Flow cytometry is useful for detection of most antibodies [2]. Typical positive and negative control reactions obtained with serum from a patient sensitive to a metabolite of the non-steroidal anti-inflammatory drug naproxen are shown in figure 3. With some, but not all, DDAbs, it is essential that soluble drug be present at all stages of the test procedure to achieve maximum sensitivity for antibody detection. Although DDAbs can be detected in the majority of patients with thrombocytopenia induced by quinine, quinidine, and certain antibiotics, when other drugs are implicated it is not uncommon to obtain negative reactions, even when the patient has a clinical picture strongly suggestive of drug-induced thrombocytopenia. One reason for this is that certain drugs are highly insoluble in the aqueous solution in which the test procedures must necessarily be carried out. A second, hypothetical possibility is that certain antibodies may recognize antigens found only on autologous platelets. A third, welldocumented explanation is that a drug metabolite can be the sensitizing agent [10] (fig. 3). In such cases, it is sometimes possible to use urine or an extract of urine from a normal subject taking the implicated medication as the source of ‘drug’ in the assay. For reasons not fully understood, glucuronide conjugates of the unmodified drug or one of its phase 1 metabolites appear to be especially prone to induce DDAbs. Treatment The most important ‘therapeutic’ maneuver is discontinuation of the sensitizing medication, af-

Aster

(3.2)

Urine

(34.4)

Naproxen glucuronide

(60.3)

Urinary metabolites

(5.1)

Desmethyl naproxen

(2.8)

Naproxen

(4.1)

Buffer

Fig. 3. Drug-dependent binding to platelets of an antibody from a patient who experienced acute thrombocytopenia after taking the non-steroidal anti-inflammatory drug naproxen. Serum from the patient contained an IgG antibody that recognized normal platelets in the presence of naproxen glucuronide in a flow cytometric assay. Antibody binding was not promoted by naproxen or by urine from a normal individual taking naproxen.

However, positive reactions were obtained with drug metabolites isolated from the same urine (‘urinary metabolites’) and with purified naproxen glucuronide. No reaction was obtained with 6-0-desmethyl naproxen, another metabolite of naproxen. Mean fluorescence intensity values are shown in parentheses. From Bougie and Aster [10].

ter which bleeding symptoms typically subside within a day or two and platelet counts return to normal within 1 week. In patients taking multiple drugs, it is appropriate to discontinue all of them for 1 week and, where necessary, substitute pharmacological equivalents that have different chemical structures. Since fatalities have resulted from intracranial and intrapulmonary hemorrhage, platelet transfusions are indicated in patients with serious bleeding symptoms even though a significant post-transfusion elevation of the platelet count may not be achieved. Intravenous -globulin and even plasma exchange have been used in rare instances, but their effectiveness is not established. Corticosteroids are often administered on the supposition that the patient may have autoimmune thrombocytopenia, but there is no evidence that they affect the outcome. In rare instances, re-challenge with the suspected

sensitizing drug may be justified to establish a diagnosis. However, only very small quantities of the suspect medication should be used in the initial trial. Remission, once achieved, is sustained indefinitely unless the patient is inadvertently reexposed to the drug at a later date. Since antibodies often disappear within a few months, re-induction of thrombocytopenia at a later date may not occur until several doses have been taken over a period of several days or a week.

Blood Dyscrasias Caused by Hypersensitivity to Drugs

Induction of Platelet-Specific Autoantibodies Autoimmune thrombocytopenia (AITP) is a relatively common condition ordinarily not directly associated with drug sensitivity. However, clinical and laboratory observations suggest that certain medications can induce platelet-reactive autoantibodies, leading to a clinical picture indistinguishable from AITP [11]. Gold salts, formerly

311

Prednisone TMP + SMX

250

Platelets (per µl × 10–3)

Fig. 4. Chronic AITP in a patient who presented initially with sulfamethoxazole (SMX)-dependent, platelet-reactive antibodies. SMX-dependent antibodies were identified in acute phase serum together with GPIIb/ IIIa-specific non-drug-dependent autoantibodies. Persistent non-drugdependent antibodies reactive with autologous platelets were identified during weeks 1, 5, and 9. SMX = Sulfamethoxazole; ICH = intracranial hemorrhage; IvIgG = intravenous  -globulin. From Aster [11].

IvIgG

200

Profuse bleeding

150

Possible ICH

100 50 0 0

used widely for the treatment of rheumatoid arthritis, provide the most convincing example of a probable cause-and-effect relationship. Other drugs implicated as possible triggers for AITP include levodopa, procainamide, penicillamine, and sulfamethoxazole. An example of a patient with AITP apparently triggered by sulfamethoxazole is illustrated in figure 4. It has been speculated that gold salts and possibly other drugs may perturb processing of platelet membrane glycoproteins by macrophages in such a way that ‘cryptic peptides’ are produced that trigger an autoimmune response in rare individuals [11]. Patients with drug-induced AITP respond to the same therapies used to treat idiopathic AITP. Immune Thrombocytopenia Induced by Ligand-Mimetic Fibrinogen Receptor Antagonists Description Platelets carry about 80,000 copies of the integrin IIb/
3 (GPIIb/IIIa). ‘Activation’ of this protein complex by intracellular signals leads to a conformational change that renders it capable of binding the coagulation protein, fibrinogen, leading to platelet aggregation, an essential step in normal hemostasis. Fibans are a class of drugs that bind tightly to an arginine-glycine-aspartic acid

312

Prednisone

2

4

6

8

10

12

14

Weeks

(RGD) recognition site on the GPIIb/IIIa complex and inhibit its interaction with fibrinogen, thereby preventing platelet aggregate formation. Fibans are RGD mimetics that bind with high affinity to GPIIb/IIIa, thereby inhibiting fibrinogen binding and platelet aggregation. Two drugs of this class, eptifibatide and tirofiban, are widely used to prevent re-thrombosis in patients undergoing coronary angioplasty. Other fiban-type drugs are in development. A unique feature of the fiban drugs is that between 0.2 and 2% of treated patients experience acute thrombocytopenia after the first exposure to these agents. Patients with fiban-induced thrombocytopenia may be asymptomatic or have only mild bleeding, even when the platelet count is very low. However, lifethreatening bleeding and fatalities have been described [12]. Pathogenesis As noted, fiban-induced acute thrombocytopenia differs from quinine-type thrombocytopenia in that an acute drop in the platelet count occurs within a few hours of the first exposure to drug. Serologic studies have shown that fiban-induced thrombocytopenia is caused by naturally occurring antibodies that react with the GPIIb/IIIa

Aster

complex when its RGD-binding site is occupied by the fiban being administered. It is known that binding of RGD or an RGD mimetic to GPIIb/ IIIa induces significant conformational changes in the heterodimeric molecule that are recognized by certain murine monoclonal antibodies specific for epitopes known as LIBS (ligand-induced binding sites). Evidence suggests that antibodies causing thrombocytopenia in patients treated with fiban drugs are specific for such determinants, but this is not yet formally established. Why such antibodies are found naturally in a significant number of normal individuals and whether they serve any physiologic purpose is an interesting and unresolved question. Thrombocytopenia Induced by Abciximab Description Abciximab, the first therapeutic agent approved for human use that is derived from a murine monoclonal antibody, is a chimeric (human/ mouse) Fab fragment that binds with high affinity to an epitope on the IIIa component of GPIIb/ IIIa close to a site that is essential for fibrinogen binding [13]. Like the fiban drugs, abciximab inhibits platelet aggregation by competing with fibrinogen for binding to activated GPIIb/IIIa. The drug is usually infused for 12–24 h following coronary angioplasty to reduce the likelihood of restenosis. About 1–2% of patients given abciximab for the first time and more than 10% given the drug a second time develop thrombocytopenia within a few hours of starting treatment. After discontinuing abciximab, platelet levels return to normal within a few days. Many patients, even those with profound thrombocytopenia (platelets !10,000/l) are asymptomatic during the thrombocytopenic episode but a few experience life-threatening bleeding. The reason for these extreme differences in hemorrhagic manifestations is unknown. In a subset of patients given abciximab, severe thrombocytopenia develops 6–8 days after the original infusion and resolves 3–7 days later.

Blood Dyscrasias Caused by Hypersensitivity to Drugs

Experimental evidence indicates that abciximab-induced thrombocytopenia following treatment with abciximab is caused by antibodies specific for the murine component of the chimeric abciximab molecule [14, 15]. Patients who experience thrombocytopenia after a second exposure have antibodies with the same specificity as those that cause thrombocytopenia after first exposure. Delayed thrombocytopenia following a first exposure is caused by the reaction of newly formed antibodies with abciximab-coated platelets, which remain in the circulation for 7–14 days after treatment [14]. Diagnosis Abciximab-specific antibodies capable of causing thrombocytopenia can be detected on the basis of their reaction with abciximab-coated platelets utilizing a flow cytometric endpoint [15]. However, special controls are needed to distinguish antibodies capable of causing thrombocytopenia from common, naturally occurring antibodies that recognize the C-terminus of the chimeric abciximab Fab fragment at the site where it is cleaved from an intact IgG molecule by papain in the manufacturing process [15]. For unknown reasons, these naturally occurring ‘Fab-specific’ antibodies appear not to be capable of causing platelet destruction although they are known to coat platelets in patients given abciximab. Treatment In the minority of patients who experience significant bleeding symptoms, platelet transfusion is recommended since, even if the transfused platelets do not elevate the platelet count, they will accelerate the clearance of transfused abciximab and shorten the duration of thrombocytopenia. Immune Complex-Mediated Platelet Destruction (Heparin-Induced Thrombocytopenia) Two to five percent of patients given unfractionated heparin develop thrombocytopenia after 5–

313

7 days of treatment. The drop in the platelet level is usually quite mild (50–100,000 platelets/l) and insufficient to cause bleeding symptoms. However, venous or arterial thromboses develop in a subset of patients, sometimes leading to loss of a limb or a fatal outcome [16, 17]. Patients who experience this complication have antibodies specific for complexes made up of heparin and platelet factor 4, a basic heparin-binding protein found normally in platelet  storage granules. The reaction of antibody with heparin/platelet factor 4 creates immune complexes which, when formed at or near the platelet surface, can activate platelets via their Fc receptors and cause platelet destruction. Thrombosis results, at least in part, from the procoagulant activity of activated platelets and platelet-derived microparticles. Other mechanisms are probably operative as well. A more detailed description of the pathogenesis, diagnosis and treatment of heparin-induced thrombocytopenia can be found in recent publications [16, 17].

Drug-Induced Immune Hemolytic Anemia

Description As with DITP, drug-induced immune hemolytic anemia (DIHA) typically develops after exposure to a sensitizing drug for at least 1 week or intermittently over a longer period of time. Symptoms range from virtually none to acute, life-threatening hemolysis associated with hypoxia, hypotension, hemoglobinemia, hematuria, and acute renal failure. Typical laboratory findings include anemia, elevated reticulocyte count, bilirubinemia, decreased haptoglobin and elevated lactic dehydrogenase levels, reflecting release of the enzyme from damaged red cells. The direct antiglobulin test (AGT) is usually positive for IgG and often for components of complement. Like DITP, DIHA can be triggered by several different mechanisms and by many different drugs (table 3). For a more complete list of implicated

314

Table 3. Drug implicated definitely as causes of immune hemolytic anemia1 Mechanism

Examples

Hapten-dependent antibody

Penicillin, cephalosporins?

‘Quinine-type’ antibody

Quinine, quinidine, cefotetin, ceftriaxone, ceftizoxime, tolmetin, etodolac, chlorpropamide, nomifensine, carboplatin, cisplatin, oxaliplatin, piperacillin, probenecid

Autoantibody induction

-Methydopa, L-dopa, procainamide

1

For a more extensive listing, see references 17–21.

medications, the reader is referred to references 18–21. It will be apparent that there are significant parallels between DITP and DIHA. Hapten-Induced Immune Hemolytic Anemia Pathogenesis As already noted, penicillin and other antibiotics containing a
-lactam structure can link spontaneously to proteins and the resulting adduct can stimulate antibodies mainly specific for the drug itself. Patients treated with penicillin often produce antibodies that recognize red cells chemically linked to penicillin. Using sensitive detection methods, similar antibodies can be detected in about 10% of normal individuals [22]. When penicillin is administered at conventional doses, penicillin-specific antibodies rarely cause hemolysis. When the drug is given in massive quantities, however, clinically significant but usually mild hemolysis occurs in a subset of patients [22]. Because penicillin is rarely used at high dosage at the present time, this type of immune hemolysis is now rare. Patients given second- and thirdgeneration penicillin derivatives (cephalosporins) also produce drug-specific antibodies, but

Aster

whether these cause clinically significant hemolysis is uncertain because such patients often have ‘quinine-type’ antibodies (see below) in addition to hapten-specific ones. Diagnosis Hapten-specific antibodies usually recognize red cells ‘coated’ with the sensitizing drug and can therefore be detected by a conventional AGT. However, a conclusive diagnosis requires the demonstration of an antibody with similar properties in an eluate prepared from the patient’s own red cells. In contrast to ‘quinine-type’ antibodies, reactions of hapten-specific antibodies with their targets can often be inhibited by soluble drug at high concentration [22]. Immune Hemolytic Anemia Associated with ‘Quinine-Type’ Antibodies Pathogenesis A more severe type of DIHA, sometimes leading to a fatal outcome, is associated with antibodies that react with red cells in the same way quininedependent antibodies bind to platelets (see above). In other words, the antibodies bind to red cells only when the sensitizing drug is present in soluble form and fail to react with cells precoated with the drug and then washed. When drug enters the circulation upon being ingested, these antibodies bind to red cells and promote their destruction by macrophages in the spleen and other tissues. Complement-mediated intravascular hemolysis also contributes to red cell destruction in many cases. The mechanism by which drugs promote binding of these antibodies to red cell membrane glycoproteins has not been characterized at the molecular level, but is likely to be similar or identical to the one responsible for platelet destruction in DITP (see above). Similarly, how drugs induce the formation of this type of antibody is unknown. However, a clue is provided by the fact that occasional antibodies are selective in their specificity in that, in the presence of drug, they recognize only red cells carrying certain al-

Blood Dyscrasias Caused by Hypersensitivity to Drugs

loantigens of the Rh, Kell or Kidd blood group systems [21]. This provides evidence that the trigger for B-cell proliferation may be a complex formed between the drug and a polymorphic red cell membrane glycoprotein. Diagnosis The direct AGT is often positive. As with platelets, ‘quinine-type’ antibodies in plasma can be documented by showing that they bind to red cells in the presence of soluble drug. Direct agglutination and/or an AGT are optimal for antibody detection. Antibodies can sometimes, but not always be identified in an elute prepared from the patient’s own red cells. As with platelets, drug metabolites sometimes produce the immunogenic stimulus. In such cases, it may be possible to detect antibodies using urine from an individual taking the implicated medication as the source of ‘drug’ [23, 24]. Treatment and Prognosis Upon discontinuing the sensitizing medication, most patients recover within a few days. Corticosteroids are often administered, but there is no evidence that they are effective. In acutely ill patients, supportive measures including red cell transfusions and hemodialysis are often required. Renal failure is usually reversible, but persistent renal dysfunction has been described. As with DITP, patients should avoid re-exposure to the implicated medication. Unfortunately, it is not rare for patients to be diagnosed as having ‘autoimmune hemolytic anemia’ on their first admission. Failure to recognize the possible role of drug sensitivity in such cases creates the possibility of re-exposure to the responsible drug at a later date, leading to more severe hemolysis and a possible fatal outcome [24]. Drug-Induced Autoantibodies Various drugs are capable of inducing true autoantibodies that recognize red cells in the absence of the drug itself. The prototype agent is -methyldopa, a drug formerly used widely for the treat-

315

ment of hypertension. About 10–40% of patients given -methyldopa for several months developed positive direct AGTs and 1–2% experienced hemolysis, usually quite mild [18]. Other drugs that appear to be capable of inducing red cell-specific autoantibodies are listed in table 3. The mechanism by which red-cell specific autoantibodies are induced by certain drugs is unknown. However, -methyldopa and levodopa have been shown to interact physically with isolated red cell membranes and it has been speculated that the drugs modify certain proteins, particularly those carrying Rh antigens in such a way that they become immunogenic. Another suggestion, that -methyldopa inhibits suppressor cells that normally inhibit immune responses, has not withstood the test of time. The diagnosis of druginduced autoimmmune hemolytic anemia is made on the basis of a positive direct AGT, usually IgG only, and the demonstration of red cellreactive antibodies, some of which have relative specificity for Rh proteins, in an eluate prepared from autologous red cells. No specific treatment, other than discontinuation of the drug, is usually necessary. More severely affected patients should be treated with measures appropriate for idiopathic autoimmune hemolytic anemia. After discontinuation of the implicated drug, weeks or months are sometimes required for the autoantibody to become undetectable.

dictable side effect of many medications. In contrast to immune thrombocytopenia and hemolytic anemia, in which drugs are implicated only in a minority of cases, it is thought that up to three-quarters of all cases of severe, idiosyncratic neutropenia are caused by drug sensitivity [25]. However, the evidence for this is based largely on epidemiologic grounds because reliable diagnostic tests capable of identifying the responsible drug in individual cases are not available.

Drug-Induced Neutropenia

Pathogenesis At least two general mechanisms, one non-immune and the other immune, appear to be responsible for idiosyncratic DINP. The former is more or less dose-related and involves interference by the drug with a critical cell function such as protein synthesis or mitosis. Since this side effect occurs only in a small number of exposed individuals, those affected are presumably predisposed to this complication by undefined genetic or acquired abnormalities. Examples of drugs that appear to cause neutropenia through

Description Various medications used in the treatment of cancer suppress hematopoiesis and may reduce neutrophil levels to a point at which the risk of infection is greatly increased (!500 cells/l). This side effect is a fairly predictable, dose-related complication of treatment with chemotherapeutic and immunosuppressive drugs, which will not be discussed here. Rather, we will consider idiosyncratic drug-induced neutropenia (DINP) – an unpre-

316

Epidemiology International surveys carried out over multi-year periods have estimated the incidence of severe idiosyncratic neutropenia to be in the range of 1–10 per million ambulatory individuals per year [25– 28]. By comparing drug exposure of affected individuals with that of unaffected persons, it has been possible to show that individuals taking certain medications have a significantly higher risk of idiosyncratic neutropenia than unexposed persons [25, 29]. A relative risk of 1100 and an incidence of 10 cases per million users per week was found for patients taking antithyroid drugs, procainamide, or sulfasalazine in one study [25]. Additional drugs associated with a lower risk were implicated in this and other studies [26–31]. Drugs shown convincingly to be associated with neutropenia/agranulocytosis on epidemiologic grounds are listed in table 4. More extensive lists of implicated drugs can be found in references 25–30.

Aster

Table 4. Drugs implicated as causes of neutropenia/ agranulocytosis1 Class of drug

Examples

Analgesic/antiinflammatory

Diclofenac, dipyrone, ibuprofen, indomethacin, phenylbutazone, sulfapyridine, sulfasalazine

Antibiotic

Ampicillin, cefotaxime, ceftazidine, ciprofloxacin, methicillin, sulfathiazole, sulfamethoxazole

Anticonvulsant

Carbamazepine, felbamate, mesantoin, phenytoin, valproic acid

Antihistamine

Cimetidine, ranitidine

Antimalarial

Amodiaquin, chloroquine, quinine

Antipsychotic/ tranquilizer

Chlorpromazine, clozapine, mianserin

Antithyroid

Carbimazole, propylthiouracil, thiocyanate

Other

Procainamide, gold salts, ticlopidine

1 Drugs shown to be associated with relative risk >10 are italicized.

non-immune mechanisms are the antipsychotic drug, clozapine, the phenothiazine group of tranquilizers, and antithyroid drugs such as propylthiouracil. Among these drugs, clozapine is exceptional in causing neutropenia in 2–3% of patients and complete failure of neutrophil production (agranulocytosis) in 0.5–1% within the first 6 months of treatment. A reactive metabolite of clozapine induces apoptosis of neutrophils at concentrations that could be achieved in vivo [32], but whether this is the mechanism by which clozapine causes neutropenia and why only a small fraction of treated patients experiences this complication is not yet resolved. The interesting suggestion has been made that reactive drug metabolites may cause neutropenia by damaging bone marrow stromal cells required to support myeloid development, rather than hematopoietic cells themselves [33].

Blood Dyscrasias Caused by Hypersensitivity to Drugs

There is little doubt that drugs can also cause neutropenia and agranulocytosis by immune mechanisms that appear to be less dosedependent [28]. In the light of numerous individual reports of patients who experienced severe neutropenia after 1 week or more of drug exposure and of the convincing demonstration in some cases of antibodies that reacted with neutrophils only when the implicated drug was present [34], it is likely that the mechanisms responsible for DINP are similar to those described above for DITP and DIHA. Some drugs such as the antithyroid agent propylthiouracil may cause neutropenia by an immune mechanism in some patients [35] and a non-immune mechanism in others. In one study of patients with quinine-induced thrombocytopenia, it was shown that quinine-dependent antibodies recognized several unknown surface glycoproteins ranging from 32 to 85 kDa in molecular weight [34]. Absence of neutrophils in the bone marrow of patients who present with agranulocytosis suggests that the immune process can affect precursor cells committed to a myeloid lineage but this has not been demonstrated experimentally. Unfortunately, very few studies characterizing drug-dependent, neutrophil-reactive antibodies have been reported, owing at least in part to technical difficulties encountered in working with granulocytes in the laboratory. Presentation Typically, patients will have taken the provocative drug for a week or more before presenting with signs of infection such as pharyngitis, stomatitis, pneumonia or sepsis. Blood studies demonstrate neutropenia but platelet levels and hematocrit are usually normal. Bone marrow findings range from total absence of identifiable neutrophil precursors to maturation arrest in the neutrophil series without abnormalities in megakaryocytes or erythroid precursors.

317

Diagnosis A high degree of suspicion and a careful history of prior drug exposure can be helpful in pinpointing the likely causative agent. Testing for the presence of antibodies that bind to neutrophils in the presence of drug using flow cytometry can sometimes be useful, but few laboratories are capable of performing such assays reliably. Treatment and Prognosis It is reasonable to advise patients beginning treatment with drugs associated with a high relative risk for agranulocytosis to contact their physician if unexplained fever, sore throat or other signs of infection develop, especially in the first month. In patients suspected of having DINP, each of the potentially responsible drugs should be discontinued and pharmacologic equivalents substituted where necessary. Treatment consists of supportive care, including appropriate antibiotic therapy when required. Prophylactic antibiotics are often used but their effectiveness has not been fully established. Granulocyte colony-stimulating factor can be helpful in selected cases [27]. Mortality rates in patients with agranulocytosis have ranged from 10 to 25% in different studies. Empirical evidence suggests that patients who have detectable neutrophil precursors on the initial marrow examination are more likely to recover than those with a total absence of myeloid elements. Because of the infrequency with which any individual drug causes neutropenia, it is impractical to monitor neutrophil levels routinely in patients taking medications known to be capable of inducing neutropenia. An exception is clozapine, for which regular blood counts are recommended by the manufacturer.

Drug-Induced Aplastic Anemia

Description Aplastic anemia is a rare, but often fatal, disorder characterized by failure of the bone marrow to produce red blood cells, neutrophils, and plate-

318

lets. A marrow biopsy usually reveals absence of hematopoietic cell precursors and replacement with fat. Epidemiology In various studies, the incidence of aplastic anemia has been estimated at between 1 and 5 cases per million per year [25, 36–38]. Idiopathic aplastic anemia appears to be more common in developing countries of Africa and Asia than in industrial nations. However, drugs are implicated as causative agents more often in developed countries, owing presumably to more extensive use of medication. As with DINP, no reliable diagnostic tools are available with which to document a cause-and-effect relationship between drug exposure and marrow aplasia in individual patients. Accordingly, various drugs have been implicated by epidemiologic studies in large populations. Drugs generally accepted to be capable of triggering aplastic anemia are listed in table 5. Additional information can be found in references 25 and 36–38. Pathogenesis As with neutropenia/agranulocytosis, drugs appear to induce aplastic anemia through both nonimmune and immune mechanisms. Examples of agents that appear to act by compromising critical cell functions through non-immune mechanisms are benzene, to which many industrial workers were formerly exposed, and chloramphenicol, an antibiotic shown to be associated with aplastic anemia in the 1950s. Although the likelihood of these drugs causing marrow aplasia appears to be doserelated, it is probable that underlying genetically determined or acquired factors predispose certain individuals to this complication. Many drugs associated with aplastic anemia are known to be metabolized to reactive intermediates which could react with proteins to produce immunogenic adducts capable of eliciting a pathologic immune response. However, experimental evidence for a cause-andeffect relationship is almost totally lacking. It is now recognized that most cases of idiopathic aplas-

Aster

Table 5. Drugs implicated as causes of aplastic anemia1 Class of drug

Examples

Analgesic/antiinflammatory

Dipyrone, felbamate, diclofenac, indomethacin, naproxen, phenylbutazone, piroxicam, sulfasalazine

Antibiotic

Cephalosporins, chloramphenicol, methicillin, sulfonamides

Anticonvulsant

Carbamazepine, phenytoin

Antihistamine

Cimetidine, chlorpheniramine, ranitidine

Antimalarial

Chloroquine, quinacrine

Antipsychotic/ tranquilizer

Chlorpromazine, prochloperazine

Antithyroid

Carbimazole, methimazole, propylthiouracil

Other

Allopurinol, chlopropamide, gold salts, penicillamine, tolbutamide

1 Drugs shown to be associated with relative risk >5 are italicized.

tic anemia (unrelated to drugs) are the result of cellular autoimmune responses mounted against hematopoietic cells and/or stromal cells that support hematopoiesis and that clinical improvement can often be achieved by immunosuppressive therapy [37]. The fact that patients with drug-induced aplastic anemia respond to immunosuppressive therapy in the same proportion as patients whose disease is apparently not drug-related favors the

possibility that both types of marrow aplasia often have an underlying immune etiology. Presentation Presenting symptoms include general weakness associated with anemia, infectious complications of neutropenia, and bleeding symptoms secondary to thrombocytopenia. Where a drug is thought to have triggered the disease, exposure for weeks or months usually precedes the onset of clinical symptoms. However, the onset of the disease is often insidious, making it difficult to accurately define the relationship between drug exposure and the onset of marrow failure. Diagnosis Since no laboratory tool is presently available to identify the responsible drug in an individual patient, the diagnosis is made on clinical grounds. Treatment and Prognosis Although it is impossible to identify the drug responsible for marrow aplasia with certainty in any particular patient, medications being taken prior to the onset of the disease should be discontinued. Anemia, bleeding, and neutropenia are treated by appropriate supportive measures. Immunosuppressive therapy leads to significant improvement in about two-thirds of all patients. Marrow transplantation is curative in about 80% of patients younger than 25 years but in only 50% of those older than 40 years. The prognosis is worse in patients who present with profound thrombocytopenia and/or neutropenia and in older age groups.

References 1 Aster R: Drug-induced thrombocytopenia; in Michelson A (ed): Platelets. New York, Academic Press, 2002, pp 593–606. 2 Aster RH: Drug-induced immune cytopenias. Toxicology 2005;209:149–153. 3 Warkentin TE, Kelton JG: Thrombocytopenia due to platelet destruction or hypersplenism; in Hoffman R, Benz EJ,

Shattil SJ, et al (eds): Hematology: Basic Principles and Practice. Philadelphia, Churchill Livingstone, 2004, vol 4, pp 2305–2325. 4 George JN, Raskob GE, Shah SR, et al: Drug-induced thrombocytopenia: a systematic review of published case reports. Ann Intern Med 1998;129: 886–890.

Blood Dyscrasias Caused by Hypersensitivity to Drugs

5 Kaufman DW, Kelly JP, Johannes CB, et al: Acute thrombocytopenic purpura in relation to the use of drugs. Blood 1993; 82:2714–2718. 6 Christie DJ, Mullen PC, Aster RH: Fabmediated binding of drug-dependent antibodies to platelets in quinidineand quinine-induced thrombocytopenia. J Clin Invest 1985;75:310–314.

319

7 Bougie DW, Wilker PR, Aster RH: Patients with quinine-induced immune thrombocytopenia have both ‘drugdependent’ and ‘drug-specific’ antibodies. Blood 2006;108:922–927. 8 Pichler WJ: Direct T-cell stimulations by drugs – bypassing the innate immune system. Toxicology 2005;209:95– 100. 9 Reddy JC, Shuman MA, Aster RH: Quinine/quinidine-induced thrombocytopenia: a great imitator. Arch Intern Med 2004;164:218–220. 10 Bougie D, Aster R: Immune thrombocytopenia resulting from sensitivity to metabolites of naproxen and acetaminophen. Blood 2001;97:3846–3850. 11 Aster RH: Can drugs cause autoimmune thrombocytopenic purpura? Semin Hematol 2000;37:229–238. 12 Bougie DW, Wilker PR, Wuitschick ED, et al: Acute thrombocytopenia after treatment with tirofiban or eptifibatide is associated with antibodies specific for ligand-occupied GPIIb/IIIa. Blood 2002;100:2071–2076. 13 Artoni A, Li J, Mitchell B, et al: Integrin
3 regions controlling binding of murine mAb 7E3: implications for the mechanism of integrin IIb
3 activation. Proc Natl Acad Sci USA 2004;101: 13114–13120. 14 Curtis BR, Divgi A, Garritty M, Aster RH: Delayed thrombocytopenia after treatment with abciximab: a distinct clinical entity associated with the immune response to the drug. J Thromb Haemost 2004;2:985–992. 15 Curtis BR, Swyers J, Divgi A, McFarland JG, Aster RH: Thrombocytopenia after second exposure to abciximab is caused by antibodies that recognize abciximab-coated platelets. Blood 2002;99:2054–2059. 16 Warkentin TE: An overview of the heparin-induced thrombocytopenia syndrome. Semin Thromb Hemost 2004; 30:273–283. 17 Davoren A, Aster RH: Heparin-induced thrombocytopenia and thrombosis. Am J Hematol 2006;81:36–44.

18 Petz LD: Drug-induced autoimmune hemolytic anemia. Transfus Med Rev 1993;7:242–254. 19 Rosse W: Autommune hemolytic anemia; in Handin RI, Lux SE, Stossel TP, et al (eds): Blood: Principles and Practice of Hematology. Philadelphia, Lippincott, 1995, vol 1, pp 1819–1945. 20 Arndt PA, Garratty G: The changing spectrum of drug-induced immune hemolytic anemia. Semin Hematol 2005;42:137–144. 21 Packham C: Hemolytic anemia resulting from immune injury; in Lichtman M, Beutler E, Kipps TJ, et al (eds): Williams Hematology. New York, McGrawHill, 2006, vol 7, pp 729–750. 22 Garratty G: Immune cytopenia associated with antibiotics. Transfus Med Rev 1993;7:255–267. 23 Salama A, Mueller-Eckhardt C: The role of metabolite-specific antibodies in nomifensine-dependent immune hemolytic anemia. N Engl J Med 1985; 313:469–474. 24 Cunha PD, Lord RS, Johnson ST, Wilker PR, Aster RH, Bougie DW: Immune hemolytic anemia caused by sensitivity to a metabolite of etodolac, a non-steroidal anti-inflammatory drug. Transfusion 2000;40:663–668. 25 Kaufman DW, Kelly JP, Jurgelon JM: Drugs in the aetiology of agranulocytosis and aplastic anaemia. Eur J Haematol 1996;60(suppl):55–59. 26 Van der Klauw MM, Wilson JH, Stricker BH: Drug-associated agranulocytosis: 20 years of reporting in The Netherlands (1974–1994). Am J Hematol 1998;57:206–211. 27 Bhatt V, Saleem A: Review: Drug-induced neutropenia – pathophysiology, clinical features, and management. Ann Clin Lab Sci 2004;34:131–137. 28 Dale D: Neutropenia and neutrophilia; in Lichtman M, Beutler E, Kipps TJ, et al (eds): Williams Hematology. New York, McGraw-Hill, 2006, vol 7, pp 907–919.

29 IAAAS: Risks of agranulocytosis and aplastic anemia. A first report of their relation to drug use with special reference to analgesics. The International Agranulocytosis and Aplastic Anemia Study. JAMA 1986;256:1749–1757. 30 Stroncek DF: Drug-induced immune neutropenia. Transfus Med Rev 1993;7: 268–274. 31 Dinauer MC, Coates TD: Disorders of phagocyte function and number; in Hoffman R, Benz EJ, Shattil SJ, et al (eds): Hematology: Basic Principles and Practice. Philadelphia, Churchill Livingstone, 2005, vol 4, pp 787–829. 32 Williams DP, Pirmohamed M, Naisbitt DJ, Uetrecht JP, Park BK: Induction of metabolism-dependent and -independent neutrophil apoptosis by clozapine. Mol Pharmacol 2000;58:207–216. 33 Guest I, Uetrecht J: Drugs that induce neutropenia/agranulocytosis may target specific components of the stromal cell extracellular matrix. Med Hypoth 1999;53:145–151. 34 Stroncek DF, Herr GP, Maguire RB, Eiber G, Clement LT: Characterization of the neutrophil molecules identified by quinine-dependent antibodies from two patients. Transfusion 1994;34:980–985. 35 Fibbe WE, Claas FH, Van der Star-Dijkstra W, Schaafsma MR, Meyboom RH, Falkenburg JH: Agranulocytosis induced by propylthiouracil: evidence of a drug dependent antibody reacting with granulocytes, monocytes and haematopoietic progenitor cells. Br J Haematol 1986;64:363–373. 36 DeLoughery T: Drug-induced immune hematologic disease. Immunol Allergy Clin North Am 1998:829–841. 37 Young NS: Aplastic anemia; in Hoffman R, Benz EJ, Shattil SJ, et al (eds): Hematology: Basic Principles and Practice. Philadelphia, Churchill Livingstone, 2004, vol 4, pp 381–417. 38 Siegel GB, Lichtman MA: Aplastic anemia; in Lichtman MA, Benz EJ, Kipps TJ, et al (eds): Williams Hematology. New York, McGraw-Hill, 2006, vol 7, pp 419–436.

Dr. Richard H. Aster Blood Research Institute, Blood Center of Wisconsin, and Medical College of Wisconsin 8727 Watertown Plank Rd Milwaukee, WI 53226-3548 (USA) Tel. +1 414 937 6338, Fax +1 414 937 6284 E-Mail [email protected]

320

Aster

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 321–339

Allergy and Pseudoallergy to Drugs and Vaccines in Children Claude Ponvert  Smail Hadj-Rabia  Pierre Scheinmann René Descartes-Paris V University, Paediatrics Department, Pulmonology, Allergology & Dermatology Unit, Sick Children’s Hospital, Paris, France

Abstract Five to 20% of subjects of all ages report suspected allergic reactions to drugs and biological substances. Children may be less affected than adults, but this difference is disputed and probably results from differences in drug exposure. Antibiotics, mainly
-lactams, antipyretics, analgesics and non-steroidal anti-inflammatory drugs, are most frequently incriminated. The most common reactions are morbilliform/maculopapular rashes, urticaria and angioedema. Other cutaneous and systemic reactions and severe anaphylactic/anaphylactoid reactions are rare. Diagnosis is based on a detailed clinical history, skin and/or in vitro tests (if possible and if validated) and challenge tests (if indicated). Prevention of relapse is based on a rigorous avoidance of the responsible and cross-reacting drugs. The results of the allergological work-up, including challenge, are often negative, suggesting that, in many instances, the drug can be tolerated again. This is especially the case in mild to moderately severe reactions such as maculopapular exanthems, non-immediate urticarias and unidentified rashes. In these cases, the reaction may result from a complex interaction between ‘danger’ signals provided by the underlying disease and a weak and transient drug-specific immune reaction. However, severe reactions like anaphylaxis (caused by myorelaxants, etc.), bullous skin reactions or drug-induced rash with eosinophilia and systemic symptoms (antiepileptics), require the same precautions as in adults and strict avoidance of the incriminated drug, even if the test results are negative. Copyright © 2007 S. Karger AG, Basel

Introduction

Five to 20% of patients report allergic-like reactions to drugs or biological substances [1, 2]. Except for antibiotics [3–5], antipyretics, analgesics and non-steroidal anti-inflammatory drugs (NSAIDs) [6–8], vaccines [9–13], myorelaxants/ neuromuscular blocking agents (NMBA) [14, 15] and latex [16, 17], only a few epidemiological data are available in children [18–24]. In the study of Menniti-Ippolito et al. [23], the incidence of adverse drug reactions (ADRs) was 15.1 per 1,000 children. The highest incidence was found in 0to 1-year-old children (34.1 vs. 7.4 in 7- to 14-yearold children) (table 1). The incidence of suspected allergic reactions (mainly cutaneous) was low (^5.18 per 1,000 defined daily doses) for most drugs, except for calcipotriol (66.67) and cefpodoxime (71.43) (table 2). Drug hypersensitivity (HS) was suspected in only 1.14–5.4% of hospitalized children [21, 22, 24]. Most reactions were mild to moderately severe, except for a few cases of erythema multiforme (EM) and anaphylaxis. It has been suggested that children develop drug HS less frequently than adults, but this difference is disputed and may result from a lower exposure to drugs [25, 26]. Although there are only a few available epidemiological studies in children with underlying diseases requiring frequent treat-

Table 1. Distribution and incidence of reported adverse drug reactions (ADRs) by age group in children (Italy 1996–1997) [23] Age group years

Children

ADRs

Incidence per 1,000 children

ADRs from vaccines

Incidence per 1,000 children

0–1 1–4 4–7 7–14

469 1,815 2,086 3,510

16 48 29 26

34.1 26.3 13.9 7.4

5 6 2 1

10.7 3.3 0.9 0.3

Total

7,890

119

15.1

14

1.8

ments, it has been suggested that the frequency of (suspected) allergy to drugs, mainly to antibiotics, was higher in children with cystic fibrosis than in other children [27]. Similar results have been reported in adult patients [28, 29]. The most frequently incriminated drugs in children are antibiotics and antipyretics/NSAIDs [1, 3, 5, 19, 23]. The commonest reactions are morbilliform/maculopapular rashes (MPR: 60– 80%) and urticaria and/or angioedema (20–30%). Serum sickness-like disease (SSLD) occurs above all in children treated with first-generation cephalosporins [30, 31], but may also occur with other drugs such as minocyclin during long-term treatment of acne [32]. There are also rare severe reactions which can be fatal. They include StevensJohnson syndrome (SJS) and its maximal form, toxic epidermal necrolysis (TEN), and acute generalized exanthematic pustulosis (AGEP). In addition, there are rare systemic drug reactions like drug-induced reaction with eosinophilia and systemic symptoms (DRESS), which occurs mainly with antiepileptics, but also with minocyclin treatment for acne, both conditions not so infrequent in children and adolescents [19, 20, 23, 32, 33]. Respiratory symptoms are less frequent than skin reactions [1], although they are reported in 17–24% of children with suspected HS to NSAIDs [34, 35]. Drug-induced anaphylactic and anaphylactoid reactions are rare [36–38], and occur mainly in the context of NMBA and antibiotic administration. However, in a recent survey of fa-

322

tal outcomes attributed to drugs (excluding overdose and vaccines), the main cause of death was hepatitis (mostly with valproic acid in the frame of DRESS), while anaphylaxis accounted only for 5.4% of death (mainly due to anesthesics and antibiotics) [36]. Although, except for a few types of reactions (anaphylaxis and immediate/very accelerated reactions, potentially harmful toxidermias) and for very specific drugs and biological substances (i.e. NMBA and latex), skin tests and challenge are negative in most reactions to commonly used drugs and biological substances in children, the allergological work-up for drug allergy diagnosis is similar in children and adults and is based on clinical examination at the time of the reaction (if possible), on a detailed anamnesis of the clinical history [39], skin [40] and/or in vitro tests with the suspected and cross-reacting drugs (if possible and validated), and on challenge (if indicated) (table 3).

Clinical Manifestations

Mild Skin Reactions Maculopapular rashes (fig. 1) are characterized by pink-red papules appearing on the upper trunk and progressing distally in a symmetric fashion. Symptoms usually begin at the end of the first week of treatment, are not or only a little pruriginous, and resolve spontaneously without

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

Table 2. Incidence of reported ADRs by substance in children (Italy 1996–1997) [23] ADR

DDD

Incidence per Substances 1,000 DDD

Type of ADR

2 1 1 1 12

28 15 77 193 4,373

71.43 66.67 12.99 5.18 2.74

2 4 4 5 1 5 21 1 2 7

798 1,969 2,210 3,540 711 4,102 18,374 915 1,836 6,710

2.51 2.36 1.81 1.41 1.41 1.22 1.15 1.09 1.09 1.04

ceftibuten cefuroxime axetil azithromycin cotrimoxazole ferrous gluconate erythromycin amoxicillin + clavulanic acid nimesulide ketoprofen cefaclor

1 1 1 15

1,165 1,292 1,380 27,813

0.86 0.77 0.72 0.54

cyproheptadine domperidone loratadine amoxicillin

4 1 1

15,513 7,369 33,114

0.26 0.14 0.03

salbutamol nedocromil beclometasone

cutaneous cutaneous gastrointestinal gastrointestinal cutaneous, gastrointestinal, eosinophilia cutaneous, gastrointestinal cutaneous, gastrointestinal cutaneous, gastrointestinal cutaneous, neurological gastrointestinal gastrointestinal cutaneous, gastrointestinal cutaneous hematuria, hypertranspiration cutaneous, gastrointestinal, neurological cutaneous cutaneous neurological cutaneous, gastrointestinal, edema, fever cutaneous, neurological cough angioedema cough

cefpodoxime calcipotriol josamycin cefatrizine clarithromycin

ADR = Adverse drug reaction; DDD = defined daily dose.

sequelae in 5–10 days [41]. The most frequently suspected drugs are anti-infectious drugs,
-lactams and sulfonamides in particular. Most children with suspected
-lactam-induced MPR did not relapse during subsequent treatments with the suspected
-lactams [42]. Therefore, and because drug-specific allergy tests are usually negative, it is assumed that most MPR in children are a consequence of drug use in combination with the infectious and/or inflammatory diseases for which the drugs have been prescribed [43]. The lesions of urticaria are pink-red and blanching, edematous and geographic, pruritic papules or plaques of varying sizes (fig. 2). The plaques may become confluent and/or polycyclic. The lesions are transient. Urticaria may be asso-

Drug Allergy in Children

ciated with angioedema, either localized (distal extremities, eyelids and face), or generalized. Facial edema is frequent in NSAID-induced reactions [44]. In ‘hemorrhagic’ urticaria, residual purpura may persist several days, and mimic the target lesions of EM [33, 41]. The most frequently suspected drugs are anti-infectious drugs and NSAIDs. The probability for the drug to be the cause of the reaction is increased in immediate and accelerated reactions [34, 41, 45, 46]. However, most urticaria and/or angioedema do not result from drug HS [47, 48]. In the prospective study of Mortureux et al. [48], drug HS was suspected in less than 10% of children with acute urticaria/angioedema, and more than 80% of the reactions were associated with (viral) infections.

323

Table 3. Schematic diagnostic work-up in children with suspected allergy to drugs (mainly anti-infectious drugs) (1) In children with reactions suggestive of immediate-type HS Severe anaphylaxis and immediate reactions ⴢ Eviction of all the drugs in the same structurally related family (high risk of cross-reactivity) until allergological work-up ⴢ Allergological work-up compulsory: immediate-reading skin tests (if possible and validated) 8 in vitro tests (IgE, HR, CAST, BAT) 8 challenge (if necessary) Mild to moderate accelerated urticaria and angioedema ⴢ Eviction of the suspected (class of) drug until allergological work-up ⴢ No contraindication of the other (classes of) drugs in the same structurally related family ⴢ Allergological work-up advisable: immediate-reading (pricks and ID tests) and, if negative, non-immediatereading skin tests (ID and/or patch tests) (if possible and validated) 8 in vitro tests (IgE, HR, CAST, BAT, LPT) 8 challenge (if necessary) (2) In children with reactions suggestive of non-immediate HS Mild to moderate delayed urticaria and angioedema ⴢ Eviction of the suspected (class of) drug until allergological work-up ⴢ No contraindication for other (classes of) drugs in the same structurally related family ⴢ Allergological work-up advisable: non-immediate-reading skin tests (ID and/or patch tests) (if possible and validated) 8 in vitro tests (LPT) 8 challenge (if necessary) Serum sickness-like reactions ⴢ Eviction of the suspected drug (and structurally related drugs?) until allergological work-up ⴢ No contraindication for other (classes of) drugs in the same structurally related family ⴢ Allergological work-up (controversial): non-immediate-reading skin tests (ID and/or patch tests) (if possible and validated) 8 challenge (if necessary) Erythema multiforme/potentially harmful toxidermias ⴢ Diagnostic tests for infectious etiology Herpes viruses and mycoplasmas >> coxsackies and other viruses (EM, SJS) Staphylococcus >> streptococcus (TEN) ⴢ If diagnostic tests for infection are negative or have not been performed Eviction of the suspected (and structurally related) drugs Allergological work-up (controversial): non-immediate-reading skin tests (ID and/or patch tests) (if possible and validated) 8 in vitro tests (LPT). Challenge strictly forbidden (except for mild to moderately severe EM, if necessary) Other reactions (MPR, unidentified mild rashes) ⴢ Challenge with the suspected drug (very low risk of relapse) ⴢ Allergological work-up in case of relapse during challenge: non-immediate-reading skin tests (ID and/or patch tests) (if possible and validated) 8 in vitro tests (LPT). If positive, challenge with other (classes of) drugs in the same structurally related family BAT = Basophil activation test; CAST = cellular allergen stimulation test; IgE = specific IgE determination; HR = histamine release test; ID = intradermal; LPT = lymphocyte proliferation test; MPR = maculopapular rash.

According to Carder [41], fixed drug eruptions (FDE) (fig. 3) account for 22% of pediatric drug eruptions. Lesions begin as round, red-purple, pruritic or burning plaques of the lips, face, trunk, extremities and genitalia, a few hours to a few days after the beginning of treatment. The

324

evolution towards bullous or necrotic lesions is not rare. Usually, erythema disappears and leaves pigmented round lesions that may persist for months or years, and relapse when the children are treated again with the responsible drug. The most common drugs involved are sulfonamides,

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

Fig. 1. Maculopapular rash in a child.

Fig. 3. Fixed drug eruption in an adolescent boy.

non-opioid analgesics, aminopenicillins, and macrolides. However, FDE are often under-recognized [31, 49]. One of the rather severe forms is EM (fig. 4). It is an acute cutaneous eruption characterized by ‘target’ symmetric lesions of the upper extremities. Lesions may also affect lower extremities and trunk. Usually, the lesions resolve in 1–4 weeks (mean 1–2 weeks). Purpuric target-like plaques of acute hemorrhagic edema and duskycentered lesions of hemorrhagic urticaria are a frequent clinical differential diagnosis of EM. The most common causes of EM are infections by herpes simplex virus and Mycoplasma pneumoniae, drug HS being rarely suspected [20, 41, 50, 51].

Fig. 2. a, b Acute urticaria in a child.

Drug Allergy in Children

Severe Skin and Systemic Symptoms SJS and TEN are serious illnesses that involve skin and several mucosal sites (conjunctiva, oral mucosa and genitalia in particular). They have been classified according to the extent of detached/detachable epidermis (^10% for SJS; 10– 30% for SJS-TEN; 630% for TEN). Prodromic symptoms (fever, headache, malaise, sore throat, and diarrhea) occur in more than 50% of the patients. Initially, cutaneous lesions of SJS are simi-

325

Fig. 4. Erythema multiforme in a child.

lar to those of EM. However, generalization occurs rapidly, and the lesions evolve into bullae and extensive areas of skin necrosis and separation [51]. Drugs are the major cause for SJS-TEN, and infections are only rarely implicated [50, 51]. Respiratory and gastrointestinal mucosae can also be involved in SJS-TEN and TEN. Anticonvulsants, sulfonamides,
-lactams and NSAIDs are the most frequent suspected causal agents of SJS-TEN and TEN. Typically, symptoms occur 2 weeks after the beginning of treatment with the responsible drug, but may occur sooner in case of previous exposure to the offending drug. TEN in adults is associated with a high mortality rate (;30%), essentially due to sepsis and hypovolemia. Prognosis is dependent on the extension of skin and mucosal lesions, the underlying diseases and older age. Importantly, some sequelae, like

326

eye damage, may remain permanently. Promptness of treatment in an experienced center is therefore necessary. Drug-induced rash with eosinophilia and systemic symptoms (also called drug hypersensitivity syndrome, anticonvulsant hypersensitivity syndrome, etc.) is a severe multiorgan syndrome associating fever, rash (ca. 70%), eosinophilia (ca. 70%), lymphadenopathy and hepatitis. Kidneys, lungs, colon, heart, thyroid, brain, and blood may also be involved. Because some symptoms mimic generalized viral infection, an initial misdiagnosis of infection is frequent. This may delay the discontinuation of the responsible drug and substantially aggravate the clinical picture. DRESS has been reported in children treated with anticonvulsants and sulfonamides, and represents a major cause of death due to adverse drug effects,

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

mainly due to fulminant hepatitis [33, 41]. An important role of drug-specific T cells in the immunopathology of DRESS has been elucidated in recent years by analyzing lamotrigine- or carbamazepine-specific T-cell clones [52, 53]. As long as the massive stimulation of the immune system persists, patients may be affected by an intolerance to other xenobiotics. A frequent complication is a reactivation of herpes viruses (HHV-6, CMV), mainly in week 3 after the onset of the symptoms [54, 55]. AGEP is characterized by a sudden onset of fever and a rapidly progressive edematous, burning and/or pruritic erythema, beginning a few hours to days (!24 h in 50% of the cases) after the administration of the responsible drug. Scattered whitish little pustules develop rapidly. The lesions first appear on the face, neck, trunk and folds, and generalize within a few days to the whole trunk and the upper and lower limbs. Usually, spontaneous healing occurs in less than 2 weeks. Although AGEP may be caused by viruses (enteroviruses, cytomegaloviruses) and mercury salts, most cases are induced by drugs and relapse if the offending drugs (anti-infectious drugs, pristinamycin and amoxicillin in particular) are re-administered [56]. SSLD is characterized by cutaneous lesions (e.g. urticaria) associated with arthralgias and moderate fever. SSLD usually begins several days to 3 weeks (average 1 week) after the beginning of treatment, and resolves spontaneously in 7–10 days without sequelae. Unlike true serum sickness, no lymphadenopathy, circulating immune complexes and proteinuria are detected [41]. The most commonly suspected drugs are sulfonamides, macrolides and
-lactams, cefaclor in particular. The prevalence of cefaclor-associated SSLD in children is estimated between 0.02– 0.05% and 1.1–3.4% of treated children [30, 57– 60]. Most reactions occur in children younger than 5–6 years of age. Although prior treatment with the suspected drug is not required, SSLD usually occurs during the second or third course

Drug Allergy in Children

of therapy [30, 41, 58]. In general practice, re-administration of the suspected drug is contraindicated. However, although drug-induced SSLD has been diagnosed in children with suspected
-lactam allergy, based on late responses in skin or oral challenge (OC) tests, OC with a complete course of the suspected
-lactam has been tolerated in most children with suspected
-lactaminduced SSLD and negative responses in immediate and non-immediate-reading skin tests [61, and Ponvert et al., unpubl. results]. Ocular and Respiratory Reactions Isolated conjunctivitis, rhinitis and/or asthma are not frequent, except in children with NSAID HS [34, 35]. However, ocular involvement (formation of synechiae) is a dangerous complication in SJS-TEN and TEN, and requires special care. Anaphylaxis Drug-induced severe anaphylactic reactions, associated with respiratory (laryngeal and/or bronchial) and cardiovasculatory (hypotension, shock) distress are very rare in the general pediatric population. They may occur in the context of operations, due to NMBA, plasma expanders and perioperative use of antibiotics. In addition, one should always be aware that latex allergy may cause anaphylaxis as well, in particular in children having undergone multiple operations, which contributed to the sensitization. Overall, drugs were the suspected cause of 6– 11% of anaphylactic reactions in children [38, 62]. However, in the prospective study of Dibs and Baker [37], 43% of 55 anaphylactic reactions in 50 children were attributed to drugs (antibiotics in particular) and latex.

Drug-Specific-Induced HS

Anti-Infectious Drugs Seven to 20% of the children report suspected allergic reactions to anti-infectious drugs, the
-

327

lactams being the most frequently involved (50– 75%). Other antibiotics are suspected in 25–30%, and sulfonamides in 10–20% of the cases [1, 3–5]. The most frequently reported reactions are MPR (;75%), and urticaria and/or angioedema (20– 25%). Other reactions are rare, except for SSLD, which occurs above all in children treated with cefaclor [30, 41, 57–60]. Skin and challenge tests have shown that 2– 60% (mean 13–15%) of the children with suspected anti-infectious drug HS are really allergic to these drugs [42, 45, 46, 61, 63–66]. Immediatetype HS is diagnosed in 25–60% of the children with immediate and severe anaphylactic reactions [45, 46, 63]. Only a few children with nonimmediate mild reactions are diagnosed allergic to antimicrobial drugs. In most of the children, non-immediate reactions probably result from the combination of the infectious diseases for which the drugs have been prescribed and transient weak allergy [20, 43, 48, 67]. In a few children, however, the reactions may result from allergic or non-allergic HS to other drugs, such as antipyretics, analgesics and NSAIDs, or from food or excipient allergic or non-allergic HS [46].
-Lactams Diagnosis of
-lactam allergy is based on a detailed clinical history, skin tests and, if necessary, challenge tests. The diagnostic value of in vitro tests for immediate-type (serum-specific IgE determination, basophil activation test, histamine and leukotriene release tests) and non-immediate type (lymphocyte proliferation test) HS to
-lactams is low [68–72]. Thus, a negative in vitro test cannot exclude allergy diagnosis. It is important to combine the in vivo and in vitro tests as well as the history for the final diagnosis, since none of the tests per se has a high enough sensitivity and specificity to prove or disprove a sensitization. Although 10% of the children frequently treated with penicillins have immediate responses in skin tests [73], immediate-reading skin tests have good diagnostic value in children with suspected

328


-lactam HS. OC tests diagnosed immediatetype HS to
-lactams in a few children with negative skin tests only [46, 63]. A well-tolerated OC may theoretically boost a latent sensitization, and be responsible for a reaction during subsequent treatment [61, 74, 75]. However, the risk of resensitization by OC appears very low [76, 77]. Finally, skin tests and OC diagnosed cross-reactivity between penicillins and cephalosporins themselves, and penicillins and cephalosporins, in 25– 84.2% of the children with immediate-type HS [46, 63]. For an unknown reason, the frequency of cross-reactivity appears much higher in children than in adults. Non-immediate responses in intradermal (ID) and/or patch tests have been reported in children with MPR, non-immediate urticaria/angioedema, SSLD and AGEP [46, 66, 78, 79]. Those tests are not standardized, and may give false-positive [80, and Ponvert et al., unpubl. results] and falsenegative responses [46, 78]. In our studies, about two thirds of non-immediate sensitizations to
lactams were diagnosed by means of OC [46, 78, 79]. In contrast to immediate reactions, the risk of cross-reactivity between
-lactams in the same and other classes is low [46, 80–82]. In clinical practice: (1) In children with anaphylactic and immediate reactions, all classes of
-lactams should be withdrawn because there is a surprisingly high risk of cross-sensitization to several
-lactams in the same and different classes. A work-up, based on skin tests with the suspected
-lactam(s) and other
-lactams from the same and other classes, and challenge (if necessary), should be performed to confirm or refute the diagnosis of allergy, and to determine if the child is sensitized to one or several (classes of)
-lactam(s). (2) In non-immediate mild to moderate urticaria and/or angioedema, the allergologic workup is also recommended. The risk of
-lactam HS is low, and cross-reactivity to several
-lactams is rare. Thus,
-lactams from other classes than that suspected are not contraindicated.

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

(3) In SSLD, the diagnostic value of non-immediate-reading skin tests is disputed. It is generally assumed that, whether skin tests are positive or negative, the suspected
-lactam is contraindicated, with no contraindication for other
lactams in the same and other classes [30, 83]. However, most children with SSLD and negative responses in skin tests tolerate prolonged OC with the suspected
-lactams [61, and Ponvert et al., unpubl. results]. (4) EM and SJS in children are mainly due to viral infections. If drug HS is suspected, ID and patch tests may be performed. Negative skin tests do not completely exclude, but probably reduce the likelihood of non-immediate
-lactam HS diagnosis. It is generally assumed that the suspected
-lactam(s) should be contraindicated. In our experience, at variance with classical opinion, most children with EM or SJS and negative skin tests tolerate prolonged OC with the suspected
-lactams [Ponvert et al., unpubl. results]. (5) Non-immediate-reading skin tests may be performed in children with FDE (preferentially in the affected area, if possible and if skin tests are necessary). They can also be performed in SJS/ TEN, DRESS and AGEP, but challenge tests are contraindicated because there is a high risk of relapse.
-Lactams of the same class are contraindicated, but not
-lactams of other classes. (6) In the other cases, such as unidentified mild rashes and MPR, skin tests are probably useless. The diagnosis of
-lactam HS is often ruled out based on tolerance of the challenge (full daily therapeutic dose ! several days, at home) with the suspected drug. However, an allergologic work-up is needed in the rare children relapsing during challenge. Other Antimicrobial Drugs Epidemiological and allergological data have been published in several detailed reviews [1, 5, 84]. Allergic-like reactions to antimicrobial drugs other than
-lactams are less frequent than those

Drug Allergy in Children

reported to
-lactams. In vitro tests for immediate and non-immediate HS to non-
-lactam antimicrobial drugs are mainly experimental and positive results can be valuable, but need to be controlled in non-sensitized individuals. Diagnostic value of immediate responses in skin tests with these drugs is controversial [45, 85–87]. Diagnostic and predictive values of skin tests for non-immediate HS depend on the type of the reaction. Late responses in patch and ID tests with non-
-lactam antibiotics have been reported in adult patients with non-immediate urticaria and/ or angioedema, MPR and potentially harmful toxidermia [88], but the sensitivity is probably low and the specificity of these tests needs to be controlled in each case. Allergic-like reactions to macrolides are rare (;0.5% of treatments). In adults, immediate responses in skin tests with these antibiotics had a low sensitivity, most patients being diagnosed allergic by means of OC [85]. In two pediatric studies, all children with suspected macrolide allergy tolerated OC, suggesting that the reactions did not result from macrolide HS [45, 89]. When clinical history is highly suggestive of HS to a particular macrolide, treatment with macrolides from other classes may be recommended, because cross-reactivity between macrolides is rare [90]. Most reactions to quinolones have been reported in adults. In patients with suspected immediate HS, the diagnostic value of immediatereading skin tests with these drugs is low [86], but an experimental IgE assay proved to be valuable [91]. Drug HS was suspected in 8.5% of the children treated with sulfonamides [3]. Immediate responses in skin tests (pricks and ID) have been reported in patients with immediate and accelerated urticaria and angioedema, but the methodology of immediate-reading skin tests with sulfonamides is not standardized. The diagnostic value of these tests is low [87]. In our studies, immediate-reading skin tests and OC were negative in all children with suspected immediate HS to sulfon-

329

amides, suggesting that immediate HS to sulfonamides in children is a rare condition [89, and Ponvert et al., unpubl. results]. However, in another study, OC were positive in 12.5 and 20.8% of children with immediate or accelerated and delayed reactions to sulfonamides, respectively [45]. To our knowledge, there is no reported case of rifampicin/rifamycin allergy in children. In adults, immediate-reading skin tests with these antibiotics have a good sensitivity and a good specificity, provided they are performed with very low concentrations of the drugs [92]. Based on immediate responses in skin tests, we diagnosed rifampicin/rifamycin allergy in 3 children with immediate urticaria and/or angioedema induced by oral rifampicin and rifamycin eye drops [Ponvert et al., unpubl. results]. The Multiple Antibiotic Sensitivity Syndrome Ten to 15% of the patients with suspected antibiotic allergy report allergic-like reactions to several antimicrobial drugs [93–95]. If multiple antibiotic sensitivity is defined by simple intolerance of various antibiotics, it is not a rare condition in children, as up to 40% of the children with suspected
-lactam HS reported suspected allergic reactions to other drugs such as macrolides and sulfonamides [46]. In these children, risk of
lactam allergy was not dependent on the clinical history of reactions to other antimicrobial drugs. OC with the suspected antibiotics were negative in most cases, which is consistent with studies showing that
-lactam allergy and non-
-lactam allergy develop independently [94]. Recurrent reactions to several drugs are often some transient, urticarial rashes, which may result from an overstimulation of mast cells, probably due to infections, intolerance to excipients, coincident allergy or pseudoallergy to foods, etc. Antipyretics, Analgesics and NSAIDs The estimated prevalence of suspected HS to these drugs in the general population is between 0.3 and 1% [7, 96, 97]. No allergic or allergic-like

330

reaction was reported in 27,000 randomized children prospectively treated with ibuprofen or paracetamol [6]. In another study, only 25% of the patients with suspected NSAID HS were children [7]. Finally, HS to NSAIDs, which is more an intolerance than an allergy involving the specific immune system, has been suspected in only 2% of 754 ADRs in hospitalized children [8]. The most frequently reported reactions affect the skin, as urticaria and/or angioedema, facial edema in particular, while respiratory reactions (rhinitis and asthma), particularly in connection with polyposis nasi and asthma, are extremely rare in children. Anaphylactic or anaphylactoid reactions are rare. Potentially harmful toxidermias have a different pathomechanism and can be seen only with certain NSAIDs (piroxicams, diclofenac), and very rarely only [8, 35, 98]. Studies, mainly based on OC, show that 13–50% of the children with suspected HS to NSAIDs are really intolerant to these drugs [34, 35, 45]. Risk factors for HS to antipyretics, analgesics and NSAIDs are a personal atopy, age and severity of the reactions [7, 34, 99]. A few cases of immediate or non-immediate HS with positive responses in skin and/or in vitro tests to a specific (family of) NSAIDs have been reported in children [100] and adults [101–103]. However, most reactions result from a non-specific HS, as suggested by a high degree (25–100%) of cross-reactivity between the various structurally unrelated families of NSAIDs, and 2–50% of cross-reactivity with drugs with a low cyclooxygenase-1 (COX-1) inhibitory activity, such as paracetamol and coxibs [7, 34, 44, 104, 105]. In our experience, 12.3% of the children with NSAID intolerance were intolerant to paracetamol, and all the children intolerant to therapeutic doses of paracetamol were intolerant to NSAIDs [34, 104]. Diagnostic value of immediate-reading skin tests with NSAIDs is not validated [102, 106, 107]. A self-made RAST has detected specific anti-aspirin IgE in a few patients with immediate reac-

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

tions to aspirin, but the sensitivity and specificity of the test are unknown [101, 103]. Finally, the diagnostic value of other in vitro tests (histamine and cysteinyl-leukotriene release, flow cytometric basophil activation) is highly variable upon the drug tested [69, 102, 108]. Diagnosis of allergic or non-allergic HS to antipyretics, analgesics and NSAIDs is primarily based on a convincing clinical history (immediate and/or severe reactions, recurrent reactions to isolated intake of the drugs) or challenge tests [45, 100, 107]. Challenges with drugs with a low COX-1 inhibitory activity are tolerated in 50–98% of the patients with a positive response in challenge test with a drug with a high COX-1 inhibitory activity [7, 34, 44, 105]. However, treatments with high doses of NSAIDs with a low COX-1 inhibitory activity may induce relapses in about 10% of the patients [109, and Ponvert et al., unpubl. results]. Prevention of relapse is based on eviction of the offending drug(s) and treatment with tolerated NSAIDs. Other drugs, such as opioids, azopropazone and benzydamine are an alternative to NSAIDs in NSAID-intolerant patients [110, 111]. However, the anti-inflammatory activity of these drugs is low. Finally, the efficacy of ‘desensitization’ with NSAIDs is disputed, especially in patients with cutaneous and anaphylactic or anaphylactoid reactions [98, 112]. Vaccines Most frequent reactions are large local inflammatory reactions, and mild to moderate non-immediate urticaria/angioedema and non-urticarial rashes induced by diphtheria (D) and tetanus (T) toxoid-containing vaccines and hepatitis B (HBV) vaccines [113]. Other reactions, such as anaphylactic and anaphylactoid reactions, serum sickness, vasculitis, eczema, persistent nodules, recurrent abscesses, EM, etc., are rare. Vaccine components may induce IgE-dependent HS reactions, such as anaphylaxis and immediate urticaria and angioedema to toxoid- [114], ovalbumin- [115–117], gelatin- [13, 118–120] and pneu-

Drug Allergy in Children

mococcal antigen-containing vaccines [121]. Dextran-containing vaccines (BCG) may induce anaphylactoid reactions [122, 123]. In a very large prospective study, in more than 2 million US children and adolescents, the mean rate of anaphylactic/anaphylactoid reactions to vaccines was 0.65 per 1 million vaccine doses [124]. The highest risk rate (21.2) was found for the DT vaccine. The risk rate was 3.4–3.5 for measlesmumps-rubella (MMR) and DTP-Hib vaccines, and lower than 2 for polio, DTP, Hib and HBV vaccines. Finally, excipients of vaccines and toxoid-containing vaccines may induce non-immediate-type HS reactions, either local, such as Arthus-type reactions [125] and recurrent abscesses [126] induced by toxoids, or generalized, such as delayed urticaria, angioedema and non-urticarial rashes induced by gelatin-containing vaccines [127–129]. Reactions at the Site of Vaccine Injection Arthus-type reactions occur in children hyperimmunized by previous injections of the vaccines, and are easily diagnosed by clinical history and serum-specific antitoxoid antibody (IgM/ IgG) determination, with high levels of these antibodies 3–4 weeks after the reaction. Large local inflammatory reactions may also occur in nonimmunized children and after booster injections of vaccines containing high concentrations of diphtheria toxoid or aluminium hydroxide, independently of their serum-specific antitoxoid IgG levels. Thus, most large local inflammatory reactions probably result from non-specific activation of the inflammatory system by high doses of aluminium salts and/or microbial components, with good tolerance of booster sequential injections of the vaccines [114]. Late responses in ID tests with tetanus toxoid suggested the diagnosis of delayed-type HS to tetanus toxoid in a child with recurrent abscesses induced by booster injections of T-containing vaccines [126]. Finally, 1–8% of subjects of all ages receiving HBV vaccines report mild to moderately severe inflam-

331

matory reactions, and Arthus-like reactions are reported in patients receiving booster injections of pneumococcal vaccine. In those cases, immunoallergological studies have not been performed. Aluminium hydroxide, mercurothiolate, and formaldehyde contained in vaccine may induce mild to moderate inflammatory reactions, resolving in a few days. Aluminium hydroxide in vaccines (and in allergenic extracts) also induces subcutaneous nodules. Nodules usually resolve in a few months, but rare cases of persistent nodules have been reported. Diagnosis is based on clinical history. Patch tests with aluminium salts are usually negative [130], probably because most subcutaneous nodules result from non-specific reaction to foreign substance. However, in the study of Bergfors et al. [131], 0.8% of children vaccinated with a new aluminium hydroxidecontaining vaccine reported persistent nodules, and patch tests with aluminium salts were positive in 77% of these children. Generalized Reactions Urticaria, angioedema and unidentified rashes are reported in 5–13% of patients receiving toxoid-containing vaccines. Immunoallergological studies based on immediate-, semi-late-, and latereading skin tests, specific antitoxoid antibody determination (IgM/IgG and IgE) and challenge suggest that most mild to moderate generalized reactions result from non-specific activation of the inflammatory system by high doses of bacterial components, with good tolerance of booster injections of the suspected vaccines [132]. However, urticaria and angioedema resulting from immediate-type or immune-complex-related HS to toxoids have been reported in children [114, 133]. In the study of Ponvert et al. [114], 19 children reported generalized skin reactions to toxoid-containing vaccines. Skin test and CAPRAST results diagnosed immediate HS to diphtheria or tetanus toxoid in 6 children with mild to moderately severe immediate and accelerated ur-

332

ticaria and angioedema to booster injections of the vaccine. In contrast, immediate-, semi-lateand late-reading skin tests were negative in the other children with non-immediate and non-urticarial reactions, and those children tolerated booster sequential injections of the vaccines. However, IgE specific for diphtheria and tetanus toxoids were found in sera of children and adolescents who had tolerated immunizations with DTcontaining vaccines [134]. Since the introduction of highly purified toxoids, anaphylactic reactions to toxoid-containing vaccines have become rare. In a study of 784 DT immunizations and 15,752 DTP (pertussis) immunizations in children, no severe urticaria or anaphylaxis was reported [10]. Jacobs et al. [125] reported only 1 patient with anaphylaxis induced by injection of a tetanus toxoid-containing vaccine, diagnosed allergic to tetanus toxoid with skin tests. In the study of Ponvert et al. [114], 6 children reported severe anaphylactic reactions induced by booster injections of DT-containing vaccines. Immediate responses in skin tests and detection of serum-specific IgE diagnosed immediate HS to tetanus or diphtheria toxoid in 4 children, which is consistent with results of other studies in a few patients with anaphylaxis to booster injections of DT-containing vaccines [135]. Unidentified rashes and arthralgias have been reported in subjects immunized with HBV vaccines, but their frequency is not significantly different in placebo groups [136]. Immediate urticaria and/or Quincke edema, asthma, and anaphylaxis have been reported in a few adults immunized with a Saccharomyces cerevisiae-derived recombinant vaccine. Immunoallergological tests were not performed, except for 1 patient diagnosed allergic to S. cerevisiae, based on immediate response in skin tests and specific IgE determination [137]. In the study of Bakonde et al. [138], children with accelerated urticaria, angioedema, and asthma were diagnosed non-allergic to HBV vaccines. All the children had nega-

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

tive responses in skin tests and tolerated booster injection of the vaccine, except for a child with chronic urticaria. Allergic or allergic-like reactions to pneumococcal vaccines are rare, and include Arthus-like reactions in patients receiving booster injections, AGEP, and anaphylactic or anaphylactoid reactions. Immediate responses in skin tests and selfmade CAP-RAST were positive in a child with severe immediate anaphylaxis to a first injection of 23-valent pneumococcal vaccine [121]. Skin tests and CAP-RAST were negative in 10 vaccinated and non-vaccinated control children, except for 1 non-vaccinated child. These results strongly suggest that immediate responses in skin tests and specific IgE determination have a good diagnostic value in children with severe IgE-dependent reactions to pneumococcal vaccines. They also suggest that occult sensitization to pneumococcal antigens may occur in non-vaccinated children, probably due to carriage or occult infection with Streptococcus pneumoniae. Vaccines, such as MMR vaccines (single or associated), and influenza, yellow fever, and tickborne encephalitis vaccines may contain variable amounts of ovalbumin (from 0–1 ng/ml in MMR to 1–45 g/ml in influenza and yellow fever vaccines), and be responsible for anaphylaxis, including anaphylactic death [116, 139–142] in eggallergic children. Lavi et al. [143] showed that all children with negative skin tests with vaccines tolerated a complete dose of ovalbumin-containing vaccines. In contrast, generalized reactions to vaccine injections performed using a ‘desensitization’ procedure were observed in 12.5% of children with immediate responses in skin tests. Most children with positive skin tests reported severe reactions to egg, whereas most children with negative skin tests reported mild to moderate reactions, such as atopic dermatitis and isolated urticaria and/or angioedema. Anaphylactic reactions have been reported in children immunized with gelatin-containing vaccines, such as MMR vaccines, and in patients

Drug Allergy in Children

immunized with other gelatin-containing vaccines such as Japanese encephalitis virus and varicella vaccines [13, 118–120]. Clinical history of allergy to gelatin-containing foods was found a posteriori in several patients, and gelatin allergy developed after reaction to vaccine in 20–25% of patients. Diagnosis is based on immediate responses in skin tests with gelatin-containing vaccine and gelatin, and on serum-specific anti-gelatin IgE determination. Non-immediate urticaria, angioedema and non-urticarial rashes have also been reported in patients immunized with gelatin-containing vaccines [127–129]. Severe anaphylactoid reactions have been reported in neonates receiving a first injection of BCG. Diagnosis of dextran HS was suggested by high levels of anti-dextran antibody (IgM/IgG) in mother’s serum and cord blood [123]. High levels of anti-dextran antibodies were also found in the serum of an adolescent, 4 weeks after a generalized reaction induced by a booster injection of the BCG vaccine [122]. However, anti-dextran antibodies are found in a high proportion of the population, most of which tolerate dextran-containing vaccines [144]. Conclusion: Prevention of (Suspected) Allergic Reactions to Vaccines Prevention of allergic and pseudoallergic reactions to vaccines is based on judicious consideration of children’s clinical history (i.e. history of ovalbumin, gelatin or mould allergy; chronology, type and severity of reaction to previous vaccine injection), of the characteristics of the suspected vaccine, and of the risk/clinical benefit ratio of the vaccine: (1) If vaccination or booster immunization is not essential (children with high levels of serumspecific IgM/IgG, vaccination not compulsory, high risk of reaction versus low benefit of the vaccine), withholding of injection of the vaccine is advised. (2) If vaccination is compulsory or essential in children with proven or highly suspected allergy,

333

choice of a vaccine that does not contain the responsible substance is advised. If such vaccine does not exist, management of patients depends on the nature and severity of the reaction: in children with large local inflammatory reactions to multivalent vaccines, booster immunization based on sequential (intramuscular) injections of vaccines containing a single or a limited number of vaccinating agent(s), at 7- to 10-day intervals, is usually well tolerated [114]. Intramuscular injection of the vaccines prevents relapse of aluminium hydroxide-induced persistent nodules [145]. In children with IgE-dependent urticaria, angioedema, and anaphylactic reactions, the responsible vaccine should be injected using an appropriate ‘desensitization’ procedure, performed under medical supervision at the hospital. Recommendations for prevention of immediate-type HS reactions in egg-allergic children depends on the ovalbumin amount of the vaccines. Vaccines grown in cultures of chick embryo fibroblasts (MMR vaccines, single or associated) contain a very low ovalbumin amount, and injection of a full-dose of the vaccine is generally well tolerated. However, hospital pediatric supervision is recommended in children with a past history of severe egg anaphylaxis [146]. For the other vaccines (influenza, yellow fever, etc.), cultured on extraembryonic fluid of chicken embryos, that may contain important amounts of egg proteins, immediate-reading skin tests with the vaccines should be performed before vaccination in children with severe IgE-dependent allergic reactions to egg, but they are useless in children with non-threatening reactions and negative skin prick test and/or serum-specific IgE determination to egg. In children with immediate responses in skin tests with the vaccines, immunization should be performed using a ‘desensitization’ procedure in the hospital [142]. Finally, to our knowledge, there is no validated method for prevention of other reactions to vaccines, such as sterile abscesses and anaphylactoid reactions to dextran-containing vaccines.

334

Drugs Used in Anesthesia and Surgery The incidence of IgE-dependent anaphylactic reactions during anesthesia and surgery in the general population is estimated between 1/4,000 and 1/25,000 in France, England, Australia and New Zealand. Based on our experience, the estimated frequency of IgE-dependent anaphylaxis during anesthesia in French children is 1 in 2,100 anesthetic procedures [14]. The most common cause of anaphylaxis in the general population are myorelaxants/NMBA (62%). Other frequent causes are latex (16.5%), antibiotics (4.7–8%), and plasma substitutes (3.6%), gelatins in particular [147]. Although in our experience 60.8 and 27% of 68 children explored for anaphylactic reactions during anesthesia were diagnosed allergic to myorelaxants and latex, respectively [14], it is generally assumed that latex allergy is the most common cause of anaphylaxis in anesthesia in children (;70–75% vs. 10–15% for NMBA) [15]. To our knowledge, there is no specific epidemiologic study of NMBA sensitization and allergy in children. In the general population, atopy is not considered to be a risk factor for sensitization and allergy to NMBA [148]. Diagnosis of NMBAinduced anaphylaxis is based on immediatereading skin tests (pricks and ID) with the suspected drug and other NMBA, because there is a high risk of cross-reactivity between several NMBA. Serum-specific IgE determination may also be useful. Prevention of relapses is based on a rigorous avoidance of NMBA that give positive responses in skin tests and specific IgE determination, and use of other NMBA [147]. An important differential diagnosis of perioperative anaphylaxis is latex allergy, which occurs mainly in multioperated children, especially in children with spina bifida (SB) or myelomeningocele [17, 149–155]. Apart from surgery and anesthesia, latex sensitization and allergy in children may be caused by (occult) exposure to latex-containing objects such as nipples and pacifiers, toys, etc. [156–158]. Diagnosis of latex al-

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

lergy is based on positive responses in prick tests and/or determination of latex-specific IgE in serum of the patients. Preoperative allergological screening, based on skin tests and specific IgE determination, should be performed in all children with suspected latex allergy and in highrisk children (atopic children exposed to latex, children allergic to exotic fruits, and multioper-

ated children, especially SB and myelomeningocele children). In latex-allergic children, prevention of potentially severe reactions is based on a rigorous eviction of latex. Latex eviction is also recommended in children with SB and myelomeningocele in order to avoid the risk of sensitization [159].

References 1 Anderson JA: Allergic reactions to drugs and biological agents. JAMA 1992;268:2845–2857. 2 Haddi E, Charpin M, Tafforeau M, Kulling G, Lanteaume A, Kleisbauer JP, Vervloet D: Atopy and systemic reactions to drugs. Allergy 1990;45:236– 239. 3 Ibia EO, Schwartz RH, Widermann BL: Antibiotic rashes in children: a survey of private practice setting. Arch Dermatol 2000;136:849–854. 4 Lin RY: A perspective on penicillin allergy. Arch Intern Med 1992;152:930– 937. 5 Shepherd GM: Allergy to
-lactam antibiotics. Immunol Allergy Clin North Am 1991;11:611–633. 6 Lesko SM, Mitchell AA: The safety of acetaminophen and ibuprofen among children younger than two years old. Pediatrics 1999;104:e39–e44. 7 Sanchez-Borges M, Capriles-Behrens E, Caballero-Fonseca F: Hypersensitivity to non-steroidal anti-inflammatory drugs in childhood. Pediatr Allergy Immunol 2004;15:376–380. 8 Titchen T, Cranswick N, Beggs S: Adverse drug reactions to non-steroidal anti-inflammatory drugs, COX-2 inhibitors and paracetamol in a paediatric hospital. Br J Clin Pharmacol 2005; 59:718–723. 9 Bernstein DI, Smith VE, Schiff VH, Rathfon HM, Boscia JA: Comparison of acellular pertussis vaccine with whole cell vaccine in children 15–18 months and 4–6 years of age. Pediatr Infect Dis 1993;12:131–135. 10 Cody CI, Baraff U, Cherry JD: Nature and rates of adverse reactions associated with DTP and DT immunizations in infants and children. Pediatrics 1981;68:650–660.

Drug Allergy in Children

11 Decker MD, Edwards KM, Steinhoff MC, Rennels MB, Pichichero ME, Englund JA, Anderson EL, Deloria MA, Reed GF: Comparison of 13 acellular pertussis vaccines: adverse reactions. Pediatrics 1995;96(suppl):557–566. 12 Long SS, Deforest A, Smith DG: Longitudinal study of adverse reactions following diphtheria-tetanus-pertussis vaccine in infancy. Pediatrics 1990; 294–302. 13 Sissman NJ: Allergic reactions to MMR vaccines. Pediatrics 1992;88:168–169. 14 Karila C, Brunet-Langot D, Labbez F, Jacqmarcq O, Ponvert C, Paupe J, Scheinmann P, de Blic J: Anaphylaxis during anesthesia: results of a 12-year survey at a French pediatric center. Allergy 2005;60:828–834. 15 Murat I: Anaphylactic reactions during paediatric anesthesia: results of the survey of the French Society of Paediatric Anaesthesia (ADARPEF), 1991– 1992. Paediatr Anaesth 1993;3:339– 343. 16 Bernardini R, Novembre E, Ingargiola A, Veltroni M, Mugnaini L, Cianferoni A, Lombardi E, Vierucci A: Prevalence and risk factors of latex sensitization in an unselected pediatric population. J Allergy Clin Immunol 1998;101:621– 625. 17 Valentine JP, Kurinczuk JJ, Loh RVS, Chauvel PJ: Latex allergy in an Australian population of children and adolescents with spinal dysfunction. Med J Aust 1999;170:15–18. 18 Bigby M: Rates of cutaneous reactions to drugs. Arch Dermatol 2001; 137:765– 770. 19 Buajordet I, Wesenberg F, Brørs O, Langslet A: Adverse drug events in children during hospitalization and hospital discharge in a Norwegian Uni-

20

21

22

23

24

25

26

27

28

versity Hospital. Acta Paediatr 2002;91: 88–94. Forman R, Koren G, Shear NH: Erythema multiforme, Stevens-Johnson syndrome and toxic epidermal necrolysis in children: a review of 10 years’ experience. Drug Saf 2002;25:965–972. Gonzalez-Martin A, Caroca CM, Paris E: Adverse drug reactions in hospitalized pediatric patients: a prospective study. Int J Clin Pharmacol Ther 1998; 36:530–533. Martinez-Mir I, Garcia-Lopez M, Palop V, Ferrer JM, Rubio E, Morales-Olivas FJ: A prospective study of adverse drug reactions in hospitalized children. Br J Clin Pharmacol 1999;47:681–688. Menniti-Ippolito F, Paschetti R, Dacas R, Giaquinto C, Cantarutti L: Active monitoring of adverse drug reactions. Lancet 2000;355:1613–1614. Weiss J, Krebs S, Hoffmann C, Werner U, Neubert A, Brune K, Rascher W: Survey of adverse drug reactions on a pediatric ward: a strategy for early and detailed detection. Pediatrics 2002;110: 254–257. Beltrani VS: Cutaneous manifestations of adverse drug reactions. Immunol Allergy Clin North Am 1998;18:867– 895. Gurwitz JH, Avorn J: The ambiguous relation between aging and adverse drug reactions. Ann Intern Med 1991; 114:956–966. Moss RB, Babin S, Hsu YP, BlessingMoore J, Lewiston NJ: Allergy to semisynthetic penicillins in cystic fibrosis. J Pediatr 1984;104:460–466. Moller NE, Ahlstedt S, Skov PS, Norn S: Allergological examination of cystic fibrosis patients with skin reactions during carbenicillin treatment. Allergy 1980;35:135–138.

335

29 Pleasants RA, Walker TR, Samuelson WM: Allergic reactions to parenteral
-lactam antibiotics in patients with cystic fibrosis. Chest 1994;106:1224– 1228. 30 Hebert AA, Sigman ES, Levy ML: Serum sickness-like reactions from cefaclor in children. J Am Acad Dermatol 1991;25:805–808. 31 Shin HT, Chang MW: Drug eruptions in children. Curr Probl Pediatr 2001;31: 207–234. 32 Grasset L, Guy C, Ollagnier M: Cyclins and acne: pay attention to adverse drug reactions! A recent literature review. Rev Méd Interne 2003;24:305–316. 33 Carroll MC, Yueng-Yue KA, Esterly NB, Drolet BA: Drug-induced hypersensitivity syndrome in pediatric patients. Pediatrics 2001;108:485–492. 34 Hassani A, Ponvert C, Karila C, Le Bourgeois M, de Blic J, Scheinmann P: Hypersensitivity to non-steroidal antiinflammatory drugs in children: results of a study of 164 cases (abstract). Proc XXVth Congress of EAACI, Vienna, June 10–14, 2006, p 188. 35 Schubert B, Grosse Perdekamp MT, Pfeuffer P, Raith P, Bröcker EB, Trautmann A: Non-steroidal anti-inflammatory drug hypersensibility: fable or reality? Eur J Dermatol 2005;15:164–167. 36 Clarkson A, Choonara I: Surveillance for fatal suspected adverse drug reactions in the UK. Arch Dis Child 2002; 87:462–467. 37 Dibs SD, Baker MD: Anaphylaxis in children: a 5-year experience. Pediatrics 1997;99:e71–e76. 38 Mehl A, Wahn U, Niggemann B: Anaphylactic reactions in children: a questionnaire-based survey in Germany. Allergy 2005;60:1440–1445. 39 Demoly P, Kropf R, Bircher A, Pichler WJ: Drug hypersensitivity questionnaire. Allergy 1999;54:999–1003. 40 Brockow K, Romano A, Blanca M, Ring J, Pichler WJ, Demoly P: General considerations for skin test procedures in the diagnosis of drug hypersensitivity. Allergy 2002;57:45–51. 41 Carder KR: Hypersensitivity reactions in neonates and infants. Dermatol Ther 2005;18:160–175. 42 Graff-Lonnevig V, Hedlin G, Linfors A: Penicillin allergy: a rare paediatric condition. Arch Dis Child 1988;63: 1342–1346. 43 Ramsey M, Reacher M, O’Flynn CO, Butterey R, Hadden C, Cohen B, Knowles W, Wreghitt T, Brown D:

336

44

45

46

47

48

49

50

51

52

53

Causes of morbilliform rash in a highly immunized English population. Arch Dis Child 2002;87:202–206. Kidon MI, Kang MW, Chin CW, Hoon LS, See Y, Goh H, Poh Lin JT, Chay OM: Early presentation with angioedema and urticaria in cross-reactive hypersensitivity to non-steroidal anti-inflammatory drugs among young, Asian, atopic children. Pediatrics 2005; 16:e675–e680. Martin-Munoz F, Moreno-Ancillo A, Dominguez-Noche C, Diaz-Pena JM, Garcia-Ara C, Boyano T, Ojeda JA: Evaluation of drug-related hypersensitivity reactions in children. Invest Allergol Clin Immunol 1999;9:172– 177. Ponvert C, Le Clainche L, de Blic J, Le Bourgeois M, Scheinmann P, Paupe J: Allergy to
-lactam antibiotics in children. Pediatrics 1999;104:e45–e54. Cousin F, Phillips K, Favier B, Bienvenu J, Nicolas JF: Drug-induced urticaria. Eur J Dermatol 2001;11:181–187. Mortureux P, Leauté-Labrèze C, Legrain-Lifermann V, Lamireau T, Sarlangue J, Taieb A: Acute urticaria in infancy and early childhood: a prospective study. Arch Dermatol 1998; 134:319–323. Morelli JG, Tay YK, Rogers M, Halbert A, Krafchik B, Weston WL: Fixed drug eruptions in children. J Pediatr 1999; 134:365–367. Lam NS, Yang YH, Wang LC, Lin YT, Chiang BL: Clinical characteristics of childhood erythema multiforme, Stevens-Johnson syndrome and toxic epidermal necrolysis in Taiwanese children. J Microbiol Immunol Infect 2004; 37:366–370. Salman SM, Kibbi AG: Vascular reactions in children. Clin Dermatol 2002; 20:11–15. Naisbitt DJ, Farrell J, Wong G, Depta JP, Dodd CC, Hopkins JE, Gibney CA, Chadwick DW, Pichler WJ, Pirmohamed M, Park BK: Characterization of drug-specific T cells in lamotrigine hypersensitivity. J Allergy Clin Immunol 2003;111:1393–1403. Naisbitt DJ, Britschgi M, Wong G, Farrel J, Depta JP, Chadwick DW, Pichler WJ, Pirmihamed M, Park BK: Hypersensitivity reactions to carbamazepine: characterization of the specificity, phenotype, and cytokine profile of drugspecific T-cell clones. Mol Pharmacol 2003;63:732–741.

54 Leeder JS: Mechanisms of idiosyncratic hypersensitivity reactions to antiepileptic drugs. Epilepsia 1998;39:S8–S16. 55 Suzuki Y, Inagi R, Aono T, Yamanishi K, Shiohara T: Human herpes virus 6 infection as a risk factor for the development of severe drug-induced hypersensitivity syndrome. Arch Dermatol 1998;134:1108–1112. 56 Roujeau C, Bioulac-Sage P, Bourseau C, Guillaume JC, Bernard P, Lok C, Plantin P, Claudy A, Delavierre C, Vaillant L: Acute generalized exanthematous pustulosis. Analysis of 63 cases. Arch Dermatol 1991; 127:1333– 1338. 57 Kearns GL, Wheeler JG, Childress SH, Letzig LG: Serum sickness-like reactions to cefaclor: role of hepatic metabolism and individual susceptibility. J Pediatr 1994;125:805–811. 58 Levine LR: Quantitative comparison of adverse reactions to cefaclor vs. amoxicillin in a surveillance study. Pediatr Infect Dis 1985;4:358–361. 59 Parshuram CS, Phillips RJ: Retrospective review of antibiotic-associated serum sickness in children presenting to a paediatric emergency department. Med J Aust 1998; 169:116. 60 Vial T, Pont J, Pham E, Rabilloud M, Descotes J: Cefaclor-associated serum sickness-like disease: eight cases and review of the literature. Ann Pharmacother 1992;26:910–914. 61 Pichichero ME, Pichichero DM: Diagnosis of penicillin, amoxicillin and cephalosporin allergy: reliability of examination assessed by skin testing and oral challenge. J Pediatr 1998;132: 137–143. 62 Novembre E, Cianferoni A, Bernardini R, Mugnaini L, Caffarelli C, Cavagni G, Giovane A, Vierucci A: Anaphylaxis in children: clinical and allergologic features. Pediatrics 1998;101:e8–e16. 63 Astanaskovic-Marjovic M, CirkovicVelikovic T, Gavrovic-Jankulovic M, Vuckovic O, Nestorovic B: Immediate allergic reactions to cephalosporins and penicillins and their cross-reactivity in children. Pediatr Allergy Immunol 2005;16:341–347. 64 Erffmeyer JE: Reactions to antibiotics. Immunol Allergy Clin North Am 1992; 12:633–648. 65 Mendelson LM, Ressler C, Rosen JP, Selcow JE: Routine selective penicillin allergy testing in children and adolescents: study of sensitization. J Allergy Clin Immunol 1984;73:76–81. 1

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

66 Romano A, Quaratino D, Papa G, Di Fonso M, Venuti A: Aminopenicillin allergy. Arch Dis Child 1997;76:513– 517. 67 Léauté-Labrèze C, Lamireau T, Chawki D, Maleville J, Taieb A: Diagnosis, classification of erythema multiforme and Stevens-Johnson syndrome. Arch Dis Child 2000;83:347–352. 68 Blanca M, Mayorga C, Torres MJ, Reche M, Moya MC, Rodriguez JL, Romano A, Juarez C: Clinical evaluation of Pharmacia CAP System FEIA amoxicilloyl and benzylpenicilloyl in patients with penicillin allergy. Allergy 2001;56:862– 870. 69 Lebel B, Messaad D, Kvedariene K, Rougier M, Bousquet J, Demoly P: Cysteinyl-leukotriene release test in the diagnosis of immediate drug reactions. Allergy 2001;56:688–692. 70 Luque I, Leyva L, Torres JM, Rosal M, Mayorga C, Segura JM, Blanca M, Juarez C: In vitro T-cell responses to
lactam drugs in immediate and nonimmediate allergic reactions. Allergy 2001;56:611–618. 71 Sanz ML, Garcia BE, Prieto I, Tabar A, Oehling A: Specific IgE determination in the diagnosis of
-lactam allergy. Invest Allergol Clin Immunol 1996; 6: 89–93. 72 Saurat JH, Ponvert C, Burtin C, Soubrane C, Lebel B, Beucher F, Paupe J: Lymphocyte transformation, leucocyte migration, specific IgE, IgM and IgG, before, during and after penicillin treatment without adverse reaction: a follow-up study. Acta Allergol 1976; 31: 1–17. 73 Cetinkaya F, Cag Y: Penicillin sensitivity among children without a positive history for penicillin allergy. Pediatr Allergy Immunol 2005;15:278–280. 74 Lopez-Serrano MC, Caballero MT, Barranco P, Martinez-Alzamora F: Booster responses in the study of allergic reactions to
-lactam antibiotics. J Invest Allergol Clin Immunol 1996;6:30–35. 75 Parker PJ, Parinello JT, Condemi JJ, Rosenfeld SI: Penicillin resensitization among hospitalized patients. J Allergy Clin Immunol 1991;88:213–217. 76 Ponvert C, Weilenmann C, Wassenberg J, Walecki P, Le Bourgeois M, De Blic J, Scheinmann P: Allergy to
-lactam antibiotics in children: a prospective follow-up study in retreated children after negative responses in skin and challenge tests. Allergy 2007;62:42–46.

Drug Allergy in Children

77 Solensky R, Earl MS, Gruchalla RS: Lack of penicillin resensitization in patients with a history of penicillin allergy after receiving repeated penicillin courses. Arch Intern Med 2002;162: 822–826. 78 Ponvert C, Chedevergne F, Le Bourgeois M, de Blic J, Paupe J, Scheinmann P: Allergy to
-lactams in children: diagnosis of non-immediate hypersensitivity by means of skin tests and oral challenge. Rev Fr Allergol Immunol Clin 2001;41:544–554. 79 Ponvert C, Le Bourgeois M, Karila C, de Blic J, Scheinmann P: Allergy to
-lactam antibiotics in child: diagnosis of non-immediate hypersensitivity by means of intradermal and patch tests and challenge. Rev Fr Allergol Immunol Clin 2004;44:379–381. 80 Romano A, Quaratino D, Di Fonso M, Papa G, Venuti A, Gasbarrini G: A diagnostic protocol for evaluating nonimmediate reactions to aminopenicillins. J Allergy Clin Immunol 1999;103: 1186–1190. 81 Novalbos A, Sastre J, Cuesta J, De Las Heras M, Lluch-Bernal M, Bombin C, Quirce S: Lack of allergic cross-reactivity to cephalosporins among patients allergic to penicillins. Clin Exp Allergy 2001;31:438–443. 82 Terrados S, Blanca M, Garcia J, Vega J, Torres MJ, Carmona MJ, Miranda A, Moya M, Juarez C, Fernandez J: Nonimmediate reactions to
-lactams: prevalence and role of the different penicillins. Allergy 1995;50:563–567. 83 Kearns GL, Wheeler JG, Rieder MJ, Reid J: Serum sickness-like reactions to cefaclor: lack of in vitro cross-reactivity with loracarbacef. Clin Pharmacol Ther 1998;63:686–693. 84 Ponvert C, Scheinmann P: Antibiotic allergy and pseudo-allergy in child. Rev Fr Allergol Immunol Clin 2003; 43: 385–392. 85 Benahmed S, Scaramuzza C, Messaad D, Sahla A, Demoly P: The accuracy of the diagnosis of macrolide antibiotic hypersensitivity: results of a singleblinded trial. Allergy 2004;59:1130– 1133. 86 Campi P, Pichler WJ: Quinolone hypersensitivity. Curr Opin Allergy Clin Immunol 2003;3:275–281. 87 Slatore CG, Tilles GA: Sulfonamide hypersensitivity. Immunol Allergy Clin North Am 2004;24:477–490.

88 Barbaud A, Reichert-Penetrat S, Trechot P, Jacquin-Petit MA, Ehlinger A, Noirez V, Faure GC, Schmutz JL, Béné MC: The use of skin testing in the investigation of cutaneous adverse reactions. Br J Derrmatol 1998; 139: 49–58. 89 Paupe J, Le Bourgeois M, Bidat E: Unjustly excluded drugs: challenge tests. Rev Fr Allergol 1996;36:155–161. 90 Demoly P, Benahmed S, Valembois M, Sahla H, Messaad D, Godard P, Michel FB, Bousquet J: L’allergie aux macrolides: revue de la littérature. Presse Méd 2000;29:321–326. 91 Manfredi M, Severino M, Testi S, Macchia D, Ermini G, Pichler WJ, Campi P: Detection of specific IgE to quinolones. J Allergy 2004;113:155– 160. 92 Buergin S, Scherer K, Hausermann P, Bircher AJ: Immediate hypersensitivity to rifampicin in three patients: diagnostic procedures and induction of clinical tolerance. Int Arch Allergy Immunol 2006;140:20–26. 93 Kamada M, Twarog FJ, Leung DYM: A comparison of multiple versus single antibiotic sensitivity in pediatric patients. Allergy Proc 1991;12:347–350. 94 Khoury L, Warrington R: The multiple drug allergy syndrome: a matched-control retrospective study in patients allergic to penicillin. J Allergy Clin Immunol 1996;98:462– 464. 95 Sullivan TJ, Ong RC, Gilliam LK: Studies of the multiple drug allergy syndrome. J Allergy Clin Immunol 1989;83:270. 96 Botey J, Ibero M, Malet A, Marin A, Eseverri JL: Aspirin-induced recurrent urticaria and recurrent angioedema in non-atopic children. Ann Allergy 1984;53:265–267. 97 Settipane RA: Aspirin and allergic diseases: a review. Am J Med 1983;74: 102–109. 98 Stevenson DD: Adverse reactions to non-steroidal anti-inflammatory drugs. Immunol Allergy Clin North Am 1998;18:773–798. 99 Speer F, Denison TR, Baptist JE: Aspirin allergy. Ann Allergy 1981;46:123– 126. 100 Vieluf D, Vieluf I, Brockow B, Abeck D: Positive intracutaneous and oral provocation tests in a child with an anaphylactoid reaction to paracetamol (abstract). J Allergy Clin Immunol 1996;97:356.

337

101 Daxun Z, Becker WM, Schulz KH, Schlaak M: Sensitivity to aspirin: a new serological diagnostic method. J Invest Allergol Clin Immunol 1993; 3: 72–78. 102 Gamboa PM, Sanz ML, Caballero MR, Antépara I, Urrutia I, Jauregui I, Gonzalez G, Dieguez I, De Weck AL: Use of CD63 expression as a marker of in vitro basophil activation and leukotriene determination in metamizol-allergic patients. Allergy 2003;58:312–317. 103 Zhu DX, Zhao L, Mo L, Li HL: Drug allergy: identification and characterization of IgE reactivities to aspirin and related compounds. J Invest Allergol Clin Immunol 1997;7:160–168. 104 Boussetta K, Ponvert C, Karila C, Le Bourgeois M, de Blic J, Scheinmann P: Hypersensitivity reactions to paracetamol in children. Allergy 2005;60: 1174–1177. 105 Debley JS, Carter ER, Gibson RL, Rosenfeld M, Redding GJ: The prevalence of ibuprofen-sensitive asthma in children: a randomized controlled bronchoprovocation challenge study. J Pediatr 2005;147:233–238. 106 Cirstea M, Cirje M, Suhaciu G: The diagnostic value of intradermal skin tests with penicilloyl-dextran and aspirin-polylysine. Physiologie 1986; 23:237–243. 107 Del Pozzo MD, Lobera T, Blasco A: Selective hypersensitivity to diclofenac. Allergy 2000;55:412–413. 108 Sanz ML, Gamboa P, De Weck AL: A new combined test with flow cytometric basophil activation and determination of sulfidoleukotrienes is useful for in vitro diagnosis of hypersensitivity to aspirin and other nonsteroidal anti-inflammatory drugs. Int Arch Allergy Immunol 2005;136: 58–72. 109 Settipane RA, Stevenson DD: Crosssensitivity with acetaminophen in aspirin-sensitive subjects with asthma. J Allergy Clin Immunol 1989;84: 26–33. 110 Gutgesell C, Fuchs T: Azopropazone in aspirin intolerance. Allergy 1999; 54:897–898. 111 Karakaya G, Kalyoncu AF: Old and forgotten but cheap and safe alternatives for analgesic-intolerant patients: shall the allergists remember that they are still in the market? (Abstract) Proc XXIIth Congress of EAACI, Paris, June 7–11, 2003, pp 349–350.

338

112 Bircher AJ: Drug-induced urticaria and angioedema caused by non-IgEmediated pathomechanisms. Eur J Dermatol 1999;9:657–663. 113 Ponvert C, Scheinmann P: Vaccine allergy and pseudo-allergy. Eur J Dermatol 2003;13:10–15. 114 Ponvert C, Scheinmann P, Karila C, Bakonde VB, Le Bourgeois M, de Blic J: Allergy to multivalent vaccines in children: a study of 30 cases using immediate, semi-late and late responses in skin tests, specific antibody determinations, and challenge with monovalent and bivalent vaccines. Rev Fr Allergol Immunol Clin 2001;41:701–711. 115 Aukrust L, Almeland TL, Refsum D, Aas K: Severe hypersensitivity or intolerance reactions to measles vaccine in six children. Allergy 1980;35:581. 116 Baxter DN: Measles immunization in children with a history of egg allergy. Vaccine 1996;14:131–134. 117 Vourga V, Tapratzi P, Panayatopoulou K: Measles-mumps-rubella vaccine in egg-allergic children. Allergy 1997;52: 96. 118 Kelso JM, Jones RT, Yunginger JW: Anaphylaxis to measles, mumps and rubella vaccine mediated by IgE to gelatin. J Allergy Clin Immunol 1993; 91:867–872. 119 Kumagai T, Yamanaka T, Wataya Y, Umetsu A, Kawamura N, Ikeda K, Furukawa H, Kimura K, Chiba S, Saito S, Sugawara N, Kurimoto F, Sa kaguchi M, Inouye S: Gelatin-specific humoral and cellular immune responses in children with immediate- and non-immediate-type reactions to live measles, mumps, rubella and varicella vaccines. J Allergy Clin Immunol 1997;30:1430–1435. 120 Sakaguchi M, Nakayama T, Inouye S: Food allergy to gelatin in children with systemic immediate-type reactions, including anaphylaxis, to vaccines. J Allergy Clin Immunol 1996; 98:1058–1061. 121 Ponvert C, Ardelean-Jaby D, ColinGorski AM, Soufflet B, Hamberger C, de Blic J, Scheinmann P: Anaphylaxis to the 23-valent pneumococcal vaccine in child: a case-control study based on immediate responses in skin tests and specific IgE determination. Vaccine 2001;19:4588–4591.

122 Fauquert JL, Tridon A, Labbé A, Perrier C: Anaphylactic shock after injection of intradermal vaccine BCG. Rev Fr Allergol Immunol Clin 2001;41: 412–414. 123 Ponnighaus JM, Fine PEM, Moreno C: Hypersensitivity to dextran in BCG vaccine. Lancet 1991;337:1039. 124 Bohlke K, Davis RL, Marcy SM, Braun MM, DeStefano F, Black SB, Mullooly JP, Thompson RS, for the Vaccine Safety Datalink Team: Risk of anaphylaxis after vaccination of children and adolescents. Pediatrics 2003;112: 815–820. 125 Jacobs RL, Lowe RS, Lanier BQ: Adverse reactions to tetanus toxoid. JAMA 1982;247:40–42. 126 Church JA, Richard W: Recurrent abscess formation following DTP immunizations: association with hypersensitivity to tetanus toxoid. Pediatrics 1985;75:899–900. 127 Sakaguchi M, Miyazawa H, Inouye S: Sensitization to gelatin in children with systemic non-immediate-type reactions to varicella vaccines. Ann Allergy Asthma Immunol 2000;84: 341–344. 128 Sakaguchi M, Miyazawa H, Inouye S: Specific IgE and IgG to gelatin in children with systemic cutaneous reactions to Japanese encephalitis vaccines. Allergy 2001;56:536–539. 129 Taniguchi K, Fujisawa T, Ihara T, Kamiya H: Gelatin-induced T-cell activation in children with non-anaphylactic-type reactions to vaccines containing gelatin. J Allergy Clin Immunol 1998;102:1028–1032. 130 Netterlid E, Bruze M, Hindsén M, Isaksson M, Olin P: Persistent itching nodules after the fourth dose of diphtheria-tetanus toxoid vaccines without evidence of delayed hypersensitivity to aluminium. Vaccine 2004;22: 3698–3706. 131 Bergfors E, Trollfors B, Inerot A: Unexpected high incidence of persistent itching nodules and delayed hypersensitivity to aluminium in children after the use of adsorbed vaccines from a single manufacturer. Vaccine 2003;22:64–69. 132 Gold M, Goodwin H, Botham S, Burgess M, Nash M, Kempe A: Revaccination of 422 children with a past history of an adverse reaction in a special immunization service. Arch Dis Child 2000;83:128–131.

Ponvert ⴢ Hadj-Rabia ⴢ Scheinmann

133 Lewis K, Jordan SC, Cherry JD, Sakai RS: Petechiae and urticaria after DTP vaccination: detection of circulating immune complexes containing vaccine-specific antigens. J Pediatr 1986; 109:1009–1012. 134 Dannemann A, Van Ree R, Kulig M, Bergmann RL, Bauer P, Forster J, Guggenmoos-Holzmann I, Aalberse RC, Wahn U: Specific IgE and IgG4 immune responses to tetanus and diphtheria toxoid in atopic and nonatopic children during the first two years of life. Int Arch Allergy Immunol 1996;111:262–267. 135 Skov PS, Pelck I, Ebbesen F, Poulsen LK: Hypersensitivity to the diphtheria component in the Di-Te-Pol vaccine: a type I allergic reaction demonstrated by basophil histamine release. Pediatr Allergy Immunol 1997;8:156–158. 136 Szmuness W, Stevens CE, Harley EJ, Zang EA, Oleszko WR, William DC, Sadovsky R, Morrison JM, Kellner A: Hepatitis B vaccine: demonstration of efficacy in a controlled clinical trial in a high-risk population in the United States. N Engl J Med 1980;303:833– 841. 137 Brightman CA, Scadding GK, Dumbreck LA, Latchman Y, Brostoff J: Yeast-derived hepatitis B vaccine and yeast hypersensitivity. Lancet 1989;22:903. 138 Bakonde VB, Ponvert C, Le Clainche L, Brunet D, Scheinmann P, Paupe J: Reactions to hepatitis B vaccines in children: results of a study of 4 cases. Rev Fr Allergol 1998;38:315–318. 139 Beck S, Williams LW, Shirrell MA, Burks AW: Egg hypersensitivity and measles-mumps-rubella vaccine administration. Pediatrics 1991;88:913– 917. 140 Fasano MB, Wood RA, Cooke SK, Sampson HA: Egg hypersensitivity and adverse reactions to measles, mumps and rubella vaccine. J Pediatr 1992;120:878–881.

141 Herman JJ, Radin R, Schneiderman R: Allergic reactions to measles (rubeola) vaccine in patients hypersensitive to egg protein. J Pediatr 1983;102: 196–199. 142 Zeiger RS: Current issues with influenza vaccination in egg allergy. J Allergy Clin Immunol 2002;110:834– 840. 143 Lavi S, Zimmerman B, Koren G, Gold R: Administration of measles, mumps and rubella virus vaccine (live) to egg-allergic children. JAMA 1990; 263:269–271. 144 Hedin H, Richter W, Ring J: Dextraninduced anaphylactoid reactions in man. Role of dextran-reactive antibodies. Int Arch Allergy Appl Immunol 1976;52:145–159. 145 Sjølin-Frederiksen M, Tofte H: Immunisation with aluminium-containing vaccine of a child with itching nodule following previous vaccination. Vaccine 2004;23:1–2. 146 Khakoo GA, Lack G: Guidelines for measles vaccination. Clin Exp Allergy 2000;30:288–293. 147 Mertes PM, Laxenaire MC, Lienhart A, and the Working Group for the SFAR and Aberer W, Ring J, Pichler WJ, Demoly P, for ENDA and the EAACI Interest Group on Drug Hypersensitivity: Reducing the risk of anaphylaxis during anaesthesia: guidelines for clinical practice. J Invest Allergol Clin Immunol 2005; 15: 91–109. 148 Cremer R, Chen Z, Bläker F, Baur X: Pattern of IgE antibodies against latex allergens in children with spina bifida. Eur J Pediatr Surg 1999;9:49–50. 149 Cremer R, Hoppe A, Korsch E, KleineDiepenbruck U, Bläker F: Natural rubber latex allergy: prevalence and risk factors in patients with spina bifida compared with atopic children and controls. Eur J Pediatr 1998;157: 13–16.

150 Bernardini R, Novembre E, Lombardi E, Mezzetti P, Cianferoni A, Danti DA, Mercurella A, Vierucci A: Risk factors for latex allergy in patients with spina bifida and latex sensitization. Clin Exp Allergy 1999;29:681–686. 151 Kelly KJ, Pearson ML, Kurup VP, Havens PL, Byrd RS, Setlock MA, Butler JC, Slater JE, Grammer LC, Resnick A: A cluster of anaphylactic reaction in children with spina bifida during general anesthesia: epidemiologic features, risk factors and latex hypersensitivity. J Allergy Clin Immunol 1994; 94:53–61. 152 Mazon A, Nieto A, Linana JJ, Montoro J, Estornell F, Garcia-Ibarra F: Latex sensitization in children with spina bifida: follow-up comparative study after two years. Ann Allergy Asthma Immunol 2000;84:207–210. 153 Porri F, Pradal M, Lemiere C, Birnbaum J, Mege JL, Lanteaume A, Charpin D, Vervloet D, Camboulives J: Association between latex and repeated latex exposure in children. Anesthesiology 1997;86:599–602. 154 Szépfaluzi Z, Seidl R, Benert G, Dietrich W, Spitzauer S, Urbanek R: Latex sensitization in spina bifida appears disease-associated. J Pediatr 1999; 134:344–348. 155 Theissen U, Theissen JL, Mertes N, Brehler R: IgE-mediated hypersensitivity to latex in childhood. Allergy 1997;52:665–669. 156 Chabane H, Leynadier F: Mise au point. Allergy to latex: don’t forget allergic patients in short pants. Rev Fr Allergol Immunol Clin 2000;40:548– 551. 157 Makinen-Kiljunen S, Sorva R, Juntunen-Backman K: Latex dummies as allergens. Lancet 1992;339:1608– 1609. 158 Venuta A, Bertolani P, Pepe P, Francomano M, Piovano P, Ferrari P: Do pacifiers cause latex allergy? Allergy 1999;54:1007. 159 Cremer R, Kleine-Dippenbruck H, Hoppe A, Bläker F: Latex allergy in spina bifida patients: prevention by primary prophylaxis. Allergy 1998; 53:709–711.

Dr. Claude Ponvert René Descartes-Paris V University, Paediatrics Department Pulmonology, Allergology & Dermatology Unit, Sick Children’s Hospital 149, rue de Sèvres FR–75015 Paris (France) Tel. +33 1 4449 4842, Fax +33 1 4273 0211 E-Mail [email protected]

Drug Allergy in Children

339

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 340–349

Hypersensitivity to Aspirin and Other NSAIDs: Mechanisms, Clinical Presentation and Management A. Szczeklik
E. Niżankowska-Mogilnicka
M. Sanak Department of Medicine, Jagiellonian University School of Medicine, Kraków, Poland

Abstract In about 10% of adult asthmatics and 20–40% of patients with chronic idiopathic urticaria, aspirin can precipitate or aggravate symptoms of the disease. These hypersensitivity reactions are the markers of two distinct clinical syndromes, which develop according to a characteristic pattern and follow the specific clinical course. The first, called aspirin-induced asthma, is characterized by a chronic eosinophilic rhinosinusitis with nasal polyposis and persistent, often severe, asthma. There are no in vitro tests, and diagnosis can be established only by provocation challenges with aspirin. Acute hypersensitivity reactions to aspirin are elicited via inhibition of cyclooxygenase-1 by NSAIDs. These two syndromes, cared for by different specialists, are characterized at the biochemical level by profound alterations in eicosanoid biosynthesis, best exemplified by cysteinyl leukotriene overproduction. They should be clearly differentiated from other adverse reactions to NSAIDs with allergic background. Copyright © 2007 S. Karger AG, Basel

from asphyxia due to aspirin was described. The association of aspirin sensitivity, asthma and nasal polyps was reported by Widal et al. [2] in 1922, and the syndrome was popularized by Samter and Beers [3], who presented a perspective description of its clinical course. In the 1970s the link between precipitation of asthmatic attacks and inhibition of cyclooxygenase (COX) by aspirin and other NSAIDs was described [4]. In the following three decades other alterations in arachidonic acid metabolism were discovered, and found recently to be common also in aspirin-induced angioedema/urticaria. Adverse reactions to aspirin and NSAIDs may have different clinical symptoms and different pathogenesis. The two most common presentations of aspirin hypersensitivity are bronchial asthma and urticaria/angioedema.

History and Types of Reactions Aspirin-Induced Asthma

Aspirin was introduced into medicine in 1899, and already 3 years later it was implicated as the cause of an anaphylactic reaction. Hirschberg [1] presented the first case of a transient acute angioedema/urticaria, occurring shortly after ingestion of aspirin. The acute bronchospasm was first reported in 1919, and a year later the first death

Definition This term refers to a distinct clinical syndrome of intractable inflammation in both the upper and lower respiratory tract, which is characterized by chronic eosinophilic sinusitis with nasal polyposis and asthma. Aspirin and other NSAIDs that

inhibit COX-1 exacerbate this condition. At the biochemical level, profound alterations in arachidonic acid metabolism are characteristic (fig. 1). The disease runs a protracted course, even if COX-1 inhibitors are avoided, and at least half of the patients require systemic corticosteroids to control their sinusitis and asthma. Prevalence In the general population the prevalence of aspirin hypersensitivity ranges from 0.5 to 2.5% [5, 6]. In three large questionnaire surveys, carried out in Finland, Poland and Australia, aspirin hypersensitivity was reported in 8.8, 4.3 and 11% of asthmatics, respectively [for references, see 7]. The incidence of aspirin-induced reactions in patients with asthma referred to allergy and respiratory centers is much higher and reaches 21% [8]. In the population of patients with chronic sinusitis and nasal polyposis the prevalence of aspirin hypersensitivity is even more elevated, reaching 30–40%. Aspirin hypersensitivity is more prevalent in females, who outnumber men by about 2.5 to 1 [9], and is unusual in children. In an American study, bronchospastic reactions to ibuprofen were found in 2% of children exposed to oral challenge with this drug [10]. Aspirin-induced asthma (AIA) seems to be underdiagnosed worldwide. The reason for underreporting the hypersensitivity may include deliberate avoidance of NSAIDs by asthmatics aware of the risk of adverse reactions, lack of recognition by patients of mild NSAID-induced reactions and lack of routine aspirin challenge testing of asthmatic patients. In 1–6% of cases there is a family history of aspirin hypersensitivity [11]. Natural History and Presentation AIA, more frequently than other types of asthma, is preceded by chronic, intractable rhinitis. The average age of patients experiencing their first nasal symptoms was 34 and 29 years in two

Hypersensitivity to Aspirin and Other NSAIDs

large studies involving European [11] and American [12] patients. AIA begins with perennial, watery rhinitis more often in women than men. It is likely that a viral infection may initiate the vicious circle of inflammatory events leading to AIA in genetically susceptible subjects [13]. Usually within 2–3 years later, asthma is diagnosed and about this time the first clinical reaction develops, following exposure to aspirin or other NSAIDs inhibiting COX-1, in patients who previously for many years tolerated these drugs very well. The reactions include symptoms from the upper and lower respiratory tract. Bronchospasm is accompanied by profuse rhinorrhea, nasal congestion, sneezing, nasal and ocular itching, ocular ‘injection’, rarely periorbital swelling. Some patients experience skin rash and erythema of the head and neck. The severity of reactions to aspirin ranges from isolated rhinitis to life-threatening anaphylactic reactions. Individual patients following aspirin administration suffer from nausea, stomach cramps or retrosternal pain. Asthma is accompanied by blood and sputum eosinophilia. It is usually severe and runs a protracted course despite avoidance of aspirin and other NSAIDs. Chronic systemic corticotherapy is necessary in about half of the patients. In a recent Japanese study, patients with multiple exacerbations had more frequently hypersensitivity to aspirin and other NSAIDs, required higher doses of corticosteroids and reported more asthma-related hospitalizations [14]. In another study, multiple regression analysis confirmed that aspirinsensitive asthma was strongly associated with a near-fatal asthma [15]. Many patients with AIA suffer from total nasal blockade and post-nasal drip. Loss of smell indicates a development of nasal polyposis, usually filling all sinuses and finally destructing bone structures. Chronic hyperplastic eosinophilic sinusitis with nasal polyposis is often exacerbated by sinus infections: on average 5–6 episodes are diagnosed each year [12].

341

Cross-Reactions with Aspirin and NSAIDs A hallmark of aspirin hypersensitivity is a crossreactivity of aspirin and all NSAIDs that preferentially inhibit COX-1 [4, 16]. In the USA, the most common NSAIDs causing hypersensitive reaction were aspirin (80%) and ibuprofen (41%) [12]. Acetaminophen is the weak inhibitor of COX-1 and stands for a safe therapeutic alternative for almost all patients with aspirin hypersensitivity in doses not exceeding 500–1,000 mg. According to recent data, less than 2% of asthmatics are sensitive to both aspirin and acetaminophen [8]. Meloxicam and nimesulide, preferential inhibitors of COX-2, were found to be well tolerated by patients with AIA, when given at low doses [17]. Higher doses elicited hypersensitive reactions. The highly selective COX-2 inhibitors (rofecoxib, celecoxib, valdecoxib, etoricoxib, parecoxib and lumiracoxib) were characterized by an excellent tolerance by AIA patients in series of placebo-controlled clinical trials. Unfortunately, rofecoxib and some other coxibs were withdrawn from the market subsequently to reports of increased risk of cardiovascular events. However, some patients, extremely sensitive to aspirin, may develop hypersensitive reactions to celecoxib and other coxibs [17–19]. Diagnosis Although the history of adverse reactions to aspirin and other NSAIDs, as well as the typical clinical presentation of this disease may raise a suspicion of aspirin hypersensitivity, the diagnosis of AIA should be based on aspirin challenges. There are four routes of provocation challenges: oral, bronchial inhalation, nasal and intravenous [20]. In vitro tests with aspirin are not currently recommended as a routine practice, although the search for such tests continues [21, 22]. Oral, bronchial and intravenous aspirin challenges must be performed in the hospital under supervision of the patient by the responsible health professionals skilled in performing provocation tests with aspirin. Emergency resuscitative

342

equipment should be readily available. Before the test the patient should be in a stable clinical condition with a baseline FEV1 at least 70% of the predicted value. The challenge should always be preceded by a placebo day. There are various protocols of aspirin challenge differing between centers and standardization of those procedures is pending. In our department the increasing doses of aspirin (27, 44, 117, 312 mg) are administered spaced 1.5– 2 h until a cumulative dose of 500 mg is reached. If the patient with a very strong suspicion of aspirin hypersensitivity shows no reaction to the last dose, 312 mg of aspirin, optionally 500 mg, may be additionally given 1.5 h (or 2 h) following 312 mg of aspirin [20]. Oral challenges are most commonly performed because the oral route mimics natural exposure and the test does not require special equipment. In bronchial challenge lysine-aspirin is administrated by inhalation of lysine-aspirin (L-ASA). The challenge is safer and faster to perform than the oral one, although it is less sensitive. Both oral and bronchial inhalation challenges with aspirin are interrupted if a decrease of FEV1 620% of baseline occurs (a positive reaction), or when the maximum cumulative dose of aspirin is reached (500–1,000 mg or 182 mg, respectively) without a fall in FEV1 620% and the symptoms of aspirin hypersensitivity (a negative reaction). The test could also be regarded as positive when severe extrabronchial symptoms of aspirin hypersensitivity appear. The treatment of respiratory reactions due to aspirin has been reviewed in detail elsewhere [20]. Nasal aspirin challenges with L-ASA [20] or ketorolac solutions are unfortunately far less popular than the oral or bronchial ones in diagnosing aspirin hypersensitivity. They may be recommended in patients who run a high risk of developing serious bronchospastic reactions when exposed to oral or bronchial aspirin challenges. The nasal aspirin challenges are safe and also far less time-consuming and cheaper procedures

Szczeklik ⴢ Niżankowska-Mogilnicka ⴢ Sanak

arachidonic acid

NSAIDs

5-lipoxygenase

5-HETE

cyclooxygenases

PGH2

PGE2

5-HPTE

leukotriene C4 synthase

leukotriene A4

leukotriene A4 hydrolase

leukotriene B4 leukotriene C4 -GTP

PGD2

PGF2

leukotriene D4

leukotriene E4

dipeptidase

Fig. 1. Principal enzymatic transformations of arachidonic acid.

than the oral or bronchial ones. Objective measurements include: rhinomanometry, acoustic rhinometry and/or peak nasal inspiratory flow. Mediators and Mechanisms Prostaglandins The central role of arachidonic acid mediators in aspirin hypersensitivity was deduced from the pharmacological properties of the drugs that precipitate the symptoms [4]. A common pharmacodynamic profile, shared by these non-steroid anti-inflammatory drugs, was inhibition of COX [4, 7, 9]. This enzyme uses arachidonic acid, released by a phospholipase A2, and oxidizes it to an intermediate – prostaglandin H2, which is further metabolized to bioactive prostaglandins E2, F2 and D2 and prostacyclin or thromboxane

Hypersensitivity to Aspirin and Other NSAIDs

A2. Specific synthases of prostaglandins differ in tissue distribution but prostaglandin E2 is produced at low levels by most human cells. The discovery of the second isoenzyme of COX, namely COX-2, permitted to narrow the characteristics of hypersensitivity-precipitating drugs. Lack of a tolerance to drug in susceptible patients correlated with its activity against the non-inducible COX – COX-1, which in contrast to COX-2 has rather a constitutive expression. Leukotrienes Cysteinyl leukotrienes are peptidyl derivatives of arachidonic acid specifically oxidized by 5-lipoxygenase. This enzyme is abundant in granulocytes but only limited types of cells express the synthase of leukotriene C4, a key downstream en-

343

zyme controlling production of cysteinyl leukotrienes: C4, D4 and E4. Precipitation of aspirin hypersensitivity symptoms is marked by a significantly increased production of cysteinyl leukotrienes. Leukotriene E4 was repeatedly demonstrated to rise in urine samples and other body fluids including nasal exudate and bronchoalveolar lavage fluid and exhaled air condensates, following exposure to aspirin [7]. The cells producing cysteinyl leukotrienes were identified as eosinophils and mast cells. A distinct feature of bronchial biopsies in asthmatics with aspirin hypersensitivity was an increased expression of leukotriene C 4 synthase; its protein immunoreactivity correlated negatively with hypersensitivity to the drug, measured as a provocative dose [23]. Cellular Source of Mediators Eosinophils and mast cells closely cooperate as a functional unit in allergic inflammation. It was surprising then to find this common inflammatory mechanism operating in aspirin hypersensitivity, separately from allergic sensitization. As the mast cell specifically produces prostaglandin D2, a stable metabolite of this prostaglandin – 9-,11-PGF2 – was measured in plasma and urine and found to rise during the hypersensitivity reaction [24]. Another accepted marker of activation of mast cells, tryptase, also increased following the aspirin reactions. At baseline, biosynthesis of cysteinyl leukotrienes is increased to the point that asthmatic patients with aspirin hypersensitivity could be distinguished from aspirintolerant asthmatics, even during a symptom-free period [25]. This observation suggests a steady release of cysteinyl leukotrienes from mast cells and eosinophils, which only becomes exaggerated upon exposure to NSAIDs. In fact, there is a correlation between involvement of the upper airways and the amount of cysteinyl leukotrienes produced which suggest that this alteration is rather an effect directly related to the mass of cells capable of producing cysteinyl leukotrienes [26].

344

Lipoxins This third, less known, class of mediators derived from arachidonic acid is also altered in aspirin hypersensitivity. Lipoxins are produced by an enzymatic oxidation of arachidonic acid mostly through transcellular mechanisms. Patients with aspirin hypersensitivity have deficiency in biosynthetic capacity of lipoxins, a trait which was observed mostly in vitro following activation of peripheral blood granulocytes [7]. Lipoxins were proven to diminish inflammatory reaction by inhibition of granulocytes transmigration to the site of inflammation. Alterations in eicosanoids metabolism are not unique for aspirin hypersensitivity. Cellular mechanisms of aspirin hypersensitivity are linked to mast cells and eosinophils. These effector mechanisms differ rather quantitatively than qualitatively from those observed in allergic reactions. Both increased production of cysteinyl leukotrienes and diminished biosynthesis of lipoxins were reported in severe asthma. Upregulation of the lipoxygenase pathway only partially contributes to aspirin hypersensitivity, as promoter variants of 5-lipoxygensae gene (ALOX5) and leukotriene C4 synthase (LTC4S) associate with severe asthma in general. The reason for difficulties in delineation between aspirin hypersensitivity manifested by airway disease and other types of severe asthma is the engagement of the same mechanisms of inflammation. So far, a triggering mechanism for aspirin hypersensitivity, which is the inhibition of COX-1 by a drug, remains the most characteristic feature of the disease. Search for common genetic variants of the genes encoding COX-1 or COX-2 did not reveal any peculiarities in aspirin hypersensitivity. Studies on genetic expression of COXs reported lower levels of COX-2-specific transcripts and protein in the airways [7]. These results, however, might be confused by a substantial inducibilty of the enzyme and its relatively high expression in the airway epithelia. In contrast, no difference in COX-1 levels in bronchial or na-

Szczeklik ⴢ Niżankowska-Mogilnicka ⴢ Sanak

sal mucosa was reported in relation to aspirin hypersensitivity. EP2 Receptor for Prostaglandin E2 and Origin of the Underlying Respiratory Disease Some effects of prostaglandin E2 are mediated by EP2 receptor, activating adenosine triphosphate cyclase. It is of potential importance that in a large genomic study only the polymorphisms within a distant regulatory region of EP2 gene (PTGER2) were associated with aspirin hypersensitivity in asthmatic patients [27]. Deficiency in signaling through EP2 receptor would be difficult to prove functionally in aspirin-hypersensitive patients, as this receptor frequently co-localizes with EP3type receptor, the one with the opposite effects on cyclic AMP accumulation. It was demonstrated, however, that inflammatory cells in aspirin-sensitive rhinosinusitis had lower expression of prostaglandin E2 receptor subtype 2 [28]. Aspirin hypersensitivity is an acquired disease of adulthood. A characteristic and slow onset, reminding of a protracted viral infection of the respiratory tract, raised suspicion on contribution of infectious agents to the disease [13]. As demonstrated recently, asthmatic patients are susceptible to protracted rhinoviral infection due to deficiency in interferon response [29]. In addition, studies on genetic variability of human leukocytes antigens class II loci revealed that in two ethnically distant populations of Slavians and Koreans, a significant association existed between HLA-DRB locus and aspirin hypersensitivity [7]. The immunological restriction of HLA class II system can sum up to immunological incompetence in overcoming the viral disease in patients with respiratory symptoms of aspirin hypersensitivity. Prevention and Treatment Patients with asthma should be alerted to the possibility of aspirin hypersensitivity by their physicians and pharmacists. Once diagnosed with AIA, patients must not use aspirin and any other

Hypersensitivity to Aspirin and Other NSAIDs

Table 1. Common NSAIDs inhibiting COX-1, precipitating reactions in AIA Generic

Brand names

Piroxicam Indomethacin Sulindac Tolmetin Ibuprofen Naproxen Naproxen sodium Fenoprofen Meclofenamate Mefenamic acid Flurbiprofen Diflunisal Ketoprofen Diclofenac Ketorolac Etodolac Nabumetone Oxaprozin

Feldene Indocin Clinoril Tolectin Motrin, Rufen, Advil Naprosyn Anaprox, Aleve Nalfon Meclomen Ponstel Ansaid Dolbid Orudis, Oruval Voltaren, Cataflam Toradol Lodine Relafen Daypro

NSAIDs that inhibit COX-1. Education of patients is crucial. They all should receive a list of contraindicated NSAIDs and safe (well-tolerated) analgesics (table 1). Acetaminophen in doses not exceeding 1,000 mg, coxibs and codeine are optimal preferential choice in patients with acute pain. In general, treatment of AIA follows rules set by international guidelines. AIA is considered as a marker of severe, cortico-dependent asthma. The advent of antileukotrienes, such as zileuton, (5-LO inhibitor) and montelukast, zafirlukast or pranlukast (all Cys-LT1 receptor inhibitors) provided a new opportunity of simultaneous management of aspirin-induced symptoms of both the upper and lower respiratory tract. The clinical efficacy of zileuton in AIA patients was documented in a double-blind placebo-controlled treatment trial [30]. Montelukast was also shown useful in controlling AIA as evidenced by improvement in symptoms, increase in FEV1, fewer asthma exacerbations, better and asthma-specific quality of life. Against expectations, aspirin-

345

hypersensitive asthmatics did not seem to respond better to leukotriene receptor antagonists than aspirin-tolerant asthmatics. The effect of Cys-leukotriene receptor antagonists was better in the carriers of the variant C allele of LTC4S [31] and HLA-DPB1*0301 marker [32]. All patients with aspirin hypersensitivity should use on long-term basis nasal corticosteroids to delay formation of nasal polyps. Some patients require bursts of oral corticosteroids to control their symptoms. Oral and nasal decongestants and antihistamines give additional relief. During acute bacterial sinus infections, extended courses of broad-spectrum antibiotics are frequently required. Chronic desensitization to aspirin, i.e. incremental administration of oral doses of aspirin, can lead to a state of tolerance but daily administration of 600–1,200 mg of aspirin must be continued to maintain this state. Aspirin desensitization is believed most effective in AIA patients with predominating symptoms of chronic sinusitis and recurrent nasal polyps requiring repeated polypectomies. It should be considered also in patients with hypersensitivity to aspirin and other drugs inhibiting COX-1 controlled only with unacceptably high doses of systemic corticosteroids or those who require NSAIDs for the treatment of other diseases, particularly ischemic heart disease. More recently, nasal aspirin challenges were used to study the clinical effectiveness of topical L-ASA in 11 patients with aspirin-sensitive nasal polyposis. Multivariate analysis did not reveal any significant clinical benefit of L-ASA as compared to placebo. In many patients, nasal potency can be restored only through surgery. Simple polypectomy results in scarring which make inevitable revisional procedures more difficult. More extensive procedures, such as functional endscopic sinus surgery, must be undertaken to re-establish proper drainage from sinuses (‘pansinus surgery’). The subjective success rate for nasal symp-

346

toms after ‘pansinus surgery’ reaches 80%, however, the recurrence of nasal polyposis is observed in up to 40% of cases. A retrospective analysis showed long-term postoperative improvement of asthma in 94% of patients subjected to sinus surgery. In a proportion of patients, a burst of corticosteroids may be an optional treatment. In a first randomized prospective study aiming to compare surgical versus medical therapy of chronic rhinosinusitis and concomitant asthma, overall asthma control was better maintained after medical therapy as evidenced by an increase in FEV1 and decrease in exhaled nitric oxide.

Urticaria/Angioedema

Definition, Prevalence Aspirin and other NSAIDs can induce or exacerbate skin eruptions in 20–40% of patients with chronic idiopathic urticaria. This particular kind of urticaria is called aspirin-induced urticaria and its mechanism is related to inhibition of COX-1. Clinical Presentation The reaction may occur in just 15 min or up to 24 h following aspirin ingestion, but on average it develops within 1–4 h. Most cases settle in a few hours, but in severe reactions, bouts of multiform skin eruptions covering most of the body may continue for 10 days after aspirin ingestion [19]. Clinical morphology of the lesions [7] consists of angioedema, transient wheals and confluent evanescent pale erythematous macules. Occasionally, the patients develop lesions described as annular, gyrate and figurate erythematous plaques, sometimes with raised borders. The skin rashes with pleomorphic clinical morphology last longer (3–10 days), are most prominent over the head and neck region, but may also involve trunk and extremities. Standardized scales have been developed to assess the intensity of skin eruptions.

Szczeklik ⴢ Niżankowska-Mogilnicka ⴢ Sanak

Histopathology Histological spectrum of cutaneous reactions to aspirin in chronic idiopathic urticaria was recently reported [33]. Aspirin up to a 500-mg dose induced a restricted range of histological responses with a classic pattern of urticarial tissue reaction occurring in the majority of (12 of 16) cases. Histological criteria of urticaria included the presence of dermal edema, lymphatic dilatation, a perivascular mononuclear infiltrate and interstitial eosinophils and neutrophils. Mast cells around blood vessels and in the interstitial space were identified in all cases. There was no evidence of leukocytoclasis, endothelial swelling, red blood cell extravasation, or fibrin deposition to suggest vasculitis. Two of 16 biopsies showed an interstitial fibrohistocytic (granuloma/annular-like) reaction pattern. One case showed only a sparse perivascular lymphocyte infiltrate, and paucicellurar dermal mucinosis was observed in 1 case. Polymorphism of histological patterns induced by aspirin suggests that in addition to the drug-specific mechanisms triggering eruptions, individual host factors also play a role in determining the ultimate histological phenotype of a drug response. Mechanisms Patients with a chronic idiopathic urticaria, who reacted to the aspirin challenge with a skin rash, shared quite a similar eicosanoid profile with aspirin-induced asthmatics [34]. In fact, about 10% challenged individuals also developed airways manifestation, e.g. nasal congestion and less frequently dyspnea. When contrasted to aspirintolerant chronic urticaria patients, hypersensitivity to aspirin correlated with increased leukotriene E4 excretion in urine at baseline and a few hours following appearance of symptoms after provocation tests. Likewise, the prostaglandin D2 metabolite increased in plasma but within a shorter time course. The extent of the eruption following the challenge correlated with urinary leukotriene E4.

Hypersensitivity to Aspirin and Other NSAIDs

Diagnosis History suggests the diagnosis, which, however, can be established only by oral challenge tests. Such standardized tests have been recently described [20]. Prevention and Treatment Aspirin and all drugs that inhibit COX-1 must be avoided. Coxibs are usually well tolerated, although sporadic adverse reactions were reported [19]. In treatment of the reactions, antihistamines are usually sufficient but in more severe cases adrenaline and corticosteroids may be warranted.

Allergic Reactions to NSAIDs

Although pharmacological inhibition of COX and overproduction of cysteinyl leukotrienes represent one of the most important and frequent pathomechanisms of the adverse reactions to aspirin and other NSAIDs, in a subset of subjects the reactions may have a ‘classic’ allergic background [35]. These reactions are drug-specific, independent of COX inhibition, and probably immune-mediated. Allergic-type responses to NSAIDs include the occurrence of generalized pruritus, erythema, urticaria, angioedema, rhinorrhea, asthma and anaphylactic shock. Pyrazolone drugs (metamizole, noramidopiryne, aminophenazone, propyphenazone) can precipitate reactions ranging from urticaria and angioedema to anaphylactic shock, presumably by immunological, IgE-mediated mechanism [36, 37]. Contrary to patients with AIA, the patients with exclusive hypersensitivity to pyrazolone drugs may take aspirin with impunity. In these cases the results of skin tests (intradermal and prick) with increasing concentrations of aminophenazone/ noramidopiryne are positive. However, intradermal skin tests correlate poorly with the clinical history of hypersensitivity reactions and in vitro tests are not useful in establishing a diagnosis

347

[37]. A genetic predisposition to pyrazolone hypersensitivity reactions linked to the HLA-DQ locus was suggested [38]. Apart from pyrazolone drugs, anaphylactic reactions to other specific NSAIDs have also been reported. Like many drugs, they can function as drug haptens to induce immune sensitization, followed by probable IgE-mediated reactions involving mast cells and basophil activation. A small number of patients

may rarely develop aseptic meningitis after ingestion of a specific NSAID (ibuprofen, sulindac, tolmetin and naproxen) [35]. In this case, cross-reactions among NSAIDs did not occur and aspirin has not been reported to cause aseptic meningitis. Some NSAIDs can cause hypersensitivity pneumonitis; in these cases disappearance of pulmonary infiltrates spontaneously or following corticosteroid therapy could be observed [35].

References 1 Hirschberg VGS: Mittelung über einen Fall von Nebenwirkung des Aspirin Dtsch Med Wochenschr 1902;28:416– 417 and A case report on the side effects of aspirin. Allergy Proc 1990;11: 249–250. 2 Widal MF, Abramin P, Lermoyez J: Anaphylaxie et idiosyncrasie. Presse Méd 1922;30:189–192. 3 Samter S, Beers RF Jr: Intolerance to aspirin. Clinical studies and consideration of its pathogenesis. Ann Intern Med 1968;68:975–983. 4 Szczeklik A, Gryglewski RJ, Czerniawska-Mysik G: Relationship of inhibition of prostaglandin biosynthesis by analgesics to asthma attacks in aspirinsensitive patients. Br Med J 1975;1:67– 69. 5 Kasper L, Sladek K, Duplaga M, et al: Prevalence of asthma with aspirin hypersensitivity in the adult population of Poland. Allergy 2003;58:1064–1066. 6 Gomes E, Cardoso MF, Praca F, Gomes L, Marino E, Demoly P: Self-reported drug allergy in a general adult Portuguese population. Clin Exp Allergy 2004;34:1597–1601. 7 Szczeklik A, Sanak M: The broken balance in aspirin hypersensitivity. Eur J Pharmacol 2006;533:145–155. 8 Jenkins C, Costello J, Hodge L: Systematic review of prevalence of aspirininduced asthma and its implications for clinical practice. BMJ 2004;328: 434–437. 9 Szczeklik A, Stevenson DD: Aspirininduced asthma: advances in pathogenesis diagnosis, and management. J Allergy Clin Immunol 2003;111:913–921.

348

10 Debley JS, Carter ER, Gibson RL, Rosenfeld M, Redding GJ: The prevalence of ibuprofen-sensitive asthma in children: a randomized controlled bronchoprovocations challenge study. J Pediatr 2005;147:233–238. 11 Szczeklik A, Nizankowska E, Duplaga M: Natural history of aspirin-induced asthma. AIANE Investigators. European Network on Aspirin-Induced Asthma. Eur Respir J 2000;16:432–436. 12 Berges-Gimeno M, Simon RA, Stevenson DD: The natural history and clinical characteristics of aspirin exacerbated respiratory disease. Ann Allergy Asthma Immunol 2002;89:474–478. 13 Szczeklik A: Aspirin-induced asthma as a viral disease. Clin Allergy 1988;18: 15–20. 14 Koga T, Oshita Y, Kamimura T, Koga H, Aizawa H: Characterisation of patients with frequent exacerbation of asthma. Respir Med 2006;100:273–278. 15 Yoshimine F, Hasegawa T, Suzuki E, Terada M, Koya T, Kondoh A, Arakawa M, Yoshizawa H, Gejyo F: Contribution of aspirin-intolerant asthma to near fatal asthma based on a questionnaire survey in Niigata Prefecture, Japan Respirology 2005;10:477–484. 16 Szczeklik A, Gryglewski RJ, Czerniawska-Mysik G: Clinical patterns of hypersensitivity to nonsteroidal anti-inflammatory drugs and their pathogenesis. J Allergy Clin Immunol 1977;60:276–284. 17 Bavbek S, Celik G, Ozer F, Mungan D, Misirligil Z: Safety of selective COX-2 inhibitors in aspirin/NSAID-intolerant patients: comparison of nimesulide, meloxicam and rofecoxib. J Asthma 2004;41:67–75.

18 Baldassarre S, Schandene L, Choufani G, Michils A: Asthma attacks induced by low doses of celecoxib, aspirin and acetaminophen. J Allergy Clin Immunol 2006;117:215–217. 19 Mastalerz L, Sanak M, Gawlewicz A, Gielicz A, Faber J, Szczeklik A: Different eicosanoid profile of the hypersensitivity reactions triggered by aspirin and celecoxib in a patient with sinusitis and asthma. J Allergy Clin Immunol 2006;118:957–958. 20 Niżankowska-Mogilnicka E, Bochenek G, Mastalerz L, et al: EAACI/GA2LEN guideline: aspirin provocation tests for diagnosis of aspirin hypersensitivity. Allergy 2007, in press. 21 Kowalski ML, Ptasinska A, Jedrzejczak M, Bienkiewicz B, Cieslak M, Grzegorczyk J, Pawliczak R, Dubuske L: Aspirintriggered 15-HETE generation in peripheral blood leukocytes is a specific and sensitive Aspirin-Sensitive Patients Identification Test (ASPI Test). Allergy 2005;60:1139–1145. 22 Gamboa P, Sanz M, Caballero MR, Urrutia I, Antepara I, Esparza R, et al: The flow-cytometric determination of basophil activation by aspirin and other non-steroidal anti-inflammatory drugs (NSAIDs) is useful for in vitro diagnosis of the NSAID hypersensitivity syndrome. Clin Exp Allergy 2004;34: 1448–1457. 23 Cowburn AS, Sładek K, Soja J, Adamek L, Niżankowska E, Szczeklik A, Lam BK, Penrose JF, Austen FK, Holgate ST, Sampson AP: Overexpression of leukotriene C4 synthase in bronchial biopsies from patients with aspirin-intolerant asthma. J Clin Invest 1998;101:834–846.

Szczeklik ⴢ Niżankowska-Mogilnicka ⴢ Sanak

24 Bochenek G, Nagraba K, Niżankowska E, Szczeklik A: A controlled study of 9,11-PGF2 (a prostaglandin D2 metabolite) in plasma and urine of patients with bronchial asthma and healthy controls after aspirin challenge. J Allergy Clin Immunol 2003; 111:743–749. 25 Sanak M, Kiełbasa B, Bochenek G, Szczeklik A: Exhaled eicosanoids following oral aspirin challenge in asthmatic patients. Clin Exp Allergy 2004; 34:1899–1904. 26 Higashi N, Taniguchi M, Mita H, Kawagishi Y, Ishi T, Higashi A, Osame M, Akiyama K: Clinical features of asthmatic patients with increased urinary leukotriene E 4 excretion (hyperleukotrienuria): involvement of chronic hyperplastic rhinosinusitis with nasal polyposis. J Allergy Clin Immunol 2004;113:277–283. 27 Jinnai M, Sakagani T, Sekigawa T, Kakihara M, Nakajima T, Yoshud K, Goto S, Hasegawa T, Koshino T, Hasegawa Y, Inoue H, Suzuki N, Sano Y, Inoue I: Polymorphisms in the prostaglandin E2 receptor subtype 2 confer susceptibility to aspirin-intolerant asthma: a candidate gene approach. Hum Mol Genet 2004;13:3203–3217. 28 Ying S, Meng Q, Scadding G, Parikh A, Corrigan CJ, Lee TH: Aspirin-sensitive rhinosinusitis is associated with re-

29

30

31

32

duced E-prostanoid 2 receptor expression on nasal mucosal inflammatory cells. J Allergy Clin Immunol 2006;117: 312–318. Contoli M, Message SD, Laza-Stanc V, Edwards MR, Wark PAB, Bartlett NW, Kebadze T, Mallia P, Stanciu LA, Parker HL, Slater L, Lewis-Antes A, Kon OM, Holgate ST, Davies DE, Kotenko SV, Papi A, Johnston SL: Role of deficient type III interferon- production in asthma exacerbations. Nat Med 2006;12:1023–1026. Dahlen B, Nizankowska E, Szczeklik A, Zetterstrom O, Bochenek G, Kumlin M, et al: Benefits from adding the 5-lipoxygenase inhibitor zileuton to conventional therapy in aspirin-intolerant asthmatics. Am J Respir Crit Care Med 1998;157:1187–1194. Mastalerz L, Nizankowska E, Sanak M, Mejza F, Pierzchalska M, Bazan-Socha S, et al: Clinical and genetic features underlying the response of patients with bronchial asthma to treatment with a leukotriene receptor antagonist. Eur J Clin Invest 2002;32:949–955. Park HE, Kim SH, Sampson A, Lee KW, Park CS: The HLA-DPB1*0301 marker might predict the requirement for leukotriene receptor antagonist in patients with aspirin-intolerant asthma. J Allergy Clin Immunol 2004;114:688–689.

33 Zembowicz A, Mastalerz L, Setkowicz M, Radziszewski W, Szczeklik A: Histological spectrum of cutaneous reactions to aspirin in chronic idiopathic urticaria. J Cutan Pathol 2004;31:323– 329. 34 Mastalerz L, Setkowicz M, Sanak M, Szczeklik A: Hypersensitivity to aspirin: common eicosanoid alterations in urticaria and asthma. J Allergy Clin Immunol 2004;113:771–775. 35 Stevenson DD, Simon RA, Zuraw BL: Sensitivity to aspirin and nonsteroidal anti-inflammatory drugs; in Adkinson NF, Yunginger JW, Busse WW, et al (eds): Middleton’s Allergy Principles and Practise. St Louis, Mosby, 2003, pp 1695–1710. 36 Czernawiska-Mysik G, Szczeklik A: Idiosyncrasy to pyrazolone drugs. Allergy 1981;36:381–384. 37 Szczeklik A, Niżankowska E, Czerniawska-Mysik G: Tolerance of azopropazone: allergic and pseudoallergic reactions and comparison with other NSAIDs; in Rainsford KD (ed): Azopropazone. Over Two Decades of Clinical Use. Dordrecht, Kluver Academic, 1989, pp 265–274. 38 Kowalski ML, Woszczek G, Bienkiewicz B, Mis M: Association of pyrazolone drug hypersensitivity with HLADQ and DR antigens. Clin Exp Allergy 1998;28:1153–1158.

Dr. Andrzej Szczeklik Department of Medicine, Jagiellonian University School of Medicine Skawińska 8, PL–31-066 Kraków (Poland) Tel. +48 12 656 2840, Fax +48 12 656 5786 E-Mail [email protected]

Hypersensitivity to Aspirin and Other NSAIDs

349

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 352–365

Approach to the Patient with a Drug Hypersensitivity Reaction – Clinical Perspectives Andreas J. Bircher Allergy Unit, Department of Dermatology, University Hospital Basel, Basel, Switzerland

Abstract

Introduction

Adverse drug reactions are a common clinical problem. Drug hypersensitivity reactions are systemic immune reactions which often involve the skin. Indeed, the skin may be the only organ affected or it may herald a more serious multiorgan reaction. The cutaneous manifestations include a wide spectrum such as wheals, angioedema, macular, papular, vesicular, bullous, vasculitic and hemorrhagic lesions. Similar processes may also occur in internal organs, mainly in the liver and the kidney, but sometimes lung, heart or pancreas may be involved. The first step in the clinical approach is the recognition or exclusion of danger signs, which relies on a detailed morphological diagnosis of skin symptoms and signs, and which should also include some laboratory tests to detect the involvement of internal organs. Cofactors like an underlying viral infection should be particularly searched for. The morphological diagnosis helps to exclude other dermatological diagnoses, and it is important to establish a pathogenetic hypothesis, which helps to select the correct diagnostic tools to identify the incriminated drug. To identify the potentially eliciting drug or drugs, the time pattern (chronology) of the reaction, the dose and duration of drug intake, previous exposure and cofactors have to be considered. Validated diagnostic tools are, with the exception of the
-lactam antibiotics, generally or commercially not available. Nevertheless, a battery of drugs can be tested in vivo (prick, intradermal and patch tests) and in vitro – whereby, because of the low sensitivity of these tests, only positive tests are relevant, while negative tests cannot rule out a sensitization. We propose a stepwise, clinically oriented approach to the patient with a suspected cutaneous hypersensitivity reaction, based on the primary clinical manifestation and the chronology of the reaction.

Drug hypersensitivity reactions are a common clinical problem which may affect a considerable number of the treated patient population. Between 10 and 15% of patients may suffer from an unwanted drug reaction, 2–5% of these have to be hospitalized, and in 1–3% of hospitalized patients mortality may result [1]. Drug hypersensitivity reactions are systemic reactions which often involve the skin. Cutaneous manifestations may be the only clinically relevant presentation of a drug hypersensitivity, but the skin can also act to herald for example a systemic hypersensitivity reaction such as anaphylaxis or the severe drug hypersensitivity syndrome (drug rash with eosinophilia and systemic symptoms (DHS/DRESS) [2]. Particularly the hypersensitivity reactions of type B (bizarre) present with symptoms and signs which are typically not related to pharmacologic doses and effects of the eliciting drugs. Therefore, drug hypersensitivity reactions present a complex challenge for treating physicians and the diagnosing allergologist. To arrive at a useful final diagnosis, a structured stepwise approach is required. In principle, two steps can be differentiated: (1) handling the acute drug hypersensitivity reac-

Copyright © 2007 S. Karger AG, Basel

B.. Morphological differential diagnosis 1. Drug-related exanthems 2. Other skin disorders

A.. Clinical diagnosis ·Patient history ·Drugs ·Clinical findings ·Ev. histology and laboratory

B2

C. Other diagnosis 1. Infectious exanthem 2. Skin disease 3. Relapse of another disease

B1

D. Severity 1. Low (macular-papular) 2. High (bullous, vasculitis, angioedema, urticaria)

D1

D2

C1/2 E. Etiological differential diagnosis 1. One drug 2. Several drugs

F1

E1

G. Procedure 1. Immediate drug withdrawal 2. Symptomatic treatment 3. Consider alternative drug

H2

E2

H. Urgent indication 1. Consider continuing therapy under strict clinical observation 2. In case of exacerbation G

I2

F.. Relationship 1. Certain, probable 2. Possible, doubtful 3. None

F2/3 I. Procedure 1. Continue therapy (with unsuspected drugs) 2. Worsening of symptoms

H

J. Diagnostic procedures After the healing of clinical signs (3–4 weeks to 6–12 months)

Fig. 1. Stepwise clinical approach to the patient with a suspected acute drug hypersensitivity reaction [according to 11, 15].

tion: it comprises the classification as well as the definition of the severity of the reaction, its documentation, the choice of an alternative drug, and the symptomatic treatment, and (2) identification of the eliciting drug(s).

Acute Phase

Drug hypersensitivity reactions can occur during all types of medical treatments, including biologicals [3], their additives [4, 5], phytotherapeutic remedies [6] or their adulterated contents [7, 8]. It is of utmost importance that the initial diagnosis is done carefully including some laboratory analysis, in particular an involvement of

Approaching Patients with a Drug Hypersensitivity Reaction

blood cells, and the affection of internal organs, such as liver and kidneys, should be excluded by appropriate blood tests (differential blood count, liver enzyme analysis, etc.). Actually, it is advisable to consult a dermatologist or alternatively to take a photograph of the exanthem and of the particular cutaneous efflorescences, and to note all drugs taken during the last 2 weeks and particularly all newly introduced drugs taken within the last 4 weeks. A skin biopsy for histological examination may be helpful to further differentiate the exanthem. This documentation should give a clue to the severity of the reaction [2], which is decisive for prognosis, treatment and future pharmacotherapy.

353

Table 1. Drug-induced cutaneous symptoms/signs Pruritus Urticaria, angioedema Exanthems Macular Papular Vesicular Pustular Bullous Fixed drug exanthem Hemorrhagic-necrotic Lichenoid Psoriasiform Pityriasiform Exfoliative Erythroderma Mucous membranes Erosions, aphthous lesions, ulcers Stomatitis Genital lesions Conjunctivitis Rhinitis Photoinduced exanthems Cutaneous vasculitis Erythema nodosum Cutaneous lupus erythematodes

The morphological differential diagnosis should exclude other causes, the most probably eliciting drugs should be immediately stopped and symptomatic treatment should be started if necessary, particularly when danger signs such as severe anaphylactic symptoms, facial edema with drug treatments causing DHS/DRESS or bullous manifestation, vasculitis or important mucosal signs are present [2]. An algorithm is proposed in figure 1. After the complete clearing of the clinical manifestations and normalization of laboratory values, an allergologic investigation is recommended. This should be best done between 3 weeks to 6 months after the incident.

354

Morphological Diagnosis Initially, a comprehensive clinical diagnosis including clinical status (fever, malaise lymphadenopathy, liver enlargement) and a detailed description of the cutaneous signs are necessary. The clinical investigation must include a complete examination of the skin and its appendages, including the mucous membranes of the mouth, eyes and genitals. The type and distribution of the lesions, sparing for example of light-protected areas, involvement of the face or the oral mucosa should be noted. Pruritus of variable severity is typically present. A clear and exact description of the clinical manifestations (table 1) is paramount for the correct identification of a cutaneous hypersensitivity reaction. Based on this information, a pathogenetic hypothesis of the reaction can be established. This is important for the later selection of the correct diagnostic tools, which may help to identify the eliciting drug or drugs. The suspected drugs should be immediately stopped, particularly if danger signs such as bullous or hemorrhagic lesions or mucosal affections or involvement of internal organs are present (table 2) [2]. The morphological presentations of cutaneous hypersensitivity reactions are manifold and variable. Although the skin has only a limited repertoire of inflammatory reaction patterns, a wide spectrum such as pruritus, erythema, erythroderma, urticaria, angioedema, macular, papular, vesicular, bullous, vasculitic and hemorrhagic exanthems (table 1) may be encountered [9, 10]. Some are characteristic for adverse skin reactions such as for example fixed drug eruption, whereas others may present a difficult challenge in the differential diagnosis towards other skin disorders, particularly infectious exanthems. Therefore, for cutaneous drug hypersensitivity reactions the term ‘the great imitator’ has been applied, as it has been used in former times for syphilis [11].

Bircher

Table 2. Danger signs for severe cutaneous reactions [according to 2] Clinical manifestations Immediate hypersensitivity reactions Sudden flush (face) Severe urticaria Severe angioedema (particularly of the oropharynx) with difficulty swallowing, hoarseness DHS/DRESS Centrofacial edema Involvement of extended body surface (erythroderma) Bullous skin disorders Painful skin Nikolsky’s sign positive Epidermolysis Vesicles, bullous lesions Mucosal erosions or aphthous lesions Atypical target lesions Hemorrhagic or necrotic lesions (vasculitis) Systemic signs High fever Malaise Laboratory findings Eosinophilia Neutrophilia Lymphoblasts Blood cytopenia (thrombopenia, lymphopenia, agranulocytosis) Elevation of liver enzymes and bilirubin (hepatopathy) Elevation of creatinine, proteinuria, nephritic sediment (nephritis)

Laboratory Analyses In the acute phase, skin tests and laboratory analyses are rarely helpful to identify the eliciting drug but might be essential to detect or exclude affections of internal organs such as liver and kidneys. Eosinophilia, neutrophilia, presence of lymphoblasts or cytopenia should be searched for – they may indicate a very strong immune reaction or the involvement of antibody-mediated cell death or bone marrow toxicity. A histological analysis of the cutaneous lesions may help to further specify the manifestation and to differentiate it from other skin disorders.

Approaching Patients with a Drug Hypersensitivity Reaction

Etiology

The diagnosis of drug hypersensitivity aims, firstly, to define the disease as drug allergy, and, secondly, to identify the eliciting drug or drugs. History and experience (from various textbooks listing drugs and their side effects) are the most widely used and important pillars on which the identification of a potentially eliciting drug is based. Dose, the duration of treatment, the chronology of the appearance of symptoms, and the identification of cofactors such as underlying disorders, for example viral infections, hepatopathy or renal diseases help for the identification of the relevant drug. The application of a standardized approach, for example with a drug questionnaire, is recommended [12], in particular for persons less experienced in drug hypersensitivity. Specific Diagnostic Approach The specific etiologically oriented allergy diagnosis should usually be started the earliest 3–4 weeks after complete clearing of all clinical symptoms and signs. However, some recent data from Japan suggest that in bullous skin diseases the relevant skin and in vitro tests are more frequently positive in the first 3 weeks of the disease [see chapter of Shiohara et al., pp 251–266]. On the other hand, after a time interval of more than 6– 12 months some drug tests (like circulating IgE) may already have become negative, resulting in false-negative results. T-cell reactions tend to persist longer: some were found to be still positive 12 years after the event, but other reactions tend to diminish over the years. According to the clinical manifestations and their chronology, a hypothesis of the putative pathogenesis [13, 14] should be generated to select appropriate test procedures (table 3) [15]. Chronology (Time Patterns) The time period between the first dose and the occurrence of first symptoms may vary. Sensitiza-

355

Table 3. From clinical manifestations to a pathogenetically orientated diagnosis [according to 15] Clinical manifestation

Potential pathogenesis

Diagnostic tests

Urticaria, angioedema

Pseudoallergy, IgE-mediated

Prick, intradermal skin tests, specific IgE, basophil activation tests

Anaphylaxis

Pseudoallergy, IgE-mediated, rarely immune complex activation (C3a, C5a)

Prick, intradermal skin tests, IgE, basophil activation tests

Maculopapular exanthem

T-cell-mediated (IVb, IVc)

Patch and intradermal tests (reading at 24/48 h), lymphocyte transformation tests

Vesicular-bullous exanthem

Cytotoxic T cells (type IVc), NK and NK-T cells (?); other immunological mechanisms

None, basophil activation tests

Pustular exanthem

T cells (type IVc, IVd)

Patch and intradermal tests (reading at 24/48 h), lymphocyte transformation test

Fixed drug eruption

Some type IV (?)

Patch test in affected area

tion to a drug or its metabolites is thought to take 5–7 days but for some reason it may take up to several weeks or months until the clinical allergy develops, in particular in DHS/DRESS [16] and in patients with TEN under treatment with systemic corticosteroids [17, 18]. This time lag between the start of pharmacotherapy until the beginning of first symptoms is the induction interval. Once an individual is sensitized, the time between the last dose and the first occurrence of symptoms (socalled reaction interval) varies depending on the pathomechanism [11, 19]. In IgE-mediated and pseudoallergic reactions the typical eliciting or reaction period varies between a few minutes to several hours. In T-cell-mediated (some exanthema) reactions the time delay is typically between a few hours to 2 days. Typically, there is also a certain inherent dynamism in the reaction. For example, an exanthema may set out with macules within hours, which later turn into papules and peak at 48–72 h. In SJS and TEN, macules on painful skin may turn into epidermolysis and bullae within 6–12 h, that is, not only the distribution and extent of the exanthem, but also of the single lesions vary over time. In immediate reactions

356

this point is less important, however, also in IgEmediated reactions the delay after antigen contact may be several hours [20]. Many factors, among them the route of administration, drug metabolism and other cofactors (food intake, drug interactions, stress or effort situations), play a modifying role. For this reason the somewhat rigid separation between immediate and non-immediate reactions at 1 h [21, 22] may be problematical, because there is a considerable overlap of the particular chronologies of the different reactions. Diagnostic Tests The primary diagnosis is always based on a detailed clinical history and thorough examination to make a correct morphological diagnosis. The clinical judgment remains the mainstay, because due to low sensitivity of the available tests for a considerable number of drug reactions they remain negative, and provocation tests are not always feasible [23]. Practical Procedure A suspected drug allergy should be diagnosed on a concept of an underlying pathogenetic mecha-

Bircher

A.. Documentation 1. Patient history 2. Detailed clinical manifestation 3. All drugs and cofactors

B.. Hypothesis pathomechanism 1. Type I 2. Type IV 3. Pseudoallergy/intolerance 4. Other mechanisms/unknown

C.. Uncertain/unknown pathomechanism According to situation go to D, E, F (G, H)

B4

B2

B1

B3

D.. Investigations type I 1. Prick/intradermal tests 2. Serology (IgE antibodies) 3. Consider basophil activation tests 3. Consider provocation test

E. Investigations type IV 1. Patch/intradermal tests 2. Lymphocyte transformation test 3. Consider provocation test _

+

_ +

+ G. Trigger identified 1. Patient information 2. Consider provocation test with alternative drugs 3. Report/allergy passport

F.. Investigations intolerance 1. Consider provocation test 2. Mediator releasing tests 3. Basophil activation tests

_

H.. Trigger uncertain 1. Report/temporary allergy passport 2. Consider provocation test with alternative drugs

I. Drug therapy indicated again 1. Alternative drug 2. In case of lacking alternative: Consider desensitization Consider premedication

Fig. 2. Stepwise approach for a pathogenetically based allergologic diagnosis of a drug hypersensitivity reaction [according to 11, 15].

nism [13, 14]; an exact history and description of the clinical manifestations are mandatory to be able to select the adequate test procedures. It should include all drugs that have been taken before the reaction occurred and also potential cofactors should be taken into account. A diagnostic algorithm is proposed in figure 2. However, only a comprehensive judgment of all factors may allow to make a final etiological diagnosis. Since neither the pathogenesis of some reactions, nor the optimal skin test concentrations for many drugs are known exactly, and as validation of the

Approaching Patients with a Drug Hypersensitivity Reaction

few in vitro tests is still lacking for many drug reactions, the diagnosis of drug hypersensitivity reactions remains a challenge. A true-positive skin or in vitro test is an indication of a hypersensitivity to a particular drug or drug allergen. The clinical relevance should be established based on the history, the clinical manifestation, the chronology and the likelihood of the drug to elicit such a hypersensitivity reaction. In case of negative skin and in vitro tests, hypersensitivity to the potentially eliciting drug or drug allergen is not definitely excluded. There-

357

Table 4. Major clinical features of some drug-induced syndromes AGEP

DHS/DRESS

SDRIFE/ (baboon syndrome)

Localization

Disseminated generalized

Generalized exanthem centrofacial edema

Flexural/intertriginous Disseminated exanthem exanthem generalized

Morphology

Erythroderma pustules (small, numerous)

Maculopapular lesions Erythroderma

Erythema Papules (pustules) Rarely bullae

Macular lesions Bullous lesions Epidermal necrolysis Stomatitis, conjunctivitis

Involvement of internal organs

Possible, rare

Obligatory Lymphadenopathy Hepatitis Pneumonitis Nephritis

Not affected

Common Pneumonitis Mucosal lesions (GIT)

Systemic signs

High fever

Fever Malaise Fatigue

None

Fever Malaise Fatigue

Pertinent laboratory findings

Neutrophilia

Eosinophilia (⬃90%) Transaminases, -GT elevated atypical lymphocytes

None

Lymphopenia

fore, in doubt, the potentially eliciting drugs have to be withdrawn from future pharmacotherapeutic treatments. Provocation tests are reserved for vital indications only and physicians experienced in this particular field of drug allergy should only perform this particular test [23]. Its advantage is the testing of the patient with its individual metabolism and immunogenetic background. It currently remains the gold standard; however, because severe relapses may occur, its indication is restricted. Finally the patient should be informed about his or her drug hypersensitivity and an appropriate allergy passport should be issued, which includes the generic and the common commercial names of the eliciting drug and the date and the manifestations of the clinical reaction. Optimally it should also include recommendations for alternative drugs.

358

SJS/TEN

Major Clinical Manifestations

The clinical characteristics of the more common and of some of the severe drug reactions are briefly presented. Some drug-induced syndromes have been proposed to differentiate them from other diseases (table 4). More detailed information is presented in other chapters of this book. General Symptoms In drug hypersensitivity reactions, often nonspecific symptoms, such as general malaise, fatigue, headaches, and vomiting, may mimic other diseases such as infections. But they can also be alarming symptoms or danger signals for DRESS or anaphylaxis. Drug Fever Fever may be the only manifestation of a drug hypersensitivity reaction; its frequency is esti-

Bircher

mated at a few percent of all unwanted effects. Many patients suffer from additional symptoms such as headaches and myalgias. The pathogenesis is variable: stimulation of IL-6, TNF-  in the frame of a strong immune response may cause fever, as well as formation of immune complexes. It is important to screen for CRP – which may be low in some severe reactions (e.g. DHS/DRESS), but high in others (e.g. amoxicillin-induced interstitial pneumopathy). Drugs which may elicit an isolated ‘drug fever’ are antibiotics, cytostatics, sulfonamides, antidepressants, and many others [24]. On the other hand, fever may precede or accompany particularly the more severe drug-induced exanthems such as TEN or DRESS and thus represent an important danger sign. Associated and Isolated Affections of Internal Organs Particularly the liver and the circulating blood cells may be affected in drug-induced reactions, either as only manifestation or in complex syndromes (table 4). Hepatitis is probably a more important cause of death than anaphylaxis. Lymphadenopathy, nephritis, pneumonitis, and pancreatitis have been observed and documented and, dependent on the drug, other organs such as the myocardium, the thyroid gland and the gastrointestinal tract may be affected as well (e.g. in allopurinol- or abacavir-induced hypersensitivity reactions). Some of these organs may be involved in the context of drug-induced autoimmune disorders such as lupus [25, 26] and vasculitis [27]. In addition, a variety of drugs elicit exclusive reactions in liver, kidney or lung, and often without skin symptoms: such reactions are difficult to recognize, which often delays the diagnosis. An eosinophilia may be an indication for it [28]. Urticaria and Angioedema Urticaria is a disseminated skin eruption characterized by rapidly emerging, migrating and itching wheals. A wheal is a rather flat, pale-red ele-

Approaching Patients with a Drug Hypersensitivity Reaction

vation caused by edema in the upper part of the dermis. Wheals mainly affect trunk and extremities, after clearing the skin surface remains intact. The lesions appear typically within minutes and usually persist less than 24 h at one location. Angioedema is a tense non-pitting edema of the deeper skin layers, affecting mainly the face, lips, oral mucosa and male genitals. Urticaria and angioedema can result from a type I (IgE antibody) or a type III (immune complex) reaction according to Coombs and Gell, but also from non-antibody-related, so-called pseudoallergic mechanisms as for example in aspirin intolerance. Eliciting drugs include antibiotics, NSAIDs, radiocontrast media, general anesthetic drugs and many others. Histamine is the principal mediator of urticaria; it leads to itching and increased vascular permeability. In angioedema, bradykinin may play a more important role [29]. A vast array of factors can trigger mast cells: beside drugs, also food additives, complement split products generated during infections, IgE-mediated allergies to food proteins, neurotransmitters like substance P, etc., can cause mast cell degranulation, which renders the etiological diagnosis difficult. Anaphylaxis Pruritus, urticaria and angioedema may present as initial or partial symptoms of full-blown anaphylaxis. Sudden pruritus of the capillitium, the palmae and plantae, or the large folds may be an early symptom. Also facial flush, conjunctivitis and erythema of the upper trunk may be initial symptoms of anaphylaxis. Other signs are angioedema, urticaria, rhinitis, bronchospasm, nausea, vomiting, abdominal pain, diarrhea, tachycardia, hypotension, shock and cardiorespiratory arrest. Cerebral hypoxia may result in headaches, disorientation, epilepsy and death. Anaphylaxis is typically IgE-mediated, however, for some drugs other pathomechanisms are discussed. In such cases the term anaphylactoid reaction is used. Typically, anaphylactic symp-

359

toms start within minutes up to a few hours after exposure to the eliciting substance. Serum Sickness Serum sickness classically manifests itself with fever, macular and urticarial exanthems, lymphadenopathy, arthralgias and sometimes peripheral edemas. Classical serum sickness is an example of an immune complex-mediated type III reaction. Antibiotics such as penicillins, cotrimoxazole, cefaclor and rifampicin may induce a serum sickness-like disease without measurable immune complexes or complement activation among others. The disease has a reaction latency period of 6–8 h [30]. Drug Hypersensitivity Syndrome For the so-called drug-induced hypersensitivity syndrome, the acronym DRESS (drug-related eosinophilia with systemic symptoms) has been proposed [16]. Typical for this multiorgan syndrome is a macular and/or papular, more rarely hemorrhagic or bullous exanthem and an erythematous centrofacial swelling. Fever, general malaise, and lymphadenopathy may be accompanied by hepatitis (in 50%), nephritis (in 10%) and more rarely pneumonitis. In the peripheral blood, eosinophilia (in 90%), lymphocytosis, agranulocytosis, and anemia may be observed [31]. Mortality may be up to 8% [32]. Common eliciting drugs include sulfonamides, aromatic antiepileptics such as phenytoin and carbamazepine, allopurinol and minocycline. Reactivation of human herpes family viruses during the usual evolution over 3 weeks of the disease has recently been demonstrated [33–36]. The period between the intake of the eliciting drug and the start of clinical manifestations may last from 6 to 12 weeks [for details, see chapter of Shiohara et al., pp 251– 266]. Maculopapular Exanthems These exanthems manifest with erythematous macules and infiltrated papules, affecting partic-

360

ularly the trunk and the proximal extremities. Confluent exanthems are possible, resulting in erythroderma. Typically, a more or less pronounced desquamation occurs after clearing of the lesions. The differential diagnosis includes the classical infectious exanthems such as measles and rubella; therefore these exanthems were formerly called morbilliform and rubeoliform. Exanthems are the result of a T-cell-mediated drug allergy. Therefore, in a sensitized individual, symptoms typically appear 24–48 h after antigen exposure. However, first signs such as a discrete erythema may already appear after a few hours. The exanthems show distinct clinical features dependent on different cytokine production by the drug-specific T cells [13]. In all forms drug-specific cytotoxic T cells are involved [13]. Common elicitors include antibiotics such as aminopenicillins [21], cephalosporins, sulfonamides, quinolones [37], antiepileptic drugs [38], radiocontrast media [39, 40], among many others. Pseudolymphomatous Reactions Macular exanthems may proceed into infiltrated lesions with the histological presentation of a pseudolymphoma. They have a slower and more protracted time course and, therefore, the relation to the eliciting drug may be obscured. Cases from carbamazepine, phenytoin [41], ACE inhibitors and quinidine have been reported [42]. Pustular Exanthems Isolated pustules may be observed in maculopapular exanthems, however for disseminated forms the acronym acute generalized exanthematic pustulosis (AGEP) has been proposed [43]. On a generalized erythema, multiple 1- to 3-mm measuring sterile pustules, rarely developing into larger bullae [44], associated with pruritus or burning, are observed. The initial lesions may start in the face or in the intertriginous areas and may disseminate within hours.

Bircher

b

a

Fig. 3. Clinical manifestation of flexural exanthem (SDRIFE) to an iodinated contrast medium [59] . Typical distribution with affection of the inguinal, perigenital area (a) and large body folds (b).

Fever and leukocytosis are always present (table 4). A diagnostic algorithm for this not uncommon drug hypersensitivity reaction has been proposed [45]. The reaction is most often elicited by
-lactam antibiotics, particularly aminopenicillins and cephalosporins, but has also been observed with a large number of other drugs and rarely with infections [43, 45]. Flexural Exanthem (So-Called Baboon Syndrome) In some instances a maculopapular exanthem may manifest in a symmetrical flexural distribution. It is a rare but possibly underreported manifestation, and in contrast to most other drug exanthems, males are more often affected than females. A sharply delineated erythema of the perigenital and perianal area is associated with the affection of the large folds, such as the axilla, the elbows and the knee folds (fig. 3a, b). Rarely pustules or vesicles may be observed. The reaction may later result in a considerable desquamation. This particular drug reaction is mainly elicited by aminopenicillins, but also other drugs have been implicated. Systemic symptoms and signs are typically absent (table 4) [46]. To delineate this particular drug re-

Approaching Patients with a Drug Hypersensitivity Reaction

action from systemic contact dermatitis to systemically absorbed contact allergens such as mercury or nickel, the neutral acronym SDRIFE has been proposed [46, 47]. SDRIFE stands for symmetrical drug-related intertriginous and flexural exanthem. Fixed Drug Eruptions The clinical manifestation of a fixed drug eruption is virtually pathognomonic for a drug-induced reaction. It is a rather common cutaneous adverse reaction [48]. A livid-erythematous, sometimes edematous plaque often with a later developing bullous center is observed mainly at acral localizations (fig. 4). Typically, a residual hyperpigmentation persists, rarely there are nonpigmenting forms [48]. Isolated mucosal lesions in the oral cavity or on the male genitals may be present. The reaction always occurs in the same localization upon re-exposure to the eliciting drug. The fixed drug eruption is typically singular, but multilocular disseminated manifestations have been observed [10]. If they are bullous, they have to be differentiated from SJS/TEN, where the patients are sicker as they have far more general symptoms. A T-cell-mediated mechanism has been proposed [49]. Common eliciting

361

4

5 Fig. 4. Fixed drug eruption of the tongue with erythema and central erosion due to blistering. Fig. 5. Cutaneous hypersensitivity vasculitis with purpuric papules.

agents include sulfonamides, tetracyclines, pyrazolones and carbamazepine [48]. Vesicular and Bullous Exanthems The severe vesicular-bullous exanthems are not an entity, but include a range of disorders. Among these are erythema exsudativum multiforme, minor and major form (EEM), Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN or Lyell’s disease) [10, 30, 50]. They are differentiated from each other by the presence or absence of typical target lesions and particularly the extent of the skin and mucosal manifestations and associated symptoms [51]. EEM is actually rarely caused by drugs. Affected persons are younger patients who often suffer from repeated viral infections, in particular to herpes virus type I. In SJS and TEN, the skin may be initially painful before bullous lesions evolve, and affection of the mucosa is always present. This may involve the conjunctival, oral, genital and perianal mucosa as well as mucosal surfaces of inter-

362

nal organs. Lethal outcome is present in approximately 30% of TEN cases [for details, see chapter of Allanore and Roujeau, pp 267–277]. Vasculitis Drugs rarely elicit a cutaneous or systemic vasculitis. A typical manifestation is the cutaneous socalled allergic or hypersensitivity vasculitis [27]. It is characterized by inflammation and necrosis of the walls of blood vessels occurring within a few weeks of drug initiation. Cutaneous lesions are characterized by palpable purpuric papules, favoring dependent areas, particularly the legs (fig. 5), but occurring anywhere on the body. Less commonly, urticarial plaques or hemorrhagic blisters followed by ulcerations are seen. Affection of internal organs such as the gastrointestinal tract, the kidneys, the liver and the central nervous system may be life-threatening and has to be excluded [52]. Hypersensitivity vasculitis must be distinguished from other forms such as Henoch-Schönlein purpura, cryoglobulinemia,

Bircher

infection-associated and systemic vasculitis (polyarteritis nodosa, Wegener’s granulomatosis). Histopathological findings of leukocytoclastic vasculitis affecting superficial cutaneous vessels are seen in hypersensitivity vasculitis and other small-vessel vasculitides. Immunofluorescence evaluation of a skin biopsy may aid in differentiation from other conditions. Immediate discontinuation of offending agents is mandatory.

IgA bullous dermatosis is a less common variant that is still not well characterized. Histologically subepidermal cleavage with a neutrophilic infiltrate is seen, direct immunofluorescence reveals granular IgA and C3 deposition along the dermal-epidermal junction. Among the various implicated causative drugs, vancomycin is the most commonly associated drug [53, 54].

Conclusions

Drug-Induced Autoimmune Diseases Drug-induced lupus erythematosus is characterized by sudden onset of fever, malaise, myalgia, arthralgia and arthritis, generally several weeks after drug initiation. The skin is affected in approximately a quarter of cases, characterized by erythematous, possibly scaling, or atrophic eruptions often on light-exposed surfaces. Any of the other features of systemic lupus erythematosus, including cytopenia, anemia, hypocomplementemia, hypergammaglobulinemia, positive rheumatoid factor and antiphospholipid antibodies may occasionally be seen. Antinuclear antibodies directed at nuclear histone H2B are often positive in drug-induced lupus erythematosus. Antibodies to double-stranded DNA may also be positive [25, 26]. Linear IgA bullous dermatosis is a rare autoimmune blistering disorder with clinical features that can overlap with bullous pemphigoid and dermatitis herpetiformis. Drug-induced linear

Drugs and drug allergens may elicit a wide variety of clinical manifestations that are based on various pathogenetic mechanisms. However, there still remain many clinical reactions where the pathomechanism is not known and therefore, no validated diagnostic tools are available. The diagnosis of a drug hypersensitivity reaction is based on clinic, history and experience with the drug. In addition, skin [55] and in vitro tests are available and some rules how to perform them have been published [21, 22, 56]. The preparation of drug allergens for skin tests, the performance of in vitro tests [57, 58], correctly conducted provocation tests [23] and the required validation procedures are time-consuming and cost-intensive. Despite these limitations, an investigation into a drug hypersensitivity reaction is important and useful to prevent recurrences and it may help to select safe drug alternatives in future pharmacological therapies.

References 1 Demoly P, Hillaire-Buys D: Classification and epidemiology of hypersensitivity drug reactions. Immunol Allergy Clin North Am 2004;24:345–356, v. 2 Bircher AJ: Symptoms and danger signs in acute drug hypersensitivity. Toxicology 2005;209:201–207. 3 Pichler WJ: Adverse side-effects to biological agents. Allergy 2006;61:912– 920.

4 Barbaud A: Place of excipients in drugrelated allergy. Clin Rev Allergy 1995; 13:253–263. 5 Bigliardi PL, Izakovic J, Weber JM, et al: Anaphylaxis to the carbohydrate carboxymethylcellulose in parenteral corticosteroid preparations. Dermatology 2003;207:100–103. 6 De Smet PAGM: Health risks of herbal remedies. Drug Saf 1995;13:81–93.

Approaching Patients with a Drug Hypersensitivity Reaction

7 Bircher AJ, Hauri U, Niederer M, et al: Stealth triamcinolone acetonide in a phytocosmetic cream. Br J Dermatol 2002;146:531–532. 8 Bircher AJ, Surber C: Unregulated alternative medicine. J Drugs Dermatol 2003;2:58–61. 9 Nigen S, Knowles SR, Shear NH: Drug eruptions: approaching the diagnosis of drug-induced skin diseases. J Drugs Dermatol 2003;2:278–299.

363

10 Roujeau JC: Clinical heterogeneity of drug hypersensitivity. Toxicology 2005; 209:123–129. 11 Bircher AJ: Arzneimittelallergie und Haut. Stuttgart, Thieme, 1996. 12 Demoly P, Kropf R, Bircher A, et al: Drug hypersensitivity: questionnaire. Allergy 1999;54:999–1003. 13 Lerch M, Pichler WJ: The immunological and clinical spectrum of delayed drug-induced exanthems. Curr Opin Allergy Clin Immunol 2004;4:411–419. 14 Pichler WJ: Direct T-cell stimulations by drugs – bypassing the innate immune system. Toxicology 2005;209:95– 100. 15 Bach S, Bircher AJ: Drug hypersensitivity reactions: from clinical manifestations to an allergologic diagnosis. Allerg Immunol (Paris) 2005;37:213–218. 16 Bocquet H, Bagot M, Roujeau JC: Druginduced pseudolymphoma and drug hypersensitivity syndrome (drug rash with eosinophilia and systemic symptoms: DRESS). Semin Cutan Med Surg 1996;15:250–257. 17 Ghislain PD, Roujeau JC: Treatment of severe drug reactions: Stevens-Johnson syndrome, toxic epidermal necrolysis and hypersensitivity syndrome. Dermatol Online J 2002;8:5. 18 Roujeau JC: Treatment of severe drug eruptions. J Dermatol 1999;26:718– 722. 19 Bircher AJ: Arzneimittelallergie; in Schultze-Werninghaus G, Fuchs T, Bachert C, Wahn U (eds): Manuale allergologicum. München-Deisenhofen, Dustri-Verlag, 2004, pp 677–714. 20 Tas E, Pletscher M, Bircher AJ: IgEmediated urticaria from formaldehyde in a dental root canal compound. J Invest Allergol Clin Immunol 2002; 12: 130–133. 21 Romano A, Blanca M, Torres MJ, et al: Diagnosis of nonimmediate reactions to
-lactam antibiotics. Allergy 2004; 59:1153–1160. 22 Torres MJ, Blanca M, Fernandez J, et al: Diagnosis of immediate allergic reactions to
-lactam antibiotics. Allergy 2003;58:961–972. 23 Aberer W, Bircher A, Romano A, et al: Drug provocation testing in the diagnosis of drug hypersensitivity reactions: general considerations. Allergy 2003;58:854–863.

364

24 Johnson DH, Cunha BA: Drug fever. Infect Dis Clin North Am 1996;10:85– 91. 25 Quiceno GA, Cush JJ: Iatrogenic rheumatic syndromes in the elderly. Clin Geriatr Med 2005;21:577–588, vii. 26 Schlienger R, Bircher AJ, Meier CR: Minocycline-induced lupus. Dermatology 2000;200:223–231. 27 Wiik A: Clinical and laboratory characteristics of drug-induced vasculitic syndromes. Arthritis Res Ther 2005;7: 191–192. 28 Spanou Z, Keller M, Britschgi M, et al: Involvement of drug-specific T cells in acute drug-induced interstitial nephritis. J Am Soc Nephrol 2006;17:2919– 2927. 29 Bircher AJ: Drug-induced urticaria and angioedema. Eur J Dermatol 1999;9: 657–663. 30 Wolf R, Orion E, Marcos B, et al: Lifethreatening acute adverse cutaneous drug reactions. Clin Dermatol 2005; 23: 171–181. 31 Prussick R, Knowles S, Shear NH: Cutaneous drug reactions. Curr Probl Dermatol 1994;6:85–122. 32 Wolkenstein P, Revuz J: Drug-induced severe skin reactions. Incidence, management and prevention. Drug Safety 1995;13:56–68. 33 Descamps V, Bouscarat F, Laglenne S, et al: Human herpesvirus 6 infection associated with anticonvulsant hypersensitivity syndrome and reactive haemophagocytic syndrome. Br J Dermatol 1997;137:605–608. 34 Descamps V, Valance A, Edlinger C, et al: Association of human herpesvirus 6 infection with drug reaction with eosinophilia and systemic symptoms. Arch Dermatol 2001;137:301–304. 35 Kano Y, Hiraharas K, Sakuma K, et al: Several herpesviruses can reactivate in a severe drug-induced multiorgan reaction in the same sequential order as in graft-versus-host disease. Br J Dermatol 2006;155:301–306. 36 Seishima M, Yamanaka S, Fujisawa T, et al: Reactivation of human herpesvirus (HHV) family members other than HHV-6 in drug-induced hypersensitivity syndrome. Br J Dermatol 2006; 155: 344–349. 37 Scherer K, Bircher AJ: Hypersensitivity reactions to fluoroquinolones. Curr Allergy Asthma Rep 2005;5:15–21.

38 Hyson C, Sadler M: Cross-sensitivity of skin rashes with antiepileptic drugs. Can J Neurol Sci 1997;24:245–249. 39 Brockow K, Christiansen C, Kanny G, et al: Management of hypersensitivity reactions to iodinated contrast media. Allergy 2005;60:150–158. 40 Kanny G, Pichler W, Morisset M, et al: T-cell-mediated reactions to iodinated contrast media: evaluation by skin and lymphocyte activation tests. J Allergy Clin Immunol 2005;115:179–185. 41 Van Renterghem D, De Vries EA: Pseudo-lymphoma and anticonvulsant hypersensitivity syndrome during the use of anti-epileptic agents. Tijdschr Geneeskd 1997;53:399–403. 42 Bocquet H, Bagot M, Roujeau JC: Druginduced pseudolymphoma and drug hypersensitivity syndrome (drug rash with eosinophilia and systemic symptoms: DRESS). Semin Cut Med Surg 1996;15:250–257. 43 Roujeau JC, Bioulac-Sage P, Bourseau C, et al: Acute generalized exanthematous pustulosis. Analysis of 63 cases. Arch Dermatol 1991;127:1333–1338. 44 Häusermann P, Scherer K, Weber M, et al: Ciprofloxacin-induced acute generalized exanthematous pustulosis mimicking bullous drug eruption confirmed by a positive patch test. Dermatology 2005;211:277–280. 45 Sidoroff A, Halevy S, Bavinck JN, et al: Acute generalized exanthematous pustulosis – a clinical reaction pattern. J Cutan Pathol 2001;28:113–119. 46 Häusermann P, Harr T, Bircher AJ: Baboon syndrome resulting from systemic drugs: is there strife between SDRIFE and allergic contact dermatitis syndrome? Contact Dermatitis 2004;51: 297–310. 47 Häusermann P, Bircher AJ: SDRIFE – another acronym for a distinct cutaneous drug exanthema. Do we really need it? Dermatology 2007;214:1–2. 48 Lee AY: Fixed drug eruptions. Incidence, recognition, and avoidance. Am J Clin Dermatol 2000; 1:277–285. 49 Shiohara T, Mizukawa Y, Teraki Y: Pathophysiology of fixed drug eruption: the role of skin-resident T cells. Curr Opin Allergy Clin Immunol 2002; 2:317–323. 50 Roujeau JC, Stern RS: Severe adverse cutaneous reactions to drugs. N Engl J Med 1994;331:1272–1285.

Bircher

51 Auquier-Dunant A, Mockenhaupt M, Naldi L, et al: Correlations between clinical patterns and causes of erythema multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis: results of an international prospective study. Arch Dermatol 2002; 138:1019–1024. 52 Doyle MK, Cuellar ML: Drug-induced vasculitis. Expert Opin Drug Saf 2003; 2:401–409. 53 Navi D, Michael DJ, Fazel N: Drug-induced linear IgA bullous dermatosis. Dermatol Online J 2006;12:12.

54 Onodera H, Mihm MC Jr, Yoshida A, et al: Drug-induced linear IgA bullous dermatosis. J Dermatol 2005; 32:759– 764. 55 Brockow K, Romano A, Blanca M, et al: General considerations for skin test procedures in the diagnosis of drug hypersensitivity. Allergy 2002;57:45– 51. 56 Romano A, Di Fonso M, Papa G, et al: Evaluation of adverse cutaneous reactions to aminopenicillins with emphasis on those manifested by maculopapular rashes. Allergy 1995;50:113–118.

57 De Weck AL, Sanz ML: Cellular allergen stimulation test (CAST) 2003, a review. J Investig Allergol Clin Immunol 2004;14:253–273. 58 Pichler WJ, Tilch J: The lymphocyte transformation test in the diagnosis of drug hypersensitivity. Allergy 2004;59: 809–820. 59 Arnold AW, Häusermann P, Bach S, et al: Recurrent flexural exanthema (SDRIFE or baboon syndrome) after administration of two iodinated contrast media. Dermatology 2007; 214:89– 93.

Dr. Andreas J. Bircher Allergy Unit, Department of Dermatology University Hospital, Petersgaben 4 CH–4031 Basel (Switzerland) Tel. +41 61 265 2525, Fax +41 61 265 4885 E-Mail [email protected]

Approaching Patients with a Drug Hypersensitivity Reaction

365

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 366–379

Place of Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions Annick Barbaud Dermatology Department, Fournier Hospital, Nancy, France

Abstract Skin tests with drugs can be helpful in determining the cause of cutaneous adverse drug reactions (CADR). Patch tests and prick tests can be done with any commercialized form of a drug, intradermal tests (IDT) normally rely on the availability of a sterile solution used for parental applications. There has been some experience with different concentrations and techniques, but a complete standardization has not been done with these tests. Prick tests and IDT are used to investigate immediate hypersensitivity, patch tests and IDT with delayed readings are helpful for symptoms appearing delayed. They can be positive in different skin symptoms, with or without involvement of internal organs (e.g. hepatitis). A great variety of drugs can be tested, including antibiotics, antiepileptics, drugs used in oncology, drugs used for diagnosis etc. Except for IDT, skin tests with drugs rarely induce adverse reactions. False positive results can occur and should be considered, if a product for skin tests is evaluated for the first time. The negative predictive value of skin tests with drugs is a crucial point that will have to be studied in detail in the near future. Copyright © 2007 S. Karger AG, Basel

Cutaneous adverse drug reactions (CADR) are a frequent problem in clinical medicine. Since patients are often on multiple drug regimens, it is often difficult to pinpoint the relevant drug from history alone. Beside clinical and chronological parameters, drug skin testing with the suspected compound has been reported to be helpful in determining the cause of a CADR [1–10] and in

studying the pathophysiological mechanisms involved in these reactions [1, 7, 11–13]. The results of skin tests depend on the drug tested and the clinical features of the initial CADR [2, 5, 6]. Drug patch tests and prick tests can be done with any commercialized form of drugs while intradermal tests (IDT) need to be performed with an injectable form or with a pure and sterile preparation of the drug.

How to Perform Drug Skin Tests?

Different guidelines for performing skin tests (patch tests, prick tests and IDT) with drugs in the investigation of cutaneous adverse drug reactions (CADR) have been published [3, 14]. Unfortunately, in reading recent literature, many other techniques have been used in performing these drug tests, especially IDT, and their readings. Consequently, it is impossible to compare the results from one center to another. Thus, it is not easy to determine the thresholds of specificity of the IDT. It is advised to perform drug skin tests during the 6 months following the CADR, as we do not know whether positive results will persist, and whether some drug reactivities last longer [1, 3, 15].

Patch Tests Drug Concentration in Drug Patch Tests Since the threshold of sensitivity for many pure substances is not yet determined, in drug patch tests, it is advised to use a 10% concentration in petrolatum and if necessary in other vehicles, although for some drugs, smaller concentrations may be sufficient [3]. Using a pulverized tablet, 30% is the highest concentration possible to get a homogeneous dilution in petrolatum, in water or in alcohol [3]. When the commercialized form of the drug is used, it is advised to use a 30% concentration of the final product. Tests with acyclovir, carbamazepine or pseudoephedrin have occasionally reproduced the generalized CADR symptoms during patch testing (table 1) [1, 2]. Therefore, it is recommended that patch tests are performed, first diluted at 1% and when negative, up to 10% (either with the commercialized form of the drug or the pure substance). Moreover, to avoid false positives, some drugs have to be tested with higher dilutions. The content of the capsules of celecoxib (Celebrex
) should be tested at 5% or at 10% in petrolatum and not with higher concentrations [16]. Desloratadine has to be tested at 1% in petrolatum (table 1) [17]. Colchicin at 10% induces false-positive results, the threshold of specificity is unknown [18]. Captopril in commercialized forms diluted at 1% and chloroquine in commercialized forms diluted at 30% in petrolatum induce sometimes false-positive results [18]. Misoprostol in commercialized forms has to be diluted at 1% in petrolatum [18]. In investigating a photosensitivity reaction induced by a drug, both drug patch tests and drug photopatch tests with the responsible drug have to be performed. The irradiation for drug photopatch tests is performed on day 1, or for practical reasons can be performed on day 2 with a 5 J/cm2 UVA irradiation [2, 3]. It has been demonstrated that photoscratch patch tests are more irritating and do not have a better value than photopatch tests [19].

Material Preparation for Drug Patch Tests If the commercialized form of the drug is tested, pills must have their coating removed. The substance has to be smashed to a very fine powder. The powder contained in capsules is tested at 30% in petrolatum and at 30% in water. The gel jacket portion of the capsules is moistened and tested as is. Liquid preparations are tested both as is and diluted at 30% in water. With commercialized forms of the drugs, each preparation is done for only 1 patient and kept no longer than 24 h [1, 3]. Whenever possible, preservatives, coloring agents and excipients should also be tested, undiluted or diluted at 10% in petrolatum, or in the vehicles and concentrations usually proposed for testing allergic contact dermatitis. Vehicles Used for Drug Patch Tests The best vehicle to prepare a drug patch test has not yet been determined. There are a few studies in which each drug has been tested, diluted in different excipients. Comparing the results obtained with drugs diluted in petrolatum, water or ethyl alcohol, there were no great differences found [2, 3]. Petrolatum seemed to be convenient in most of the cases. Steroid hormones have to be tested diluted in alcohol as false-negative results have been observed in testing estrogens diluted in water or petrolatum [3, 20]. We also observed a maculopapular rash (MPR) with positive patch tests with corticosteroids diluted in water and alcohol but negative when diluted in petrolatum [A. Barbaud, unpubl. data]. Positive patch tests were obtained with ganciclovir diluted at 20% in water but negative with the drug diluted at 20% in petrolatum [21]. In fixed drug eruptions (FDE) due to cotrimoxazole, Ozkaya-Bayazit et al. [22] reported on 25 positive reactions in 27 patients, in situ tested with cotrimoxazole, diluted in dimethylsulfoxide (DMSO), while 19 patients had negative results when the drug was diluted in petrolatum. There are no extensive studies concerning the value of other vehicles used in drug patch tests.

Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions

367

Table 1. Drugs with reported positive patch tests in investigating systemic cutaneous adverse drug reactions Drug or drug class

References

Concentrations used and controls (when available)

Relapse of the CADR due to drug patch tests

Acyclovir

31, 47

as is, 20, 10 and 1% in petrolatum (commercialized form?) [31]; acyclovir 10% in petrolatum [47]

2

Allylisopropylacetylurea

84

Amoxicillin

2, 4, 6, 8–11, 10% in petrolatum seems to be specific 31, 38, 51–53

immediate reactions in case of anaphylaxis


-Lactam antibiotics

2, 31, 36

immediate reactions in case of anaphylaxis

Captopril

2, 85

at 1 and 10% in petrolatum [85]; false-positive results [2, 18]

Carbamazepine

61–70, 125

10% in petrolatum

Cefcapene pivoxil

86

10 and 1% in petrolatum

Celecoxib; frequent false-positive reactions

16, 87

has to be tested at 10 or 1%; higher concentrations can induce false-positive results [16] false-positive reactions; the threshold of specificity has to be determined [18]

Chloroquine; no true-positive reaction Chlorphenamine

89

Clindamycin

31

Clobazam

88

Codeine

90

diluted at 1 and 5% in petrolatum (2 negative controls) [90]

Colchicine; false-positive results

18

diluted at 10% in petrolatum, 80% of 29 negative controls developed false-positive results [18]

Corticosteroids

2, 54–56

if negative, have to be diluted in ethyl alcohol [1]

Cotrimoxazole sulfamethoxazole/trimethoprim

22, 25

10, 20 or 50% in dimethylsulfoxide (DMSO) but frequently negative when diluted in petrolatum [20]

Cyamemazine

2

30% petrolatum (commercialized form)

Desloratadine

17

diluted at 10% in petrolatum in 8/10 volunteers, seems to be specific when tested diluted at 1% in petrolatum (7 negative controls) [17]

Diazepam

52

Diclofenac

28, 91

1% in petrolatum [28]

Diltiazem

2, 71–73

10% in petrolatum

Enoxoparin

2

pure

Estrogens

20

if negative, tested diluted in alcohol

Famciclovir

83

50% in petrolatum (commercialized form)

Fluindione

92

5 and 30% in petrolatum (commercialized form), doubtful in water

Fluoroquinolones

2, 93, 94

30% in petrolatum or water (commercialized form)

368

61

20% in petrolatum (commercialized form?) 15 negative controls

88

at 1% petrolatum in an anaphylactic shock [28]

Barbaud

Table 1 (continued) Drug or drug class

References

Concentrations used and controls (when available)

Fusafungine

2

30% in petrolatum (commercialized form)

Ganciclovir

21

as is, 20% in petrolatum

Gentamycin

31

Heparin derivatives

12, 49, 58, 59

non-specific results do to sensitization to excipients (benzyl alcohol) [18]

Hydroxyzine

2, 74

10% in petrolatum

Isoniazid

31, 96

at 50% in petrolatum (10 negative controls) [96]

Lamisil® (terbinafine)

2

as is, detail of the test remained negative [2]

Meprobamate

2

30% in petrolatum (commercialized form)

Metamizole

97, 98

at 1 and 10% in petrolatum

Metronidazole

31

Mexiletine hydrochloride

99

diluted at 10 and 20% in petrolatum [99]

Misoprostol; no true-positive reaction

16

false-positive results in 9/10 negative controls at the day 2 reading, no false-positive results at the day 4 reading or when Cytotec® was diluted at 1% in petrolatum [16]

Nimesulide

86

10% in petrolatum (commercialized form?) [86]

Nystatin

87

at 10% in petrolatum (10 negative controls) [87]

Omeprazole; no true-positive reaction

18

30% in petrolatum or water (commercialized form)

Oxicams

104, 122, 123 1% in petrolatum [104] or 10% in petrolatum [123]

Paracetamol (acetaminophen)

103

relapse of an AGEP? 103]

Pristinamycin

2, 40, 60

2

Pseudoephedrine

2, 75, 76

tested at 1% in petrolatum to avoid any relapse of the CADR

Radiocontrast medium

44–47

pure

Rifampicin

119

Teicoplanin

37

Tetrazepam

2, 26, 77–81

30% in commercialized form, 10% in petrolatum

Triamcinolone

2

30% in petrolatum (commercialized form)

Valaciclovir

21, 83

30% in petrolatum, water or alcohol [83], commercialized form; as is, 20, 10, 1% commercialized form? [21]

Vancomycin

37

at 0.005% in water

Vitamin K1

105

10 mg/ml in olive oil

Vitamin K3

105

10 mg/ml in olive oil

4% water

Relapse of the CADR due to drug patch tests

to study cross-reactions with acyclovir

2

97

2, 75, 76

20 negative controls [37]

2

20 negative controls [37]

Concentrations and excipients used in preparing the material for these positive tests are reported as drugs published as being able to induce a relapse of the adverse reaction during patch testing.

Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions

369

Which Site to Test? Under chambers, patch tests are performed on the upper back but it could also be of value to test on the most affected site of the initial CADR. In FDE, it is well known that patch tests [23] or repeated application tests [24] with the suspected drug are positive only when performed on the residual pigmented skin site of the CADR but not when applied on the non-previously affected skin of the back. Testing in the affected area could also be of value in other forms of CADR, as reported in toxic epidermal necrolysis in 1 case with cotrimoxazole [25] or with tetrazepam in a MPR [26]. Readings of Drug Patch Tests The results of patch testing are reported according to the International Contact Dermatitis Research Group (ICDRG) criteria for patch test reading [27]. As drug patch tests can induce immediate positive reactions, especially with
-lactam antibiotics, these tests have to be read at 20 min in patients who have developed urticaria or anaphylactic shock. Immediate reactions on patch tests have been reported with
-lactam antibiotics, neomycin, gentamycin, bacitracin [1] and recently with diclofenac [28]. Immediate positive results can be associated with generalized anaphylactic reactions. Prick Tests Prick tests are performed on the volar forearm with the commercialized form of the drug. Whenever possible, both the pure drug and excipients have to be tested. If urticaria has developed in the patient, prick tests are done using sequential increasing dilutions. From recent studies, we can consider that the reading of prick tests has been standardized [3, 14, 29]. Reactions are considered positive when a wheal with a diameter larger than that of the negative control (0.9% saline), with a diameter 63 mm and with a surrounding erythema [3, 14, 29], is present 20 min later.

370

In all cases, a delayed reading has to be performed 24 h after the initial tests, because delayed positive results can occur in prick tests [2, 3, 11]. Intradermal Tests IDT are performed only when prick tests show negative results 20 min after testing with the suspected drug. As they could re-induce the adverse drug reactions, IDT are not performed in patients with Stevens-Johnson syndrome, Lyell’s syndrome or leukocytoclastic vasculitis on histological examination and have to be considered with caution in drug reactions with eosinophilia and systemic symptoms (DRESS) [3]. As they can induce a relapse of the CADR, both immediate [3, 30] and delayed [3] reactions in 10% of the tested patients, IDT have to be done under hospital surveillance. It is necessary to obtain sterile forms of the drug for IDT. Even if it is not recommended in guidelines [3, 14], some authors also do IDT with non-injectable drugs. They use the powder contained in capsules or obtained by removing the external layer of tablets with a scalpel. After weighing the powder, they prepare solutions under a laminar flow and sterilize them by filtration through single-use devices [29]. It is recommended to prepare dilutions under a laminar flow, no more than 2 h before administration. All necessary precautions have to be taken and emergency treatment should be ready. The IDT are performed using a sterile solution of the suspected drug, diluted sequentially in saline. A negative control can be performed with saline (0.9% NaCl) or the excipient used for dilutions. Some authors also perform a positive control with histamine at 1 mg/ml [29]. The site of injection (back, forearm, extensor surface of the arm), the injected volume (0.02 ml [29, 30], 0.04 ml [3] to 0.05 ml [14]) or a volume which leads to a certain diameter of the injection papula (i.e., 3 mm [14], 1 to 2 mm [32]), and the time for immediate reading (in most of the cases 20 or 30 min) are not

Barbaud

yet standardized [3, 14, 33–35]. Moreover, criteria for readings in case of immediate positive reactions are not standardized. According to different authors or guidelines, IDT are considered positive on immediate readings when the diameter of the reaction is equal or more than twice the diameter of the injection papula [3], 610 mm [2], 68 mm [32]; for the injected wheal (from 1 to 2 mm large), an increase in wheal diameter by 3 mm or greater is expected [14, 29, 35]. In most of the papers, delayed reactions, occurring after 24 h or later, are considered positive when there is an infiltrated erythematous reaction [3, 29, 33]. As the techniques used in performing IDT are not standardized, thresholds for specificity are impossible to compare from one group to another [18, 29, 33]. Considering the two papers written on guidelines in doing drug skin tests [3, 14], it could be advised to determine the volume, whatever the site of injection, which leads to a diameter of the injection wheal of 3 mm. Read at 20 min, IDT would be considered as having positive results when the diameter of the reaction would be 66 mm.

Value of Drug Skin Tests

In populations with various or unknown imputabilities of the tested drug, drug patch tests were positive in 89/826 patients (10.8%) [31]. In patients with a high imputability that a certain drug causes the CADR, drug (photopatch) patch tests had positive results in 50% of 108 patients [4]. Lammintausta and Kortekangas-Savolainen [31], with an unknown imputability of the tested drugs, obtained 10 (1.1%) positive reactions on 935 prick tests. Except for
-lactam antibiotics, the usefulness of prick tests and IDT in determining the cause of CADR has not been evaluated as much as the usefulness of drug patch tests. In some papers the specificity of the IDT cannot be validat-

ed as negative controls are lacking. With a high imputability for a drug having induced a CADR, positive results were obtained in 24% of 46 cases with prick tests and in 64% of the 30 patients undergoing IDT [3]. Among 60 patients with CADR and negative patch tests with the suspected drug, 35 (58%) had positive results on IDT [3]. Among 94 patients with a suspected delayed sensitization to
-lactam antibiotics, 36% had both positive patch tests and IDT but 8 had positive IDT with negative patch tests [36]. In a CADR due to vancomycin, prick tests as well as IDT done with glycopeptide antibiotics remained negative even on delayed readings, while drug patch tests were positive and specific (20 negative controls) [37]. This investigation emphasizes that delayed reactions to IDT are not always sufficient to investigate a CADR due to delayed hypersensitivity to drugs, but that drug patch tests can supplement the analysis and be positive in spite of a negative IDT. Usefulness of Drug Skin Tests according to the Clinical Features of the CADR Patch tests are of value in determining the responsible drug in generalized eczema, systemic contact dermatitis, baboon syndrome (recently also called SDRIFE, see chapter of Bircher, pp 352–365], MPR [1, 3, 4, 9, 11, 38], AGEP [39] and FDE [23, 24]. They also seem to be of value in DRESS. Photopatch tests may be useful in studying drug photosensitivity. On the other hand, they are of less value in investigating urticaria [1, 2], Stevens-Johnson or Lyell’s syndromes [39], pruritus or vasculitis [1, 2]. In urticaria, prick tests can have immediate positive results as demonstrated with
-lactam antibiotics, anesthetics or synergistins [40]. Even if this is rare, recent papers confirm that prick tests may also have delayed positive results with vaccines,
-lactam antibiotics and also synergistins [4, 40], pseudoephedrin [3] or minocyclin [Barbaud, unpubl. data].

Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions

371

Immediate positive results on IDT are not unusual in case of urticaria, angioedema or anaphylactic shock. It has been mainly studied for
-lactam antibiotics, e.g. penicillins [34], cephalosporins [29] and anesthetics [32]. Recent studies have confirmed the value of IDT in investigating delayed hypersensitivity to drugs. In MPRs, delayed positive reactions of IDT have been reported with
-lactam antibiotics [3, 33], synergistins [40], platinum salts [41– 43] or radiocontrast medias [44–48]. Delayed positive results of IDT have also been reported in investigating localized eczema or MPR due to heparins [3, 49, 50]. Usefulness of Drug Skin Tests according to the Tested Drug The usefulness of drug skin tests also depends on the tested drug. The most frequent reports of positive drug patch tests are related to the following drugs (table 1):
-lactam antibiotics, especially amoxicillin [2–4, 6, 8–11, 31, 36, 38, 51–53], cotrimoxazole [22, 25], corticosteroids [2, 3, 54– 57], heparin derivatives even though IDT have highest usefulness [12, 49, 58, 59], pristinamycin (20 positive patch tests/29 cases; 69%) [2, 40, 60], carbamazepine [2, 4, 61–70], diltiazem [2, 71–73], hydroxyzine [2, 74], pseudoephedrine [1, 2, 75, 76] or tetrazepam [2, 26, 77–81]. Even though they are not classical drugs, recent studies emphasize that radiocontrast medium can induce delayed CADR. In such cases, skin tests can be of value [46–48]. Even though IDT with delayed readings are more frequently positive, in these delayed reactions, patch tests could also be of value [44, 46, 47]. The value of patch tests in delayed hypersensitivity to contrast medium varied from 2/15 [47] to 46/58 [82], but IDT have a better value as 8/15 patients had delayed positive reactions on IDT [47]. Miscellaneous drugs have been reported as yielding CADR with positive drug patch tests [2, 4, 31]: acyclovir and valacyclovir [21, 83], allylisopropylacetylurea in a FDE [84], captopril at 1 and

372

10% in petrolatum [2, 85], cefcapene pivoxil [86], celecoxib [12, 87], clindamycin [31], clobazam [88], chlorphenamine (20% in petrolatum) [89], codeine diluted at 1 and 5% in petrolatum (2 negative controls) [90], cotrimoxazole [22, 25], cyamemazine [23], diazepam [3, 52], diclofenac [28, 91], enoxoparin [2, 3], estrogens [20], fluindione [92], fluoroquinolones [2, 3, 93–95], fusafungine [2], gentamycin [31], isoniazide at 50% in petrolatum (10 negative controls) [31, 96], Lamisil
[2, 3], meprobamate [2, 3], metamizole at 1 and 10% in petrolatum [97, 98], metronidazole [31], mexiletine hydrochloride diluted at 10 and 20% in petrolatum [99], minocyclin [100], nimesulide [101], nystatin at 10% in petrolatum (10 negative controls) [102], paracetamol (acetaminophen) [103], piroxicam [104], vancomycin at 0.005% in water [37] and vitamin K1 (10 mg/ml in olive oil) and vitamin K3 [105]. Most of the papers published on the usefulness of prick tests and IDT concern CADR due to
-lactam antibiotics [3, 15, 24, 28, 29, 33–35, 53, 68–70, 106–108]. Of 998 skin tests done with
lactam, 14.7% of the patients had positive results [30]. Among 166 subjects with aminopenicillin-associated MPR, patch tests and delayed-reading IDT with amoxicillin and ampicillin were positive in 52.4 and 54.2%, respectively [53]. In immediate reactions to
-lactam antibiotics among 290 patients, the sensitivity of prick tests and IDT was 22% for benzylpenicilloyl poly-L-lysine, 21% for minor determinants mixture, 43% for amoxicillin, and 33% for ampicillin [108]. Prick tests and IDT have also been performed with other drugs such as local anesthetics (all negative in 236 patients, with only one localized reaction among 236 subcutaneous challenges [109, 110]), carboplatin (126 patients tested) [42] or platin salts [43], clindamycin (2 delayed positive reactions/31 tested patients) [111], corticosteroids [72, 112], dexchlorpheniramine (immediate positivity in 1 case) [113], estradiol [3], gadoterate meglumine [114], heparin

Barbaud

derivatives [38, 49–52], insulin [115, 116], quinolones [95, 117], radiocontrast media [44, 46–48], rifampicin (3 cases and 24 controls) [118, 119], and synergistins [40] with 3 positive results on 9 prick tests and 4/5 delayed positive results on IDT. Drug skin tests do not seem to have any value in investigating intolerance to non-steroidal antiinflammatory drugs, flush due to corticosteroids, angioedema due to angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, adverse reactions due to hyaluronic acid [120] or urticaria due to bupropion [121]. Usefulness of Drug Skin Tests to Study Cross-Reactivity between Drugs Skin tests can help to study the ability of drugs to elicit symptoms due to cross-reactivity, e.g. between
-lactam antibiotics [36] such as penicillins and cephalosporins [29]. In MPR due to diltiazem, patch tests with other calcium channel blockers are rarely positive, and no cross-reactions between dihydropyridine calcium channel blockers and ‘non-dihydropyridine’ calcium channel blockers are found (verapamil, diltiazem) [73]. No or rare cross-reactivity between tetrazepam and diazepam was detected [81]. With patch tests and drug re-administration, cross-reactions between acyclovir, valacyclovir and famciclovir are frequent in patients suffering from CADR due to acyclovir [21, 83]. It could be due to common chemical structure, as the 2-aminopurine nucleus is also found in ganciclovir but not in foscarnet or cidofovir [83]. Cross-reactions between synergistins are very frequent as demonstrated in 29 cases of CADR due to pristinamycin [40]. Cross-reactions between synergistins occurred in patients sensitized to pristinamycin, in 9/22 with virginiamycin and in 7/8 cases with dalfopristin-quinupristin. In piroxicam reactions, the profile of cross-sensitization may differ in distinct clinical features of the CADR. In photosensitization, photopatch tests with piroxicam were positive in 27/31. But there

was no cross-reaction with tenoxicam and lenoxicam and only 1/31 had positive photopatch tests with meloxicam. In contrast, in 1 case with a photosensitization to piroxicam, all photopatch tests done with piroxicam, tenoxicam, droxicam and meloxicam were positive [122]. However, in 8 patients with FDE, cross-reactions between oxicams were very frequent [123]. There is cross-reactivity among pseudoephedrin and different sympathomimetic drugs, e.g. phenylephrine or ephedrine [124]. Drug patch tests can also be used to study reactions between the native drug and its metabolites, as done for carbamazepine and its main metabolite carbamazepine oxide [125]. With patch tests, prick tests and IDT, crossreactions have also been demonstrated between radiocontrast medias [47], insulins [116] or heparins and heparinoids [2, 3, 49, 58, 59] and platin salts [43]. With all these drugs they differ from one patient to another and have absolutely to be tested to find a replacement drug belonging to the same therapeutic class.

Negative Predictive Value of Drug Patch Tests

The predictive value of a negative drug skin test is unknown. There are a lot of published cases of CADR with negative drug patch tests but positive delayed IDT [2, 3]. This was also reported with rifampicin [119]. In a recent study [J. Waton and A. Barbaud, pers. commun., European Academy of Allergy and Clinical Immunology, Vienna, 2006], we demonstrated that in patients having had a CADR 18/140 drug re-administrations (13%) were positive in patients who had negative patch tests, prick tests and IDT. Lammintausta and Kortekangas-Savolainen [31] performed drug challenges in 229 patients who had negative patch tests or prick tests. They had a positive challenge in 22/229 patients (9.6%). Considering challenges in
-lactam hypersensitivity, Romano et al. [126] did not observe

Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions

373

any positive challenge in 49 patients having negative patch tests, prick tests, IDT and specific IgE against
-lactams. In our study [J. Waton and A. Barbaud, EAACI, Vienna, 2006], 3/18 (17%) of the patients with negative patch tests, prick tests and IDT had a positive oral challenge. With negative prick tests and IDT, without patch tests, oral provocation tests to
-lactam antibiotics were positive in 49/89 (55%) [127] but in only 4/153 (2.6%) [3], or in 35/416 patients (8.4%) in other studies [128]. With macrolides, 2 of 6 (33%) patients with drug reactions but negative patch tests, prick tests and IDT had a positive challenge (own results), while 2/2 patients with drug reactions (urticaria) to macrolides had a negative IDT and a negative provocation to the eliciting macrolide (spiramycin and josamycin) [129]. With paracetamol, 3/18 (17%) drug challenges were positive in our patients. With corticosteroids 1/9 cases (11%) had a positive challenge, as Padial et al. [112] observed 19 positive reactions among 30 patients (63%) with CADR due to corticosteroids. In delayed reactions occurring with contrast medium, the negative predictive value seems to be low. Indeed, even with negative patch tests, prick tests and IDT, 5/12 patients had a relapse during the re-administration of a contrast medium, that had negative skin tests to the reapplied contrast medium [47]. Therefore, in non-severe CADR with a delayed hypersensitivity mechanism, such as MPR, it is advisable to perform first drug patch tests, then, if negative, the other drug skin tests with immediate and delayed readings. In immediate hypersensitivity, prick test and IDT can be helpful. There is no value for the patient in performing patch tests when positive delayed reactions have been previously obtained in testing the same drug with IDT. In such cases, drug patch tests are very frequently also positive.

374

Safety of Drug Skin Tests

Drug patch tests can re-induce the CADR (table 1). This has been reported with acyclovir [2, 3], amoxicillin and other
-lactam antibiotics (with immediate reactions in case of anaphylaxis), carbamazepine [61], clobazepam [88], diclofenac at 1% petrolatum in an anaphylactic shock [28], hydroxyzine [2], metamizole [97], maybe due to paracetamol (acetaminophen) in an acute generalized exanthematous pustulosis [103], pristinamycin [40], pseudoephedrin [2, 3, 75, 76] and triamcinolone [2, 3]. The relapse of the CADR is more frequent with IDT [3]. There is a risk of inducing a relapse of initial CADR especially when performing IDT. Among 30 IDT performed, 3 minor incidents were observed [3]. With
-lactam antibiotics, side effects would occur in 0.1–2% of tested patients [30]. Among 998 skin tests, in patients with a suspected hypersensitivity to
-lactam antibiotics, 15% of them had positive skin tests and 13 (8.8%) had a systemic reaction during these tests. Anaphylaxis and a short delay of !1 h between the last drug intake and the reaction were more frequent in reactors than in 135 patients who had positive skin tests without systemic reactions [30]. Among 126 patients tested with platinum salts, only one experienced erythroderma, chills, chest discomfort and mild dyspnea a few minutes after IDT to platinum salts [42]. Even with negative results, a relapse of a pruritic rash occurred following a prick test with pristinamycin [40]. Thus, prick and IDT can induce a systemic reaction, although their results are negative.

Relevance and Specificity of Drug Patch Tests

One crucial point is the interpretation of the results of skin tests with drugs. As a negative control, the vehicle used to dilute the material for patch tests should be tested: false-positive drug

Barbaud

patch tests are rare but can be observed due to a sensitization to ethyl alcohol but also to petrolatum [130]! False-positive results are rather frequent in performing IDT [18], but false-positive or nonrelevant results can also be obtained in drug patch testing, namely with the commercialized form of drugs containing sodium laurylsulfate in their formulation, with colchicine diluted at 10% in petrolatum (80% of 29 negative controls developed false-positive results), with Cytotec
containing misoprostol diluted at 30% in petrolatum at the day 2 reading [18] or with captopril, chloroquine and omeprazole [18]. Kleinhans et al. [16] reported on the irritant reactions observed while patch testing Celebrex
(celecoxib) with concentrations higher than 10% in petrolatum (17/19 false-positive results on the 48-h readings) but if a 5 or 10% solution in petrolatum was used, patch tests with Celebrex
seem to be specific. False-positive results can occur in testing desloratadine diluted at 10% in petrolatum as observed in 8/10 volunteers [17]. Diluted at 1% in petrolatum, desloratadine did not induce any irritation or positive reaction in 7 other volunteers. This emphasizes the necessity to compare skin test results of patients with those obtained in negative controls [2, 3, 18]. A drug patch test may be positive due to a contact dermatitis to a drug or excipient without any relevance to the CADR [18]. Drug patch tests can be positive for a commercialized form of a drug and for one component of this product (iodine or avocado oil), but with a well-tolerated oral re-administration of the drug. Drug patch tests are positive because the patients have previously developed a contact eczema due to an antiseptic containing iodine or to avocado oil contained in a wound-healing ointment [18]. Similarly, nonrelevant positive drug patch tests with commercialized forms of drugs have been reported where preservatives or stabilizers were causative, but not the drug itself, e.g. sodium sulfite in certain

formulations [18] or benzyl alcohol in patients with heparin intolerance. These patients had previously developed contact allergy either to sulfites or to benzyl alcohol but had a good tolerance to systemically administered drugs containing these excipients. False-positive prick tests can be observed in testing with codeine derivatives or pure spiramycin. With most of the drugs, the used dosage leads to irritative, immediate reaction in performing IDT [18]. But as the practices used for performing IDT and their readings are not standardized, it is impossible at the moment to compare the thresholds for specificity from one study to another.

Conclusion

Drug skin tests can be helpful in determining the cause of a CADR. They induce only rarely adverse reactions and they can be done with any commercialized form of a drug. IDT have a greater value, but it will be necessary to standardize the techniques used, to determine and compare their thresholds of specificity. Falsepositive results can occur and should be considered by testing new products. The negative predictive value of drug skin tests is a crucial point that will have to be studied in more detail in the near future.

Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions

375

References 1 Barbaud A: Drug patch testing in systemic cutaneous drug allergy. Toxicology 2005;209:209–216. 2 Barbaud A, Reichert-Penetrat S, Trechot P, Jacquin-Petit MA, Ehlinger A, Noirez V, Faure GC, Schmutz JL, Bene MC: The use of skin testing in the investigation of cutaneous adverse drug reactions. Br J Dermatol 1998;139:49–58. 3 Barbaud A, Gonçalo M, Bruynzeel D, Bircher A: Guidelines for performing skin tests with drugs in the investigation of cutaneous adverse drug reactions. Contact Dermatitis 2001;45:321–328. 4 Barbaud A, Bene MC, Faure G, Schmutz JL: Tests cutanés dans l’exploration des toxidermies supposées de mécanisme immuno-allergique. Bull Acad Natle Med 2000;184:47–63. 5 Osawa J, Naito S, Aihara M, Kitamura K, Ikezawa Z, Nakajima H: Evaluation of skin test reactions in patients with non-immediate type drug eruptions. J Dermatol 1990;17:235–239. 6 Bruynzeel DP, Van Ketel WG: Patch testing in drug eruptions. Semin Dermatol 1989;8:196–203. 7 Barbaud A: Tests cutanés dans l’investigation des toxidermies: de la physiopathologie aux résultats des investigations. Thérapie 2002;57:258–262. 8 Bruynzeel DP, Von Blomberg-van der Flier M, Scheper RJ, Van Ketel WG, De Haan P: Penicillin allergy and the relevance of epicutaneous tests. Dermatologica 1985;171:429–434. 9 Bruynzeel DP, Van Ketel WG: Skin tests in the diagnosis of maculopapular drug eruptions. Semin Dermatol 1987;6: 119–124. 10 Romano A, Di Fonso M, Pietrantonio F, Pocobelli D, Giannarini L, Del Bono A, Fabrizi G, Venuti A: Repeated patch testing in delayed hypersensitivity to
-lactam antibiotics. Contact Dermatitis 1993;28:190. 11 Barbaud A, Bene MC, Schmutz JL, Ehlinger A, Weber M, Faure GC: Role of delayed cellular hypersensitivity and adhesion molecules in maculopapular rashes induced by amoxicillin. Arch Dermatol 1997;133:481–486. 12 Barbaud A, Bene MC, Reichert-Penetrat, Jacquin-Petit MA, Schmutz JL, Faure GC: Immunocompetent cells and adhesion molecules in 14 cases of cutaneous drug reactions induced by antibiotics. Arch Dermatol 1998;134:1040–1041.

376

13 Britschgi M, Steiner UC, Schmid S, Depta JPH, Senti G, Bircher A, Burkhart C, Yawalkar N, Pichler WJ: T-cell involvement in drug-induced acute generalized exanthematous pustulosis. J Clin Invest 2001;107:1433–1441. 14 Brockow K, Romano A, Blanca M, Ring J, Pichler W, Demoly P: General considerations for skin test procedures in the diagnosis of drug hypersensitivity. Allergy 2002;57:45–51. 15 Fernandez T, Torres MJ, R-Pena R, Fuentes MS, Robles S, Mayorga C, Blanca M: Decrease of selective immunoglobulin E response to amoxicillin despite repeated administration of benzylpenicillin and penicillin V. Clin Exp Allergy 2005;35:1645–1650. 16 Kleinhans M, Linzbach L, Zedlitz S, Kaufmann R, Boehncke WH: Positive patch test reactions to celecoxib may be due to irritation and do not correlate with the results of oral provocation. Contact Dermatitis 2002;47:100–102. 17 Barbaud A, Bursztejn AC, Vernassiere C, Schmutz JL, Trechot P: Toxidermie à l’Aerius
: les patch-tests avec la desloratadine induisent des faux positifs. Ann Dermatol Vénéréol 2005;132:S71. 18 Barbaud A, Trechot P, Reichert-Penetrat S, Commun N, Schmutz JL: Relevance of skin tests with drugs in investigating cutaneous adverse drug reactions. Contact Dermatitis 2001;45: 265–268. 19 Conilleau V, Dompmartin A, Michel M, Verneuil L, Leroy D: Photoscratch testing in systemic drug-induced photosensitivity. Photodermatol Photoimmunol Photomed 2000; 16:62–66. 20 Gonçalo M, Oliveira HS, Monteiro C, Clerins I, Figueiredo A: Allergic and systemic contact dermatitis from estradiol. Contact Dermatitis 1999;40:58– 59. 21 Lammintausta K, Mäkela L, Kalimo K: Rapid systemic valacyclovir reaction subsequent to acyclovir contact allergy. Contact Dermatitis 2001;45:181. 22 Ozkaya-Bayazit E, Bayazit H, Ozarmagan G: Topical provocation in 27 cases of cotrimoxazole-induced fixed drug eruption. Contact Dermatitis 1999;41: 185–189. 23 Alanko K, Stubb S, Reitamo S: Topical provocation of fixed drug eruption. Br J Dermatol 1987;116:561–567.

24 Alanko K: Topical provocation of fixed drug eruption: a study of 30 patients. Contact Dermatitis 1994;31:25–27. 25 Klein CE, Trautmann A, Zillikens D, Brocker EB: Patch testing in an unusual case of toxic epidermal necrolysis. Contact Dermatitis 1995;33:448–449. 26 Barbaud A, Trechot P, Reichert-Penetrat S, Granel F, Schmutz JL: Drug patch testing: the usefulness in testing on the most previously affected site in a systemic cutaneous adverse drug reaction to tetrazepam. Contact Dermatitis 2001;44:259–260. 27 Wilkinson DS, Fregert S, Magnusson B, Bandmann HJ, Calnan CD, Cronin E, Hjorth N, Maibach HJ, Malten KE, Meneghini CL, Pirilä V: Terminology of contact dermatitis. Acta Derm Venereol 1970;50:287–292. 28 Jonker MJ, Bruynzeel D: Anaphylactic reaction elicited by patch testing with diclofenac. Contact Dermatitis 2003; 49:114–115. 29 Romano A, Guéant-Rodriguez RM, Viola M, Amoghly F, Gaeta F, Nicolas JP, Guéant JL: Diagnosing immediate reactions to cephalosporins 2005;35: 1234–1242. 30 Co Minh HB, Bousquet PJ, Fontaine C, Kvedariene V, Demoly P: Systemic reactions during skin tests with
-lactams: a risk factor analysis. J Allergy Clin Immunol 2006;117:466–468. 31 Lammintausta K, Kortekangas-Savolainen O: The usefulness of skin tests to prove drug hypersensitivity. Br J Dermatol 2005;152:968–974. 32 Laxenaire MC, Moneret-Vautrin DA: Anaphylactic reactions to rocuronium. Br J Anaesth 2000;85:325–326. 33 Torres MJ, Sanchez-Sabate E, Alvarez J, Mayorga C, Fernandez J, Padial A, Cornejo-Garcia JA, Bellon T, Blanca M: Skin test evaluation in nonimmediate allergic reactions to penicillins. Allergy 2004;59:219–224. 34 Rodriguez-Bada JL, Montanez MI, Torres MJ, Mayorga C, Canto G, PerezInestrosa E, Suau R, Blanca M: Skin testing for immediate hypersensitivity to
-lactams: comparison between two commercial kits. Allergy 2006;61:947– 951. 35 Antunez C, Blanca-Lopez N, Torres MJ, Mayorga C, Perez-Inestrosa E, Montanez MI, Fernandez T, Blanca M: Immediate allergic reactions to cephalo-

Barbaud

36

37

38

39

40

41

42

43

44

sporins: evaluation of cross-reactivity with a panel of penicillins and cephalosporins. J Allergy Clin Immunol 2006; 117:404–410. Romano A, Viola M, Mondino C, Pettinato R, Di Fonso M, Papa G, Venuti A, Montuschi P: Diagnosing nonimmediate reactions to penicillins by in vivo tests. Int Arch Allergy Immunol 2002; 129:169–174. Bernedo N, Gonzalez I, Gastaminza G, Audicana M, Fernandez E, Munoz D: Positive patch test in vancomycin allergy. Contact Dermatitis 2001;45:43. Romano A, Di Fonso M, Papa G, Pietrantonio F, Federico F, Fabrizi G, Venuti A: Evaluation of adverse cutaneous reactions to aminopenicillins with emphasis on those manifested by maculopapular rashes. Allergy 1995;50: 113–118. Wolkenstein P, Chosidow O, Flechet ML, Robbiola O, Paul M, Dume L, Revuz J, Roujeau J.C: Patch testing in severe cutaneous adverse drug reactions, including Stevens-Johnson syndrome and toxic epidermal necrolysis. Contact Dermatitis 1996;35:234–236. Barbaud A, Trechot P, Weber-Muller F, Ulrich G, Commun N, Schmutz JL: Drug skin tests in cutaneous adverse drug reactions to pristinamycin: 29 cases with a study of cross-reactions between synergistins. Contact Dermatitis 2004;50:22–26. Menczer J, Barda G, Glezerman M, Hyat H, Brenner Y, Brickman CM, Tanai A: Hypersensitivity reaction to carboplatin. Results of skin tests. Eur J Gynaecol Oncol 1999;20:214– 216. Markman M, Zanotti K, Peterson G, Kulp B, Webster K, Belinson J: Expanded experience with an intradermal skin test to predict for the presence or absence of carboplatin hypersensitivity. J Clin Oncol 2003 15;21:4611–4614. Leguy-Seguin V, Jolimoy G, Coudert B, Pernot C, Dalac S, Vabres P, Collet E: Diagnostic and predictive value of skin testing in platinum salt hypersensitivity. J Allergy Clin Immunol 2007;119: 726–730. Brockow K, Kiehn M, Kleinheinz A, Vieluf D, Ring J: Positive skin tests in late reactions to radiographic contrast media. Allerg Immunol 1999;31:49–51.

45 Brockow K, Becker EW, Worret WI, Ring J: Late skin test reactions to radiocontrast medium. J Allergy Clin Immunol 1999;104:1107–1108. 46 Brockow K, Christiansen C, Kanny G, Clement O, Barbaud A, Bircher A, Dewachter P, Gueant JL, Rodriguez Gueant RM, Mouton-Faivre C, Ring J, Romano A, Sainte-Laudy J, Demoly P, Pichler WJ; ENDA; EAACI interest group on drug hypersensitivity: management of hypersensitivity reactions to iodinated contrast media. Allergy 2005;60:150–158. 47 Vernassiere C, Trechot P, Schmutz JL, Barbaud A: Results and negative value of skin tests in investigating delayed reactions to radio-contrast media. Contact Dermatitis 2004;50:359–366. 48 Gueant-Rodriguez RM, Romano A, Barbaud A, Brockow K, Gueant JL: Hypersensitivity reactions to iodinated contrast media. Curr Pharm Des 2006; 12:3359–3372. 49 Figarella I, Barbaud A, Lecompte T, De Maistre E, Reichert-Penetrat S, Schmutz JL: Réaction cutanée d’hypersensibilité retardée avec polysensibilisation aux héparines et héparinoides. Ann Dermatol Vénéreol 2001;128: 25–30. 50 Ludwig RJ, Beier C, Lindhoff-Last E, Kaufmann R, Boehncke WH: Tolerance of fondaparinux in a patient allergic to heparins and other glycosaminoglycans. Contact Dermatitis 2003;49:158– 159. 51 Vega JM, Blanca M, Carmona MJ, Garcia J, Claros A, Juarez C, Moya MC: Delayed allergic reactions to
-lactams. Four cases with intolerance to amoxicillin or ampicillin and good tolerance to penicillin G and V. Allergy 1991;46: 154–157. 52 Felix RH, Comaish JS: The value of patch and other skin tests in drug eruptions. Lancet 1974;1:1017–1019. 53 Romano A, Viola M, Mondino C, Pettinato R, di Fonso M, Papa G et al: Diagnosing nonimmediate reactions to penicillins by in vivo tests. Int Arch Allergy Immunol 2002, 129:169–174. 54 Dooms-Goossens A: Allergy to inhaled corticosteroids: a review. Am J Contact Dermatitis 1995;6:1–3. 55 Bircher AJ, Pelloni F, Langauer-Messmer S, Müller D: Delayed hypersensitivity reactions to corticosteroids applied to mucous membranes. Br J Dermatol 1996;135:310–313.

Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions

56 Kilpio K, Hannuksela M: Corticosteroid allergy in asthma. Allergy 2003; 58:1131–1135. 57 Buettiker U, Keller M, Pichler WJ, Braathen LR, Yawalkar N: Oral prednisolone-induced acute generalized exanthematous pustulosis due to corticosteroids of group A confirmed by epicutaneous testing and lymphocyte transformation tests. Dermatology 2006;213:40–43. 58 Bircher AJ, Flückiger R, Buchner SA: Eczematous infiltrate plaque to subcutaneous heparin: a type IV allergic reaction. Br J Dermatol 1990; 123:507– 514. 59 Moreau A, Dompmartin A, Esnault P, Michel M, Leroy D: Delayed hypersensitivity at the injection sites of a lowmolecular-weight heparin. Contact Dermatitis 1996;34:31–34. 60 Michel M, Dompmartin A, Szczurko C, Castel B, Moreau A, Leroy D: Eczematous-like drug eruption induced by synergistins. Contact Dermatitis 1996; 34:86–87. 61 Vaillant L, Camenen I, Lorette G: Patch testing with carbamazepine: reinduction of an exfoliative dermatitis. Arch Dermatol 1989;125:299. 62 Silva R, Machado A, Brandao M, Goncalo S: Patch test diagnosis in carbamazepine erythroderma. Contact Dermatitis 1982;8:283–284. 63 Camarasa JG: Patch test diagnosis of exfoliative dermatitis due to carbamazepine. Contact Dermatitis 1985;12: 49. 64 Romaguera C, Grimalt F, Vilaplana J, Azon A: Erythroderma from carbamazepine. Contact Dermatitis 1989;20: 304–305. 65 Jones M, Fernandez-Herrera J, Dorado JM, Sols M, Ruiz M, Garcia-Diez A: Epicutaneous test in carbamazepine cutaneous reactions. Dermatology 1994;188:18–20. 66 Rodriguez-Mosquera M, Iglesias A, Saez A, Vidal C: Patch test diagnosis of carbamazepine sensitivity? Contact Dermatitis 1991;25:137–138. 67 Alanko K: Patch testing in cutaneous reactions caused by carbamazepine. Contact Dermatitis 1993;29:254–257. 68 Corazza M, Mantovani L, Casetta I, Virgili A: Exfoliative dermatitis caused by carbamazepine in a patient with isolated IgA deficiency. Contact Dermatitis 1995;33:447.

377

69 Puig L, Nadal C, Fernandez-Figueras MT, Alomar A: Carbamazepine-induced drug rashes: diagnostic value of patch tests depends on clinicopathologic presentation. Contact Dermatitis 1996;34:435–437. 70 Lee AY, Choi J, Chey WY: Patch testing with carbamazepine and its main metabolite carbamazepine epoxide in cutaneous adverse drug reactions to carbamazepine. Contact Dermatitis 2003; 48:137–139. 71 Romano A, Pietrantonio F, Garcovich A, Rumi C, Bellocci F, Caradonna P, Barone C: Delayed hypersensitivity to diltiazem in two patients. Ann Allergy 1992;69:31–32. 72 Barbaud A, Trechot P, Gillet-Terver MN, Zannad F, Schmutz JL: Investigations allergologiques dans une toxidermie au diltiazem (Tildiem 300 LPR). Thérapie 1993;48:498–499. 73 Cholez C, Trechot P, Schmutz JL, Faure G, Bene MC, Barbaud A: Maculopapular rash induced by diltiazem: allergological investigations in four patients and cross-reactions between calcium channel blockers. Allergy 2003;58: 1207–1209. 74 Michel M, Dompmartin A, Louvet S, Szczurko C, Castel B, Leroy D: Skin reactions to hydroxyzine. Contact Dermatitis 1997;36:147–149. 75 Tomb RR, Lepoittevin JP, Espinassouze F, Heid E, Foussereau J: Systemic contact dermatitis from pseudoephedrine. Contact Dermatitis 1991;24:86–88. 76 Sanchez TS, Sanchez-Perez J, Aragues M, Garcia-Diaz A: Flare-up reaction of pseudoephedrine baboon syndrome after positive patch test. Contact Dermatitis 2000;42:312–313. 77 Camarasa JG, Serra-Baldrich E: Tetrazepam allergy detected by patch test. Contact Dermatitis 1990;22:246. 78 Collet E, Dalac S, Morvan C, Sgro C, Lambert D: Tetrazepam allergy once more detected by patch test. Contact Dermatitis 1992;26:281. 79 Tomb R, Grosshans E, Defour E, Heid E: Allergic skin reaction to tetrazepam detected by patch testing. Eur J Dermatol 1993;3:116–118. 80 Calkin JM, Maibach H: Delayed hypersensitivity drug reactions diagnosed by patch testing. Contact Dermatitis 1993; 29:223–233.

378

81 Pirker C, Misic A, Brinkmeier T, Frosch PJ: Tetrazepam drug sensitivity – usefulness of the patch test. Contact Dermatitis 2002;47:135–138. 82 Akiyama M, Nakada T, Sueki H, Fujisawa R, Iijima M: Drug eruption caused by non-ionic iodinated X-ray contrast media. J Acad Radiol 1998;5:S159– S161. 83 Vernassiere C, Barbaud A, Trechot P, Weber-Muller F, Schmutz JL: Systemic acyclovir reaction subsequent to acyclovir contact allergy: Which systemic antiviral drug should then be used? Contact Dermatitis 2003;49:155–157. 84 Sakakibara T, Hata M, Numano K, Kawase Y, Yamanishi Tkawana S, Tsuboi N: Fixed-drug eruption caused by allylisopropylacetylurea. Contact Dermatitis 2001;44:189–190. 85 Gaig P, San Miguel-Moncin MM, Bartra J, Bonet A, Garcia-Ortega P: Usefulness of patch tests for diagnosing selective allergy to captopril. J Investig Allergol Clin Immunol 2001;11:204– 206. 86 Kawada A, Aragane Y, Maeda A, Asai M, Shiraishi H, Tezuka T: Drug eruption induced by cefcapene pivoxil hydrochloride. Contact Dermatitis 2001; 44:197. 87 Verbeiren S, Morant C, Charlanne H, Ajebbar K, Caron J, Modiano P: Toxidermie due au celecoxib avec tests épicutané positif. Ann Dermatol Vénéreol 2002;129:203–205. 88 Machet L, Vaillant L, Dardaine V, Lorette G: Patch testing with clobazam: relapse of generalized drug eruption. Contact Dermatitis 1992;26:347–348. 89 Brown VL, Orton DI: Cutaneous adverse drug reaction to oral chlorphenamine detected with patch testing. Contact Dermatitis 2005;52:49–50. 90 Estrada JL, Alvarez Puebla MJ, Ortiz de Urbina JJ, Matilla B, Rodriguez Priet MA, Gozalo F: Generalized eczema due to codeine. Contact Dermatitis 2001;44:185. 91 Romano A, Pietrantonio F, Di Fonso M, Garcovich A, Chiarelli C, Venuti A, Barone C: Positivity of patch tests in cutaneous reaction to diclofenac. Allergy 1994;49:57–59. 92 Frouin E, Roth B, Grange A, Grange F, Tortel MC, Guillaume JC: Syndrome d’hypersensibilité à la fluindione (Préviscan
). Positivité des tests épicutanés. Ann Dermatol Vénéreol 2005;132:1000–1002.

93 Rodriguez-Morales A, Alonso Llamazares A, Palacios Benito R, Martinez Cocera C: Fixed drug eruption from quinolones with a positive lesional patch test to ciprofloxacin. Contact Dermatitis 2001;44:255. 94 Scherer K, Bircher AJ: Hypersensitivity reactions to fluoroquinolones. Curr Allergy Asthma Rep 2005;5:15– 21. 95 Schmid DA, Depta JP, Pichler WJ: Tcell-mediated hypersensitivity to quinolones: mechanisms and cross-reactivity. Clin Exp Allergy 2006;36: 59–69. 96 Rebello S, Sanchez P, Vega JM, Sedano E, Sanchis ME, Asensio T, Callejo A: Hypersensitivity syndrome from isoniazid with positive patch test. Contact Dermatitis 2001;45:306. 97 Gonzalo-Garijo MA, de Arila D, Rodriguez-Nevado I: Generalized reaction after patch testing with metamizol. Contact Dermatitis 2001;45: 180. 98 Quinones Estevez D, Fernandez Schmitz C: Exanthema to metamizole. Allergy 2001;56:262–263. 99 Sasaki K, Yamamoto T, Kishi M, Yokozeki H, Nishioka K: Acute exanthematous pustular drug eruption induced by mexiletine. Eur J Dermatol 2001;11:469–471. 100 Antunes A, Davril A, Trechot P, Grandidier M, Truchetet F, Cuny JF: Syndrome d’hypersensibilité à la minocycline. Ann Dermatol Vénéreol 1999; 126:518–521. 101 Malheiro D, Cadinha S, Rodrigues J, Vaz M, Castel-Branco MG: Nimesulide-induced fixed drug eruption. Allergol Immunopathol (Madr) 2005; 33: 285–287. 102 Barranco R, Tornero P, De Barrio M, De Frutos C, Rodriguez A, Rubio M: Type IV hypersensitivity to oral nystatin. Contact Dermatitis 2001;45:60. 103 Mashiah J, Brenner S: A systemic reaction to patch testing for the evaluation of acute generalized exanthematous pustulosis. Arch Dermatol 2003; 139:1181–1183. 104 Montoro J, Diaz M, Genis C, Lozano A, Bertomeu F: Non-pigmenting cutaneous-mucosal fixed drug eruption due to piroxicam. Allergol Immunopathol 2003;31:53–55.

Barbaud

105 Wong DA, Freeman S: Cutaneous allergic reaction to intramuscular vitamin K1. Australas J Dermatol 1999; 40:147–152. 106 Antunez C, Martin E, Cornejo-Garcia JA, Blanca-Lopez N, R-Pena R, Mayorga C, Torres MJ, Blanca M: Immediate hypersensitivity reactions to penicillins and other
-lactams. Curr Pharm Des 2006;12:3327–3333. 107 Bonadonna P, Schiappoli M, Senna G, Passalacqua G: Delayed selective reaction to clavulanic acid: a case report. J Investig Allergol Clin Immunol 2005; 15:302–304. 108 Torres MJ, Romano A, Mayorga C, Moya MC, Guzman AE, Reche M et al: Diagnostic evaluation of a large group of patients with immediate allergy to penicillins: the role of skin testing. Allergy 2001;56:850–856. 109 El-Qutob D, Morales C, Pelaez A: Allergic reaction caused by articaine. Allergol Immunopathol (Madr) 2005; 33:115–116. 110 Berkun Y, Ben-Zvi A, Levy Y, Galili D, Shalit M: Evaluation of adverse reactions to local anesthetics: experience with 236 patients. Ann Allergy Asthma Immunol 2003;91:342–345. 111 Notman MJ, Phillips EJ, Knowles SR, Weber EA, Shear NH: Clindamycin skin testing has limited diagnostic potential. Contact Dermatitis 2005; 53:335–338. 112 Padial A, Posadas S, Alvarez J, Torres MJ, Alvarez JA, Mayorga C, Blanca M: Nonimmediate reactions to systemic corticosteroids suggest an immunological mechanism. Allergy 2005;60: 665–670. 113 Caceres Calle O, Fernandez-Benitez M: Allergy to dexchlorpheniramine. Study of a case. Allergol Immunopathol (Madr) 2004;32:306–309.

114 Beaudouin E, Kanny G, Blanloeil Y, Guilloux L, Renaudin JM, MoneretVautrin DA: Anaphylactic shock induced by gadoterate meglumine (Dotarem). Allerg Immunol (Paris) 2003; 35:382–385. 115 Adachi A, Fukunaga A, Horikawa T: A case of human insulin allergy induced by short-acting and intermediateacting insulin but not by long-acting insulin. Int J Dermatol 2004; 43:597– 599. 116 Lee AY, Chey WY, Choi J, Jeon JS: Insulin-induced drug eruptions and reliability of skin tests. Acta Derm Venereol 2002;82:114–117. 117 Takahama H, Tsutsumi Y, Kubota Y: Anaphylaxis due to levofloxacin. Int J Dermatol 2005;44:789–790. 118 Buergin S, Scherer K, Hausermann P, Bircher AJ: Immediate hypersensitivity to rifampicin in three patients: diagnostic procedures and induction of clinical tolerance. Int Arch Allergy Immunol 2006;140:20–26. 119 Strauss RM, Green ST, Gawkrodger DJ: Rifampicin allergy confirmed by an intradermal test, but with a negative patch test. Contact Dermatitis 2001;45:108. 120 Patel VJ, Bruck MC, Katz BE: Hypersensitivity reaction to hyaluronic acid with negative skin testing. Plast Reconstr Surg 2006;117:92e–94e. 121 Fays S, Trechot P, Schmutz JL, Cuny JF, Truchetet F, Barbaud A: Bupropion and generalized acute urticaria: eight cases. Br J Dermatol 2003; 148:177– 178. 122 Trujillo MJ, de Barrio M, Rodriguez A, Moreno-Zazo M, Sanchez I, Pelta R, Tornero P, Herrero T: Piroxicaminduced photodermatitis. Cross-reactivity among oxicams. A case report. Allergol Immunopathol (Madr) 2001; 29:133–136.

123 Gonçalo M, Figueiredo A: Cross-reactions among oxicams in cutaneous adverse drug reactions. Contact Dermatitis 2002;46:s31. 124 Barranco R, Rodriguez A, de Barrio M, Trujillo MJ, de Frutos C, Matheu V, Tornero P, Herrero T: Sympathomimetic drug allergy: cross-reactivity study by patch test. Am J Clin Dermatol 2004, 5:351–355. 125 Lee AY, Choi J, Chey WY: Patch testing with carbamazepine and its main metabolite carbamazepine epoxide in cutaneous adverse drug reactions to carbamazepine. Contact Dermatitis 2003;48:137–139. 126 Romano A, Quaratino D, Papa G, Di Fonso M, Venuti A: Aminopenicillin allergy. Arch Dis Child. 1997;76:513– 517. 127 Torres MJ, Mayorga C, Leyva L, Guzman AE, Cornejo-Garcia JA, Juarez C, Blanca M: Controlled administration of penicillin to patients with a positive history but negative skin and specific serum IgE tests. Clin Exp Allergy 2002;32:270–276. 128 Messaad D, Sahla H, Benahmed S, Godard P, Bousquet J, Demoly P: Drug provocation tests in patients with a history suggesting an immediate drug hypersensitivity reaction. Ann Intern Med 2004 15;140:1001–1006. 129 Demoly P, Benahmed S, Sahla H, Messaad D, Valembois M, Godard P, Michel FB, Bousquet J: Allergy to macrolides. 21 cases. Presse Méd 2000; 29: 294–298. 130 Ulrich G, Schmutz JL, Trechot P, Commun N, Barbaud A: Sensitization to petrolatum: an unusual cause of false-positive drug patch tests. Allergy 2004;59:1006–1009.

Dr. Annick Barbaud Dermatology Department, Fournier Hospital 36 quai de la Bataille, FR–54000 Nancy (France) Tel. +33 3 8385 2465, Fax +33 3 8385 2412 E-Mail [email protected]

Drug Skin Tests in Investigating Systemic Cutaneous Drug Reactions

379

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 380–390

In vitro Tests of T-Cell-Mediated Drug Hypersensitivity Andreas Beeler
Werner J. Pichler Division of Allergology, Department for Rheumatology and Clinical Immunology/Allergology, Inselspital, Bern, Switzerland

Abstract In clinical practice, side effects of drugs are a major problem. A broad range of drugs can elicit many different immunemediated diseases with distinct pathomechanisms. In this chapter, we focus on current in vitro techniques for the diagnosis of T-cell-mediated drug hypersensitivity. Specifically, we discuss the most recent findings regarding the diagnosis of T-cell-mediated drug hypersensitivity reactions that can be applied to the development of new tests which would enable a more conclusive diagnosis. Copyright © 2007 S. Karger AG, Basel

Introduction

Drugs are capable of inducing all of the immunological hypersensitivity reactions such as IgEmediated type-I reactions including anaphylaxis, urticaria, rhinitis, angioedema and bronchoconstriction, type-II and -III reactions such as hemolytic anemia, thrombocytopenia, purpura and hypersensitivity vasculitis, and type-IV delayed reactions such as allergic contact dermatitis, maculopapular, bullous or pustular exanthemas and fixed drug eruptions. A conclusive diagnosis of drug allergy has two aims: to confirm a drug-hypersensitivity reaction, and to identify its causative agent.

In vivo tests, such as the patch, prick and intracutaneous tests, often do not yield positive reactions, even in patients with well-documented histories of drug-allergic reactions [1, 2]. Challenge tests, which are considered to be the gold standard, are frequently not tolerated by the patients, bear the risk of severe reactions, and do not differentiate between allergic and pseudo-allergic reactions [3]. In vitro tests have the advantage of being a safe procedure for the patient, avoiding possible new or recall sensitizations to the drug or other adverse drug effects, and may provide insight into the underlying pathomechanisms. However, they are often still a research tool, are not standardized, fit only to certain type of drug allergies, and are dependent on the availability of the pure substance and fresh blood cells or frozen mononuclear cells which are not always at hand. Nevertheless, recent progress in drug allergy research as well as the urgent need for such tests has revived the interest in this topic. In this chapter, different readout systems for in vitro tests are discussed, which detect drug-reacting T cells in the peripheral blood of drug-allergic patients and identify the culprit drug.

Pathomechanisms of T-Cell-Mediated Drug Hypersensitivity

There is increasing evidence that drug-specific T cells play an important role in the pathogenesis of non-IgE-mediated drug hypersensitivity reactions. Positive patch test reactions after 24–48 h, positive in vitro reactivity in lymphocyte transformation tests (LTT; 3H-thymidine incorporation), and in particular the generation of drugspecific T-cell lines and T-cell clones (TCCs) from the blood of patients with different forms of drug hypersensitivities such as maculopapular exanthema (MPE), drug rash with eosinophilia and systemic symptoms (DRESS), acute generalized exanthematous pustulosis (AGEP), StevensJohnson Syndrome (SJS) and toxic epidermal necrolysis (TEN) clearly demonstrate that drugspecific T cells can be detected in the affected tissue and peripheral blood of drug-allergic patients [4–10]. In addition, a newer study clearly showed that 1:250 to 1:10,000 of T cells in the peripheral blood of drug-allergic patients are reactive to the relevant drug. This frequency of drug-reactive T cells is even higher than the frequency of T cells able to recognize tetanus toxoid in the same subjects. This study also demonstrated the persistence of drug-reactive T cells in blood for up to 12 years [5]. We assume that the long persistence of memory drug-specific T cells, in spite of strict drug avoidance, is eventually related to some crossreactivity of the drug-reactive T cells, as drugreactive T cells frequently possess another specificity, as revealed by the high frequency of cross-reactivity with allogeneic major histocompatibility complex (MHC)-peptide complexes [11]. However, the exact nature of these putative cross-reactive antigens, which might maintain the drug-reactive memory pool, is unknown. The functional dissection of drug-reactive T cells obtained from different clinical pictures revealed that the pronounced heterogeneity of drug

In vitro Tests of T-Cell-Mediated Drug Hypersensitivity

hypersensitivity reactions in the skin as well as in other organs such as the kidney are due to different T-cell functions [12].

Sequence of Events in T-Cell Activation – Possibility for Testing

The complete stimulation of T cells requires the formation of the immunological synapse involving the MHC, T-cell receptor (TCR) and various adhesion and costimulatory molecules. This activation goes along with a sequence of events on the membrane and inside the T cell, which can be monitored and might help to identify the T cells with particular antigen specificity. After T cells are triggered by a specific antigen, they enter a maturation process of several days that involves commitment, proliferation, and functional differentiation (fig. 1). Commitment occurs rapidly after binding of the TCR to a specific antigen presented by MHC class I or II, is accompanied by an increase in intracellular Ca2+, generation of inositol 1,4,5-triphosphate, and followed by the expression of a number of early genes within half an hour of antigen recognition, including the proto-oncogene c-fos, cellular oncogenes c-myc and c-jun, and the transcription factors NFAT and NF-B. Moreover, within 1–2 h of antigen recognition, genes encoding for IL-2, 3, 4, 5, and 6, IFN-, TGF-, CD25 (IL-2R), CD40L, the early activation marker CD69 and numerous other proteins are expressed. In the initial proliferation phase around 1–2 days after T-cell activation, additional gene expression is induced including HLA-DR, CD80, and various adhesion molecules like VLA [13– 15] and DNA are synthesized. 3–5 days after activation, T cells enter the phase of terminal differentiation, characterized by the expression of functional genes, and production of perforin and granzyme B molecules important for cytotoxic effector cells, while FasL expression occurs earlier [16–18].

381

COMMITMENT

PROLIFERATION

FUNCTIONAL DIFFERENTION

c-fos/c-jun NFAT Myc/NF-␬B IL-2 / IFN-␥ TGF-␤ / IL-3 IL-2 R␣ CD69/CD40L IL-4 to IL-6 DNA synthesis HLA-Class DR / CD80 CTL Serine esterase / FasL / Perforin / GranzymeB

VLA HOURS 0

1

DAYS 2

3

4

5

//

1

2

3

4

5

6

7

Fig. 1. Starting time of gene expression events after T-cell activation by antigen.

These very early activation steps are not further discussed in this chapter. However, measurement of these early signs of T-cell activation could theoretically be used to detect drug-reacting T cells in vitro. In the following sections of this chapter, tests detecting the various sequences of activation-induced events of T cells are discussed. They include upregulation of activation markers, cytokine secretion, proliferation, and cytotoxicity of drug-reactive cells.

Diagnostic Tests for Drug-Specific T-Cell-Mediated Hypersensitivity Reactions

When to Perform the in vitro Tests? In vitro tests to identify the drug involved in drug hypersensitivity are normally not done during

382

the acute stage of the disease, as the immune system is still strongly activated, causing a high background proliferation and activity [10]. Skin tests like patch/prick and in vitro tests are performed after a time interval that allows for the resolution of clinical symptoms and the clearance of the incriminated drugs and anti-allergic medications from the circulation. In general, peripheral blood mononuclear cells (PBMC) isolated from an acute drug-allergic patient are highly activated. For example, in an LTT, these activated cells would still proliferate and therefore, high background counts per minute are measured (unstimulated condition), whereas in the drugstimulated conditions, the further enhancement of proliferation is only marginal and thus not easy to detect. However, recent data by Shiohara et al. [pp 251–266] suggested that this concept is not always applicable. They observed that the

Beeler ⴢ Pichler

carbamazepine 1 µg/ml

culture medium PerCp-A: CD4 PerCp-A

CD4 – PerCP

10,000

99.79

0.09

99.31

carbamazepine 10 µg/ml

carbamazepine 5 µg/ml

0.55

98.34

1.49

98.06

1.74

0.19

0.01

Tetanus Toxoid 10 µg/ml 99.29

0.39

0.30

0.02

1,000 100 10 0.12

1 1

0 10 100 1,000 10,000 PE-A: CD69 PE-A

0.14 1

0 10 100 1,000 10,000 PE-A: CD69 PE-A

0.17 1

0 10 100 1,000 10,000 PE-A: CD69 PE-A

1

10 100 1,000 10,000 PE-A: CD69 PE-A

1

10 100 1,000 10,000 PE-A: CD69 PE-A

CD69 – PE

Fig. 2. Measurement of CD69 upregulation after drug stimulation. Flow cytometric analysis of PBMC from a carbamazepine-allergic patient gated on CD4+ T cells after stimulation for 48 h with culture medium alone

(negative control), different concentrations of carbamazepine (1, 5, and 10 g/ml) or tetanus toxoid (positive control).

LTT in SJS/TEN is more often positive in the acute state than after remission at 6 weeks, while in DRESS a later testing might be more appropriate. At present, one cannot predict whether the reactivity of the drug-specific T cells in an individual patient will persist for months or even years, and whether those which have lost their in vitro reactivity will tolerate the drug again. On one hand, we and others have observed positive in vitro reactivity of T cells after drug stimulation more than 12 years after the original treatment with -lactams or sulfamethoxazole which had originally caused a delayed or immediate reaction [5]. On the other hand, some patients appear to lose reactivity within 1–3 years after the reaction. Thus many groups carry out tests after a minimal time interval of 3 weeks but more than 3–6 months after the acute event. A special situation is found in children, in whom drug hypersensitivity often cannot be demonstrated even if the history is quite convincing. Reasons for this could be that young children may be incapable of developing an appropriate memory immune response at all, and the antigen/drug-activated cells undergo rapid apoptosis [19; Ponvert et al., pp 321–339].

Flow Cytometry Analysis of Different Surface Markers In the peripheral blood of acute patients with drug allergy, circulating T cells are activated in the commitment phase [10]. It has also been shown for in vitro cultures that specific, drug-activated lymphocytes express a number of molecules on their surface, such as CD25, CD69, CD40L and later CD71 and HLA-DR, which are usually expressed minimally or are even absent on resting cells [11]. Such activation antigens are easily detected by immunofluorescence and flow cytometric analysis and offer the advantage of identifying the T-cell subpopulations involved in the activation process within the bulk lymphocyte cultures. Indeed, our preliminary studies on activation markers in different drug-allergic patients measured by flow cytometry show that, upon drug stimulation of PBMC isolated from a drug-allergic patient during remission, an enhanced expression of CD69 (fig. 2), CD25 and CD71 is seen on the surface of a relatively high number of T cells, while in non-sensitized individuals this upregulation was not observed. Actually, the frequency of CD69+ T cells after drug activation was much higher than the frequency of drug-spe-

In vitro Tests of T-Cell-Mediated Drug Hypersensitivity

383

cific T cells in the peripheral blood of drug-allergic patients determined by the enzyme-linked immunospot technique and carboxyfluorescein diacetate succinimidyl ester (CFSE) measurements [5]. Nevertheless, the upregulation was specific as the same drug inducing upregulation in sensitized individuals failed to elicit a CD69 upregulation in non-sensitized patients or control individuals. Preliminary data indicate that CD69 upregulation on T cells has two components: first, the early activation of drug-specific T cells by TCR-mediated activation, and then a further amplification of this response by secreted cytokines inducing CD69 on neighbor T cells, which are not drug-specific (own observation). Thus, this amplification occurs only if ‘real’ drug-reactive T cells are present in the pool of T cells. Thus, measurement of CD69 upregulation might be a useful tool to detect the presence of drug-specific and cross-reactive T cells in the peripheral blood of drug-allergic patients, because amplification by IL-2 renders the detection much easier by flow cytometric measurement but is not more unspecific (good sensitivity and specificity). Measurement of Cytokines Produced by Cells after Drug-Stimulation Due to their critical role in orchestrating all phases of an immune response, cytokines are important mediators of effector functions. In general, the amount and pattern of cytokines in vivo and in vitro correlates with the type of immune response evolving after antigen contact [20, 21]. During the acute phase of a drug-allergic reaction, in vivo monitoring of the cytokine patterns in immediate- and delayed-type reactions to drugs reveals a Th1 or Th2 cytokine profile, depending on the clinical entity involved. In subjects with delayed-type reactions to -lactams, a high production of IL-2, IFN-, and TNF-  in PBMC has been reported. On the other hand, in confirmed IgE-mediated reactions such as anaphylaxis to -lactams, this monitoring has shown

384

a Th2-biased cytokine profile with preferential expression of IL-4 [22]. Most studies investigating the role of cytokines in different forms of cutaneous drug hypersensitivity reactions combine immunohistological analyses of acute skin lesions with quantification of cytokines released by drug-specific TCCs generated in vitro from hypersensitive patients in remission. In general, TCCs from drug-allergic patients show a very heterogeneous pattern of cytokines. Nevertheless, an elevated IFN- production is often observed in patients with clinical pictures as diverse as MPE, DRESS and AGEP to carbamazepine, lamotrigine, celecoxib and other drugs [6, 9, 23–25]. IFN- may thus play a dominant role in the pathophysiology of different forms of exanthemas and serve as a marker cytokine. Its high production also provides an explanation for the upregulation of MHC class II on keratinocytes. Such activated keratinocytes subsequently enable drug presentation to CD4+ T cells [26, 27]. Another rather dominant cytokine in various forms of drug hypersensitivity reactions is IL-5. In acute DRESS and MPE patients, high IL-5 levels were found in the serum, and secretion of IL-5 and eotaxin (CCL-11, highly expressed in tissue sections) are common characteristics of drugspecific TCCs [28–32]. Eotaxin and IL-5 are both known to be key factors in regulating the growth, differentiation and activation of eosinophils. Both were also detected already in PBMC cultures with the incriminated drug from patients with such diseases [33, 34]. Consequently, measurement of IL-5 in combination with the LTT has been proposed to be a better readout parameter for detecting a sensitization to a drug than the LTT alone [35]. On the other hand, IL-5 expression was not increased in a study investigating the cytokine pattern in immediate and non-immediate reactions, and the analysis of TCC to carbamazepine and lamotrigine from patients with DRESS did not reveal a high IL-5 production, but rather a high IFN- production [6, 9, 22].

Beeler ⴢ Pichler

In a particular drug hypersensitivity reaction like AGEP, analysis of IL-8 (CXCL8) might be useful as this chemokine is produced by T cells causing this form of drug hypersensitivity often together with IFN- and GM-CSF [24]. Based on all these findings it would be advisable to screen for IFN-- and IL-5-secreting T cells in most drug hypersensitivity reactions. Such an approach might be equally or even more sensitive than proliferation assays, in particular in diseases with eosinophilia. The assessment of cytokine production can be performed by molecular biology techniques such as polymerase chain reaction or competitive reverse transcription polymerase chain reaction. At the protein level, cytokines are measured in cell culture supernatants by bioassays and immunoassays such as ELISA, radioimmunoassay or the enzyme-linked immunospot technique. Flow cytometric analysis of intracellular cytokines with fluorescent conjugated monoclonal antibodies can also be used in vitro or ex vivo. Luminex’s xMAP
technology or BioplexTM, which are multiplexed, particle-based, flow cytometric assays capable of measuring a whole panel of cytokines, will probably provide useful albeit expensive tools considering the heterogeneous cytokine profile released following drug-specific activation of PBMC. Cell Proliferation Assays The LTT is currently the most widely used standard test to detect a T-cell sensitization to drugs for diagnosis of T-cell-mediated drug hypersensitivity. It is applied for delayed-type drug hypersensitivity reactions but can also detect T cells in the peripheral blood of a patient with immediate reactions [4]. The LTT measures the proliferative response of T cells and PBMC to the drug under in vitro conditions [36]. The specificity of the drug-specific proliferative response was demonstrated by establishing drug-specific T-cell lines and predominantly CD4+ TCCs from LTT cultures or patch test sites against different drugs

In vitro Tests of T-Cell-Mediated Drug Hypersensitivity

such as -lactams, lidocaine, mepivacaine, sulfamethoxazole, celecoxib, lamotrigine, carbamazepine, p-phenylenediamine or ciprofloxacin [8, 11, 25, 33, 37–40]. In the LTT, PBMC are obtained from a drug-sensitized patient and cultured in the presence of the suspected drug for 6 days. Sensitized lymphocytes undergo blastogenesis and generate cytokines such as IL-2, followed by a proliferative response that can be measured by the incorporation of 3H-thymidine during DNA synthesis. The result can be expressed as a stimulation index, which is the ratio of cell proliferation with antigen divided by the proliferation without antigen (fig. 3). Normally, a stimulation index of 12 and for -lactams of 13 is considered positive. A retrospective analysis of 923 case histories of patients with suspected drug allergy indicated that the sensitivity of the LTT depends on the drug tested. 78 of 100 patients with a very likely drug allergy had a positive reactivity in the LTT, yielding a sensitivity of 78%. When only allergies to -lactam antibiotics were analyzed, the sensitivity was around 74% and thus higher than that of skin tests (62%). The negative and positive predictive values of the LTT can only be calculated to a certain extent for penicillin preparations, in particular amoxicillin. Assuming an incidence of 5% sensitization to amoxicillin in exposed individuals, the negative predictive value would be 98.6% and the positive predictive value 79%. In agreement with this finding was the observation that those patients classified to have a lower likelihood of drug hypersensitivity had a lower incidence of a positive LTT, while patients with an unlikely history had the lowest incidence of positive reactivity in the LTT [36]. In a similar study, PBMC from 62% of 50 patients with welldocumented delayed or immediate -lactam hypersensitivity proliferated in vitro in response to at least one of the penicillins tested independent of the type of reaction [4]. The sensitivity of the test was higher for the immediate reaction group (64.5%) than for the non-immediate reaction

385

WITH DRUG WITH DRUG +

STIMULATION DAY 0

NO DRUG NO DRUG –

Peripheral blood mononuclear cells are separated over a density gradient. Distribution of the cell suspension in triplicates in a microtiter plate with 96 wells. Stimulation with different concentrations of drug, tetanus toxoid as positive control and medium alone as negative control. Incubation at 37°C in a CO2 incubator for 6 days.

DAY 6 Drug stimulated PBMC are pulsed with 3H-thymidine for 8 –14 h.

DAY 7

3H 3H

3H

3H

PBMC are transfered on a filter, drugspecific proliferation is measured as 3Hthymidine uptake in counts/min (cpm). The mean cpm‘s of the triplicates to the different drug concentrations are put into the relation to the baseline (negative control) cpm. This is expressed as stimulation index (SI).

3H 3H

3H

Proliferating drug-reactive T cell with 3Hthymidine incorporation Non-specific T / B cell Drug-specific T cell Monocyte

Fig. 3. Lymphocyte transformation test (LTT) from days 0–7.

group (57.9%), but the proliferative responses in the immediate reaction group tended to be less intense than in the non-immediate reaction group. Everness et al. [41] found a sensitivity of 92% in 66 patients with a nickel allergy, and Warrington et al. [42] reported a sensitivity of 51% in 69 patients with clinically established allergies to several drugs. Lower sensitivities of 33% were ob-

386

served by Barna et al. [43] and 38% by Berg et al. [44]. In the frame of our retrospective analysis of 923 case histories, the specificity of the LTT was evaluated as well. 15 of 102 presumably nondrug-allergic patients had a positive LTT. Thus, the specificity of the LTT was 85%. However, this relatively low specificity might be due to the fact

Beeler ⴢ Pichler

that any positive LTT to an NSAID was calculated as false-positive, as it was considered a pseudo-allergic reaction even though some reactions, e.g. to diclofenac, might have been real allergies. Other analyses indeed showed a far better specificity for the LTT using various drugs [4, 6, 9, 10, 45]. Recent analyses of LTT to antiepileptic drugs eliciting DRESS revealed that the LTT in drug-allergic patients showed a sensitivity of 97% and a specificity of 100% for carbamazepine and lamotrigine hypersensitivity, while nonallergic persons showed none or only a weak proliferation to the drug. This high specificity was recently confirmed for -lactam hypersensitivity as only sensitized patients reacted, resulting in a specificity of 93% [4, 6, 9]. Different modifications and simplifications of the LTT have been proposed. One of these modified tests is the memory lymphocyte immunostimulation assay (MELISA
), which is mainly applied to test sensitization to metals. Its only difference compared to the usual LTT is a depletion step of the monocytes before culture with the drug is started. In conclusion, the LTT, with an overall sensitivity of 60–70%, is still a controversial test because of several disadvantages: it requires sterile cell cultures; it takes a long time and is cumbersome; the drug-specific proliferation depends on the quality of the culture medium, and cells are labeled radioactively. However, our own experience suggests that, by complementing other tests, it is a helpful tool in the diagnosis of drug allergies [46]. In the hands of experienced technicians, it can provide valuable additional information in the diagnosis of drug hypersensitivity: a positive LTT helps to define the incriminated drug in the drug hypersensitivity reaction, but a negative test does not rule out drug hypersensitivity. To evaluate the proliferative in vitro response of cells to drugs, flow cytometry can also be used. Cell samples containing drug-specific cells are first stained with a fluorescent dye such as CFSE which binds to amino groups of intracellular pro-

In vitro Tests of T-Cell-Mediated Drug Hypersensitivity

teins or PKH26 which becomes integrated into the lipid membrane of the cells. Upon cell division after stimulation, cells become 2 times less fluorescent at each division due to partition of the dye between daughter cells. The number of cell cycles can thus directly be deduced from the decreasing intensity of fluorescence. It has been shown that the CFSE fluorescent intensity decreased substantially in drug-stimulated cultures, but not in control cultures without drug addition. In order to validate the specificity of the expanding T cells in response to drugs, cells have been sorted, cloned and tested in a drug-specificity assay. The majority of the drug-stimulated CD4+ T cells reacted in a specific way to the drug, supporting the notion that the expanded T cells were indeed drug-specific [5]. As an advantage, this method allows the identification of the dividing cell population involved in the reactivity to the drug. The sensitivity and specificity of the test needs still to be evaluated. Cytotoxicity The typical feature of a type-IVc reaction is the participation of cytotoxic T cells [12]. It appears that cytotoxic T cells are involved to a certain extent in all forms of T-cell-mediated drug allergies, even if they are clinically quite distinct. Cytotoxic T cells are often located at the dermoepidermal junction zone or in the epidermis, along with signs of keratinocyte cell damage, but may also be present in other organs like the kidney (interstitial nephritis) or liver (hepatitis) [13, 47; Keller et al., pp 295–305]. Most data indicate that drug-specific cytotoxic T cells kill target cells mainly by perforin/granzyme B, which are important mediators of cell-mediated cytotoxic reactions [14, 27]. It has also been demonstrated that perforin levels, as measured by quantitative polymerase chain reaction in PBMC, correlate with the severity of the allergic reactions during the acute phase, e.g. in TEN [15]. Interestingly, in vitro analysis of drug-specific TCCs and T-cell lines from a patient with comparably mild exan-

387

thema has also shown some FasL-mediated killing [10, 26, 27, 48–50]. Since cytotoxic T cells are presumably causing the most severe forms of drug hypersensitivity, namely SJS/TEN and hepatitis, their detection in drug hypersensitivity reactions has a high priority. Interestingly, recent data from our group showed that some highly potent drug-specific cytotoxic T cells were not detected by proliferation assays but could only be identified as drug-specific by their cytotoxic potential in a cytotoxicity test [48]. Regretfully, such cytotoxicity tests are rather cumbersome and require a radioactive labeling of target cells (e.g. 51 Cr-labelling of autologous, Epstein-Barr virustransformed target cells). The lack of proliferation may explain why, in diseases thought to be due to cytotoxic cells like TEN [51], the LTT is often negative. To date, no simple in vitro test has been established which would allow identification of in vitro drug-activated T cells with cytotoxic capacity [10, 26, 27, 48–50], but the preliminary data from our group to detect cytotoxic T cells after drug stimulation look promising.

Conclusion

The three cornerstones of drug allergy diagnosis are an exact history, skin tests and supplementary laboratory investigations (differential blood count,

liver function tests), which help to define the severity of the reaction – even though an ever-increasing armamentarium of tests is available to identify the culprit drug and to confirm the diagnosis of drug allergy by in vitro tests. As almost all tests pinpointing the relevant drug have a moderate sensitivity, it is advisable to combine skin and in vitro tests to increase the overall sensitivity. Due to the peculiar circumstances of the allergic reaction, it is difficult to validate the respective tests as provocation tests are rarely done. Nevertheless, in the majority of cases (
70%) the combined approach of history, skin and in vitro tests enables reliable identification of the relevant drug. The knowledge of drug hypersensitivity reactions has increased in the last 10 years and new concepts explaining some ‘unexplainable’ findings have been developed. It will take some time to implement this new knowledge in the routine diagnosis of drug allergies. If this knowledge is combined with the rapid progress of cellular diagnostic tests using flow cytometry and the growing knowledge of the human genome, we will undoubtedly not only identify novel genetic predisposing factors for drug hypersensitivity reactions, but may also have new tests available to identify the relevant drug. Together, these developments will hopefully lead to a better diagnosis and, consequently, to a reduced incidence of severe drug hypersensitivity reactions.

References 1 Torres MJ, Blanca, M, Fernandez J, et al: Diagnosis of immediate allergic reactions to beta-lactam antibiotics. Allergy 2003;58:961–972. 2 Romano A, Blanca M, Torres MJ, et al: Diagnosis of nonimmediate reactions to beta-lactam antibiotics. Allergy 2004;59:1153–1160. 3 Aberer W, Bircher, A, Romano A, et al: Drug provocation testing in the diagnosis of drug hypersensitivity reactions: general considerations. Allergy 2003;58:854–863.

388

4 Luque I, Leyva, L, Jose Torres M, et al: In vitro T-cell responses to beta-lactam drugs in immediate and nonimmediate allergic reactions. Allergy 2001;56:611– 618. 5 Beeler A, Engler O, Gerber, BO, Pichler WJ: Long-lasting reactivity and high frequency of drug-specific T cells after severe systemic drug hypersensitivity reactions. J Allergy Clin Immunol 2006;117:455–462.

6 Naisbitt DJ, Britschgi M, Wong G, et al: Hypersensitivity reactions to carbamazepine: characterization of the specificity, phenotype, and cytokine profile of drug-specific T cell clones. Mol Pharmacol 2003;63:732–741. 7 Naisbitt DJ, Farrell J, Chamberlain PJ, et al: Characterization of the T-cell response in a patient with phenindione hypersensitivity. J Pharmacol Exp Ther 2005;313:1058–1065.

Beeler ⴢ Pichler

8 Yawalkar N, Hari Y, Frutig K, et al: T cells isolated from positive epicutaneous test reactions to amoxicillin and ceftriaxone are drug specific and cytotoxic. J Invest Dermatol 2000; 115:647– 652. 9 Naisbitt DJ, Farrell J, Wong G, et al: Characterization of drug-specific T cells in lamotrigine hypersensitivity. J Allergy Clin Immunol 2003;111:1393– 1403. 10 Hari Y, Frutig-Schnyder K, Hurni M, et al: T cell involvement in cutaneous drug eruptions. Clin Exp Allergy 2001; 31:1398–1408. 11 Mauri-Hellweg D, Bettens, F, Mauri D, et al: Activation of drug-specific CD4+ and CD8+ T cells in individuals allergic to sulfonamides, phenytoin, and carbamazepine. J Immunol 1995;155: 462–472. 12 Pichler WJ: Delayed drug hypersensitivity reactions. Ann Intern Med 2003; 139:683–693. 13 Lee WM: Drug-induced hepatotoxicity. N Engl J Med 2003;349:474–485. 14 Stepp SE, Mathew PA, Bennett M, de Saint Basile G, Kumar V: Perforin: more than just an effector molecule. Immunol Today 2000;21:254–256. 15 Posadas SJ, Padial A, Torres MJ, et al: Delayed reactions to drugs show levels of perforin, granzyme B, and Fas-L to be related to disease severity. J Allergy Clin Immunol 2002;109:155–161. 16 Song A, Nikolcheva T, Krensky AM: Transcriptional regulation of RANTES expression in T lymphocytes. Immunol Rev 2000;177:236–245. 17 Ullman KS, Northrop JP, Verweij CL, Crabtree GR: Transmission of signals from the T lymphocyte antigen receptor to the genes responsible for cell proliferation and immune function: the missing link. Annu Rev Immunol 1990; 8:421–452. 18 Ortiz BD, Nelson PJ, Krensky AM: Switching gears during T-cell maturation: RANTES and late transcription. Immunol Today 1997;18:468–471. 19 Thornton CA, Upham JW, Wikstrom ME, et al: Functional maturation of CD4+CD25+CTLA4+CD45RA+ T regulatory cells in human neonatal T cell responses to environmental antigens/ allergens. J Immunol 2004;173:3084– 3092.

20 Del Prete GF, De Carli M, D’Elios MM, et al: Allergen exposure induces the activation of allergen-specific Th2 cells in the airway mucosa of patients with allergic respiratory disorders. Eur J Immunol 1993;23:1445–1449. 21 Romagnani S: Induction of TH1 and TH2 responses: a key role for the ‘natural’ immune response? Immunol Today 1992;13:379–381. 22 Posadas SJ, Leyva L, Torres MJ et al: Subjects with allergic reactions to drugs show in vivo polarized patterns of cytokine expression depending on the chronology of the clinical reaction. J Allergy Clin Immunol 2000;106:769– 776. 23 Naisbitt DJ, Gordon SF, Pirmohamed M, et al: Antigenicity and immunogenicity of sulphamethoxazole: demonstration of metabolism-dependent haptenation and T-cell proliferation in vivo. Br J Pharmacol 2001;133:295–305. 24 Britschgi M, Pichler WJ: Acute generalized exanthematous pustulosis, a clue to neutrophil-mediated inflammatory processes orchestrated by T cells. Curr Opin Allergy Clin Immunol 2002;2: 325–331. 25 Britschgi M, Steiner UC, Schmid S, et al: T-cell involvement in drug-induced acute generalized exanthematous pustulosis. J Clin Invest 2001;107:1433– 1441. 26 Schnyder B, Frutig K, Mauri-Hellweg D, et al: T-cell-mediated cytotoxicity against keratinocytes in sulfamethoxazol-induced skin reaction. Clin Exp Allergy 1998;28:1412–1417. 27 Yawalkar N, Egli F, Hari Y, et al: Infiltration of cytotoxic T cells in drug-induced cutaneous eruptions. Clin Exp Allergy 2000;30:847–855. 28 Hari Y, Urwyler A, Hurni M, et al: Distinct serum cytokine levels in drugand measles-induced exanthema. Int Arch Allergy Immunol 1999;120:225– 229. 29 Sachs B, Erdmann S, Al-Masaoudi T, Merk HF: In vitro drug allergy detection system incorporating human liver microsomes in chlorazepate-induced skin rash: drug-specific proliferation associated with interleukin-5 secretion. Br J Dermatol 2001; 144:316–320. 30 Pichler WJ, Zanni M, von Greyerz S, et al: High IL-5 production by human drug-specific T cell clones. Int Arch Allergy Immunol 1997;113:177–180.

In vitro Tests of T-Cell-Mediated Drug Hypersensitivity

31 Pichler WJ, Schnyder B, Zanni MP, Hari Y, von Greyerz S: Role of T cells in drug allergies. Allergy 1998;53:225– 232. 32 Brugnolo F, Annunziato F, Sampognaro S, et al: Highly Th2-skewed cytokine profile of beta-lactam-specific T cells from nonatopic subjects with adverse drug reactions. J Immunol 1999; 163:1053–1059. 33 Zanni MP, Mauri-Hellweg D, Brander C, et al: Characterization of lidocainespecific T cells. J Immunol 1997;158: 1139–1148. 34 Choquet-Kastylevsky G, Intrator L, Chenal C, et al: Increased levels of interleukin 5 are associated with the generation of eosinophilia in drug-induced hypersensitivity syndrome. Br J Dermatol 1998;139:1026–1032. 35 Merk HF: Diagnosis of drug hypersensitivity: lymphocyte transformation test and cytokines. Toxicology 2005; 209:217–220. 36 Nyfeler B, Pichler WJ: The lymphocyte transformation test for the diagnosis of drug allergy: sensitivity and specificity. Clin Exp Allergy 1997;27:175–181. 37 Hertl M, Geisel J, Boecker C, Merk HF: Selective generation of CD8+ T-cell clones from the peripheral blood of patients with cutaneous reactions to beta-lactam antibiotics. Br J Dermatol 1993;128:619–626. 38 Brander C, Mauri-Hellweg D, Bettens F, et al: Heterogeneous T cell responses to beta-lactam-modified self-structures are observed in penicillin-allergic individuals. J Immunol 1995;155:2670– 2678. 39 Sieben S, Kawakubo Y, Al Masaoudi T, Merk HF, Blomeke B: Delayed-type hypersensitivity reaction to paraphenylenediamine is mediated by 2 different pathways of antigen recognition by specific alphabeta human T-cell clones. J Allergy Clin Immunol 2002; 109:1005–1011. 40 Schmid DA, Depta JP, Pichler WJ: T cell-mediated hypersensitivity to quinolones: mechanisms and cross-reactivity. Clin Exp Allergy 2006;36:59–69. 41 Everness KM, Gawkrodger DJ, Botham PA, Hunter JA: The discrimination between nickel-sensitive and non-nickelsensitive subjects by an in vitro lymphocyte transformation test. Br J Dermatol 1990;122:293–298.

389

42 Warrington RJ, Tse KS: Lymphocyte transformation studies in drug hypersensitivity. Can Med Assoc J 1979;120: 1089–1094. 43 Barna BP, Gogate P, Deodhar SD, Moeder M: Lymphocyte transformation and radioallergosorbent tests in drug hypersensitivity. Am J Clin Pathol 1980;73:172–176. 44 Berg PA, Brattig N, Diao G-J, SchuffWerner P: Diagnose arzneimittelbedingter Nebenwirkungen mit Hilfe des Lymphozytentransformationstests. Allergologie 1983;6:77–81.

45 Schnyder B, Pichler WJ: Skin and laboratory tests in amoxicillin- and penicillin-induced morbilliform skin eruption. Clin Exp Allergy 2000;30:590– 595. 46 Pichler WJ, Tilch J: The lymphocyte transformation test in the diagnosis of drug hypersensitivity. Allergy 2004;59: 809–820. 47 Rossert J: Drug-induced acute interstitial nephritis. Kidney Int 2001;60:804– 817. 48 Kuechler PC, Britschgi M, Schmid S, et al: Cytotoxic mechanisms in different forms of T-cell-mediated drug allergies. Allergy 2004;59:613–622.

49 Schmid S, Kuechler PC, Britschgi M, et al: Acute generalized exanthematous pustulosis: role of cytotoxic T cells in pustule formation. Am J Pathol 2002; 161:2079–2086. 50 Traidl C, Sebastiani S, Albanesi C, et al: Disparate cytotoxic activity of nickelspecific CD8+ and CD4+ T cell subsets against keratinocytes. J Immunol 2000; 165:3058–3064. 51 Roujeau JC, Albengres E, Moritz S: Lymphocyte transformation test in drug-induced toxic epidermal necrolysis. Int Arch Allergy Appl Immunol 1985;78:22–24.

Prof. Dr. Werner J. Pichler Division of Allergology, Department for Rheumatology and Clinical Immunology/Allergology Inselspital CH–3010 Bern (Switzerland) Tel. +41 31 632 2264, Fax +41 31 632 2747, E-Mail [email protected]

390

Beeler ⴢ Pichler

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 391–402

In vitro Tests: Basophil Activation Tests María L. Sanz a  Pedro M. Gamboa b  A.L. De Weck a a

Department of Allergology and Clinical Immunology, University Clinic, School of Medicine, University of Navarra, Pamplona, and b Servicio de Alergia, Hospital Basurto, Bilbao, Spain

Abstract In recent years, the quantification of basophil activation by flow cytometry (basophil activation test, BAT) has proven to be a useful tool for the assessment of immediate-type responses to allergens mediated by IgE or other mechanisms in drug-allergic patients. To date, most BAT studies reported in the literature have used CD69 or CD203c as markers to quantify basophil activation after antigen-specific stimulation. Technical variations, such as the use of whole blood or isolated leukocytes, the addition of IL-3, the conditions of storage of the blood sample, the time of incubation with allergens and their concentration, can affect the results of the BATs. The BAT is more sensitive and specific than other in vitro diagnostic techniques in drug allergy. In various studies, its sensitivity in allergy to musclerelaxant drugs ranges between 36 and 97.7%, with a specificity of around 95%. For
-lactam antibiotics, BAT sensitivity is 50% and its specificity 90%. For NSAIDs, sensitivity varies between 66 and 75%; specificity is about 93%. BAT is also a useful technique in the diagnosis of isolated cases of hypersensitivity to various other drugs and substances used in some therapeutic and diagnostic procedures.

other kinds of allergic or pseudo-allergic reactions in which other activation mechanisms such as complement activation, non-IgE-mediated stimulation or non-immunological mechanisms are implicated [1]. Basophils represent less than 0.5% of the total leukocytes in peripheral blood, which makes their purification difficult. Since these cells play an important role in immediate allergic reactions, some functional in vitro tests have been developed which detect their activation. One of the first was the histamine release test, a technique that did not find extensive clinical application due to its insufficient sensitivity and specificity [2]. This is why in the past few years several groups have taken advantage of flow cytometry and developed new tools to monitor basophil activation after antigen-specific stimulation using the expression of various membrane surface markers [3, 4].

Copyright © 2007 S. Karger AG, Basel

Historical Background Introduction

Peripheral blood basophils and tissue mast cells are primary effector cells in IgE-mediated immediate allergic reactions such as rhinitis, asthma and anaphylaxis. They may also be involved in

The background of the flow cytometric basophil activation test (BAT) goes back to early studies performed by Nakagawa et al. [5], Knol et al. [6], Gane et al. [7], among others, and its application to allergy diagnosis by Sainte-Laudy et al. [3].

Basophil phenotype at rest: IgE+ CD63–

Anti-CD63 PE

Basophil phenotype after activation: IgE+ CD63+ Anti-CD63 PE

IgE

IgE Ag

CD63+ Gp53+

CD63– gp53–

Others

Anti-IgE FITC

Anti-IgE FITC

Fig. 1. Basophil phenotype at rest.

Fig. 2. Basophil phenotype after activation.

Basis of the Basophil Activation Test

ated with degranulation, can be applied in flowassisted allergy diagnosis and/or improve the technique remains to be established [8]. The basophils are able to release the content of their granules after an activation process dependent on the antigenic stimulus. Bridging of IgE receptors by the action of di- or plurivalent allergens provokes the intracytoplasmatic fusion of the granules and the fusion of their membrane with the plasmatic cell membrane, followed by exocytosis of the granules (degranulation). Molecules present on the granule membrane, such as the CD63 molecule, are then expressed on the basophil membrane upon activation (fig. 1, 2). CD63 is a tetraspan, 53-kDa granular protein that is expressed not only on basophil granules but also on monocytes, macrophages and platelets. The expression of this marker correlates with degranulation and histamine release, which makes it an ideal marker of basophil activation [14]. For in vitro stimulation with allergen, the peripheral blood cells are incubated with the suspected allergen for 15–40 min at 37 ° C. After stopping the reaction, the cells are labelled with antiCD63-PE and anti-IgE-FITC monoclonal antibodies. Two controls are used: a negative control in which the cells are incubated with the stimulation buffer used in the assay and that often con-

Flow cytometry is a useful tool for the analysis of different cellular types and can be used to identify specific cell populations, even those which are present in low amounts. It has proven to be useful in the study of allergen-induced activation [8]. In a first step the basis of these assays is the identification of basophils by specific fluorescent antibodies such as anti-IgE, anti-CD123 (IL-3 receptor) and anti-HLA-DR or anti-CCR3, and in a second step the demonstration of certain membrane phenotypes that appear after exposure to allergen. Upregulation has been described for expression of CD45 [7], CD11b and CD11c molecules, but downregulation in CD62L expression [3], as well as a decrease in the mean fluorescent intensity of IgE-carrying basophils after in vitro stimulus with the allergen [9]. However, most studies in the literature make reference to the expression of CD63 [10–12] or CD203c [13] on basophils after their in vitro activation. CD203c antibodies recognize a type-II transmembrane protein which is increased on the surface after activation. Whether in the future the novel basophil identification antigen, CRTH2 (DP2) and the activation markers CD13, CD164 and CD107a, associ-

392

Sanz ⴢ Gamboa ⴢ De Weck

tains IL-3 (negative control alias basal stimulation), and as a positive control, an anti-IgE or an anti-IgE receptor antibody can be used. With the help of receiver-operated curves (ROC; optimal sensitivity versus specificity), the determination of a positivity cutoff must be made for each allergen in order to evaluate the results. The BAT technique has generally been considered to be useful for in vitro allergologic diagnosis [8, 15, 16] and has been validated clinically for various allergens: such as inhalants [12, 17, 18], hymenoptera venoms [11, 19], latex [20–22], muscle relaxants [9, 23],
-lactam antibiotics [24–26], pyrazolones [27, 28], and NSAIDs [28, 29]. Recently, its diagnostic reliability has also been studied using recombinant allergens [30] in the monitoring of antigen-specific immunotherapy [31], in food allergy diagnosis [10, 32, 33] and in chronic urticaria [34, 35].

Some Technical Aspects

Use of Whole Blood or Isolated Leukocytes The flow cytometry technique can be used with either whole blood or isolated leukocytes. Whole blood simplifies the manipulation (fewer centrifugation steps), but it has several disadvantages such as a decreased basophil recovery, possible interference with serum components (e.g. allergen-specific IgE and IgG antibodies) that can lead to decreased sensitivity, or serum components causing unspecific activation in controls [36], and interference by aggregated platelets also expressing CD63 markers, which can be responsible for inaccurate flow cytometric counting [3, 37, 38]. For protein allergens, these differences have not appeared to markedly affect the clinical diagnostic results. For drug allergy, on the other hand, several reports of results with a whole blood technique show a lower sensitivity than those obtained with isolated leukocytes (table 3).

In vitro Tests: Basophil Activation Tests

Factors Affecting the Negative Control It is desirable to obtain an as low as possible negative control, particularly when investigating allergens causing a low specific stimulation, as is the case with drugs. In general, the negative control remains below 5% in 80% of the cases (of 504 cases: stimulation 0–5% in 79.9%; 5–10% in 13.6%; 110% in 6.5% of the cases). Natural exposure in vivo to the allergen can cause high basal activation, for example in a pollen-allergic patient studied during the pollen season [39], although some authors disagree [18]. In the latter case, the use of whole blood with the corresponding increase in pollen-specific IgG during the season may have led to misinterpretation. High basal values have also been observed when a food-allergic patient has suffered a recent reaction or is presumably continuously exposed [10] and in patients with venom allergy undergoing immunotherapy [31]. There are also several causes likely to be responsible in vitro for a high basal value, particularly pyrogens and endotoxins that could contaminate the water used in the technique or other reactives such as heparin, preservatives or even some plastic tubes or microplates. It is therefore important to use ultrapure water and cell culture grade plastic material [40]. Factors Affecting the Positive Control Most studies use mono- or polyclonal anti-IgE in the assay, but it is known that a percentage of patients do not react to anti-IgE, either by histamine release [3] or sulfidoleukotriene production [41, 42]. Polyclonal anti-IgE is recommended because many monoclonal anti-IgE antibodies are poor activators. The percentage of nonreactors ranges between 15 and 25% depending on the authors [10, 43, 44], but this seems to apply essentially to histamine release. For BAT, the percentage of non-responders reported is usually lower, near or below 10% [8]. The sensitivity of the positive control can be improved using a monoclonal anti-IgE Fc recep-

393

tor (FcR1) antibody instead of anti-IgE: this increases the activation percentage and the number of reactors [9]. In the largest series of patients investigated so far (n = 504) with the same reagents (FLOW CAST) within the frame of multicentric studies [45, 46], the percentage of true non-responders (negative for BAT and CAST) was 3.2% for CAST and 2.8% for BAT. An additional 10.5% were found to be negative for BAT but positive for CAST: they were then not true non-responders. It was later found that this dichotomy was due to dilution of the anti-IgE receptor antibody in a buffer not containing Ca+ and Mg+. It was then confirmed that BAT is more sensitive than CAST to a low external Ca+ concentration (results not shown). If the positive control with anti-IgE is negative, a negative result with antigen cannot be interpreted properly. A negative control with antiIgE seems to be more frequent in non-atopic patients [12]. However, there is no apparent correlation between the total IgE levels and the degree of basophil activation by anti-IgE (n = 104; r = 0.002; p = n.s., results not shown), which reflects that there is no relation between the basophil reactivity determined by BAT and the IgE level. This finding is in contradiction with what has been reported for histamine release and shows once more that both manifestations of basophil activation should not be entirely amalgamated. Effect of Blood Storage Blood sampling and storage for cellular tests, such as BAT, require some special conditions in order to obtain good cellular viability and functionality. The recovery of an acceptable number of reactive basophils depends on the medium, the storage time and the temperature of the blood sample. It has been shown that EDTA and ACD blood samples kept at 4 ° C maintain a suitable viability for at least 24 h, which is less the case for heparinized blood. In normal conditions at room

394

temperature, IgE-mediated reactions decrease markedly faster, probably due to the release of IgE suffered by the cells obtained ex vivo [10, 40]. At 48 h and 4 ° C, a sizeable response can still be observed, but for lower sensitivities, as it is the case with drugs, a greater number of false-negatives can be observed [10, 40]. Effect of Time of Incubation with Allergen IgE-mediated activation is a relatively short process which reaches its peak within 10–15 min. Therefore, maximum activation of the basophils can be expected within this time range. However, if BAT is to be performed concomitantly with sulfidoleukotriene determination in the supernatant, a longer incubation time is recommended [47]. Some other manifestations of basophil activation, such as expression of CD203c [8] or activation by non-IgE mechanisms (e.g. C5a, fMLP) are faster and require only a few minutes to reach their peak. Effect of Allergen Concentration It is very important to determine the optimal concentrations that provoke the maximum cellular activation for each allergen by means of dose-response curves. If the concentration that provokes antigen-specific stimulation is narrow but shows great individual variations, it will be necessary to use several concentrations per allergen in diagnostic tests. If, on the contrary, the activation by allergen takes place over a broad concentration range, with few individual variations, it becomes possible to use a single protein allergen concentration. Here too, there seems to be some contrast between BAT and what has been recommended for histamine release tests. In contrast to histamine release, the dose-response curves for BAT appear to be sigmoid in most instances and not bell-shaped. For small molecule-like drugs, the range of concentrations to which the cells respond is usually narrow; for example, 0.001–0.0001 mM for muscle relaxants [9]. In any case, for drugs, the

Sanz ⴢ Gamboa ⴢ De Weck

use of at least two consecutive concentrations around the optimal range is recommended. It is essential to use allergens or drugs free of preservative agents, such as glycerol, pyrogens, and any substances that could provoke unspecific activation or, on the contrary, cytotoxicity. For each new drug to be tested, it is necessary to include appropriate healthy controls tolerating the drug. ROC must be established in order to determine the correct cutoff points. Preactivation with IL-3 Pre- or simultaneous incubation with IL-3 has been found essential for optimal sulfidoleukotriene production (CAST) [47] and also increases histamine release [47], but there is no general consensus on its need for improving the sensitivity of BAT. Some authors have reported that it increases CD63 expression, increasing thereby the assay’s sensitivity, which is relevant for allergens causing little specific stimulation, such as drugs [8]. Reproducibility Intra-assay reproducibility assessed by duplicate determinations for different allergens is very high, which allows performance of a single test for each allergen concentration [48]. The reproducibility of BAT was found in drug allergy to be quite remarkable with an intra-assay variation coefficient of 0.89 (p ! 0.0001) for
-lactams, 0.99 (p ! 0.0001) for metamizol and 0.92 (p ! 0.0001) for NSAIDs. Evaluation of Results For appropriate evaluation of the results, two values should be taken into account: (1) the absolute number of basophils evaluated, which should be over 150, and (2) the percentage of activated basophils. In the negative (non-stimulated) control, the percentage of activated basophils is usually below 5%. The positive control after activation of the cells with anti-IgE anti-receptor has been discussed above.

In vitro Tests: Basophil Activation Tests

Some authors also describe a decreased mean fluorescence emitted by the cells activated in vitro by allergen [9, 19, 49]; a phenomenon which does not seem to affect the percentage of basophils expressing CD63. BAT Positivity Criteria To establish cutoff points for each allergen, it is necessary to set up ROC. It is also important to take into account the following considerations: (1) the negative control can show variable values, although it is usually !5%, and (2) some allergens, particularly foods, can provoke unspecific stimulations. Other allergens, such as drugs, usually cause lower responses than those obtained with protein allergens. In our experience, the cutoff points offering the highest specificity and sensitivity values determined by ROC [25, 28] are the following: for inhalant allergens 115%; food allergens 115%; latex 110%, hymenoptera venoms 110%;
-lactam antibiotics 15% and SI 12; metamizol 15% and SI 15, and aspirin (ASA) and NSAIDs 15% and SI 12 (SI: stimulation index = stimulation by allergen/basal stimulation or negative control). Some authors found a correlation between the degree of clinical sensitivity and the percentage of activated basophils [36], but other authors have not confirmed this observation [10]. For muscle relaxants, a correlation between the degree of skin sensitivity and CD63 expression has been reported [9]. Comparison with Other in vitro Diagnostic Techniques Comparisons among various in vitro tests have been made for several allergens. The BAT usually shows a good correlation with histamine release and/or sulfidoleukotriene determinations (CAST) [10, 12, 25, 50], but BAT seems more sensitive and specific than the other tests [10, 51]. For diagnosis of immediate reactions to drugs in general [49] and for immediate reactions to
-

395

Table 1. BAT in allergy to muscle relaxants Reference

Clinical manifestations

Number of patients

BAT sensitivity, %

BAT specificity, %

Abuaf et al. [9] Monneret et al. [53] Sudheer et al. [55] Kvedariene et al. [56] Ebo et al. [57]

Anaphylaxis Anaphylaxis Anaphylaxis Anaphylaxis Anaphylaxis

41 39 21 47 14

64 54 79 36 91.7

93 100 100 93 100

lactams in particular, the joint use of BAT and CAST offers better results and allows detection of up to 80% of the cases. The same holds true for pyrazolones [51].

Clinical Applications of BAT to Diagnosis of Drug Allergy

BAT in Diagnosis of Hypersensitivity Reactions to Neuromuscular-Blocking Agents Several reports have been published relating the usefulness of the BAT technique in the diagnosis of hypersensitivity reactions to muscle relaxants (table 1). Abuaf et al. [9] studied basophil activation to muscle relaxants, such as suxamethonium, gallamine, vecuronium, and pancuronium. In a series of 41 patients, they observed a characteristic dose-response curve, but indicated that high drug concentrations can act as unspecific basophil activators. The sensitivity of the technique reported by them is 64% and the specificity 93%. Monneret et al. [4] applied a three-colored flow cytometry assay (IgE, CD45, CD63) in 4 patients with muscle relaxant-induced anaphylaxis, observing that the results correlate with intradermal tests. In a subsequent study in 39 patients with reactions to muscle relaxants, the same authors found a sensitivity of BAT of 54% with a specificity of 100% [23, 53].

396

Aly Hassan et al. [54] published good results of the BAT technique in suxamethonium allergy. More recently, Sudheer et al. [55] compared the CD63 and CD203c expressions and measured histamine release in 21 patients who were referred with possible perioperative anaphylaxis. The sensitivity of CD63, CD203c, basophil histamine release and skin prick test for the muscle relaxants was found to be 79, 36, 36 and 64%, respectively. In 47 patients Kvedariene et al. [56] observed a sensitivity of 36% and a specificity of 93% when performing the Basotest (CD63 in whole blood). On the other hand, Ebo et al. [57] reported the sensitivity of a similar BAT technique to be 91.7% and 100% specificity for rocuronium in a study including 14 patients with anaphylaxis and positive skin tests to rocuronium. BAT in
-Lactam Allergy We studied the diagnostic reliability of the BAT and CAST techniques in a series of 81 patients with immediate allergy to
-lactams and in 30 healthy controls [25, 48]. BAT (FLOW-CAST) and sulfidoleukotriene determination by CAST ELISA (Bühlmann Laboratories, Allschwil, Switzerland), as well as specific IgE determination (CAP-FEIA, Phadia, Uppsala, Sweden) were performed. The results are discussed in detail below. BAT in
-Lactam-Allergic Patients with Positive Skin Tests The patients included in this study presented with anaphylaxis or urticaria-angioedema with-

Sanz ⴢ Gamboa ⴢ De Weck

in a few minutes or hours after administration of
-lactams. 58 of these patients had positive skin tests at least to one of the allergens tested: a minor determinant mixture, penicilloyl polylysine, penicillin, ampicillin, amoxicillin or cephalosporin that was sometimes the culprit drug. As control group, 30 non-allergic patients were included. All of them had negative skin tests to
-lactams and tolerated their administration. The overall sensitivity of BAT was 50.7% (35/69) and its specificity 93.3% (28/30) for a cutoff point of basophil activation of 15% and a SI of 12 (SI = activation with drug/basal activation). Amoxicillin shows the highest sensitivity and specificity values of all the
-lactams tested, with significant differences compared with penicillin, ampicillin and a minor determinant mixture. The highest positive and negative predictive values were also obtained with amoxicillin (table 2). The positive predictive value of BAT was 6.4% and the negative predictive value 97.7%. These values are higher than those obtained with other in vitro techniques. In our study the specific IgE determination by CAP (Pharmacia) showed a sensitivity of 36.7% and a specificity of 83.3%. CAST presented a sensitivity of 47.7% and a specificity of 83.3%. The joint use of CAP (Pharmacia) and BAT allows identification of 65.2% of the
-lactam-allergic patients with a specificity of 83.3%. BAT in
-Lactam-Allergic Patients with Negative Skin Tests 23
-lactam-allergic patients with negative skin tests were included in this group; 5 of them had a positive CAP to
-lactams, 12 a negative CAP but a positive challenge test, and 6 with negative CAP had presented more than two episodes of anaphylaxis/urticaria after administration of amoxicillin. Among the 23
-lactam-allergic patients with negative skin tests, BAT was positive in 39.1% with a specificity of 93.3%. CAST was positive in 22.7% with a specificity of 83.3%, and CAP showed a sensitivity of 21.7% with a specificity of

In vitro Tests: Basophil Activation Tests

Table 2. BAT sensitivity (SE), specificity (SP) and positive (PPV) and negative (NPV) predictive values in
-lactamallergic patients with positive skin test

Penicillin Ampicillin Amoxicillin PPL MDM Global

SE, %

SP, %

PPV, %

NPV, %

20* 22.22* 33.3 15.25 16.67* 50.7

93.33 93.33 96.67 95.45 95.45 93.3

8.48 9.34 23.63 9.39 10.17 6.4

97.41 97.48 97.91 97.32 97.37 97.7

PPL = Penicilloyl polylysine; MDM = minor determinant mixture. * p < 0.05 compared to amoxicillin.

83.3%. The association of the three techniques allows diagnosis in 60.9% of the patients of this group, with a specificity of 70%. Using the Basotest Torres et al. [24] found similar results. In their study, the Basotest with different
-lactam allergens was performed in 70 patients with immediate allergic reactions, and classified into 3 groups: (A) positive skin test independent of the CAP/RAST immunoassay; (B) negative skin test but positive to CAP/RAST, and (C) negative skin test and CAP/RAST but positive drug provocation test. They obtained the following results: among the 70 patients, 34 (48.6%) were positive to the Basotest (sensitivity 48.6%), 31 (44.3%) to CAP/RAST, and 46 (65.7%) positive to either one or both. Considering the different groups, the Basotest was positive in 50.9% of patients in group A, 60% in group B and 14.3% in group C. Specificity was 91.3%. Positivity to the
-lactam allergens was 28.6% to amoxicillin, 21.7% to BP, 20% to benzylpenicilloyl-poly-L-lysine, 12.5% to ampicillin and 2.2% to the minor determinant mixture. In patients with cephalosporin reactions, the Basotest to the culprit cephalosporin was positive in 77.7%. In table 3, we summarize the results of the studies discussed, including some additional re-

397

Table 3. BAT results in immediate allergic reactions to
-lactam and in NSAID hypersensitivity Reference

Antigen

BAT Skin tests technique

Torres et al. [24] Sanz et al. [25] Gamboa et al. [48] Arnoux et al. [58] Kvedariene et al. [59] Erdmann et al. [60] De Weck et al. [45] Sanz et al. [28], Gamboa et al. [29] Kvedariene et al. [59] Erdmann et al. [60] De Weck et al. [46]


-Lactams
-Lactams
-Lactams
-Lactams
-Lactams
-Lactams
-Lactams

Basotest FLOW FLOW Basotest Basotest Basotests FLOW

NSAIDs NSAIDs NSAIDs NSAIDs

FLOW Basotest Basotest FLOW

Positive and negative Positive Negative Positive Positive Positive Positive and negative

ports provided in abstract form [58, 59]. It is to be noted that, with the exception of Torres et al. [24], the sensitivities obtained in whole blood (Basotest) appear to be markedly lower [58–60] than those obtained with isolated leukocytes (FLOWCAST technique). Our original results have recently been confirmed by a multicentric study involving 10 European groups, 181
-lactam-allergic patients and 80 controls [45]. In that study, BAT sensitivity was 47.8% but increased to 67% when combined with CAST. Specificity was 92%. The sensitivity of specific IgE, on the other hand, was only 38%. Therefore these results show that, although not yet optimally sensitive, BAT and CAST, particularly when combined, are the most reliable in vitro tests to confirm a diagnosis of immediate-type
-lactam allergy. BAT in Allergy to Pyrazolones (Metamizol) In a study by our group [61], we assessed the diagnostic reliability of BAT in a group of 26 metamizol-allergic patients and in 30 controls compared with other in vitro techniques (CAST, CAP). These patients were strictly only reacting to metamizol and not to other NSAIDs; they also had positive skin tests. It is therefore assumed that they belong to the IgE-mediated group of al-

398

Number of BAT BAT patients sensitivity, % specificity, % 70 58 23 35 22 20 181

48.6 50.7 39.1 9 34 35 46

60 15 20 150

74 25 30 76

91.3 93.3 93.3 100 83 93 92 80

Table 4. TAB sensitivity (SE), specificity (SP), and positive (PPV) and negative (NPV) predictive values in NSAID hypersensitivity

ASA Paracetamol Metamizol Diclofenac Naproxen Global1 1 ASA

SE, %

SP, %

PPV, %

NPV, %

43.3 11.7 15 43.3 54.8 66.3

100 100 100 93.3 74.1 93.3

100 100 100 6.1 2 8.7

99.4 100 99.1 99.3 99.3 99.6

+ paracetamol + metamizol + diclofenac.

lergies due to the analgesics and not to the NSAID hypersensitivity syndrome. BAT was positive in 11 of the 26 patients, whereas all the controls were negative. The sensitivity of the technique was 42.3% and the specificity 100% for a cutoff of basophil activation of 15% and a SI of 12. Assuming a prevalence in the general population similar to that of the
-lactam antibiotics (1%), the positive predictive value of BAT is 100% and the negative predictive value 99.4%. CAP was negative in the 17 cases tested, and CAST offered a sensitivity of 52% with a specificity of 90% (table 4).

Sanz ⴢ Gamboa ⴢ De Weck

The joint use of BAT and CAST allows detection of 76% of the cases, whereas the association of BAT, CAST and skin tests only increases the reliability of diagnosis slightly (76.9%). BAT in the NSAID Hypersensitivity Syndrome The syndrome of hypersensitivity to NSAID is characterized by a triad of asthma, rhinitis and nasal polyposis, on the one hand, and by urticaria and angioedema on the other hand, elicited by a large number of NSAIDs. The current hypothesis is that these reactions are due to the pharmacological effect of the eliciting drugs, namely inhibition of cyclooxygenase type 1 (COX-1). Up to now, it was generally contended that no in vitro diagnostic test exists for that condition, although several reports, in part contested, have suggested that such cases may yield a positive CAST test to NSAIDs [42, 47]. Following some anecdotal reports [3, 27, 38], our group [29] studied 60 patients diagnosed with NSAID hypersensitivity (38 with skin manifestations, 20 with respiratory manifestations and 2 with both), and 30 controls. BAT was performed in all and the production of sulfidoleukotrienes by CAST (Bühlmann Laboratories, Switzerland) was determined concomitantly with different NSAIDs: ASA, paracetamol, metamizol, diclofenac and naproxen at different concentrations. In this group of patients, BAT with ASA showed a sensitivity of 43.3% and a specificity of 100%, a positive predictive value of 100% and a negative predictive value of 99.4%. The sensitivity and specificity values were 11.7 and 100% for paracetamol, 15 and 100% for metamizol, 43.3 and 93.3% for diclofenac. The overall sensitivity for the four NSAIDs combined was 63.3% and the specificity 93.3%. If we limit the drugs tested to ASA and diclofenac for practical reasons, the sensitivity of BAT is 58.3% and the specificity 93.3% (table 4). In that study, CAST showed a sensitivity of 38.3% and a specificity of 76.7%.

In vitro Tests: Basophil Activation Tests

The joint use of BAT and CAST in this group did not significantly increase the diagnostic reliability of BAT. These results have recently been confirmed by a large multicenter study including 12 European groups, 150 NSAID-hypersensitive patients and 163 controls [46]. In that study, the sensitivity of BAT when combining ASA, diclofenac and naproxen was found to be 75% with a specificity of around 90%. However, it was also observed that the mode of isolation of leukocytes may influence the reactions in ASA-tolerant controls: some normal individuals may also show dose-dependent basophil activation to various NSAIDs, which appears related to their pharmacological COX-1 inhibiting effect. Therefore the NSAIDhypersensitivity syndrome appears related to a shift in the normal pharmacological dose response. An optimal differentiation in vitro by BAT is obtained by testing two concentrations each of ASA, diclofenac and naproxen (the socalled ASA/diclofenac/naproxen index). BAT in the Diagnosis of Hypersensitivity Reactions to Various Drugs In our hands flow cytometric analysis of activated basophils has proven to offer a confirmatory test in a number of drug hypersensitivities for which no other in vitro tests are available. The BAT technique is also useful in the study of isolated cases of hypersensitivity to various substances and in some therapeutic or diagnostic procedures such as the diagnosis of hemodialysis-associated anaphylactic and anaphylactoid reactions [62]. It has proven to be useful in the in vitro diagnosis of allergy to plasma expanders such as hydroxyl ethyl starch [63], gelatins [64] or gelofusine [65]. In the diagnosis of an anaphylactic reaction after artificial insemination, the involvement of bovine serum albumin (BSA) in the implantation medium was confirmed by means of BAT [66]. Two cases showed the usefulness of BAT in the diagnosis of allergic reactions to pat-

399

ent blue dye [67, 68]. Other studies showed its reliability in the diagnosis of anaphylaxis to cyclosporine [69]; in omeprazol allergy [70], with heparins [71]; with chlorhexidine [72]; with hyaluronidase [73], and in one case of anaphylaxis to viscotoxins of mistletoe (Viscum album) [74].

Conclusions

According to the results published to date, the BAT constitutes a diagnostic tool of interest in the study of immediate-type hypersensitivity reactions to drugs [8, 64, 75].

This technique offers several advantages: (1) it reproduces in vitro hypersensitivity mechanisms involved in immediate-type allergic reactions; (2) it allows the diagnosis of allergic and pseudoallergic reactions particularly for drugs which are often not detectable by serological techniques, such as determination of specific IgE; (3) the results are obtained within a few hours; (4) several different drugs can be tested simultaneously with a small amount of blood, and (5) it may frequently make it possible to avoid performing exposition tests with the drug.

References 1 Schleimer RP, Fox CC, Togias A, Norman P, Lichtenstein LM: Role of human basophils and mast cells in the pathogenesis of allergic diseases: J Allergy Clin Immunol 1985;76:369–374. 2 Hamilton RG, Adkinson NF: Clinical laboratory assessment of IgE-dependent hypersensitivity. J Allergy Clin Immunol 2003;111(suppl):S687–S701. 3 Sainte-Laudy J, Sabbah A, Vallon C, Guerin JC: Analysis of anti-IgE and allergen induced human basophil activation by flow Cytometry. Comparison with histamine release. Inflamm Res 1998,47:401–408. 4 Monneret G, Gutowski MC, Bienvenu J: Detection of allergen-induced basophil activation by expression of CD63 antigen using a tricolour flow cytometric method. Clin Exp Immunol 1999;115: 393–396. 5 Nakagawa T, Stadler BM, De Weck AL: Flow-cytometric analysis of human basophil degranulation. I. Quantification of human basophils and their degranulation by flow-cytometry. Allergy 1981;36:39–47. 6 Knol EF, Mul FPJ, Jansen H, Calafat J, Roos D: Monitoring human basophil activation via CD63 monoclonal antibody 435. J Allergy Clin Immunol 1991; 88:328–338. 7 Gane P, Pecquet C, Lambin P, Abuaf N, Leynadier F: Flow cytometry evaluation of human basophils. Cytometry 1993;14:344–348.

400

8 Ebo DG, Sainte-Laudy J, Bridts CH, Mertens CH, Hagendorens MM, Schuerwegh AJ, De Clerck LS, Stevens WJ: Flow-assisted allergy diagnosis: current applications and future perspectives. Allergy 2006;61:1028–1039. 9 Abuaf N, Rajoely B, Ghazouani E, Levy DA, Pecquet C, Chabane H, Leynadier F: Validation of flow cytometry assay detecting in vitro basophil activation for the diagnosis of muscle relaxant allergy. J Allergy Clin Immunol 1999; 104:411–418. 10 Moneret-Vautrin DA, Sainte Laudy J, Kanny J, Fremont S: Human basophil activation measured by CD 63 expression and LTC4 release in IgE-mediated food allergy. Ann Allergy Asthma Immunol 1999;82:33–40. 11 Sainte Laudy J, Drouet M, Lauret MG, Loiry M, Sabbah A: Test d’activation des basophiles par cytometrie en flux dans l’allergie aux venins d’hyménoptères. Allerg Immunol (Paris) 1999; 31:11–14. 12 Sanz ML, Sánchez G, Vila L, Uasuf C, Chazot M, Diéguez I, Oehling A, de Weck AL: Ag-induced basophil activation: CD63 cell expression detected by flow cytometry in patients allergic to Dermatophagoides pteronyssinus and Lolium perenne. Clin Exp Allergy 2001;31:1–8. 13 Boumiza R, Debard AL, Monneret G: The basophil activation test by flow cytometry: recent developments in

14

15

16

17

18

19

clinical studies, standardization and emerging perspectives. Clin Mol Allergy 2005;3:3–9. Bochner BS: Systemic activation of basophils and eosinophils: markers and consequences. J Allergy Clin Immunol 2000;160:S292–S302. Ebo DG, Hagendorens MM, Brits CH, Scherwegh AJ, De Clerck LS, Stevens WJ: In vitro allergy diagnosis: should we follow the flow? Clin Exp Allergy 2004;34:332–339. Kleine-Tebbe J, Erdmann S, Knol EF, Macglashan DW Jr, Poulsen KL, Gibbs BF: Diagnostic tests based on human basophils: potentials, pitfalls and perspectives. Int Arch Allergy Immunol 2006;141:79–90. Paris-Kohler A, Demoly P, Persi L, Lebel B, Bousquet J, Arnoux B: In vitro diagnosis of cypress pollen allergy by using cytofluorometric analysis of basophils (Basotest). J Allergy Clin Immunol 2000,105:339–345. Saporta M, Kamei S, Persi L, Bousquet J, Arnoux B: Basophil activation during pollen season in patients monosensitized to grass pollens. Allergy 2001;56: 442–445. Sainte-Laudy J, Sabbah A, Drouet M, Lauret MG, Loiri M: Diagnosis of venom allergy by flow cytometry. Correlation with clinical history, skin tests, specific IgE, histamine and leukotriene C4 release. Clin Exp Allergy 2000;30: 1166–1171.

Sanz ⴢ Gamboa ⴢ De Weck

20 Sanz ML, Gamboa PM, Garcia-Aviles C, Vila L, Dieguez I, Antepara I, de Weck AL: Flow-cytometric cellular allergen stimulation test in latex allergy. Int Arch Allergy Immunol 2003;130: 33–39. 21 Sanz ML, García-Avilés MC, Gamboa PM, Maselli JP, Diéguez I, De Weck AL: The use of a flow-cytometric basophil activation test (FAST) for the in vitro diagnosis of latex allergy (abstract 869). J Am Cancer Inst 2002;109. 22 Ebo DG, Lechkar B, Schuerwegh AJ, Bridts CH, De Clerk LS, Stevens WJ: Validation of a two-color flow cytometric assay detecting in vitro basophil activation for the diagnosis of IgE-mediated natural rubber latex allergy. Allergy 2002;57:706–712. 23 Monneret G, Benoit Y, Debard AL, Gutowski MC, Topenot I, Bienvenu J: Monitoring of basophil activation using CD63 and CCR3 in allergy to muscle relaxant drugs. Clin Immunol 2002; 102:192–199. 24 Torres MJ, Padial, Mayorga C, Fernandez T, Sanchez-Sebate E, Cornejo-Garcia JA, Antunez C, Blanca M: The diagnostic interpretation of basophil activation test in immediate allergic reactions to betalactams. Clin Exp Allergy 2004;34:1768–1775. 25 Sanz ML, Gamboa PM, Antépara I, Uasuf C, Vila L, Garcia-Aviles C, Chazot M, De Weck AL: Flow-cytometric basophil activation test by detection of CD63 expression in patients with immediate type reactions to betalactam antibiotics. Clin Exp Allergy 2002;32:277–286. 26 Sanz ML, De Weck AL, Uasuf C, Gamboa PM, Chazot M: Use of flow cytometry to assess basophil activation in patients allergic to betalactam antibiotics. Correlation between flow cytometric allergen stimulation test (FAST) and other in vivo and in vitro tests. Int Arch Allergy Immunol 2001;124:307–308. 27 Sabbah A, Sainte-Laudy J, Drouet M, Lauret MG, Loiry M: Réactions anaphylactiques ou anaphylactoïdes au paracétamol. A propos de 3 cas. Allerg Immunol (Paris) 1997;29:60–62. 28 Sanz ML, Gamboa PM, De Weck AL: A new combined test with flowcytometric basophil activation and determination of sulfidoleukotrienes is useful for in vitro diagnosis of hypersensitivity to aspirin and other nonsteroidal antiinflammatory drugs. Int Arch Allergy Immunol 2005;136:58–72.

In vitro Tests: Basophil Activation Tests

29 Gamboa PM, Sanz ML, Caballero MR, Urrutia I, Antépara I, Esparza R, De Weck AL: The flow-cytometric determination of basophil activation induced by aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) is useful for in vitro diagnosis of the NSAID hypersensitivity syndrome. Clin Exp Allergy 2004;34: 1448–1457. 30 Erdmann SM, Sachs B, Schmidt A, Merck HF, Scheiner O, Moll-Slodowy S, Sauer I, Kwiecien R, Maderegger B, Hoffmann-Sommergruber K: In vitro analysis of birch-pollen-associated food allergy by use of recombinant allergens in the basophil activation test. Int Arch Allergy Immunol. 2005;136: 230–238. 31 Erdmann SM, Sachs B, Kwiecien R, Moll-Slodowy S, Sauer I, Merk HF: The basophil activation test in wasp venom allergy: sensitivity, specificity and monitoring specific immunotherapy. Allergy 2004;59:1102–1109. 32 Ebo DG, Hagendorens MM, Bridts CH, Schuerwegh AJ, De Clerck LS, Stevens WJ: Flow cytometric analysis of in vitro activated basophils, specific IgE and skin tests in the diagnosis of pollenassociated food allergy. Cytometry B Clin Cytom 2005;64:28–33. 33 Garcia-Aviles C, Sanchez-Lopez G, Sanz ML, Uasuf C, Dieguez I, Chazot M, De Weck AL: Flow cytometric allergen stimulation test (FAST) in the in vitro diagnosis of food allergy. Allergy 2000;55(suppl 63):128. 34 Wedi B, Novacovic V, Koerner M, Kapp A: Chronic urticaria serum induces histamine release, leukotriene production, and basophil CD63 surface expression-inhibitory effects of anti-inflammatory drugs. J Allergy Clin Immunol 2000;105:552–560. 35 Gyimesi E, Sipka S, Danko K, Kiss E, Hidvegi B, Gal M, Hunyadi J, Irinyi B, Szegedi A: Basophil CD63 expression assay on highly sensitized atopic donor leucocytes. A useful method in chronic autoimmune urticaria. Br J Dermatol 2004;151:288–296. 36 Stevens S, Drouet M, Lauret MG, Loiry M, Sabbah A: Basophil activation test using flow cytometry in hymenoptera venom allergy. Allerg Immunol (Paris) 1999;31:11–14.

37 Metzelaart M, Wijngaard PLJ, Peters PJ, Sioxma JJ, Nieuwenhuist HK, Clevers HC: CD63 antigen: a novel lysosome membrane glycoprotein, cloned by a screening procedure for intracellular antigens in eukaryotic cells. J Biol Chem 1991;266:3239–3245. 38 Sainte-Laudy J: Application of flow cytometry to the analysis of activation of human basophils. Immunologic validation of the method. Allerg Immunol (Paris) 1998;30:41–43. 39 Sabbah A, Drouet M, Sainte-Laudy J, Lauret MG, Loiry M: Apport de la cytométrie en flux dans le diagnostic allergologique. Allerg Immunol (Paris) 1995;29:15–21. 40 de Weck AL, Sanz ML: Flow cytometric cellular allergen stimulation test (FAST/FLOW-CAST). Technical and clinical evaluation of a new diagnostic test in allergy and pseudo-allergy. ACI Int 2002;14:204–215. 41 Crockard AD, Ennis M: Laboratorybased allergy diagnosis: should we go with the flow? Clin Exp Allergy 2001; 31:975–977. 42 De Weck AL, Sanz ML: Cellular allergen stimulation test (CAST) 2003, a review. J Investig Allergol Clin Immunol 2004;14:253–273. 43 Sabbah A, Sainte-Laudy J: Flow cytometry applied to the analysis of lymphocyte and basophil activation. ACI Int 1996;8:116–119. 44 Sainte-Laudy J, Valon C, Guérin JC: Analyse de l’expression membranaire du marqueur CD63 par activation du basophile humain. Application au diagnostic allergologique. Allerg Immunol (Paris) 1994;26:211–214. 45 De Weck AL, Sanz ML, Gamboa PM, Aberer W, Bilo B, Blanca M, Torres M, Mayorga L, Campi P, Drouet M, SainteLaudy J, Romano A, Jermann T, Weber JM; ENDA group: Diagnosis of immediate-type betalactam allergy in vitro by flowcytometry (Flow-CAST) and sulphidoleukotriene production. A multicentric study (abstract). 25th Congr EAACI, Vienna, 2006, p 196. 46 De Weck AL, Sanz ML, Gamboa PM, Aberer W, Sturm G, Blanca M, Torres M, Mayorga L, Correia S, Kowalski M, Medrala W, Sczceklik A; ENDA group: Diagnosis of hypersensitivity to nonsteroidal anti-inflammatory drugs (NSAIDs) in vitro by flowcytometry (Flow-CAST) and sulfidoleukotriene assay. A multicentric study (abstract). 25th Congr EAACI, Vienna, 2006, p 188.

401

47 De Weck AL: Zellulärer Allergen Stimulierungstest (CAST). Eine Uebersicht und kritische Auswertung der klinischen Anwendung in der Allergiediagnose. Allergologie 1997;20:487– 502. 48 Gamboa PM, Garcia-Aviles MC, Urrutia I, Antepara I, Esparza R, Sanz ML: Basophil activation and sulfidoleukotriene production in patients with immediate allergy to betalactam antibiotics and negative skin tests. J Investig Allergol Clin Immunol 2004;14:278– 283. 49 Gane P, Pecquet C, Crespeau H, Lambin P, Abuaf N, Leynadier F, Rouger P: Flow cytometric monitoring of allergen induced basophil activation. Cytometry 1995;19:361–365. 50 Sainte-Laudy J, Le Provost A, André C, Vallon C: Comparison of four methods for human basophil activation measurement: alcian blue staining, histamine and LTC4 release and flow cytometry; in Basomba A, Hernandez MD, De Rojas F (eds): Proc. 16th EAACI Congress, 1995, pp 257–260. 51 De Weck AL, Gamboa PM, Sanz ML: Clinical evaliationon in vitro tests in diagnosis of immediate allergic reactions to betalactam antibiotics. ACI Int 2002;14:185–192. 52 Ebo DG, Hagendorens MM, Bridts CH, Stevens WJ: Immediate-type allergy to drugs and related compounds: evaluation and management. Acta Clin Belg 2005;60:350–361. 53 Monneret G, Benoit Y, Gutowski MC, Bienvenu J: Detection of basophil activation by flow cytometry in patients with allergy to muscle-relaxant drugs. Anesthesiology 2000;92:275–277. 54 Aly Hassan IM, Crockard AD, Asghar MS, Edgar JD, Atkinson S: Basotest and suxamethonium allergy. Allergy 2001; 56:1016–1017. 55 Sudheer PS, Hall JE, Read GF, Rowbottom AW, Williams PE: Flowcytometric investigation of peri-anaesthetic anaphylaxis using CD63 and CD203c. Anaesthesia 2005;60:251–256.

56 Kvedariene V, Kamey S, Ryckwaert Y, Rongier M, Bousquet J, Demoly P, Arnoux B: Diagnosis of neuromuscular blocking agent hypersensitivity reactions using cytofluorimetric analysis of basophils. Allergy 2006;61:311–315. 57 Ebo DG, Brits CH, Hagendorens MM, Mertens CH, De Cleck LS, Stevens WJ: Flow-assisted diagnostic management of anaphylaxis from rocuronium bromide. Allergy 2006;61:935–939. 58 Arnoux B, Kvedariene V, Rongier M, Bousquet J, Demoly P: Betalactam allergy in clinical practice. Relevance of clinical history and/or laboratory tests (abstract). J Allergy Clin Immunol 2004;113:s71. 59 Kvedariene V, Arnoux B, Rongier M, Bousquet PF, Demoly P: Benefit of the basophil activation test for the diagnosis of immediate drug hypersensitivity reaction (abstract 371). ACI Int J World Allergy Org 2005(suppl 1):128. 60 Erdmann SM, Ventocilla S, Moll-Slodowy S, Sauer L, Merk HF: Basophil activation tests in the diagnosis of drug reactions. Hautarzt 2005;56:38–43. 61 Gamboa PM, Sanz ML, Caballero MR, Antépara I, Urrutia I, Jáuregui I, González G, Diéguez I, De Weck AL: Use of CD63 expression as a marker of in vitro basophil activation and leukotriene determination in metamizol allergic patients. Allergy 2003;58:312–317. 62 Ebo DG, Bosmans JL, Couttenye MM, Stevens WJ: Haemodialysis-associated anaphylactic and anaphylactoid reactions. Allergy 2006;61:211–220. 63 Ebo DG, Schuerwegh A, Stevens WJ: Anaphylaxis to starch. Allergy 2000;55: 1098–1099. 64 Sanz ML, Maselli JP, Gamboa PM, García-Avilés MC, Oehling A, Diéguez I, de Weck AL: Flow-cytometric basophil activation test. A review. J Investig Allergol Clin Immunol 2002;12:143–154. 65 Apostolou E, Deckert K, Puy R, Sandrini A, de Leon MP, Douglas JA, Rolland JM, O’Heir RE: Anaphylaxis to gelofusine confirmed by in vitro basophil activation test: a case series. Anesthesia 2006;61:264–268.

66 Orta M, Ordoqui E, Artanzabal A, Fernández C, Bartolome B, Sanz ML: Anaphylactic reaction after artificial insemination. Ann Allergy Asthma Immunol 2003;90:446–451. 67 Cottineau C, Beydon L, Drouet M, Renier G, Nicolie B: Allergic reaction to patent blue dye with positive detection of basophil activation by flow cytometry Ann Fr Anesth Reanim 2005;24: 541–542. 68 Ebo DG, Wets RD, Spiessens TK, Bridts CH, Stevens WJ: Flow-assisted diagnosis of anaphylaxis to patent blue. Allergy 2005;60:703–704. 69 Ebo DG, Piel GC, Conraads V, Stevens WJ: IgE-mediated anaphylaxis after first intravenous infusion of cyclosporine. Ann Allergy Asthma Immunol 2001;87:243–245. 70 Gamboa PM, Sanz ML, Urrutia I, Jauregui I, Antepara I, Dieguez I, De Weck AL: CD63 expression by flow cytometry in the in vitro diagnosis of allergy to omeprazole. Allergy 2003;58:538– 539. 71 Ebo DG, Haine SE, Hagendorens MM, Bridts CH, Conraads VM, Vorlat A: Hypersensitivity to nadroparin calcium: case report and review of the literature. Clin Drug Invest 2004;24:421– 426. 72 Ebo DG, Bridts CH, Stevens WJ: Anaphylaxis to an urethral lubricant: chlorhexidine as the ‘hidden’ allergen. Acta Clin Belg 2004;59:358–360. 73 Ebo DG, Goossens S, Opsomer F, Bridts CH, Stevens WJ: Flow-assisted diagnosis of anaphylaxis to hyaluronidase. Allergy 2005;60:1333–1334. 74 Bauer C, Oppel T, Rueff F, Przybilla B: Anaphylaxis to viscotoxins of mistletoe (Viscum album) extracts. Ann Allergy Asthma Immunol 2005;94:86–89. 75 Sanz ML, Gamboa PM, Garcia-Aviles C, De Weck A: Drug hypersensitivities: which room for biological tests? Allerg Immunol (Paris) 2005;37:230–235.

Prof. María L. Sanz Departamento de Alergología e Inmunología Clínica Clínica Universitaria de Navarra Universidad de Navarra, Apartado 4209 ES–31080 Pamplona (Spain) Tel. +34 948 255 400, Fax +34 948 296 500, E-Mail [email protected]

402

Sanz ⴢ Gamboa ⴢ De Weck

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 404–412

Desensitization with Antibiotics Roland Solensky Division of Allergy and Immunology, The Corvallis Clinic, Corvallis, Oreg., USA

Abstract

Introduction

Rapid antibiotic desensitization is a type of induction of tolerance by which patients with IgE-mediated allergies to antibiotics may safely receive these drugs. Only patients with type I allergy are candidates for this procedure. Rapid desensitization is intended for clinical situations in which there is an absolute need for an antibiotic and alternate noncross-reacting antibiotics cannot be substituted. Desensitization protocols involve stepwise administration of gradually increasing doses of the antibiotic, starting at a sub-allergenic dose and progressing to the full dose. Successful antibiotic desensitization has been described via the oral, intravenous, intramuscular and inhaled routes. Penicillin desensitization has the longest and largest published body of evidence, but similar regimens have been successfully applied to virtually every other class of antibiotics. Antibiotic desensitization is a relatively safe procedure, and allergic reactions occur in a minority of patients; the vast majority of individuals are able to complete the procedure. Patients undergoing desensitization need to be continually monitored, and medical staff in attendance should be able to rapidly treat potential anaphylaxis. Graded challenge differs from desensitization because the immune response to an antibiotic is not modified. Patients who are unlikely to be allergic to an antibiotic are candidates for graded challenge, whereas desensitization is indicated for those who probably have IgE-mediated allergy. Copyright © 2007 S. Karger AG, Basel

Adverse reactions to antibiotics are commonly encountered in clinical practice. Penicillin allergy alone is reported by about 10% of the population [1]. Many patients with a history of antibiotic allergies, following complete evaluation, are found to not be allergic and are able to tolerate the drug in question. In the case of penicillin, approximately 90% of patients with a history of ‘penicillin allergy’ carry this label unnecessarily and can safely be treated with penicillins. Patients who are confirmed to have an IgE-mediated antibiotic allergy are generally advised to avoid this and other cross-reacting antibiotics. For patients with IgE-mediated antibiotic allergies who require treatment with the antibiotic to which they are allergic, the only safe method of administration is rapid desensitization. This chapter will focus on various aspects of rapid desensitization with penicillin and other antibiotics, including patient selection, possible mechanisms, side effects, safety issues, and procedure schemes. The chapter will conclude with a discussion of graded challenge, its indications and how it differs from desensitization. For a discussion of trimethoprim-sulfamethoxazole desensitization protocols in HIV-positive patients, the reader is referred to another chapter of this textbook. Similarly, rapid desensitization with

oncologic drugs and aspirin desensitization are covered in other chapters of this textbook [see chapters by Castells, pp 413–425, and by Sczceklik et al., pp 340–349].

Patient Selection

Rapid desensitization is an induction of tolerance procedure during which patients with type I antibiotic allergy are converted to a state which allows them to be safely treated with the antibiotic. Rapid antibiotic desensitization is intended only for patients who are confirmed, or strongly suspected, to have antibiotic-specific IgE antibodies. Rapid desensitization does not prevent development of non-IgE-mediated reactions. In fact, there are reports of interstitial nephritis, hemolytic anemia and serum sickness-like reactions in patients who were successfully desensitized [2– 5]. Antibiotic desensitization should be considered only in instances when an alternative noncross-reacting antibiotic cannot be substituted, or for patients who have failed treatment with an alternative antibiotic. In such cases, the risk of not properly treating the underlying infection should outweigh the risk of desensitization. For penicillin, the most common clinical scenario in which an absolute need for penicillin arises is treatment of syphilis during pregnancy, since alternative antibiotics such as erythromycin and clindamycin have inferior cure rates and may not cross the placenta in sufficient quantities. Cystic fibrosis (CF) patients commonly require antibiotic desensitization, because they have a high incidence of antibiotic allergies and, due to resistant organisms, have few effective antibiotic choices to treat respiratory infections [6]. Penicillins When available, penicillin skin testing should guide the clinical approach to patients with a history of penicillin allergy. Patients with a positive skin test response to any penicillin antigenic de-

Desensitization with Antibiotics

terminant are presumed to have a type I allergy. They should avoid penicillin class antibiotics, but if an absolute need for penicillin arises, they may receive the antibiotic via rapid desensitization. A negative penicillin skin test result (when appropriate major and minor antigenic determinants are available) has high negative predictive value, and therefore these patients may safely be treated with penicillins [1]. The group of M. Blanca from Spain postulates, however, a far lower sensitivity of skin tests (only
70%), partly due to selective reactivity to amoxicillin side chains [see chapter by Torres et al., pp 190–203]. At the time of this writing, both major and minor penicillin skin testing determinants are commercially unavailable in the USA, while the company Diater ([email protected]) made them available in Europe again. As a result, evaluation of possible penicillin allergy is far more challenging and driven by the patient history. Without the availability of penicillin skin testing, penicillin should be administered via rapid desensitization to those patients who, based on their history, are likely to have a type I allergy. Administration of penicillin via graded challenge (instead of desensitization) may be an option for some patients, depending on the reaction history. Since penicillin-specific IgE antibodies are known to wane over time, patients with distant reaction histories are less likely to still be allergic than those with recent reactions [7]. While patients with vague reaction histories are less likely to be allergic than those with convincing reaction histories, patients with vague reaction histories make up a one third of all penicillin skin test-positive subjects [8]. Consequently, caution needs to be exercised in the approach to any patient presenting with a history of penicillin allergy. Patients who experienced clearly non-IgE-mediated reactions are not candidates for rapid desensitization, and those with histories of severe non-IgE-mediated reactions (such as toxic epidermal necrolysis, interstitial nephritis, etc.) should not receive penicillin under any circumstances.

405

Table 1. Non-irritating intradermal skin test concentrations of selected intravenous antibiotics [from 9] Antibiotic

Full-strength concentration mg/ml

Non-irritating concentration (dilution from full-strength)

Cefotaxime Cefuroxime Cefazolin Ceftazidime Ceftriaxone Tobramycin Ticarcillin Clindamycin Gentamycin Trimethoprim-sulfa Levofloxacin Erythromycin Nafcillin Vancomycin Azithromycin

100 100 330 100 100 40 200 150 40 80 (sulfa component) 25 50 250 50 100

10-fold 10 10 10 10 10 10 10 10 100-fold 1,000-fold 1,000 10,000-fold 10,000 10,000

Non-Penicillin Antibiotics Selection of patients for desensitization with nonpenicillin antibiotics is mostly based on patient history. Fully validated skin testing for non-penicillin
-lactams and other antibiotics is not available or not well documented [but see also chapter by Campi et al., pp 216–232], and hence the reaction history is an essential component of whether to pursue desensitization, graded challenge or avoid the drug altogether. Patients with a convincing history of an immediate-type reaction (such as anaphylaxis) should undergo rapid desensitization if re-administration of the antibiotic is indicated. Unlike with penicillins, the natural history of type I allergy with other antibiotics is unknown. Because drug-specific IgE antibodies may not diminish over time to the same extent as they do with penicillin, patients with distant reaction histories should be approached more cautiously. Patients with vague reaction histories may be candidates to receive the antibiotic in question via graded challenge, rather than desensitization. As with penicillin, patients who

406

report severe non-IgE-mediated reactions should not receive the antibiotic again. Skin testing with a non-irritating concentration of native antibiotics can be useful and should be used during an evaluation of patients with history of reactions to non-penicillin antibiotics. A positive response suggests the presence of drugspecific IgE antibodies and would be an indication for rapid desensitization. A negative skin test response using a non-irritating concentration of these antibiotics has an uncertain negative predictive value and does not rule out possible immediate-type allergy. Hence, depending on the reaction history, patients who test negative may nevertheless need to receive the antibiotic via rapid desensitization. The amount of antibiotic tolerated during skin testing can serve as the starting point for desensitization. Many case reports or case series have described non-irritating concentrations for certain antibiotics, but only one study systematically determined non-irritating concentrations for many commonly-used antibiotics [9]. Results of this study are summarized in table 1.

Solensky

Desensitization Procedure

Penicillin Penicillin desensitization was first reported in 1946 [10] and in the ensuing three decades there were a number of additional published case reports [11–15]. Most of these procedures involved parenteral administration of the antibiotic and were associated with significant allergic reactions either during or immediately after the procedure. Sullivan and colleagues [3, 4, 16] were the first to describe large series of patients with skin test-proven IgE-mediated allergy who underwent penicillin desensitization. This procedure utilized oral penicillin, with the usual starting dose being about 1/10,000 of the full dose (table 2). Doubling doses were administered every 15 min until the full dose was reached. At the completion of desensitization, patients received penicillin via the desired route. For example, women treated for gestational syphilis received intramuscular benzathine penicillin G following desensitization with oral penicillin V [16]. In order to maintain the desensitized state, patients must continue treatment with penicillin. Unless a long-acting preparation (such as benzathine penicillin G) is used, this typically translates to oral penicillin taken 2–3 times daily [4, 10, 13, 17–19]. Discontinuation of penicillin treatment following successful desensitization causes a reversal of the state of tolerance and again places the patient at risk of penicillin-induced anaphylaxis. Hence, re-administration of penicillin in such a scenario necessitates that rapid desensitization be repeated. Penicillin desensitization can successfully be accomplished in analogous fashion via the intravenous route [5, 20, 21]. The dosing may be bolus every 15 min or via continuous infusion, with an increase in the rate every 15 min (tables 3, 4). This author prefers the oral route due to its ease of administration and probable increased safety (which is discussed in more detail later in this chapter). Additionally, when costs of both routes

Desensitization with Antibiotics

Table 2. Penicillin oral desensitization protocol [from 33] Step1

Penicillin mg/ml

Amount ml

Dose mg

Cumulative dose, mg

1 2 3 4 5 6 7 8 9 10 11 12 13 14

0.5 0.5 0.5 0.5 0.5 0.5 0.5 5 5 5 50 50 50 50

0.1 0.2 0.4 0.8 1.6 3.2 6.4 1.2 2.4 5 1 2 4 8

0.05 0.1 0.2 0.4 0.8 1.6 3.2 6 12 25 50 100 200 400

0.05 0.15 0.35 0.75 1.55 3.15 6.35 12.35 24.35 49.35 100 200 400 800

Observe patient for 30 min, then give full therapeutic dose by the desired route. 1 Interval between doses is 15 min.

Table 3. Penicillin intravenous desensitization protocol with drug added by piggyback infusion [from 33] Step1

Penicillin mg/ml

Amount ml

Dose mg

Cumulative dose, mg

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

0.1 0.1 0.1 0.1 0.1 1 1 1 10 10 10 10 100 100 100 1,000 1,000

0.1 0.2 0.4 0.8 1.6 0.32 0.64 1.2 0.24 0.48 1 2 0.4 0.8 1.6 0.32 0.64

0.01 0.02 0.04 0.08 0.16 0.32 0.64 1.2 2.4 4.8 10 20 40 80 160 320 640

0.01 0.03 0.07 0.15 0.31 0.63 1.27 2.47 4.87 10 20 40 80 160 320 640 1,280

Observe patient for 30 min, then give full therapeutic dose by the desired route. 1 Interval between doses is 15 min.

407

Table 4. Penicillin intravenous desensitization protocol using a continuous infusion pump Step1

Penicillin mg/ml

Flow rate, Dose ml/h mg

Cumulative dose, mg

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

0.01 0.01 0.01 0.1 0.1 0.1 0.1 0.1 0.1 10 10 10 10 10 10 10

6 12 24 5 10 20 40 80 160 3 6 12 25 50 100 200

0.015 0.045 0.105 0.23 0.48 1 2 4 8 15 30 60 123 250 500 1,000

0.015 0.03 0.06 0.125 0.25 0.5 1 2 4 7.5 15 30 62.5 125 250 500

Observe patient for 30 min, then give full therapeutic dose by the desired route. 1 Interval between doses is 15 min.

Table 5. Ciprofloxaxin oral desensitization protocol [from 28] Step1

Cipro mg/ml

Amount ml

Dose mg

Cumulative dose, mg

1 2 3 4 5 6 7 8 9 10 11 12 13

0.1 0.1 0.1 0.1 0.1 1 1 1 1 10 10 10 N/A

0.5 1 2 4 8 1.6 3.2 6.4 12.8 2.5 5 10 1 tablet

0.05 0.1 0.2 0.4 0.8 1.6 3.2 6.4 12.8 25 50 100 250

0.05 0.15 0.35 0.75 1.55 3.15 6.35 12.75 25.55 50.55 100.55 200.55 450.55

1

Interval between doses is 15 min.

408

of penicillin desensitization were compared at a single medical center, the oral regimen was found to be approximately half as expensive as the intravenous regimen [20]. Non-Penicillin Antibiotics The principles learned from penicillin desensitization have been applied to other classes of antibiotics to which patients have IgE-mediated allergies. In the last two decades, successful desensitizations have been described in patients with type I allergies to virtually all classes of antibiotics – including cephalosporins, monobactams, carbapenems, aminoglycosides, sulfonamides, quinolones, macrolides, and vancomycin [22–31]. These reports include both oral and intravenous regimens. There is a single case report of successful rapid desensitization to inhaled tobramycin [31]. Among intravenous desensitization protocols, some investigators administered the antibiotic via bolus dosing whereas others utilized continuous infusion pumps with stepwise increases in the rate. While the reported desensitization protocols are not uniform, they generally parallel penicillin desensitization with progressively increasing doses of the antibiotic being administered in frequent time intervals until the full dose is reached. If a given antibiotic is available in both oral and intravenous formulations, the oral desensitization route is preferred. For example, Lantner [28] administered doubling doses of oral ciprofloxacin every 15 min starting at 0.05 mg (1/10,000 of the full dose) until the full 500-mg dose was reached (table 5). As with penicillin, if the patient is to remain in a desensitized state, treatment with the antibiotic needs to be continued.

Safety Issues

Rapid desensitization should be performed only by physicians familiar with the procedure and with treatment of anaphylaxis. Medication and

Solensky

equipment to treat potential allergic reactions should be in close vicinity. Patients should have intravenous access, and they should be continuously observed for the appearance of IgE-mediated signs or symptoms. Vital signs and peak inspiratory flow values should be monitored regularly. A physician should remain present throughout the entire procedure. Since rapid desensitization has proven to be relatively safe, it is reasonable to perform the procedure in an outpatient setting, assuming the patient is stable and generally healthy. If more caution is required, desensitization may be performed in a general ward or in an intensive care unit. No life-threatening allergic reactions have been reported to occur during antibiotic desensitizations. Ideally, prior to rapid desensitization, patients should be clinically stable. Patients with asthma or other pulmonary disorders should be optimally controlled. Treatment with
-adrenergicblocking medications must be discontinued prior to desensitization, but caution and involvement of the cardiologist is advised if
-blockers are used for arrhythmia and not for elevated blood pressure. Pretreatment with systemic corticosteroids and antihistamines is generally contraindicated, although some patients may need to remain on these drugs for underlying conditions, such as chronic urticaria or autoimmune diseases. Pretreatment with antiallergy medications is potentially problematic because they may mask early signs of an allergic reaction. On the other hand, if a mild allergic reaction occurs during desensitization, it should be treated accordingly. Assuming the reaction subsides, the last dose should be repeated and then the dose may be advanced further. Most patients who undergo antibiotic rapid desensitization tolerate the procedure without evidence of allergic reactions. In Sullivan and colleagues’ [3, 4, 16] three major publications on oral penicillin desensitization, approximately one third of patients developed mild systemic allergic reactions. Most of the reactions were mild

Desensitization with Antibiotics

and easily treated, and the vast majority appeared following the completion of desensitization, rather than during the procedure [3, 4, 16]. Only 1 subject (out of 70) with CF could not complete desensitization due to worsening lower respiratory tract symptoms [4]. Other investigators have likewise reported very high success rates for completion of penicillin desensitization [5, 20, 21]. Only 2 desensitizations out of combined 58 had to be terminated due to allergic reactions [5, 20, 21]. In these series of patients, only 1/7 experienced allergic reactions related to desensitization. Because most published literature on rapid desensitization with non-penicillin antibiotics consists of single case reports, it is difficult to draw conclusions regarding success rates or frequency of reactions. One limitation is that reporting of failed desensitizations may not be pursued, and therefore publishing bias may lead to overrepresentation of successful desensitization procedures. Another confounding factor is that CF patients make up a majority of published nonpenicillin antibiotic desensitization case series. They are not representative of all desensitization subjects because they appear to have higher reaction rates and are unable to complete the procedure more often than non-CF patients [6]. With those limitations in mind, Turvey et al. [30] recently described outcomes of 57 consecutive antibiotic desensitizations to 12 different antibiotics, in 21 patients (19 of whom had CF), at a single medical center over a 5-year period. All but one of the regimens were via the intravenous route and the other one was oral. Eleven desensitizations (19%) had to be terminated due to allergic reactions, but 7 out of the 11 were felt to be due to non-IgE-mediated mechanisms. Interestingly, a single patient accounted for 6 of the 11 unsuccessful desensitizations. While there are no randomized trials comparing the safety of intravenous and oral routes of antibiotic desensitization, the oral route would appear to be safer. Sullivan [2] hypothesized that

409

oral desensitization is safer because it is less likely to expose patients to multivalent antibiotic conjugates and polymers, which are likely to play an important role in IgE-mediated allergic reactions. Epidemiological evidence indicates that oral penicillin is less commonly associated with allergic reactions than when it is administered parenterally [32]. Additionally, there are less than 10 reported anaphylactic deaths secondary to oral penicillin, compared to thousands following parenteral administration [2]. Therefore, for situations in which a given antibiotic is available in both oral and parenteral forms, it is preferable to desensitize via the oral route. One exception to this may be CF patients or other patients who have poor gastrointestinal absorption.

Mechanism

Rapid desensitization essentially ‘fools’ the immune system into not reacting against an antibiotic to which specific IgE antibodies are present. The mechanism by which desensitization renders mast cells unresponsive to the relevant allergic determinants remains elusive. One possible explanation is that cross-linking of drug-specific IgE on the surface of mast cells occurs in gradual fashion, thereby keeping the intracellular signal (which would normally cause degranulation) below a clinical threshold. Another theory is that univalent drug-carrier protein molecules prevent cross-linking of surface IgE and hence the transmission of an intracellular signal. While the exact mechanism of rapid desensitization is unclear, the process is antigen-specific. Following penicillin desensitization, a number of investigators have demonstrated that skin test responses to penicillin determinants diminish or disappear altogether in a majority of patients [2, 4, 12–14, 17, 18]. Similarly, a loss of skin test reactivity to a non-irritating concentration of vancomycin has been reported by two groups of researchers following successful vancomycin de-

410

sensitization [23, 24]. Furthermore, antibiotic desensitization does not affect the ability of mast cells to respond to other IgE stimuli. Sullivan [2] showed that skin test responses to aeroallergens, histamine and compound 48/80 (a chemical inducer of mast cell degranulation) were unchanged after penicillin desensitization. These findings also indicate that desensitization does not result in depletion of mast cell mediators or tachyphylaxis to the effects of these mediators.

Graded Challenge

Graded challenge, also called incremental test dosing, differs from rapid desensitization because the immunologic response to the drug is not modified by the procedure. Graded challenge is a method of cautiously administering a medication to patients who have a low likelihood of being allergic to it. In contrast, patients who undergo desensitization are assumed to have a type I allergy. The intent of graded challenge is to introduce the medication cautiously and hence not cause a severe allergic reaction. By administering the drug in stepwise fashion, if an allergic reaction occurs, it is hoped that it is minor and easily treated. Patients who tolerate an antibiotic given via graded challenge have proven themselves to not be allergic to that drug. Patients who experience allergic reactions during graded challenge should subsequently receive the antibiotic only via rapid desensitization. Most graded challenges may be carried out in an outpatient setting without intravenous access, but with full preparedness to treat potential anaphylactic reactions. The starting dose is generally about 1/100 of the full dose, which is higher than for desensitization, and the incremental dose increases are also relatively larger – usually 5- to 10-fold until the full dose is reached. Doses are usually administered at 30–60 min intervals, and prior to each dose, the patient should be questioned and examined for symptoms and signs of

Solensky

Table 6. Graded challenge with oral amoxicillin in a patient with distant reaction history to penicillin

Table 7. Graded challenge with telithromycin in a patient with history of mild reaction to azithromycin

Step1

Amoxicillin mg/ml

Amount ml

Dose mg

Step1

Telithromycin mg/ml

Amount ml

Dose mg

1 2 3

5 50 50

0.5 0.5 5

2.5 25 250

1 2 3 4

40 40 tablet tablet

0.25 1.25 1/2 2

10 50 200 800

1 Interval between doses is 30 min. Observe patient for 60 min after last dose. If an allergic reaction occurs, amoxicillin should be administered by rapid desensitization instead.

an allergic reaction. The pace of the challenge and degree of caution exercised depend on the likelihood that the patient may be allergic, the physician’s experience and comfort level with graded challenges, and the clinical stability of the patient. The number of steps in a graded challenge ranges from two steps to several. Too many dose escalations in a graded challenge may possibly convert the procedure to desensitization instead. This is problematic because if the patient was unknowingly desensitized, simple full dose re-administration of the antibiotic sometime in the future would result in an allergic reaction. A decision of whether to pursue rapid desensitization or graded challenge with an antibiotic is based on the patient’s evaluation, which may include history, review of medical records, physical examination, and possibly skin testing or in vitro testing. Only if a patient is deemed to have a low likelihood of IgE-mediated allergy should the antibiotic be administered via graded challenge. Otherwise, the antibiotic should be avoided or given via rapid desensitization. One scenario in which graded challenge may be appropriate is patients with distant type I allergic reactions to penicillin (table 6). This is because penicillinspecific IgE antibodies are known to wane over time and there might be a commercial unavailability of penicillin skin test reagents (which

Desensitization with Antibiotics

1

Interval between doses is 30 min. Observe patient for 60 min after last dose. If an allergic reaction occurs, telithromycin should be administered by rapid desensitization instead.

could otherwise help rule out IgE-mediated allergy). Graded challenge also may be useful when patients who have experienced a reaction to a given antibiotic require treatment with another one, and there is a potential of allergic cross-reactivity among them. For example, allergic cross-reactivity between macrolides and telithromycin is uncertain. If a patient with history of a mild type I reaction to azithromycin requires treatment with telithromycin, it would be reasonable to administer telithromycin via graded challenge (table 7).

411

References 1 Solensky R: Hypersensitivity reactions to
-lactam antibiotics. Clin Rev Allergy Immunol 2003;24:201–219. 2 Sullivan TJ: Antigen-specific desensitization of patients allergic to penicillin. J Allergy Clin Immunol 1982;69:500– 508. 3 Sullivan TJ, Yecies LD, Shatz GS, Parker CW, Wedner HJ: Desensitization of patients allergic to penicillin using orally administered
-lactam antibiotics. J Allergy Clin Immunol 1982;69: 275–282. 4 Stark BJ, Earl HS, Gross GN, Lumry WR, Goodman EL, Sullivan TJ: Acute and chronic desensitization of penicillin-allergic patients using oral penicillin. J Allergy Clin Immunol 1987;79: 523–532. 5 Borish L, Tamir R, Rosenwasser LJ: Intravenous desensitization to
-lactam antibiotics. J Allergy Clin Immunol 1987;80:314–319. 6 Moss RB: Drug allergy in cystic fibrosis. Clin Rev Allergy 1991;9:211–229. 7 Blanca M, Torres MJ, Garcia JJ, Romano A, Mayorga C, de Ramon E, Vega JM, Miranda A, Juarez C: Natural evolution of skin test sensitivity in patients allergic to
-lactam antibiotics. J Allergy Clin Immunol 1999;103:918–924. 8 Solensky R, Earl HS, Gruchalla RS: Penicillin allergy: prevalence of vague history in skin test-positive patients. Ann Allergy Asthma Immunol 2000; 85:195–199. 9 Empedrad R, Darter AL, Earl HS, Gruchalla RS: Nonirritating intradermal skin test concentrations for commonly prescribed antibiotics. J Allergy Clin Immunol 2003;112:629–630. 10 O’Donovan WJ, Klorfajn I: Sensitivity to penicillin: anaphylaxis and desensitization. Lancet 1946;ii:444–446. 11 Green GR, Peters GA, Gerach JE: Treatment of bacterial endocarditis in patients with penicillin hypersensitivity. Ann Intern Med 1967;67:235–249.

12 Gillman SA, Korotzer JL, Haddad ZH: Penicillin desensitization: correlation of clinical and immunological events using an in vitro model. Clin Allergy 1972;2:63–68. 13 Graybill JR, Sande MA, Reinarz JA, Shapiro SR: Controlled penicillin anaphylaxis leading to desensitization. South Med J 1974;67:62–64. 14 Fellner MJ, Van Hecke E, Rozan M, Baer RL: Mechanisms of clinical desensitization in urticarial hypersensitivity to penicillin. J Allergy 1970;45:55–61. 15 Pedersen-Bjergaard J: Specific hyposensitization of patients with penicillin allergy. Acta Allergol 1969;24:333–361. 16 Wendel GD, Stark BJ, Jamison RB, Molina RD, Sullivan TJ: Penicillin allergy and desensitization in serious infections during pregnancy. N Engl J Med 1985;312:1229–1232. 17 Naclerio RM, Mizrahi EA, Adkinson NJ: Immunologic observations during desensitization and maintenance of clinical tolerance to penicillin. J Allergy Clin Immunol 1983;71:294–301. 18 Brown LA, Goldberg ND, Shearer WT: Long-term ticarcillin desensitization by the continuous oral administration of penicillin. J Allergy Clin Immunol 1982;69:51–54. 19 Ganier M, Lieberman P: Infantile agammaglobulinemia and immediate hypersensitivity to penicillin G. JAMA 1977;237:1852–1853. 20 Chisholm CA, Katz VL, McDonald TL, Bowes WA: Penicillin desensitization in the treatment of syphilis during pregnancy. Am J Perinatol 1997; 14: 553–554. 21 Moss RB, Babin S, Hsu YP, BlessingMoore J, Lewiston NJ: Allergy to semisynthetic penicillins in cystic fibrosis. J Pediatr 1984;104:460–466. 22 Win PH, Brown H, Zankar A, Ballas ZK, Hussain I: Rapid intravenous cephalosporin desensitization. J Allergy Clin Immunol 2005;116:225–228.

23 Lin RY: Desensitization in the management of vancomycin hypersensitivity. Arch Intern Med 1990;150:2197–2198. 24 Anne’ S, Middleton E, Reisman RE: Vancomycin anaphylaxis and successful desensitization. Ann Allergy 1994; 73:402–404. 25 Wazny LD, Daghigh B: Desensitization protocols for vancomycin hypersensitivity. Ann Pharmacother 2001;35: 1458–1464. 26 Earl HS, Sullivan TJ: Acute desensitization of a patient with cystic fibrosis allergic to both
-lactam and aminoglycoside antibiotics. J Allergy Clin Immunol 1987;79:477–483. 27 Gorman SK, Zed PJ, Dhingra VK, Ronco JJ: Rapid imipenem/cilastatin desensitization for multidrug-resistant Acinetobacter pneumonia. Ann Pharmacother 2003;37:513–516. 28 Lantner RR: Ciprofloxacin desensitization in a patient with cystic fibrosis. J Allergy Clin Immunol 1995;96:1001– 1002. 29 Gea-Banacloche JC, Metcalfe DD: Ciprofloxacin desensitization. J Allergy Clin Immunol 1996;97:1426–1427. 30 Turvey SE, Cronin B, Arnold AD, Dioun AF: Antibiotic desensitization for the allergic patient: 5 years of experience and practice. Ann Allergy Asthma Immunol 2004;92:426–432. 31 Spigarelli MG, Hurwitz ME, Nasr SZ: Hypersensitivity to inhaled TOBI following reaction to gentamicin. Pediatr Pulmonol 2002;33:311–314. 32 Herman R, Jick H: Cutaneous reaction rates to penicillins: oral versus parental. Cutis 1979;24:232–234. 33 Sullivan TJ: Drug allergy; in Middleton E, Reed CE, Ellis EF, Adkinson NF, Yunginger JW (eds): Allergy: Principles and Practice, ed 4. St Louis, Mosby, 1993, pp 1726-1746.

Dr. Roland Solensky The Corvallis Clinic, 3680 NW Samaritan Drive Corvallis, OR 97330 (USA) Tel. +1 541 754 1260, Fax +1 541 758 2680 E-Mail [email protected]

412

Solensky

Pichler WJ (ed): Drug Hypersensitivity. Basel, Karger, 2007, pp 413–425

Drug Desensitization in Oncology: Chemotherapy Agents and Monoclonal Antibodies Mariana Castells Harvard Medical School, Boston, Mass., USA

Abstract

Introduction

The need to offer first-line therapy for primary and recurrent cancers and for treating chronic inflammatory diseases has spurred the clinical development of rapid desensitizations for chemotherapy and monoclonal antibodies. Rapid desensitization allows patients to be treated with medications to which they have presented hypersensitivity reactions, including anaphylaxis. Rapid desensitization induces temporary clinical tolerization and is achieved by administering small incremental doses of the drug inducing the hypersensitivity reaction up to the full therapeutic dose within a few hours. Protocols are available for most chemotherapy agents, including taxenes, platins, doxorubicin, monoclonal antibodies and others. Candidate patients include those who present type I hypersensitivity reactions, mast cell/IgEdependent, including anaphylaxis, during the chemotherapy infusion or shortly after. Anaphylactoid reactions are amenable to treatment with the same rapid desensitization protocols as for type I hypersensitivity reactions. Idiosyncratic reactions and T-cell-mediated reactions are not amenable to rapid desensitization. The indication for rapid desensitization can only be made by allergy and immunology specialists and can only be performed in settings with oneto-one nurse-patient care and where resuscitation personnel and resources are readily available. Repeated desensitizations can be safely performed in outpatient settings with similar conditions, which allows cancer patients to remain in clinical studies and permits the treatment of cancer and chronic rheumatologic and gastrointestinal conditions. Copyright © 2007 S. Karger AG, Basel

All chemotherapy agents can cause hypersensitivity reactions [1], which has limited the use of critical drugs in very sick patients for fear of inducing a more severe reaction and possibly death [2]. The choice of an alternative chemotherapy regimen is often limited by tumor sensitivity and, because of the increasing number of cancer survivors, exposure to multiple courses of the same or similar chemotherapy agents can lead to sensitization and to hypersensitivity reactions. One third of the patients exposed to seven or more cycles of carboplatin develop hypersensitivity reactions, including anaphylaxis, and deaths have been reported with re-exposure [3]. Hypersensitivity reactions to humanized monoclonal antibodies are rare but their frequency is increasing as more cancer and rheumatologic patients are exposed to multiple courses. Rapid desensitization is the induction of temporary unresponsiveness to drug antigens, thereby allowing patients to be treated with medications to which they have presented hypersensitivity reactions. Gradual re-introduction of small amounts of drug antigens up to full therapeutic doses has been successfully achieved using dif-

ferent protocols [4, 5]. The term rapid desensitization indicates that full therapeutic doses can be reached in a relatively short time, typically 4– 12 h, unlike bacterial or viral vaccination or classical allergen immunotherapy in which weeks to months are required to provide immune protection. Despite the clinical expansion and success of rapid desensitizations, the cellular and molecular inhibitory mechanisms inducing the temporary tolerization of patients are still incompletely understood.

Mechanism of Adverse Drug Reactions to Chemotherapy Agents and Monoclonal Antibodies Amenable to Desensitization

Anaphylactic Reactions: Type I Hypersensitivity Reactions, Mast Cell/IgE-Dependent, and Their Diagnosis Drug-induced type I hypersensitivity reactions result from the release of mediators from IgEsensitized mast cells and/or basophils and can affect all organ systems, leading to anaphylaxis and, if untreated, to death. Drug antigens can sensitize patients after multiple courses, and repeated exposures are needed for the development of specific IgE [6]. Among chemotherapy drugs, platins, such as carboplatin, cisplatin and oxaliplatin, can induce IgE formation by a mechanism described in metal workers exposed to low molecular weight platinum salts by inhalation and skin contact, acting as haptens. Sensitization to platinum salts can be demonstrated by skin testing and in vitro platin-specific IgE antibodies (radioallergosorbent test). Symptoms are induced by platinum salts cross-linking of specific IgE bound to high-affinity IgE receptors, Fc
RI, on mast cells and/or basophils, with the release of membrane and granule mediators. These include vasoactive amines such as histamine, proteolytic enzymes such as tryptase, and pro-inflammatory mediators such as prostaglandins and leukotrienes.

414

Cross-linking of IgE by drug antigens can lead to limited skin reactions (flushing, pruritus, urticaria, angioedema) or multi-organ systems involvement (sneezing, sinus and nasal congestion, cough, shortness of breath, wheezing, abdominal pain, nausea, vomiting, diarrhea) with hypotension and cardiovascular collapse during anaphylaxis. Reactions can occur within minutes of exposure, and minimal amounts of a drug can induce severe reactions in highly sensitized individuals, such as laryngeal edema with asphyxiation. Disseminated intravascular coagulation is a rare complication of anaphylaxis. Measurements of tryptase in serum [7] and histamine in urine can confirm the diagnosis. Epinephrine is the only treatment that can reverse anaphylaxis [8]. The diagnosis of type I hypersensitivity reactions to drugs relies on the demonstration of in vivo or in vitro drug-specific IgE. In a population of 126 patients who received over 6 courses of carboplatin for recurrent ovarian cancer and were skin tested before each course, only 10 patients with negative skin test presented a hypersensitivity reaction, indicating that the rate of false-negative skin test is as low as 1.5% [9]. In the same population, 7 out of 41 patients who volunteered to have carboplatin despite a positive skin test presented anaphylaxis [9]. Other studies indicate that in patients with hypersensitivity reactions to carboplatin, 80–90% have a positive skin test [9]. The likelihood of a severe hypersensitivity reaction is very high in skin test-positive patients, and re-challenging those patients is not indicated. Anaphylactoid Reactions Anaphylactoid reactions are hypersensitivity reactions induced by drug antigens upon initial exposure, without prior sensitization and with similar clinical presentation and symptoms as IgEmediated reactions. They can result from the release of mediators from mast cells and/or basophils, without known IgE mechanism and with negative skin test. Among chemotherapy drugs,

Castells

taxenes can induce severe hypersensitivity symptoms with cardiovascular collapse within a few minutes of first exposure in patients who present negative skin test. Mechanisms implicated in such reactions include the activation of complement by the diluent cremophor [40] and/or the direct release of mediators. Anaphylactoid reactions respond to epinephrine and antihistamines indicating that mast cells and/or basophil mediators are released among other mediators.

Principles of Rapid Desensitization and Cellular and Molecular Targets

The goal of rapid desensitization is to induce tolerization with few or no side effects while escalating to the therapeutic dose. An initial dose is required, followed by dose increments delivered at fixed time intervals or as a continuous infusion until reaching the target dose. The starting dose can range from 10- to 10,000-fold lower than the target dose, and dose increments are given every 10–30 min [5]. Doubling or tripling dose increments at each time interval has been more successful at reducing side effects than 10fold increments. All available clinical protocols are empirical and based on trial and error clinical experiences since the cellular and molecular targets of this inhibitory process are not completely understood. Mast cells and/or basophils are thought to be the major cellular players, since suboptimal doses of antigen administered prior to an optimal dose renders those cells unresponsive to antigen but not to other activating stimuli. Suboptimal doses can provide excessive monomeric antigen incapable of cross-linking surface Fc
RI receptors or it can induce rapid internalization of antigen cross-linked receptors depleting the cell surface. In vitro rapid desensitization of human mast cells induces the decreased levels of signal transducing molecules, such as syk, due to ubiquitinilation and degradation [10]. Naturally occurring syk-deficient ba-

Desensitization in Oncology

sophils are unresponsive to drug antigens, indicating that syk is critical for activation and for desensitization [11]. In recent studies, STAT6, a signal transducer and transcription activator responsible for the transcription of IL-4 and IL-13, has been involved in rapid desensitizations. STAT-6-deficient mast cells are capable of releasing mediators during the early phase of IgE cell activation but cannot release late cytokines, such as TNF-  and IL-6, and cannot be desensitized to antigen [12].

Protocols for Desensitization to Chemotherapy Agents and Monoclonal Antibodies

We, along with others [13–15], have generated protocols for desensitization to chemotherapy and monoclonal antibodies based on published protocols for desensitizations to antibiotics and in vitro data. We generated a standardized, 3-solution (A–C), 12-step protocol which allows for gradual increases in the infusion rate and drug concentration and allows for the infusion of the target dose over 5.8 h, as seen in table 1. The three solutions containing X/100, X/10, and X mg, respectively, were diluted in 250 ml of 5% dextrose in water, and were used in sequence of increasing concentrations. The concentration of solutions A, B, and C were (X/100)/250, (X/10)/250, and (X)/250 mg/ml, respectively. Solution A was used for steps 1–4, solution B for steps 5–8, and solution C for steps 9–12. The concentration of solution C for paclitaxel was adjusted to the maximum concentration allowed to avoid precipitation (see below). The rate of the infusion was changed every 15 min, with each step delivering approximately twice the dose of the previous step. The final step (step 12) maintained a constant rate of infusion to deliver the remainder of the total dose. Risk stratification of the patients was performed and included the assessment of associated cardiac and pulmonary conditions,

415

Table 1. Inpatient desensitization to carboplatin (350 mg)

Solution A Solution B Solution C

Full dose: 350.0 mg

Total mg to be injected in each bottle

250 ml of 0.014 mg/ml 250 ml of 0.140 mg/ml 250 ml of 1.389 mg/ml

3.500 35.000 347.246

Note to pharmacy: the total mg injected is more than the final dose because solutions A and B are not completely infused. Step

Solution

Rate ml/h

Time min

Administered dose, mg

Cumulative dose, mg

1 2 3 4

A A A A

2 5 10 20

15 15 15 15

0.0070 0.0175 0.0350 0.0700

0.0070 0.0245 0.0595 0.1295

5 6 7 8

B B B B

5 10 20 40

15 15 15 15

0.1750 0.3500 0.7000 1.4000

0.3045 0.6545 1.3545 2.7545

9 10 11 12

C C C C

10 20 40 75

15 15 15 186

3.4725 6.9449 13.8898 322.9383

6.2270 13.1719 27.0617 350.0000

Total time, min

usage of -blockers, and the severity of the reaction. For patients with increased cardiac risk and/or near death anaphylactic/anaphylactoid reactions a 16-step protocol was generated by the addition of a 4th bag which contained 250 ml of diluted drug (X/1,000) and was infused prior to bag 3, in 4 steps as for the other 3 bags [16]. Patients were hospitalized in the medical intensive care unit (MICU) for the initial desensitization, as required by hospital policy for desensitizations. One-to-one nurse-patient care was provided for each desensitization, and nurses were trained by the allergy team in how to administer the protocol and how to recognize the early symptoms of hypersensitivity reactions and anaphylaxis.

416

351

Once a patient completed a successful desensitization in the MICU, all subsequent courses of chemotherapy were given in the outpatient facility with a desensitization-trained chemotherapy nurse in one-to-one attendance. Patients did not receive specific premedication for the desensitization procedure; rather, they were premedicated with the standard oncology regimen. The outpatient protocol was similar to the inpatient protocol except for the volume of the solutions. Instead of 250 ml, solutions A, B, and C each contained 100 ml of 5% dextrose in water. Because of the increased concentration of the solutions, the time required to administer the total dose was reduced from 5.8 h for inpatients to 3.8 h for outpatients. For outpatient desensitiza-

Castells

tions of paclitaxel, the volume of the solutions ranged between 100 and 250 ml to provide a concentration of ^1.3 mg/ml, given the potential for drug precipitation at higher concentrations. Rapid Desensitization for Hypersensitivity Reactions to Taxenes Paclitaxel is a widely used antineoplastic agent with activity against ovarian, breast, and other solid tumors. It was initially isolated from the bark of the Pacific yew tree (Taxus brevifolia) in the 1970s, and its antimitotic activity is due to the bundling of microtubules, which arrests cell division. Docetaxel is a semisynthetic taxene originally extracted from the needles of the European yew tree (Taxus baccata), whose antimitotic activity is similar to that of paclitaxel [13]. A high incidence of hypersensitivity reactions was observed with paclitaxel in early clinical trials involving flushing, hemodynamic changes, dyspnea, musculoskeletal pain, paresthesias, and gastrointestinal symptoms, with fatalities reported. Symptoms frequently occurred during the first course of therapy, within seconds to minutes of beginning the infusion, indicating a lack of need for prior sensitization. Slower infusion rates and premedication with H1, H2 antihistamine receptor antagonists and corticosteroids have decreased the incidence of hypersensitivity reactions to less than 10%. Despite such interventions, there is a subset of patients with taxene-responsive cancers who present hypersensitivity reactions, in whom rapid desensitization is indicated [13, 14, 17]. Attempts at using docetaxel in patients with paclitaxel hypersensitivity reactions have not proven universally successful. Hypersensitivity reactions to taxenes resemble anaphylactic reactions induced by the acute release of mast cell/basophil mediators. Skin tests have been negative, indicating that the diluent cremophor, the generation of reactive metabolites capable of activating complement,

Desensitization in Oncology

or mast cells/basophils directly may be responsible. Multiple protocols have been described for the desensitization of patients with hypersensitivity reactions to taxenes [13, 14, 18]. The standard desensitization protocol developed at the Brigham and Women’s Hospital in Boston, Massachusetts (described in table 1 for carboplatin), was used to perform 176 desensitizations in 40 patients who presented a hypersensitivity reaction after the first or second infusion of paclitaxel or docetaxel [16, 19, 20]. The presenting symptoms of hypersensitivity reactions to taxenes are described in table 2 and include flushing, pruritus, urticaria, chest pain, hypotension and hypertension, loss of consciousness, gastrointestinal symptoms, musculoskeletal pain (chest or lower back), and dyspnea with O2 desaturation. The initial hypersensitivity reaction in those patients was immediate (!10 s) or occurred within a maximum of 15 min. One patient who switched to docetaxel developed a similar hypersensitivity reaction, indicating that cremophor, the vehicle for paclitaxel, was not responsible for the hypersensitivity reaction. All 40 patients were successfully desensitized initially in the MICU and had repeated desensitizations, completing their chemotherapy cycles. Breakthrough reactions occurred in 12% of the desensitizations, and the reactions were less severe than the initial reaction and did not preclude the completion of any treatment course. We noted that the incidence of allergic disease in these patients (seasonal allergic rhinitis, asthma, allergy to other drugs, venom sensitivity) was 57%, which far exceeds the 15–20% reported for the general population [16]. An earlier review of 19 patients with paclitaxel-induced hypersensitivity reactions found a statistically significant difference in the rate of hymenoptera venom sensitivity [21]. Patients with allergic conditions may be at higher risk for hypersensitivity reactions to taxenes.

417

Table 2. Initial anaphylactic and anaphylactoid symptoms in patients reacting to platins, taxenes, and liposomal doxorubicin (adapted from Lee et al. [20]) Percent of patients

Flushing Pruritus Urticaria Hypo-/hypertension Loss of consciousness O2 desaturation Nausea/vomiting Abdominal pain Dyspnea Chest pain Musculoskeletal pain Throat tightness

anaphylactic

anaphylactoid

platins (n = 31)

taxenes (n = 22)

doxorubicin (n = 6)

54.8 77.4 29 12.9 6.5 16.1 19.4 16.1 38.7 25.8 3.2 9.7

86.4 4.5 4.5 27.3 18.2 31.8 4.5 27.3 45.5 68.2 45.5 18.2

16.7 33.3 0 16.7 0 0 16.7 0 0 50 16.7 0

Rapid Desensitization for Hypersensitivity Reactions to Platins Carboplatin Carboplatin is an effective and well-tolerated cytotoxic agent used as standard front-line chemotherapy for ovarian cancer. Many patients achieve a complete clinical remission with the platinumbased regimen but later develop recurrent disease within 3 years of diagnosis. For patients with platinum-sensitive recurrent cancer, disease relapsing after at least a 6-month disease-free interval, platinum-based chemotherapy remains the most active regimen. Therefore, the ability to administer carboplatin safely as front-line therapy and in the relapse setting provides a significant clinical benefit to the patient. Patients treated with multiple courses of carboplatin experience increased incidence of hypersensitivity reactions, which are uncommon during the initial courses. The incidence of reactions increases to 27% in patients receiving more than seven cycles of carboplatin [9, 22]. Symptoms of hypersensitivity reactions vary from cutaneous reactions such as

418

flushing and urticaria to life-threatening respiratory and cardiovascular compromise, including bronchospasm, chest pain, and hypotension, with more than 50% of patients developing at least moderately severe symptoms [9]. Hypersensitivity reactions to carboplatin often prompt its permanent discontinuation. They are thought to be mast cell/IgE-mediated since skin test performed on the volar surface of the forearm with a drop of a non-irritating concentration of carboplatin at 1–10 mg/ml is positive in over 80% of reactive patients [23, 24]. Multiple protocols have been used for the desensitization of patients with carboplatin hypersensitivity reactions [15, 25, 26]. We performed 162 desensitizations in 54 patients with severe hypersensitivity reactions to platins with the standardized desensitization protocol described in table 1, the same as for taxene desensitizations with 3 solutions and 12-steps for 5.8 h [20, 23]. Positive skin test was demonstrated in 80.8% of the patients (table 3). Most of the patients (46.2%) reacted at the intradermal 1 mg/ml carboplatin

Castells

Table 3. Skin testing in 26 patients presenting symptoms of hypersensitivity while infused with carboplatin (adapted from Lee et al. [20])

Total Prick (10 mg/ml) Intradermal (1 mg/ml) Intradermal (10 mg/ml)

Positive

Negative

21 (80.8%) 1 (3.8%) 12 (46.2%) 8 (30.8%)

5 (19.2%)

concentration, and a few at prick testing with 10 mg/ml. In addition, a few patients reacted at intradermal 10 mg/ml carboplatin concentration. Patients received a median of 8 courses before developing their initial hypersensitivity reaction. Most patients had reactions during therapy for recurrent cancer, not in their initial cancer treatment. Typically, after recurrence of their disease, patients were re-exposed to carboplatin during the seventh cycle and developed reactions during the eighth cycle, which suggests that a prolonged period of sensitization is required before the onset of hypersensitivity reactions. The reaction profile consistent with type I hypersensitivity reactions [20, 23] includes flushing, pruritus, urticaria, nausea, dyspnea, tachycardia, hypertension or hypotension, and chest pain. Most patients had their initial hypersensitivity reaction during the infusion with 6 patients experiencing symptoms within 15 min of infusion. No delayed reactions were observed. Cutaneous manifestations were present in 96% of the patients, and extracutaneous symptoms were present in 77% of the patients, including hypotension and loss of consciousness. In contrast to patients presenting reactions to paclitaxel, there was a low incidence of musculoskeletal pain, including back pain. All patients successfully received all planned courses of carboplatin through the desensitization program. Mild breakthrough reactions occurred in 12% of the desensitizations, none of which resulted in cardiovascular collapse or death. Positive skin test to carboplatin became negative after the

Desensitization in Oncology

desensitization, demonstrating the inhibition of cutaneous mast cell reactivity, consistent with other studies [27]. Cisplatin and Oxaliplatin Reactions to other platins such as cisplatin and oxaliplatin in patients with hypersensitivity reactions to carboplatin have been described, and although cross-reactivity is thought to be high, it is not universal [28]. Substituting cisplatin in carboplatin-sensitive patients was done in 5 out of 6 patients with no side effects, but no skin test was available to provide evidence of the cross-sensitivity. Reactions to cisplatin and oxaliplatin can also be predicted by skin test, such as seen with carboplatin, and in 2 patients with hypersensitivity reactions to cisplatin who presented positive skin test, premedication with steroids and antihistamines did not prevent the anaphylactic symptoms and desensitization was performed [29]. Anaphylactic reactions to oxaliplatin have been described, and in most cases the hypersensitivity reaction has precluded the use of the drug since premedication is ineffective. Successful desensitizations have been reported in 2 patients with colorectal cancer who presented severe anaphylactic reactions and positive skin tests [30, 31] and another patient who received a slow 24-hour infusion after developing a hypersensitivity during the sixth cycle. Further studies of IgE crossreactivity among platins are needed but the role of rapid desensitizations is clearly established. Desensitizations in Children Platins The use of carboplatin has been limited in children with unresectable low-grade glioma due to a high rate of hypersensitivity reactions to carboplatin, and few studies have validated the use of alternative drugs. A series of 9 patients with hypersensitivity reactions to carboplatin were treated with vinblastine weekly with acceptable toxicity/efficacy ratio, but no guidelines have been provided for children’s desensitization to carbo-

419

platin. Children with neurofibromatosis type 1 develop optic tumors which are treated with carboplatin and up to 30% develop hypersensitivity reactions after multiple exposures. Recently, rapid desensitizations have been done successfully, with limited side effects and with remarkable good tumor responses [32]. Children with astrocytomas develop similar rates of hypersensitivity reactions after exposure to several cycles of carboplatin and desensitizations have been successful in 6 out of 26 children with hypersensitivity reactions, and in 2 children who received multiple desensitization cycles and had stable remission of their astrocyomas. A girl with optic glioma developed a hypersensitivity reaction after 9 courses of carboplatin and was successfully desensitized [25]. The use of rapid desensitizations in the children’s cancer population is rapidly emerging as the only way to administer a curing medication. L-Asparaginase L-Asparaginase derived from Escherichia coli has

been used effectively for the treatment of children with acute lymphoblastic leukemia (ALL), but after repeated courses, up to 43% of patients can develop hypersensitivity reactions, including anaphylaxis [1]. Skin tests have been positive [33] and tryptase levels have been elevated in 2 patients with severe anaphylactic events when treated for central nervous system relapsing disease [34], indicating an IgE and mast cell dependence. Switching to different asparaginase preparations has been done, but some patients develop reactions to all preparations. Desensitizations to L-asparaginase by using an intravenous infusion starting at 100-fold dilution of the concentration giving a positive skin test was successfully done in a 2-year-old girl with ALL who presented an anaphylactic event during her second chemotherapy course [33]. Other desensitization protocols have been used recently for patients with ALL who develop hypersensitivity reactions to asparaginase [35], with no major side effects providing safety evidence.

420

Rapid Desensitization for Hypersensitivity Reactions to Doxorubicin/Adriamycin Liposomal doxorubicin (formerly adriamycin) was developed to bypass a surface-membrane permeability glycoprotein overexpressed in ovarian cancer patients with multidrug resistance and to decrease cumulative cardiotoxicity [36, 37]. Acute infusions reactions have been described with liposomal doxorubicin at the first or second infusion, similar to the reactions observed with paclitaxel, in up to 5–10% of patients. The reactions include flushing, shortness of breath, swelling, headache, chills, chest pain, tightness in chest or throat, pruritus, rash, tachycardia and, in rare cases, cyanosis, apnea and syncope (table 2). Premedication and reducing the infusion rate [38] has decreased the reactions to less than 1%, but discontinuation of therapy has been required in some patients with severe reactions. We have desensitized 6 patients with severe reactions to liposomal doxorubicin for 26 courses with the protocol described in table 4. The patients presented symptoms at first exposure, indicating the lack of sensitization and the possibility that the reaction was a direct effect of the liposomal preparation. The symptoms included flushing, shortness of breath, and severe flushing with presyncopal episodes. All 6 patients tolerated the procedure without major side effects and were desensitized once in the MICU and then received all remaining cycles in the outpatient setting. No skin test was performed due to the local toxicity of liposomal doxorubicin. Rapid Desensitization for Hypersensitivity Reactions to Monoclonal Antibodies Anti-Her2 (Herceptin/Trastuzumab), Anti-CD20 (Rituxan/Rituximab), and Anti-TNF- (Humira/Adalimumab, Remicade/Infliximab, Enbrel/Etanercept) Hypersensitivity reactions to humanized monoclonal antibodies are rare but their frequency is increasing as more cancer and rheumatologic pa-

Castells

Table 4. Inpatient desensitization to liposomal doxorubicin (102 mg)

Solution A Solution B Solution C

Full dose: 102.0 mg

Total mg to be injected in each bottle

250 ml of 0.004 mg/ml 250 ml of 0.041 mg/ml 250 ml of 0.405 mg/ml

1.020 10.200 101.197

Note to pharmacy: the total mg injected is more than the final dose because solutions A and B are not completely infused. Step

Solution

Rate ml/h

Time min

Administered dose, mg

Cumulative dose, mg

1 2 3 4

A A A A

2 5 10 20

15 15 15 15

0.0020 0.0051 0.0102 0.0204

0.0020 0.0071 0.0173 0.0377

5 6 7 8

B B B B

5 10 20 40

15 15 15 15

0.0510 0.1020 0.2040 0.4080

0.0887 0.1907 0.3947 0.8027

9 10 11 12

C C C C

10 20 40 75

15 15 15 186

1.0120 2.0239 4.0479 94.1135

1.8147 3.8387 7.8865 102.0000

Total time, min

tients are exposed to multiple courses. Patients with hypersensitivity reactions to rituximab (Rituxan), to trastuzumab (Herceptin) [39], and to anti-TNF- monoclonal antibodies [40, 41] have been described. In patients with reactions compatible with a type I hypersensitivity reaction, in whom a mast cell/IgE mechanism can be demonstrated, desensitizations have been reported for humanized monoclonal antibodies [42]. We used the protocol described in table 1 to desensitize patients to rituximab and trastazumab [20 ; and manuscript in preparation]. The initial reaction was severe in both cases and included anaphylaxis. An IgE mechanism was demonstrated by positive skin test to non-irritant concentrations of the drugs. Breakthrough symptoms

Desensitization in Oncology

351

were less severe than the initial reaction and allowed the patients to complete their courses. Patients received their first desensitization in the intensive care setting and for repeated desensitizations were transitioned to the outpatient clinic with one-to-one nurse support. One patient desensitized to trastazumab presented side effects during the first two initial desensitizations and was later transitioned to the outpatient setting, where she underwent 8 more courses without side effects, with a modified protocol. Outcomes of Desensitization and Cancer Progression Until now, patients have been declined further therapy once they have presented a hypersensitiv-

421

Table 5. Symptoms during desensitizations to platins, taxenes, trastuzumab and liposomal doxorubicin (adapted from Lee et al. [20]) Number of patients

Flushing Pruritus Urticaria Hypo-/hypertension Loss of consciousness O2 desaturation Nausea/vomiting Abdominal pain Dyspnea Chest pain Musculoskeletal pain Throat tightness

ity reaction to a chemotherapy agent, and the impact of using an alternative drug has not been well established. Through rapid desensitizations, patients can receive the drugs to which they presented a hypersensitivity reaction, allowing them to continue using first-line therapy for their cancer or chronic inflammatory condition. Whether chemotherapy desensitizations are effective at tumor killing or control needs to be defined. In a small population of 26 adult patients receiving carboplatin desensitization for recurrent cancer, 10 (38.5%) had a radiographic response (partial or complete response) and/or 150% drop of initial CA125 value, 11 (42.3%) had stable disease radiographically and/or CA125 response (!50% drop), and 5 (19.2%) had progressive disease after 1–2 cycles of carboplatin. Of 16 adult patients receiving paclitaxel desensitization for newly diagnosed cancer, 16 patients (100%) achieved clinical remission. Of the 3 adult patients receiving paclitaxel desensitization for recurrent cancer, 1 had clinical response to therapy, 1 stable disease, and 1 progressive disease. Such rates are expected for cancer patient populations not receiving chemotherapy desensitizations [20]. In the children’s

422

platins (n = 11)

taxenes (n = 6)

trastuzumab (n = 1)

doxorubicin (n = 6)

4 7 3 1/2 0 1 0 1 3 1 0 1

3 1 1 1/1 0 0 0 1 0 2 0 0

1 1 1 0 0 0 0 0 0 0 0 1

0 2 0 0 0 0 0 0 0 1 0 0

population, 6 of 26 patients with astrocytoma treated with desensitization had similar rates of remissions as for non-desensitized patients [32, 43].

Rapid Desensitization Procedures: Indications, Location and Safety

The indications for rapid desensitizations can only be established by allergy and immunology specialists who can understand the nature and mechanism of the initial hypersensitivity reaction, can perform skin testing to provide evidence of IgE/mast cell sensitization, and can assign a risk for the desensitization procedure. In the series of Feldweg and Lee [16], all patients received their first desensitization in the intensive care setting and once they had completed a desensitization with no or minimal side effects, they were transitioned safely to the outpatient clinics with one-toone nurse support. One patient desensitized to trastazumab presented side effects during the first 2 initial desensitizations and was later transitioned to the outpatient setting, where she under-

Castells

Table 6. Indications and contraindications for rapid desensitization to chemotherapy drugs and monoclonal antibodies Indications

Drug most frequently used

Disease

Anaphylactic IgE-mediated reactions

Chemotherapy drugs

Primary and recurrent metastatic cancers (breast, ovarian, colon, others)

Flushing, pruritus, urticaria, angioedema

Platins: carboplatin, oxaliplatin, cisplatin

Laryngeal edema, rhinorrhea, conjunctivitis, shortness of breath, wheezing, bronchospasm Nausea, vomiting, diarrhea Hypotension

Monoclonal antibodies: rituximab, trastuzumab

Chronic inflammatory diseases, leukemias, breast, ovarian cancers

Anaphylactoid Direct mast cell/basophil Complement, others

Chemotherapy drugs

Primary and recurrent metastatic cancers (breast, ovarian, colon)

Flushing, pruritus, urticaria, angioedema

Taxenes: paclitaxel, docetaxel

Throat tightness, shortness of breath, O2 desaturation

Doxorubicin

Nausea, vomiting, diarrhea Hypotension, hypertension, syncope Back, chest and/or abdominal pain Contraindications Delayed reactions (>24 h) Erythema multiforme Stevens-Johnson syndrome Toxic epidermal necrolysis Foot-and-mouth syndromes Fever

went 8 more courses without side effects, with a modified protocol. Symptoms occurring during desensitization to platins, taxenes, trastazumab, and liposomal doxorubicin in 24 patients are presented in table 5. 92% (22 of 24) of patients completed the desensitization with mild reactions, cutaneous in majority, which only required the use of antihistamines. 8.3% (2 of 24) of patients had a severe reaction requiring epinephrine. All patients completed their chemotherapy courses. Similar safety data have been provided by other groups and no deaths have been reported in the last 10 years since the implementation of rapid de-

Desensitization in Oncology

sensitizations for cancer therapy and for therapy of chronic inflammatory conditions [13, 32].

Conclusions

Rapid desensitization protocols are available for the treatment of patients with cancer and rheumatologic diseases who present hypersensitivity reactions to their chemotherapy. Candidate patients include those who present type I hypersensitivity, mast cell/IgE-dependent reactions and those who present anaphylactoid reactions during

423

the chemotherapy infusion or shortly after (table 6). Symptoms include pruritus, flushing, urticaria, angioedema, respiratory and gastrointestinal distress, changes in blood pressure including hypotension and shock. Associated musculoskeletal symptoms and pain can be present in patients reacting to taxenes and liposomal doxorubicin. Rapid desensitizations are not indicated in patients with late reactions (124 h), idiosyncratic reactions (fever), erythema multiforme, StevensJohnson syndrome or toxic epidermal necrolysis. During rapid desensitization, drug antigens are re-introduced in a slow and incremental fashion allowing for full therapeutic doses to be delivered with minor or no side effects. Temporary tolerization is achieved within hours. Initial rapid desensitization should only be done in settings with one-to-one nurse-patient care and where resuscitation personnel and resources are readily available. After the first desensitization, standard protocols are available for safe repeated desensitizations in outpatient settings with similar conditions, which provides flexibility and allows patients to remain in clinical studies, or receive chronic treat-

ments. Breakthrough symptoms are less severe than the initial hypersensitivity reactions in all series reviewed, and deaths have not been reported. Managing breakthrough symptoms with antihistamines and steroids and decelerating the dose escalation with intermediate infusion steps successfully improves the tolerability of desensitization protocols. Initial outcome data suggest that desensitized adults and children have equal responses to therapy as non-desensitized patients. Failed desensitizations in which patients cannot achieve the target therapeutic dose are rare and require modified protocols. Education of nurses, pharmacists, oncology, rheumatology and allergy specialists will lead to the universal use of rapid desensitization protocols for all cancer patients and patients with chronic inflammatory conditions with hypersensitivity reactions to first-line therapy agents. Basic research is needed to uncover the cellular and molecular mechanisms underlying the temporary tolerization induced by rapid desensitization, so that pharmacological interventions can improve its safety and efficacy.

References 1 Weiss RB, Bruno S: Hypersensitivity reactions to cancer chemotherapeutic agents. Ann Intern Med 1981;94:66–72. 2 Vervloet D, Durham S: Adverse reactions to drugs. BMJ 1998;316:1511–1514. 3 Zweizig S, Roman LD, Muderspach LI: Death from anaphylaxis to cisplatin: a case report. Gynecol Oncol 1994;53: 121–122. 4 Wendel G, Stark B, Jamison RB, Molina RD, Sullivan TJ: Penicillin allergy and desensitization in serious infections during pregnancy. N Engl J Med 1985; 312:1229–1232. 5 Sullivan TJ: Antigen-specific desensitization of patients allergic to penicillin. J Allergy Clin Immunol 1982;69:500– 508. 6 Solensky R: Drug hypersensitivity. Med Clin North Am 2006;90:233–260. 7 Schwartz LB, Metcalfe DD, Miller JS, Earl H, Sullivan T: Tryptase levels as an

424

8

9

10

11

indicator of mast-cell activation in systemic anaphylaxis and mastocytosis. N Engl J Med 1987;316:1622–1626. Sampson HA, Munoz-Furlong A, Bock SA, Schmitt C, Bass R, Chowdhury BA et al: Symposium on the definition and management of anaphylaxis: summary report. J Allergy Clin Immunol 2005; 115:584–591. Markman M, Kennedy A, Webster K, Elson P, Peterson G, Kulp B et al: Clinical features of hypersensitivity reactions to carboplatin. J Clin Oncol 1999; 17:1141. Macglashan D, Miura K: Loss of syk kinase during IgE-mediated stimulation of human basophils. J Allergy Clin Immunol 2004;114:1317–1324. Kepley CL, Youssef L, Andrews RP, Wilson BS, Oliver JM: Syk deficiency in nonreleaser basophils. J Allergy Clin Immunol 1999;104:279–284.

12 Morales AR, Shah N, Castells M: Antigen-IgE desensitization in signal transducer and activator of transcription 6deficient mast cells by suboptimal doses of antigen. Ann Allergy Asthma Immunol 2005;94:575–580. 13 Markman M, Kennedy A, Webster K, Kulp B, Peterson G, Belinson J: Paclitaxel-associated hypersensitivity reactions: experience of the gynecologic oncology program of the Cleveland Clinic Cancer Center. J Clin Oncol 2000;18:102–105. 14 Markman M, Kennedy A, Webster K, Peterson G, Kulp B, Belinson J: An effective and more convenient drug regimen for prophylaxis against paclitaxelassociated hypersensitivity reactions. J Cancer Res Clin Oncol 1999;125:427– 429.

Castells

15 Markman M, Hsieh F, Zanotti K, Webster K, Peterson G, Kulp B et al: Initial experience with a novel desensitization strategy for carboplatin-associated hypersensitivity reactions: carboplatinhypersensitivity reactions. J Cancer Res Clin Oncol 2004;130:25–28. 16 Feldweg AM, Lee CW, Matulonis UA, Castells M: Rapid desensitization for hypersensitivity reactions to paclitaxel and docetaxel: a new standard protocol used in 77 successful treatments. Gynecol Oncol 2005;96:824–829. 17 Peereboom DM, Donehower RC, Eisenhauer EA, McGuire WP, Onetto N, Hubbard JL et al: Successful re-treatment with taxol after major hypersensitivity reactions. J Clin Oncol 1993;11: 885–890. 18 Fishman A, Gold T, Goldberg A, Confino-Cohen R, Beyth Y, Menczer J et al: Effective desensitization protocol to paclitaxel following hypersensitivity reaction. Int J Gynecol Cancer 1999;9: 156–159. 19 Price KS, Castells MC: Taxol reactions. Allergy Asthma Proc 2002;23:205–208. 20 Lee CW, Matulonis UA, Castells MC: Rapid inpatient/outpatient desensitization for chemotherapy hypersensitivity: standard protocol effective in 57 patients for 255 courses. Gynecol Oncol 2005;99:393–399. 21 Grosen E, Siitari E, Larrison E, Tiggelaar C, Roecker E: Paclitaxel hypersensitivity reactions related to bee-sting allergy. Lancet 2000;355:288–289. 22 Kook H, Kim KM, Choi SH, Choi BS, Kim HJ, Chung SY et al: Life-threatening carboplatin hypersensitivity during conditioning for autologous PBSC transplantation: successful rechallenge after desensitization. Bone Marrow Transplant 1998;21:727–729. 23 Lee CW, Matulonis UA, Castells MC: Carboplatin hypersensitivity: a 6-hour 12-step protocol effective in 35 desensitizations in patients with gynecological malignancies and mast cell/IgE-mediated reactions. Gynecol Oncol 2004;95: 370–376.

Dr. Mariana Castells Harvard Medical School, One Jimmy Fund Way Smith Building, Room 626D Boston, MA 02115 (USA) Tel. +1 617 525 1265, Fax +1 617 525 1310 E-Mail [email protected]

Desensitization in Oncology

24 Zanotti KM, Rybicki LA, Kennedy AW, Belinson JL, Webster KD, Kulp B et al: Carboplatin skin testing: a skin-testing protocol for predicting hypersensitivity to carboplatin chemotherapy. J Clin Oncol 2001;19:3126–3129. 25 Sims-McCallum RP: Outpatient carboplatin desensitization in a pediatric patient with bilateral optic glioma. Ann Pharmacother 2000;34:477–478. 26 Morgan M, Bowers DC, Gruchalla RS, Khan DA: Safety and efficacy of repeated monthly carboplatin desensitization. J Allergy Clin Immunol 2004;114: 974–975. 27 Markman M, Zanotti K, Peterson G, Kulp B, Webster K, Belinson J: Expanded experience with an intradermal skin test to predict for the presence or absence of carboplatin hypersensitivity. J Clin Oncol 2003;21:4611–4614. 28 Ottaiano A, Tambaro R, Greggi S, Prato R, Di Maio M, Esposito G et al: Safety of cisplatin after severe hypersensitivity reactions to carboplatin in patients with recurrent ovarian carcinoma. Anticancer Res 2003;23:3465–3468. 29 Goldberg A, Altaras MM, Mekori YA, Beyth Y, Confino-Cohen R: Anaphylaxis to cisplatin: diagnosis and value of pretreatment in prevention of recurrent allergic reactions. Ann Allergy 1994;73:271–272. 30 Gammon D, Bhargava P, McCormick MJ: Hypersensitivity reactions to oxaliplatin and the application of a desensitization protocol. Oncologist 2004;9: 546–549. 31 Mis L, Fernando NH, Hurwitz HI, Morse MA: Successful desensitization to oxaliplatin. Ann Pharmacother 2005;39:966–969. 32 Ogle SK, Rose MM, Wildes CT: Development and implementation of a carboplatin desensitization protocol for children with neurofibromatosis, type 1 and hypersensitivity reactions in an outpatient oncology clinic. J Pediatr Oncol Nurs 2002;19:122–126.

33 Bonno M, Kawasaki H, Hori H, Umemoto M, Komada Y, Sakurai M: Rapid desensitization for L-asparaginase hypersensitivity. J Allergy Clin Immunol 1998;101:571–572. 34 Stone HD Jr, DiPiro C, Davis PC, Meyer CF, Wray BB: Hypersensitivity reactions to Escherichia coli-derived polyethylene glycolated-asparaginase associated with subsequent immediate skin test reactivity to E. coli-derived granulocyte colony-stimulating factor. J Allergy Clin Immunol 1998;101:429–431. 35 Guo Y: Desensitization therapy of acute lymphocytic leukemia with injection of L-asparaginase (report of 5 cases) (in Chinese). Zhonghua Er Ke Za Zhi 2005; 43:309–310. 36 Booser DJ, Hortobagyi GN: Anthracycline antibiotics in cancer therapy. Focus on drug resistance. Drugs 1994;47: 223–258. 37 Rigatos SK, Tsavdaridis D, Athanasiadis A, Stathopoulos JG, Stathopoulos GP: Paclitaxel and liposomal doxorubicin (Caelyx) combination in advanced breast cancer patients: a phase II study. Oncol Rep 2003;10:1817–1819. 38 Harrison M, Tomlinson D, Stewart S: Liposomal-entrapped doxorubicin: an active agent in AIDS-related Kaposi’s sarcoma. J Clin Oncol 1995;13:914–920. 39 Tada K, Ito Y, Hatake K, Okudaira T, Watanabe J, Arakawa M et al: Severe infusion reaction induced by trastuzumab: a case report. Breast Cancer 2003;10:167–169. 40 Nikas SN, Voulgari PV, Drosos AA: Urticaria and angiedema-like skin reactions in a patient treated with adalimumab. Clin Rheumatol 2006;1–2. 41 Chavez-Lopez MA, Delgado-Villafana J, Gallaga A, Huerta-Yanez G: Severe anaphylactic reaction during the second infusion of infliximab in a patient with psoriatic arthritis. Allergol Immunopathol (Madr) 2005; 33:291–292. 42 Stallmach A, Giese T, Schmidt C, Meuer SC, Zeuzem SS: Severe anaphylactic reaction to infliximab: successful treatment with adalimumab – report of a case. Eur J Gastroenterol Hepatol 2004;16:627–630. 43 Broome CB, Schiff RI, Friedman HS: Successful desensitization to carboplatin in patients with systemic hypersensitivity reactions. Med Pediatr Oncol 1996;26:105–110.

425

Author Index

Allanore, L. 267 Almeida, C.-A. 95 Aster, R.H. 306 Azukizawa, H. 129

Hadj-Rabia, S. 321 Hung, S.-I. 105

Barbaud, A. 366 Beeler, A. 380 Bertoli, R. 278 Bircher, A.J. 352 Blanca, M. 190

Kano, Y. 251 Kawabata, T.T. 140 Keller, M. 295

Campi, P. 151, 216 Castells, M. 413 Cavani, A. 74 Cerny, A. 278 Chen, Y.-T. 105 Christiansen, C. 233 Chung, W.-H. 105

Rebelo Gomes, E. 2 Romano, A. 2 Roujeau, J.-C. 267

Itami, S. 129

Mallal, S. 95 Manfredi, M. 216 Martin, S.F. 34 Mayorga, C. 190 Mockenhaupt, M. 18 Naisbitt, D.J. 55 Niżankowska-Mogilnicka, E. 340 Nolan, D. 95

Sanak, M. 340 Sanderson, J.P. 55 Sanz, M.L. 391 Scheinmann, P. 321 Severino, M. 216 Shenton, J.M. 115 Shiohara, T. 251 Solensky, R. 404 Spanou, Z. 295 Szczeklik, A. 340 Takahashi, R. 251 Torres, M.J. 190 Uetrecht, J.P. 115

Demoly, P. 2 De Pità, O. 74 Dewachter, P. 204 De Weck, A.L. 391 Gamboa, P.M. 391 Gerber, B.O. 66

Park, B.K. 55 Phillips, E. 95 Piccotti, J.R. 140 Pichler, W.J. 66, 151, 168, 295, 380 Pirmohamed, M. 84 Ponvert, C. 321 Popovic, M. 115

Viola, M. 2 Weltzien, H.U. 47 Yawalkar, N. 242

Subject Index

Abacavir genetic susceptibility studies in Han Chinese association study candidate genes 108 HLA allele frequency 109 single-nucleotide polymorphisms 108, 109 drug specificity 110 ethnic specificity 110, 111 fine mapping 109 genome-wide study 109, 110 human immunodeficiency virus and hypersensitivity reactions clinical features 97 formulations 95 genetic susceptibility 88, 89, 97–102 immunological markers 101 overview 88, 89 prospects for study 102, 103 screening 102 mechanism of action 95, 96 metabolism 96, 97 severe reactions 183 structure 95, 96 Abatacept, adverse reactions 158 Abciximab, drug-induced thrombocytopenia 313 Acetaminophen, patch testing 370 Acute generalized exanthematous pustulosis (AGEP) clinical features 182, 184 definition 19 diagnosis 360, 361 epidemiology data sources 20, 21 etiologic factors 27, 28 genetic epidemiology 28 incidence 22 risk estimation methodology 22, 23 pediatric drug hypersensitivity 327

Acyclovir, patch testing 368 Adalimumab, adverse reactions in therapy 158 Adverse drug reactions (ADRs) biological adverse reactions versus drug reactions 152, 153 classification 2, 105 Allergic contact dermatitis (ACD) allergen testing 41–43 haptens 131, 132 immune response polarization by contact allergens 38, 39 innate immunity 34–38 mouse model, see Contact hypersensitivity model nickel allergy, see Nickel hypersensitivity regulatory T cell role 34 treatment options 43 Allopurinol genetic susceptibility studies in Han Chinese association study candidate genes 108 HLA allele frequency 109 single-nucleotide polymorphisms 108, 109 drug specificity 110 ethnic specificity 110, 111 fine mapping 109 genome-wide study 109, 110 severe cutaneous adverse reactions 24–26, 106, 183 Allylisopropylacetylurea, patch testing 368 Amidotrizoate, structure 235 Aminopyrine hypersensitivity reactions 224 trade names 222 Amoxicillin drug-induced liver injury 289 patch testing 368 pediatric drug hypersensitivity 323 severe cutaneous adverse reactions 27 structure 192

427

Amphotericin B, cutaneous specific immunoglobulin E 219 Ampicillin severe cutaneous adverse reactions 27 structure 192 Amprenavir, human immunodeficiency virus and hypersensitivity reactions 90 Anakinra, adverse reactions 158 Anaphylaxis clinical features 359, 360 definition 204, 205 pediatric 327 perioperative, see Perioperative anaphylaxis Anesthetics, perioperative anaphylaxis induction 207 Angiotensin-converting enzyme (ACE) inhibitors host factors in drug hypersensitivity reactions 12 severe cutaneous adverse reactions 26 Animal models, see Non-clinical testing; see also specific diseases Antibodies adverse drug reactions to therapy anaphylactic reactions 414 anaphylactoid reactions 414, 415 desensitization indications and contraindications 422, 423 outcomes 421, 422 principles 415 technique 420, 421 biological agents 154 protein recognition 57 pseudoallergic drug reactions 174 therapeutics, see specific agents Antigen, definition 55 Antipyrine hypersensitivity reactions 224 trade names 222 Antiseptics, perioperative anaphylaxis induction 207 Apazone, trade names 222 Aplastic anemia, see Drug-induced aplastic anemia Aprotinin, perioperative anaphylaxis induction 207 Arginine vasopressin (AVP), perioperative anaphylaxis management 214 L-Asparaginase, desensitization 420 Aspirin hypersensitivity asthma clinical features 341 cross-reactivity 342 definition 340, 341 diagnosis 342, 343 EP2 receptor 345 mediators cellular sources 344 leukotrienes 343, 344 lipoxins 344, 345 prostaglandins 343, 345

428

prevalence 341 prevention 345 treatment 345, 346 overview 217, 340 urticaria clinical presentation 346 diagnosis 347 epidemiology 346 histopathology 347 prevention and treatment 347 Association study, drug hypersensitivity susceptibility studies in Han Chinese candidate genes 108 HLA allele frequency 109, 111, 112 single-nucleotide polymorphisms 108, 109 Asthma, see Aspirin hypersensitivity Atopy definition 168 predisposition 168, 169 Azithromycin, pediatric drug hypersensitivity 323 Aztreonam cross-reactivity 201, 202 mechanism of action 191 structure 196 Bacitracin, cutaneous specific immunoglobulin E 219 Basiliximab, adverse reactions in therapy 158, 160 Basophil activation test (BAT) advantages 400 betalactam hypersensitivity diagnosis 199 betalactam testing negative skin test patients 397, 398 positive skin test patients 396, 397 historical perspective 391 metamizol testing 398, 399 neuromuscular blocking agent testing 396 non-steroidal anti-inflammatory drug testing 399 principles 392, 393 technical aspects allergen concentration 394, 395 blood storage effects 394 comparison with other techniques 395, 396 cutoff points 395 incubation time 394 interleukin-3 preactivation 396 negative control 393 positive control 393, 394 reproducibility 395 whole blood versus isolated cells 393 Basophil mediator release, betalactam hypersensitivity diagnosis 199 Beclometasone, pediatric drug hypersensitivity 323 Betalactams, see Cephalosporins; Penicillins Biological agent adverse reactions agent classification 154

Subject Index

classification agent-related side effects 155 target-related side effects 155 type
155, 156 type  156–159 type  163 type  163 type  159–163 versus drug reactions 152, 153 Blood dyscrasias, see Drug-induced aplastic anemia; Drug-induced immune hemolytic anemia; Drug-induced neutropenia; Drug-induced thrombocytopenia Bumadizon, trade names 222 Calcipotriol, pediatric drug hypersensitivity 323 Captopril, patch testing 368 Carbamazepine genetic susceptibility studies in Han Chinese association study candidate genes 108 HLA allele frequency 109, 111, 112 single-nucleotide polymorphisms 108, 109 drug specificity 110 ethnic specificity 110, 111 fine mapping 109 genome-wide study 109, 110 phenotype specificity 112 screening 112, 113 patch testing 368 severe cutaneous adverse reactions 24, 25 severe reactions 183 Carbenicillin, structure 192 Carboplatin, desensitization 418, 419 Cefaclor, pediatric drug hypersensitivity 323 Cefadroxil, structure 194 Cefalexin, structure 194 Cefamandole, structure 194 Cefatrizine, pediatric drug hypersensitivity 323 Cefazolin, structure 194 Cefcapene pivoxil, patch testing 368 Cefepime, structure 195 Cefixime, structure 195 Ceflacor, structure 194 Cefonicid, structure 194 Cefoperazone, structure 195 Ceforanide, structure 195 Cefotetan, structure 194 Cefoxime, structure 195 Cefoxitin, structure 194 Cefpodoxime pediatric drug hypersensitivity 323 structure 195 Ceftazidime, structure 195 Ceftibuten, pediatric drug hypersensitivity 323

Subject Index

Ceftriaxone, structure 195 Cefuroxime, structure 194 Cefuroxime axetil, pediatric drug hypersensitivity 323 Celecoxib patch testing 368 severe reactions 183 Cephalosporins, see also specific drugs clinical course of hypersensitivity 197 cross-reactivity 201, 202 diagnosis of hypersensitivity 197–199 pathophysiology of drug reactions 191, 193, 195–197 pediatric drug hypersensitivity 328, 329 perioperative anaphylaxis induction 206 pseudoallergic drug reactions 174 resensitization 200, 201 severe reactions 183 structures 191, 194, 195 treatment of hypersensitivity 202 Cephalothin, structure 194 Cetuximab, adverse reactions 158, 163 Challenge test, see Drug provocation test Chemical interactions, types 56 Chemotherapy agents adverse drug reactions anaphylactic reactions 414 anaphylactoid reactions 414, 415 desensitization indications and contraindications 422, 423 outcomes 421, 422 principles 415 technique children 419, 420 doxorubicin 420 overview 415–417 platins 418, 419 taxanes 417 Children, see Pediatric drug hypersensitivity Chlorhexidine basophil activation testing 400 perioperative anaphylaxis induction 207 Chlormezanone, trade names 222 Chloroquine, patch testing 368 Chlorphenamine, patch testing 368 Ciprofloxacin maculopapular exanthem 181, 182 non-irritating intradermal test concentration 221 Cisplatin, desensitization 419 Clarithromycin, pediatric drug hypersensitivity 323 Clavulanic acid drug-induced liver injury 289 pediatric drug hypersensitivity 323 structure 196 Clindamycin, patch testing 368 Clobazem, patch testing 368

429

Cloxacillin, structure 192 Codeine, patch testing 368 Colchicine, patch testing 368 Colloids basophil activation testing 399 perioperative anaphylaxis induction 207 Competition inhibition assays, betalactam hypersensitivity diagnosis 199, 200 Contact dermatitis, see Allergic contact dermatitis Contact hypersensitivity model (CHS) costimulatory molecules 39 immune regulation and tolerance induction 40, 41 immune response polarization by contact allergens 38, 39 innate immunity in pathogenesis 35–38 Langerhans’ cell role 39, 40 non-clinical testing 146 tolerance studies 75, 76 weak allergens and effector/regulatory cells 41 Contrast agents, see Radio contrast media Corticosteroids drug-induced thrombocytopenia management 311 drug reaction with eosinophilia and systemic symptom management 263, 264 patch testing 368 Co-trimoxazole cutaneous-specific immunoglobulin E 219 human immunodeficiency virus and hypersensitivity reactions 85–88 pediatric drug hypersensitivity 323 severe cutaneous adverse reactions 25 severe reactions 183 Coumarins, drug-induced liver injury 289, 290 Cyamemazine, patch testing 368 Cyclophosphamide, epidermal necrolysis disease management 276 Cyclosporine, basophil activation testing 400 Cyproheptadine, pediatric drug hypersensitivity 323 CYPs, drug metabolism 282, 283 Cytokine release syndrome, adverse reactions in therapy 155, 156 Cytomegalovirus (CMV), immune reconstitution syndrome 261–263 Daclizumab, adverse reactions in therapy 158 Dapsone, severe reactions 183 DC, see Dendritic cell Delavirdine, human immunodeficiency virus and hypersensitivity reactions 89 Dendritic cell (DC) allergic contact dermatitis role 37–39 maturation 81 penicillin allergy role 61 tolerance role 80

430

Desensitization chemotherapy agents indications and contraindications 422, 423 outcomes 421, 422 principles 415 technique children 419, 420 doxorubicin 420 overview 415–417 platins 418, 419 taxanes 417 graded challenge 410, 411 mechanism 410 monoclonal antibodies indications and contraindications 422, 423 outcomes 421, 422 principles 415 technique 420, 421 non-beta lactam antibiotics patient selection 406 technique 230, 231, 408 penicillins patient selection 405 technique 407, 408 safety 408–410 Desloratadine, patch testing 368 Diazepam, patch testing 368 Diclofenac, patch testing 368 Dicloxacillin, structure 192 Diltiazem patch testing 368 severe cutaneous adverse reactions 27, 183 Dipyrone hypersensitivity reactions 224 trade names 222 Docetaxel, desensitization 417 Domperidone, pediatric drug hypersensitivity 323 Doxorubicin, desensitization 420 Drug hypersensitivity reactions (DHRs) acute phase clinical approach 353, 354 laboratory analysis 355 morphological diagnosis 354 classification 2, 168–170, 172–177 clinical presentation and diagnosis acute generalized exanthematic pustulosis 360, 361 anaphylaxis 359, 360 drug fever 358, 359 drug-induced lupus erythematosus 363 drug reaction with eosinophilia and systemic symptoms 360 fixed drug eruptions 361, 362 flexural exanthem 361 general symptoms 358

Subject Index

maculopapular exanthem 360 organ systems 359 pseudolymphomatous reactions 360 serum sickness 360 severe skin reactions 362 urticaria and angioedema 359 vasculitis 362, 363 etiology evaluation algorithm 356–358 chronology 355, 356 specific diagnostic approach 355 host factors in drug hypersensitivity reactions 13 incidence and prevalence hospital-based population studies 306 mortality data 7–9 outpatient and general population studies 6, 7 phases 132, 133 risk factors drug-related factors 9, 10 host-related factors 10–13 treatment regimen 10 Drug-induced aplastic anemia clinical presentation 319 description 318 diagnosis 319 epidemiology 318 pathogenesis 318, 319 prognosis 319 treatment 319 Drug-induced hypersensitivity syndrome, see Drug reaction with eosinophilia and systemic symptoms Drug-induced immune hemolytic anemia (DIHA) autoantibodies 315, 316 clinical features 175 description 314 hapten-induced disease 314, 315 quinine type 315 Drug-induced interstitial nephritis (DIN) clinical presentation extrarenal symptoms 301, 302 onset 299, 301 renal symptoms 302 cofactors 297 definition 295 diagnosis 302, 303 drugs 296, 297 epidemiology 295, 297 histopathology 297 pathogenesis 298, 299 prognosis 304 treatment 303, 304 Drug-induced liver injury (DILI) cholestatic liver injury 279, 280 classification 279, 280

Subject Index

definition 279 diagnosis 290, 291 drug risks amoxicillin 289 clavulanic acid 289 coumarins 289, 290 halothane 285, 286 isoniazid 288 lamotrigine 290 nevirapine 115, 116 sulfamethoxazole 287, 288 tienilic acid 286, 287 troglitazone 290 ximelagatran 290 epidemiology 280 genetic susceptibility 282 immune system of liver 283, 284 metabolism of drugs 282, 283 overview 278 pathogenesis 284, 285 prospects for study 291 risk factors 281 Drug-induced lupus erythematosus features 363 Drug-induced neutropenia (DINP) clinical presentation 317 description 316 diagnosis 318 epidemiology 316 pathogenesis 316, 317 prognosis 318 treatment 318 Drug-induced thrombocytopenia (DITP) abciximab 313 autoimmune thrombocytopenia 311, 312 classification 306, 307 clinical features 175 glycoprotein IIb/IIIa antagonist induction 312, 313 hapten-dependent antibodies 307 heparin 313, 314 quinine type clinical presentation 310 diagnosis 310 epidemiology 307, 308 pathogenesis 308–310 treatment 310, 311 Drug provocation test betalactam hypersensitivity diagnosis 200 non-betalactam antibiotics 229, 230 perioperative anaphylaxis diagnosis 213 Drug reaction with eosinophilia and systemic symptoms (DRESS) clinical features 186, 252–254, 360 cytokine storm 183 definition 19, 20

431

Drug reaction (continued) diagnosis 256, 257 epidemiology data sources 20, 21 etiologic factors 28 genetic epidemiology 28 incidence 22, 252 risk estimation methodology 22, 23 herpesviruses in reactivation 257, 258 history of study 251, 252 immune reconstitution syndrome 261–263 laboratory findings 254, 255 pathological findings 255, 256 pediatric drug hypersensitivity 326, 327 prognosis 263 T cells antiviral T cells in pathogenesis 258 detection 259–261 treatment 263, 264 Dyes, perioperative anaphylaxis induction 208 Efalizumab, adverse reactions 158 Efavirenz, human immunodeficiency virus and hypersensitivity reactions 89 Enoxaparin, patch testing 368 Eosinophilia, drug allergy 187 Epinephrine, perioperative anaphylaxis management 213, 214 Erythema multiforme, pediatric drug hypersensitivity 325 Erythromycin cutaneous specific immunoglobulin E 219 pediatric drug hypersensitivity 323 serum immunoglobulin E 220 Erythropoietin, adverse reactions in therapy 157 Esomprazole, drug-induced interstitial nephritis 296, 297, 301, 302 Etanercept, adverse reactions 158, 162 Famciclovir, patch testing 368 Fas, toxic epidermal necrolysis role 137, 138 Feprazone, trade names 222 Ferrous gluconate, pediatric drug hypersensitivity 323 Flexural exanthem, features 361 Flow cytometry basophil activation test, see Basophil activation test T-cell-mediated drug hypersensitivity reaction testing 383, 384 Fluconazole, cutaneous specific immunoglobulin E 219 Fluindione, patch testing 368 Fosfomycin, cutaneous specific immunoglobulin E 219 Framycetin, cutaneous specific immunoglobulin E 219 Furantoin, drug allergy 187 Fusafungine, patch testing 369

432

Ganciclovir, patch testing 369 Gemifloxacin, maculopapular exanthem 181, 182 Gentamicin cutaneous-specific immunoglobulin E 219 patch testing 369 Glutathione severe drug reaction protection 185 sulfamethoxazole hypersensitivity role 87 Glycoprotein IIb/IIIa antagonists, drug-induced thrombocytopenia 312, 313 Graft-versus-host disease (GVHD) herpesviruses in reactivation 258, 259 toxic epidermal necrolysis association 129, 130 Halothane, drug-induced liver injury 285, 286 Hapten definition 55 drugs as haptens and prohaptens 66, 171 nickel 49, 52 Hemolytic anemia, see Drug-induced immune hemolytic anemia Heparin basophil activation testing 400 drug-induced thrombocytopenia 313, 314 patch testing 369 Herpesviruses drug reaction with eosinophilia and systemic symptoms, reactivation 257, 258 immune reconstitution syndrome 261, 262 Histamine, perioperative anaphylaxis pathophysiology 208 HLA alleles abacavir hypersensitivity 88, 89, 97–102 drug hypersensitivity susceptibility studies in Han Chinese 109, 111, 112 host factors in drug hypersensitivity reactions 12, 13 screening 102, 112, 113 severe cutaneous adverse reactions 28 Human immunodeficiency virus (HIV) abacavir hypersensitivity clinical features 97 formulations 95 genetic susceptibility 88, 89, 97–102 immunological markers 101 overview 88, 89 prospects for study 102, 103 screening 102 co-trimoxazole hypersensitivity reactions 85–88 drug-induced interstitial nephritis 297 drug-induced liver injury 281 epidemiology of drug hypersensitivity 84, 85 immune reconstitution syndrome 261–263 interferon- levels 88

Subject Index

non-nucleoside reverse transcriptase inhibitor hypersensitivity, see Nevirapine protease inhibitor hypersensitivity 90 Hydroxyzine, patch testing 369 Hypersensitivity syndrome, see Drug reaction with eosinophilia and systemic symptoms Imipenem, structure 196 Immune reconstitution syndrome (IRS), drug reaction with eosinophilia and systemic symptoms 261–263 Immunogen, definition 55 Immunoglobulin-E-mediated drug reactions, see also specific drugs clinical features 172 pathogenesis 170 Immunoglobulin-G-mediated drug reactions hemolytic anemia 175 thrombocytopenia 175 type II reactions 174, 175 type III reactions 175, 176 Infliximab, adverse reactions in therapy 157–161 Interferons adverse reactions in therapy 158, 159, 161, 163 interferon- levels in human immunodeficiency virus 88 T-cell-mediated drug hypersensitivity reaction testing 384, 385 Interleukin-3 (IL-3), basophil activation test preactivation 395 Interleukin-4 (IL-4), polymorphisms in drug hypersensitivity reactions 13 Interleukin-5 (IL-5) cytokine release syndrome 155 T-cell-mediated drug hypersensitivity reaction testing 384, 385 Interleukin-13 (IL-13), polymorphisms in drug hypersensitivity reactions 13 Interstitial lung disease, drug allergy 187 Interstitial nephritis, see Drug-induced interstitial nephritis Intravenous immunoglobulin (IVIG), epidermal necrolysis disease management 276 Iobitridol, structure 235 Iodinated contrast agents, see Radio contrast media Iodixanol, structure 235 Iohexol, structure 235 Iomeprol, structure 235 Iopamidol, structure 235 Iopentol, structure 235 Iopromide, structure 235 Iothalamate, structure 235 Iotrolan, structure 235 Ioversol, structure 235 Ioxaglate, structure 235

Subject Index

Ioxithalamate, structure 235 Ioxitol, structure 235 Isoniazid cutaneous specific immunoglobulin E 219 drug-induced liver injury 288 patch testing 369 serum immunoglobulin E 220 Isosulfan blue, perioperative anaphylaxis induction 207 Josamycin, pediatric drug hypersensitivity 323 Ketoconazole, cutaneous specific immunoglobulin E 219 Ketoprofen, pediatric drug hypersensitivity 323 Kikuchi-Fujimoto disease, drug reaction with eosinophilia and systemic symptoms, differential diagnosis 256 Lamotrigine drug-induced liver injury 290 severe reactions 183 Langerhans’ cell, allergic contact dermatitis role 39, 40 Lansoprazole, drug-induced interstitial nephritis 296, 297, 301, 302 Latex, perioperative anaphylaxis induction 206 Lenalidomide, T cell costimulation 56 Leukotrienes aspirin-induced asthma 343, 344 perioperative anaphylaxis pathophysiology 208 Levofloxacin, non-irritating intradermal test concentration 221 Lipoxins, aspirin-induced asthma 344, 345 Liver injury, see Drug-induced liver injury Loracarbef, structure 194 Loratadine, pediatric drug hypersensitivity 323 Lymph node proliferation assay (LNPA), non-clinical testing 146 Lymphocyte transformation test (LTT) drug-induced liver injury diagnosis 291 drug reaction with eosinophilia and systemic symptoms 260 maculopapular exanthem diagnosis 248 T-cell-mediated drug hypersensitivity reaction testing 385–387 Maculopapular exanthem (MPE) causative drugs and cofactors 243 clinical features 181, 182, 245, 246 diagnosis 248, 360 differential diagnosis 246, 247 epidemiology 242 histopathology 247, 248 pathophysiology 243–245 treatment 249

433

Major histocompatibility complex (MHC), see also HLA alleles non-hapten drugs as antigens 67 peptide antigen presentation 57, 63 p-i concept 67, 69, 71 Meclocillin, structure 193 Medical history, non-betalactam antibiotic adverse reactions 227 Meprobamate, patch testing 369 Meropenem, structure 196 Metamizol basophil activation testing 398, 399 patch testing 369 Methicillin, structure 192 Metronidazole, patch testing 369 Mexiletine, patch testing 369 Minocyclin, severe reactions 183 Misoprostol, patch testing 369 Monoclonal antibodies, see Antibodies Multiple drug hypersensitivity syndrome, clinical features 187, 188 Muromunab, adverse reactions in therapy 156 Nafcillin, structure 192 Natural killer T cell, allergic contact dermatitis role 37, 38, 40 Nedocromil, pediatric drug hypersensitivity 323 Neuromuscular blocking agents (NMBAs) basophil activation testing 396 cross-reactvity 212, 213 epidemiology of drug hypersensitivity reactions 5 perioperative anaphylaxis induction 205, 206 pseudoallergic drug reactions 174 Neutropenia, see Drug-induced neutropenia Nevirapine human immunodeficiency virus and hypersensitivity reactions 89, 90 metabolites 124–126 Norway rat model of rash CD4+ T cell role 118–120 inflammatory infiltrates 120, 122 overview 117, 118 predictive testing 126, 127 prevention and treatment of rash 122, 124 risk factors in rash and hepatotoxicity 115, 116 severe reactions 183 skin reactions 85, 115 tolerance 124 Nickel hypersensitivity cross-linking model for nickel recognition 50, 51 haptens 49, 52, 60 nickel carriers 51 protein complexes 48 T cell response 48–52, 60

434

Nifenazone, trade names 222 Nimesulide patch testing 369 pediatric drug hypersensitivity 323 Nitrofurantoin, cutaneous specific immunoglobulin E 219 Nitrohalobenzenes, sensitization 58–60 Non-betalactam antibiotics, see also Pyrazolones; Quinolones allergic-type hypersensitivity reactions incidence 216, 217 management 226–231 pyrazolones 221–226 quinolones 218, 220, 221 serum-specific immunoglobulin E findings 218, 220 skin test studies 217–219 desensitization patient selection 406 technique 230, 231, 408 pediatric drug hypersensitivity 329, 330 Non-clinical testing in vitro testing, see T-cell-mediated delayed drug hypersensitivity reactions lymph node proliferation assay 146 popliteal lymph node assay direct assay 143, 144 reporter antigen assay 145, 147 secondary and adoptive assay 144, 145 prospects 148, 149 rationale 140 reactive metabolite screening 142 scenarios 147 total dose and structural alerts 141 toxicity studies 142, 143 Non-nucleoside reverse transcriptase inhibitors, see Nevirapine Non-steroidal anti-inflammatory drugs (NSAIDs), see also specific drugs basophil activation testing 399 epidemiology of drug hypersensitivity reactions 5, 6 overview of allergic reactions 340, 347, 348 pediatric drug hypersensitivity 322, 327, 330, 331 pseudoallergic drug reactions 172, 174 severe cutaneous adverse reactions 24, 183 urticaria clinical presentation 346 diagnosis 347 epidemiology 346 histopathology 347 prevention and treatment 347 Nystatin, patch testing 369

Subject Index

Ofloxacin, non-irritating intradermal test concentration 221 Omalizumab, adverse reactions 158 Omeprazole basophil activation testing 400 drug-induced interstitial nephritis 296, 297, 301, 302 patch testing 369 Opioids, perioperative anaphylaxis induction 207 Oxacillin, structure 192 Oxaliplatin, desensitization 419 Oxyphenbutazone hypersensitivity reactions 224 trade names 222 Oxytetracycline, cutaneous specific immunoglobulin E 219 Paclitaxel, desensitization 417 Pancreatitis, drug allergy 187 Pantoprazole, drug-induced interstitial nephritis 296, 297, 301, 302 Patch tests, see Skin tests Pediatric drug hypersensitivity antibiotic induction betalactams 328, 329 multiple antibiotic sensitivity syndrome 330 non-betalactams 329, 330 chemotherapy agent desensitization 419, 420 clinical manifestations anaphylaxis 327 mild skin reactions 322–325 ocular and respiratory reactions 327 severe skin reactions and systemic symptoms 325–327 diagnosis 324 epidemiology 321–323 non-steroidal anti-inflammatory drugs 322, 327, 330, 331 vaccines generalized reactions 332, 333 local reactions 331, 332 prevention 333–335 Penicillins, see also specific drugs clinical course of hypersensitivity 197 cross-reactivity 201, 202 desensitization patient selection 405 technique 407, 408 diagnosis of hypersensitivity 197–199 immune response 61 pathophysiology of drug reactions 191, 193, 195–197 pediatric drug hypersensitivity 328, 329 perioperative anaphylaxis induction 206 protein modification 61

Subject Index

pseudoallergic drug reactions 174 resensitization 200, 201 severe reactions 183 structures 191–193 treatment of hypersensitivity 202 Pentamidine, cutaneous specific immunoglobulin E 219 Perioperative anaphylaxis clinical features 209, 210 comorbidity 210 definition 204, 205 diagnosis 210–212 epidemiology 205 inducers anesthetics 207 antibiotics 207 antiseptics 207 aprotinin 207 colloids 207 dyes 207 hypnotics 207 latex 206 neuromuscular blocking agents 205, 206 opioids 207 protamine sulfate 207 radio contrast media 208 pathophysiology 208 risk factors 210 screening 213 treatment 213, 214 Phenobarbital, severe cutaneous adverse reactions 24, 25 Phenylbutazone, hypersensitivity reactions 224 Phenytoin, severe reactions 24, 25, 183 p-i concept cross-reactivity between drug and peptide antigens 70, 71 molecular aspects 67, 69, 70 overview 67, 169 prospects for study 71, 72 Piperacillin, structure 193 Polymyxin B, cutaneous specific immunoglobulin E 219 Popliteal lymph node assay contact hypersensitivity model 146 direct assay 143, 144 reporter antigen assay 145, 147 secondary and adoptive assay 144, 145 Pristinamycin cutaneous specific immunoglobulin E 219 patch testing 369 severe cutaneous adverse reactions 27, 183 Propofol, perioperative anaphylaxis induction 206 Propyphenazone hypersensitivity reactions 224 trade names 222

435

Prostaglandins aspirin-induced asthma 343, 345 perioperative anaphylaxis pathophysiology 208 Protamine sulfate, perioperative anaphylaxis induction 207 Pseudoallergic drug reactions, features 172–174 Pseudoephedrine, patch testing 369 Pyrazolones, see also specific drugs historical perspective 221 hypersensitivity reactions cross-reactivity 226 genetic susceptibility 226 incidence 223 serum-specific immunoglobulin E 223–226 skin tests 223 types and trade names 222 Quinolones, see also specific drugs hypersensitivity reactions 218–221 non-irritating intradermal test concentrations 221 severe reactions 183 Rabeprazole, drug-induced interstitial nephritis 296, 297, 301, 302 Radioallergosorbent test (RAST) betalactam hypersensitivity diagnosis 199 perioperative anaphylaxis diagnosis 211 Radio contrast media (RCM) excretion 234 hypersensitivity reactions epidemiology 5, 6, 8 immediate hypersensitivity reactions clinical manifestations 234 pathophysiology 234, 236 prevalence 234 prevention and treatment 236, 237 late hypersensitivity reactions clinical manifestations 237 pathophysiology 238, 239 prevalence 237, 238 prevention and treatment 239–241 patch testing 369 perioperative anaphylaxis induction 208 structures 233, 235 Regulatory T cell (Treg) allergic contact dermatitis role 34, 41 chemical tolerance role 76–80 chemokine receptors 77 subsets 77, 78 suppressive activity 78, 79 toxic epidermal necrolysis role 136, 137 therapeutic targeting 138 Rifampicin cutaneous specific immunoglobulin E 219

436

patch testing 369 serum immunoglobulin E 220 Rifamycin SV cutaneous specific immunoglobulin E 219 serum immunoglobulin E 220 Rituximab adverse reactions in therapy 158 desensitization 420, 421 Rocurinium, perioperative anaphylaxis induction 205 Roxithromycin, cutaneous specific immunoglobulin E 219 Salbutamol pediatric drug hypersensitivity 323 perioperative anaphylaxis management 214 Serum-sickness-like disease (SSLD) clinical features 360 pediatric drug hypersensitivity 322, 327 Severe cutaneous adverse reactions, see specific diseases Single-nucleotide polymorphisms (SNPs), drug hypersensitivity susceptibility studies in Han Chinese 108, 109 Skin tests betalactam hypersensitivity diagnosis 198, 199 intradermal tests 370, 371 non-betalactam antibiotic adverse reactions 227, 228 patch tests drug concentration 366, 367 interpretation 367, 370 material preparation 367 negative predictive value 373 safety 374 sites 367 specificity and relevance 374, 375 vehicles 367 perioperative anaphylaxis diagnosis 211, 212 prick tests 370 utility 371–373 Spiramycin, cutaneous specific immunoglobulin E 219 Stevens-Johnson syndrome (SJS) clinical features early 271, 272 late 272, 273 overview 184, 185 course and complications 273 definition 18, 19, 106 diagnosis 274 epidemiology data sources 20, 21 etiologic factors 23–27 genetic epidemiology 28 incidence 21, 22 risk estimation methodology 22, 23 epidermal necrolysis

Subject Index

animal models 270 drug-specific immune response 269, 270 genetic susceptibility 271 pathomechanisms 269 reactive metabolites 271 virus infection 271 pediatric drug hypersensitivity 325, 326 population differences in susceptibility 106, 107 prognosis 273, 274 risk factors 274, 275 sequelae 274 treatment 275, 276 Streptomycin, cutaneous specific immunoglobulin E 219 Succinylcholine, perioperative anaphylaxis induction 205, 206 Sulfamethoxazole (SMX) cutaneous specific immunoglobulin E 219 drug-induced liver injury 287, 288 human immunodeficiency virus and hypersensitivity reactions 86–88 hypersensitivity mechanisms 61, 62 metabolism 86 prohapten activity 67 serum immunoglobulin E 220 T-cell-mediated delayed drug hypersensitivity reactions 181 Sulfasalazine, severe reactions 183 Sulfinpyrazone, trade names 222 Suxibuzone hypersensitivity reactions 224 trade names 222 T-cell-mediated delayed drug hypersensitivity reactions clinical features acute generalized exanthematous pustulosis 182, 184 maculopapular exanthem 181, 182 multiple drug hypersensitivity syndrome 187, 188 overview 179, 181 Stevens-Johnson syndrome 184, 185 systemic reactions 185–187 toxic epidermal necrolysis 184, 185 pathomechanisms 381 subclassification 176, 177 testing in vitro activation events 380, 381 cytokine assays 384, 385 cytotoxicity assay 387, 388 flow cytometry of cell surface markers 383, 384 lymphocyte transformation test 385–387 prospects 388 rationale 380 timing 382, 383

Subject Index

tolerance mechanisms 177, 179 T cell receptor (TCR) drug and hapten recognition 47, 48, 52 nickel carriers 51 cross-linking model for nickel recognition 50, 51 nickel-induced epitopes 49, 50 non-hapten drugs as antigens 67 peptide antigen presentation 57, 63 p-i concept 67, 69, 71 Teicoplanin cutaneous specific immunoglobulin E 219 patch testing 369 Telithromycin, cutaneous specific immunoglobulin E 219 Terbinafine patch testing 369 severe reactions 27, 183 Tetracycline, cutaneous specific immunoglobulin E 219 Tetrazepam, patch testing 369 TGN-1412 adverse reactions 157 cytokine storm 183 T cell costimulation 56 Thiopental, perioperative anaphylaxis induction 206 Thrombocytopenia, see Drug-induced thrombocytopenia Ticarcillin, structure 192 Tienilic acid, drug-induced liver injury 286, 287 Tobramycin, cutaneous specific immunoglobulin E 219 Tolerance animal models 74–76 breaking 80 induction and restoration 81 nevirapine 124 T cell-mediated delayed drug hypersensitivity reaction mechanisms 177, 179 T regulatory cell phenotypes and functions 76–80 Toll-like receptors (TLRs), allergic contact dermatitis role 35–37 Toxic epidermal necrolysis (TEN) clinical features early 271, 272 late 272, 273 overview 184, 185 course and complications 273 definition 18, 19, 106 diagnosis 274 epidemiology data sources 20, 21 etiologic factors 23–27 genetic epidemiology 28 incidence 21, 22 risk estimation methodology 22, 23

437

Toxic epidermal necrolysis (TEN) (continued) epidermal antigen-specific cytotoxic T lymphocytes 133 epidermal necrolysis animal models 270 drug-specific immune response 269, 270 genetic susceptibility 271 pathomechanisms 269 reactive metabolites 271 virus infection 271 Fas modulation 137, 138 graft-versus-host disease association 129, 130 inductive model 133 K5-mOVA.OT-I double transgenic mouse model 135, 136 pediatric drug hypersensitivity 325, 326 population differences in susceptibility 106, 107 prognosis 273, 274 regulatory T cells role 136, 137 therapeutic targeting 138 risk factors 274, 275 sequelae 274 T cell attack on keratinocytes 130, 131 three-phase model 134 treatment 275, 276 Transforming growth factor- (TGF-), tolerance role 81 Trastuzumab adverse reactions in therapy 158, 160 desensitization 420, 421 Triamcinolone, patch testing 369

438

Trimethoprim cutaneous specific immunoglobulin E 219 human immunodeficiency virus and hypersensitivity reactions 86 serum immunoglobulin E 220 Troglitazone, drug-induced liver injury 290 Tumor necrosis factor-
(TNF-
), cytokine release syndrome 155 Ultraviolet B, immunosuppression 75 Urticaria aspirin and non-steroidal anti-inflammatory drug induction clinical presentation 346 diagnosis 347 epidemiology 346 histopathology 347 prevention and treatment 347 clinical features 359 Vaccines, pediatric hypersensitivity generalized reactions 332, 333 local reactions 331, 332 prevention 333–335 Valaciclovir, patch testing 369 Vancomycin cutaneous specific immunoglobulin E 219 patch testing 369 serum immunoglobulin E 220 Vasculitis, features 362, 363 Vitamin K, patch testing 369 Ximelagatran, drug-induced liver injury 290

Subject Index