American Heart Journal, Volume 161, Issue 2, Feb 2011

MHJ161_2_toc 2/2/11 2:24 PM Page 1 American Heart Journal www.ahjonline.com Table of Contents February 2011, Volume 1...

Author: Robert M. Califf | Daniel B. Mark | Patricia K. Hodgson | Rebecca L. Hines

360 downloads 1507 Views 7MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

MHJ161_2_toc 2/2/11 2:24 PM Page 1

American Heart Journal www.ahjonline.com

Table of Contents

February 2011, Volume 161, Number 2

Editorial

Curriculum in Cardiology

221 Challenge of rehospitalizations for heart failure:

241 Atrial fibrillation, anticoagulation, fall risk, and outcomes

in elderly patients

Potential of natriuretic doses of mineralocorticoid receptor antagonists Robert W. Schrier, MD and Mihai Gheorghiade, MD, Denver, CO; and Chicago, IL

Matthew B. Sellers, MD and L. Kristin Newby, MD, MHS, Durham, NC 247 Primary percutaneous coronary intervention for acute

myocardial infarction: Is it worth the wait?: The risktime relationship and the need to quantify the impact of delay

Special Articles 224 Clinical development of pharmacologic agents for acute

Giuseppe Tarantini, MD, PhD, Frans Van de Werf, MD, PhD, Claudio Bilato, MD, PhD, and Bernard Gersh, MB, ChB, DPhil, FRCP, Padua, Italy; Leuven, Belgium; and Rochester, NY

heart failure syndromes: A proposal for a mechanistic translational phase Mihai Gheorghiade, MD, Peter S. Pang, MD, Christopher M. O’Connor, MD, Krishna Prasad, MD, John McMurray, MD, John R. Teerlink, MD, Mona Fiuzat, PharmD, Hani Sabbah, PhD, and Michel Komajda, MD, Chicago, IL; Durham, NC; London, England; Glasgow, United Kingdom; San Francisco, CA; Detroit, MI; and Paris, France 233 Lessons learned from a pediatric clinical trial: The

Pediatric Heart Network Angiotensin-Converting Enzyme Inhibition in Mitral Regurgitation Study Jennifer S. Li, MD, MHS, Steven D. Colan, MD, Lynn A. Sleeper, ScD, Jane W. Newburger, MD, MPH, Victoria L. Pemberton, RNC, MS, Andrew M. Atz, MD, Meryl S. Cohen, MD, Fraser Golding, MD, Gloria L. Klein, Ronald V. Lacro, MD, Elizabeth Radojewski, RN, Marc E. Richmond, MD, and L. LuAnn Minich, MD, Durham, NC; Watertown and Boston, MA; Bethesda, MD; Charleston, SC; Philadelphia, PA; Ontario, Canada; New York, NY; and Salt Lake City, UT

Trial Design 254

Design and rationale of the RadIal Vs. femorAL access for coronary intervention (RIVAL) trial: A randomized comparison of radial versus femoral access for coronary angiography or intervention in patients with acute coronary syndromes Sanjit S. Jolly, MD, MSc, Kari Niemelä, MD, PhD, Denis Xavier, MD, Petr Widimsky, MD, Andrzej Budaj, MD, PhD, Vicent Valentin, MD, Basil S. Lewis, MD, Alvaro Avezum, MD, PhD, Philippe Gabriel Steg, MD, Sunil V. Rao, MD, John Cairns, MD, Susan Chrolavicius, BScN, Salim Yusuf, MBBS, D.Phil, and Shamir R. Mehta, MD, MSc, Ontario and Vancouver, Canada; Tampere, Finland; Bangalore, India; Prague, Czech Republic; Warsaw, Poland; Valencia, Spain; Haifa, Israel; Sao Paulo, Brazil; Paris, France; and Durham, NC

American Heart Journal (ISSN 0002-8703) is published monthly by Mosby, 360 Park Avenue South, New York, NY 10010-1710. Periodicals postage paid at New York, NY and additional mailing offices. POSTMASTER: Send address changes to American Heart Journal, Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043, USA.

American Heart Journal

1A

Contents 261

continued

A randomized, partially blinded, multicenter, active-controlled, dose-ranging study assessing the safety, efficacy, and pharmacodynamics of the REG1 anticoagulation system in patients with acute coronary syndromes: Design and rationale of the RADAR Phase IIb trial Thomas J. Povsic, MD, PhD, Mauricio G. Cohen, MD, Roxana Mehran, MD, Christopher E. Buller, MD, Christoph Bode, MD, Jan H. Cornel, MD, Jaroslaw D. Kasprzak, MD, Gilles Montalescot, MD, Diane Joseph, William A. Wargin, PhD, Christopher P. Rusconi, PhD, Steven L. Zelenkofske, DO, Richard C. Becker, MD, and John H. Alexander, MD, MHS, Durham, and Chapel Hill, NC; Miami, FL; New York, NY; Ontario, Canada; Freiberg, Germany; Alkmaar, Netherlands; Lódz´, Poland; Paris, France; and Basking Ridge, NJ

269

Associations between cardiovascular parameters and uteroplacental Doppler (blood) flow patterns during pregnancy in women with congenital heart disease: Rationale and design of the Zwangerschap bij Aangeboren Hartafwijking (ZAHARA) II study Ali Balci, MD, MSc, Krystyna M. Sollie, MD, Barbara J. M. Mulder, MD, PhD, Monique W. M. de Laat, MD, PhD, Jolien W. Roos-Hesselink, MD, PhD, Arie P. J. van Dijk, MD, PhD, Elly M. C. J. Wajon, MD, Hubert W. Vliegen, MD, PhD, Willem Drenthen, MD, PhD, Hans L. Hillege, MD, PhD, Jan G. Aarnoudse, MD, PhD, Dirk J. van Veldhuisen, MD, PhD, and Petronella G. Pieper, MD, PhD, Groningen, Amsterdam, Rotterdam, Nijmegen, Enschede, Leiden, and Utrecht, The Netherlands Clinical Investigations

283 The influence of time from symptom onset and reper-

fusion strategy on 1-year survival in ST-elevation myocardial infarction: A pooled analysis of an early fibrinolytic strategy versus primary percutaneous coronary intervention from CAPTIM and WEST Cynthia M. Westerhout, PhD, Eric Bonnefoy, MD, Robert C. Welsh, MD, Philippe Gabriel Steg, MD, Florent Boutitie, PhD, and Paul W. Armstrong, MD, Edmonton, Canada; and Lyon and Paris, France 291 Has the ClOpidogrel and Metoprolol in Myocardial

Infarction Trial (COMMIT) of early -blocker use in acute coronary syndromes impacted on clinical practice in Canada? Insights from the Global Registry of Acute Coronary Events (GRACE) Jeremy Edwards, MD, Shaun G. Goodman, MD, MSc, Raymond T. Yan, MD, Robert C. Welsh, MD, Jan M. Kornder, MD, J. Paul DeYoung, MD, Denis Chauret, MD, Jean-Pierre Picard, MD, Kim A. Eagle, MD, and Andrew T. Yan, MD, Ontario, Alberta, British Columbia, and Quebec, Canada; and Ann Arbor, MI 298

Incidence and clinical consequences of acquired thrombocytopenia after antithrombotic therapies in patients with acute coronary syndromes: Results from the Acute Catheterization and Urgent Intervention Triage Strategy (ACUITY) trial Adriano Caixeta, MD, PhD, George D. Dangas, MD, PhD, Roxana Mehran, MD, Frederick Feit, MD, Eugenia Nikolsky, MD, PhD, Alexandra J. Lansky, MD, Jiro Aoki, MD, PhD, Jeffrey W. Moses, MD, Steven R. Steinhubl, MD, Harvey D. White, DSc, E. Magnus Ohman, MD, Steven V. Manoukian, MD, Martin Fahy, MSc, and Gregg W. Stone, MD, New York, NY; New Haven, CT; Lexington, KY; Auckland, New Zealand; Durham, NC; and Nashville, TN

Acute Ischemic Heart Disease 276 Prehospital triage in the ambulance reduces infarct

size and improves clinical outcome Sonja Postma, MSc, Jan-Henk E. Dambrink, MD, PhD, Menko-Jan de Boer, MD, PhD, A. T. Marcel Gosselink, MD, PhD, Gerrit J. Eggink, MD, Henri van de Wetering, MANP, Frans Hollak, RN, Jan Paul Ottervanger, MD, PhD, Jan C. A. Hoorntje, MD, PhD, Evelien Kolkman, MSc, Harry Suryapranata, MD, PhD, and Arnoud W. J. van ‘t Hof, MD, PhD, Zwolle, The Netherlands

2A

American Heart Journal

Valvular and Congenital Heart Disease 307 Pregnancy in women with corrected tetralogy of Fallot: Occurrence and predictors of adverse events Ali Balci, MD, MSc, Willem Drenthen, MD, PhD, Barbara J. M. Mulder, MD, PhD, Jolien W. RoosHesselink, MD, PhD, Adriaan A. Voors, MD, PhD, Hubert W. Vliegen, MD, PhD, Philip Moons, RN, PhD, Krystyna M. Sollie, MD, Arie P. J. van Dijk, MD, PhD, Dirk J. van Veldhuisen, MD, PhD, and Petronella G. Pieper, MD, PhD, Groningen, Utrecht, Amsterdam, Rotterdam, Leiden, and Nijmegen, The Netherlands; and Leuven, Belgium

Contents

continued

Imaging and Diagnostic Testing 314 Left atrial reverse remodeling and functional improve-

ment after mitral valve repair in degenerative mitral regurgitation: A real-time 3-dimensional echocardiography study Nina Ajmone Marsan, MD, Francesco Maffessanti, MS, Gloria Tamborini, MD, Paola Gripari, MD, Enrico Caiani, PhD, Laura Fusini, MD, Manuela Muratori, MD, Marco Zanobini, PhD, Francesco Alamanni, MD, and Mauro Pepi, MD, Milan, Italy Congestive Heart Failure 322 Certoparin versus unfractionated heparin to prevent

venous thromboembolic events in patients hospitalized because of heart failure: A subgroup analysis of the randomized, controlled CERTIFY study Ulrich Tebbe, MD, Sebastian M. Schellong, MD, Sylvia Haas, MD, Horst Eberhard Gerlach, MD, Claudia Abletshauser, PhD, Christian Sieder, MSc, Peter Bramlage, MD, and Hanno Riess, MD, Detmold, Dresden, München, Mannheim, Nürnberg, Mahlow, and Berlin, Germany 329

A randomized controlled trial evaluating the safety and efficacy of cardiac contractility modulation in advanced heart failure Alan Kadish, MD, Koonlawee Nademanee, MD, Kent Volosin, MD, Steven Krueger, MD, Suresh Neelagaru, MD, Nirav Raval, MD, Owen Obel, MD, Stanislav Weiner, MD, Marc Wish, MD, Peter Carson, MD, Kenneth Ellenbogen, MD, Robert Bourge, MD, Michael Parides, PhD, Richard P. Chiacchierini, PhD, Rochelle Goldsmith, PhD, Sidney Goldstein, MD, Yuval Mika, PhD, Daniel Burkhoff, MD, PhD, and William T. Abraham, MD, Chicago, IL; Inglewood, CA; Philadelphia, PA; Lincoln, NE; Amarillo, Dallas, and Tyler, TX; Atlanta, GA; Fairfax, and Richmond, VA; Birmingham, AL; New York, and Orangeburg, NY; Detroit, MI; and Columbus, OH

338

Effects of n-3 polyunsaturated fatty acids on malignant ventricular arrhythmias in patients with chronic heart failure and implantable cardioverter-defibrillators: A substudy of the Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca (GISSI-HF) trial Andrea A. Finzi, MD, Roberto Latini, MD, Simona Barlera, MSc, Maria G. Rossi, MD, Albarosa Ruggeri, MD, Alessandro Mezzani, MD, Chiara Favero, BSc, Maria G. Franzosi, BiolD, Domenico Serra, MD, Donata Lucci, MSc, Francesca Bianchini, BSc, Roberto Bernasconi, Aldo P. Maggioni, MD, Gianluigi Nicolosi, MD, Maurizio Porcu, MD, Gianni Tognoni, MD, Luigi Tavazzi, MD, and Roberto Marchioli, MD, Milano, Reggio Calabria, Veruno, Florence, Pordenone, Cagliari, S Maria Imbaro, and Cotignola, Italy; and Lugano, Switzerland

Coronary Artery Disease 344 Common oral mucosal diseases, systemic inflammation, and cardiovascular diseases in a large cross-sectional US survey Stefano Fedele, DDS, PhD, Wael Sabbah, BDS, MSc, Nikos Donos, DDS, MS, PhD, Stephen Porter, BSc, MD, PhD, and Francesco D’Aiuto, London, United Kingdom Prevention and Rehabilitation 351 Reducing cardiovascular disease risk in medically

underserved urban and rural communities Alfred A. Bove, MD, PhD, FACC, William P. Santamore, PhD, Carol Homko, RN, PhD, Abul Kashem, MD, PhD, Robert Cross, MD, Timothy R. McConnell, PhD, Gail Shirk, RN, and Francis Menapace, MD, Philadelphia, Bloomsburg, and Danville, PA Interventional Cardiology 360 Incidence and clinical outcome of minor surgery in the

year after drug-eluting stent implantation: Results from the Evaluation of Drug-Eluting Stents and Ischemic Events Registry Emmanouil S. Brilakis, MD, PhD, David J. Cohen, MD, MSc, Neal S. Kleiman, MD, Michael Pencina, PhD, Deborah Nassif, PhD, Jorge Saucedo, MD, Robert N. Piana, MD, Subhash Banerjee, MD, Michelle J. Keyes, PhD, Chen-Hsing Yen, MS, and Peter B. Berger, MD, Dallas, and Houston, TX; Kansas City, MO; Boston, MA; Oklahoma City, OK; Nashville, TN; and Danville, PA Continued on page 4A

American Heart Journal

3A

Contents

continued

367 Qualitative assessment of neointimal tissue after drug-

391 Impact of baseline thrombocytopenia on the early and

eluting stent implantation: Comparison between follow-up optical coherence tomography and intravascular ultrasound

late outcomes after ST-elevation myocardial infarction treated with primary angioplasty: Analysis from the Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction (HORIZONSAMI) trial

Sung Woo Kwon, MD, Byeong-Keuk Kim, MD, Tae-Hoon Kim, MD, Jung-Sun Kim, MD, Young-Guk Ko, MD, Donghoon Choi, MD, Yangsoo Jang, MD, and Myeong-Ki Hong, MD, Seoul, Korea 373

Standard versus high loading doses of clopidogrel in Asian ST-segment elevation myocardial infarction patients undergoing percutaneous coronary intervention: Insights from the Korea Acute Myocardial Infarction Registry Cheol Ung Choi, MD, Seung-Woon Rha, MD, Dong Joo Oh, MD, Kanhaiya L. Poddar, MBBS, Jin Oh Na, MD, Jin Won Kim, MD, Hong Euy Lim, MD, Eung Ju Kim, MD, Chang Gyu Park, MD, Hong Seog Seo, MD, Taek Jong Hong, MD, Jong-Seon Park, MD, Young Jo Kim, MD, Seung Ho Hur, MD, In Whan Seong, MD, Jei Keon Chae, MD, Myeong Chan Cho, MD, Jang Ho Bae, MD, Dong Hoon Choi, MD, Yang Soo Jang, MD, In Ho Chae, MD, Hyo Soo Kim, MD, Chong Jin Kim, MD, Jung Han Yoon, MD, Tae Hoon Ahn, MD, SeungJea Tahk, MD, Wook Sung Chung, MD, Ki Bae Seung, MD, Shung Chall Chae, MD, Seung Jung Park, MD, Young Keun Ahn, MD, and Myung Ho Jeong, MD, Seoul, Pusan, Daegu, Daejeon, Jeonju, Chongju, Bundang, Wonju, and Gwangju, South Korea

383 Comparison of 2 point-of-care platelet function tests,

VerifyNow Assay and Multiple Electrode Platelet Aggregometry, for predicting early clinical outcomes in patients undergoing percutaneous coronary intervention Young-Guk Ko, MD, Jung-Won Suh, MD, PhD, Bo Hyun Kim, BA, Chan Joo Lee, MD, Jung-Sun Kim, MD, PhD, Donghoon Choi, MD, PhD, Myeong-Ki Hong, MD, PhD, Myung-Ki Seo, MD, Tae-Jin Youn, MD, PhD, In-Ho Chae, MD, PhD, Dong Joo Choi, MD, PhD, and Yangsoo Jang, MD, PhD, Seoul, and Gyeonggi-do, Korea

4A

Diaa A. Hakim, MD, PhD, George D. Dangas, MD, PhD, Adriano Caixeta, MD, PhD, Eugenia Nikolsky, MD, PhD, Alexandra J. Lansky, MD, Jeffrey W. Moses, MD, Bimmer Claessen, MD, Elias Sanidas, MD, Harvey D. White, DSc, E. Magnus Ohman, MD, Steven V. Manoukian, MD, Martin Fahy, MSc, Roxana Mehran, MD, and Gregg W. Stone, MD, New York, NY; New Haven, CT; Auckland, New Zealand; Durham, NC; and Atlanta, GA

American Heart Journal

397

Temporal changes in the outcomes of patients with diabetes mellitus undergoing percutaneous coronary intervention in the National Heart, Lung, and Blood Institute dynamic registry Elizabeth M. Holper, MD, MPH, J. Dawn Abbott, MD, Suresh Mulukutla, MD, Helen Vlachos, MSc, Faith Selzer, PhD, Darren McGuire, MD, MHSc, David P. Faxon, MD, Warren Laskey, MD, Vankeepuram S. Srinivas, MD, Oscar C. Marroquin, MD, and Alice K. Jacobs, MD, Dallas, TX; Providence, RI; Pittsburgh, PA; Boston, MA; Albuquerque, NM; and New York, NY

Surgery 404 Clopidogrel loading dose and bleeding outcomes in patients undergoing urgent coronary artery bypass grafting Nicholas L. M. Cruden, PhD, MBChB, MRCP, Kristin Morch, Daniel R. Wong, MD, MPH, FRCSC, W. Peter Klinke, MD, FRCPC, John Ofiesh, MD, FRCSC, and J. David Hilton, MD, FRCPC, British Columbia, Canada; and Edinburgh, United Kingdom

Contents

continued

Pediatrics 411 Factors associated with the physical activity level of

children who have the Fontan procedure Patricia E. Longmuir, PhD, Jennifer L. Russell, MD, FRCP(C), Mary Corey, PhD, Guy Faulkner, PhD, and Brian W. McCrindle, MD, FRCP(C), Toronto, Canada 418

Corrections

Letters to the Editor e5 An important indirect drug interaction between

dronedarone and warfarin that may be extrapolated to other drugs that can alter gastrointestinal function James A. Reiffel, MD, New York, NY e7 Shirolkar’s reply to Reiffel’s letter to the editor

Shailesh C. Shirolkar, MD, Durham, NC

American Heart Journal

5A

Editor in Chief: Robert M. Califf, MD, Durham, NC Editor: Daniel B. Mark, MD, MPH, Durham, NC Executive Editor: Patricia K. Hodgson, Durham, NC Managing Editor: Rebecca L. Hines, Durham, NC Editorial Assistant: Brenda McCoy, Durham, NC Issue Manager: Todd C. Reiss, New York, NY

AHJ

American Heart Journal www.ahjonline.com

Founded in 1925

Associate Editors Sana M. Al-Khatib, MD, Durham Thomas M. Bashore, MD, Durham Richard C. Becker, MD, Durham Jeffrey S. Berger, MD, MS New York Salvador Borges - Neto, MD, Durham Mark P. Donahue, MD, Durham Mark H. Drazner, MD, MSc, Dallas

G. Mike Felker, MD, Durham Christopher B. Granger, MD, Durham Robert A. Harrington, MD, Durham John K. Harrison, MD, Durham Paul A. Heidenreich, MD, MS, Palo Alto Adrian F. Hernandez, MD, Durham Jennifer S. Li, MD, MHS, Durham

Renato D. Lopes, MD, PhD, MHS, Durham Kenneth W. Mahaffey, MD, Durham Darren K. McGuire, MD, MHSc, Dallas Rajendra H. Mehta, MD, Durham L. Kristin Newby, MD, Durham E. Magnus Ohman, MD, Durham Manesh R. Patel, MD, Durham

Thomas J. Povsic, MD, Durham Sunil V. Rao, MD, Durham Matthew T. Roe, MD, MHS, Durham Joseph G. Rogers, MD, Durham Lynda A. Szczech, MD, Durham David J. Whellan, MD, MHS Philadelphia

Editorial Board Frank V. Aguirre, MD John H. Alexander, MD Karen P. Alexander, MD Andrew S. Allen, PhD Larry A. Allen, MD Kevin J. Anstrom, PhD Elliott M. Antman, MD Paul W. Armstrong, MD Alvaro Avezum, MD, PhD Steven R. Bailey, MD Gust H. Bardy, MD Gregory W. Barsness, MD Eric R. Bates, MD Jeroen J. Bax, MD, PhD George A. Beller, MD Peter B. Berger, MD Deepak L. Bhatt, MD Philip F. Binkley, MD John A. Bittl, MD Vera Bittner, MD, MSPH Roger S. Blumenthal, MD Eric Boersma, MSc, PhD Robert O. Bonow, MD Sorin J. Brener, MD Ralph G. Brindis, MD, MPH Christopher E. Buller, MD Javed Butler, MD, MPH Alfred E. Buxton, MD Christopher P. Cannon, MD Warren J. Cantor, MD Blase A. Carabello, MD Melvin D. Cheitlin, MD W. Randolph Chitwood Jr, MD David J. Cohen, MD, MSc Marc Cohen, MD Mauricio G. Cohen, MD Lawrence H. Cohn, MD Heidi M. Connolly, MD Stuart J. Connolly, MD Anne B. Curtis, MD Ralph J. Damiano, MD James P. Daubert, MD 6A American Heart Journal

Charles J. Davidson, MD James A. De Lemos, MD Elizabeth R. DeLong, PhD David L. DeMets, PhD J. Michael DiMaio, MD John P. DiMarco, MD, PhD Pamela S. Douglas, MD Vladimir Dzavik, MD Kim A. Eagle, MD Kenneth A. Ellenbogen, MD Stephen G. Ellis, MD Nathan R. Every, MD, MPH David P. Faxon, MD Gregg C. Fonarow, MD Gary S. Francis, MD John K. French, MB, PhD Thomas C. Gerber, MD, PhD Bernard J. Gersh, MB, ChB, DPhil Myron C. Gerson, MD Jalal K. Ghali, MD Mihai Gheorghiade, MD Raymond J. Gibbons, MD C. Michael Gibson, MS, MD Robert P. Giugliano, MD Donald D. Glower, MD Lee Goldman, MD Pascal J. Goldschmidt-Clermont, MD John R. Guyton, MD Frank E. Harrell Jr, PhD David Hasdai, MD Edward P. Havranek, MD Robert C. Hendel, MD Timothy D. Henry, MD Charles A. Herzog, MD James A. Hill, MD Mark A. Hlatky, MD Judith S. Hochman, MD David R. Holmes Jr, MD Michael P. Hudson, MD, MHSc Alice K. Jacobs, MD James L. Januzzi, MD Marc E. Jolicoeur, MD

David E. Kandzari, MD Padma Kaul, PhD Dean J. Kereiakes, MD Morton J. Kern, MD Neal S. Kleiman, MD George J. Klein, MD Richard A. Krasuski, MD William E. Kraus, MD Joel Kupersmith, MD Gervasio A. Lamas, MD Kerry L. Lee, PhD Thomas H. Lee Jr, MD A. Michael Lincoff, MD Gregory Y.H. Lip, MD James E. Lock, MD Thomas H. Marwick, MB, PhD Frederick A. Masoudi, MD, MSPH Barry M. Massie, MD Charles Maynard, PhD George A. Mensah, MD D. Douglas Miller, MD Julie M. Miller, MD Todd D. Miller, MD David J. Moliterno, MD Fred Morady, MD David A. Morrow, MD, MPH Joseph Brent Muhlestein, MD Debabrata Mukherjee, MD Sean M. O’Brien, PhD Christoper M. O’Connor, MD Gerald T. O’Connor, MD Elizabeth O. Ofili, MD Peter M. Okin, MD Brian Olshansky, MD Catherine M. Otto, MD Douglas L. Packer, MD Richard L. Page, MD Patricia A. Pellikka, MD Carl J. Pepine, MD Eric D. Peterson, MD Marc A. Pfeffer, MD, PhD Matthias E. Pfisterer, MD

Karen S. Pieper, MS Louise Pilote, MD, MPH, PhD Bertram Pitt, MD Jeanne E. Poole, MD Jeffrey J. Popma, MD Eric N. Prystowsky, MD Paolo Raggi, MD Rita F. Redberg, MD, MSc Harmony Reynolds, MD Vera H. Rigolin, MD Veronique L. Roger, MD, MPH Melvyn Rubenfire, MD John S. Rumsfeld, MD, PhD Thomas Ryan, MD Gillian D. Sanders, PhD Hartzell V. Schaff, MD Kevin A. Schulman, MD Svati H. Shah, MD Leslee Shaw, PhD Thomas N. Skelton, MD Steven R. Steinhubl, MD Lynne W. Stevenson, MD William G. Stevenson, MD Carla A. Sueta, MD, PhD W.H. Wilson Tang, MD Allen J. Taylor, MD James E. Tcheng, MD Alan N. Tenaglia, MD Pierre Theroux, MD Eric J. Topol, MD Jack V. Tu, MD, PhD Hector O. Ventura, MD David A. Vorchheimer, MD Lars Wallentin, MD, PhD Andrew Wang, MD Tracy Y. Wang, MD, MHS W. Douglas Weaver, MD William S. Weintraub, MD J. Marcus Wharton, MD Harvey D. White, DSc Clyde W. Yancy, MD James B. Young, MD

AHJ

American Heart Journal www.ahjonline.com

Founded in 1925

Editorial Communications If you have questions or comments regarding original articles and editorial management, please contact Robert M. Califf, MD, Editor in Chief, American Heart Journal, Duke Clinical Research Institute Communications, 300 West Morgan Street, Suite 800, Durham, NC 27701, Telephone 919-668-5948.

Author Inquiries For inquiries relating to the submission of articles (including electronic submission where available) please visit this journal’s homepage at http://www.ahjonline. You can track accepted articles at http://www.elsevier. com/trackarticle and set up e-mail alerts to inform you of when an article’s status has changed. Also accessible from here is information on copyright, frequently asked questions and more. Contact details for questions arising after acceptance of an article, especially those relating to proofs, will be provided by the publisher.

American Heart Journal (ISSN 0002-8703) is published monthly by Mosby, 360 Park Avenue South, New York, NY 10010-1710. Periodicals postage paid at New York, NY, and additional mailing offices. POSTMASTER: Send address changes to American Heart Journal, Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043, USA.

Funding Body Agreements and Policies Elsevier has established agreements and developed policies to allow authors whose articles appear in journals published by Elsevier, to comply with potential manuscript archiving requirements as specified as conditions of their grant awards. To learn more about existing agreements and policies please visit http://www.elsevier.com/fundingbodies.

Annual Subscription Rates United States and possessions: Individual $298; Institution $699; Student/ Resident: $142. All other countries: Individual $395; Institution $803; Student/Resident $188. Single Issues $67. Prices include postage and are subject to change without notice. For additional information contact Elsevier Inc. Customer Support Department, P.O. Box 945, New York, NY 10010. Tel: (212) 633-3730. Toll-Free (for customers in North America): 1-888-4ES-INFO. Fax: (212) 633-3680. E-mail: [email protected]. For author support contact: [email protected].

Copyright ©2011 Mosby Inc. All rights reserved. This journal and the individual contributions contained in it are protected under copyright by Mosby Inc., and the following terms and conditions apply to their use:

Information for Readers

Orders, Claims, and Journal Inquiries Please contact the Elsevier Customer Service Department nearest you: St. Louis: Elsevier Customer Service Department, 3251 Riverport Lane, Maryland Heights, MO 63043, USA; phone: (800) 654-2452 [toll free within the USA]; (+1) (314) 447-8871 [outside the USA]; fax: (+1) (314) 447-8029; e-mail: [email protected]. Oxford: Elsevier Customer Service Department, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK; phone: (+44) (1865) 843434; fax: (+44) (1865) 843970; e-mail: JournalsCustomerServiceEMEA@ elsevier.com. Tokyo: Elsevier Customer Service Department, 4F Higashi-Azabu, 1-Chome Bldg, 1-9-15 Higashi-Azabu, Minato-ku, Tokyo 106-0044, Japan; phone: (+81) (3) 5561 5037; fax: (+81) (3) 5561 5047; e-mail: [email protected]. Singapore: Elsevier Customer Service Department, 3 Killiney Road, #08-01 Winsland House I, Singapore 239519; phone: (+65) 63490222; fax: (+65) 67331510; e-mail: [email protected]. Advertising Information Advertising orders and inquiries can be sent to: USA, Canada, and South America: Cunningham Associates, 180 Old Tappan Road, Old Tappan, NJ 07675, USA; phone (201) 767-4170; fax (201) 767-8065. Classified advertising orders and inquiries can be sent to Aileen Rivera, Marketing Solutions, Elsevier, 360 Park Avenue South, New York, NY 10010-1710, USA; phone (212) 633-3721; fax (212) 633-3820. Europe and the rest of the world: Julie Toop; phone +44(0) 1865 843016; fax +44 (0) 1865 843976; e-mail: [email protected]. Reprints To order 100 or more reprints for educational, commercial, or promotional use, contact the Commercial Reprints Department, Elsevier Inc., 360 Park Avenue South, New York, NY 10010-1710; E-mail: [email protected]. Author Information For full and complete Author Information, please go to: http://www.ahjonline. com/authorinfo.

Photocopying Single photocopies of single articles may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. For information on how to seek permission visit www.elsevier.com/permissions or call: (+44) 1865 843830 (UK) / (+1) 215 239 3804 (USA). Derivative Works Subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. Permission of the Publisher is required for resale or distribution outside the institution. Permission of the Publisher is required for all other derivative works, including compilations and translations (please consult www.elsevier.com/permissions). Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this journal, including any article or part of an article (please consult www.elsevier.com/permissions). Except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. Although all advertising material is expected to conform to ethical (medical) standards, inclusion in this publication does not constitute a guarantee or endorsement of the quality or value of such product or of the claims made of it by its manufacturer. The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper) American Heart Journal 7A

Editorial

Challenge of rehospitalizations for heart failure: Potential of natriuretic doses of mineralocorticoid receptor antagonists Robert W. Schrier, MD, a and Mihai Gheorghiade, MD b Denver, CO; and Chicago, IL

There are N1 million hospitalizations for heart failure (HF) each year in the United States. During the 60 to 90 days postdischarge for HF, readmission rates are approximately 30%. Taken together, recurrent hospitalizations account for N75% of the US $46 billion in annual HF expenditures. Three fourths of hospitalizations are in patients with exacerbations of previously diagnosed HF. Prospective randomized studies in patients with severe HF have shown improved survival by blocking the mineralocorticoid receptor, primarily with nonnatriuretic doses of spironolactone (RALES).1 Nevertheless, the results of treating patients hospitalized for HF are still disappointing. According to the results of the Acute Decompensated Heart Failure Registry in hospitalized patients, at least 50% of these patients are discharged with continued symptoms. Despite treatment with intravenous loop diuretics in 90% of these patients, 33% are discharged with ≤5 lb of weight loss; and 16% are actually discharged with an increase in body weight.2 Thirty percent of these patients were considered to be resistant to diuretic therapy. Moreover, loop diuretics block sodium chloride entry into the macula densa, which causes further stimulation of the renin-angiotensin-aldosterone system (RAAS). Because angiotensin and aldosterone have both been shown to contribute to cardiac fibrosis and remodeling, this may be a negative component to loop diuretics therapy in HF patients. However, 84% of the patients are admitted with dyspnea, 67% with rales, and 66% with peripheral edema.2 Moreover, these physical findings predict 1-year cardiovascular rehospitalization and mortality; and currently, loop diuretics are the therapy of choice for these symptoms.3 There is, therefore, a need for additional approaches to treat congestive HF, a problem that will no doubt increase with the aging of the population. Gheorghiade et al4 have suggested that early intervention should be undertaken in these patients. Hemodynamic congestion, defined as a high left ventricular filling pressure, generally precedes clinical congestion that then leads to hospitalization for HF. Whether earlier diuretic intervention in patients with

From the aUniversity of Colorado Denver School of Medicine, Denver, CO, and b Northwestern University Feinberg School of Medicine, Chicago, IL. Submitted September 23, 2010; accepted October 29, 2010. 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.039

hemodynamic congestion could have a major impact on hospitalizations and readmissions for clinical congestion remains to be proven. The rate of fluid removal must, however, be judicious because mobilization of interstitial fluid in HF patients may be limited to 12 to 14 mL/min.5 Thus, theoretically, a diuresis in excess of this rate with a loop diuretic may worsen arterial underfilling with diminished cardiac and renal function. There is a more modest diuretic approach that is not often used to treat HF patients with either hemodynamic or clinical congestion, namely, natriuretic doses of mineralocorticoid receptor antagonists. This approach has been used successfully in cirrhotic patients for many years.6 The pathophysiology of sodium and water retention in HF and cirrhosis is very similar.7 The arterial baroreceptors sense arterial underfilling in HF secondary to a decrease in cardiac output, whereas, in cirrhosis, unloading of these arterial baroreceptors occurs secondary to primary systemic arterial vasodilation. Approximately 50% of HF patients have preserved left ventricular function. This is, however, at the expense of increased filling pressure that is associated with increased ventricular wall stress, endomyocardial ischemia, decreased cardiac venous drainage, and secondary mitral and even tricuspid insufficiency.4,8 In cirrhosis, the arterial vasodilation occurs primarily in the splanchnic circulation and is associated with a secondary increase in cardiac output. The pathophysiology of high-output cardiac failure, for example, beriberi and thyrotoxicosis, has a similar pathogenesis as cirrhosis. With this arterial underfilling in both HF and cirrhosis, the neurohumoral axis is stimulated to maintain circulatory integrity; but tradeoffs occur that can be deleterious. Perhaps most important are the increases in the RAAS, sympathetic activity, and the nonosmotic release of arginine vasopressin. Because of the effects of angiotensin and α-adrenergic stimulation to enhance tubular epithelial sodium reabsorption in the more proximal sites in the nephron, the normal escape from the sodium-retaining effects of aldosterone in the renal collecting duct does not occur in patients with severe heart or liver failure. Thus, secondary hyperaldosteronism is pivotal in both congestion in HF and ascites formation in cirrhosis. On this background, natriuretic doses of mineralocorticoid receptor antagonists, not loop diuretics, are the initial diuretics of choice in cirrhotic patients with

222 Schrier and Gheorghiade

American Heart Journal February 2011

Figure 1

Diuretics, digoxin, and angiotensin-converting enzyme inhibitors were withdrawn 4 days before admission to the General Clinical Research Center. The subjects were placed on a constant daily diet of 100 mEq sodium and 60 mEq potassium. A (white bars), Positive cumulative sodium balance in the 6 patients (4 ischemic heart disease, 1 idiopathic cardiomyopathy, and 1 aortic valvular disease). B (black bars), In the same patients, significant negative cumulative sodium balance during 200 mg BID spironolactone (P b .01). C, Increase in urine Na−:K+ concentration ratio during spironolactone in all 6 patients (P b .05), a finding compatible with aldosterone antagonism. Mean plasma potassium increased from 3.86 ± 0.2 to 4.1 ± 0.2 mEq/L during spironolactone treatment (P b .05). Mean systolic blood pressure (112 ± 7 vs 110 ± 5 mm Hg, P = not significant) and creatinine clearance (87 ± 7 vs 87.2 ± 8 mL/min, not significant) did not change with spironolactone treatment. Plasma human atrial natriuretic peptide decreased significantly with spironolactone (147 ± 58 vs 83 ± 30 mg/L, P b .05). Fluid intake was not restricted, and a mean of 2-kg weight loss occurred.

ascites.6 The International Ascites Club has designated diuretic resistance in cirrhosis as a failure to respond to 400 mg of spironolactone and 160 mg of furosemide. In

practice, however, the spironolactone doses used to treat cirrhotic patients with ascites are rarely N100 to 200 mg/d. Because these patients are not receiving

American Heart Journal Volume 161, Number 2

RAAS inhibitors and are generally receiving a potassiumlosing loop diuretic, hyperkalemia is generally not a problem. In contrast, the standard mean daily dose of spironolactone used in HF patients, based on the RALES study, is 25 mg. The dose-finding RALES article demonstrated no increase in urinary sodium excretion with 25 mg of spironolactone. Therefore, the interpretation of the RALES results to improve survival in HF was an effect of spironolactone to block the nongenomic, nonnatriuretic effects of aldosterone on fibrosis, oxidant injury, and remodeling of the heart.1 There is a paucity of results using natriuretic doses in HF patients.9 A publication in 1961 demonstrated a natriuretic response to 100 mg of spironolactone in 3 patients with HF.10 There was also a small study in diuretic-resistant patients on modest dose of angiotensinconverting enzyme inhibitor that demonstrated a natriuretic response to 100 mg/d of spironolactone.11 Another prospective study was performed on 6 patients with severe HF.12 All other treatments were discontinued, and the HF patients were shown to be avidly retaining sodium. Spironolactone therapy 200 mg BID then was shown to reverse totally the renal sodium retention over a 4-day period (Figure 1). In these small, short-term studies, hyperkalemia was not observed. These studies did not have a clinically important rise in serum potassium concentration. A recent study from Scotland using a record linkage database examined all patients receiving one or more prescriptions for spironolactone in patients with and without HF between 1994 and 2007. Despite a large increase in spironolactone prescriptions, there was no increase in hospitalizations for hyperkalemia.13 Thus, the role of secondary hyperaldosteronism and failure of aldosterone escape in HF patients with hemodynamic or clinical congestion is of pivotal importance. Careful titration of mineralocorticoid receptor antagonists to natriuretic doses (eg, 50-100 mg/d), while monitoring serum potassium concentration and urinary sodium-potassium ratio, is necessary. In this setting, the HF patients should be on a potassium-restricted diet and a potassium-losing loop diuretic without potassium supplementation before starting natriuretic doses of mineralocorticoid antagonists. There should be particular caution for hyperkalemia (N5.5 mEq/L)-related arrhythmias when instituting natriuretic doses of spironolactone in HF patients, particularly those with an estimated glomerular filtration rate b45 ml/min and baseline plasma potassium concentrations N4.5 mEq/L.14 A prospective randomized study needs to be undertaken to test the hypothesis that early treatment of hemodynamic congestion or clinical congestion with natriuretic doses of mineralocorticoid receptor antagonists can decrease symptomology, decrease hospitalizations and readmissions, and be safe and cost-effective in HF patients. A dose-range study of spironolactone in HF was performed by the RALES investigators, and a

Schrier and Gheorghiade 223

significant decrease was observed in sodium retention with the 50 mg and 75 mg/day doses at 9 days.15 The HF population in the United States is approximately 6 million, and effectively blocking the sodium retention secondary to hyperaldosteronism could have an important impact both medically and economically.

References 1. Pitt B, Zannad F, Remme WJ, et al. for the Randomized Aldactone Evaluation Study Investigators. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med 1999;341:709-17. 2. Adams Jr KF, Fonarow GC, Emerman CL, et al. For the ADHERE Scientific Advisory Committee and Investigators. Characteristics and outcomes of patients hospitalized for heart failure in the United States: rationale, design, and preliminary observations form the first 100,000 cases in the Acute Decompensated Failure National Registry (ADHERE). Am Heart J 2005;149:209-16. 3. Dunlay SM, Gheorghiade M, Reid K, et al. Critical elements of clinical follow-up after hospital discharge for heart failure: insights from the EVEREST trial. Eur J Heart Fail 2010;12:367-74. 4. Gheorghiade M, Filippatos G, De Luca L, et al. Congestion in acute heart failure syndromes: an essential target of evaluation and treatment. Am J Med 2006;119:S3-S10. 5. Fauchauld P. Effects of ultrafiltration of body fluid and transcapillary colloid osmotic gradient in hemodialysis patients, improvements in dialysis therapy. Contrib Nephrol 1989;74:170-5. 6. Ginès P, Schrier RW. Renal failure in cirrhosis. N Engl J Med 2009; 361:1279-90. 7. Bansal S, Lindenfeld JA, Schrier RW. Sodium retention in heart failure and cirrhosis: potential role of natriuretic doses of mineralocorticoid antagonist? Circ Heart Fail 2009;2:370-6. 8. Schrier RW. Role of diminished renal function in cardiovascular mortality: marker or pathogenetic factor? J Am Coll Cardiol 2006;47: 1-8. 9. Schrier RW, Masoumi A, Elhassan E. Aldosterone: role in edematous disorders, hypertension, chronic renal failure, and metabolic syndrome. Clin J Am Soc Nephrol 2010;5:1132-40. 10. Braunwald E, Plauth Jr WH, Morrow AGA. Method for the detection and quantification of impaired sodium excretion. Results of an oral sodium tolerance test in normal subjects and in patients with heart disease. Circulation 1965;32:223-31. 11. Van Vliet AA, Donker AJ, Nauta JJ, et al. Spironolactone in congestive heart failure refractory to high-dose loop diuretic and low-dose angiotensin-converting enzyme inhibitor. Am J Cardiol 1993;71:21A-8A. 12. Hensen J, Abraham WT, Durr JA, et al. Aldosterone in congestive heart failure: analysis of determinants and role in sodium retention. Am J Nephrol 1991;11:441-6. 13. Li W, Struthers AD, Fahey T, et al. Spironolactone use and renal toxicity: population based longitudinal analysis. BMJ 2010: 340-c1768. 14. Weir MR, Rolfe M. Potassium homeostasis and renin-angiotensinaldosterone system inhibitors. Clin J Am Soc Nephrol 2010;5: 531-48. 15. The RALES Investigators. Effectiveness of spironolactone added to an angiotensin-converting enzyme inhibitor and a loop diuretic for severe chronic congestive heart failure (The Randomized Aldactone Evaluation Study [RALES]. Am J Cardiol 1996;78:902-7.

Special Articles

Clinical development of pharmacologic agents for acute heart failure syndromes: A proposal for a mechanistic translational phase Mihai Gheorghiade, MD, a Peter S. Pang, MD, a,b Christopher M. O'Connor, MD, c Krishna Prasad, MD, d John McMurray, MD, e John R. Teerlink, MD, f Mona Fiuzat, PharmD, c Hani Sabbah, PhD, g and Michel Komajda, MD h Chicago, IL; Durham, NC; London, England; Glasgow, United Kingdom; San Francisco, CA; Detroit, MI; and Paris, France

Hospitalization for acute heart failure syndromes (AHFS) predicts a poor prognosis, with postdischarge mortality and rehospitalization rates reaching 45% within 60 to 90 days. Despite the use of evidence-based therapies and adherence to national process measures, these event rates have largely remained the same over the past decade. Given the current and growing burden of AHFS, there exists a substantial unmet need for novel therapies that improve outcomes. However, attempts to improve symptoms and/or reduce postdischarge events have failed to produce positive results, either because of safety and/or efficacy. These negative results may be related to the drug itself, the protocol in terms of patient selection and/or end points, and/or the trial execution. Although experts may not agree on the exact reasons to explain the lack of success to date of phase III trials in AHFS, there is agreement that clinical benefits observed in phase II trials were not reproduced in phase III trials. A different approach may be needed. In November of 2009, a meeting was held at the Food and Drug Administration with the primary purpose of identifying the reasons why benefits observed during phase II did not translate into benefits in phase III to improve future trial design. Although multiple domains of trial design were discussed, the participants identified a lack of in-depth understanding of novel molecules before pivotal trials in AHFS as a possible contributor to the disappointing results of recent large trials. In this brief report, we outline the T1 or translational phase of research for AHFS clinical development as an important first step toward greater success in AHFS clinical trials. (Am Heart J 2011;161:224-32.)

Acute heart failure syndromes (AHFS) have been defined as the gradual or rapid onset of signs or symptoms of heart failure (HF) requiring urgent therapy.1 The cost of AHFS is high; N1 million admissions occur every year in the United States, consuming the majority of the US $39 billion spent on HF care, with similar numbers in Europe.2-5 Heart failure is the most common reason for rehospitalization and the most expensive From the aExperimental Therapeutics, Center for Cardiovascular Innovation, Northwestern University Feinberg School of Medicine, Chicago, IL, bDepartment of Emergency Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL, cDuke University Medical Center, Durham, NC, dMHRA/St Thomas' Hospital, London, England, eWestern Infirmary and the British Heart Foundation Glasgow Cardiovascular Research Centre, University of Glasgow, Glasgow, United Kingdom, fSection of Cardiology, San Francisco Veterans Affairs Medical Center and School of Medicine, University of California-San Francisco, San Francisco, CA, gDivision of Cardiology, Wayne State University, Henry Ford Hospital, Detroit, MI, and hDepartment of Cardiology, Hôpital Pitié-Salpêtrière, Paris, France. Submitted August 26, 2010; accepted October 15, 2010. Reprint requests: Mihai Gheorghiade, MD, Division of Cardiology, Center for Cardiovascular Innovation, Experimental Therapeutics, Northwestern University Feinberg School of Medicine, 645 N. Michigan Avenue, Suite 1006 Chicago, IL 60611. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.023

hospital diagnosis for Medicare beneficiaries.6 Most importantly, the postdischarge mortality and rehospitalization rates of patients admitted with AHFS can be as high as 15% and 30%, respectively, within 60 to 90 days.7-9 Despite the use of evidence-based therapies and adherence to national process measures, these event rates have largely remained the same over the past decade.7-9 Given the current and growing burden of AHFS, there exists a substantial unmet need for novel therapies. Unfortunately, every large trial conducted to date in AHFS has failed to produce positive results in terms of efficacy and/or safety.10-18 These neutral or negative results may be related to the drug itself, the trial design in terms of patient selection and/or end points, and/or the trial execution.

Issues related to clinical trials conducted to date Novel therapies in AHFS are developed to improve signs and symptoms during hospitalization, and/or prevent death or rehospitalization postdischarge. There is no simple, all-encompassing answer to explain why

American Heart Journal Volume 161, Number 2

Gheorghiade et al 225

Table I. Contributors to lack of success in phase III AHFS trials Pathophysiology and epidemiology

The therapy (experimental drug or device)

The protocol

Study execution

1. 2. 3. 4. 5. 6.

Poor understanding of the pathophysiology in AHFS Heterogeneous patient population in terms of pathophysiology, etiology, and clinical presentation Uncertain relationship between hemodynamics and neurohormonal changes and clinical outcomes Clinical course (particularly soon after discharge) of AHFS patients has not been well studied. Cardiac and noncardiac comorbid conditions influence the outcome and interaction with therapy. The transition from animal studies to clinical studies has occurred without comprehensive understanding of the mechanistic properties of the drug in specific patient subgroups. Not “knowing” the drug 7. Not having the correct dose 8. Possible variation of efficacy and safety with time (given significant fluctuations in symptoms, hemodynamics, neurohormones, renal function, and myocardial injury during the course of hospitalization and postdischarge, the efficacy and/or safety of the drug may be dependent on the time of the invention) 9. The majority of drugs tested thus far reduce systemic BP, which may potentially decrease coronary perfusion, thereby contributing to myocardial and/or kidney injury 10. Patient selection 11. Surrogate end points and clinical outcomes in phase II do not predict the results of a phase III trial. 12. Choice of end points 13. Because most patients' signs and symptoms improve with standard therapy, it is difficult to prove that novel therapies are producing further and substantial improvements. 14. Selection of incorrect patients and/or less than ideal follow-up

Stage A trials include those short-term therapies (hours, days) targeted at treatment during the initial presentation, Stage B trials are conducted during hospitalization in those patients who continue to have signs and symptoms despite initial therapies (therapy administered for a few days), Stage C trials are therapies initiated shortly before or after discharge (therapy administered over longer duration; weeks or months). BP, Blood pressure.

trials to date have not safely achieved either of these goals (Table I). Although experts may not agree on the exact reasons to explain the lack of success to date of phase III trials in AHFS, there is agreement that clinical benefits observed in phase II trials were not reproduced in phase III trials. Several examples are listed in Table II. After a decade of efforts and hundreds of millions of research dollars invested, it still is not a certainty that effective therapies can be developed for AHFS. It is clear however that a different approach is needed. In November of 2009, a meeting was held at the Food and Drug Administration (FDA) with the goal of identifying the reasons why benefits observed during phase II did not translate into benefits in phase III to improve future trial design. This meeting was not sponsored by any company, organization, or governmental entity; and no extramural funding was used to support this work. The authors are solely responsible for the drafting and editing of the paper and its contents. Principal investigators from 4 major drug development programs (tolvaptan, tezosentan, levosimendan, and rolofylline) took part, as well as representatives from the European Society of Cardiology, the European Medicines Agency, and the FDA. A critical review of the clinical development of these drugs from phase II onward was presented in detail. Although multiple domains of trial design were discussed, the participants identified the lack of in-depth understanding of molecules before pivotal trials in AHFS as a possible contributor to the disappointing results seen to date. The Clinical Research Roundtable at the Institute of Medicine outlined 2 overarching obstacles to the realization of tangible health benefits, which might be

derived from the success of basic science research. They named these obstacles translational blocks42 (Figure 1). The first block, described as T1 or the first translational block,43 is “the transfer of new understandings of disease mechanisms gained in the laboratory into the development of new methods for diagnosis, therapy, and prevention and their first testing in humans.”42 The second translational block, or T2, is “the translation of results from clinical studies into everyday clinical practice and health decision making.”42 From the perspective of clinical trial design in AHFS, overcoming the T1 block was the primary focus of discussion at the FDA meeting; specifically, whether early T1 mechanistic studies—translating the beneficial effects of experimental agents observed in animal models of HF into patients with HF—might increase the likelihood of successful development.

Purpose of studies before pivotal testing Once successful animal model studies have been completed, traditional drug development follows a paradigm of sequential phases of development (Figure 2), culminating in a pivotal phase III trial. Variations on this traditional paradigm, or combination or adaptive designs have also been used in the past. After initial phase I trials, primarily conducted for safety in healthy subjects, phase II studies are designed for initial hypothesis testing in the desired target population to understand the mechanistic properties of the drug, selection of dose, and broad safety. Detailed knowledge of these broad domains, although not a guarantee, maximizes the likelihood of success while minimizing risk to patients. Knowing who to target and when is important

American Heart Journal February 2011

226 Gheorghiade et al

Table II. Clinical Trials in AHFS - From Phase II to Phase III Development program/ phase II studies Tolvaptan ACTIV in CHF19 (n = 319)

ECLIPSE20 (n = 181)

METEOR21 (n = 240)

Tezosentan RITZ-1 22 –symptom study (n = 669) RITZ-2 23 –hemodynamic study (n = 292) RITZ-4 24 –AHFS and ACS (n = 193) RITZ-5 25 –pulmonary edema study (n = 84) Levosimendan Dose ranging versus placebo/dobutamine26 (n = 151)

Dose escalation27 (n = 146) Hemodynamic study29 (n = 85) RUSSLAN 30 –LV failure post-MI (n = 504) LIDO 32 –low-output HF versus dobutamine (n = 203) CASINO33 (n = 227) REVIVE-134,35 (n = 100) Rolofylline CKI-20136 (n = 159) (146 received tx) CKI-202 36 (n = 35) (descriptive study in diuretic-resistant patients) CKI-20337 (n = 32) PROTECT–pilot38 (n = 301)

Nesiritide Nesiritide Efficacy Trial40 (n = 127) Nesiritide Comparative Trial (n = 305) VMAC18 (n = 489)

Phase II primary end point reached?

Primary end point(s)

1. In-hospital outcome: change in body weight at 24 h 2. Outpatient outcome: worsening HF (death, hospitalization, or unscheduled visit for HF) at 60 d

PCWP peak change from baseline within 3 to 8 h after treatment administration (80% power to detect 3–mm Hg difference) Change from baseline in LV end-diastolic volume index at 54-wk visit

Yes—for body weight reduction. No—for worsening HF or postdischarge mortality (in retrospect, patients with severe congestion, hyponatremia, and abnormal renal function had decreased mortality in response to oral tolvaptan) Yes (is borderline)

No. In retrospect, patients assigned to tolvaptan had a lower rate of mortality and hospitalizations.

Dyspnea at 24 h

No

Mean change in CI at 6 h

Yes

Composite of death, worsening HF, recurrent ischemia, and recurrent or new myocardial infarction within 72 h Change in oxygen saturation from baseline to 1 h

No

Proportion achieving in each treatment group at least one of the following: (1) N15% increase in SV at 23 to 24 h; (2) N25% decrease in PCWP (and N4 mm Hg) at 23 to 24 h; (3) N40% increase in CO (with change in HR b20%); (4) N50% decrease in PCWP during 2 consecutive measurements Proportion of patients with N25% increase in SV or decrease in PCWP at 6 h Changes in hemodynamics between 6 and 24 h Change in hemodynamics between 24 and 48 h Hypotension or myocardial ischemia of clinical significance Hemodynamic improvement (increase of N30% in CO and a decrease of N 25% in PCWP) at 24 h Combined death or rehospitalization for worsening HF Global outcomes at 5 d

Yes—positive dose-response relationship for all 5 doses

No

Yes Yes Yes (noninferior) Yes Yes Yes—positive trends

Total urine output at 6 h after first dose

Yes

Change in hourly urine output over 24 h and change in CrCl

Yes—positive trends

Improvement in renal function (GFR) and RBF Trichotomous classification of patients: Success: improvement in dyspnea (Likert scale— moderately or markedly better) at 24 and 48 h or day of discharge, and not meeting criteria for treatment failure. Failure: death, HF readmission within 7 d, worsening HF (by physician assessment by day 7), or persistent renal impairment. Unchanged: neither criteria for success or failure.

Yes Yes–positive trends

Change in PCWP at 6 h Global clinical status and clinical symptoms

Yes—Efficacy Trial Neutral for Comparative Trial

1. Change in PCWP at 3 h 2. Change in dyspnea (Likert) at 3 h

Yes However, retrospective analysis suggests worsening renal function and increased mortality.

American Heart Journal Volume 161, Number 2

Phase III studies Tolvaptan EVEREST12 (n = 4133)

EVEREST Short Term Trials11

Gheorghiade et al 227

Phase III primary end point reached?

Primary end point(s)

Contributors to lack of success per Table I⁎ 4, 6, 7. 8, 11

Long term 1. All-cause mortality (superiority and noninferiority) 2. Cardiovascular death or hospitalization for HF (superiority only)

No

Short term Composite of global clinical status and body weight reduction

Yes (driven by body weight reduction)

2, 3, 4, 5, 7, 8, 9, 10,13

Tezosentan

VERITAS13 (n = 1448)

1. Change in dyspnea (visual analog scale) over 24 h (in the individual trials) 2. Death or worsening HF at 7 d (in both trials combined)

No

3, 5, 6, 7, 9, 11

Levosimendan

REVIVE-2 28 (n = 600)

Composite of clinical signs/symptoms of HF and: 1. Patient reported moderately or markedly improved at 6 h, 24 h, and 5 d (and no criteria for worsening) 2. Worsening (death, patient reported moderate or severe deterioration at any time point) or worsening symptoms at any time, or persistent severe symptoms after 24 h requiring rescue therapy (ie, intravenous diuretic, vasodilator or inotropic agents) 3. Unchanged

Yes—excess hypotension, atrial and ventricular arrhythmias, and trends toward early mortality in levosimendan

SURVIVE31 (n = 1327)

All-cause mortality at 180 d

No

PROTECT39 (n = 2033)

No Trichotomous classification of patients: Success: improvement in dyspnea (Likert scale–moderately or markedly better) at 24 and 48 h or day of discharge, and not meeting criteria for treatment failure. Failure: death, HF readmission within 7 d, worsening HF (by physician assessment by day 7), or persistent renal impairment. Unchanged: neither criteria for success or failure.

3, 4, 7, 8, 9, 12, 13

Nesiritide

ASCEND-HF41

1, 2, 4, 5, 6, 7, 8, 10, 11, 12, 13

1. Composite all-cause mortality + HF rehospitalization at 30 d 2. Dyspnea at 6 and 24 h

No

American Heart Journal February 2011

228 Gheorghiade et al

Figure 1

Figure 2

T1 Block: Translation from Basic Sciences to Human Studies

Basic Biomedical Research

T2 Block: Translation of New Knowledge into Clinical Practice & Health Decision Making

Clinical Science & Knowledge

Improved Health

Proposal for AHFS Clinical Development: The T1 Phase

Translational blocks (adapted and reproduced with permission from Sung et al42).

both for efficacy and safety. Study design is drug/class specific. “Hard” clinical outcomes such as mortality and/or rehospitalization are not the purpose of phase II studies, which are, by design, underpowered to answer such questions. Outcomes analyses will continue to be commonly conducted, however, as pressures to continue late phase development often require such analysis. Overall, well-designed phase II studies will continue to be an important benchmark in clinical development. As previously mentioned, success in phase II has not led to success in phase III. After careful review of past major AHFS development programs, a new hypothesis was proposed: improved phase II trial design preceded by indepth mechanistic testing of the drug after animal studies but before phase II would increase the probability of success in phase III. Broadly, the purpose of such earlyphase studies would be to safely replicate and explore in detail the mechanistic efficacy seen in animal models through translational research in humans. Pharmacokinetic and pharmacodynamic studies would also be conducted because, unlike animal models to date, patients with HF are commonly on chronic background therapy for HF and have significant cardiac and noncardiac comorbidities, which might influence safety, efficacy, and outcomes. Although large animal testing remains a fundamental step in novel drug development, this phase does not fully replicate the conditions of a patient with HF. Human HF has many etiologies and comorbidities; in addition, a patient with HF typically receives extensive background therapy, both cardiac and noncardiac, before being exposed to an experimental drug. Heterogeneity of the HF population combined with long-term exposure to background therapy can potentially attenuate both the safety and efficacy signal of an experimental drug, particularly when compared with highly controlled

Current paradigm for drug development.

preclinical studies. Some of these difficulties can be overcome by exposing animals to background therapy and assessing the safety and efficacy of new drugs as “additive” to background therapy; however, this does not fully resolve the issue. Patients with HF are often on years of background therapy, a state difficult to reproduce in preclinical models.

A proposal: the T1 or mechanistic translational phase (Figure 3B) The transition from preclinical to clinical development of a drug for HF requires the utmost of care. There is an increased awareness that, absent attention to detail, specifically regarding mechanisms of drug action, patient selection and background therapy during this transition can derail an otherwise promising drug development program. The preclinical paradigm, performed under tightly controlled conditions, has not been fully explored in human AHFS clinical development, despite careful

American Heart Journal Volume 161, Number 2

Figure 3

Gheorghiade et al 229

Table III. T1 concept 1. A more thorough understanding of all of a molecule's effects on the heart (effects on viable but noncontractile myocardium, coronary perfusion, diastolic function, etc) is important. 2. Reproduce the results obtained in large animal HF models in homogeneous group of patients taking into account systolic and diastolic dysfunction, extent and severity of CAD, viable but dysfunctional myocardium, etc. 3. These in-depth evaluations should take advantage of recent progress made in noninvasive methods of assessment of cardiac function and structure (echocardiography, MRI spectroscopy, etc). 4. These studies would also expand our understanding of the pharmacokinetic and pharmacodynamic properties of novel molecules because, unlike animal models to date, patients with HF are commonly on background therapy for HF and have substantial comorbid conditions that might influence safety, efficacy, and outcomes. 5. These studies should be conducted in dedicated centers that have the patient population, technology, and expertise to conduct such technically challenging studies.

The current schema (A) and the proposed paradigm (B) for drug development.

inclusion/exclusion criteria. Examples from recent clinical trials are listed below. Selective vasopressin antagonists, such as tolvaptan, an oral V2 receptor antagonist, were developed without fully knowing the effects of unopposed V1 activity on the heart and the vasculature as a result of increasing vasopressin levels (Table II). Although the hypotensive effects of levosimendan were well known, the potential deleterious effects of decreased coronary perfusion were not well studied, especially in patients with coronary artery disease (CAD) (Table II). Although promising renal-protective effects of adenosine blocking agents were seen in small, early-phase trials, a well-powered study to assess the effects of these agents on renal function were not performed before pivotal trials assessing whether renal protection would lead to improved outcomes (Table II). The hemodynamic effects of tezosentan at relatively low doses were not tested in-depth before its hemodynamic properties were studied in a large trial.44 Patient selection to maximize homogeneity during early phases of T1 research forms the basis of our proposal below. Front-loading development programs to determine which patients are more likely to respond as well as

which patients to avoid will establish a foundation upon which to build the later phases of clinical development. In essence, homogeneity strengthens the signal to noise ratio. Although positive signals for efficacy and safety are ideal, clarity of the signal is the critical need during early phases of development. Positive signals may be particularly alluring, yet may be falsely positive if the negative signal was minimal or absent in a small, but heterogeneous sample size. A robust negative signal might halt development, or allow for a careful pause, before continued utilization of limited resources. Conversely, promising novel drugs may have been prematurely halted when negative signals in a small, but heterogeneous sample size drown out the positive effects. For example, patients with myocardial hibernation due to chronic ischemia may respond differently to an inotrope compared with patients with a primary cardiomyopathy. Similarly, patients with viable but dysfunctional myocardium due to causes other than ischemia may respond differently to a drug compared with those who have myocardial scarring. A dedicated mechanistic phase of development, what we have termed the T1 phase to distinguish from phase II, is proposed to ensure a seamless transition from animal studies to phase II (Table III). This phase would occur after initial human studies demonstrate safety to proceed with further development (phase 0, phase I). T1 studies will be designed in an analogous manner to animal studies, with relatively small numbers of homogenous patients for initial hypothesis testing. For example, patients entered into a T1 study would have the same HF etiology, a narrow range of age, ejection fraction, background therapy, kidney function, etc—in other words, a “homogenous” population in whom one would expect similar outcomes of similar magnitude. Such a strong emphasis on homogeneity should minimize background noise. Importantly, multiple homogenous groups will be

American Heart Journal February 2011

230 Gheorghiade et al

studied to best capture which subgroup within the overall heterogeneous HF population would be the initial target population for larger studies. However, emphasizing homogeneity will be a significant hurdle to recruitment, which should be mitigated by the relatively small sample size required for T1 studies. As these studies are mechanistic in nature, clearly defined end points (eg, change in pulmonary capillary wedge pressure, renal function) would be described. The small sample size would also allow for the introduction of uniform study standards across centers to assess cardiac function and performance and minimize geographic variation.

Technical expertise Given the mechanistic focus of T1 and the small numbers of patients, T1 studies would be very technical in nature and would use the latest advances in cardiac imaging and measurement. These advances, from magnetic resonance imaging (MRI) to new biomarkers, have not been well represented in clinical trial design to date. In terms of experience and expertise, studies requiring invasive hemodynamics are a good example, as right heart catheterization becomes less and less part of routine clinical practice. Other examples include echocardiographic expertise (eg, tissue Doppler imaging, 3-dimensional echocardiography) or other modalities (eg, cardiac magnetic resonance imaging [MRI], nuclear imaging/ testing, novel use of biomarkers, coronary perfusion, positron emission tomography, myocardial metabolism, myocardial spectroscopy) as technology and cardiovascular science continue to advance. In addition to cardiac studies, renal function, renal and coronary perfusion, and/or renal and myocardial injury should be studied in detail. With a focus on technical quality, highly skilled examiners would provide bedside quality control of all collected measures. Experts other than clinical trialists are essential to successful T1 studies. These include, but are not limited to, experts in acute kidney injury, biomarkers, imaging modalities, electrophysiology, and angiography. Finally, a small but robust consortium of experienced centers with ready access to a large HF population will be needed and is currently being created. Standardization Traditionally, both the active and placebo arms in clinical trials have used “standard” therapy as background therapy, which has generally been left to the discretion of individual investigators and study sites. However, standard therapy is not uniform, reflecting geographical norms and patient types.45 To minimize variability, standardizing therapy to the extent possible across T1 centers is recommended. This has already been attempted in a large phase IV study, ASCEND-HF.46 As a

general rule, studies involving novel agents should be placebo controlled.

Dose finding We propose, at minimum, 3 distinct doses plus placebo during T1 studies, with a clear rationale for both how the dosing was chosen and, if not, why such studies were not performed. It is necessary to ensure that doses studied and chosen have sufficient pharmacologic rationale with demonstrable dose response. The “correct” dose, by itself, may not lead to improved outcomes. However, the “wrong” dose may lead to a negative or neutral study—a good drug, good target, but wrong dose scenario. Further dose studies may be needed in phase II.

Other considerations Should we go straight into phase III? Depending on the drug, one may proceed more quickly to phase III, by either skipping phase II altogether or doing adaptive studies, where confirmatory phase II studies are designed to segue right into phase III, which has certain appeal. In such a design, pilot study results may count toward the overall pivotal trial. Well-designed, large clinical trials might better identify the “real-world” effect of novel therapies. However, at the present time, we would advocate against proceeding straight to phase III, given the lack of success to date. A more methodological approach may be necessary until a better pathophysiologic understanding of AHFS is achieved. Such due diligence needs to be carefully considered in light of key business development milestones, such as patent expiration. However, if T1 phase testing is considered early during strategic planning, appropriate timelines and milestones can be created. Patent pressures, regulatory landscape, and industry collaboration As each of the authors will attest, multiple factors are brought into play when outlining the overall strategy and then subsequent tactical plan for clinical development. The time limits of patent protection, regulatory considerations, as well as justifiable concerns by for-profit small biotech and larger pharmaceutical corporations all play a role and are just a fraction of the overall contributors and influences to drug development. Although it would be easiest to “blame” industry pressures circumventing “quality” science for the disappointing results seen to date, this not only would be an oversimplification, but would in fact be wrong. The lack of success to date in AHFS clinical trials, combined with the myriad of opinions and advice from key opinion leaders, has been cited as one reason why some trials never go forward at all. Overall, it is clear that every clinical development program ultimately remains a

American Heart Journal Volume 161, Number 2

hypothesis to be tested and that clarity, teamwork, consensus, and a relentless focus on execution, guided by the best science, are all requirements for successful development, thus, our proposal for a T1 phase of development. Although this may add a step to the overall timeline, this additional time would be offset by the substantial benefits such knowledge would bring, at a fraction of the overall cost of a large phase III trial.

Next steps An upcoming consensus meeting is currently being planned to discuss T1 phase trial design in more detail, which will include respective experts in various domains, such as biomarkers, renal injury, and statistical analysis, given the technical nature of T1 studies. In addition, consensus regarding end points for later stages of development is critical, that is, whether the focus should be on efficacy during the acute phase (inpatient phase), where it appears that standard therapy improves signs and symptoms, or the recovery/subacute phase (early postdischarge period), where there appears to be a vulnerable phase in terms of morbidity/mortality.47

Conclusion A significant unmet need exists for novel therapies for AHFS, given the high postdischarge mortality and rehospitalization rates despite available therapy. Although clinical trials to date have largely disappointed, much has been learned. In hindsight, phase II trials have been inconsistent in methodological rigor. This is likely a major contributor to the lack of success in phase III. A comprehensive, mechanistic understanding of a molecule before pivotal trials is critical. Such knowledge will facilitate identification of the right patient population as well as end points. We propose the T1 or translational phase as a means to achieve this end. This will require a small, but highly motivated, skilled, and experienced network of investigators to perform early translational AHFS research. The T1 phase should be followed by a carefully crafted and well-executed phase II and III trial.

Acknowledgements We would like to thank Norman Stockbridge, MD, PhD, for hosting the meeting held at the FDA in November 2009 and for his thoughtful review of this manuscript.

References 1. Gheorghiade M, Zannad F, Sopko G, et al. Acute heart failure syndromes: current state and framework for future research. Circulation 2005;112:3958-68.

Gheorghiade et al 231

2. Dickstein K, Cohen-Solal A, Filippatos G, et al. ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur Heart J 2008;29: 2388-442. 3. Fang J, Mensah GA, Croft JB, et al. Heart failure-related hospitalization in the U.S., 1979 to 2004. J Am Coll Cardiol 2008; 52:428-34. 4. Hunt SA, Abraham WT, Chin MH, et al. Focused update incorporated into the ACC/AHA 2005 guidelines for the diagnosis and management of heart failure in adults: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in collaboration with the International Society for Heart and Lung Transplantation. J Am Coll Cardiol 2009;53:e1-e90. 5. Lloyd-Jones D, Adams RJ, Brown TM, et al. Heart disease and stroke statistics—2010 update: a report from the American Heart Association. Circulation 2010;121:e46-e215. 6. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009; 360:1418-28. 7. Bueno H, Ross JS, Wang Y, et al. Trends in length of stay and short-term outcomes among Medicare patients hospitalized for heart failure, 1993-2006. JAMA 2010;303:2141-7. 8. Gheorghiade M, Abraham WT, Albert NM, et al. Systolic blood pressure at admission, clinical characteristics, and outcomes in patients hospitalized with acute heart failure. JAMA 2006;296: 2217-26. 9. Fonarow GC, Peterson ED. Heart failure performance measures and outcomes: real or illusory gains. JAMA 2009;302:792-4. 10. Cuffe MS, Califf RM, Adams Jr KF, et al. Short-term intravenous milrinone for acute exacerbation of chronic heart failure: a randomized controlled trial. Journal of the American Medical Association 2002;287:1541-7. 11. Gheorghiade M, Konstam MA, Burnett Jr JC, et al. Short-term clinical effects of tolvaptan, an oral vasopressin antagonist, in patients hospitalized for heart failure: the EVEREST clinical status trials. JAMA 2007;297:1332-43. 12. Konstam MA, Gheorghiade M, Burnett Jr JC, et al. Effects of oral tolvaptan in patients hospitalized for worsening heart failure: the EVEREST outcome trial. JAMA 2007;297:1319-31. 13. McMurray JJV, Teerlink JR, Cotter G, et al. Effects of tezosentan on symptoms and clinical outcomes in patients with acute heart failure: the VERITAS randomized controlled trials. JAMA 2007;298: 2009-19. 14. Mebazaa A, Nieminen MS, Packer M, et al. Levosimendan vs dobutamine for patients with acute decompensated heart failure: the SURVIVE randomized trial. JAMA 2007;297: 1883-1891. 15. Sackner-Bernstein J, Kowalski M, Fox M, et al. Short-term risk of death after treatment with nesiritide for decompensated heart failure. JAMA 2005;293:1900-5. 16. Sackner-Bernstein JD, Kowalski M, Fox M, et al. Short-term risk of death after treatment with nesiritide for decompensated heart failure: a pooled analysis of randomized controlled trials. JAMA 2005;293: 1900-5. 17. Sackner-Bernstein JD, Skopicki HA, Aaronson KD. Risk of worsening renal function with nesiritide in patients with acutely decompensated heart failure. Circulation 2005;111:1487-91.

American Heart Journal February 2011

232 Gheorghiade et al

18. VMAC Investigators. Intravenous nesiritide vs nitroglycerin for treatment of decompensated congestive heart failure: a randomized controlled trial. JAMA 2002;287:1531-40. 19. Gheorghiade M, Gattis WA, O'Connor CM, et al. Effects of tolvaptan, a vasopressin antagonist, in patients hospitalized with worsening heart failure: a randomized controlled trial. JAMA 2004; 291:1963-71. 20. Udelson JE, Orlandi C, Ouyang J, et al. Acute hemodynamic effects of tolvaptan, a vasopressin V2 receptor blocker, in patients with symptomatic heart failure and systolic dysfunction: an international, multicenter, randomized, placebo-controlled trial. J Am Coll Cardiol 2008;52:1540-5. 21. Udelson JE, McGrew FA, Flores E, et al. Multicenter, randomized, double-blind, placebo-controlled study on the effect of oral tolvaptan on left ventricular dilation and function in patients with heart failure and systolic dysfunction. J Am Coll Cardiol 2007;49:2151-9. 22. Coletta AP, Cleland JG. Clinical trials update: highlights of the scientific sessions of the XXIII Congress of the European Society of Cardiology—WARIS II, ESCAMI, PAFAC, RITZ-1 and TIME. Eur J Heart Fail 2001;3:747-50. 23. Torre-Amione G, Young JB, Colucci WS, et al. Hemodynamic and clinical effects of tezosentan, an intravenous dual endothelin receptor antagonist, in patients hospitalized for acute decompensated heart failure. J Am Coll Cardiol 2003;42:140-7. 24. O'Connor CM, Gattis WA, Adams Jr KF, et al. Tezosentan in patients with acute heart failure and acute coronary syndromes: results of the Randomized Intravenous TeZosentan Study (RITZ-4). J Am Coll Cardiol 2003;41:1452-7. 25. Kaluski E, Kobrin I, Zimlichman R, et al. RITZ-5: randomized intravenous TeZosentan (an endothelin-A/B antagonist) for the treatment of pulmonary edema: a prospective, multicenter, doubleblind, placebo-controlled study. J Am Coll Cardiol 2003;41: 204-10. 26. Nieminen MS, Akkila J, Hasenfuss G, et al. Hemodynamic and neurohumoral effects of continuous infusion of levosimendan in patients with congestive heart failure. J Am Coll Cardiol 2000;36: 1903-12. 27. Slawsky MT, Colucci WS, Gottlieb SS, et al. Acute hemodynamic and clinical effects of levosimendan in patients with severe heart failure. Study Investigators. Circulation 2000;102:2222-7. 28. Cleland JG, Freemantle N, Coletta AP, et al. Clinical trials update from the American Heart Association: REPAIR-AMI, ASTAMI, JELIS, MEGA, REVIVE-II, SURVIVE, and PROACTIVE. Eur J Heart Fail 2006; 8:105-10. 29. Kivikko M, Lehtonen L, Colucci WS. Sustained hemodynamic effects of intravenous levosimendan. Circulation 2003;107:81-6. 30. Moiseyev VS, Poder P, Andrejevs N, et al. Safety and efficacy of a novel calcium sensitizer, levosimendan, in patients with left ventricular failure due to an acute myocardial infarction. A randomized, placebo-controlled, double-blind study (RUSSLAN). Eur Heart J 2002;23:1422-32. 31. Mebazaa A, Nieminen MS, Packer M, et al. Levosimendan vs dobutamine for patients with acute decompensated heart failure: the SURVIVE Randomized Trial. JAMA 2007;297:1883-91. 32. Follath F, Cleland JG, Just H, et al. Efficacy and safety of intravenous levosimendan compared with dobutamine in severe low-output heart failure (the LIDO study): a randomised double-blind trial. Lancet 2002;360:196-202. 33. Cleland JG, Ghosh J, Freemantle N, et al. Clinical trials update and cumulative meta-analyses from the American College of Cardiology:

34.

35.

36.

37.

38.

39.

40.

41.

42.

43. 44.

45.

46.

47.

WATCH, SCD-HeFT, DINAMIT, CASINO, INSPIRE, STRATUS-US, RIO-Lipids and cardiac resynchronisation therapy in heart failure. Eur J Heart Fail 2004;6:501-8. Garratt CPM, Colucci W, Fisher L, et al. Development of a comprehensive new endpoint for the evaluation of new treatments for acute decompensated heart failure: results with levosimendan in the REVIVE I study (abstract). Presented at: 24th International Symposium on Intensive Care and Emergency Medicine, Brussels, Belgium; 2004. http://ccforum.com/content/8/S1/P89. Earl GL, Fitzpatrick JT. Levosimendan: a novel inotropic agent for treatment of acute, decompensated heart failure. Ann Pharmacother 2005;39:1888-96. Givertz MM, Massie BM, Fields TK, et al. The effects of KW-3902, an adenosine A1-receptor antagonist, on diuresis and renal function in patients with acute decompensated heart failure and renal impairment or diuretic resistance. J Am Coll Cardiol 2007;50: 1551-60. Dittrich HC, Gupta DK, Hack TC, et al. The effect of KW-3902, an adenosine A1 receptor antagonist, on renal function and renal plasma flow in ambulatory patients with heart failure and renal impairment. J Card Fail 2007;13:609-17. Cotter G, Dittrich HC, Weatherley BD, et al. The PROTECT pilot study: a randomized, placebo-controlled, dose-finding study of the adenosine A1 receptor antagonist rolofylline in patients with acute heart failure and renal impairment. J Card Fail 2008;14: 631-40. Massie BM, O'Connor CM, Metra M, et al. Rolofylline, an adenosine A1-receptor antagonist, in acute heart failure. N Engl J Med 2010; 363:1419-28. Colucci WS, Elkayam U, Horton DP, et al. Intravenous nesiritide, a natriuretic peptide, in the treatment of decompensated congestive heart failure. Nesiritide Study Group. N Engl J Med 2000;343: 246-53. Hernandez AF, O'Connor CM, Starling RC, et al. Rationale and design of the Acute Study of Clinical Effectiveness of Nesiritide in Decompensated Heart Failure Trial (ASCEND-HF). Am Heart J 2009; 157:271-7. Sung NS, Crowley Jr WF, Genel M, et al. Central challenges facing the national clinical research enterprise. JAMA 2003;289: 1278-87. Woolf SH. The meaning of translational research and why it matters. JAMA 2008;299:211-3. Cotter G, Kaluski E, Stangl K, et al. The hemodynamic and neurohormonal effects of low doses of tezosentan (an endothelin A/B receptor antagonist) in patients with acute heart failure. Eur J Heart Fail 2004;6:601-9. Blair JE, Zannad F, Konstam MA, et al. Continental differences in clinical characteristics, management, and outcomes in patients hospitalized with worsening heart failure results from the EVEREST (Efficacy of Vasopressin Antagonism in Heart Failure: Outcome Study with Tolvaptan) program. J Am Coll Cardiol 2008;52: 1640-8. Ezekowitz JA, Hernandez AF, Starling RC, et al. Standardizing care for acute decompensated heart failure in a large megatrial: the approach for the Acute Studies of Clinical Effectiveness of Nesiritide in Subjects with Decompensated Heart Failure (ASCEND-HF). Am Heart J 2009;157:219-28. Gheorghiade M, Bonow RO. Heart failure: early follow-up after hospitalization for heart failure. Nat Rev Cardiol 2010;7: 422-4.

Lessons learned from a pediatric clinical trial: The Pediatric Heart Network Angiotensin-Converting Enzyme Inhibition in Mitral Regurgitation Study Jennifer S. Li, MD, MHS, a ,j Steven D. Colan, MD, b,j Lynn A. Sleeper, ScD, b,j Jane W. Newburger, MD, MPH, c,j Victoria L. Pemberton, RNC, MS, d,j Andrew M. Atz, MD, e,j Meryl S. Cohen, MD, f,j Fraser Golding, MD, g,j Gloria L. Klein, b,j Ronald V. Lacro, MD, c,j Elizabeth Radojewski, RN, g,j Marc E. Richmond, MD, h,j and L. LuAnn Minich, MD i,j Durham, NC; Watertown and Boston, MA; Bethesda, MD; Charleston, SC; Philadelphia, PA; Ontario, Canada; New York, NY; and Salt Lake City, UT

Background Mitral regurgitation is the most common indication for reoperation in children following repair of atrioventricular septal defect (AVSD). We hypothesized that angiotensin-converting enzyme inhibitor therapy would decrease the severity of mitral regurgitation and limit left ventricular volume overload in children following AVSD repair. Methods The Pediatric Heart Network designed a placebo-controlled randomized trial of enalapril in this population. The primary aim was to test the effect of enalapril on the change in left ventricular end-diastolic dimension body surface area– adjusted z score. Before the launch of the trial, a feasibility study was performed to estimate the number of patients with at least moderate mitral regurgitation following AVSD repair. Trial experience Seventeen months after the start of the study, 349 patients were screened, 8 were trial eligible, and only 5 were enrolled. The study was subsequently terminated because of low patient accrual. Several factors led to the problems with patient accrual, including (1) the use of criteria to assess disease severity in the feasibility study that were not identical to those used in the trial, (2) failure to achieve equipoise for the study among clinicians and referring physicians, (3) reliance on methodology developed in adult populations with different disease mechanisms, and (4) absence of adequate data to define the natural history of the disease process under study. Progress in the treatment of children with cardiovascular disease will depend on the future of multicenter collaborative clinical trials. The lessons learned from this study may contribute to improvements in this research. (Am Heart J 2011;161:233-40.) Randomized clinical trials have resulted in remarkable advances in cardiovascular care. During the 1990s, results from more cardiovascular clinical trials were published than in the previous 3 decades combined, ushering in the current era of “evidence-based cardiovascular medicine.”1 Cardiovascular controlled trials have established novel treatments resulting in major improvements in patient outcomes and have also enhanced our understanding of From the aDuke University Medical Center, Durham, NC, bNew England Research Institute, Watertown, MA, cChildren's Hospital, Boston, MA, dNational Heart, Lung, and Blood Institute, NIH, Bethesda, MD, eMedical University of South Carolina, Charleston, SC, f Children's Hospital of Philadelphia, Philadelphia, PA, gHospital for Sick Children, Toronto, Ontario, Canada, hColumbia University, New York, NY, and iUniversity of Utah, Salt Lake City, UT. j For the Pediatric Heart Network Investigators. ClinicalTrials.gov Identifier: NCT00113698. J. Michael DiMaio, MD served as guest editor for this article. Submitted July 21, 2010; accepted October 19, 2010. Reprint requests: Jennifer S. Li, MD, MHS, Division of Cardiology, Department of Pediatrics, Duke University Medical Center, Duke Clinical Research Institute, PO Box 17969, Durham, NC 27715. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.030

heart disease and the impact of risk factors and adverse events. Certain populations, however, have been underrepresented in cardiovascular clinical trials, including women, the elderly, minority populations, and children.1-3 Many barriers impede the design and conduct of randomized clinical trials in children, including the relative rarity of specific diseases, disease heterogeneity, incompletely defined natural history, lack of research infrastructure, ethical issues in pediatric research, and difficulty in identifying valid clinical end points. Systematic controlled studies of medications in children with congenital heart disease have thus been limited, and most medications are not labeled for pediatric use.4 Therefore, treatment decisions in this population are often based on clinical experience, small observational studies, and extrapolation from adult data, rather than clinical trial evidence. In an attempt to narrow the scientific gap in the pediatric cardiovascular population, the National Heart, Lung, and Blood Institute (NHLBI) established the Pediatric Heart Network (PHN) in 2001.5 In 2004, the PHN launched a placebo-controlled randomized trial of the angiotensin-converting enzyme (ACE) inhibitor enalapril for use in infants and children

234 Li et al

with mitral regurgitation following atrioventricular septal defect (AVSD) repair. Our failure to complete this trial has led us to reflect on crucial issues that arose in its design and execution. These issues should be of interest to investigators, clinical practitioners, research networks, participating institutions, sponsors, data and safety monitoring boards (DSMBs), and families of the children whom we serve. Our purpose is to encourage discussion so that standards can be developed to better inform future trials in pediatric medicine.

The PHN ACE Inhibition in Mitral Regurgitation experience Rationale Mitral regurgitation remains the most common indication for reoperation (about 10% of patients) in children following repair of AVSD. Postoperatively, moderate mitral regurgitation occurs in 15% to 30% and severe regurgitation occurs in 5% to 19% of patients.6-10 Residual mitral regurgitation causes volume overload and places a hemodynamic burden on the left ventricle that induces a series of compensatory adjustments, including up-regulation of the renin-angiotensin-aldosterone system.11,12 Although these adjustments may initially restore cardiac output, the disease process is often progressive; and compensatory mechanisms fail. Reports are conflicting from trials of ACE inhibitor therapy for the treatment of mitral regurgitation in both the adult and pediatric populations, and use of ACE inhibitors is not recommended for treatment of mitral regurgitation in the current consensus paper from the American College of Cardiology and the American Heart Association.13-17 Despite this, we have observed an escalating trend toward adoption of this therapy in children, based primarily on data indicating a favorable response to acute ACE inhibitor therapy in children.14,15 We hypothesized that ACE inhibition therapy would decrease the severity of mitral regurgitation and thus limit left ventricular volume overload in children following AVSD repair. We designed the ACE Inhibition in Mitral Regurgitation (ACEi in MR) trial to evaluate the efficacy and safety of the ACE inhibitor enalapril for the treatment of significant mitral regurgitation in children (ClinicalTrials.gov ID: NCT00113698) (Figure 1). Children who had residual mitral (left atrioventricular valve) regurgitation after repair of an AVSD were selected as the study population for 2 reasons: (1) the regurgitant orifice in AVSD responds dynamically to left ventricular size and therefore might respond to medical therapy, and (2) this is a reasonably homogeneous group with a relatively high incidence of at least moderate mitral regurgitation postoperatively. Study design Before the trial, we performed a feasibility study. Each network center undertook a chart review of the

American Heart Journal February 2011

assessment of mitral regurgitation in patients following AVSD repair according to local echocardiography reports from 2003 to 2004. The degree of mitral regurgitation was collected at each center from the reports determined by subjective assessment of the color Doppler width or area (eg, none, trivial, mild, moderate, or severe). The image data were not re-reviewed, and data on left ventricular size were not collected for the purposes of the feasibility study. A total of 1,005 reports covering a 5-year period from July 1, 1997, to June 30, 2002, were reviewed; and 195 patients were identified as having at least moderate mitral regurgitation based on subjective assessment following AVSD repair. Based on these reports, we estimated the incidence of at least moderate mitral regurgitation to be 20% in postoperative AVSD patients followed at PHN centers. The primary aim of the trial (drug phase) was to test the effect of enalapril therapy on the 6-month change in left ventricular end-diastolic dimension (LVEDD) body surface area (BSA)–adjusted z score. Secondary aims included assessment of the effects of enalapril on changes in echocardiographic measures of left ventricular geometry and hemodynamics (ejection fraction, regurgitant fraction, fiber stress, sphericity index, and BSA-adjusted z scores for mass and end-diastolic and end-systolic dimensions and volumes), the change in level of neurohormonal activation evaluated by measurement of B-type natriuretic peptide, and the incidence of adverse effects. Two groups of patients who had undergone repair of an AVSD were targeted for the study. Inclusion and exclusion criteria are shown in Table I. Group 1 patients were to be evaluated at the time of AVSD repair. Eligible patients ≤5 years of age were to be enrolled and observed for 6 months in an observational phase to allow the heart to adapt to the surgical intervention and thus to allow the effects of ACE inhibition to be assessed independently of the acute results of the surgery. Group 2 patients were those who had had their surgery performed either before the study launch or at non-PHN institutions but were referred to a PHN center, so that they could not be enrolled into the observational phase. At ≥6 months after surgery, subjects in both groups with at least moderate mitral regurgitation were to be randomized to receive enalapril or placebo for 6 months. We defined moderate mitral regurgitation on the study echocardiogram used to assess trial eligibility as (1) either a proximal regurgitant jet area ≥6 mm2/m2 or a regurgitant fraction ≥30% (extrapolating from data on adults with mitral regurgitation) and (2) left ventricular volume overload as shown by the presence of an LVEDD BSA-adjusted z score of ≥2.0. 18 An echocardiography core laboratory centrally assessed all study echocardiograms. We assumed an SD of 0.75 for the change in LVEDD BSA-adjusted z score in conjunction with a 2-sided 0.05

American Heart Journal Volume 161, Number 2

Li et al 235

Figure 1

Angiotensin-Converting Enzyme Inhibition in Mitral Regurgitation trial.

type I error rate, 85% power, a minimum clinically significant group difference in mean z score change of 0.5, and one interim look at the data. These assumptions resulted in a total required sample size of 82 subjects. A loss to follow-up rate of 10% was assumed to occur over the 6 months. Thus, the total required sample size was 92 subjects (46 per group) for the drug phase. All participating centers received Institutional Review Board approval. A DSMB appointed by the NHLBI monitored the conduct of the study. Written informed consent was obtained from the parent or guardian of each patient participating in either phase of the study. Assent from the subject was obtained when age appropriate.

Enrollment The observational phase of the ACEi in MR trial began on June 1, 2004; and the start date for randomization in the drug phase was 6 months later, on December 1,

2004. Recruitment in the drug phase was estimated to require 2 years. Two groups of patients were screened for eligibility—those who completed the observational phase (group 1) and those who were already ≥6 months postoperative from their repair (group 2). By April 2005, 257 patients from both groups were screened; but only one patient had been randomized in the drug phase. With the goal of increasing study enrollment, the protocol was amended to broaden the drug phase eligibility criteria by (1) extending the upper age limit for eligibility from 5 to 18 years and (2) lowering the LVEDD BSA z score eligibility criterion from 2.0 to 1.5 to be more inclusive of borderline left ventricular dilation. In addition, community outreach and advertising to referring cardiologists and advocacy groups were implemented. An additional 92 patients were subsequently screened for a total of 349 patients screened by November 2005 (Figure 2).

American Heart Journal February 2011

236 Li et al

Table I. Inclusion and exclusion criteria Observational phase

Inclusion criteria

Exclusion criteria

Drug phase

Inclusion criteria

Exclusion criteria

No longer requires intensive care ≤28 d after repair of an AVSD (including primum atrial septal defect, transitional AVSD, and complete AVSD) Age ≤5 y Complete transthoracic echocardiogram ≤28 d after repair Informed consent of legal guardian Tetralogy of Fallot and total or partial anomalous venous connection Associated mitral stenosis with a mean diastolic gradient of N10 mm Hg Other sources of LV volume overload, (eg, Nmild aortic regurgitation, defined as a regurgitant jet/annulus ratio N20%, or significant left-to right shunt, defined as an echocardiographically determined Qp/Qs ≥1.5 Associated obstructive LV outflow lesions that are more than trivial, defined as an LV outflow peak instantaneous gradient of N20 mm Hg Significant residual coarctation, defined as an arm/leg systolic pressure difference of N20 mm Hg Associated hypertrophic obstructive cardiomyopathy ≥6 m postoperation from repair of an AVSD (primum atrial septal defect, transitional AVSD, and complete AVSD) Age ≤5 y At least moderate mitral regurgitation, defined as both LV end-diastolic dimension BSA-adjusted z score ≥2.0 AND Either a proximal regurgitant jet area ≥6 mm2/m2 or a regurgitant fraction ≥30% Asymptomatic or minimally symptomatic, defined by Ross or NYHA heart failure class I or II Atrioventricular synchrony (paced or intrinsic) Girls of child-bearing potential must either abstain from sexual activity or provide assurance of birth control. Informed consent of legal guardian and assent when appropriate Tetralogy of Fallot, total or partial anomalous venous connection, and obstructed pulmonary venous return Reoperation for mitral regurgitation after the initial repair of an AVSD LV dysfunction, defined as an ejection fraction b55% Treatment with an ACE inhibitor within 6 m of randomization Associated mitral stenosis with a mean diastolic gradient of N10 mm Hg Other sources of LV volume overload, including Nmild aortic regurgitation, defined as a regurgitant jet/annulus ratio N20%, or significant left-to right shunt, defined as an echocardiographically determined Qp/Qs ≥1.5 Systemic hypertension, defined as sustained systolic or diastolic blood pressure ≥95th percentile for age Associated obstructive LV outflow with an LV outflow peak instantaneous gradient of N20 mm Hg Significant residual coarctation with an arm/leg systolic pressure difference of N20 mm Hg Associated hypertrophic obstructive cardiomyopathy History of chronic renal or hepatic dysfunction or persistent neutropenia Severe mitral regurgitation that is expected to need surgery ≤6 m Residual postoperative lesions that are expected to require additional surgery ≤6 m Need for ACE inhibitor therapy expected within 6 m in the opinion of the attending cardiologist History of intolerance to ACE inhibitors Pregnancy or intent to become pregnant

LV, Left ventricular; NYHA, New York Heart Association.

In total, only 8 patients were ultimately trial eligible and 5 were enrolled, all from group 2 (ie, subjects who were ≥6 months postoperative). The PHN ACEi in MR Subcommittee and the PHN Steering Committee reviewed these patient accrual numbers and recommended that the drug phase of this study be halted because the recruitment rate was much lower than expected. The DSMB of the PHN met on November 4, 2005, and concurred with the recommendation that was accepted by NHLBI.

Cost Pediatric Heart Network clinical sites receive an infrastructure budget that supports personnel and administrative costs for all studies. For each study, the sites also receive reimbursement for expenses related to screening, enrollment, and follow-up of subjects. For this study, the sites received $150 per patient for screening; $2,150 per patient for the observational phase; and $10,120 per patient for the drug phase. Based on 196 patients screened (but not enrolled in the observational phase),

American Heart Journal Volume 161, Number 2

Li et al 237

Figure 2 Group 1: ≤28 days postop (N = 84)

Group 2: ≥ 6 mo postop (N = 265)

Screened for Trial (N = 349)

At least mild to moderate MR * (N = 139)

Potentially trial eligible (N = 54)

At least moderate MR** (N = 9)

Residual shunt (N = 1)

ACE inhibitor use (N = 47)

(BY CORE LAB ECHO)

(BY LOCAL ECHO)

Associated complex defect (N = 8)

Less than moderate MR (N = 42)

Trial eligible & no exclusion criteria (N = 8)

Other clinical exclusion (N = 17)

<Mild MR or unknown MR (N = 210)

Not approached (N = 13)

Unknown MR (N = 3)

*Qualitative assessment **Quantitative assessment MR, mitral regurgitation

Randomized (N = 5)

Patient flow diagram.

153 patients in the observational phase, and 5 patients in the drug phase (as of November 4, 2005), a total of $408,950 was paid to the PHN centers for patient-related costs for the entire study; but only $50,600 was spent on patients enrolled in the trial. This represents just 2% of the total patient care budget for the PHN for the study period. The personnel, administrative, and data-related costs for the clinical sites, the Data Coordinating Center, and core laboratory are difficult to estimate because they are spread across multiple studies; but it is unlikely that these represented more than another $100,000 for the drug phase. The total expenditure was able to be limited because of careful tracking and early recognition of inadequate enrollment despite multiple adjustments to the protocol and the agreement of the research team to terminate the drug arm of the study.

Factors leading to underaccrual In this NHLBI-funded randomized trial of ACE inhibition to treat mitral regurgitation, we enrolled only 5% of the intended study sample and spent substantial effort before enrollment was suspended. It was disheartening that, despite screening 349 subjects, only 8 were eligible; and only 5 of the planned 92 participants were randomized. This experience, however, is not atypical: N80% of all clinical trials fall short of their accrual goals.19,20 As a result, trials are stopped, timelines are extended, or the studies accrue inadequate

sample sizes, resulting in CIs that are so wide or power so low that they are uninterpretable. As randomized trials involving patients with congenital heart disease are more commonly undertaken, it is worthwhile to analyze why this trial failed. We believe that several factors likely led to this outcome. 1. The available sample size was estimated from a retrospective chart review at participating centers using criteria that differed from trial entry criteria. Our feasibility study involved chart review of the subjective assessment of mitral regurgitation according to echocardiogram reports from the year before the design phase; the incidence of at least moderate mitral regurgitation was estimated to be 20%. In contrast, our trial entry criteria required quantitative measurements to operationalize the definition of moderate to severe mitral regurgitation instead of the qualitative grading assessment used in the chart review. Some echocardiographic measurements specified in the entry criteria, for example, regurgitant fraction and proximal regurgitant jet area, were not routinely available in the retrospective feasibility study. Other echocardiographic measurements used in the trial entry criteria, such as the degree of left ventricular dilation secondary to mitral regurgitation detectable within 6 months of surgery, were not

238 Li et al

collected in the feasibility study. If we had conducted the feasibility study using the defined trial eligibility criteria, we might have been alerted to the inadequate sample size available for this trial. 2. Clinical equipoise could not be established at the centers. The ethics of clinical research require equipoise—a state of uncertainty on the part of the clinical investigator regarding the comparative therapeutic merits of each arm in a trial.21 We, as PHN investigators, had clinical equipoise with each arm in the drug trial based on conflicting data reported from trials of ACE inhibitor therapy for the treatment of mitral regurgitation in both the adult and pediatric populations.12-16 Equipoise, however, could not be uniformly established among our colleagues at the participating centers. A total of 47 patients screened for the trial with at least mild to moderate mitral regurgitation noted in the chart were already on ACE inhibitors or had ACE inhibitor use planned by their primary cardiologists. These physicians may have felt compelled to use medical therapy because they believed that such therapy might delay the need for a second mitral valve surgery despite the lack of evidence showing benefit of ACE inhibitors in this setting. Others may have used ACE inhibitors because they were not fully informed of the study and used the medicine routinely in their patients after AVSD surgery. This culture of off-label use of drugs in pediatrics is historically based because most drugs before recent years have not been labeled for use in children. Off-label use of cardiovascular medications in children hospitalized with congenital and acquired heart disease remains common. A recent study analyzed data from 31,432 children in the Pediatric Health Information System database hospitalized during 2005 with cardiovascular disease and showed that 78% received ≥1 cardiovascular medication off-label and 31% received ≥3 cardiovascular medications off-label.4 The practice of off-label use without adequate information on drug efficacy and safety places children at risk for adverse events and ineffective therapy. In 9 reports evaluating pediatric drug-related adverse events, off-label prescriptions were involved in 23% to 60%.3 These data highlight the need for further study to determine which treatments are effective and should be used more frequently and which are unsafe or ineffective in children hospitalized with cardiovascular disease. In launching this trial, we did not fully appreciate how daunting it could be to change the culture among pediatric cardiologists from experiential practice to evidence-based medicine. 3. Quantitative eligibility criteria for the trial had not been verified in children. Of 349 patients screened, only 8 met the study definition of at least moderate mitral regurgitation and

American Heart Journal February 2011

left ventricular dilation, and had no exclusion criteria. Many quantitative methods are used to estimate the severity of mitral regurgitation including area of the regurgitant jet, size of the regurgitant orifice, regurgitant fraction, and effective regurgitant orifice area by proximal isovelocity surface area.18,22,23 For inclusion in this trial, we defined at least moderate mitral regurgitation to require either a proximal regurgitant jet area ≥6 mm2/m2 or a regurgitant fraction ≥30% with evidence of left ventricular dilation demonstrated by a LVEDD BSAadjusted z score ≥2 (later 1.5). Our definition was of necessity adapted from published guidelines in adults.18 A number of the techniques recommended in adult guidelines have not been evaluated in children with AVSD. We relied on the presence of left ventricular dilation in conjunction with either a regurgitant jet or a regurgitant fraction consistent with at least moderate regurgitation. Regurgitant jet area rather than jet width was selected because the regurgitant orifice in AVSD is nearly always noncircular. The selection of the specific criteria for the proximal regurgitant jet area was based on adjustment of the published adult criteria for BSA. The selection of these 2 criteria combined a quantitative measure of the severity of mitral regurgitation with evidence of left ventricular enlargement. We included left ventricular dilation as an entry criterion based on the data that left ventricular volume overload resulting from the mitral regurgitation is the primary cause of the clinically important outcomes that therapy is intended to avert.10,11 Of the 349 screened patients with initial local echocardiograms available for analysis (Figure 2), 40% (139/349) were judged to have greater than mild mitral regurgitation and 18% were judged to have moderate to severe mitral regurgitation by the site investigators using qualitative assessment on the local echocardiogram. Of these, however, only 54 patients had no other exclusion criteria (including ACE inhibitor use) based on the local echocardiogram and received a study echocardiogram to assess quantitative criteria (Figure 2). Subsequently, only 8 met the study definition of at least moderate mitral regurgitation (regurgitant fraction ≥30% or proximal regurgitant jet area/BSA ≥6 mm2/m2) with left ventricular dilation (defined as LVEDD z score ≥2 [later 1.5]) in the absence of any newly identified clinical exclusion criteria. For 3 patients, the degree of mitral regurgitation was indeterminate. Analysis of the 42 remaining patients showed that none had a regurgitant fraction N30% or an LVEDD z score ≥2 (later 1.5) and that 81% had a proximal regurgitant jet area/BSA N6 mm2/m2. Thus, most patients were excluded from trial entry because of the absence of left ventricular dilation. We therefore attempted to broaden this requirement by decreasing the LVEDD z score requirement to ≥1.5; but ultimately, this decrease had little effect on enrollment. The lack of left ventricular dilation may be related to the possibility that a 6-month period was not long enough to

American Heart Journal Volume 161, Number 2

observe left ventricular remodeling. Another possibility is that young children following AVSD surgery may have restrictive left ventricular physiology that inhibits dilation with short-term volume overload. Alternatively, we cannot exclude the possibility that the proximal regurgitant jet area value of 6 mm2/m2 overestimated the actual degree of mitral regurgitation because the left ventricle did not dilate. Neither this method of quantitative assessment of mitral regurgitation nor echocardiographic estimation of regurgitant fraction has been validated in children after repair of AVSD. In this study, we presumed that left ventricular dilation would be a mechanistic consequence of significant mitral regurgitation; but in fact, we failed to detect ventricular dilation in a significant number of subjects who met adult quantitative criteria for at least moderate mitral regurgitation. Because the pathophysiologic consequences of mitral regurgitation are mediated through left ventricular volume overload, it would be difficult to justify a clinical trial of medical therapy for mitral regurgitation in children who have no evidence of left ventricular volume overload. Before initiating this trial, there were insufficient natural history data to define the frequency and timing of onset of ventricular dilation as a consequence of mitral regurgitation after repair of AVSD; and this lack of natural history data continues to be a barrier to the design and execution of trials of this kind.

Conclusion Despite our detailed planning and execution of a feasibility analysis, the slow enrollment rate and ultimate failure of this trial were due to several identifiable factors: 1. Use of criteria to assess disease severity in the feasibility study that were not identical to those used in the trial. 2. Failure to achieve equipoise for the study among clinicians and referring physicians. 3. Reliance on methodology developed in adult populations with different disease mechanisms. 4. Absence of adequate data to define the natural history of the disease process under study. The PHN's sustainable infrastructure, which supports multiple simultaneous studies, meant that we did not have to dismantle an entire clinical trial structure and thus were able to minimize the financial losses associated with closing this trial. Given evidence indicating different toxicities and benefits for drugs in children compared with adults,24-26 pediatric clinical trials are necessary to appropriately assess therapeutic agents. When there is limited market value in conducting such trials, the National Institutes of Health becomes one of the few organizations willing to sponsor such randomized trials. Progress in the treatment

Li et al 239

of children with cardiovascular disease will depend on the future of these multicenter collaborative clinical trials. Studies that better define the natural history of pediatric cardiovascular conditions will facilitate the conduct of randomized clinical trials involving these conditions.27-29 We have used the lessons learned from this study to better plan subsequent studies within the PHN. This includes the ongoing randomized clinical trial of atenolol versus losartan in individuals with Marfan syndrome for which the criteria used in the feasibility phase were identical to those being used in the actual trial.30 We hope that the issues discussed and the lessons learned from this study will facilitate improved research for children.

Disclosures Supported by U01 grants from the National Heart, Lung, and Blood Institute (HL068269, HL068270, HL068279, HL068281, HL068285, HL068292, HL068290, HL068288, HL085057).

References 1. Lee PY, Alexander KP, Hammill BG, et al. Representation of elderly persons and women in published randomized trials of acute coronary syndromes. JAMA 2001;286:708-13. 2. Heiat A, Gross CP, Krumholz HM. Representation of the elderly, women, and minorities in heart failure clinical trials. Arch Intern Med 2002;162:1682-8. 3. Shah SS, Hall M, Goodman D, et al. Off-label drug use in hospitalized children. Arch Pediatr Adolesc Med 2007;161:282-90. 4. Pasquali SK, Hall M, Slonim AD, et al. Off-label use of cardiovascular medications in children hospitalized with congenital and acquired heart disease. Circ Cardiovasc Qual Outcomes 2008;1:74-83. 5. Mahony L, Sleeper LA, Anderson PA, et al. The Pediatric Heart Network: a primer for the conduct of multicenter studies in children with congenital and acquired heart disease. Pediatr Cardiol 2006; 27:191-8. 6. Hanley FL, Fenton KN, Jonas RA, et al. Surgical repair of complete atrioventricular canal defects in infancy. Twenty-year trends. J Thorac Cardiovasc Surg 1993;106:387-94 [discussion 394-7]. 7. Moran AM, Daebritz S, Keane JF, et al. Surgical management of mitral regurgitation after repair of endocardial cushion defects: early and midterm results. Circulation 2000;102(19 Suppl 3):III 160-5. 8. Crawford Jr FA, Stroud MR. Surgical repair of complete atrioventricular septal defect. Ann Thorac Surg 2001;72:1621-8 [discussion 1628-9]. 9. Michielon G, Stellin G, Rizzoli G, et al. Repair of complete common atrioventricular canal defects in patients younger than four months of age. Circulation 1997;96(9 Suppl):II-316-22. 10. Ten Harkel AD, Cromme-Dijkhuis AH, Heinerman BC, et al. Development of left atrioventricular valve regurgitation after correction of atrioventricular septal defect. Ann Thorac Surg 2005;79:607-12. 11. Pisacane C, Pacileo G, Santoro G, et al. New insights in the pathophysiology of mitral and aortic regurgitation in pediatric age: role of angiotensin-converting enzyme inhibitor therapy. Ital Heart J 2001;2:100-6. 12. Gaasch WH, Aurigemma GP. Inhibition of the renin-angiotensin system and the left ventricular adaptation to mitral regurgitation. J Am Coll Cardiol 2002;39:1380-3.

240 Li et al

13. Wisenbaugh T, Sinovich V, Dullabh A, et al. Six month pilot study of captopril for mildly symptomatic, severe isolated mitral and isolated aortic regurgitation. J of Heart Valve Dis 1994;3:197-204. 14. Calabro R, Pisacane C, Pacileo G, et al. Hemodynamic effects of a single oral dose of enalapril among children with asymptomatic chronic mitral regurgitation. Am Heart J 1999;138(5 Pt 1): 955-61. 15. Gupta DK, Kapoor A, Garg N, et al. Beneficial effects of nicorandil versus enalapril in chronic rheumatic severe mitral regurgitation: six months follow up echocardiographic study. J Heart Valve Dis 2001; 10:158-65. 16. Seguchi M, Nakazawa M, Momma K. Effect of enalapril on infants and children with congestive heart failure. Cardiol Young 1992;2: 14-9. 17. Bonow RO, Carabello B, de Leon AC, et al. ACC/AHA guidelines for the management of patients with valvular heart disease. Executive summary. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Management of Patients with Valvular Heart Disease). J Heart Valve Dis 1998;7:672-707. 18. Zoghbi WA, Enriquez-Sarano M, Foster E, et al. Recommendations for the evaluation of the severity of native valvular regurgitation with two-dimensional and Doppler echocardiography. J Am Soc Echocardiog 2003;16:777-802. 19. Rendell JM, Licht RW. Under-recruitment of patients for clinical trials: an illustrative example of a failed study. Acta Psychiatr Scand 2007; 115:337-9. 20. Gajewski BJ, Simon SD, Carlson SE. Predicting accrual in clinical trials with Bayesian posterior predictive distributions. Stat Med 2008; 17:2328-40.

American Heart Journal February 2011

21. Freedman B. Equipoise and the ethics of clinical research. N Engl J Med 1987;317:141-5. 22. Alharthi MS, Mookadam F, Tajik AJ. Echocardiographic quantitation of mitral regurgitation. Expert Rev Cardiovasc Ther 2008;6:1151-60. 23. Patel AR, Mochizuki Y, Yao J, et al. Mitral regurgitation: comprehensive assessment by echocardiography. Echocardiography 2000; 17:275-83. 24. Cuzzolin L, Atzei A, Fanos V. Off-label and unlicensed prescribing for newborns and children in different settings: a review of the literature and a consideration about drug safety. Expert Opin Drug Safety 2006;5:703-18. 25. Roberts R, Rodriguez W, Murphy D, et al. Pediatric drug labeling: improving the safety and efficacy of pediatric therapies. JAMA 2003; 290:905-11. 26. Benjamin Jr DK, Smith PB, Murphy MD, et al. Peer-reviewed publication of clinical trials completed for pediatric exclusivity. JAMA 2006;296:1266-73. 27. McCrindle BW, Zak V, Sleeper LA, et al. Laboratory measures of exercise capacity and ventricular characteristics and function are weakly associated with functional health status after Fontan procedure. Circulation 2010;121:34-42. 28. Williams IA, Sleeper LA, Colan SD, et al. Functional state following the Fontan procedure. Cardiol Young 2009;15:1-11. 29. Anderson PA, Sleeper LA, Mahony L, et al. Contemporary outcomes after the Fontan procedure: a Pediatric Heart Network multicenter study. J Am Coll Cardiol 2008;52:85-98. 30. Lacro RV, Dietz HC, Wruck LM, et al. Rationale and design of a randomized clinical trial of beta blocker therapy (atenolol) vs. angiotensin II receptor blocker therapy (losartan) in individuals with Marfan syndrome. Am Heart J 2007;154:624-31.

Curriculum in Cardiology

Atrial fibrillation, anticoagulation, fall risk, and outcomes in elderly patients Matthew B. Sellers, MD, a and L. Kristin Newby, MD, MHS a,b,c Durham, NC

Atrial fibrillation (AF) affects 2.5 million patients in the United States. The incidence of this condition increases with age, such that approximately 5% of people N65 years of age have AF. Because of the lack of organized atrial contraction and thrombus formation in the left atrium, patients with AF are at increased risk of stroke. The estimated risk of stroke among all AF patients is 5% per year. Among patients without mitral stenosis, there is a graded relationship of stroke risk with the number of CHADS2 risk factors. Warfarin is the recommended treatment for embolic stroke prophylaxis in AF in intermediate- to high-risk patients. However, elderly patients who are deemed to be at risk of falls are often not started on warfarin therapy secondary to a perceived higher risk of bleeding complications. These risks have been evaluated, but conclusive data regarding the riskbenefit trade-off are elusive. This review summarizes available data on the use of warfarin in elderly patients with AF, focusing on the risk of bleeding, and will specifically address the utility of falls risk assessment in the decision to initiate warfarin therapy for AF. (Am Heart J 2011;161:241-6.)

Atrial fibrillation (AF) is the most common cardiac arrhythmia and currently affects nearly 2.5 million people in the United States. Approximately 5% of people N65 years of age carry this diagnosis1; and with a growing geriatric population in the United States, this proportion is expected to increase by 2.5-fold over the next 50 years.2 The risk of stroke in the setting of AF increases with age and may be as high as 23.5% in patients 80 to 90 years of age.3 Furthermore, not only is AF associated with an increased risk of stroke; but the Framingham Study also demonstrated that mortality from AF-related strokes is almost double that of strokes unrelated to AF, and functional deficits after AF-related strokes were more likely to be severe.4 Thus, anticoagulation therapy to prevent stroke is of paramount importance in patients at high risk for thromboembolic stroke. However, the bleeding risk associated with warfarin therapy has led physicians to be cautious in using (and, arguably, to underuse) warfarin in older patients, especially those perceived as being at risk for falls.5,6 Some of these fears may stem from studies showing that antithrombotic therapy can double the risk of intracranial hemorrhage (ICH), especially fatal hemorrhagic events.7-9 However, several other studies have called into question From the aDepartment of Medicine, Duke University Medical Center, Durham, NC, b Division of Cardiology, Duke University Medical Center, Durham, NC, and cThe Duke Clinical Research Institute, Durham, NC. George J. Klein, MD served as guest editor for this article. Submitted April 2, 2010; accepted November 2, 2010. Reprint request: L. Kristin Newby, MD, MHS, Duke Clinical Research Institute, DUMC Box 17969, Durham, NC 27715-7969. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.002

the use of “falls risk” as a contraindication to antithrombotic therapy in the setting of AF for high-risk patients10,11; indeed, there is little agreement on how to define or assess “falls risk.” Consequently, elderly patients who have the highest risk of stroke and of worse outcomes with stroke in the setting of AF3,4 may be frequently undertreated despite a lack of evidence to support withholding therapy. This review summarizes some of the robust data behind the use of anticoagulants for stroke prevention in the setting of AF; the data behind adverse events, specifically hemorrhagic events, associated with the use of warfarin and aspirin; and what is presently known about the relationship of falls with bleeding in the context of the general risks and benefits of warfarin therapy in AF. No external sources of funding were used to support this work. The authors are solely responsible for the development of the concept for this review paper, as well as the drafting and editing of the paper and its final contents.

Stroke prevention Multiple clinical trials and subsequent meta-analyses have demonstrated the benefit of aspirin compared with placebo, as well as warfarin compared with placebo, in reducing stroke risk among AF patients (Table I). In a pooled analysis of 3 randomized controlled trials (RCTs), the Atrial Fibrillation Investigators found that the relative risk reduction for aspirin versus placebo was 21% (95% CI 0%-38%, P = .05).12 In addition, another meta-analysis of 6 RCTs of aspirin versus placebo for stroke prevention found that aspirin reduced stroke by 22% (95% CI 2%38%), with absolute risk reductions of 1.5% per year for

American Heart Journal February 2011

242 Sellers and Newby

Table I. Relative risk reduction in meta-analysis of anticoagulation treatment Relative risk reduction No. of trials 3 6 6 5

ASA versus placebo

Warfarin versus placebo

ASA versus warfarin

Reference

21% (0%-38%) 22% (2%-38%) – –

– – 62% (48%-72%) –

– – – 36% (14%-52%)

AF Investigators12 Hart et al13 Hart et al13 Hart et al13

ASA, Acetylsalicylic acid.

primary prevention and 2.5% per year for secondary prevention.13 Even larger reductions have been demonstrated for warfarin relative to placebo. In 6 trials of warfarin versus placebo, adjusted-dose warfarin resulted in a risk reduction of 62% (95% CI 48%-72%). Finally, 5 RCTs showed that adjusted-dose warfarin compared with aspirin yielded a relative risk reduction of 36% (95% CI 14%-52%).13 Patient risk in these warfarin versus aspirin trials was further stratified using the now-validated CHADS2 criteria to determine which patients should be treated with aspirin and which should be treated more aggressively with vitamin K antagonists. Following from these clinical trials, the Seventh American College of Chest Physicians guidelines and the American College of Cardiology/American Heart Association/European Society of Cardiology guidelines for AF recommend either vitamin K antagonist or aspirin therapy for stroke reduction based on the CHADS2 risk stratification scheme.14,15

Hemorrhagic complications The relative benefit of warfarin compared with aspirin in preventing embolic stroke is known12,13; however, warfarin therapy is not without risk. The association of hemorrhagic complications with warfarin use is well established, and elderly patients appear to be at higher risk.9,16 In one analysis, when compared with patients b50 years of age, the unadjusted relative risk of patients ≥80 years of age having a life-threatening or fatal bleed on warfarin was 4.5 (95% CI 1.3-15.6); after adjusting for intensity of anticoagulation, the increased risk remained (relative risk [RR] 4.6, 95% CI 1.2-18.1).16 Among patients treated with warfarin for deep venous thrombosis, age N65 years was an independent risk factor for bleeding (hazard ratio [HR] 1.3, 95% CI 1.0-1.7)9; and among patients ≥85 years of age compared with patients aged 70 to 74 years, the risk of ICH is substantially increased (adjusted odds ratio [OR] 2.5, 95% CI 1.3-4.7).17 Furthermore, among elderly patients who do have a severe hemorrhagic complication (such as an intracerebral hemorrhage), warfarin use appears to be associated with significantly higher mortality. In one prospective cohort, 3-month mortality after ICH among patients

receiving warfarin at the time of an ICH was 52%, compared with 25.8% in patients not taking warfarin. Warfarin use was an independent predictor of death, with an OR of 2.2 (95% CI 1.3-3.8)—roughly a doubling in mortality in a dose-dependent manner.8 In addition to fear of hemorrhagic complications, there may be other reasons physicians choose aspirin over warfarin therapy, such as warfarin's narrow therapeutic window, the atherosclerotic benefits of aspirin therapy, patient preference, and the lack of monitoring and ease of administration associated with aspirin. But as with all interventions, physicians must weigh the risks and benefits of treatment.

Relationship of international normalized ratio with hemorrhage and stroke Warfarin monotherapy Recommendations for anticoagulation therapy in AF patients should consider the balance of stroke risk, bleeding risk, and other complications of warfarin therapy, all of which appear to be at least in part associated with the intensity of anticoagulation. Prothrombin time ratio is a strong predictor of bleeding risk in all age groups.18 In addition, international normalized ratio (INR) may be associated with worse stroke outcomes. For example, in one study of AF patients taking warfarin who presented with stroke, patients with INR N2.0 had an increased risk of a severe stroke and an increased risk of death within 30 days relative to patients with an INR b2.0 (HR 3.4, 95% CI 1.1-10.1).19 It appears that an INR of 2.0 to 3.0 provides the best balance between bleeding risk and stroke prevention benefit. Fang et al17 showed that, compared with an INR b2.0, an INR of 3.5 to 3.9 was associated with an increased risk of ICH (adjusted OR 4.6, 95% CI 2.3-9.4), but an INR of 2.0 to 3.0 was not (adjusted OR 1.3, 95% CI 0.8-2.2). In analyses from the SPORTIF III and SPORTIF V trials, rates of bleeding were higher among patients who had poor INR control compared with those with good control (goal INR 2.0-3.0)20 (Table II). In summary, there is a critical need for appropriate monitoring of INR in patients (particularly elderly patients) taking warfarin to limit hemorrhagic complica-

American Heart Journal Volume 161, Number 2

Sellers and Newby 243

Table II. Hemorrhagic complications in anticoagulation therapy Authors

Anticoagulant

Fihn et al16 White et al9 Fang et al17 Hylek et al19 Fang et al17 White et al20 Mant et al21

Warfarin Warfarin Warfarin Warfarin Warfarin Warfarin ASA + warfarin

Analysis Age (>79 vs <51) Age (>65) Age (>84 vs 70-74) INR (>2 vs <2) INR (3.5-3.9) INR† ASA versus warfarin

Age, y

Complication

>75 versus <51 >65 >84 versus 70-74 50-98 (mean 78.3) >84 versus 70-74 70.9 (mean) >75 (mean 81.5)

RR (hemorrhage) 4.6%⁎ HR (hemorrhage) 1.3% OR (ICH) 2.5% HR (30-d mortality) 3.4% OR (ICH) 4.6% 3.85% versus 1.58% (hemorrhage) RR (hemorrhage) 0.96%

95% CI/P value 1.2%-18.1% 1.0%-1.7% 1.3%-4.7% 1.1%-10.1% 2.3%-9.4% P < .01 0.53%-1.75%

⁎ Adjusted for intensity of anticoagulation. † Poor control (INR 2-3 <60% of time) versus good control (INR 2-3 >75% of time).

tions. Anticoagulation clinics may provide one means of decreasing the rate of hemorrhagic complications from warfarin therapy. In one study, patients who were treated in an anticoagulation clinic were 59% less likely to experience a bleeding complication than patients receiving usual care (HR 0.41, 95% CI 0.24-0.70).22

Aspirin and warfarin treatment The balance of risk and benefit with combined use of aspirin and warfarin in patients with AF is also an important question, as many patients also have an indication for aspirin therapy. In the SPORTIF trials, there was no significant reduction in stroke, systemic embolism, or myocardial infarction with the use of warfarin plus aspirin; but major bleeding occurred significantly more often with the combined use of warfarin and aspirin (3.9% per year) compared with monotherapy with warfarin (2.3% per year, P b .01).23 Aspirin monotherapy Because ICH is a major concern in elderly AF patients treated with warfarin therapy (as well as in others considered to be at high risk for ICH or other bleeding complications), one approach for physicians is to use aspirin monotherapy. As discussed previously, aspirin therapy does indeed decrease the risk of stroke in AF; however, aspirin is not as effective as warfarin.12,13 Thus, electing to use aspirin instead of warfarin assumes that aspirin therapy offers a lower risk of ICH than warfarin therapy, which balances the lower efficacy for stroke prevention. However, among patients N75 years of age in the Birmingham Atrial Fibrillation Treatment of the Aged trial, there was no difference in the rates of ICH between aspirin- and warfarin-treated groups (RR 0.96, 95% CI 0.531.75) with a goal INR of 2.0 to 3.023 (Table II). However, the SPINAF II trial24 (goal INR of 4.5) and the Japanese Nonvalvular Atrial Fibrillation–Embolism Secondary Prevention trial25 both found significantly higher rates of ICH among warfarin-treated patients than among those treated with aspirin. These mixed results do not necessarily support the decision to favor aspirin therapy over warfarin therapy when treating patients with AF who are at high risk for falls or hemorrhagic complications.

Anticoagulation and risk of falls Although increasing age is consistently associated with increased bleeding risk in warfarin therapy, an evaluation that specifically focused on fall-related hemorrhagic events showed that warfarin treatment was not associated with an increased risk of bleeding complications. In this study, the cohort treated with warfarin (379 falls patients) exhibited a hemorrhagic event rate of 6%, compared with 11% among patients (2,256 falls) not treated with warfarin (P = .01).26 However, these results were likely subject to selection bias because patients who are selected for warfarin therapy are less likely to be at risk for falls5,6 and have fewer comorbid conditions, decreasing their risk of complications. In a large retrospective study of 1,245 Medicare patients, approximately 50% of whom were prescribed warfarin, patients at high risk of falls suffered ICH more than twice as often as other subjects.10 The status of high risk for falls was based on documentation in the medical record; therefore, the definition of high risk was not standardized or defined in this retrospective study. Few studies have addressed the relationship of falls or predicted fall risk with bleeding in the setting of anticoagulation for AF. A meta-analysis of antithrombotic therapy in elderly patients at risk for falls concluded that the propensity for falling in elderly patients should not be an important factor when deciding whether or not a patient is a good candidate for anticoagulation for AF.11 In this analysis, the quality-adjusted life expectancy was greatest for warfarin, followed by aspirin, followed by no therapy. This remained true unless the annual stroke risk was b2%. Considering these numbers, an elderly patient taking warfarin would have to fall approximately 300 times per year for the risk of bleeding complications from falling to outweigh the benefits for prevention of embolic stroke. However, the authors were not able to estimate similar rates for subdural and intracerebral hemorrhages because there were too few events. This of itself suggests that the risk of ICH among elderly patients at risk for falls is low. Finally, the stroke rate may have been overestimated, and complications underestimated, in the RCTs in the meta-analysis compared with clinical practice; and patients in the trials may have been monitored more intensely than is usual in clinical practice.11

American Heart Journal February 2011

244 Sellers and Newby

Figure 1

CHADS2 Score 0-1

CHADS2 Score 2-6

0.5

1

2

Hazard Ratio (95% CI)

Hazard ratio for out-of-hospital death, hospitalization for stroke, myocardial infarction or bleeding for warfarin treatment versus no warfarin treatment in atrial fibrillation patients according to CHADS2 risk score. Adapted from Gage et al.10

In another assessment of falls risk and anticoagulation therapy for AF in 19,506 patients, after accounting for baseline factors associated with risk of ICH, neither warfarin nor aspirin treatment was associated with risk of ICH (HR 1.0, 95% CI 0.8-1.4 for warfarin and HR 1.1, 95% CI 0.8-1.4 for aspirin).10 Importantly, in this study, the increased risk of stroke appeared to outweigh the risk of ICH. Among patients at high risk for falls, the HR for stroke was 1.3 (95% CI 1.1-1.6, P = .002) compared with patients who were not at high risk for falls. Furthermore, among patients at high risk for falls, the HR for stroke for each 1point increase in CHADS2 score was 1.42 (95% CI 1.371.47, P b .0001). The HR for the primary composite outcome of out-of-hospital death, hospitalization for stroke, myocardial infarction, and hemorrhage on warfarin compared with no warfarin was 0.98 (95% CI 0.56-1.72, P = .94) for a CHADS2 score of 0 to 1 and 0.75 (95% CI 0.610.91, P = .004) for a CHADS2 score of 2 to 6 (Figure 1). These results support the contention that patients at risk for falls but with concomitant increased stroke risk as manifested by a CHADS2 score of ≥2 would benefit overall from anticoagulation, specifically, warfarin therapy, even in the setting of an increased risk of hemorrhage.10 One underlying central limitation of the body of literature on falls risk in elderly patients is that there is no unifying definition of which patients are at risk of falls. Many of the documented trials rely on physician reporting of falls risk, reports that may be multifactorial and not necessarily based on actual risk of falling. This potentially introduces multiple confounders that may contribute to risk of ICH but are not related to the fall itself.

Future directions The ACTIVE trials explored the role of clopidogrel (an irreversible inhibitor of the platelet P2Y12 receptor) plus aspirin versus aspirin alone in warfarin-intolerant patients

(ACTIVE A) and clopidogrel plus aspirin versus warfarin (ACTIVE W) in AF patients who were able to take warfarin.27 These trials included patients with AF at enrollment or ≥2 episodes of AF in the previous 6 months, in addition to ≥1 of the following risk factors for stroke: age N74 years, hypertension, previous stroke or transient ischemic attack, non–central nervous system embolism, ejection fraction b45%, peripheral vascular disease, or age 55 to 74 years with diabetes or coronary artery disease. Warfarin was found to be superior to clopidogrel + aspirin in ACTIVE W28; but clopidogrel + aspirin reduced risk of stroke and systemic embolism in patients intolerant of warfarin compared with aspirin alone, albeit at the expense of increased major bleeding that was most prominent in patients N65 years of age.29 Thus, warfarin remains the cornerstone of treatment of most moderate- to high-risk patients with AF. However, several novel oral anticoagulants are expected to shift the balance of benefit and risk in anticoagulation for AF. In the RELY trial, the oral direct thrombin inhibitor dabigatran was compared at 2 doses with warfarin to a target INR of 2 to 3 in 18,113 patients with a mean age of 71.5 years.30 In this trial, low-dose dabigatran (110 mg twice daily) was noninferior to warfarin in preventing stroke or systemic embolism and exhibited a better bleeding profile than warfarin, with 2.7% of patients assigned to dabigatran experiencing a major hemorrhage (RR 0.80, 95% CI 0.69-0.93). Highdose dabigatran (150 mg twice daily) was superior to warfarin in preventing stroke or systemic embolism (RR 0.66, 95% CI 0.53-0.82), with similar rates of major bleeding (RR 0.93, 95% CI 0.81-1.07). Importantly, both the high- and low-dose dabigatran groups had significantly lower rates of ICH compared with the warfarin group: 27 (0.23% per year) intracranial bleeds in the lowdose group, 36 (0.30% per year) in the high-dose group, and 87 (0.74% per year) in the warfarin group (RR 0.31, 95% CI 0.20-0.47, low-dose vs warfarin; RR 0.40, 95% CI 0.27-0.60, high-dose vs warfarin). Dabigatran was recently approved by the US Food and Drug Administration for the prevention of stroke and embolism in AF and will be available in 2 doses: a 75-mg, twice-daily dose intended for patients with severe renal dysfunction and a 150-mg, twice-daily dose.31,32 In addition, at the 2010 European Society of Cardiology scientific sessions, results were presented from an RCT examining apixaban (an oral competitive factor Xa antagonist) compared with aspirin in warfarin-intolerant patients.33 The study identified a N50% reduction in thromboembolic complications in apixaban-treated patients with an acceptable bleeding risk, which resulted in early termination of the trial by the data monitoring committee. Full publication of the results is pending. The ongoing ROCKET AF trial is comparing the efficacy and safety of rivaroxaban (another oral factor Xa

American Heart Journal Volume 161, Number 2

inhibitor) versus warfarin in a superiority trial in patients with moderate to high CHADS2 scores34 ; and the ARISTOTLE trial is assessing the efficacy and safety of apixaban compared with warfarin across the spectrum of CHADS2 risk scores in an ongoing noninferiority trial.35 The results of both trials are expected within the year and may provide additional alternatives to warfarin anticoagulation with lower bleeding risk that may be particularly beneficial in elderly patients. Finally, pharmacogenetic guidance of warfarin dosing may improve the safety of warfarin use. Trials completed to date have been small and have not shown benefit on intermediate end points or bleeding,36 but larger RCTs are ongoing.

Conclusions The population of elderly patients with AF presents challenges with regard to the decision to provide anticoagulation treatment as well as which therapy, aspirin or warfarin, to choose. A higher likelihood of drug-drug interactions with warfarin, more adverse effects, and more comorbidities are at play in making these decisions. However, the available data suggest that physicians' decisions are guided more by their concerns over bleeding than an evaluation of the patient's risk for stroke; in many cases, their concerns regarding bleeding appear to be overemphasized in the equation. Overall, warfarin appears to be generally underused in the treatment of elderly AF patients despite fairly clear evidence that it reduces embolic and ischemic events, benefits that outweigh bleeding risk. We conclude that the risk of falling in the elderly population should not be an absolute or relative contraindication to the initiation of warfarin therapy, but that physicians should use their clinical judgment, weighing the evidence for risk and benefit with each case they are presented, including consideration of newer anticoagulants as they become clinically available.

References 1. Feinberg WM, Blackshear JL, Laupacis A, et al. Prevalence, age distribution, and gender of patients with atrial fibrillation. Analysis and implications. Arch Intern Med 1995;155:469-73. 2. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA 2001;285:2370-5. 3. Garwood CL, Corbett TL. Use of anticoagulation in elderly patients with atrial fibrillation who are at risk for falls. Ann Pharmacother 2008;42:523-32. 4. Lin HJ, Wolf PA, Kelly-Hayes M, et al. Stroke severity in atrial fibrillation. The Framingham Study. Stroke 1996;27:1760-4. 5. Dharmarajan TS, Varma S, Akkaladevi S, et al. To anticoagulate or not to anticoagulate? A common dilemma for the provider: physicians' opinion poll based on a case study of an older long-term care facility resident with dementia and atrial fibrillation. J Am Med Dir Assoc 2006;7:23-8.

Sellers and Newby 245

6. Man-Son-Hing M, Laupacis A. Anticoagulant-related bleeding in older persons with atrial fibrillation: physicians' fears often unfounded. Arch Intern Med 2003;163:1580-6. 7. He J, Whelton PK, Vu B, Klag MJ. Aspirin and risk of hemorrhagic stroke: a meta-analysis of randomized controlled trials. JAMA 1998; 280:1930-5. 8. Rosand J, Eckman MH, Knudsen KA, et al. The effect of warfarin and intensity of anticoagulation on outcome of intracerebral hemorrhage. Arch Intern Med 2004;164:880-4. 9. White RH, Beyth RJ, Zhou H, et al. Major bleeding after hospitalization for deep-venous thrombosis. Am J Med 1999;107: 414-24. 10. Gage BF, Birman-Deych E, Kerzner R, et al. Incidence of intracranial hemorrhage in patients with atrial fibrillation who are prone to fall. Am J Med 2005;118:612-7. 11. Man-Son-Hing M, Nichol G, Lau A, et al. Choosing antithrombotic therapy for elderly patients with atrial fibrillation who are at risk for falls. Arch Intern Med 1999;159:677-85. 12. The efficacy of aspirin in patients with atrial fibrillation. Analysis of pooled data from 3 randomized trials. The Atrial Fibrillation Investigators. Arch Intern Med 1997;157:1237-40. 13. Hart RG, Benavente O, McBride R, et al. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med 1999;131:492-501. 14. Fuster V, Ryden LE, Cannom DS, et al. ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of Patients With Atrial Fibrillation): developed in collaboration with the European Heart Rhythm Association and the Heart Rhythm Society. Circulation 2006;114:e257-354. 15. Singer DE, Albers GW, Dalen JE, et al. Antithrombotic therapy in atrial fibrillation: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004;126:429S-56S. 16. Fihn SD, Callahan CM, Martin DC, et al. The risk for and severity of bleeding complications in elderly patients treated with warfarin. The National Consortium of Anticoagulation Clinics. Ann Intern Med 1996;124:970-9. 17. Fang MC, Chang Y, Hylek EM, et al. Advanced age, anticoagulation intensity, and risk for intracranial hemorrhage among patients taking warfarin for atrial fibrillation. Ann Intern Med 2004;141:745-52. 18. Hart RG, Tonarelli SB, Pearce LA. Avoiding central nervous system bleeding during antithrombotic therapy: recent data and ideas. Stroke 2005;36:1588-93. 19. Hylek EM, Go AS, Chang Y, et al. Effect of intensity of oral anticoagulation on stroke severity and mortality in atrial fibrillation. N Engl J Med 2003;349:1019-26. 20. White HD, Gruber M, Feyzi J, et al. Comparison of outcomes among patients randomized to warfarin therapy according to anticoagulant control: results from SPORTIF III and V. Arch Intern Med 2007;167: 239-45. 21. Mant J, Hobbs FD, Fletcher K, et al. Warfarin versus aspirin for stroke prevention in an elderly community population with atrial fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomised controlled trial. Lancet 2007;370: 493-503. 22. Nichol MB, Knight TK, Dow T, et al. Quality of anticoagulation monitoring in nonvalvular atrial fibrillation patients: comparison of anticoagulation clinic versus usual care. Ann Pharmacother 2008;42: 62-70.

246 Sellers and Newby

23. Flaker GC, Gruber M, Connolly SJ, et al. Risks and benefits of combining aspirin with anticoagulant therapy in patients with atrial fibrillation: an exploratory analysis of stroke prevention using an oral thrombin inhibitor in atrial fibrillation (SPORTIF) trials. Am Heart J 2006;152:967-73. 24. SPAF II Investigators. Warfarin versus aspirin for prevention of thromboembolism in atrial fibrillation: Stroke Prevention in Atrial Fibrillation II Study. Lancet 1994;343:687-91. 25. Yamaguchi T. Optimal intensity of warfarin therapy for secondary prevention of stroke in patients with nonvalvular atrial fibrillation : a multicenter, prospective, randomized trial. Japanese Nonvalvular Atrial Fibrillation-Embolism Secondary Prevention Cooperative Study Group. Stroke 2000;31:817-21. 26. Bond AJ, Molnar FJ, Li M, et al. The risk of hemorrhagic complications in hospital in-patients who fall while receiving antithrombotic therapy. Thromb J 2005;3:1. 27. Connolly SJ, Yusuf S, Budaj A, et al. Rationale and design of ACTIVE: the Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events. Am Heart J 2006;151:1187-93. 28. ACTIVE Writing Group of the ACTIVE Investigators. Clopidogrel plus aspirin versus oral anticoagulation for atrial fibrillation in the Atrial fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events (ACTIVE W): a randomised controlled trial. Lancet 2006;367:1903-12. 29. Connolly SJ, Pogue J, Hart RG, et al. Effect of clopidogrel added to aspirin in patients with atrial fibrillation. N Engl J Med 2009;360: 2066-78.

American Heart Journal February 2011

30. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;361: 1139-51. 31. U.S. Food and Drug Administration Web site. FDA news release. FDA approves Pradaxa to prevent stroke in people with atrial fibrillation (October 19, 2010). Available at: http://www.fda.gov/News Events/Newsroom/PressAnnouncements/ucm230241.htm (accessed October 22, 2010). 32. Wood S, O'Riordan M. FDA approves dabigatran for stroke prevention, embolism in AF patients. October 20, 2010. Available at: http://www.theheart.org/article/1138703.do (accessed October 22, 2010). 33. Eikelboom JW, O'Donnell M, Yusuf S, et al. Rationale and design of AVERROES: apixaban versus acetylsalicylic acid to prevent stroke in atrial fibrillation patients who have failed or are unsuitable for vitamin K antagonist treatment. Am Heart J 2010;159:348-53 e1. 34. Rivaroxaban-once daily, oral, direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation: rationale and design of the ROCKET AF study. Am Heart J 2010;159:340-347 e1. 35. Lopes RD, Alexander JH, Al-Khatib SM, et al. Apixaban for reduction in stroke and other ThromboemboLic events in atrial fibrillation (ARISTOTLE) trial: design and rationale. Am Heart J 2010;159: 331-9. 36. Kangelaris KN, Bent S, Nussbaum RL, Garcia DA, Tice JA. Genetic testing before anticoagulation? A systematic review of pharmacogenetic dosing of warfarin. J Gen Intern Med 2009;24:656-64.

Primary percutaneous coronary intervention for acute myocardial infarction: Is it worth the wait? The risk-time relationship and the need to quantify the impact of delay Giuseppe Tarantini, MD, PhD, a Frans Van de Werf, MD, PhD, b Claudio Bilato, MD, PhD, a and Bernard Gersh, MB, ChB, DPhil, FRCP c Padua, Italy; Leuven, Belgium; and Rochester, NY

The efficacy of reperfusion therapy is dependent not only by the duration of symptoms before therapy but also by the baseline risk of the individual and the circumstances (time and context) of the occurrence. All these variables play a crucial role in determining the choice of best therapy (fibrinolysis or primary angioplasty [primary percutaneous coronary intervention, PPCI]), thereby confirming the admonition that one size does not fit all. It is generally accepted that patients are best served by PPCI when times to therapy are equal between PPCI and fibrinolysis, whereas pivotal issues that are less well supported by evidence include whether a single time interval is appropriate with regard to the “acceptable” PPCI-related delay and what degree of transfer-related delay is acceptable in patients presenting “early” to a non–percutaneous coronary intervention (PCI)-capable facility. The aim of this perspective is to use available data to individualize the approach to reperfusion therapy, taking into account temporal delays and the overall mortality risk on a case-by-case basis. (Am Heart J 2011;161:247-53.)

“It is not enough that we do our best; sometimes we have to do what is required” Sir Winston Churchill In the 20 years since the GISSI-1 and ISIS-2 were published, therapy for ST-segment elevation myocardial infarction (STEMI) has continuously evolved.1,2 Several meta-analyses, including randomized trials, have demonstrated that primary percutaneous coronary intervention (PPCI) is superior to in-hospital fibrinolysis for the treatment of patients with STEMI, even among patients admitted to hospitals without interventional facilities, in which interhospital transfer is necessary.3-6 It is generally accepted that PPCI is the preferred reperfusion strategy for all, but particularly for patients with STEMI with large From the aDivision of Cardiology, Department of Cardiac, Thoracic and Vascular Sciences, University of Padua Medical School, Padua, Italy, bUniversity Hospital Gasthuisberg, Leuven, Belgium, and cDivision of Cardiovascular Diseases, Mayo Clinic College of Medicine, Rochester, NY. Submitted September 30, 2010; accepted November 7, 2010. Reprint requests: Giuseppe Tarantini, MD, PhD, Division of Cardiology, Department of Cardiac, Thoracic and Vascular Sciences, Policlinico Universitario, Via Giustiniani, 2, 35128 Padova, Italy. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.003

infarct and low bleeding risk, provided it is performed within 90 minutes of first medical contact (particularly in patients presenting early) by an experienced team.7,8 However, the availability of catheterization laboratories among countries and within countries strongly affects the ability to deliver mechanical reperfusion therapy.9 Thus, according to the guidelines by the American College of Cardiology/American Heart Association7 and the European Society of Cardiology,8 patients with STEMI who cannot undergo PPCI in a timely manner, such as those presenting to a hospital without PCI capability and cannot be transferred to a PCI center, should be treated with fibrinolytic therapy within 30 minutes of hospital presentation as a system goal unless fibrinolytic therapy is contraindicated (class I, level of evidence B). Notwithstanding, controversies still exist in the care of patients with STEMI, especially in the everyday clinical practice. Should a 50-year-old male diabetic patient with a 3-hour anterior STEMI who is hemodynamically stable, is admitted to a hospital without PCI capability, and is 70 minutes away from the tertiary center with PPCI facilities be treated in the same manner as a 74-yearold male patient with a 3-hour anterior STEMI who is hemodynamically unstable (ie, heart rate N100 beat/min) and is admitted to the same hospital? These 2 cases raise a number of important questions and unresolved issues. What is the maximum acceptable PPCI-related

American Heart Journal February 2011

248 Tarantini et al

Figure 1

Table I. Risk scores

TIMI score for STEMI (0-14)

TIMI risk index CADILLAC (0-18)

Independent clinical predictors of 30-day mortality.

delay compared with fibrinolysis that can justify immediate transfer for PPCI as opposed to initial treatment with a fibrinolytic agent? Is interhospital transfer necessary for both patients? Does a specific PPCI-related delay fit both cases? Is there any role for a pharmacoinvasive strategy? The aim of this perspective is to use available data to individualize the approach to reperfusion therapy, taking into account temporal delays and the overall mortality risk.

Risk-dependent benefit of the reperfusion therapy: key modulator of the reperfusion choice? Choices among alternative therapies or decisions regarding the allocation of clinical resources are based on a baseline assessment of the patient's risk. It is well documented,10 and N80% of the prognostic information is based on 3 characteristics, namely, age, hemodynamic status (eg, heart rate, blood pressure, and Killip class), and the location of the infarction (Figure 1), whereas other variables reflecting the circumstances under which the infarction was treated (eg, time to treatment and type of reperfusion therapy) are of lesser but nonetheless modifiable importance.10,11 One of the first validated and clinically useful risk scores for STEMI was the Thrombolysis In Myocardial Infarction (TIMI) score, which was derived from fibrinolytic therapy trials.11 The TIMI score incorporates clinical and electrocardiographic (ECG) characteristics (Table I) and has a robust prognostic performance

PAMI (0-15)

GRACE (0-372)

Risk factor

Score

Age, 65-74/≥75 y Systolic blood pressure, b100 mm Hg Heart rate, N100 beat/min Killip classification II-IV Anterior STEMI or left branch bundle block Diabetes mellitus, hypertension, or angina pectoris Weight, b67 kg Time to treatment, N4 h Heart rate × (age/10)2/systolic blood pressure Baseline left ventricle ejection fraction, b40% Renal insufficiency Killip classification II-IV Final TIMI flow 0-2 Age, N65 y Anemia⁎

2/3 3 2 2 1 1

3-Vessel disease Age, N75 y Age, 65-75 y Killip classification N1 Heart rate, N100 beat/min Diabetes mellitus Anterior STEMI or left branch bundle block Age (y) b30 30-39 40-49 50-59 60-69 70-79 80-89 ≥90 Heart rate (beat/min) b50 50-69 70-89 90-109 110-149 150-199 N200 Systolic blood pressure (mm Hg) b80 80-99 100-119 120-139 140-159 160-199 N200 Creatinine (mg/dL) 0-0.39 0.4-0.79 0.8-1.19 1.2-1.59 1.6-1.99 2-3.99 N4 Killip classification I II III IV

1 1

4 3 3 2 2 2 2 7 3 2 2 2 2 0 8 25 41 58 75 91 100 0 3 9 15 24 38 46 58 53 43 34 24 10 0 1 4 7 10 13 21 28 0 20 39 59

American Heart Journal Volume 161, Number 2

Tarantini et al 249

Table I (continued ) Risk factor Cardiac arrest at admission Increased cardiac markers ST-segment deviation

Score 39 14 28

CADILLAC, Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications; PAMI, Primary Angioplasty in Myocardial Infarction; GRACE, Global Registry for Acute Coronary Events. ⁎ Anemia is defined as baseline hematocrit b39% for men and 36% for women.

across the heterogeneous spectrum of patients with STEMI.12 With the increasing adoption of PPCI as the preferred reperfusion strategy, risk scores based on PPCI trials were subsequently developed (Table I) and validated, although with a notable difference in predictive accuracy.13 The Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications risk score is based on clinical and angiographic parameters, whereas the Primary Angioplasty in Myocardial Infarction score relies on clinical and ECG characteristics (Table I). In addition, the Global Registry for Acute Coronary Events score, based on a large registry of patients across the entire spectrum of acute coronary syndromes, incorporates clinical and ECG characteristics. This risk score was validated and showed to be predictive for all acute coronary syndromes but had lower predictive accuracy for early mortality as compared with other scores.13 None of the models, however, have been tested prospectively by randomizing patients to a reperfusion strategy based on the estimated mortality at presentation. The quantitative analysis by Keeley et al3 that compared PPCI with fibrinolysis showed an absolute decrease of approximately 2% of risk of death by PPCI over in-hospital thrombolysis. However, mortality benefit is related to the risk stratification at admission, being more evident in high-risk patients (low proportion of patients), as demonstrated by a subgroup analysis of the DANAMI-2 trial.14 Indeed, individual trials and risk-benefit metaregression analyses confirm that the absolute difference in mortality at 30 days between PPCI and fibrinolysis increases in favor of PPCI as the estimated risk of mortality with fibrinolysis grows.14-16 Conversely, if the estimated mortality benefit with fibrinolysis declines, the absolute mortality benefit of PPCI decreases, although definitive evidence on the exact equipoise is still lacking. On the other hand, when the estimated mortality with fibrinolysis is extremely high (eg, patients with cardiogenic shock), compelling evidence exists favoring a PPCI strategy, as demonstrated by the SHOCK trial17 and the National Registry of Myocardial Infarction II.18 If PPCI is unavailable, the potential benefits of fibrinolytic therapy need to take into consideration the risk of a lifethreatening bleed.19

Time-dependent benefit of reperfusion therapy: overemphasized or overlooked? Time from the symptom onset The early open-artery theory suggests that benefits of reperfusion in patients with STEMI are directly related to the speed and completeness of the restoration of patency of the infarct-related coronary artery. Several clinical studies strongly confirm the relationship between achieving prompt antegrade coronary flow and improvement of myocardial salvage and clinical outcomes, both for PPCI and fibrinolysis.20,21 As shown in experimental and clinical models, the extent of transmurality and the presence of microvascular injury are strongly dependent on the duration of ischemia before reperfusion, with a close relationship among myocardial and microvascular injury, myocardial viability, left ventricular function, and clinical outcomes.22-25 A clear relationship between the extent of delay and mortality has been recently described26: for each additional 30 minutes of treatment delay, there is a 7.5% increase in mortality. Similarly, it has been recently reported that the delivery of reperfusion therapy for patients with STEMI outside guidelinerecommended delays is associated with adverse outcomes.27 According to the hypothetical construct of the association between mortality and treatment delay described by Gersh et al,28 this demonstrates a striking benefit within the first 2 to 3 hours, emphasizing the narrow “golden window of opportunity,” as reported also by Boersma et al.21 At a later stage, on the “flat part of the curve,” there is a continued but decreasing magnitude of mortality benefit of PPCI over time. The pathophysiologic factors that potentially could modulate the relationship between time and clinical outcomes are the presence of functioning coronary collaterals, ischemic preconditioning, myocardial oxygen demand, myocardial territory at risk, endothelial and microvascular function, and spontaneous coronary reperfusion. In clinical practice, symptom-onset-to-needle time or symptom-onset-to-balloon time is often and inappropriately considered as surrogate of the total ischemic time. However, it should be emphasized that the precise time from the symptom onset may be troublesome to determine by patient's history (presentation delay) because of inaccurate recollection, nonspecific symptoms, silent ischemia, intermittent spontaneous reperfusion, collateral circulation, or prodromal angina. Moreover, the time from the symptom onset cannot be a surrogate of the true ischemic time if the infarctrelated artery is already open at the time of PPCI or at the onset of fibrinolytic therapy.22 Thus, although time from the symptom onset has become central in the diagnosis, triage, and management of patients with STEMI, its prognostic role may be overemphasized because of the substantial differences between the

250 Tarantini et al

pathophysiology of STEMI in a patient and experimental models.29 Accordingly, several efforts have been made to evaluate characteristics of the ECG at the time of presentation to provide additional insight into the stages of the STEMI process and to discriminate between “pseudo early comers” (patients with perceived short time from the symptom onset but with new Q waves on the admission ECG) and “pseudo latecomers” (patients with perceived long time from the symptom onset but without Q waves). Indeed, the presence of Q waves on the admission ECG provides incremental value to the presentation delay in predicting 30-day mortality after either fibrinolysis or PPCI30 and might help to evaluate patients with STEMI for triage and potential transfer to tertiary centers for planned PCI especially if the duration of the symptoms is unclear.

Time from the symptom onset–risk interaction A significant interaction between ischemic time and baseline risk in patients with STEMI treated with PPCI has been reported with an adverse impact on 1-year mortality for any delay in reperfusion, particularly if patients are at high risk and/or preprocedural TIMI 2-3 flow is absent.31 Time-to-treatment analyses, however, may have been confused by other variables. First, it has been recently observed that early presenters have the highest risk score and the largest magnitude of cumulative ST elevation, reflecting a larger area at risk.29,32 Late presenters, on the other hand, may be considered, at least in part, survivors (eg, they did not die out of hospital); therefore, they may be at lower risk compared with early presenters, but this may be confounded by comorbities, for example, renal failure, diabetes, advanced age, etc. Second, with regard to the differential benefit of PPCI over fibrinolysis as a function of symptom duration, the current American College of Cardiology/American Heart Association guidelines state that fibrinolysis is an appropriate alternative in patients presenting with symptom onset duration ≤3 hours depending on the extent of transfer delay. This statement is based on a subgroup analysis of the CAPTIM trial,33 which demonstrated a higher mortality rate in PPCI-treated than in fibrinolysis-treated patients presenting within 2 hours from the symptom onset. On the other hand, Boersma et al,4 using individual data from randomized trials (not including the CAPTIM trial), provided evidence that lower mortality is seen in PPCItreated than in fibrinolysis-treated patients for each specific time-to-treatment period. Therefore, the key question, for which data are unavailable, is whether fibrinolysis at 60 to 70 minutes is better than PPCI at 120 to 140 minutes and in which patients. Similarly, the recent article by Lambert et al27 emphasized clearly the importance of reperfusion within the guideline-recom-

American Heart Journal February 2011

mended delays, but it is not conclusive to answer this question.

Primary percutaneous coronary intervention–related delay The mortality benefit of PPCI over fibrinolysis is also dependent on the additional PPCI-related delay. Current guidelines suggested an acceptable PPCI-related delay of b60/120 minutes based on the currently published trials.7,8 However, other authors found different times to equipoise according to different modeling of the data,34 for example, stratifying the DANAMI-2 study patients according to whether they were initially admitted to an interventional or noninterventional center or the inclusion of trials using non–fibrin-specific fibrinolytic therapy. To this regard, in a meta-analysis, Nallamothu et al35 showed that the survival advantage of PPCI was lost for PPCI delay N1 hour only when fibrin-specific agents were considered. Irrespective of whatever categorization, precise cutoff value for PPCI-related delay is, to some extent, oversimplification from a clinical standpoint. Primary percutaneous coronary intervention–related delay-risk interaction By collecting patient-specific data for most randomized trials and using a PPCI-related delay defined at hospital level, Boersma et al4 showed that PPCI remained superior to fibrinolysis when PPCI-related delay was either b35 minutes or N79 minutes. Because the mortality rates in patients with longer delays were higher, it is likely that an interaction between acceptable PPCI-related delay and severity of baseline mortality risk was present. Other large registry studies have also demonstrated that longer PPCI-related delay does not negate the benefits of PPCI,36 but all these data are subject to the confounding factors and selection bias inherent of registry studies. Recently, Pinto et al,37 using data from the National Registry of Myocardial Infarction registry, reported a time to equivalent benefit in mortality between PPCI and fibrinolysis at 114 minutes. The authors assessed also the impact of PCI-related reperfusion delay across cohorts of patients with various risk factors (anterior vs nonanterior infarction and age b65 or ≥65 years) and further classified them into those presenting before versus ≥2 hours after symptom onset. This analysis demonstrated that in young patients with STEMI, either anterior or nonanterior, within 2 hours after symptom onset with low risk of bleeding, the PPCI-related reperfusion delay leading to equivalent mortality between PPCI and fibrinolysis was b60 minutes. In older patients, this balance swayed toward a longer acceptable delay. This, although nonrandomized, registry further supports the risk in modulating the “acceptable” PCI-related reperfusion delay. The Vienna registry documented a comparable mortality in PPCI- and fibrinolysis-treated patients

American Heart Journal Volume 161, Number 2

Tarantini et al 251

Figure 2

The estimated PPCI delay to equipoise for patient 2 is 200 minutes, according to its time from symptom onset (180 minutes) and risk (TIMI risk score 5 [12.4% mortality rate]) (see Tarantini et al16 for details). Within this time interval, however, loosing time still means loosing benefit. For example, 90 minutes of extra delay from “A” to “B” determines a significant reduction of the PPCI benefit as shown by the increase of the number needed to treat. Finally$2, an expected system delay longer than acceptable should imply the choice of fibrinolysis (shift from “B” to “D”) to avoid harming the patient. NNH, Number needed to harm; NNT, number needed to treat. Modified from Tarantini et al16 with permission from Elsevier.

even for PPCI-related delay of 138 minutes, with a significant advantage in cardiogenic shock or if age was ≥75 years.38 Finally, a recent meta-regression analysis confirmed that the variability in baseline mortality risk significantly modified the acceptable time delay in choosing the appropriate reperfusion strategy and that the time delay that nullifies the survival benefit of PPCI over fibrinolysis can be calculated from the baseline mortality risk and the length of presentation delay.16 Despite the fact that increasing treatment delay, particularly the modifiable health system–related delay, results in higher mortality in high-risk patients,27,29,31,39,40 PPCI may still be the preferred strategy in high-risk subgroups presenting to hospitals without interventional facilities unless the patient presents extremely early (eg, b60 minutes).16,37,40 In this respect, a recent comprehensive registry from the province of Quebec, Canada, identified that being transferred for PPCI was a major predictor of lower 30-day mortality.27 Based on the described multifaceted risk-time relationship, it comes out that there is a clinical need of a customized quantification of the impact of treatment delay (“waiting score”) in order to choose the most appropriate reperfusion therapy on an individual basis. While waiting for more robust data on the field, how do we apply the bulk of the available evidences to the 2 patients described in the introduction? The preferred approach to the first patient would be the prompt initiation of fibrinolytic therapy, whereas in the

case of the second patient, it would be the immediate transfer for PPCI, based on the quantification of the impact of delay, that is, b1 hour and N2 hours, respectively. Figure 2 shows schematically how the survival advantage of PPCI over fibrinolysis may vary as function of the related PPCI delay in our case vignette 2. Primary percutaneous coronary intervention remains the treatment of choice even when longer delays are unavoidable, although the concept that, particularly in high-risk patients, the longer the delay the lower the absolute survival advantage of PPCI, as shown by others,31,40 remains true. In summary, although accepting that the preferred approach for most patients is PPCI, within the confines of guideline-recommended PCI-related delays, the data suggest that, in certain circumstances based on baseline levels of mortality risk, longer delays to choose PPCI instead of fibrinolysis might be acceptable.4,37 It should be noteworthy, however, that there are major shortcomings in the available data to take into account. First, the prognostic impact of the presentation delay may be underestimated because few patients are admitted within the “golden time window of opportunity.”28 Thus, a shorter time to equipoise may be expected in patients presenting very early (when the thrombus is more vulnerable to fibrinolysis) and to a drug-invasive strategy. Second, other confounders may play a role. For example, (1) longer presentation delays are associated with longer door-to-balloon or door-to-needle times,39,41 and (2) high-

252 Tarantini et al

quality and large-volume PCI centers may be associated with shorter delays because of more efficient systems of care.41,42 Third, the calculation of a definite waiting score is difficult and needs to be validated prospectively in the context of large clinical trials. Finally, patients treated with the pharmacoinvasive approach were not included in the reported analyses. The pharmacoinvasive strategy implies primary reperfusion therapy by thrombolysis with routine angiography 3 to 24 hours after apparently successful thrombolysis or as a “rescue” procedure. There is mounting evidence that a routine but nonemergent invasive strategy after fibrinolytic therapy is not only safe and effective but also the preferred approach,43 as recognized by European and US guidelines.7,8 However, few of the currently published studies compare a pharmacoinvasive strategy (with fibrinolysis) to transfer for “plain” PPCI. Therefore, if the maximum acceptable PPCI-related delay depends on the baseline risk of the patient, we need data comparing PPCI with pharmacoinvasive strategy, tailored on the baseline mortality risk. This is currently being studied prospectively in the STREAM trial in which high-risk patients presenting early (b3 hours) to an ambulance crew to a regional hospital with PPCI facilities are randomized to either transfer for PPCI or a pharmacoinvasive approach with tenecteplase, clopidogrel, and enoxaparin.44

Conclusions In patients who present to hospitals with requisite facilities and documented expertise, PPCI is the preferred strategy. Fibrinolysis without delay may provide maximal advantage in younger patients at low risk of hemorrhage and mortality presenting earlier (only if the delay to PPCI is beyond the acceptable time frame for that patient), whereas high-risk patients, except those who present very early without Q waves on the admission ECG, might benefit from PPCI, even when longer delays are unavoidable. Developing systems geared to rapid transfer may be more cost-effective overall than the development of a profusion of low-volume, PPCI-capable facilities.45

References 1. GISSI trial: early results and late follow-up. Gruppo Italiano per la Sperimentazione della streptokinasi nell'Infarto Miocardico. J Am Coll Cardiol 1987;10(5 Suppl B):33B-9B. 2. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17,187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988;2:349-60. 3. Keeley EC, Boura JA, Grines CL. Primary angioplasty versus intravenous thrombolytic therapy for acute myocardial infarction a quantitative review of 23 randomized trials. Lancet 2003;361: 13-20.

American Heart Journal February 2011

4. Boersma E and PCAT-2 Trialists Collaborative Group. Does time matter? A pooled analysis of randomized clinical trials comparing primary percutaneous coronary intervention and in-hospital fibrinolysis in acute myocardial infarction patients. Eur Heart J 2006;27: 779-88. 5. Dalby M, Bouzamondo A, Lechat P, et al. Transfer for primary angioplasty versus immediate thrombolysis in acute myocardial infarction: a meta-analysis. Circulation 2003;108:1809-14. 6. Andersen HR, Nielsen TT, Rasmussen K, et al, for the DANAMI-2 Investigators. A comparison of coronary angioplasty with fibrinolytic therapy in acute myocardial infarction. N Engl J Med 2003;349: 733-42. 7. Kushner FG, Hand M, Smith Jr SC, et al. 2009 Focused updates: ACC/AHA guidelines for the management of patients with STelevation myocardial infarction. A report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2009;120:2271-306. 8. Van de Werf F, Bax J, Betriu A, et al. Management of acute myocardial infarction in patients presenting with persistent STsegment elevation: the Task Force on the Management of ST-Segment Elevation Acute Myocardial Infarction of the European Society of Cardiology. Eur Heart J 2008;29:2909-45. 9. Nallamothu BK, Krumholz HM, Ko DT, et al. Development of systems of care for ST-elevation myocardial infarction patients: gaps, barriers, and implications. Circulation 2007;116:e68-e72. 10. Lee KL, Woodlief LH, Topol EJ, et al. Predictors of 30-day mortality in the era of reperfusion for acute myocardial infarction: results from an international trial of 41 021 patients. Circulation 1995;91:1659-68. 11. Morrow DA, Antman EM, Charlesworth A, et al. TIMI risk score for ST-elevation myocardial infarction: a convenient, bedside, clinical score for risk assessment at presentation: an Intravenous nPA for Treatment of Infarcting Myocardium Early II trial substudy. Circulation 2000;102:2031-7. 12. Morrow DA, Antman EM, Parsons L, et al. Application of the TIMI risk score for ST-elevation MI in the National Registry of Myocardial Infarction 3. JAMA 2001;286:1356-9. 13. Lev EI, Kornowski R, Vaknin-Assa H, et al. Comparison of the predictive value of four different risk scores for outcomes of patients with ST-elevation acute myocardial infarction undergoing primary percutaneous coronary intervention. Am J Cardiol 2008; 102:6-11. 14. Thune JJ, Hoefsten DE, Lindholm MG, et al, for the Danish Multicenter Randomized Study on Fibrinolytic Therapy Versus Acute Coronary Angioplasty in Acute Myocardial Infarction (DANAMI)-2 Investigators. Simple risk stratification at the admission to identify patients with reduced mortality from primary angioplasty. Circulation 2005;112: 2017-21. 15. Tarantini G, Razzolini R, Ramondo A, et al. Explanation for the survival benefit of primary angioplasty over thrombolytic therapy in patients with ST-elevation acute myocardial infarction. Am J Cardiol 2005;96:1503-5. 16. Tarantini G, Razzolini R, Napodano M, et al. Acceptable reperfusion delay to prefer primary angioplasty over fibrin-specific thrombolytic therapy is affected (mainly) by the patient's mortality risk: 1 h does not fit all. Eur Heart J 2010;31:676-83. 17. Hochman JS, Sleeper LA, White HD, et al, for the Should We Emergently Revascularization Occluded Coronaries for Cardiogenic Shock (SHOCK) Investigators. One-year survival following early revascularization for cardiogenic shock. JAMA 2001;285:190-2. 18. Wu AH, Parsons L, Every NR, et al, for the Second National Registry of Myocardial Infarction. Hospital outcomes in patients presenting with congestive heart failure complicating acute

American Heart Journal Volume 161, Number 2

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

myocardial infarction: a report from the Second National Registry of Myocardial Infarction (NRMI-2). J Am Coll Cardiol 2002;40: 1389-94. Krumholz HM, Pasternak RC, Weinstein MC, et al. Cost effectiveness of thrombolytic therapy with streptokinase in elderly patients with suspected acute myocardial infarction. N Engl J Med 1992;327: 7-13. Cannon CP, Gibson CM, Lambrew CT, et al. Relationship of symptom-onset-to-balloon time and door-to-balloon time with mortality in patients undergoing angioplasty for acute myocardial infarction. JAMA 2000;283:2941-7. Boersma E, Maas AC, Deckers JW, et al. Early thrombolytic treatment in acute myocardial infarction: reappraisal of the golden hour. Lancet 1996;348:771-5. Tarantini G, Cacciavillani L, Corbetti F, et al. Duration of ischemia is a major determinant of transmurality and severe microvascular obstruction after primary angioplasty: a study performed with contrastenhanced magnetic resonance. J Am Coll Cardiol 2005;46:1229-35. Tarantini G, Razzolini R, Cacciavillani L, et al. Influence of transmurality, infarct size, and severe microvascular obstruction on left ventricular remodeling and function after primary coronary angioplasty. Am J Cardiol 2006;98:1033-40. Basso C, Corbetti F, Silva C, et al. Morphologic validation of reperfused hemorrhagic myocardial infarction by cardiovascular magnetic resonance. Am J Cardiol 2007;100:1322-7. Francone M, Bucciarelli-Ducci C, Carbone I, et al. Impact of primary coronary angioplasty delay on myocardial salvage, infarct size, and microvascular damage in patients with ST-segment elevation myocardial infarction: insight from cardiovascular magnetic resonance. J Am Coll Cardiol 2009;54:2145-53. De Luca G, Suryapranata H, Ottervanger JP, et al. Time delay to treatment and mortality in primary angioplasty for acute myocardial infarction. Every Minute of Delay Counts Circulation 2004;109: 1223-5. Lambert L, Brown K, Segal E, et al. Association between timeliness of reperfusion therapy and clinical outcomes in ST-elevation myocardial infarction. JAMA 2010;303:2148-55. Gersh BJ, Stone GW, White HD, et al. Pharmacological facilitation or primary percutaneous coronary intervention for acute myocardial infarction. Is the slope of the curve the shape of the future? JAMA 2005;293:979-86. Terkelsen CJ, Sørensen JT, Maeng M, et al. System delay and mortality among patients with STEMI treated with primary percutaneous coronary intervention. JAMA 2010;304:763-71. Armstrong PW, Fu Y, Westerhout CM, et al. Baseline Q-wave surpasses time from symptom onset as a prognostic marker in STsegment elevation myocardial infarction patients treated with primary percutaneous coronary intervention. J Am Coll Cardiol 2009;53: 1503-9. De Luca G, Suryapranata H, Zijlstra F, et al. Symptom onset to balloon time and mortality in patients with acute myocardial infarction treated by primary angioplasty. J Am Coll Cardiol 2003; 42:991-7. Zijlstra F, Patel A, Jones M, et al. Clinical characteristics and outing come of patients with early (less than 2 h), intermediate (2-4 h) and

Tarantini et al 253

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

late (greater than 4 h) presentation treated by primary coronary angioplasty or thrombolytic therapy for acute myocardial infarction. Eur Heart J 2002;23:550-7. Bonnefoy E, Lapostolle F, Leizorovicz A, et al. Primary angioplasty versus prehospital fibrinolysis in acute myocardial infarction: a randomised study. Lancet 2002;360:825-9. Betriu A, Masotti M. Comparison of mortality rates in acute myocardial infarction treated by percutaneous coronary intervention versus fibrinolysis. Am J Cardiol 2005;95:100-1. Nallamothu BK, Antman EM, Bates ER. Primary percutaneous coronary intervention versus fibrinolytic in acute myocardial infarction does the choice of fibrinolytic agent impact on the importance of time to treatment? Am J Cardiol 2004;94:772-4. Magid DJ, Wang Y, Herrin Y, et al. Relationship between time of day, day of week, timeliness of reperfusion, and in-hospital mortality for patients with acute ST-segment elevation myocardial infarction. JAMA 2005;294:803-12. Pinto DS, Kirtane AJ, Nallamothu BK, et al. Hospital delays in reperfusion for ST-elevation myocardial infarction. Implications when selecting a reperfusion strategy. Circulation 2006;114:2019-25. Kalla K, Christ G, Karnik R, et al, for the Vienna STEMI Registry Group. Implementation of guidelines improves the standard of care: the Viennese registry on reperfusion strategies in ST-elevation myocardial infarction (Vienna STEMI Registry). Circulation 2006; 113:2398-405. Brodie BR, Hansen C, Stuckey TD, et al. Door-to-balloon time with primary percutaneous coronary intervention for acute myocardial infarction impacts late cardiac mortality in high-risk patients and patients presenting early after the onset of symptoms. J Am Coll Cardiol 2006;47:289-95. Brodie BR, Gersh BJ, Stuckey T, et al. When is door-to-balloon time critical? Analysis from the HORIZONS-AMI (Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction) and CADILLAC (Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications) trials. J Am Coll Cardiol 2010;56:407-13. Ting HH, Bradley EH, Wang Y, et al. Delay in presentation and reperfusion therapy in ST-elevation myocardial infarction. Am J Med 2008;121:316-23. Magid DJ, Calonge BN, Rumsfeld JS, et al, for the National Registry of Myocardial Infarction 2 and 3 Investigators. Relation between hospital primary angioplasty volume and mortality for patients with acute MI treated with primary angioplasty vs thrombolytic therapy. JAMA 2000;284:3131-8. Wijeysundera HC, You JJ, Nallamothu BK, et al. An early invasive strategy versus ischemia-guided management after fibrinolytic therapy for ST-segment elevation myocardial infarction: a metaanalysis of contemporary randomized controlled trials. Am Heart J 2008;156:564-72. Armstrong PW, Gershlick A, Goldstein P, et al. The Strategic Reperfusion Early After Myocardial Infarction (STREAM) study. Am Heart J 2010;160:30-5. Concannon T, Kent D, Normand SL, et al. Comparative effectiveness of STEMI regionalization strategies. Circ Cardiovasc Qual Outcomes 2010;3:506-13.

Trial Design

Design and rationale of the RadIal Vs. femorAL access for coronary intervention (RIVAL) trial: A randomized comparison of radial versus femoral access for coronary angiography or intervention in patients with acute coronary syndromes Sanjit S. Jolly, MD, MSc, a,l Kari Niemelä, MD, PhD, b,l Denis Xavier, MD, c,l Petr Widimsky, MD, d,l Andrzej Budaj, MD, PhD, e,l Vicent Valentin, MD, f,l Basil S. Lewis, MD, g,l Alvaro Avezum, MD, PhD, h,l Philippe Gabriel Steg, MD, i,l Sunil V. Rao, MD, j,l John Cairns, MD, k,l Susan Chrolavicius, BScN, a,l Salim Yusuf, MBBS, D.Phil, a,l and Shamir R. Mehta, MD, MSc a,l Ontario and Vancouver, Canada; Tampere, Finland; Bangalore, India; Prague, Czech Republic; Warsaw, Poland; Valencia, Spain; Haifa, Israel; Sao Paulo, Brazil; Paris, France; and Durham, NC

Background Major bleeding in acute coronary syndromes (ACS) is associated with an increased risk of subsequent mortality and recurrent ischemic events. Observational data and small randomized trials suggest that radial instead of femoral access for coronary angiography/intervention results in fewer bleeding complications, with preserved and possibly improved efficacy. Radial access versus femoral access has yet to be formally evaluated in a randomized trial adequately powered for the comparison of clinically important outcomes. Objectives The aim of this study is to evaluate the efficacy and safety of radial versus femoral access for coronary angiography/intervention in patients with ACS managed with an invasive strategy. Design This was a multicenter international randomized trial with blinded assessment of outcomes. 7021 patients with ACS (with or without ST elevation) have been randomized to either radial or femoral access for coronary angiography/ intervention. The primary outcome is the composite of death, myocardial infarction, stroke, or non–coronary artery bypass graft-related major bleeding up to day 30. The key secondary outcomes are (1) death, myocardial infarction, or stroke up to day 30 and (2) non–coronary artery bypass graft-related major bleeding up to day 30. Percutaneous coronary intervention (PCI) success rates will also be compared between the two access sites. Conclusions The RIVAL trial will help define the optimal access site for coronary angiography/intervention in patients with ACS. (Am Heart J 2011;161:254-260.e4.)

From the aMcMaster University and the Population Health Research Institute, Hamilton Health Sciences, Hamilton, Ontario, Canada, bTampere University Hospital, Tampere, Finland, cSt John's Medical College and Research Institute, Bangalore, India, dCharles University, Hospital Kralovske Vinohrady, Prague, Czech Republic, ePostgraduate Medical School, Department of Cardiology, Grochowski Hospital, Warsaw, Poland, fHospital Universitari Dr Peset, Valencia, Spain, gLady Davis Carmel Medical Center, Haifa, Israel, h Dante Pazzanese Institute of Cardiology, Sao Paulo, Brazil, iINSERM U-698. “Recherche Clinique en Athérothrombose.” Université Paris 7 and Assistance Publique—Hôpitaux de Paris, Paris, France, jDuke Clinical Research Institute, Duke University, Durham, North Carolina, and kUniversity of British Columbia, Vancouver, Canada. l On behalf of the RIVAL steering committee. See online Appendix B for complete listing. Reg. number NCT01014273. Marc Cohen, MD, served as guest editor for this article. Submitted October 20, 2010; accepted November 23, 2010. Reprint requests: Sanjit S. Jolly, MD, MSc, Rm C3-118, DBCVSRI Building, Hamilton General Hospital, 237 Barton St. East, Hamilton, Ontario, Canada L8L 2X2. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.021

Among patients with acute coronary syndromes (ACS) (ie, ST elevation myocardial infarction [MI] [STEMI], non– ST-segment elevation MI [NSTEMI], or unstable angina), 2% to 5% experience major bleeding,1-3 a substantial proportion of which originates from the vascular access site.4 In multiple observational studies (in both ACS and percutaneous coronary intervention [PCI]), major bleeding has been independently associated with a 2- to 10-fold increased risk of death.4-7 Vascular access for coronary angiography/intervention via the radial artery, a superficial and readily compressible site, could result in a lower risk of bleeding than that associated with access via the femoral artery site. Observational studies have suggested that radial access may be independently associated with a 50% to 60% reduction in the odds of major bleeding compared to

American Heart Journal Volume 161, Number 2

femoral access.8,9 A meta-analysis of 18 small randomized trials (n = 4,458 patients, 61 major bleeding events) showed that radial access was associated with a 73% reduction in major bleeding compared to femoral access (0.5% vs 2.3%, respectively, odds ratio [OR] 0.27, 95% CI 0.16-0.45, P b .001).10 Currently, radial access accounts for only 6% to 12% of procedures worldwide.9,11-14 Due to technical challenges related to radial artery diameter, subclavian tortuosity, and reduced guide catheter support, there are perceptions that the radial approach may be associated with a greater PCI procedural failure rate. In addition, the inability to use large bore hemodynamic support devices such as intra-aortic balloon pumps via the radial artery may lead to concerns among operators that they may not be able to respond adequately to emergencies that arise during the procedure. On the other hand, femoral access has a long history of use, allows for larger diameter catheters for complex procedures, and, during emergency situations, has the advantage of ease of access to the femoral vein for transvenous pacing and of the availability of the femoral artery route for the insertion of an intraaortic balloon pump. Femoral access allows superior guide support and therefore may be associated with greater rates of PCI success. Although femoral artery vascular closure devices allow earlier sheath removal, they have not been shown to reduce major bleeding in randomized trials.15 There is an emerging hypothesis that major bleeding may cause recurrent ischemic events because of (1) activation of the coagulation cascade, (2) adverse effects of blood transfusion, (3) cessation of antithrombotic and antiplatelet therapies or reversal of their effects, and (4) decreased ischemic threshold due to anemia or hypovolemia. Observational studies have demonstrated an association between major bleeding and subsequent ischemic events, but causality remains uncertain.4-6,16 Several randomized trials (OASIS 5, HORIZONS) have demonstrated that a reduction in early bleeding events using safer antithrombotic therapies is strongly associated with a reduction in longer-term mortality, MI, and stroke; this relationship is likely to be causal.2,17 Whether procedural methods to reduce bleeding such as radial access could reduce ischemic events and death is an unanswered question; such a relationship would help support the causal relationship between major bleeding and mortality. Although observational studies have suggested that radial access is associated with a lower risk of mortality than femoral access, potential selection bias necessitates a cautious interpretation of the findings.8,11 The meta-analysis of small randomized trials performed by our group found that radial access was associated with a trend toward less death, MI, or stroke compared to femoral access (OR 0.71, 95% CI 0.49-1.01, P = .058).10

Jolly et al 255

The results from observational studies and meta-analyses of small randomized trials require confirmation in a large adequately powered randomized trial. Accordingly, we designed RIVAL, a multicenter randomized trial to compare the 2 procedures among 7,000 patients with ACS.

The RIVAL study Primary objective Among patients with ACS (with or without ST-segment elevation), to determine whether radial access is superior to femoral access for the composite of death, MI, stroke or non–coronary artery bypass graft (CABG)-related major bleeding up to 30 days.

Design RIVAL is a multinational, multicenter, randomized, parallel group study comparing radial versus femoral access for coronary angiography/intervention among patients with non–ST-segment elevation ACS (NSTEACS) or STEMI (Figure 1). The trial began as an investigator initiated randomized substudy of the CURRENT-OASIS 718 trial, which compared 2 regimens of clopidogrel (double dose vs standard dose) and 2 regimens of aspirin (high vs low dose) among patients with NSTE-ACS or STEMI.19,20 The main CURRENT-OASIS 7 study was completed in July 2009,19 and the RIVAL study has continued as a stand-alone study.

Eligibility criteria Patients with NSTE-ACS or STEMI who are to be managed with an invasive approach are eligible i) if the interventional cardiologist is willing to proceed with either a radial or femoral approach; ii) if an operator is available, with the requisite expertise for both the radial (≥50 such procedures in the past year) and the femoral approach; and iii) if the patient has a normal Allen test (ie, confirmation of collateral flow to the hand) (Table I).

Expertise We mandated that each operator in the RIVAL study had performed ≥50 radial procedures within the previous year. An alternative approach would have been to use “expertise-based” randomization. The advantage of this approach is that the procedure is performed by experts who are completely familiar with the technique. However, there are also disadvantages to expertise-based randomization. First, linking the randomly allocated treatment to specific operators may introduce confounding, and this could threaten the very essence of randomization (which endeavors to balance known and unknown variables). For example, radial experts may have intrinsically better catheterization skills and use lower heparin doses, which may lead to a

American Heart Journal February 2011

256 Jolly et al

Figure 1

RIVAL study design: a randomized trial of radial versus femoral access in patient with NSTE-ACS and STEMI.

difference in outcomes not because of the randomized treatment. Second, expertise-based randomization may limit the external validity of the trial because only a select few, highly specialized individuals would be eligible to participate. Finally, in our meta-analysis,10 the benefits of radial compared to femoral access were seen in both trials of radial experts (operators' preferred route was radial) and nonradial experts. The cut point of a requirement of 50 procedures in the previous year is based on published data on the learning curve of previously exclusive femoral operators to proficiently use radial access with high success rates and within an acceptable procedure time.21,22 Specifically, in a study of 4 experienced femoral operators during their first 415 radial procedures, the learning curve plateaued at 50 radial procedures per operator in procedural success, fluoroscopy, and procedural time, and these results have been replicated in other studies of the radial learning curve in experienced femoral operators.21,22 In a recent randomized trial of 1,024 patients, experienced femoral operators who had performed ≥50 radial procedures demonstrated high radial procedural success rates of 96.5% (3.5% crossover to femoral).23 Annual operator volume for both radial and femoral diagnostic and PCI procedures is being recorded on the

case report forms. Based on interim data from the first 6,925 patients randomized, the procedural volumes of the operators in the RIVAL trial were high (median 300 PCI/y, first quartile (Q1) 186 and third quartile (Q3) 400 PCI/y), and they had experience with both procedures (median proportion annual PCI radial 40%, Q1 25% and Q3 70%).

Randomization Patients are randomized in equal proportions to the 2 groups using a 24-hour computer central automated voice response system located at the coordinating center in Hamilton, Canada.

Interventions 1. Radial access to perform coronary angiography and PCI (if clinically indicated); or 2. Femoral access to perform coronary angiography and PCI (if clinically indicated). The use of an arterial vascular closure device is allowed at the discretion of the treating physician.

American Heart Journal Volume 161, Number 2

Jolly et al 257

Table I. Eligibility criteria Inclusion criteria ACS patients Patients with UA or NSTEMI

Ischemic symptoms suspected to represent a non–ST-segment elevation ACS (UA or NSTEMI) were defined as clinical history consistent with new onset or a worsening pattern of characteristic ischemic chest pain occurring at rest or with minimal exertion (lasting longer than 10 min) and at least one of the following: (1) ECG changes compatible with new ischemia (ST depression of at least 1 mm or transient ST elevation or ST elevation of ≤1 mm or T wave inversion N3 mm in at least 2 contiguous leads); (2) patients N60 y with normal ECG are eligible provided that there is a high degree of certainty that presenting symptoms are because of myocardial ischemia. Such patients must have documented evidence of previous CAD with at least one of the following: (a) prior MI requiring hospitalization, (b) prior revascularization procedure (N3 m ago), (c) cardiac catheterization showing significant CAD, (d) positive exercise test, and (e) other objective evidence of atherosclerotic vascular disease; or (3) already elevated cardiac biomarkers (CK-MB or troponin T or I) above the upper limit of normal. Patients with STEMI (1) Presenting with signs or symptoms of acute MI lasting ≤20 min and (2) definite ECG changes compatible with STEMIpersistent ST elevation (≥2 mm in 2 contiguous precordial leads or N1 mm in ≤2 limb leads) or new left bundle-branch block or Q wave in 2 contiguous leads. Intent to perform same-sitting coronary angiography and PCI during index hospitalization Written informed consent Suitable candidate for either radial or femoral artery PCI, including (a) palpable radial artery with documented normal Allen test, (b) previous experience of the operator with ≥50 cases within the past year of radial artery access for coronary angiography/intervention, and (c) acceptance by operator to use whichever route is assigned by randomization Exclusion Criteria Age b18 y Active bleeding or significant increased risk of bleeding (severe hepatic insufficiency, current peptic ulceration, proliferative diabetic retinopathy) Uncontrolled hypertension Cardiogenic shock Prior CABG surgery with use of N1 internal mammary artery Documented severe peripheral vascular disease precluding a femoral approach Previously entered in the study Investigational treatment (drug or device) within the previous 30 d Medical, geographic, or social factors making study participation impractical or inability to provide written informed consent and to understand the full meaning of the informed consent UA, Unstable angina; CAD, coronary artery disease.

Procedures Patients are screened before undergoing coronary angiography with the permission of the treating interventional cardiologist and then randomized after informed consent. In patients undergoing PCI, troponin, creatine kinase (CK), and CK-MB must be drawn immediately pre PCI and at 2, 6, and 12 hours post PCI. For patients requiring CABG, blood must be drawn for CK and CK-MB immediately pre CABG and at 6 and 12 hours post CABG. All patients must have electrocardiogram (ECG) pre CABG, immediately post CABG, and at the time of discharge.

Outcomes

Other outcomes 1. Death within 30 days; 2. Components of primary outcome at 48 hours and at 30 days; 3. PCI procedural success; 4. Major vascular access site complications at 48 hours and 30 days after the procedure (major vascular access site complications include pseudoaneurysms requiring ultrasound compression, thrombin injection, or surgical repair; and large hematomas requiring prolonged hospitalization, arteriovenous fistulae, limb ischemia, or damage to adjacent nerve).

Primary outcome The primary efficacy outcome is the occurrence of death, MI, stroke or non–CABG-related major bleeding within 30 days.

Study outcome definitions are listed in Table II. Mortality is the first and most important other outcome and has the potential to confirm the link between mortality and major bleeding.

Key secondary outcomes The 2 key secondary outcomes are:

Central events adjudication

1. Death, MI, or stroke within 30 days; and 2. Non–CABG-related major bleeding within 30 days.

A committee of clinicians blinded to treatment allocation will adjudicate all primary efficacy outcomes and bleeding events.

American Heart Journal February 2011

258 Jolly et al

Table II. Outcome definitions First occurrence until day 30 of any component of death, MI, stroke, non–CABG-related major bleeding Term Non–CABG-related major bleeding

Death

MI

Definition Defined as bleeding that is (a) fatal, (b) results in transfusion of ≥2 U of red blood cells or equivalent whole blood, (c) causes significant hypotension with the need for inotropes or surgical intervention (a requirement for surgical access-site repair will constitute major bleeding only if there has been significant hypotension or transfusion of ≥2 U), (d) causes significantly disabling sequellae, or (e) is intracranial and symptomatic or intraocular and leads to significant visual loss. The primary outcome uses all-cause mortality. All deaths will be subclassified as cardiovascular and noncardiovascular. All deaths with a clear cardiovascular cause or unknown will be classified as cardiovascular (including complications of procedures and bleeding). Only deaths because of documented noncardiovascular cause (eg, cancer) will be classified as noncardiovascular. The diagnosis of new MI will depend on the timing of the event after randomization (ie, within the first 24 h of randomization, between 24 h and 7 d after randomization, and N7 d after randomization), the presence or absence of an associated MI at baseline, and whether the suspected event occurred after a revascularization procedure. No associated MI at baseline In patients with no associated baseline MI, either one of the 2 following criteria satisfies the diagnosis for an acute MI: (1) typical rise and fall of biochemical markers of myocardial necrosis including troponin, CK-MB, CK to N2× ULN (or if markers are already elevated, N50% of the lowest recovery enzyme level from the index infarction) with at least one of the following: (a) ischemic symptoms, (b) development of pathologic Q waves on the ECG, and (c) ECG changes indicative of ischemia (ST-segment elevation or depression); or (2) pathologic findings of an acute MI. MI within 24 h of randomization In UA/NSTEMI patients with an associated MI at baseline or in STEMI patients, a new MI within 24 h of randomization is defined as (1) new ischemic symptoms N20 min and (2) new or recurrent ST-segment elevation or depression N1 mm in ≥2 contiguous leads, not due to changes from evolution of the index MI or STEMI patients. MI between 24 h and 7 d In UA/NSTEMI patients with an associated MI at baseline or in STEMI patients, of randomization a new MI between 24 h and 7 d is defined as (a) new ischemic symptoms N20 min and (b) elevation or reelevation of CK-MB (or total CK if CK-MB is not available) ≥2× the upper limit of normal or N50% above the previous valley level and N2× the upper limit of normal in patients with already elevated enzymes or new or recurrent ST-segment elevation or depression N1 mm or new significant Q waves in ≥2 contiguous leads discrete from the baseline MI. MI occurring after 7 d or hospital In all patients, the definition of new MI occurring after hospital discharge or discharges, whichever comes first after 7 d, whichever comes first, will be either one of the 2 following criteria satisfies the diagnosis for an acute, evolving, or recent MI: typical rise and fall of biochemical markers of myocardial necrosis (including troponin, CK-MB, and CK) to N2× ULN (or if markers are already elevated, N50% of the lowest recovery enzyme level from the index infarction and N2× ULN) with at least one of the following: (a) ischemic symptoms, (b) development of pathologic Q waves on the ECG, and (c) ECG changes indicative of ischemia (ST-segment elevation or depression); or pathologic findings of an acute MI. MI post PCI For patients with MI within 24 h after PCI, a new MI is defined by (1) CK-MB⁎ (or total CK if CK-MB is unavailable) ≥3× the upper limit of normal or increased by 50% from the preprocedural valley level and ≥3× ULN in patients with already elevated enzymes or (2) new ST-segment elevation or development of new or significant Q waves in ≥2 contiguous leads (discrete from the baseline MI in STEMI patients). MI post CABG For patients with MI within 24 h after CABG, a new MI is defined by CK-MB (or total CK, if CK-MB is unavailable) (1) ≥5× the upper limit of normal or increased by 50% from the preprocedural valley level and ≥5× ULN in patients with already elevated enzymes and development of new pathologic Q waves in ≥2 contiguous leads or (2) CK-MB value ≥10× ULN without new pathologic Q waves. In all cases of new MI, troponin T or I may be used for the diagnosis of new MI in the absence of CK-MB at the discretion of the event adjudication committee, taking into consideration all available clinical and laboratory evidence. In addition, the event adjudication committee may request further details of the revascularization procedure such as the PCI or CABG written report or supplementary narratives to assist in ascertainment of new post PCI or CABG MI.

American Heart Journal Volume 161, Number 2

Jolly et al 259

Table II (continued) First occurrence until day 30 of any component of death, MI, stroke, non–CABG-related major bleeding Term

Definition

Stroke

Any stroke is defined as the presence of a new focal neurologic deficit thought to be vascular in origin, with signs or symptoms lasting N24 h. It is strongly recommended (but not required) that an imaging procedure such as a CT or MRI be performed. Failure: no success at dilating attempted lesion(s) and/or failure to cross/dilate/not attempted. Partial success: one of ≥2 attempted lesions was successfully dilated and procedure performed but N50% residual or TIMI flow b3 or failure. Full success: lesion(s) attempted was successfully dilated with b50% residual or TIMI 3 flow.

PCI success

Steering committee: S. Jolly (coprincipal investigator), S.R. Mehta (coprincipal investigator), A. Avezum, A. Budaj, J. Cairns, S. Chrolavicius (project manager), R. Diaz, V. Dzavik, M.G. Franzosi, C. Granger, C. Joyner (events adjudication committee chair), M. Keltai, F. Lanas, B. Lewis, K. Niemela, S. Rao, P.G. Steg, V. Valentin, P. Widimsky, D. Xavier, and S. Yusuf⁎ (steering committee chair). Data monitoring committee: P. Sleight (chair), J.L. Anderson, D.L. DeMets, J. Hirsh, D.R. Holmes, Jr, and D.E. Johnstone. Project office staff: S. Chrolavicius (project manager), R. Afzal (statistician), L. Blake, W. Chen, S. Di Diodato, C. Cramp (research coordinator), C. Horsman (research coordinator), B. Jedrzejowski (research coordinator), M. Lawrence (event adjudication coordinator), A. Lehmann, A. Robinson (research coordinator), R. Manojlovic, L. Mastrangelo, E. Pasadyn, T. Sovereign, L. Wasala, L. Xu (statistician). ⁎ If available, troponin may be used instead of CK-MB post PCI if no associated MI at baseline. CT, computed tomography; MRI, magnetic resonance imaging. TIMI, thrombolysis in myocardial infarction; ULN, upper limit of normal.

Statistical considerations

Limitations

The originally estimated sample size was 4,000 patients, which was chosen to provide 80% power for the detection of a relative risk reduction of 25% for a 10% rate of the primary outcome with a 2-sided α ≤ .05. However, it was evident by July 2009 that the aggregate event rate was much lower than originally estimated, and a revised estimate of 6% for the primary outcome at 30 days in the femoral access group led to the calculation of a revised sample size of approximately 7,000 patients to provide 80% power to detect a relative risk reduction of 25%, with 2-sided α ≤ .05. In September 2010, before the end of enrollment and before blinding, the steering committee clarified that the major bleeding component of the primary outcome refers to non-CABG major bleeding. This was because CABG-related major bleeding is unlikely to be modified by vascular access site. The protocol clarification was issued to all sites. Coronary artery bypass graft–related major bleeding will be reported as a tertiary outcome. For the primary analysis, the relative efficacy of radial access versus femoral access will be assessed on the primary outcome by a comparison of the survival curves (estimated using the Kaplan-Meier method) for the 2 treatments using the log-rank statistic (primary test of treatment effect). Treatment effect, expressed as the hazard ratio (radial access vs femoral access) and corresponding 95% CIs, will be estimated from a Cox proportional hazards model. Statistical significance will be assessed using a 2-sided α ≤ .05. There will be no adjustment for multiple comparisons of key secondary outcomes because there are only 2, and each contains components of the primary outcome. An analysis of subgroups by tertiles of both center and operators' radial procedural volume will be performed as well.

The first potential limitation of the RIVAL trial is the exclusion of higher risk patients. For example, operators may have been unwilling to randomize morbidly obese patients because of a lack of clinical equipoise in this population. A second potential limitation is that very high-proportion radial operators (N90%) may not have commonly participated in the trial potentially because of a lack clinical equipoise between the 2 procedures among these operators. However, on the other hand, very high-proportion radial operators could have higher femoral complications rates; accordingly, we designed a pragmatic trial engaging operators with expertise in both procedures. Finally, the optimal minimum annual number of radial procedures to define expertise is unknown; accordingly, there is a potential limitation in the chosen cut point of 50 radial procedures annually. However, in practice, the RIVAL trial was successful in engaging highvolume operators who had significantly exceeded these minimum requirements for radial volume and who had extensive experience with both procedures.

Study status Enrollment was completed on November 3, 2010, with 7,021 patients randomized with the expectation that 30-day data are likely to be available by the spring or summer of 2011. The recruitment by country and site is shown in an online Appendix.

Summary The RIVAL trial will be the first large randomized trial comparing radial versus femoral access for coronary angiography/intervention among patients with ACS that

260 Jolly et al

is adequately powered to determine effects on major efficacy and safety outcomes.

Disclosures Funding for the RIVAL trial was provided by SanofiAventis, Population Health Research Institute, and the Canadian Network and Center for Trials Internationally (CANNeCTIN), which is funded by the Canadian Institutes of Health Research.

References 1. Stone GW, McLaurin BT, Cox DA, et al. Bivalirudin for patients with acute coronary syndromes. N Engl J Med 2006;355:2203-16. 2. Yusuf S, Mehta SR, Chrolavicius S, et al. Comparison of fondaparinux and enoxaparin in acute coronary syndromes. N Engl J Med 2006; 354:1464-76. 3. Fox KA, Goodman SG, Klein W, et al. Management of acute coronary syndromes. Variations in practice and outcome; findings from the Global Registry of Acute Coronary Events (GRACE). Eur Heart J 2002;23:1177-89. 4. Budaj A, Eikelboom JW, Mehta SR, et al. Improving clinical outcomes by reducing bleeding in patients with non–ST-elevation acute coronary syndromes. Eur Heart J 2008. 5. Rao SV, O'Grady K, Pieper KS, et al. Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes. Am J Cardiol 2005;96:1200-6. 6. Eikelboom JW, Mehta SR, Anand SS, et al. Adverse impact of bleeding on prognosis in patients with acute coronary syndromes. Circulation 2006;114:774-82. 7. Ndrepepa G, Berger PB, Mehilli J, et al. Periprocedural bleeding and 1-year outcome after percutaneous coronary interventions: appropriateness of including bleeding as a component of a quadruple end point. J Am Coll Cardiol 2008;51:690-7. 8. Chase AJ, Fretz EB, Warburton WP, et al. The association of arterial access site at angioplasty with transfusion and mortality: the M.O.R.T. A.L study: (Mortality benefit of Reduced Transfusion After PCI via the Arm or Leg). Heart 2008. 9. Rao SV, Ou FS, Wang TY, et al. Trends in the prevalence and outcomes of radial and femoral approaches to percutaneous coronary intervention: a report from the National Cardiovascular Data Registry. JACC Cardiovasc Interv 2008;1:379-86. 10. Jolly SS, Amlani S, Hamon M, et al. Radial versus femoral access for coronary angiography or intervention and the impact on major bleeding and ischemic events: a systematic review and meta-analysis of randomized trials. Am Heart J 2009;157:132-40.

American Heart Journal February 2011

11. Montalescot G, Ongen Z, Guindy R, et al. Predictors of outcome in patients undergoing PCI. Results of the RIVIERA study. Int J Cardiol 2007. 12. Cantor WJ, Mahaffey KW, Huang Z, et al. Bleeding complications in patients with acute coronary syndrome undergoing early invasive management can be reduced with radial access, smaller sheath sizes, and timely sheath removal. Catheter Cardiovasc Interv 2007;69: 73-83. 13. Hamon M, Steg G, Faxon D, et al. Major bleeding in patients with acute coronary syndrome undergoing early invasive management can be reduced by fondaparinux, even in the context of trans-radial coronary intervention: insights from OASIS-5 trial. Circulation 2006; 114(Suppl II):552. 14. Hamon M, Rasmussen LH, Manoukian SV, et al. Choice of arterial access site and outcomes in patients with acute coronary syndromes managed with an early invasive strategy: the ACUITY trial. EuroIntervention 2009;5:115-20. 15. Biancari F, D'Andrea V, Di Marco C, et al. Meta-analysis of randomized trials on the efficacy of vascular closure devices after diagnostic angiography and angioplasty. Am Heart J 2010;159: 518-31. 16. Manoukian SV, Feit F, Mehran R, et al. Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY trial. J Am Coll Cardiol 2007;49:1362-8. 17. Stone GW, Witzenbichler B, Guagliumi G, et al. Bivalirudin during primary PCI in acute myocardial infarction. N Engl J Med 2008;358: 2218-30. 18. Mehta SR, Bassand JP, Chrolavicius S, et al. Design and rationale of CURRENT-OASIS 7: a randomized, 2 × 2 factorial trial evaluating optimal dosing strategies for clopidogrel and aspirin in patients with ST and non–ST-elevation acute coronary syndromes managed with an early invasive strategy. Am Heart J 2008;156:1080-8.e1. 19. Mehta SR, Bassand JP, Chrolavicius S, et al. Dose comparisons of clopidogrel and aspirin in acute coronary syndromes. N Engl J Med 2010;363:930-42. 20. Mehta SR, Tanguay JF, Eikelboom JW, et al. Double-dose versus standard-dose clopidogrel and high-dose versus low-dose aspirin in individuals undergoing percutaneous coronary intervention for acute coronary syndromes (CURRENT-OASIS 7): a randomised factorial trial. Lancet 2010;376:1233-43. 21. Spaulding C, Lefevre T, Funck F, et al. Left radial approach for coronary angiography: results of a prospective study. Cathet Cardiovasc Diagn 1996;39:365-70. 22. Hildick-Smith DJ, Lowe MD, Walsh JT, et al. Coronary angiography from the radial artery—experience, complications and limitations. Int J Cardiol 1998;64:231-9. 23. Brueck M, Bandorski D, Kramer W, et al. A randomized comparison of transradial versus transfemoral approach for coronary angiography and angioplasty. JACC Cardiovasc Interv 2009;2:1047-54.

American Heart Journal Volume 161, Number 2

Jolly et al 260.e1

Appendix A

BELGIUM (43 patients)

Table I. Recruitment by country Country

Argentina Australia Belgium Brazil Bulgaria Canada Chile China Croatia Czech Republic Finland France Germany Hungary India Ireland

No. of patients randomized

Country

No. of patients randomized

94 55 43 183 551 1475 146 208 1 187 1124 187 94 31 902 9

Israel Italy Latvia Lithuania Malaysia Mexico New Zealand Poland Romania Russia Singapore Slovakia Spain Sweden United Kingdom United States

239 129 11 80 1 17 9 583 12 1 5 20 432 50 20 122

Appendix B: RIVAL program list of principal investigators and recruitment by site ARGENTINA (94 patients) 252 Dr Eduardo Conrado Conci, Instituto Modelo De Cardiologia Priv. S.R.L. (3) 260 Dr L. L. Lobo Marquez, Instituto de Cardiologia de Tucuman—SRL (1) 265 Dr Guillermo Covelli, Clinica Privada del Prado (16) 266 Dr Miguel Angel Hominal, Sanatorio Medico de Diagnostico y Tratamiento (13) 271 Dr Gerardo Zapata, Instituto Cardiovascular de Rosario (5) 272 Dr J. C. Pomposiello, Hospital Privadomde Comunidad (12) 277 Dr Simon Salzberg, Hospital Juan A. Fernandez (15) 278 Dr Alejandro Sanchez, Policlinico Modelo Cipolletti (14) 280 Dr Daniel Santos, Instituto Cardiologico Especializado SRL (1) 282 Dr Daniel Piskorz, Sanatorio Britanico de Rosario (2) 284 Dr Mario Berli, Hospital Provinicial Dr Jose Maria Cull (5) 286 Dr Claudo Rodolfo Majul, Hospital Britanico de Buenos Aires (7) AUSTRALIA (55 patients) 233 Dr Matthew Worthley, Royal Adelaide Hospital (54) 240 Prof Peter Thompson, Sir Charles Gairdner Hospital (1)

936 Dr P-E Massart, Clinique et Maternite SainteElisabeth de Namur (21) 937 Prof Vincent Dangoisse, Cliniques Universitaires de Mont-Godinne (7) 938 Dr Marc Vincent, Clinique générale Saint-Jean (15) BRAZIL (183 patients) 315 Dr José Armando Mangione, Hospital Beneficiência Portuguesa de São Paulo (8) 320 Dr Sandra Andrade, Mendonça Hilgemberg Incor-hemocardio (9) 326 Dr Maria Sanali Moura de Oliveira Paiva, Natal Hospital Center (10) 328 Dr Gilmar Reis, Hospital São Francisco de Assis (17) 331 Dr José Francisco Kerr Saraiva, HMCP PucCampinas Hospital e Maternidade Celso Pierro (5) 332 Dr Ari Timerman, Instituto Dante Pazzanese de Cardiologia (48) 333 Dr Rogério Tadeu Tumelero, Hospital São Vicente de Paulo (38) 334 Dr Wladimir Faustino Saporito, Hospital Estadual Mario Covas (8) 336 Dr Roberto Vieira Botelho, Instituto do Coracao do Triangulo Mineir (40) BULGARIA (551 patients) 431 Dr Alexander Doganov, National Heart Hospital (23) 432 Prof Julia Jorgova-Makedonska, “St Ekaterina” University Hospital (21) 435 Dr Atanas Penev, UMHAT-Sveta Marina (3) 433 Dr Georgi Mazhdrakov, UMHAT “St Anna” (52) 436 Dr Ivan Manoukov, Clinic of Invasive Cardiology, University Hospital “Sveti Georgi” (452) CANADA (1,475 patients) 001 Dr Sanjit Jolly/Dr Shamir Mehta, Hamilton General Hospital (696) 002 Dr Michael Rokoss, Henderson Hospital (31) 003 Dr Omid Salehian, McMaster University Medical Centre (29) 004 Dr Gilbert Gosselin, Montreal Heart Institute (1) 005 Dr Jeffrey Pang/Dr Cam Joyner, Sunnybrook Health Sciences Centre (30) 006 Dr Sven Pallie, Niagara Health System (NHS)— St Catharines General Hospital Site (58) 007 Dr Anthony Fung, Vancouver General Hospital (150) 009 Dr Yun Kai Chan, NHS—Greater Niagara General Site (12) 013 Dr Farrukh Hussain, St Boniface General Hospital (10) 014 Dr C. Van Kieu, CSSSRY C.H. Honore Mercier (1) 018 Dr Asim Cheema, St Michael's Hospital (138)

American Heart Journal February 2011

260.e2 Jolly et al

019 Dr G. Calvin, MacCallum Memorial University Eastern Health Sciences Center (7) 020 Dr Frank Nigro, Intermountain Research Consultants (2) 026 Dr Patrick Beliveau, CHUQ—Hotel Dieu de Quebec (49) 030 Dr Denis-Carl Phaneuf, Hospital Pierre-Le Gardeur (3) 031 Dr Jean-Pierre Dery, Hôpital Laval (7) 033 Dr Vlad Dzavik, University Health Network (22) 035 Dr Salima Shariff, Surrey Memorial Hospital (27) 039 Dr Joseph Berlingieri, Joseph Berlingieri and William Nisker (JBN) Medical Diagnostic Services (19) 062 Dr Warren Cantor, York PCI Group Inc (53) 063 Dr Robert Boone, Providence Health Care, St Paul's (42) 064 Dr Francois Charbonneau, University of Calgary (88) CHILE (146 patients) 353 Dr R. Lamich Betancourt, Hospital Barros Luco (70) 354 Dr C.P.P. Jofre, Hospital Dr Herman Henriquez Aravena (52) 355 Dr E.E.G. Flores, Hospital Clinico Regional Valdivia (23) 360 Dr Misael Lopetegui, Hospital Clínico San Borja Arriarán (1) CHINA (208 patients) 906 Dr Shuyang Zhang, Peking Union Medical College Hospital (10) 912 Dr Yaling Han, The General Hospital of Shenyang Military C (11) 914 Dr Meng Wei, Shanghai No. 6 People Hospital (22) 915 Dr Daowen Wang, Tongji Hospital of Huazhong University (9) 917 Dr Jianan Wang, Second Affiliated Hospital Zhejiang University (22) 919 Dr Jiyan Chen, GuangDong Provincial People's Hospital (60) 924 Prof Tianchang Li, Beijing Tongren Hospital (1) 927 Dr Biao Xu, The Affiliated Drum Tower Hospital (31) 928 Dr Genshan Ma, Zhong Da Hospital (42) CROATIA (1 patient) 457 Dr Mijo Bergovec, Klinicka bolnica Dubrava (1) CZECH REPUBLIC (187 patients) 404 Assoc Prof Dr Pavel Cervinka, Masaryk Hospital (54) 405 MUDr Zdenek Coufal, Regional Hospital T. Bati a.s. Zlín (14) 410 Dr Petr Kala, Fakultní nemocnice Brno ( 10) 412 Prof Ales Linhart, General Faculty Hospital (3)

413 Dr Roman Ondrejcak, KKN a.s., Nemocnice Karlovy Vary, kardiologické odd (29) 421 Prof Jan Vojacek, Fakultni nemocnice Hradec Kralove (6) 422 Dr Richard Rokyta, I. Interní klinika, Fakultní (27) 430 Dr Michal Rezek, University Hospital St Anna (44) FINLAND (1,124 patients) 636 Oyl Saila Vikman, Heart Center, Tampere University Hospital (232) 637 Prof Juhani Airaksinen, Turku University Hospital (30) 640 Oyl Antti Ylitalo, Satakunta Central Hospital (5) 641 Oyl Matti Niemela, Oulu University Hospital (857) FRANCE (187 patients) 583 Dr Emile Ferrari, CHU de Nice—Hôpital Pasteur (160) 585 Dr Gilles Grollier, CHU de caen—Hôpital Côte de Nacre (6) 595 Prof Gabriel Steg, Hôpital Bichat Claude Bernard (21) GERMANY (94 patients) 508 Dr Dietrich Andresen, Vivantes Klinikum am Urban (8) 510 Dr Florin Laubenthal, Elisabeth-Krankenhaus Essen (1) 522 Dr Juergen vom Dahl, Kliniken Maria Hilf GmbH (62) 525 Dr Christoph Nienaber, Universitaet Rostock (7) 532 Dr Stefan Hoffmann, Vivantes Klinikum im Friedrichshain (16) HUNGARY (31 patients) 485 Dr Peter Polgár Josa Andras Teaching Hospital (1) 487 Dr Béla Nagybaczoni Bajcsy Zsilinszky Hospital (30) INDIA (902 patients) 851 Dr Kumar Rajendra Premchand, Krishna Institute of Medical Sciences (240) 852 Dr M. Bhaskar Rao, CARE Hospital (71) 854 Dr Keyur Parikh, S.A.L. Hospital & Medical Institute (161) 862 Dr Nakul Sinha, Sanjay Gandhi PGIMS (299) 863 Dr Hemang Baxi, The Heart Care Clinic (38) 872 Dr Brian Pinto, Holy Family Hospital (76) 874 Dr Sudhir R. Naik, Apollo Hospitals (2) 881 Dr Pradeep Kumar Shetty, Narayana Hrudayalaya (15) IRELAND (9 patients) 391 Dr Peter Crean, St James's Hospital (9) ISRAEL (239 patients) 971 Dr Yoav Turgeman, Haemek Medical Centre (17)

American Heart Journal Volume 161, Number 2

972 Prof Basil S. Lewis, Lady Davis Carmel Medical Centre (114) 980 Dr Joel Arbel, Meir Medical Center (3) 981 Dr Alon Marmor, Heart Institute (62) 983 Dr Majdi Halabi, Division of Invasive Cardiology (43) ITALY (129) 756 Dr Marco Rossi, Istituto Clinico Humanitas (1) 766 Prof Giancarlo Piovaccari, Ospedale Infermi U.O. Cardiologia (56) 767 Dott Salvatore Pirelli, Div. Di Cardiologia, Azienda Istituti Ospitalieri Cremona (43) 771 Dr Giuseppe Steffenino, Azienda Ospeddliera Santa Croce e Carle (3) 773 Dr Roberto Zanini, Azienda Ospedaliera “Carlo Poma” (12) 775 Dr Ugo Limbruno, Ospedale della Misericordia (14) LATVIA (11 patients) 396 Dr Andrejs Erglis, Pauls Stradins Clinical University Hospital (11) LITHUANIA (80 patients) 992 Dr Ramunas Unikas, Kaunas Medical University Hospital (80) MALAYSIA (1 patient) 801 Dr Chong Wei Peng, University Malaya Medical Centre (1) MEXICO (17 patients) 363 Dr Jose Luis Arenas Leon, Hospital Angeles Centro Medico Del Potos (5) 370 Dr M.S.L.Velasco, Star Medica Morelia (12) NEW ZEALAND (9 patients) 067 Dr Gerard Devlin, Waikato Hospital (6) 068 Dr Scott Harding, Wellington Hospital (3) POLAND (583 patients) 617 Prof Andrzej Budaj, Szpital Grochowski (134) 618 Dr Piotr Achremczyk, Radomski Szpital Specjalistyczny (7) 619 Dr Pawel Miekus, Szpital Miejski im. J. Brudzinskiego (6) 621 Dr Barbara Kusnierz, Wojewódzki Szpital Specjalistyczny nr 4 (19) 623 Dr Bozena Wrzosek, Wojewódzki Szpital Specjalistyczny (181) 624 Dr Jerzy Kopaczewski, Szpital Wojewodzki (16) 628 Dr Jan Wodniecki, Szpital Specjalistyczny w Zabrzu (3) 628 Dr Damian Kawecki, Szpital Specjalistyczny w Zabrzu (26)

Jolly et al 260.e3

631 Dr Maciej Dalkowski, Regionalny Osrodek Kardiologii, “Miedziowe Centrum Zdrowia” S.A. (171) 647 Prof Lech Polonski, III, Katedra i Oddzial Kliniczny Kardiologii SUM (14) 648 Prof Zbigniew Kalarus, Slaskie Center for Heart Disease (6) ROMANIA (12 patients) 442 Prof Radu Capalneanu, Institutul Inimii N. Stancioiu (12) RUSSIA (1 patient) 813 Prof Svetlana Berns, Municipal Healthcare Institution (1) SINGAPORE (5 patients) 791 Prof Tian Hai Koh, National Heart Centre (5) SLOVAKIA (20 patients) 452 Dr Roman Margozcy, Stredoslovensky ustav srdcovych a cievnych chorob (10) 454 Dr Peter Kurray, Kardiocentrum Nitra sro (10) SPAIN (432 patients) 701 Dr Vicent Valentin, Hospital Universitario Dr Peset (128) 702 Dr Nicolas Vazquez, Hospital Universitario Juan Canalejo (13) 706 Dr Juan Angel Ferrer, Hospital General Vall D'Hebron (25) 707 Dr Manel Sabate, Hospital De La Santa Creu i De Sant Pau (4) 713 Dr Inaki Lekuona, Hospital De Galdakao-Usansolo (51) 721 Dr Francisco Bosa, Hospital Universitario de Canarias (116) 722 Dr Jose Moreu, Hospital Virgen de la Salud (1) 724 Dr Andrés Iñiguez, Hospital Meixoeiro (24) 725 Dr Francisco Macaya, Hospital Clínico San Carlos (29) 730 Dr Juan Sanchis, Hospital Clinico de Valencia (4) 732 Dr Ramón López Palop, Hospital Universitario San Juan de Alicante (35) 738 Dr Ramiro Trillo, Hospital Clinico de Santiago (2) SWEDEN (50 patients) 652 Dr Aida Hot-Bjelak, Capio St Goran's Hospital (15) 653 Dr Loghman Henareh, Karolinska University Hospital (5) 654 Prof Goran Olivecrona, Lund University Hospital (30) UNITED KINGDOM (20 patients) 941 Dr Andreas Baumbach, Bristol Royal Infirmary (15)

260.e4 Jolly et al

944 Prof Adam de Belder, Royal Sussex County Hospital (2) 948 Dr Iqbal Malik, St Mary's Hospital (1) 950 Dr Mike Pitt, Heart of England NHS Foundation Trust (2) USA (122 patients) 75 Dr Mehrdad Saririan, Maricopa Integrated Health System (1)

American Heart Journal February 2011

104 Dr Ramesh Mazhari, GWU Medical Faculty Associates (7) 105 Dr Steven L. Goldberg, University of Washington (1) 172 Dr Narendra Singh, Northside Cardiology (32) 178 Dr Joseph Chambers, Endovascular Research (1) 183 Dr Stephen Thew, Heart Clinics Northwest (62) 186 Dr M. Kevin Ariani, Northridge Hospital Medical Center (1) 192 Dr Joana Magno, The Queen's Medical Center (4)

A randomized, partially blinded, multicenter, active-controlled, dose-ranging study assessing the safety, efficacy, and pharmacodynamics of the REG1 anticoagulation system in patients with acute coronary syndromes: Design and rationale of the RADAR Phase IIb trial Thomas J. Povsic, MD, PhD, a,l Mauricio G. Cohen, MD, b,l Roxana Mehran, MD, c,l Christopher E. Buller, MD, d,l Christoph Bode, MD, e,l Jan H. Cornel, MD, f,l Jarosław D. Kasprzak, MD, g,l Gilles Montalescot, MD, h,l Diane Joseph, a William A. Wargin, PhD, i Christopher P. Rusconi, PhD, j Steven L. Zelenkofske, DO, k,l Richard C. Becker, MD, a,l and John H. Alexander, MD, MHS a,l Durham, and Chapel Hill, NC; Miami, FL; New York, NY; Ontario, Canada; Freiberg, Germany; Alkmaar, Netherlands; Łódź , Poland; Paris, France; and Basking Ridge, NJ

Anticoagulants are the cornerstone of current acute coronary syndrome (ACS) therapy; however, anticoagulation regimens that aggressively reduce ischemic events are almost uniformly associated with more bleeding. REG1, an anticoagulation system, consists of RB006 (pegnivacogin), an RNA oligonucleotide factor IXa inhibitor, and RB007 (anivamersen), its complementary controlling agent. Phase I and IIa studies defined predictable relationships between doses of RB006, RB007, and degree of antifactor IX activity. The efficacy and safety of REG1 for the treatment of patients with ACS managed invasively and the safety of reversing RB006 with RB007 after cardiac catheterization are unknown. Randomized, partiallyblinded, multicenter, active-controlled, dose-ranging study assessing the safety, efficacy, and pharmacodynamics of the REG1 anticoagulation system compared to unfractionated heparin or low molecular heparin in subjects with acute coronary syndrome (RADAR) is designed to assess both the efficacy of the anticoagulant RB006 and the safety of a range of levels of RB006 reversal with RB007. The objectives of RADAR are (1) to determine the safety of a range of levels of RB006 reversal with RB007 after catheterization, (2) to confirm whether a dose of 1 mg/kg RB006 results in near-complete inhibition of factor IXa in patients with ACS, and (3) to assess the efficacy of RB006 as an anticoagulant in patients with ACS undergoing percutaneous coronary intervention. (Am Heart J 2011;161:261-268.e2.)

Anticoagulant therapies are the cornerstone of treatment of patients with acute coronary syndromes (ACS); however, they are associated with clinically significant increases

From the aDuke Clinical Research Institute, Duke University Medical Center, Durham, NC, b Miller School of Medicine, University of Miami, Miami, FL, cMount Sinai Medical Center, New York, NY, dHamilton General Hospital, Hamilton, Ontario, Canada, eUniversity of Freiberg, Freiberg, Germany, fMedisch Centrum Alkmaar, Alkmaar, Netherlands, g Medical University of Lodz, Łódź, Poland, hInstitut de Cardiologie, Pitié-Salpétrière Hospital, Paris, France, iPK-PM Associates, LLC, Chapel Hill, NC, jRegado Biosciences, Durham, NC, and kRegado Biosciences, Basking Ridge, NJ. l On behalf of the RADAR Investigators. See online Appendix B for a complete listing. RCT reg no. NCT00932100. Vladimir Dzavik, MD, served as guest editor for this article. Submitted July 1, 2010; accepted October 15, 2010.

Reprint requests: Thomas J. Povsic, MD, PhD, Alexander H. Sands Building, 303 Research Drive, Room 321, Duke University Medical Center, Durham, NC, 27710. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.022

in bleeding.1 As the clinical consequences of bleeding have become clearer,2-6 attention has focused on reducing bleeding without compromising antithrombotic efficacy. The ideal anticoagulant would prevent thrombus formation, cause minimal or no bleeding, be easy to administer and monitor, and be reversible should a clinical need (bleeding) arise or when anticoagulation is no longer necessary. Currently available parenteral anticoagulants have important limitations.7 Unfractionated heparin has variable and unpredictable anticoagulant effects, results in platelet activation, can cause heparininduced thrombocytopenia, and its reversal agent, protamine, has its own adverse effects.8-10 Low-molecular-weight heparins are long acting, lack effective reversal strategies, cannot be monitored, are associated with variability in achieving adequate levels of anticoagulation,11 and require a delay in sheath removal.12 Bivalirudin is rapidly cleared but requires an infusion for

American Heart Journal February 2011

262 Povsic et al

Figure 1

Table I. Components of the REG1 system

Half-life Clinical activity Clinical target Elimination

Structure of RB006-RB007 complex. From Dyke CK, Steinhubl SR, Kleiman NS, et al. Circulation 2006;114:2490-7.

longer treatment times and may be less effective at preventing ischemic events.13 Fondaparinux is associated with low bleeding rates but is less effective as an anticoagulant in the catheterization laboratory.14 One approach to more reliable inhibition of coagulation is to target upstream proteases that are activated in the early steps of coagulation and are present at lower and more predictable concentrations. Factor IX is a particularly attractive target, given its central role in the generation of factor Xa and thrombin and in the propagation of the clot on the surface of activated platelets.15-17 In addition, as tissue factor-mediated factor X activation proceeds through factor IX, effective factor IX inhibition may inhibit thrombin generation via both the intrinsic and extrinsic pathways.18 Finally, the pathophysiologic correlate of factor IX inhibition is hemophilia B. Patients with hemophilia B have prolonged activated partial thromboplastin times (aPTT), suggesting that aPTT might be used to monitor factor IXa inhibition with REG1 (Regado Biosciences, Basking Ridge, NJ).19,20

The REG1 anticoagulation system REG1 is a novel anticoagulation system made up of an RNA aptamer that binds to and selectively inactivates factor IXa (RB006, pegnivacogin) and its complementary controlling agent (RB007, anivamersen).21 Aptamers are oligonucleotides that bind to proteins based on their three-dimensional structure.22,23 Because of their nucleotide structure, they typically lack immunogenicity and toxicity, have tunable pharmacokinetics (PK), and can be formulated as either intravenous (IV) or subcutaneous injectables. Given their small size, they are frequently able to target protein interfaces that have proven difficult to pursue using larger peptides.

RB006

RB007

RB006-RB007 complex

100 h N30 h Factor IX Nucleases

b5 min Inert RB006 Nucleases

10 min Inert None Nucleases

Unique to aptamers is their inherent ability to code for their own reversal agents based on their oligonucleotide sequence and the high specificity and affinity afforded by Watson-Crick base pairing (Figure 1, Table I).20 This allows modulation of aptamer activity at any given time point depending on the clinical circumstances. RB006 (Figure 1) consists of a modified nucleotide aptamer attached to a polyethylene glycol chain to slow degradation and increase plasma half-life to approximately 100 hours with stable anticoagulant activity observed for 30 hours after single-dose administration (Table I).21 Extensive preclinical and early clinical modeling have demonstrated that RB006 results in reliable PK and pharmacodynamics (PD) based on the relationship between RB006 dosing, plasma concentrations, aPTT prolongation, and factor IX activity. These studies suggest that 0.75 to 1.0 mg/kg of RB006 results in consistent and near-complete (N99%) inhibition of factor IX in stable subjects and during elective percutaneous coronary intervention (PCI).21,24-26 RB007 has a half-life of b5 minutes and is rapidly cleared from the bloodstream by endogenous endonucleases. RB007 has no known biological effect other than rapid and irreversible binding to RB006 resulting in dissociation of RB006 from its factor IXa binding site and permanent reversal of anticoagulant effect.24,25 Administration of excess molar ratios of RB007 results in rapid (b10 minutes) sustained reversal of RB006 anticoagulant activity.25 Lower ratios of RB007-RB006 administration result in partial reversal of factor IX inhibition.

Clinical experience The REG1 program has 3 phase I studies assessing the PK and PD of the REG1 system in healthy volunteers and patients with coronary artery disease (CAD).21,24,25 A phase II pilot study demonstrated the feasibility of using RB006 as the sole anticoagulant therapy in patients undergoing elective PCI.26 The applicability of these findings to patients with ACS undergoing PCI remains unknown.

The RADAR trial RADAR (Clinicaltrials.gov identifier NCT00932100) is an international, multicenter, phase II, randomized, partially blinded, active control, clinical trial investigating

American Heart Journal Volume 161, Number 2

RB006 and a range of doses of RB007 in patients with ACS managed using an early invasive strategy. The purpose of this article was to describe the design and rationale for the RADAR trial. The authors are solely responsible for the design and conduct of this study, all study analyses, and the drafting and editing of the paper and its final contents.

Povsic et al 263

Figure 2

Objectives The main objectives of RADAR are (1) to determine the safety of a range of levels of RB006 reversal with RB007 after catheterization, (2) to confirm whether a dose of 1 mg/kg RB006 results in near-complete inhibition of factor IXa in patients with ACS, and (3) to assess the efficacy of RB006 as an anticoagulant in patients with ACS undergoing PCI. Secondary objectives include assessment of the safety of early post-PCI anticoagulation reversal and assessment of the feasibility of early sheath removal and patient ambulation after various degrees of RB006 reversal with RB007. Patient population A total of 800 patients will be recruited from approximately 80 sites worldwide with enrollment in Canada, France, Germany, the Netherlands, Poland, and the United States. Inclusion/exclusion criteria are listed in Figure 2. Patients are eligible if they are 18 to 80 years old and have experienced ischemic symptoms for ≥10 minutes within 72 hours of enrollment. Ischemic symptoms must be associated with (1) ST-segment depression or transient ST-segment elevation; (2) elevated troponin I, troponin T, or creatine kinase (CK)-MB; (3) documented CAD on prior angiography; or (4) a history of PCI or coronary artery bypass grafting (CABG).27 To be eligible, cardiac catheterization must be planned within 24 hours of randomization. Key exclusion criteria are acute ST-segment elevation myocardial infarction (MI), planned use of sheath sizes N7F during catheterization, and contraindications to anticoagulation including a history of intracranial bleeding or aneurysm and recent use of bivalirudin, glycoprotein (GP) IIb/IIIa inhibitors, or fibrinolytic agents. Patients receiving fondaparinux will be eligible 24 hours after their last dose of fondaparinux. Randomization and study drug Patients in RADAR will be randomized in a 2:1:1:2:2 fashion to REG1 with 25%, 50%, 75%, or 100% reversal or heparin and recommended GP IIb/IIIa inhibition (Figure 3). RADAR is designed as a partially blinded study with assignment to open-label REG1 or heparin and blinded RB007.

Inclusion and exclusion criteria.

REG1 arms RB006 dosing. Patients randomized to REG1 will receive 1 mg/kg of RB006 immediately after randomization (if no prior heparin administered) or as soon as the aPTT is b60 seconds (if prior heparin administered). Additional RB006 dosing. Given the data suggesting the consistent and high level of factor IX inhibition for up to 24 hours with 1 mg/kg of RB006, redosing is expected to be rare (Table II). Provisions have been made, however, to ensure the adequacy of anticoagulation. Local aPTTs will be drawn at 20 minutes and at 1, 4 (PK/PD substudy only), and 10 hours after RB006 administration, provided the patient has not undergone cardiac catheterization before these time points. Redosing will occur only if the patient is to undergo PCI; the most recent aPTT is b2 × upper limit of normal (ULN) of the local aPTT assay or b2 × baseline aPTT value; and no prior redosing has been performed. Redosing based on activated clotting time (ACT) will be limited to cases where aPTT data are not available and N4 hours have elapsed since RB006 administration.

American Heart Journal February 2011

264 Povsic et al

Figure 3

Study flow.

Table II A. Nomogram for RB006 redosing based on aPTT RB006 redose (mg/kg)

Relative aPTT value N2 1.81-2.0 1.61-1.8 1.41-1.6 1.21-1.4 1.0-1.2

No redose 0.2 0.4 0.6 0.6 0.8

B. Nomogram for RB006 redosing based on ACT 4-10 h post-RB006 administration (if aPTT not available) ACT (s) ≥180 b180

RB006 redose (mg/kg) No redose 0.3

N10 h post-RB006 administration (if aPTT not available) ACT (s) ≥180 b180

RB006 redose (mg/kg) No redose 0.4

Pharmacokinetics/PD substudy A key component of RADAR will be to verify the adequacy of 1 mg/kg of RB006 to achieve near-complete inhibition of factor IX activity in patients with ACS. A comprehensive PK/PD assessment will be performed in 20 patients who have not and 10 patients who have received heparin before randomization. Blood drawn

immediately before and 10 minutes after RB006 dosing and at the onset and completion of catheterization will be analyzed to ensure that RB006 concentrations are in the predicted range and that the corresponding effect on aPTT and factor IX inhibition is consistent with a high level of factor IX inhibition as was predicted from early phase studies in clinically stable patients. RB007 dosing. Patients randomized to REG1 will also be randomized to variable doses of RB007 designed to effect 25% (0.075 mg/kg, n = 200), 50% (0.2 mg/kg, n = 100), 75% (0.4 mg/kg, n = 100), or 100% (1 mg/kg, n = 200) reversal of RB006 activity (Figure 3). Upon completion of catheterization, a blinded dose of RB007 will be administered. If a patient receives RB006 but does not undergo catheterization, the blinded RB007 dose will be administered 24 hours after RB006 dosing or once the decision is made to not perform cardiac catheterization if the patient will not remain under observation for 24 hours. An aPTT will be obtained, and the sheath removed 10 minutes after administration of the randomized blinded RB007 dose. If hemostasis has not been achieved after 20 minutes or should uncontrollable bleeding occur, open-label 1 mg/kg RB007 may be administered to achieve 100% reversal of RB006.25,26

Heparin arm Patients randomized to heparin (n = 200) will be treated with enoxaparin or heparin, according to local practice, and dosed per a suggested nomogram (Table III). If a subject is to undergo PCI, unfractionated heparin will be

American Heart Journal Volume 161, Number 2

Povsic et al 265

Table III. Heparin-dosing nomogram

threatened closure, thrombus formation, distal embolization, or vessel dissection. To assess the effectiveness of RB006 to prevent thrombotic complications during PCI, the presence of intravascular thrombus on angiography, abrupt closure, threatened closure, side-branch closure, catheterassociate or intraprocedural thrombus formation, distal embolization, and no-reflow will be assessed in all patients.

Initial heparin dose Age

Male

Female

b60 60-85

14 U/kg 12 U/kg

12 U/kg 11 U/kg

Infusion alterations

aPTT

aPTT

b1.2 × ULN 1.2-1.5 × ULN 1.51-2.0 × ULN 2.01-2.4 × ULN 2.41-2.8 × ULN 2.81-4.0 × ULN N4.0 × ULN

b40 40-49 50-75 76-85 86-100 100-150 N150

Bolus (U)

Stop (min)

Rate (U/h) Δ 100 Δ 100

3000

30 60 60

Δ Δ Δ Δ

−100 −100 −200 −300

administered to achieve an ACT N200 seconds. Sheath removal, according to local practice, will be encouraged after documentation of an ACT b165 seconds. For patients treated with enoxaparin, a 1 mg/kg subcutaneous dose every 12 hours is recommended, with an additional 0.3 mg/kg IV dose to be administered if the most recent enoxaparin dose was ≥8 hours from the time of PCI.12 Sheath removal will occur 6 hours after the last subcutaneous dose or 4 hours after the last IV dose of enoxaparin, whichever is later. Concomitant medications. Aspirin is to be administered before or immediately after study enrollment. Thienopyridine therapy with loading doses (clopidogrel 600 mg or prasugrel 60 mg) is encouraged at or before study enrollment; however, if local practice stipulates later administration, alternative administration is permitted consistent with product labeling. Given the presentation of these patients with ACS, the use of GP IIb/IIIa inhibitors is recommended as the standard of care for patients undergoing PCI in the heparin group. Glycoprotein IIb/IIIa inhibitors may be started upon presentation or deferred until a decision is made regarding PCI according to local practice. All other concomitant medications are left to the discretion of the investigator.

Cardiac catheterization Cardiac catheterization and, if appropriate, PCI are to be performed within 24 hours of RB006 administration. If patients received prior heparin, cardiac catheterization will be delayed until N3 hours after REG1 administration to ensure that any heparin-related anticoagulant effect is gone by the end of the procedure. If PCI is indicated, provisional GP IIb/IIIa inhibitor use is permitted at the discretion of the operator in the event of vessel closure or

Sheath removal and bleeding assessments. Sheath removal procedures have been formalized to ensure standard bleeding assessments and processes across groups. If manual or mechanical compression is used to achieve hemostasis, bleeding assessments will be performed every 10 minutes, and the time to hemostasis recorded. Hemostasis achieved through the use of vascular closure devices (VCDs) will follow similar procedures, with deployment of the VCD and arterial closure 10 minutes after administration of RB007. Patients in whom VCDs are used will undergo identical assessment. If complete hemostasis fails after VCD deployment, manual compression will be performed for up to 20 minutes before any further interventions. Formal bleeding assessments will be performed and recorded for all treatment arms at 10 and 20 minutes and at 1, 2, 4, 8, 16, and 24 hours, provided the patient has not been discharged.

Primary and secondary outcomes The primary outcome is a composite of major and minor bleeding through 30 days using a modified ACUITY bleeding scale (online Appendix A).28 As most bleeding events in this study are likely to be access site related, the ACUITY bleeding scale was felt to be most appropriate. Secondary outcomes include ischemic events (death, nonfatal MI, recurrent ischemia in target vessel distribution, or urgent target lesion revascularization [TLR]) (online Appendix A), level of factor IXa inhibition as measured by effect on measures of coagulation, duration of hospitalization, feasibility of early sheath removal, and clinical and laboratory markers of anticoagulant activity. A blinded clinical events committee based at the Duke Clinical Research Institute will adjudicate all suspected bleeding events, MIs, urgent TLR, and ischemic events in the target vessel distribution using standard prespecified definitions. Clinical follow-up Patient assessments, including history, physical exam, concomitant medication assessment, and safety laboratories will occur at discharge and at 30-day postrandomization. To ensure capture of any bleeding occurring after hospital discharge, patients will be contacted by telephone at 7 days.

American Heart Journal February 2011

266 Povsic et al

Data analysis and statistical assumptions As a phase II, dose-finding study, RADAR is designed to test several hypotheses. All analyses will be considered significant at a nominal α level of .05 with no adjustments for multiple comparisons. With respect to the primary end point of major or minor bleeding, with a sample size of 200 patients per arm and assuming a 20% event rate in the heparin arm, RADAR has 80% power to detect a 10% absolute reduction with RB006 with 100% reversal compared with heparin.29 Similarly, assuming a 20% event rate in the RB006 with 25% reversal arm, RADAR has 80% power to detect a 10% absolute reduction with RB006 with 100% reversal compared with 25% reversal. To assess the effect of RB006 on ischemic end points, the study will compare the 600 patients randomized to RB006 with the 200 patients randomized to heparin. Formal noninferiority testing is not feasible; however, assuming an 8% event rate in the heparin arm, RADAR will be able to exclude an absolute increase of N6.2% in the REG1 group with 80% power and an increase of 7.2% with 90% power. Adaptive design An independent Data Safety Monitoring Board (DSMB) will meet after enrollment of 100, 200, and 400 patients to assess bleeding rates. They may recommend that the Steering Committee discontinue individual reversal arms if excess bleeding is observed in the low reversal arms that is consistent with a dose-dependent effect of RB007. Major bleeding rates will be compared with the historical rate of approximately 5% observed in ACUITY. If bleeding rates in the lowest reversal arm exceed this rate with 95% confidence, the DSMB may recommend discontinuation of that arm. Under these guidelines, termination of individual arms may be considered if • ≥4 events in 25 patients (rate 16%, lower 95% CI 5.02%); • ≥7 events in 50 patients (rate 14%, lower 95% CI 6.15%); • ≥9 events in 75 patients (rate 12%, lower 95% CI 5.9%); • ≥10 events in 100 patients (rate 10%, lower 95% CI 5.12%). Should open-label 100% RB007 reversal be used at a rate of ≥25% that is consistent with a dose-dependent effect of RB007, a recommendation to discontinue a reversal arm may also be considered. If an RB007 reversal arm is dropped, additional patients will be allocated to the remaining REG1 arms.

Committees An academic steering committee (online Appendix B) composed of international representation from the

academic interventional community will provide scientific direction and assess trial progress. The steering committee is responsible for trial design and publication of major trial findings. The committee will also encourage and prioritize additional analyses for publication suggested by investigators. The DSMB (online Appendix B) will consist of 2 interventional cardiologists, an expert in coagulation, a statistician, and a nonvoting member.

Discussion RADAR is the first study to address the clinical effectiveness of a drug-controlling agent combination and offers unique challenges in design and interpretation. RADAR is designed to assess the REG1 system in patients with ACS by addressing 3 critical issues. First, to determine the safety of a range of levels of RB006 reversal with RB007 after catheterization with respect to sheath removal. Second, to confirm whether a dose of 1 mg/kg RB006 results in near-complete inhibition of factor IXa based on pharmacodynamic markers in patients with ACS. Third, to preliminarily assess the efficacy of RB006 as an anticoagulant in patients with ACS undergoing PCI. Understanding these issues will be critical to the design of subsequent phase III clinical trials of REG1 in patients with the spectrum of ACS and undergoing PCI.

Drug reversal: advantages and complexities A unique feature of REG1 is that RB006 can be actively, partially, or completely reversed using RB007. One approach to achieving a desirable balance between the prevention of ischemic events and avoiding bleeding is to use a therapy that is actively reversible, allowing modulation of the level of anticoagulation to fit the individual patient's clinical circumstances. Aptamers are unique in their ability to code for their own controlling agents, which allows for active modulation of their activity.21,24,25 Assessing the efficacy and safety of a combination product is challenging because dosing of each component must be independently evaluated. Phase I studies of REG1 established the relationship between weightbased and fixed RB006 doses and degree of factor IX inhibition, as well as the ability of RB007 to completely reverse the anticoagulant effects of RB006 in stable volunteers and patients with CAD 21,24 ; phase Ic determined the relationship between RB007 dose and the degree of RB006 reversal.25 RB006 dosing in ACS The dose of RB006 being studied in RADAR is based on preclinical and early clinical data.20,21,24,25 The dose of 1 mg/kg that was chosen is expected to achieve consistent and near-complete factor IX inhibition. Given that RB006

American Heart Journal Volume 161, Number 2

is a selective inhibitor of factor IX, this dose is expected to be maximally or near maximally effective. Investigation of a high dose of RB006 is facilitated by the availability of RB007-mediated reversal should clinical bleeding occur and, at least in part, mitigates the need for extensive dose ranging for RB006. The PK/PD substudy will confirm that the dose of RB006 investigated in RADAR achieves consistent and near-complete inhibition of factor IX in patients with ACS. Finally, RADAR will begin to assess the efficacy of factor IX inhibition in patients with ACS undergoing PCI by assessing rates of ischemic events both during PCI and after RB007 reversal in patients treated with RB006. Event rates with RB006 will be compared with those of patients treated with heparin with or without a GP IIb/IIIa inhibitor in RADAR and from historical data.

RB007 reversal A key objective in RADAR is to determine the range of RB006 reversal with RB007 that is clinically feasible after cardiac catheterization. Unlike other anticoagulants, REG1 offers the possibility of tailoring the anticoagulation to the clinical circumstances of the patient. Because REG1 allows precise modulation of the level of antithrombotic therapy during the course of treatment, RADAR will be the first study to explore the consequence of variable levels of anticoagulation in the postcatheterization/PCI period. RADAR has statistical power to detect a 10% absolute difference in major or minor bleeding; an assessment of bleeding rates across all reversal arms will inform the minimal RB007 dose that will allow early sheath removal without excessive bleeding. Because patients with hemophilia B exhibit clinically evident bleeding only after losing N90% of factor IX activity, it is apparent that even partial RB006 reversal may permit arterial sheath removal. As would happen in practice, if patients have residual or recurrent bleeding, an additional dose of RB007 sufficient to fully reverse RB006 can be given. Blinding of RB006 study drug and study monitoring Given the different modes of administration, the timing of aPTT monitoring, and the timing of reversal for REG1, it is logistically challenging to fully blind patients and study personnel to treatment assignment. Therefore, a partially blinded design was chosen with open-label assignment to REG1 versus heparin and blinded randomization to the degree of reversal in REG1-treated patients. Required femoral access We required femoral access in RADAR to reduce the heterogeneity in groin management and procedurerelated bleeding. Both femoral closure and compression devices are allowed during the study. Although VCDs have been shown to reduce time to ambulation, registry

Povsic et al 267

data and meta-analyses of randomized studies are inconsistent with respect to demonstrating reductions in bleeding complications with VCDs as compared with manual compression.30-36

Conclusions RADAR is a unique phase II clinical trial investigating a combined anticoagulant thrombotic and complementary active reversal agent. The data from RADAR will answer important questions about each component of the REG1 anticoagulation system and define the best strategy to support the development of adequately powered phase III clinical trials.

Disclosures Funding: RADAR is funded by Regado Biosciences (Basking Ridge, NJ).

References 1. Harrington RA, Becker RC, Cannon CP, et al. Antithrombotic therapy for non–ST-segment elevation acute coronary syndromes: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2008;133:670S-707S. 2. Stone GW, Witzenbichler B, Guagliumi G, et al. Bivalirudin during primary PCI in acute myocardial infarction. N Engl J Med 2008;358: 2218-30. 3. Rao SV, O'Grady K, Pieper KS, et al. Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes. Am J Cardiol 2005;96:1200-6. 4. Rao SV, O'Grady K, Pieper KS, et al. A comparison of the clinical impact of bleeding measured by two different classifications among patients with acute coronary syndromes. J Am Coll Cardiol 2006;47: 809-16. 5. Rao SV, Kaul PR, Liao L, et al. Association between bleeding, blood transfusion, and costs among patients with non–ST-segment elevation acute coronary syndromes. Am Heart J 2008;155:369-74. 6. Rao SV, Jollis JG, Harrington RA, et al. Relationship of blood transfusion and clinical outcomes in patients with acute coronary syndromes. JAMA 2004;292:1555-62. 7. Hirsh J, O'Donnell M, Eikelboom JW. Beyond unfractionated heparin and warfarin: current and future advances. Circulation 2007;116: 552-60. 8. Kimmel S, Sekeres M, Berlin J, et al. Adverse events after protamine administration in patients undergoing cardiopulmonary bypass: risks and predictors of under-reporting. J Clin Epidemiol 1998;51: 1-10. 9. Seifert H, Jobes D, Ten Have T, et al. Adverse events after protamine administration following cardiopulmonary bypass in infants and children. Anesth Analg 2003;97:383-9. 10. Weiler JM, Gellhaus MA, Carter JG, et al. A prospective study of the risk of an immediate adverse reaction to protamine sulfate during cardiopulmonary bypass surgery. J Allergy Clin Immunol 1990;85: 713-9. 11. Collett JP, Montalescot G, Lison L, et al. Percutaneous coronary intervention after subcutaneous enoxaparin pretreatment in patients with unstable angina pectoris. Circulation 2001;103:658-63. 12. Ferguson JJ, Califf RM, Antman EM, et al. Enoxaparin vs unfractionated heparin in high-risk patients with non–ST-segment

268 Povsic et al

13.

14.

15.

16. 17.

18.

19.

20. 21.

22. 23. 24.

elevation acute coronary syndromes managed with an intended early invasive strategy: primary results of the SYNERGY randomized trial. JAMA 2004;292:45-54. Stone GW, Ware JH, Bertrand ME, et al. Antithrombotic strategies in patients with acute coronary syndromes undergoing early invasive management: one-year results from the ACUITY trial. JAMA 2007; 298:2497-506. Fifth Organization to Assess Strategies in Acute Ischemic Syndromes Investigators. Comparison of fondaparinux and enoxaparin in acute coronary syndromes. N Engl J Med 2006;354:1464-76. Kjalke M, Monroe DM, Hoffman M, et al. Active site-inactivated factors VIIa, Xa, and IXa inhibit individual steps in a cell-based model of tissue factor-initiated coagulation. Thromb Haemost 1998;80: 578-84. Mann K, Brummel K, Butenas S. What is all that thrombin for? J Thromb Haemost 2003;1:1504-14. Monroe DM, Hoffman M, Roberts HR. Transmission of a procoagulant signal from tissue factor-bearing cells to platelets. Blood Coagul Fibrinolysis 1996;7:459-64. Howard EL, Becker KC, Rusconi CP, et al. Factor IXa inhibitors as novel anticoagulants. Arterioscler Thromb Vasc Biol 2007;27: 722-7. Zelenkofske SL, Rusconi CP, Damiento CM, et al. Subcutaneous RB006, a direct FIXa inhibitor, provides prolonged anticoagulation with rapid reversal: the first clinical experience with the REG2 system. Poster presented at: Arteriosclerosis, Thrombosis and Vascular Biology 2010 Scientific Session Sessions; April 9, 2010; San Francisco, CA; 2010. Rusconi CP, Scardino E, Layzer J, et al. RNA aptamers as reversible antagonists of coagulation factor IXa. Nature 2002;419:90-4. Dyke CK, Steinhubl SR, Kleiman NS, et al. First-in-human experience of an antidote-controlled anticoagulant using RNA aptamer technology: a phase 1a pharmacodynamic evaluation of a drug-antidote pair for the controlled regulation of factor IXa activity. Circulation 2006;114:2490-7. Gold L, Polisky B, Uhlenbeck O, et al. Diversity of oligonucleotide functions. Ann Rev Biochem 1995;64:763-97. Nimjee SM, Rusconi CP, Sullenger BA. APTAMERS An emerging class of therapeutics. Ann Rev Med 2005;56:555-83. Chan MY, Cohen MG, Dyke CK, et al. Phase 1b randomized study of antidote-controlled modulation of factor IXa activity in patients with stable coronary artery disease. Circulation 2008;117:2865-74.

American Heart Journal February 2011

25. Chan MY, Rusconi CP, Alexander JH, et al. A randomized, repeatdose, pharmacodynamic and safety study of an antidote-controlled factor IXa inhibitor. J Thromb Haemost 2008;6:789-96. 26. Cohen MG, Purdy DA, Rossi JS, et al. First clinical application of an actively reversible direct factor IXa inhibitor as an anticoagulation strategy in patients undergoing percutaneous coronary intervention. Circulation 2010;122:614-22. 27. Thygesen K, Alpert JS, White HD. Universal definition of myocardial infarction. J Am Coll Cardiol 2007;50:2173-95. 28. Stone GW, Bertrand M, Colombo A, et al. Acute Catheterization and Urgent Intervention Triage strategY (ACUITY) trial: study design and rationale. Am Heart J 2004;148:764-75. 29. Lincoff AM, Kleiman NS, Kereiakes DJ, et al. Long-term efficacy of bivalirudin and provisional glycoprotein IIb/IIIa blockade vs heparin and planned glycoprotein IIb/IIIa blockade during percutaneous coronary revascularization: REPLACE-2 randomized trial. JAMA 2004;292:696-703. 30. Biancari F, D'Andrea V, Di Marco C, et al. Meta-analysis of randomized trials on the efficacy of vascular closure devices after diagnostic angiography and angioplasty. Am Heart J 2010;159: 518-31. 31. Vaitkus PT. A meta-analysis of percutaneous vascular closure devices after diagnostic catheterization and percutaneous coronary intervention. J Invasive Cardiol 2004;16:243-6. 32. Koreny M, Riedmüller E, Nikfardjam M, et al. Arterial puncture closing devices compared with standard manual compression after cardiac catheterization: systematic review and meta-analysis. JAMA 2004;291:350-7. 33. Sanborn TA, Ebrahimi R, Manoukian SV, et al. Impact of femoral vascular closure devices and antithrombotic therapy on access site bleeding in acute coronary syndromes: The ACUITY Trial. Circ Cardiovasc Interv 2010;3:57-62. 34. Marso JP, Amin AP, House JA, et al. Association between use of bleeding avoidance strategies and risk of periprocedural bleeding among patients undergoing percutaneous coronary intervention. JAMA 2010;303:2156-64. 35. Dauerman HL, Applegate RJ, Cohen DJ. Vascular closure devices: the second decade. J Am Coll Cardiol 2007;50:1617-26. 36. Arora N, Matheny ME, Sepke C, et al. A propensity analysis of the risk of vascular complications after cardiac catheterization procedures with the use of vascular closure devices. Am Heart J 2007;153:606-11.

American Heart Journal Volume 161, Number 2

Appendix A. End point definitions Bleeding Major bleeding is defined as intracranial, intraocular, or retroperitoneal hemorrhage; access site hemorrhage requiring radiologic or surgical intervention; 5-cm diameter hematoma at puncture site; clinically overt blood loss resulting in a decrease in hemoglobin concentration of N3 g/dL; any decrease in hemoglobin concentration of N4 g/ dL; reoperation for bleeding; use of blood product transfusion; and hemarthrosis. Minor bleeding is defined as clinically overt bleeding that does not meet criteria for major bleeding. Ischemic end points Ischemic events will be defined as all deaths, regardless of cause, nonfatal MI, urgent TLR, or recurrent ischemia in the target vessel distribution. These are defined below. Myocardial infarction. Myocardial infarction will be defined based on occurrence in the setting of recent revascularization (CABG or PCI) and its relation to recent cardiac markers. (A) No recent MI, no recent revascularization in the previous 24 hours: any elevation of troponin I or T above the ULN, OR any elevation of CK-MB above the ULN or total CK above 2 × ULN (if no troponin data available), or new significant (N0.04 seconds) Q waves in ≥2 contiguous leads, or clinical evidence of ischemia (N20 minutes) with new or recurrent ST-segment elevation of N0.1 mV in 2 contiguous limb leads or N0.2 mV in 2 contiguous precordial leads prompting urgent cardiac catheterization AND resulting in documented poor flow (Thrombolysis In Myocardial Infarction b3) in the infarct-related artery prompting urgent intervention (PCI or CABG). (B) Baseline MI present, but no recent revascularization (within 24 hours): clinical evidence of recurrent ischemic pain (N20 minutes) with new or recurrent ST-segment elevation N0.1 mV in 2 contiguous limb leads or N0.2 mV in 2 contiguous precordial leads prompting urgent cardiac catheterization AND resulting in documentation poor flow (TIMI b3) in the infarct-related artery prompting urgent intervention (PCI or CABG), or new significant (≥ 0.04 seconds) Q waves in ≥ 2 contiguous leads and discrete from enrollment MI, or cardiac markers of necrosis or electrocardiographic (ECG) evidence as follows: i. re-elevation of troponin I or T NULN (if prior level was normal); or ii. re-elevation of troponin I or T NULN and N50% above the nadir prior level (if prior level was above normal); or iii. re-elevation of CK-MB to ≥ULN (if prior level was normal); or iv. CK-MB ≥ ULN and N50% above the prior nadir level (if prior level was above normal); or

Povsic et al 268.e1

v. total CK ≥ 2 × ULN and increased by ≥50% over the previous nadir value (if CK-MB is unavailable). (C) No recent MI prior and recent revascularization: for patients undergoing PCI, CK-MB (or total CK, if CK-MB is unavailable) ≥ 3 × ULN. In the absence of CK-MB and total CK data, the troponin value must exceed N3 × ULN or new significant (N0.04 seconds) Q waves in 2 contiguous ECG leads. For patients undergoing CABG, CK-MB (or total CK, if CK-MB is unavailable) ≥ 5 × ULN. In the absence of CKMB and total CK data, the troponin value must exceed N5 × ULN or new significant (N0.04 seconds) Q waves in 2 contiguous ECG leads. (D) For patients in whom the cardiac markers of necrosis are elevated before PCI, an end point MI is defined as follows: CK-MB (or total CK, if CK-MB is unavailable) ≥ 3 × ULN and increased by at least 20% from the level before the procedure or new significant (N0.04 seconds) Q waves in 2 contiguous ECG leads. In the absence of CK-MB and total CK data, troponin value must exceed 3 × ULN and represent a N50% increase in the nadir preprocedural troponin on 2 consecutive assessments. For patients in whom the cardiac markers of necrosis are elevated before CABG, an end point MI is defined as follows: CK-MB (or total CK, if CK-MB is unavailable) ≥ 5 × ULN and increased by at least 20% from level before the procedure. In the absence of CK-MB and total CK data, the troponin value must exceed 5 × ULN and represent a N50% increase in the nadir preprocedural troponin on 2 consecutive assessments or new significant (N0.04 seconds) Q waves in 2 contiguous ECG leads. Urgent revascularization and urgent TLR. Any revascularization procedure (PCI or CABG) performed because of a clinical event that occurs after randomization will constitute an urgent revascularization. If urgent revascularization occurs at a previously treated site (PCI or CABG) as part of RADAR, this will constitute a TLR. Recurrent myocardial ischemia in target vessel distribution. Recurrent ischemia in target vessel distribution will include any symptoms of chest discomfort or equivalent ischemic symptoms of ≥10 minutes' duration not fulfilling criteria for an MI but prompting medical intervention (eg, thrombolytics, GP IIb/IIIa inhibitors, anticoagulation, prolonged hospitalization, or additional medical procedures, etc) and with clear documentation that the ischemia originates from the target or culprit vessel(s). The culprit lesion(s) should be identified by imaging studies (eg, cardiac catheterization, radionucleotide study, echocardiographic wall motion study, etc). This end point is designed to capture recurrent events in the target vessel requiring additional medical or procedural intervention (ie, intracoronary lytics and additional systemic GP IIb/IIIa inhibition) but not

268.e2 Povsic et al

resulting in MI or revascularization caused by ischemic or thrombotic complications.

American Heart Journal February 2011

Appendix B. Committee members

MD, Duke Clinical Research Institute; Jan Cornel, MD, Medisch Centrum Alkmaar; Gilles Montalescot, MD, PitiéSalpétrière Hospital; Jaroslav Kasprzak, MD, Medical University of Lodz; Steven Zelenkofske, MD, Regado Biosciences.

Academic Steering Committee John H. Alexander, MD (chair), Duke Clinical Research Institute; Roxana Mehran, MD, Mount Sinai Medical Center; Christoph Bode, MD, University of Freiberg; Mauricio G. Cohen, MD, University of Miami; Christopher Buller, MD, Hamilton General Hospital; Richard Becker, MD, Duke Clinical Research Institute; Thomas Povsic,

Data Safety Monitoring Board Ronald Waksman, MD, Washington Heart Center (chair); Stefan James, MD, Uppsala Clinical Research Center; Jack Ansel, MD, Lenox Hill Hospital; Vic Hasselblad, PhD, Duke Clinical Research Institute; and Rebecca Torguson, MPH, Washington Heart Center.

Associations between cardiovascular parameters and uteroplacental Doppler (blood) flow patterns during pregnancy in women with congenital heart disease: Rationale and design of the Zwangerschap bij Aangeboren Hartafwijking (ZAHARA) II study Ali Balci, MD, MSc, a,j,k Krystyna M. Sollie, MD, b,k Barbara J. M. Mulder, MD, PhD, d,k Monique W. M. de Laat, MD, PhD, e Jolien W. Roos-Hesselink, MD, PhD, f Arie P. J. van Dijk, MD, PhD, g Elly M. C. J. Wajon, MD, h Hubert W. Vliegen, MD, PhD, i Willem Drenthen, MD, PhD, a Hans L. Hillege, MD, PhD, a,c Jan G. Aarnoudse, MD, PhD, b Dirk J. van Veldhuisen, MD, PhD, a and Petronella G. Pieper, MD, PhD a Groningen, Amsterdam, Rotterdam, Nijmegen, Enschede, Leiden, and Utrecht, The Netherlands

Background Previous research has shown that women with congenital heart disease (CHD) are more susceptible to cardiovascular, obstetric, and offspring events. The causative pathophysiologic mechanisms are incompletely understood. Inadequate uteroplacental circulation is an important denominator in adverse obstetric events and offspring outcome. The relation between cardiac function and uteroplacental perfusion has not been investigated in women with CHD. Moreover, the effects of physiologic changes on pregnancy-related events are unknown. In addition, long-term effects of pregnancy on cardiac function and exercise capacity are scarce. Methods Zwangerschap bij Aangeboren Hartafwijking (ZAHARA) II, a prospective multicenter cohort study, investigates changes in and relations between cardiovascular parameters and uteroplacental Doppler flow patterns during pregnancy in women with CHD compared to matched healthy controls. The relation between cardiovascular parameters and uteroplacental Doppler flow patterns and the occurrence of cardiac, obstetric, and offspring events will be investigated. At 20 and 32 weeks of gestation, clinical, neurohumoral, and echocardiographic evaluation and fetal growth together with Doppler flow measurements in fetal and maternal circulation are performed. Maternal evaluation is repeated 1 year postpartum. Implications By identifying the factors responsible for pregnancy-related events in women with CHD, risk stratification can be refined, which may lead to better pre-pregnancy counseling and eventually improve treatment of these women. (Am Heart J 2011;161:269-275.e1.)

Because of improved long-term survival, most women with congenital heart disease (CHD) reach child-bearing age and many pursue pregnancy. In women with uncorrected maternal congenital heart defects or with

residual sequelae after correction, the hemodynamic changes in pregnancy can have negative effects on the health of both mother and her (unborn) child. Cardiac events are rare in healthy women (b1%), while arrhythmias occur in 4.5% and heart failure in 4.8% of women with CHD.1 In complex CHD, cardiac event

From the aDepartment of Cardiology, University Medical Centre Groningen, University of Groningen, Groningen, The Netherlands, bDepartment of Obstetrics, University Medical Centre Groningen, University of Groningen, Groningen, The Netherlands, cDepartment of Epidemiology, University Medical Centre Groningen, University of Groningen, Groningen, The Netherlands, dDepartment of Cardiology, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands, eDepartment of Obstetrics, Academic Medical

k On behalf of the ZAHARA-II investigators. See the online Appendix for complete listing. This study is supported by a grant from the Netherlands Heart Foundation (NHF) (2007B75). DJvV is clinically established investigator of the NHF (D97-017). The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents. Submitted August 26, 2010; accepted October 18, 2010.

Centre, University of Amsterdam, Amsterdam, The Netherlands, fDepartment of Cardiology, Erasmus Medical Centre, Erasmus University, Rotterdam, The Netherlands, g Department of Cardiology, Radboud University Nijmegen Medical Centre, Radboud University Nijmegen, Nijmegen, The Netherlands, hDepartment of Cardiology, Medical Spectrum Twente, Enschede, The Netherlands, iDepartment of Cardiology, Leiden University Medical Centre, University of Leiden, Leiden, the Netherlands, and jInteruniversity Cardiology Institute of the Netherlands (ICIN), Utrecht, The Netherlands.

Reprint requests: Petronella G. Pieper, MD, PhD, Department of Cardiology, University Medical Centre Groningen, University of Groningen, PO Box 30.001, 9700 RB, Groningen, The Netherlands. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.024

Background

American Heart Journal February 2011

270 Balci et al

rate can be even higher.1-4 Women with CHD are not only more susceptible to cardiac events, but obstetric and offspring events are also more prevalent.1-3 Important obstetric events are postpartum hemorrhage (8%, up to 29%), pregnancy-induced hypertension (5.5%, up to 13%), preeclampsia (PE; 3.2%, up to 10%), and preterm delivery (16%, up to 65%); whereas in healthy pregnant women, the prevalence of these events is much lower.1 Frequently observed events in offspring of women with CHD are intrauterine growth restriction (IUGR), prematurity, and mortality.1-3 The magnitude of these risks depends, at least in part, on the type and severity of maternal CHD. In the general population, both IUGR and PE are associated with a lower cardiac output (CO), an elevated total vascular resistance (TVR), and abnormal uterine and umbilical artery Doppler waveform patterns.5,6 These parameters can be used to identify women at risk for PE and IUGR. The higher incidence of these events in women with CHD may be caused by inadequate maternal hemodynamics, resulting in insufficient uteroplacental circulation.7,8 Nevertheless, the interaction between echocardiographic, hemodynamic, and neurohumoral parameters on one side and uteroplacental Doppler flow patterns on the other side in relation to pregnancy outcome have not been studied in women with CHD and have not been compared to those in healthy pregnant women. In addition, although some studies identified predictors of cardiac events in women with both acquired and CHD, prospective data relating echocardiographic parameters of ventricular or valvular function to maternal cardiac events are still scarce and little is known about the long-term effects of pregnancy on cardiac function or exercise capacity in women with CHD.2,3,9 The mid- and longterm effects of pregnancy on cardiovascular hemodynamics have been described in a few select subgroups. 10,11 In these studies, however, exercise capacity or cardiac biomarkers were not assessed. In this article, we will introduce the study design and describe the rationale of the ZAHARA II study.

Methods Study objectives The primary objective of the present study is to compare cardiovascular, neurohumoral, and uteroplacental Doppler flow changes during pregnancies of women with CHD with age- and parity-matched healthy controls and to relate these changes to the occurrence of cardiovascular and obstetric events and to offspring outcome. The secondary objective of this study is to evaluate the incidence of permanent postpartum cardiovascular deterioration in women with CHD.

Study design This is an observational prospective multicenter cohort study.

Table I. Inclusion and exclusion criteria of ZAHARA II Women with CHD Inclusion criteria - Age ≥18 y - Morphological CHD - Presentation at ≤20 wk of gestation - Presentation in one of the participating medical centers Exclusion criteria - Miscarriage or termination of pregnancy b20 wk of gestation - Alcohol abuse - Illicit drugs use Healthy controls Inclusion criteria - Age ≥18 y - Presentation at ≤20 wk of gestation Exclusion criteria - Miscarriage or termination of pregnancy b20 wk of gestation - Women who are on chronic medication - Women who are under specialist control - Alcohol abuse - Illicit drugs use

Study population Women with any morphological CHD with a pregnancy of b20 weeks duration, presenting in the participating centers, who meet all the inclusion and none of the exclusion criteria are eligible (Table I). During a 3-year period, a minimum of 160 women with CHD are enrolled, and simultaneously, 60 healthy, age- and parity-matched women are recruited from a low-risk midwife practice in Groningen and in Rotterdam, the Netherlands, to serve as controls. We will use a subclassification for our cohort where appropriate, using division in disease complexity, the adapted World Health Organization classification for estimating pregnancy risk, or clustering of the morphological and functional comparable diseases as in previous studies.2,9,12,13

Measurements Baseline data are recorded at the first prenatal visit using medical records and include underlying heart disease, prior interventions, cardiac sequelae, prior cardiac events, comorbidity, and obstetric history. Maternal age, parity, present cardiac status (including New York Heart Association functional class, physical examination, oxygen saturation, and echocardiographic data), use of medication, intoxications, educational status, and current employment are also recorded. Clinical evaluation at gestational weeks 20 and 32 as well as at 1 year postpartum is performed for follow-up data and for registration of events. During follow-up echocardiograms, electrocardiograms and 24hour electrocardiographic registrations as well as obstetric evaluation and blood and urinalysis are conducted.

Echocardiography Standardized echocardiograms according to disease-specific protocols are performed at 20 and 32 weeks of gestation and at 1 year postpartum. Echocardiograms are evaluated off-line in the University Medical Center Groningen, Groningen, the Netherlands. Morphological left and right ventricular size and function (if feasible, ejection fraction according to Simpson's

American Heart Journal Volume 161, Number 2

Balci et al 271

Table II. Adverse events as defined in ZAHARA II Primary cardiovascular events during pregnancy, up to 6 m postpartum Need for urgent invasive cardiovascular procedures Heart failure: according to the guidelines of the European Society of Cardiology and documented by the attending physician40 Any documented new-onset or symptomatic tachy- or bradyarrhythmia requiring new/extended treatment Thromboembolic events: deep vein thrombosis, pulmonary embolism, intracardiac thrombosis, arterial thrombosis, systemic arterial embolisms, or transient ischemic attack Myocardial infarction Cardiac arrest Cardiac death Endocarditis: according to the Duke criteria41 Aortic dissection Secondary cardiovascular events during pregnancy, up to 6 m postpartum NYHA class deterioration: decline of 2 points in NYHA functional class during pregnancy or within 6 m postpartum, compared with pre-pregnancy NYHA class or a persisting deterioration in NYHA functional class postpartum Primary obstetric events Assisted delivery: use of forceps, use of a vacuum extractor, or the performance of a cesarean section for delivery on maternal cardiac and medical indication, on maternal obstetric indication, or on fetal indication PIH: systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg or an increase of systolic (N30 mm Hg) or diastolic (N15 mm Hg) blood pressure, in the absence of proteinuria, occurring after ≥20 wk of gestation PE: PIH with ≥0.3 g/24 h proteinuria Eclampsia: PE with grand mal seizures Mild gestational diabetes mellitus: a fasting blood glucose b140 mg/dL (7.8 mmol/L) and 2-h postprandial 140-198 mg/dL (7.8-11 mmol/L) Severe gestational diabetes mellitus: a random serum glucose value N200 mg/dL (11.1 mmol/L) or a fasting blood glucose value N126 mg/dL (7.0 mmol/L) HELLP: hemolysis (LDH N 250 U/L), elevated liver enzymes (ASAT N40 U/L and ALAT N45 U/L), low platelets (b1.0 × 10.6/mm3) syndrome Hyperemesis gravidarum: severe, intractable nausea and vomiting, leading to dehydration, loss of weight, metabolic disorders, and hospitalization Noncardiac death: all cause mortality, except cardiac mortality Postpartum hemorrhage: blood loss N500 mL (vaginal delivery) or N1000 mL (cesarean section), requiring transfusion or leading to a drop in hemoglobin N20 g/L (1.24 mmol/L) Premature labor: spontaneous onset of labor b37 wk of gestation Preterm premature rupture of membranes: spontaneous rupture of membrane before the onset of uterine contractions and before 37 wk of gestation Abruptio placentae: premature detachment of the placenta from the wall of the uterus Secondary obstetric events Amniotomy: mechanical/artificial rupture of membranes Induction of labor Prolongation of cervix ripening: omitted dilatation of the portio vaginalis during ≥20 h (nullipara) or ≥14 h (multipara), despite adequate and regular uterine contractions Prolongation of second stage of delivery (primipara N2 h; multipara N1 h) Placenta previa: localization of the placenta partially or completely above the internal ostium of the cervix Offspring events Fetal death: intrauterine death N20 wk of gestation Extended perinatal death: the number of stillbirths from 20 wk of gestation and neonatal death up to 28 d postpartum Early neonatal death: within 6 d after birth Late neonatal death: within 7-28 days after birth Infant death: N28 d and within 1 y after birth Perinatal death: the total number of stillbirths from 20 wk of gestation and death up to 7 days postpartum Offspring death: the total number of stillbirths from 20 wk of gestation and death up to 1 y postpartum Intraventricular hemorrhage: bleeding in the fetal cerebral ventricles Neonatal respiratory distress syndrome: respiratory insufficiency caused by a developmental insufficiency of surfactant production and structural immaturity in the lungs in premature infants Infections leading to hospital admission Premature birth: birth b37 wk gestation Occurrence of CHD Occurrence of other congenital disease Small for gestational age: birth weight below the 10th percentile adjusted for gestational age and based on population values Low birth weight: birth weight b2500 g Meconium stained amniotic fluid General events Anemia: between 18 wk of gestation until 1 wk postpartum: 6.5 mmol/L or 10.5 g/dL42 Hospitalization: all cause hospitalization for more than 1 night Fever: ≥38.5°C during pregnancy up to 6 m postpartum requiring medical treatment Infection: infections during pregnancy up to 6 m postpartum requiring medical treatment ALAT, Alanine aminotransferase; ASAT, aspartate aminotransferase; LDH, lactate dehydrogenase; NYHA, New York Heart Association; PIH, pregnancy-induced hypertension.

American Heart Journal February 2011

272 Balci et al

rule), atrial size, valvular function (quantification of regurgitation and stenosis of all valves) as well as disease-specific evaluation (ie, presence and location of intracardiac shunts, evaluation of conduits, or baffles) are performed according to current recommendations and guidelines.14-20 Blood pressure and heart rate are measured during echocardiography to calculate CO and TVR as described previously.15,17,19 Prepregnancy routinely performed echocardiograms (available in most of the patients) are analyzed for comparison.

Obstetric evaluation Fetal biometry is assessed by ultrasound and uteroplacental perfusion is studied by Doppler flow measurements of uterine arteries. Umbilical artery pulsatility index, resistance index, and uterine artery flow and early diastolic notching are evaluated at 20 and 32 weeks of pregnancy according to the guidelines of the International Perinatal Doppler Society.5,21 Evaluation of the digitally stored (Doppler) ultrasound registrations is performed in the University Medical Center Groningen.

Blood and urinalysis Hematologic parameters, renal and hepatic function, hemoglobin A1c as well as nonfasting glucose, and N-terminal prohormone brain natriuretic peptide (NT-proBNP) are assessed from blood samples at 20 and 32 weeks of gestation and 1 year postpartum and compared with the values before pregnancy, if available. Proteinuria is quantitatively assessed at 20 and 32 weeks of gestation.

Cardiopulmonary aerobic capacity testing Cardiopulmonary aerobic capacity testing is performed 1 year postpartum in patients who underwent this test b2 years before pregnancy.22

Adverse events Clinical adverse events in ZAHARA II are subdivided in cardiac, obstetric, offspring, and general adverse events (Table II). We define cardiovascular events as described in previous studies (Table II). In addition, we assess changes in NTpro-BNP levels and in echocardiographic findings in patients with CHD compared to those in healthy controls. We define abnormal echocardiographic changes in pregnancies as a significant deterioration in size or function of subpulmonary or sub-aortic ventricle; new onset or aggravation of valve regurgitation ≥1 grade (mild to moderate or severe, or moderate to severe) during pregnancy and/or persisting 1 year postpartum; persistent (≥1 year) significant aggravation of valve stenosis (mild to moderate or severe, or moderate to severe); significant increase in aortic dimensions (≥5 mm) during pregnancy and/or 1 year postpartum. Furthermore, pulsatility index N95th centile in umbilical artery, uterine artery resistance index N95th centile, umbilical artery resistance index N0.58 or N90th centile, or early diastolic notching in uterine artery is considered abnormal in obstetric ultrasound evaluation.23 Finally, a significant deterioration in functional capacity and/or exercise capacity 1 year postpartum in all patients compared with pre-pregnancy values is considered abnormal.

Statistical and ethical considerations Sample size calculation. One of the primary aims of the ZAHARA II study is to compare the uteroplacental Doppler flow, expressed as pulsatility index in the umbilical artery, during pregnancy between women with CHD and healthy controls. A sample size of 160 patients and 60 controls achieves a power of 80% at a significance level of .05 to detect a difference of 5% higher mean pulsatility index in women with CHD. Statistical analysis. Continuous variables with normal distribution will be presented as mean (±SD), nonnormally distributed variables as median (with 25th and 75th percentile), and dichotomous variables will be presented as absolute numbers and percentages. Comparison of continuous variables between groups will be made by independent t tests or the Mann-Whitney U test, depending on their distribution. For the comparison of dichotomous variables, we will use the χ2 test or Fisher exact test, where applicable. Uni- and multivariable logistic regression analyses will be performed to identify predictors of adverse pregnancy outcome. For the analysis of repeated measures within the women with CHD, between the women with CHD, as well as between the women with CHD and the healthy controls, we will use multivariate general linear models. Ethical considerations. The study is conducted according to the principles of the Declaration of Helsinki and in accordance with the Medical Research Involving Human Subjects Act (Wet medisch-wetenschappelijk onderzoek met mensen). The study design, all research aims, and the specific measurements in the ZAHARA II Study have been approved by the medical ethical committee of all participating hospitals. New measurements will only be embedded in the study after approval of the medical ethical committee. All participants are asked for their written informed consent after having received written and oral information about the study. Data management and privacy protection. Data are directly entered onto written case record forms (CRF) and manually entered into an electronic database by 2 researchers. Random samples of all entered data are double checked by other research members to monitor the quality of this manual data entry process. Open text fields are copied into the electronic database exactly as they are filled in on the CRF. All measurements will be checked by examination of the data including their ranges, distributions, means, SD, outliers, and logical errors. Data outliers and missing values will be checked on the original CRF. All information in these data sets that enables identification of a participant will be excluded. The CRF as well as the data sets include subject unique identification numbers that enable feedback about one subject to the data manager but do not enable identification of that particular subject.

Discussion In the present study, we assess whether changes in cardiovascular, hemodynamic, neurohumoral parameters, and uteroplacental Doppler flow patterns during pregnancy of women with CHD differ from age- and parity-matched healthy controls. We also assess the interaction of these changes with the occurrence of cardiovascular, obstetric, and offspring events. In addition, we evaluate the incidence of permanent changes

American Heart Journal Volume 161, Number 2

in cardiovascular parameters 1 year postpartum in women with CHD and compare these with matched healthy controls. During normal pregnancy, considerable hemodynamic changes occur.24 Total vascular resistance decreases, mainly due to peripheral arterial vasodilatation, mediated by progesterone and vasodilators such as nitric oxide as well as a low-resistance flow in the uteroplacental circulation.25,26 In normal pregnancy, a gradual widening of the maternal spiral arteries occurs early in the first trimester due to the invasion of endovascular and interstitial trophoblasts that convert maternal spiral arteries closer to the intervillous space. As a consequence, plasma volume, heart rate, and CO gradually increase in the first 2 trimesters of pregnancy to maintain adequate organ and uteroplacental perfusion while TVR decreases.24,26 Blood pressure drops in the first trimester of pregnancy, reaching its lowest point at the end of the second trimester, around 20 weeks of gestation, and returns to near pre-pregnancy levels around term.24,27 Together with the rise in plasma volume, the CO rises, which is reflected by an increase in both atrial and ventricular volumes, ventricular wall thickness, and the rise of heart rate.24 In normal pregnancy, an increased NT-proBNP level can be measured.28 Cardiovascular hemodynamic state returns back to pre-pregnancy state in most women within 6 months postpartum.24,29 In women with CHD, hemodynamic changes in pregnancy can exceed the compensatory possibilities of their compromised circulation, resulting in cardiac complications such as heart failure, arrhythmias, and other cardiovascular events. Known predictors of maternal cardiovascular events are related to underlying disease as well as to pre-pregnancy hemodynamic and functional status.2,3,9 However, detailed information about changes in ventricular and valvular function as well as in CO and TVR and their relationship to the occurrence of cardiovascular events in pregnancy is not available. Echocardiographic evaluation before, during, and after pregnancy in our population with CHD and in matched healthy pregnant controls will clarify some of the confusion. NT-proBNP is a well-known marker of heart failure severity.30,31 In women with CHD, NT-proBNP has been incompletely evaluated, even outside pregnancy. However, it has been shown that NT-proBNP correlates positively with New York Heart Association functional class deterioration as well as with cyanosis and inversely with ventricular ejection fraction, even in asymptomatic women with CHD.32,33 As these parameters predict pregnancy complications in women with CHD, it is plausible to hypothesize that NT-proBNP may be an easy and useful method for stratification of cardiovascular risk in pregnancy.1-3,9 Moreover, NT-proBNP levels are divergent in hypertensive disorders of pregnancy, with a graded increase from normal pregnancy to

Balci et al 273

gestational hypertension and PE.34 The frequency of these disorders is increased in women with CHD, and NTproBNP levels may provide further insight in the relation of cardiovascular status and the occurrence of these complications.4,35 Hypertensive disorders of pregnancy, especially PE, are characterized by impaired trophoblast invasion and failure of dilatation of spiral arteries, resulting in high uteroplacental vascular resistance, leading to inadequate uteroplacental perfusion and adverse obstetric and offspring outcome, including IUGR and PE.6,7 Compromised uterine perfusion with placental dysfunction is reflected by abnormal uteroplacental Doppler waveform patterns.5 Abnormal uterine artery pulsatility index and notching in early diastolic phase predicts PE, is related to high TVR and low CO, and is associated with IUGR and small for gestational age.5 In CHD, a higher incidence of PE and IUGR is seen.1-4,35,36 This may be related to a lower CO and higher incidence of heart failure in these women that may lead to inadequate uteroplacental perfusion. Therefore, uteroplacental flow patterns may differ in women with CHD compared to healthy women and abnormal uteroplacental perfusion may be associated with cardiovascular status as well as with offspring outcome.

Late effects of pregnancy in CHD Long-term effects of pregnancy in women with CHD are incompletely described. In a cohort of women with aortic stenosis, requirement of cardiac intervention was the most important late cardiac event after pregnancy.10 In a small group of women after Mustard correction, permanent deterioration in functional class, dilatation of right ventricle, right ventricular dysfunction, and tricuspid valve regurgitation have been demonstrated.11 In women with atrioventricular septal defect, persisting deterioration of atrioventricular valve regurgitation was found as well as persisting deterioration of functional class.37 Unfortunately, most data are retrospective, and for many defects, no data on mid- and late-term outcome after pregnancy are available. In the present study, echocardiographic and clinical follow-up will be at least 1 year post pregnancy for all women with CHD. In normal pregnancy, exercise capacity declines in the early postpartum period.38,39 Despite the improvement in the following 6 months, the pre-pregnancy level is not reached. Whether exercise capacity in women with CHD declines more than in healthy pregnant women is uncertain and will be studied in our cohort.

Conclusion The current ZAHARA II study is the first “in vivo” study in women with CHD to evaluate the effect of compromised cardiac performance on the uteroplacental circulation and its relationship with the occurrence of obstetric events and adverse offspring outcome. By

274 Balci et al

identifying the components responsible for pregnancyrelated events in women with CHD, we will refine risk stratification that will lead to better pre-pregnancy counseling and may eventually improve treatment of these women.

References 1. Drenthen W, Pieper PG, Roos-Hesselink JW, et al. Outcome of pregnancy in women with congenital heart disease: a literature review. J Am Coll Cardiol 2007;49:2303-11. 2. Siu SC, Sermer M, Colman JM, et al. Prospective multicenter study of pregnancy outcomes in women with heart disease. Circulation 2001; 104:515-21. 3. Drenthen W, Boersma E, Balci A, et al. Predictors of pregnancy complications in women with congenital heart disease. Eur Heart J 2010 doi:10.1093/eurheartj/ehq200. 4. Drenthen W, Pieper PG, Ploeg M, et al. Risk of complications during pregnancy after Senning or Mustard (atrial) repair of complete transposition of the great arteries. Eur Heart J 2005;26: 2588-95. 5. Aardema MW, Lander M, Oosterhof H, et al. Doppler ultrasound screening predicts recurrence of poor pregnancy outcome in subsequent pregnancies, but not the recurrence of PIH or preeclampsia. Hypertens Pregnancy 2000;19:281-8. 6. Bosio PM, McKenna PJ, Conroy R, et al. Maternal central hemodynamics in hypertensive disorders of pregnancy. Obstet Gynecol 1999;94:978-84. 7. Gerretsen G, Huisjes HJ, Elema JD. Morphological changes of the spiral arteries in the placental bed in relation to pre-eclampsia and fetal growth retardation. Br J Obstet Gynaecol 1981;88:876-81. 8. Madazli R, Somunkiran A, Calay Z, et al. Histomorphology of the placenta and the placental bed of growth restricted foetuses and correlation with the Doppler velocimetries of the uterine and umbilical arteries. Placenta 2003;24:510-6. 9. Khairy P, Ouyang DW, Fernandes SM, et al. Pregnancy outcomes in women with congenital heart disease. Circulation 2006;113:517-24. 10. Tzemos N, Silversides CK, Colman JM, et al. Late cardiac outcomes after pregnancy in women with congenital aortic stenosis. Am Heart J 2009;157:474-80. 11. Guedes A, Mercier LA, Leduc L, et al. Impact of pregnancy on the systemic right ventricle after a Mustard operation for transposition of the great arteries. J Am Coll Cardiol 2004;44:433-7. 12. Thorne S, MacGregor A, Nelson-Piercy C. Risks of contraception and pregnancy in heart disease. Heart 2006;92:1520-5. 13. Warnes CA, Liberthson R, Danielson GK, et al. Task force 1: the changing profile of congenital heart disease in adult life. J Am Coll Cardiol 2001;37:1170-5. 14. Baumgartner H, Hung J, Bermejo J, et al. Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice. J Am Soc Echocardiogr 2009;22:1-23. 15. Easterling TR, Carlson KL, Schmucker BC, et al. Measurement of cardiac output in pregnancy by Doppler technique. Am J Perinatol 1990;7:220-2. 16. Lang RM, Bierig M, Devereux RB, et al. Recommendations for chamber quantification: a report from the American Society of Echocardiography's Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18: 1440-63.

American Heart Journal February 2011

17. Robson SC, Dunlop W, Moore M, et al. Combined Doppler and echocardiographic measurement of cardiac output: theory and application in pregnancy. Br J Obstet Gynaecol 1987;94:1014-27. 18. Rudski LG, Lai WW, Afilalo J, et al. Guidelines for the echocardiographic assessment of the right heart in adults: a report from the American Society of Echocardiography endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography. J Am Soc Echocardiogr 2010;23:685-713. 19. Vasapollo B, Novelli GP, Valensise H. Total vascular resistance and left ventricular morphology as screening tools for complications in pregnancy. Hypertension 2008;51:1020-6. 20. Vahanian A, Baumgartner H, Bax J, et al. Guidelines on the management of valvular heart disease: the Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology. Eur Heart J 2007;28:230-68. 21. Barnett SB, Maulik D. Guidelines and recommendations for safe use of Doppler ultrasound in perinatal applications. J Matern Fetal Med 2001;10:75-84. 22. ATS/ACCP. ATS/ACCP statement on cardiopulmonary exercise testing. Am J Respir Crit Care Med 2003;167:211-77. 23. Gomez O, Figueras F, Fernandez S, et al. Reference ranges for uterine artery mean pulsatility index at 11-41 weeks of gestation. Ultrasound Obstet Gynecol 2008;32:128-32. 24. Duvekot JJ, Peeters LL. Maternal cardiovascular hemodynamic adaptation to pregnancy. Obstet Gynecol Surv 1994;49:S1-14. 25. Weiner CP, Knowles RG, Moncada S. Induction of nitric oxide synthases early in pregnancy. Am J Obstet Gynecol 1994;171: 838-43. 26. Duvekot JJ, Cheriex EC, Pieters FA, et al. Early pregnancy changes in hemodynamics and volume homeostasis are consecutive adjustments triggered by a primary fall in systemic vascular tone. Am J Obstet Gynecol 1993;169:1382-92. 27. Wilson M, Morganti AA, Zervoudakis I, et al. Blood pressure, the renin-aldosterone system and sex steroids throughout normal pregnancy. Am J Med 1980;68:97-104. 28. Hameed AB, Chan K, Ghamsary M, et al. Longitudinal changes in the B-type natriuretic peptide levels in normal pregnancy and postpartum. Clin Cardiol 2009;32:E60-2. 29. van Oppen AC, Stigter RH, Bruinse HW. Cardiac output in normal pregnancy: a critical review. Obstet Gynecol 1996;87:310-8. 30. Hogenhuis J, Jaarsma T, Voors AA, et al. BNP and functional status in heart failure. Cardiovasc Drugs Ther 2004;18:507. 31. Palazzuoli A, Gallotta M, Quatrini I, et al. Natriuretic peptides (BNP and NT-proBNP): measurement and relevance in heart failure. Vasc Health Risk Manag 2010;6:411-8. 32. Tulevski II, Groenink M, van der Wall EE, et al. Increased brain and atrial natriuretic peptides in patients with chronic right ventricular pressure overload: correlation between plasma neurohormones and right ventricular dysfunction. Heart 2001;86:27-30. 33. Giannakoulas G, Dimopoulos K, Bolger AP, et al. Usefulness of natriuretic peptide levels to predict mortality in adults with congenital heart disease. Am J Cardiol 2010;105:869-73. 34. Moghbeli N, Srinivas SK, Bastek J, et al. N-terminal pro-brain natriuretic peptide as a biomarker for hypertensive disorders of pregnancy. Am J Perinatol 2010;27:313-9. 35. Drenthen W, Pieper PG, Roos-Hesselink JW, et al. Non-cardiac complications during pregnancy in women with isolated congenital pulmonary valvar stenosis. Heart 2006;92:1838-43. 36. Yap SC, Drenthen W, Meijboom FJ, et al. Comparison of pregnancy outcomes in women with repaired versus unrepaired atrial septal defect. BJOG 2009;116:1593-601.

American Heart Journal Volume 161, Number 2

37. Drenthen W, Pieper PG, van der Tuuk K, et al. Cardiac complications relating to pregnancy and recurrence of disease in the offspring of women with atrioventricular septal defects. Eur Heart J 2005;26: 2581-7. 38. Treuth MS, Butte NF, Puyau M. Pregnancy-related changes in physical activity, fitness, and strength. Med Sci Sports Exerc 2005;37: 832-7. 39. South-Paul JE, Rajagopal KR, Tenholder MF. Exercise responses prior to pregnancy and in the postpartum state. Med Sci Sports Exerc 1992;24:410-4. 40. Dickstein K, Cohen-Solal A, Filippatos G, et al. ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute

Balci et al 275

and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur Heart J 2008;29: 2388-442. 41. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings. Duke Endocarditis Service. Am J Med 1994;96:200-9. 42. Verstappen WHJM, Jans SMPJ, Van Egmond N, et al. Nationwide Primary Care Cooperation Agreement on anemia during pregnancy and puerperium. (Landelijke Eerstelijns Samenwerkings Afspraak Anemie tijdens zwangerschap en kraamperiode). Huisarts Wetenschap 2007;50:S17-20.

American Heart Journal Volume 161, Number 2

Appendix. Supplementary Data The ZAHARA II investigators From the Departments of 1Cardiology, 2Obstetrics and 3 Epidemiology, University Medical Centre Groningen, University of Groningen, Groningen, The Netherlands; 4 Interuniversity Cardiology Institute of the Netherlands (ICIN)/Royal Dutch Academy of Science (KNAW), Utrecht, The Netherlands: Ali Balci, MD, MSc1,4, Willem Drenthen, MD, PhD1, Joost van Melle, MD, PhD1, Elke Hoendermis, MD, PhD1, Adriaan A. Voors, MD, PhD1, Dirk J. van Veldhuisen, MD, PhD1, Petronella G. Pieper, MD, PhD1, Krystyna M. Sollie, MD2, Jan G. Aarnoudse, MD, PhD2, Hans L. Hillege, MD, PhD1,3 From the Departments of 5Cardiology and 6Obstetrics, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands: Barbara J.M. Mulder, MD, PhD5, Berto J. Bouma, MD, PhD5, Maarten Groenink, MD, PhD5, Michiel M. Winter, MD, PhD5, Jeroen C. Vis, MD, PhD5, Paul Luijendijk, MD5, Zelia Koyak, MD5, Piet de Witte, MD5, A. Carla Zomer, MD5, Monique W. M. de Laat, MD, PhD6, Manja H.T.L Bunschoten, RN6 From the Departments of 7Cardiology and 8Obstetrics, Erasmus Medical Centre, Erasmus University, Rotterdam, The Netherlands: Jolien W. Roos-Hesselink, MD, PhD7,

Balci et al 275.e1

Maarten Witsenburg, MD, PhD7, Judith A.A.E. Cuypers, MD7, Eric A.P. Steegers, MD, PhD8, J. Cornette, MD, PhD8 From the Departments of 9Cardiology and 10Obstetrics University Medical Centre Nijmegen St Radboud, Radboud University Nijmegen, Nijmegen, The Netherlands: Arie P.J. van Dijk, MD, PhD9, W. Marc Waskowsky, MD, PhD9, Marc Spaanderman, MD, PhD10 From the Department of 11Cardiology, Medical Spectrum Twente, Enschede, The Netherlands: Elly M.C.J. Wajon, MD11, Lodewijk J. Wagenaar, MD, PhD, Jeannine A.J.M. Hermens, MD11 From the Departments of 12Cardiology and 13Obstetrics, Leiden University Medical Centre, University of Leiden, Leiden, The Netherlands: Hubert W. Vliegen, MD, PhD12, Monique R.M. Jongbloed, MD, PhD12, Marjolein S. Verhart, RN13, Jos J.M. van Roosmalen, MD, PhD13 From the Departments of 14Cardiology and 15Obstetrics, University Medical Centre Utrecht, The Netherlands: Gertjan T. Sieswerda, MD, PhD14, A. Carla C van Oppen, MD, PhD15, Martijn A. Oudijk, MD, PhD15 From the Departments of 16Cardiology and 17Obstetrics, Maastricht University Medical Centre, University of Maastricht, Maastricht, The Netherlands: Jan L.M. Stappers, MD, PhD16, Jos P.M. Offermans, MD, PhD17

Clinical Investigations

Acute Ischemic Heart Disease

Prehospital triage in the ambulance reduces infarct size and improves clinical outcome Sonja Postma, MSc, a Jan-Henk E. Dambrink, MD, PhD, b Menko-Jan de Boer, MD, PhD, b A. T. Marcel Gosselink, MD, PhD, b Gerrit J. Eggink, MD, c Henri van de Wetering, MANP, b,c Frans Hollak, RN, c Jan Paul Ottervanger, MD, PhD, b Jan C. A. Hoorntje, MD, PhD, b Evelien Kolkman, MSc, a Harry Suryapranata, MD, PhD, a,b and Arnoud W. J. van 't Hof, MD, PhD b Zwolle, The Netherlands

Background We evaluated the effect of prehospital triage (PHT) in the ambulance on infarct size and clinical outcome and studied its relationship to the distance of patient's residence to the nearest percutaneous coronary intervention (PCI) center. Methods All consecutive ST-segment elevation myocardial infarction patients who were transported to the Isala klinieken from 1998 to 2008 were registered in a dedicated database. Of these, 2,288 (45%) were referred via a spoke center and 2.840 (55%) via PHT. Results PHT patients were more often treated within 3 hours after symptom onset (46.2% vs 26.8%, P b .001), more often had a post-procedural thrombolysis in myocardial infarction (TIMI) 3 flow (93.0% vs 89.7%, P b .001) had a smaller infarct size (peak creatine kinase 2,188 ± 2,187 vs 2,575 ± 2,259 IU/L, P b .001) and had a lower 1-year mortality (4.9% vs 7.0%, P = .002). After multivariate analysis, PHT was independently associated with ischemic time less than 3 hours (OR 2.45, 95% CI 2.13-2.83), a peak creatine kinase less than the median value (OR 1.19, 95% CI 1.04-1.36) and a lower 1-year mortality (OR 0.67, 95% CI 0.50-0.91). The observed differences between PHT patients and the spoke group were more pronounced in the subgroup of patients living N38 km from the PCI center. Conclusion PHT in the ambulance is associated with a shorter time to treatment, a smaller infarct size and a more favorable clinical outcome, especially with longer distance from the patient's residence to the nearest PCI center. Therefore, PHT in the ambulance may reduce the negative effect of living at a longer distance from the PCI center. (Am Heart J 2011;161:276-82.)

It is well known that rapid restoration of coronary blood flow in patients with ST-segment elevated myocardial infarction (STEMI) is crucial1 to limit myocardial damage.2,3 There are 2 established therapeutic strategies to achieve this: fibrinolysis and primary percutaneous coronary intervention (pPCI), of which the latter is considered more effective.2,4–7 However, there is still debate about the time interval during which pPCI can be recommended. According to the American College of Cardiology/American Heart Association (ACC/AHA), pPCI is preferred if the delay between first medical

From the aDiagram, Zwolle, Netherlands, bIsala Klinieken, Zwolle, Netherlands, and cRAV IJsselland, Zwolle, The Netherlands. Submitted January 21, 2010; accepted October 18, 2010. Reprint requests: Arnoud W.J. van't Hof, MD, PhD, FESC, Isala klinieken, location Weezenlanden, Department of Cardiology, Groot Wezenland 20, 8011JW Zwolle, The Netherlands. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.028

contact and pPCI does not exceed 90 minutes8 and according to the European Society of Cardiology (ESC), this delay should not exceed 120 minutes.9 If the anticipated delay exceeds 90 minutes (ACC/AHA) or 120 minutes (ESC), fibrinolytic therapy may be an alternative because it can be administered with a shorter delay in the ambulance or in a nearby spoke center.10,11 One of the possibilities to shorten delays and maximize the number of patients eligible for pPCI is to transport STEMI patients directly to a PCI center after prehospital triage (PHT) in the ambulance. Several studies have shown that this strategy can significantly reduce ischemic times compared to patients being referred via a spoke center.12–15 Recently, Pedersen et al15 have demonstrated that PHT also improves outcome. However, a recent analysis from the HORIZONS AMI study did not show a benefit for patients immediately transported to a PCI center.16 In addition, it is less well known if PHT is beneficial for patients living at a longer distance from a PCI center. Therefore, we addressed both the question on efficacy and on distance in a large cohort of nonselected STEMI patients who

American Heart Journal Volume 161, Number 2

were directed to our center either via a referral center (spoke) or via PHT.

Methods Population Since the early nineties, STEMI patients referred to the Isala Klinieken were treated by pPCI. The PHT project was initiated to improve the logistics of STEMI patients. This has gradually been implemented in the region, starting in 1998. At that time 13 spoke centers referred their STEMI patients to our PCI center however, this number decreased to 8 spoke centers since new PCI centers opened. During the project, all consecutive STEMI patients who were transported to our PCI center from 1998 until 2008 and underwent pPCI were prospectively registered in a dedicated database. Criteria for the diagnosis of STEMI were: (1) history of cardiac symptoms of at least 10 minutes in the last 24 hours before presentation at the spoke or PCI center, (2) elevated levels of creatine kinase (CK) or CK-MB, and (3) concurrent electrocardiogram (ECG) changes: ST-segment elevation of N1 mV in at least 2 adjacent electrocardiogram leads.17 Information whether patients were transferred via referral centers (spoke group) in the network or via PHT (PHT group) was recorded. In addition, information on infarct size, angiographic outcome, and short- and long-term clinical outcome were registered as well. Infarct size was calculated as the peak level of CK (peak CK) within 48 hours after admission.18 The distance via motorway to the nearest PCI center was computerized using the postal codes of the patient's residence and the PCI center. Subsequently, the percentiles (25-75) of the computed distances were calculated. The PHT and spoke groups were subdivided in short (≤38 km) and long distance (N38 km) based on the median distance from patient's residence to the PCI center. Patients were excluded if they did not have a confirmed diagnosis of acute myocardial infarction (CK b200 and no evidence of unstable plaque or culprit lesion at coronary angiography), the distance from patient's residence to the PCI center was not available, or if the distance from patient's residence to the PCI center was ≥120 km (outer bound of referring area).

Triage for pPCI PHT: The algorithm of PHT has been described previously.19 In brief, after patients dialed the emergency number, patients were triaged in the ambulance, if the ambulance was equipped with the PHT equipment. If STEMI was suspected, an ECG was made by highly trained paramedics and the computerized algorithm revealed an outcome. If a diagnosis of STEMI was made, the ambulance went straight to the catheterization laboratory of the PCI center, bypassing the emergency departments of the spoke center. Spoke: If the ambulance was not equipped with the PHT equipment, the ambulance went to the nearest spoke center where diagnosis and triage was performed. For diagnosis, an ECG was made immediately upon arrival. In case of a STEMI diagnosis, patients were transported to the catheterization laboratory of the PCI center as soon as possible. Walk-ins at the PCI center were excluded because they did not receive PHT.

Postma et al 277

pPCI procedure In both situations (PHT and spoke), the staff of the catheterization laboratory of the PCI center was preinformed about the estimated time of arrival of the patient and was activated well before the arrival of the patient. In case the staff lived more than 30 minutes away from the PCI center, they had to stay in the PCI center when being on-call. Patients were pretreated with an intravenous bolus of 5000 IU unfractionated heparin, 500 mg aspirin intravenously, and subsequently with 600 mg clopidogrel and/or tirofiban (25 μg/kg bolus, 0.15 μg kg−1 min−1 maintenance infusion). Prehospital triage patients were pretreated in the ambulance and spoke patients were pretreated at the spoke center and/or in the ambulance that transferred the patient form the spoke center to the PCI center.

Time intervals Six different time intervals were evaluated: (1) time from symptom onset (SO) to infarct diagnosis (time diagnostic ECG) either in the ambulance or at a spoke center (SO-diagnosis); (2) time from diagnosis till arrival at the PCI center (diagnosis-door PCI); (3) time from diagnosis to balloon inflation (BI) (diagnosisBI); (4) time from arrival spoke center to arrival PCI center (door-to-door, or D2D); (5) time from arrival at the PCI center to BI for PHT patients and time from arrival spoke hospital to BI for spoke patients (door-to-balloon, or D2B); and (6) total ischemic time defined as the time from SO to BI.

Statistical analysis Statistical analysis was performed with SPSS 17.0 (SPSS, Chicago, IL). Continuous data were expressed as mean ± SD or median and interquartile range. Categorical data were presented as percentage. Analysis of variance was used for continuous data and Pearson χ2 test was used for the categorical data, respectively. We tested the associations between the variable “PHT” and other baseline characteristics using univariate logistic regression. The Mann-Whitney test has been used to calculate the time intervals between the PHT group and the spoke group, because they were non-Gaussian distributed. To assess independent predictors of PHT, multivariate analysis was performed using a logistic regression analysis. In this analysis, univariate variables with a P value b.10 were included. In all the statistical analyses, P values ≤.05 were considered as statistical significant. The study was conducted according to the principles of the Declaration of Helsinki and the protocol was approved by the local institutional review board. No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses and the drafting and editing of the paper.

Results Ambulance triage versus referral by spoke center From 1998, the first year of implementation of PHT in the ambulance, until 2008, 5,674 patients were referred to our institution with the intention to perform pPCI. Three hundred twenty-three patients (5.7%) were false positives (PHT and spoke patients) and 223 patients

American Heart Journal February 2011

278 Postma et al

Table I. Baseline characteristics Characteristic

Total (N = 5128)

PHT (n = 2840)

Spoke (n = 2288)

P

Age (y) ± SD Male gender Risk factors Hypertension DMII Smoking Hypercholesterolemia Family history Killip class N1 Previous MI Previous PCI Previous CABG Previous CVA TIMI risk score

62.17 ± 12.14 3800/5125 (74.1%)

62.80 ± 12.12 2110/2838 (74.3 %)

61.40 ± 12.12 1690/2287 (73.9%)

b.001 .713

1668/5063 (32.9%) 557/5108 (10.9%) 2433/5015 (48.5%) 1085/4852 (22.4%) 2058/4928 (41.8%) 346/5098 (6.8%) 499/5109 (9.8%) 364/5074 (7.2%) 132/5119 (2.6%) 146/5114 (2.9%) 2.16 ± 1.82 2 (1-3)

926/2811 (32.9%) 296/2833 (10.4%) 1310/2785 (47.0%) 563/2682 (21.0%) 1151/2731 (42.1%) 171/2831 (6.0%) 232/2830 (8.2%) 214/2821 (7.6%) 84/2837 (3.0%) 78/2834 (2.8%) 2.17 ± 1.84 2 (1-3)

742/2252 (32.9%) 261/2275 (11.5%) 1123/2230 (50.4%) 522/2170 (24.1%) 907/2197 (41.3%) 175/2267 (7.7%) 267/2279 (11.7%) 150/2253 (6.7%) 48/2282 (2.1%) 68/2280 (3.0%) 2.15 ± 1.80 2 (1-3)

.996 .243 .019 .011 .542 .018 b.001 .203 .054 .623 .948

CABG, Coronary artery bypass grafting; CVA, cerebrovascular accident; DMII, diabetes mellitus type II; MI, myocardial infarction.

Figure 1

The different time intervals are shown for the total group, for patients living ≤38 km from a PCI center and for patients living N38 km from a PCI center. These groups were subdivided for PHT and spoke patients.

(3.9%) were walk-ins at the PCI center. From the remaining 5,128 patients who actually underwent pPCI, 2,288 patients (45%) were referred via a spoke center and 2,840 patients (55%) via PHT. Patients from the spoke group were younger, more often smoked, more often had hypercholesteremia, more often had a previous myocardial infarction, more often presented in Killip class N1 and less often had previous coronary artery bypass grafting (CABG) (Table I). The median distance from patient's residence to the nearest PCI center was 38 km (25-49), 28 km (16-41) for the PHT

group and 43 km (37-60) for the spoke group (P b .001). The different time intervals for the 2 groups are presented in Figure 1. Total median ischemic time was 209 minutes, 184 minutes for the PHT group and 260 minutes for the spoke group (P b .001). The percentage of patients treated within 3 hours of SO was 46.2% in the PHT group compared with 26.8% in the spoke group (P b .001) (Table II). The time from SO-diagnosis was shorter for the PHT group as compared to that for the spoke group (82 vs 129 minutes, P b .001). The median D2D for the spoke group was 85 minutes. The D2B was longer for the spoke

American Heart Journal Volume 161, Number 2

Postma et al 279

Table II. Distance, total time delays, indication of pPCI, and distance of patient's residence to PCI center PHT Total Distance (km)⁎ Ischemic time (median, IQR, n = 2728) Ischemic time b3 h (%, n = 5028) SO-diagnosis (median, IQR, n = 2531) Diagnosis-door PCI (median, IQR, n = 1982) Diagnosis-BI (median, IQR, n = 2625) D2D (median, IQR, n = 984) D2B (median, IQR, n = 2664) Indication pPCI ACC/AHA (%, n = 2625) Indication pPCI ESC (%, n = 2625) Distance home-PCI center ≤38 km Distance (km)⁎ Ischemic time (median, IQR, n = 1308) Ischemic time b3 h (n = 2522) SO-diagnosis (median, IQR, n = 1206) Diagnosis-door PCI (median, IQR, n = 974) Diagnosis-BI (median, IQR, n = 1240) D2D (median, IQR, n = 285) D2B (median, IQR, n = 1266) Indication pPCI ACC/AHA (%, n = 1245) Indication pPCI ESC (%, n = 1245) Distance home-PCI center N38 km Distance (km)⁎ Ischemic time (median, IQR, n = 1420) Ischemic time b3 h (%, n = 2506) SO-diagnosis (median, IQR, n = 1325) Diagnosis-door PCI (median, IQR, n = 1326) Diagnosis-BI (median, IQR, n = 1358) D2D (median, IQR, n = 699) D2B (median, IQR, n = 1398) Indication pPCI ACC/AHA (%, n = 1385) Indication pPCI ESC (%, n = 1385)

28.1 184 1285/2784 82 52 94

(15.9-14.8) (140-274) (46.2%) (47-150) (37-68) (77-119)

44 (30-69) 719/1637 (43.9%) 1230/1637 (75.1%) 19.0 187 850/1908 87 46 90

(13.1-29.0) (139-282) (44.5%) (49-158) (33-61) (74-118)

46 (31-74) 484/962 (49.7%) 731/962 (76.0%) 45.9 180 435/876 76 60 101

(41.0-63.0) (142-256) (49.7%) (45-136) (45-79) (82-122)

40 (27-61) 241/675 (35.7%) 488/675 (73.9%)

Spoke

P

42.6 (37.3-59.9) 260 (187-393) 601/2244 (26.8%) 129 (68-252) 83 (59-116) 123 (98-160) 85 (60-123) 132 (101-189) 173/988 (17.5%) 466/988 (47.2%)

b.001 b.001 b.001 b.001 b.001 b.001

32.6 (26.8-36.0) 255 (177-407) 171/614 (27.9%) 132 (69-259) 78 (54-114) 118 (93-159) 78 (53-120) 127 (98-198) 56/283 (20.1%) 149/283 (52.7%)

b.001 b.001 b.001 b.001 b.001 b.001

49.6 261 430/1630 125 85 124 89 133 117/710 322/710

(41.6-66.7) (190-391) (26.4%) (66-251) (61-117) (100-160) (62-126) (103-188) (16.5%) (45.4%)

b.001 b.001 b.001

b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001

IQR, Interquartile range. ⁎ Data expressed as median and IQR.

group as compared to that for the PHT group (132 vs 44 minutes, P b .001). Significantly more patients of the PHT group were treated according to the most recent guidelines of the ACC/AHA and the ESC as compared to the spoke group (ACC/AHA: PHT 43.9% vs spoke 17.5%, ESC: PHT 75.1% vs spoke 47.2%, P b .001 for both comparisons) (Table II).8,9 PHT patients had also better angiographic and clinical outcome. They more often had post-procedural TIMI 3 flow (93.0% vs 89.7%, P b .001), had a smaller infarct size (peak CK: 2188 ± 2187 vs 2575 ± 2259 IU/L, P b .001) and a lower 1-year mortality (4.9% vs 7.0%, P = .002) (Table III). After correction for differences in baseline characteristics, including the difference in distance, PHT was independently associated with an ischemic time of less than 3 hours (OR 2.45, 95% CI 2.13-2.83), infarct size of less than the median value (OR 1.19, 95% CI 1.04-1.36) and a lower 1-year mortality (OR 0.67, 95% CI 0.50-0.91) (Table IV).

Distance to PCI center Figure 1 illustrates the time intervals in association with the distance from the patient's residence to the PCI center. PHT patients had a shorter time from SO-diagnosis irrespective of distance to the PCI center. Surprisingly, total ischemic time was even somewhat shorter in the PHT patients living at a distance N38 km as compared to ≤38 km. For both groups, however, the time from diagnosis-door PCI was longer for patients living at a longer distance from the PCI center (PHT from 44 to 55 minutes, spoke from 78 to 89 minutes, P b .001 for both comparisons). A longer distance led to a decrease in the percentage of patients treated according to the guidelines of the ACC/AHA (≤38 km: 540/1245 [43.4%] vs N38 km: 358/1385 [25.6%], P b .001) and the ESC (≤ 38 km: 880/1245 [70.7%] vs N38 km: 810/1385 [58.5%], P b .001) (Table II).8,9 In patients living ≤38 km from the PCI center, apart from a higher post-procedural TIMI 3 flow in the PHT group, no significant differences in angiographic

American Heart Journal February 2011

280 Postma et al

Table III. Angiographic and clinical outcome PHT (n = 2840)

Spoke (n = 2288)

Total TIMI 3 flow pre-PCI 570/2826 (20.2%) 423/2286 (18.5%) TIMI 3 flow post-PCI 2625/2822 (93.0%) 2042/2276 (89.7%) Infarct size ± SD⁎ 2188±2187 2575±2259 (IU/L, n = 4986) 30-d mortality 86/2798 (3.1%) 90/2249 (4.0%) 1-y mortality 134/2729 (4.9%) 151/2164 (7.0%) Distance residence-PCI center ≤38 km TIMI 3 flow pre-PCI 359/1935 (18.6%) 125/626 (20.0%) TIMI 3 flow post-PCI 1792/1932 (92.8%) 561/623 (90.0%) Infarct size ± SD⁎ 2154 ± 2161 2359±2146 (IU/L, n = 2482) 30-d mortality 60/1913 (3.1%) 28/615 (4.6%) 1-y mortality 98/1878 (5.2%) 41/581 (7.1%) Distance residence-PCI center N38 km TIMI 3 flow pre-PCI 211/891 (23.7%) 298/1660 (18.0%) TIMI 3 flow post-PCI 833/890 (93.6%) 1481/1653 (89.6%) Infarct size ± SD⁎ 2261 ± 2241 2656±2295 (IU/L, n = 2504) 30-d mortality 26/885 (2.9%) 62/1634 (3.8%) 1-y mortality 36/851 (4.2%) 110/1583 (6.9%)

P

.134 b.001 b.001 .074 .002 .432 .030 .054 .096 .094 b.001 b.001 b.001 .264 .007

⁎ Peak CK.

parameters, infarct size, or clinical outcome were found. However, in patients living N38 km away from the PCI center, PHT patients more often had a higher initial and post-procedural TIMI 3 flow, had a smaller infarct size and had a lower 1-year mortality compared to the spoke group (Table III).

Discussion So far, this is the largest study showing that PHT in the ambulance with immediate transportation to the nearest PCI center result in a shorter time to treatment, a reduction in infarct size and a better angiographic and clinical outcome as compared to referral via a spoke center. This was most evident for patients living at a longer distance from the PCI center. In addition, PHT significantly increased the number of patients treated according to the ACC/AHA and the ESC guidelines.8,9 These results suggest that living at a longer distance from a PCI center may not negatively influence angiographic and clinical outcomes when PHT is available. Our study showed that 43.9% and 75.1% of the patients, respectively, fulfill the ACC/AHA and the ESC criteria for pPCI when PHT is available. When diagnosis and triage was performed at a spoke center, only 17.5% (ACC/AHA) and 47.2% (ESC) of the patients were treated according to the guidelines. Figure 1 shows that the largest difference between the groups is seen in the time from SO-diagnosis. Because the ambulance that brought the patient initially to a spoke center was not able to make the infarct diagnosis (no trained personnel, no

ECG equipment), diagnosis was only made after arrival of the patient in the spoke center at a median of 129 minutes after SO, whereas after PHT, diagnosis was made 47 minutes (36%) earlier. This earlier diagnosis, together with the early initiation of potent antiplatelet and antithrombotic agents in the ambulance (unfractionated heparin, aspirin, and clopidogrel), may have led to the higher initial patency and better angiographic outcome in PHT patients.20 Despite the fact that all efforts were taken to arrange further or a new transport to the PCI center as soon as possible, the diagnosis-door PCI time was considerably longer in the spoke group as compared to the diagnosis-door PCI time in the PHT group (83 vs 52 minutes, P b .001) (Figure 1). As a consequence, the D2B time was significantly longer for the spoke group as compared to the PHT group (132 vs 44 minutes, P b .001). These findings correspond with the results of Le May et al.21

Effect of distance Our study shows that time to treatment is substantially reduced by PHT in the ambulance and shows that incorporating this triage significantly increases the number of patients who are candidates for primary angioplasty instead of thrombolysis according to the guidelines, especially for patients who live further than 38 km away from a PCI center. These findings suggest that the logistics of arranging diagnosis and immediate transportation is more important than the distance of the patient's residence from a PCI center: for PHT patients, total ischemic times remained very short despite a longer distance from the PCI center. A longer distance from patients' residence to the PCI center had very limited effect on total ischemic time despite a significantly increased diagnosis-door PCI time. This was due to a shorter SO-diagnosis time. Other investigators have also studied the effect of distance on clinical outcome in STEMI patients. In a cohort study, Wei et al22 demonstrated that patients with a first myocardial infarction who were living at more than 9 miles (14.5 km) from the PCI center had a higher mortality compared to patients who lived closer to the PCI center. Nevertheless, in our study, the effect of distance on outcome was related to the type of patient triage, which was not reported in the study from Wei et al. According to a recent study, D2B may also be reduced by performing pPCI at a PCI center without on-site cardiac surgery.23 However, this study was stopped prematurely and a third arm with routine ambulance triage was lacking.24 It might therefore be true that the opening of extra PCI centers is not necessary when routine ambulance triage systems are being developed. Because PHT patients have better outcomes compared to spoke patients, more centers should implement the

American Heart Journal Volume 161, Number 2

Postma et al 281

Table IV. Output multivariate analyses Treatment within 3 h Variables Type triage Gender Hypertension Family history DMII Age TIMI risk score Previous CABG Previous PCI IRV/LAD Distance

OR (95% CI) 2.45 1.21 1.20 1.04 1.01 1.03 0.61 0.70 1.55 1.27 1.00

(2.13-2.83) (1.04-1.41) (1.04-1.41) (0.91-1.18) (0.81-1.26) (1.02-1.04) (0.57-0.65) (0.46-1.08) (1.21-1.99) (1.13-1.47) (1.00-1.00)

Infarct size < median Variables Type triage Gender Hypertension Hypercholesterolemia Smoking Age Previous MI Previous CABG Previous PCI Killip class N1 IRV/CX IRV/graft IRV/LAD IRV/RCA Distance

OR (95% CI) 1.19 0.67 0.96 0.99 0.86 1.01 1.29 1.03 1.47 0.61 2.10 3.98 1.85 4.05 0.99

(1.04-1.36) (0.58-0.78) (0.84-1.10) (0.85-1.16) (0.75-0.98) (1.00-1.01) (1.01-1.65) (0.58-1.83) (1.12-1.93) (0.47-0.78) (0.67-6.64) (1.02-15.6) (0.59-5.81) (1.29-12.7) (0.99-1.00)

One-year mortality Variables Type triage Gender Hypertension Hypercholesterolemia DMII Age TIMI risk score Family history Previous MI Pervious CVA Killip class N1 Distance IRV/LAD

OR (95% CI) 0.67 0.90 1.42 0.63 1.06 1.04 1.21 0.70 1.23 2.29 3.07 1.00 1.05

(0.50-0.91) (0.66-1.21) (1.06-1.92) (0.43-0.93) (0.72-1.57) (1.02-1.05) (1.09-1.35) (0.51-0.96) (0.90-1.67) (1.28-4.10) (1.99-4.72) (0.99-1.01) (0.80-1.39)

CABG, Coronary artery bypass grafting; CX, circumflex; DMII, diabetes mellitus type II; IRV, infarct related vessel; LAD, left anterior descendents; MI, myocardial infarction; RCA, right coronary artery.

PHT for STEMI patients and use ECG equipment with a computerized electrographic algorithm or with telemedicine. It is also important to make patients and general practitioners aware of the fact that PHT with ambulance transport has better outcomes for STEMI patients instead of self-referring, referring via general practitioners, or referring via a spoke center.

Limitations Several limitations of this study need to be acknowledged. First, because the project was not randomized and dispersed over 10 years (the percentage of PHT patients increased from 6.0% in 1998, 51.6% in 2004, to 68.7% in 2008), consequently, the risk of unknown confounders exists. Some selection of patients has occurred. PHT patients were older and more often lived closer to the PCI center. During the PHT project, more remote ambulance services started participating, while at the beginning these patients all were transported via a spoke center. Second, we did not investigate the effect of fibrinolytics for patients living at great distance. Therefore, we cannot state that PHT is the best way to treat all STEMI patients living at a great distance however, in our population PHT patients have better outcomes when living at N38 km from a PCI center compared to patients referred via a spoke center. Third, we have no exact numbers of patients who were self-referred or came in by ambulance at the spoke center. Most patients came in by ambulance at the spoke center however, these ambulances were not operational with ECG equipment and the EMS personnel were not trained to make STEMI diagnosis. Subsequently, diagnosis was made in the spoke hospital however, first medical contact took place in the ambulance.

Fourth, information of the time of first medical contact is lacking. We expect these times to be the same in both groups however, we do not have solid evidence. Fifth, although peak CK has been shown to be a reliable parameter for the estimation of infarct size, data on the area under the CK release curve are lacking. Sixth, we have used the postal codes of patient's residence to calculate the distance to the PCI center because the exact location of the place where the patient was picked up by an ambulance was not available. Finally, more research has to be performed to the favor of PHT for patients living at a long distance from the PCI center, so more understanding and verification of this improvement can be achieved.

Conclusion In conclusion, PHT in the ambulance with immediate transportation to the nearest PCI center is associated with a significantly shorter time to treatment, reduced infarct size, and better angiographic and clinical outcome when compared to referral via a nearby spoke center. This beneficial effect is more apparent with longer distance from the patient's residence to the PCI center. PHT also significantly increased the percentage of patients that fall within the time window in which pPCI is the preferred treatment according to the ACC/AHA and ESC guidelines. Therefore, PHT may reduce transportation delays in patients who live at a longer distance from the PCI center.

Disclosures Conflict of interest: none declared.

282 Postma et al

References 1. Keeley EC, Hillis LD. Primary PCI for myocardial infarction with STsegment elevation. N Engl J Med 2007;356:47-54. 2. Boersma E. The Primary Coronary Angioplasty vs. Thrombolysis (PCAT)-2 Trialists' Collaborative Group. Does time matter? A pooled analysis of randomized clinical trials comparing primary percutaneous coronary intervention and in-hospital fibrinolysis in acute myocardial infarction patients. Eur Heart J 2006;27:779-88. 3. De Luca G, Suryapranata H, Ottervanger JP, et al. Time delay to treatment and mortality in primary angioplasty for acute myocardial infarction. Every minute of delay counts. Circulation 2004;109: 1223-5. 4. Dalby M, Bouzamondo A, Lechat P, et al. Transfer for primary angioplasty versus immediate thrombolysis in acute myocardial infarction. A meta-analysis. Circulation 2003;108:1809-14. 5. Henry TD, Unger BT, Sharkey SW, et al. Design of a standardized system for transfer of patients with ST-elevation myocardial infarction for percutaneous coronary intervention. Am Heart J 2005;150: 373-84. 6. Keeley EC, Boura JA, Grines CL. Primary angioplasty versus intravenous thrombolytic therapy for acute myocardial infarction: a quantitative review of 23 randomised trials. Lancet 2003;361:13-20. 7. Andersen HR, Nielsen TT, Rasmussen K, et al. A comparison of coronary angioplasty with fibrinolytic therapy in acute myocardial infarction. N Eng J Med 2003;349:733-42. 8. ACC/AHA. 2007 focussed update of the ACC/AHA 2004 guidelines for the management of patients with ST elevated myocardial infarction. JACC 2008;51:210-47. 9. The task force on the management of ST-segment elevation acute myocardial infarction of the European Society of Cardiology: Task force members. Management of acute myocardial infarction in patients presenting with persistent ST-segment elevation. Eur Heart J 2008;29:2909-45. 10. Pinto DS, Kirtane AJ, Nallamothu BK, et al. Hospital delays in reperfusion for ST-elevation myocardial infarction: implications when selecting a reperfusion strategy. Circulation 2006;114: 2019-25. 11. Nallamothu BK, Bates ER, Herrin J, et al. Times to treatment in transfer patients undergoing primary percutaneous coronary intervention in the United States: National Registry of Myocardial Infarction (NRMI)3/4 analysis. Circulation 2005;111:761-7. 12. Terkselsen CJ, Lassen JF, Nørgaard BJ, et al. Reduction of treatment delay in patients with ST-elevated myocardial infarction: impact of pre hospital diagnosis and direct referral to primary coronary intervention. Eur Heart J 2005;26:770-7.

American Heart Journal February 2011

13. Terkelsen CJ, Nørgaard BJ, Lassen JF, et al. Prehospital evaluation in ST-elevation myocardial infarction patients treated with primary percutaneous coronary intervention. J of Electrocard 2005;38: 187-92. 14. Clemmensen P, Sejersten M, Sillesen M, et al. Diversion of STelevation myocardial infarction patients for primary angioplasty based on wireless prehospital 12-lead electrocardiographic transmission directly to the cardiologist's handheld computer: a progress report. J of Electrocard 2005;38:194-8. 15. Pedersen SH, Galatius SG, Hanse PR, et al. Field triage reduces treatment delay and improves long-term clinical outcome in patients with acute ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention. J Am Coll Card 2009; 54:2296-302. 16. O'Riordan M. Should patients go directly to a PCI-capable hospital? Two studies, different results. Paper presented at: Transcatheter Cardiovascular Therapeutics; September 23, 2009, San Fransico, CA. 17. Van ‘t Hof AWJ, Liem A, de Boer MJ, et al. Clinical value of 12-lead electrocardiogram after successful reperfusion therapy for acute myocardial infarction. The Lancet 1997;350:615-9. 18. Nienhuis MB, Ottervanger JP, de Boer MJ, et al. Prognostic importance of creatine kinase and creatine kinase–MB after primary percutaneous coronary intervention for ST-elevation myocardial infarction. Am Heart J 2008;155:673-9. 19. Van't Hof AWJ, Rasoul S, Wetering H, et al. Feasibility and benefit of prehospital diagnosis, triage, and therapy by paramedics only in patients who are candidates for primary angioplasty for acute myocardial infarction. Am Heart J 2006;151:1255.e1-5. 20. Zijlstra F, Patel A, Jones M, et al. Clinical characteristics and outcome of patients with early (b2 h), intermediate (2-4 h) and late (N4 h) presentation treated by primary coronary angioplasty or thrombolytic therapy for acute myocardial infarction. Eur Heart J 2002;23:550-7. 21. Le May MR, Derek YS, Dionne R, et al. A citywide protocol for primary PCI in ST-segment elevated myocardial infarction. N Eng J Med 2008;358:231-40. 22. Wei L, Lang CC, Sullivan FM, et al. Impact on mortality following first acute myocardial infarction of distance between home and hospital: cohort study. Heart 2008;94:1141-6. 23. Peels HO, Swart de H, van der Ploeg T, et al. Percutaneous coronary intervention with off-site cardiac surgery backup for acute myocardial infarction as a strategy to reduce door-to-balloon time. Am J Card 2007;100:1353-8. 24. de Boer MJ, Bronzwaer JGF, Boers M. Percutaneous coronary intervention with off-site cardiac surgical backup. Am J Card 2008; 101:1522.

The influence of time from symptom onset and reperfusion strategy on 1-year survival in ST-elevation myocardial infarction: A pooled analysis of an early fibrinolytic strategy versus primary percutaneous coronary intervention from CAPTIM and WEST Cynthia M. Westerhout, PhD, a,d Eric Bonnefoy, MD, b,d Robert C. Welsh, MD, a,d Philippe Gabriel Steg, MD, c,d Florent Boutitie, PhD, b,d and Paul W. Armstrong, MD a,d Edmonton, Canada; and Lyon and Paris, France

Background The CAPTIM trial suggested a survival benefit of prehospital fibrinolysis (FL) compared to primary percutaneous coronary intervention (PCI) in patients with ST-elevation myocardial infarction (STEMI) with a presentation delay of b2 hours. We examined the relationship between reperfusion strategy and time from symptom onset on 1-year mortality in a combined analysis of 1,168 patients with STEMI. Methods Individual patient data from CAPTIM (n = 840, 1997-2000) and the more recent WEST trial (n = 328, 20032005) were pooled. Results Median age was 58 years, 81% were men, and 41% had anterior myocardial infarction; 640 patients were randomized to FL versus 528 patients to PCI. Both arms received contemporary adjunctive medical therapy. Presentation delay (ie, symptom onset to randomization) was similar in FL and PCI patients (median 105 [72-158] vs 106 [74-162] minutes, P = .712). Rescue PCI after FL occurred in 26% and 27%, and 30-day PCI, in 70% and 71% in CAPTIM and WEST, respectively. Mortality was not different between FL and PCI (4.6% vs 6.5%, P = .263); however, the interaction between presentation delay and treatment was significant (P = .043). Benefit with FL was observed with time b2 hours (2.8% [FL] vs 6.9% [PCI], P = .021, hazard ratio [HR] 0.43, 95% CI 0.20-0.91), whereas beyond 2 hours, no treatment difference was observed (6.9% [FL] vs 6.0% [PCI], P = .529, HR 1.23, 95% CI 0.61-2.46). Conclusions A strategy of early FL demonstrated a reduction in 1-year mortality compared to primary PCI in early presenters. Time from symptom onset should be a key consideration when selecting reperfusion therapy for STEMI. (Am Heart J 2011;161:283-90.)

The optimal reperfusion therapy for patients with STelevation myocardial infarction (STEMI) has attracted great interest, stimulated vigorous controversy, and led to constructive enhancement of health care systems aimed at abbreviating time from symptom onset to reperfusion. Although contemporary STEMI guidelines recommend primary percutaneous coronary intervention (PCI) as the

From the aUniversity of Alberta, Edmonton, Canada, bUniversite Lyon, Lyon, France, and c Universite Paris, Paris, France. ClinicalTrials.gov Identifier: NCT00121446 (WEST). d For the CAPTIM/WEST Investigators. Submitted August 25, 2010; accepted October 15, 2010. Reprint requests: Paul W. Armstrong, MD, 251 Medical Sciences Building, University of Alberta, Edmonton, Canada T6G 2H7. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.033

preferred strategy, provided it can be delivered promptly in an expert facility, this is often not feasible.1,2 However these guidelines also emphasize the desirability of a total ischemic time of b2 hours and support the use of fibrinolytic therapy if primary PCI cannot be performed within 90 minutes of first medical contact. It follows that patient and situational specific reperfusion strategies, that is, fibrinolysis (FL) or primary PCI, require consideration, and the optimal approach remains an active source of debate in many circumstances. Although some trials and a systematic overview conclude that PCI is the preferred therapy, there are notable caveats arguing against a “one size fits all” strategy favoring a more nuanced approach.3,4 An important and well-recognized modulator of prognosis after STEMI is time from symptom onset until reperfusion,5,6 which is now a required consideration in therapeutic decision making.1,2 Indeed, recent STEMI guidelines indicate a goal of curtailing total ischemic time

American Heart Journal February 2011

284 Westerhout et al

to b2 hours.1 Major advancements in the delivery of STEMI care have been aimed at minimizing time to treatment, particularly in prehospital diagnosis, early delivery of reperfusion,7 and timely triage to an appropriate hospital. The CAPTIM trial was the first large-scale trial to evaluate the efficacy of prehospital FL versus PCI8 and, importantly, provided evidence of an early window of benefit with prehospital FL when time from symptom onset was within 2 hours and mechanical coronary cointervention was frequently employed.9 However, the CAPTIM trial was hypothesis-generating in this regard and concluded enrollment (after 840 patients) before its target sample of 1,200 patients. The current study extends this observation through collaboration between CAPTIM and the WEST trialists, the latter adding over 300 similarly randomized contemporaneous patients with STEMI, thereby approximating the initial CAPTIM enrollment.10 Our primary goal was to examine the influence of reperfusion choice on 1-year mortality and the extent to which it is modulated by time from symptom onset. In addition, we explored this issue on the prespecified events of recurrent myocardial infarction (MI), cardiogenic shock, and safety.

Methods The details and primary results of the CAPTIM and WEST trials have been previously published.8,10 In brief, the CAPTIM trial randomized eligible patients at the site of initial management to prehospital FL or direct transfer for primary PCI. All patients received an intravenous bolus of 5,000 U heparin and 250 to 500 mg aspirin. Patients assigned to prehospital FL received an intravenous bolus of alteplase followed by an infusion over 90 minutes. Coronary angiography and subsequent revascularization were allowed in the FL group at the discretion of the responsible physicians and, when appropriate, rescue angioplasty was done. Patients assigned to primary PCI were transported immediately to the hospital for coronary angiography and angioplasty, if indicated. Angioplasty was done according to local standards with the intention of restoring blood flow in the infarct-related artery as soon as possible. After randomization, heparin was continued for at least 48 hours. Those receiving stents were treated with a thienopyridine for 1 month. In addition to aspirin, the protocol recommended use of atenolol in all patients and lisinopril in those with anterior infarcts. The WEST trial had a parallel-group design that randomized patients into one of the following 3 treatment arms, at the earliest point of care, including the prehospital setting and participating study hospitals: (1) usual care: optimal pharmacologic therapy (prompt administration of tenecteplase [TNK], aspirin, and enoxaparin); (2) early invasive strategy: identical pharmacologic therapy and early invasive strategy including mandatory rescue PCI; or (3) primary PCI (after aspirin, enoxaparin, and 300 mg clopidogrel). Abciximab was recommended for all PCI procedures unless performed within 3 hours of fibrinolytic therapy, and clopidogrel was used in patients in the pharmacologic therapy groups according to American College of Cardiology/ American Heart Association PCI guidelines.

Patients The CAPTIM trial enrolled patients if they presented within 6 hours of symptom onset (ie, characteristic pain lasted for at least 30 minutes, not responsive to nitrates, with electrocardiographic [ECG] ST-segment elevation of at least 0.2 mV in ≥2 contiguous leads, or left bundle branch block) in mobile emergency care units (Service d'Aide Médicale d'Urgence) with 27 affiliated tertiary hospitals in France. Patients could be excluded if the transfer time to hospital was expected to be N60 minutes. Like CAPTIM, the WEST trial enrolled patients with STEMI within 6 hours of symptom onset. Specifically, patients were sought in whom primary PCI could not be delivered within 60 minutes but for whom primary PCI and/or transfer for rescue PCI was feasible within 3 hours of randomization.11 Symptoms presumed secondary to STEMI lasting at least 20 minutes were required, accompanied by ECG evidence of high risk, which included ≥2 mm of ST-elevation in ≥2 contiguous precordial leads or limb leads or ≥1-mm STelevation in ≥2 limb leads coupled with ≥1-mm STdepression in ≥2 contiguous precordial leads (total STdeviation ≥4 mm) or presumed new left bundle branch block. An emphasis on the earliest possible randomization was a cardinal feature of WEST and prehospital ECG; randomization and initiation of therapy (using paramedics vs physicians in CAPTIM) were strongly encouraged for those patients using 911/ambulance access to health care facilities. Forty-four percent of patients were randomized prehospital: 45.7% with FL and 41.1% with primary PCI. The enrollment period was extended beyond the prespecified sample of 304 patients to include an additional 24 patients to expand the prehospital randomization. Fifteen sites within 4 metropolitan areas in Canada (Edmonton, Halifax, Montreal, and Vancouver) were involved in enrollment. The maximum follow-up in WEST was 1 year.

Statistical analysis Discrete variables are reported as counts and percentages of nonmissing cases; median and 25th and 75th percentiles are reported for continuous variables. Differences between studies and between study treatments were tested by χ2, WilcoxonMann-Whitney U, and Kruskal-Wallis tests as appropriate. Analyses were conducted according to the intention-to-treat principle, which included all patients who gave informed consent and were randomized to study treatment, irrespective of whether treatment was actually received. In the WEST trial, the 2 FL arms (TNK + usual care and TNK + early angiography) were not statistically different with respect to the primary end point of the trial10; thus, these 2 arms were combined with the fibrinolytic arm in CAPTIM. The primary end point of this study was all-cause mortality within 1 year. Eight patients were lost to follow-up in the WEST study, whereas CAPTIM had a complete 1-year followup. Kaplan-Meier survival curves were constructed to illustrate the time to death within 1 year after randomization, with between-group differences evaluated by the log-rank test (and accounting for the between-trial variation by stratification). Via stratified Cox proportional hazards regression, the adjusted HR and corresponding 95% CI for 1-year mortality were estimated for study treatment, time from symptom onset, and the

American Heart Journal Volume 161, Number 2

Westerhout et al 285

Figure 1

Derivation of study cohort.

Table I. Selected baseline characteristics CAPTIM

n Age, y Male, % Systolic BP, mm Hg Pulse, beat/min Weight, kg Height, cm Killip Class N1, % Hypertension, % History of diabetes, % History of angina, % Prior MI, % Previous PCI, % Current smoker, % Anterior MI, %

WEST

CAPTIM-WEST

FL

Primary PCI

FL

Primary PCI

All

FL

Primary PCI

419 58 (49-68.5) 82.5 125 (110-140) 75 (64-84) 75 (68-85) 170 (165-175) 8.7 34.1 11.1 13.5 8.2 5.3 52.7 40.1

421 58 (50-68) 81.5 128 (111-140) 75 (66-88) 75 (67-84) 170 (165-175) 12.2 34.8 13.6 14.6 6.7 4.3 49.2 42.7

221 58 (50-68.5) 77.4 140 (122.3-160) 72 (62-85) 83 (73-94) 173 (167-178) 5.4 45.2 11.8 28.5 12.7 8.1 49.5 38.9

107 60 (49-71) 80.4 143 (124-158.5) 74 (62-88) 82 (73-93) 173 (167.5-179) 6.5 34.6 15.0 18.7 13.1 4.7 39.3 42.1

1168 58 (49-69) 81.0 131 (23.5) 75 (65-86) 77 (69-86) 170 (165-176) 9.1 36.5 12.5 17.2 9.0 5.4 49.5 41.0

640 58 (49-68.3) 80.7 130 (115-147) 74 (63-84) 78 (70-88) 171 (165-176) 7.5 38.0 11.3 18.7 9.8 6.3 51.6 39.7

528 58.5 (50-69) 81.3 130 (115-145) 75 (65-88) 75 (68-85) 170 (165-176) 11.0 34.8 13.9 15.4 8.0 4.4 47.1 42.9

Continuous variables as median (25th to 75th percentile).

interaction of these 2 factors, again accounting for betweentrial variation. Time from symptom onset was treated as a discrete variable (ie, time from symptom onset to randomization b2 vs ≥2 hours), which was prespecified based on previous analyses of time from symptom onset and is in accord with current STEMI guidelines relating to reperfusion options and total ischemic time.1,2,5,9 The relationship between time from symptom onset, treatment, and 1-year mortality was also expressed with time in a continuous fashion. The linearity and proportional hazard assumptions

were evaluated. Adjustment of the relationship was performed via backward selection using an inclusion criterion of α = .05 and exclusion criterion of α = .10 (and confirmed with forward selection). Baseline patient characteristics considered in adjustment included age, sex, systolic blood pressure, heart rate, Killip class, hypertension, history of diabetes, history of angina, prior MI, previous PCI, current smoking status, and anterior MI. Because the DANAMI-2 study reported that the benefit observed for PCI in their study was restricted to patients with

American Heart Journal February 2011

286 Westerhout et al

Table II. Timing intervals CAPTIM

n Symptom onset to randomization, min b2 h, % ≥2 h, % Symptom onset to first medical contact, min First medical contact to randomization, min Symptom onset to FL, min Symptom onset to primary PCI, min Randomization to FL, min Randomization to primary PCI, min Length of stay, d

WEST

CAPTIM-WEST

FL

Primary PCI

FL

Primary PCI

All

FL

Primary PCI

419 107 (75.8-158.3) 55.3 44.7 78 (48-135)

421 108 (76-162) 55.0 45.0 77 (48-135)

221 105 (66-159) 60.2 39.8 52.5 (29-101.5)⁎

107 101 (70-163.8) 56.1 43.9 54 (29.3-108.8)⁎

1168 105.5 (73-161) 56.1 43.9 72 (42-130)

640 105 (72.5-158.5) 57.0 43.0 70 (40-125)

528 106 (74-161.5) 55.3 44.7 73 (44.8-131)

25 (18-34)

26 (19-35)

38 (28-59)⁎

39 (26-60)⁎

29 (20-40)

30 (21-41)

28 (20-38)

130 (95-180)

–

117 (75-179.3)⁎

–

–

122 (85-180)

–

–

–

–

9 (6-15)⁎

189 (150.5-293) –

–

16 (10-23)

190 (148.5-255) –

–

13 (8-21)

189 (148.3-256.8) –

–

72 (60-88)

–

90 (64.5-97.3)⁎

–

–

76 (62-94)

8 (6-11)

8 (6-10)

3 (2-5)

3 (2-5)

7 (3-10)

7 (3-10)

7 (4-9)

⁎ P b .05 for comparison between studies within treatment.

high Thrombolysis In Myocardial Infarction (TIMI) risk, we undertook a secondary analysis to examine this in high- versus low-risk patients as defined by the TIMI risk score (high risk ≥5 points, low risk b5 points).12,13 Additional analysis was undertaken on prespecified events, that is, recurrent MI and cardiogenic shock, and the incidence of major systemic bleeding and nonfatal intracranial hemorrhage (ICH) inhospital. All statistical comparisons were done at the 5% level of significance using a 2-sided alternative hypothesis, unless stated otherwise. Analyses were performed using SAS (version 9.2, SAS Institute, Cary, NC).

in both trials with 56% of patients randomized within 2 hours of symptom onset. Although there was a shorter interval from symptom onset to first medical contact in WEST than in CAPTIM (median 53 vs 78 minutes), CAPTIM's time from first medical contact to randomization was shorter at 26 minutes versus 38 minutes in WEST. Overall, time from symptom onset to treatment with FL was shorter in WEST compared to CAPTIM (P = .006); however, time to PCI was comparable between the studies (Table II). The median length of stay was 8 days in CAPTIM and 3 days in WEST.

No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses, and the drafting and editing of the paper and its final contents.

One-year mortality Overall, 63 (5.4%) of 1,160 patients with complete follow-up died within 1 year of randomization. There was no difference in overall 1-year mortality between FL and primary PCI ([29/633] 4.6% vs [34/527] 6.5%, unadjusted HR 0.75 [FL vs PCI], 95% CI 0.46-1.24, P = .546) (Figure 2). When 1-year mortality was further examined according to the TIMI risk score (high risk ≥5 points [17.8% of patients], low risk b5 points [82.2%]), mortality was comparable in low-risk patients (FL [12/513] 2.3% vs PCI [9/405] 2.2%, P = .899). Oneyear mortality was nominally higher in PCI than in FL in high-risk patients, but this did not reach statistical significance (FL [13/99] 13.1% vs PCI [18/101] 17.8%, P = .373). However, when mortality was examined according to time from symptom onset to randomization (Figures 3 and 4), the interaction was statistically significant (P = .043). Patients randomized within 2 hours of symptom onset had improved survival with FL compared to those receiving primary PCI ([10/358] 2.8% vs [20/288] 6.9%, P = .021, HR 0.43, 95% CI 0.20-

Results Individual patient data from the CAPTIM, which enrolled 840 patients with STEMI between 1997 and 2000, and WEST trials, which enrolled 328 patients with STEMI between 2003 and 2005, were obtained for the current study (Figure 1). Baseline patient characteristics according to study treatment are presented in Table I for the individual trials as well as the pooled cohort. Overall, the patients enrolled in the CAPTIM and WEST trials were comparable across the trials and between the study treatments: median age of 58 years, 81% men, 9% with prior MI, and median weight of 77 kg.

Time from symptom onset In Table II, the overall median time from symptom onset to randomization was 106 minutes and was similar

American Heart Journal Volume 161, Number 2

Figure 2

Kaplan-Meier curve of 1-year survival according to study treatment (P = .263) (FL, dashed; PCI, solid line).

Westerhout et al 287

Figure 4

Association between time from symptom onset and study treatment according to 1-year mortality.

Figure 3 Figure 5

Kaplan-Meier curve of 1-year survival according to study treatment and time from symptom onset (P = .021, FL b2 hours vs PCI b2 hours; P = .529, FL ≥2 hours vs PCI ≥2 hours) (FL b2 hours, dotted; FL ≥2 hours, dashed; PCI b2 hours, thick solid; PCI ≥2 hours, thin solid line).

0.91). Beyond 2 hours, however, no treatment difference in 1-year mortality was observed (FL [19/274] 6.9% vs PCI [14/234] 6.0%, P = .529, HR 1.23, 95% CI 0.61-2.46). After adjustment for age, systolic blood pressure, heart rate, diabetes, and Killip class, a significant interaction between time from symptom onset and study treatment remained (P = .037) such that FL patients had a (relative) 42% lower hazard of 1year mortality than PCI patients with symptom onset within 2 hours (ie, in time b2 hours, FL vs PCI, adjusted HR 0.58, 95% CI 0.26-1.29). In patients randomized beyond 2 hours, there appeared to be an excess hazard associated with FL versus PCI (ie, FL vs PCI, adjusted HR 1.81, 95% CI 0.87-3.77). When examining time from symptom onset in a continuous fashion (Figure 5), the point estimates for FL versus PCI suggest a survival benefit for FL until approximately 127 minutes from

Fibrinolysis versus PCI on 1-year mortality according to increasing time from symptom onset. The hazard of mortality for FL approximated that of PCI at 127 minutes (note that time b127 minutes, 59.1% of all patients; time b240 minutes, 91.6% of all patients). Hazard ratio indicated by thick solid line and 95% CI in thick gray lines.

when symptom onset had elapsed; thereafter, a reversal occurred with improved survival with PCI.

Inhospital and 30-day events and interventions Table III provides the 30-day occurrence of cardiogenic shock and re-MI and inhospital safety events. Although 30-day cardiogenic shock appeared to be higher in primary PCI patients than in FL patients, this did not achieve statistical significance (5.3% vs 3.8%, P = .203); re-MI, however, was over 2-fold higher in FL patients (2.0% vs 5.3%, P = .004). Intracranial hemorrhage was rare in this patient population. Although major systemic bleeds were nominally higher in primary PCI than in FL patients, this did not reach statistical significance. In FL

American Heart Journal February 2011

288 Westerhout et al

Table III. Cardiogenic shock, re-MI, ICH, and major systemic bleeding CAPTIM

n 30-day cardiogenic shock, n (%) Onset to randomization b2 h, n (%) Onset to randomization ≥2 h, n (%) 30-day re-MI, n (%) Onset to randomization b2 h, n (%) Onset to randomization ≥2 h, n (%) Inhospital ICH, n (%) Inhospital major systemic bleeding, n (%)

WEST

CAPTIM-WEST

FL

Primary PCI

FL

Primary PCI

All

FL

Primary PCI

419 12 (2.9) 5 (2.2) 7 (3.7) 15 (3.7) 9 (4.0) 6 (3.4) 2 (0.5) 2 (0.5)

421 19 (4.5) 11 (4.8) 8 (4.3) 7 (1.7) 3 (1.4) 4 (2.2) 0 (0.0) 8 (1.9)

221 12 (5.4) 9 (6.8) 3 (3.4) 18 (8.1) 11 (8.3) 7 (8.0) 0 (0.0) 3 (1.4)

107 9 (8.4) 6 (10.0) 3 (6.4) 3 (2.8) 3 (5.0) 0 (0.0) 0 (0.0) 1 (0.9)

1168 52 (4.5) 31 (4.7) 21 (4.1) 43 (3.8) 26 (4.1) 17 (3.5) 2 (0.2) 14 (1.2)

640 24 (3.8) 14 (3.8) 10 (3.6) 33 (5.3) 20 (5.6) 13 (4.9) 2 (0.3) 5 (0.8)

528 28 (5.3) 17 (5.9) 11 (4.7) 10 (2.0)⁎ 6 (2.2)⁎ 4 (1.8) 0 (0.0) 9 (1.7)

⁎ P b .05.

patients, rescue PCI occurred in 26% of patients and PCI within 30 days in 70% of patients.

Discussion Our novel findings not only support the critical relationship between time from symptom onset and 1year survival after reperfusion, but also provide new evidence about the impact of this relationship on the relative efficacy of the 2 standard modes of therapy. A survival advantage existed for patients treated with FL within 2 hours of symptom onset relative to those treated with primary PCI (P interaction [unadjusted] = .043). When time from symptom onset was explored in a continuous fashion, a progressive attenuation of the benefit of FL was observed until approximately 127 minutes had elapsed; thereafter, the survival advantage of PCI appeared and increased with time (Figure 5). What might account for the difference in our findings and those previously reported in a systematic overview and meta-analysis?3,5 A key distinguishing characteristic of the current versus prior trials comparing pharmacologic and mechanical reperfusion was the short time between symptom onset and randomization (including prehospital randomization) such that more than one half of the patients achieved this within 2 hours. The influence of time to treatment in myocardial salvage in patients treated with primary PCI versus FL has been examined by Schomig et al.14 Using myocardial scintigraphy to quantify the salvage index in 264 patients from 2 randomized trials, this group demonstrated that, although a similar salvage index occurred within the first tertile of time from symptom onset (approximately 165 minutes), there was a progressive decrease in salvage with FL versus PCI. Noteworthy and perhaps accounting for these findings was that median door-to-needle time was 35 minutes, that is, beyond current recommendations, whereas the door-to-balloon time was a remarkably short 65 minutes, thus resulting in an unusually brief half-hour difference between reperfusion strategies. Important additional distinguishing features of patients

in the CAPTIM-WEST analysis include systematic use of fibrin-specific pharmacologic therapy, rescue PCI in 26% of patients, and frequent cointervention with PCI in 70% of the fibrinolytic-treated patients by 30 days. The earlier time-to-treatment and more frequent cointervention likely account for the better outcome of fibrinolytictreated patients in the current study versus the DANAMI-2 trial; thereafter, mechanical rescue and cointervention were discouraged and infrequent in the fibrinolytictreated patients. Although prior data and guidelines suggest an advantage for PCI in high-risk patients, they represented a minority of the overall STEMI population and our combined cohort (17.8% of patients), whereas they represented 26% of the DANAMI-2 trial.12,13 Importantly, both the CAPTIM and WEST trials also used contemporary adjunctive antithrombotic and antiplatelet therapy in patients assigned to the primary PCI arms to ensure optimal overall results in both treatment arms. Our findings are in accordance with those of the GRACIA-2 investigators who demonstrated that routine stent/angioplasty within 3 to 12 hours of FL proved both safe and equivalent to primary stenting in preserving myocardial function.15 Although these investigators found better myocardial perfusion in the fibrinolytictreated patients, their study of 212 patients was underpowered to address whether there was associated clinical improvement and called for “a larger clinical outcome study for confirmation.” We observed an increase in the occurrence of inhospital re-MI among fibrinolytic-treated patients. This finding could relate not only to surveillance bias and the challenge of defining and detecting periprocedural MI after PCI for STEMI, but also to the possibility that early fibrinolytic-treated patients had more salvaged but subsequently jeopardized myocardium. Fibrinolytic therapy was well tolerated by this patient population with low ICH rates and minimal major bleeding that was numerically greater in the PCI group as has been observed previously.3 The low rate of ICH in the current report likely relates the relatively young age of our patients and the careful exclusion of prior

American Heart Journal Volume 161, Number 2

stroke, a feature that has not been uniformly applied in other studies. Our data are also consistent with recent observational reports from the French registry of acute ST-elevation myocardial infarction and respond to the recent call for confirmation of those findings in a randomized design.16,17 They are also aligned with the 3-state regional approach from the Mayo Clinic and the Vienna STEMI registry, which both attest to the excellent results after fibrinolytic administration in the early post–symptomonset period.18,19 The primacy of time from symptom onset as a modulator of outcome is underscored by the recent data of Francone et al,20 which showed marked attenuation of the potential for myocardial salvage in patients undergoing primary PCI N90 minutes after symptom onset. Hence, the relatively flat relationship between survival and time from symptom onset that we observed in primary PCI-treated patients is perhaps not surprising, given that, even under the optimal circumstances existing for timely access to expert PCI in both CAPTIM and WEST, the median time to achieve it was 189 minutes; hence, only one quarter of our patients underwent PCI within 148 minutes of symptom onset.

Limitations and strengths Some limitations of our analysis should be noted. Although combining these 2 trials was not prespecified, we believe it to be warranted given the demonstrated similar baseline characteristics and times to randomization and reperfusion. The populations studied were part of clinical trials; hence, caution regarding the generalizability of our results to a broader clinical population would be prudent. However, they are in accordance with substantial data acquired in registries.21 Although the interaction between treatment and delay was evaluated on 5-year mortality in the CAPTIM trial, this relationship did not achieve statistical significance.22 By merging the WEST and CAPTIM trials, we increased the size of the study population and events resulting in a statistically significant interaction for 1year mortality, indicating that we had sufficient power to test this association.

Conclusion Given persisting evidence of failure to meet guidelinesuggested times to PCI, especially among patients presenting to non-PCI hospitals, our data provide additional evidence to support the efficacy of an alternative reperfusion strategy, that is, fibrinolytic therapy (in patients without contraindications, coupled with contemporary adjunctive therapy, timely rescue PCI, and subsequent revascularization), especially in those presenting early after symptom onset either in the prehospital setting or in the non-PCI hospital setting. Such an approach is also likely to be welcome in

Westerhout et al 289

countries and regions where there is a relative paucity of PCI centers and where climate and geography may complicate immediate transfer for primary PCI.

Acknowledgements The authors would like to acknowledge the expert editorial assistance of Jo-An Padberg.

Disclosures The original WEST study was supported by an unrestricted grant from Hoffmann-LaRoche Limited; Aventis Pharma, a member of the Sanofi-Aventis group; as well as Eli-Lilly Canada. The original CAPTIM study was supported by a grant from the French Ministry of Health (Projet Hospitalier de Recherche Clinique, 96/045), by the Hospices Civils de Lyon, and by an unrestricted research grant from AstraZeneca France. Biotronic GmbH provided balloons and guidewires free of charge. Drs Armstrong and Welsh received Clinical Trial funding from Boehringer-Ingelheim.

References 1. Antman EM, Hand M, Armstrong PW, et al. 2007 focused update of the ACC/AHA 2004 guidelines for the management of patients with ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association task force on practice guidelines: developed in collaboration with the Canadian Cardiovascular Society endorsed by the American Academy of Family Physicians. Circulation 2008;117:296-329. 2. The Task Force on the management of ST-segment elevation acute myocardial infarction of the European Society of Cardiology. Management of acute myocardial infarction in patients presenting with persistent ST-segment elevation. Eur Heart J 2008;29:2909-45. 3. Keeley EC, Boura JA, Grines CL. Primary angioplasty versus intravenous thrombolytic therapy for acute myocardial infarction: a quantitative review of 23 randomized trials. Lancet 2003;361:13-20. 4. Willerson JT. Editor's commentary: one size does not fit all. Circulation 2003;107:2543-4. 5. Boersma E and the Primary Coronary Angioplasty versus Thrombolysis (PCAT)-2 trialists' collaborative group. Does time matter? A pooled analysis of randomized clinical trials comparing primary percutaneous coronary intervention and in-hospital fibrinolysis in acute myocardial infarction patients. Eur Heart J 2006;27:779-88. 6. Armstrong PW, Westerhout CM, Welsh RC. Duration of symptoms is the key modulator of the choice of reperfusion for ST-elevation myocardial infarction. Circulation 2009;119:1293-303. 7. Solis P, Amsterdam EA, Bufalino V, et al. Development of systems of care for ST-elevation myocardial infarction patients. Policy recommendations. Circulation 2007;116:e73-6. 8. Bonnefoy E, Lapostolle F, Leizorovicz A, et al, on behalf of the Comparison of Angioplasty and Prehospital Thrombolysis in Acute Myocardial Infarction (CAPTIM) study group. Primary angioplasty versus prehospital fibrinolysis in acute myocardial infarction: a randomized study. Lancet 2002;360:825-9. 9. Steg PG, Bonnefoy E, Chabaud S, et al, for the Comparison of Angioplasty and Prehospital Thrombolysis in acute Myocardial infarction (CAPTIM) Investigators. Impact of time to treatment on

American Heart Journal February 2011

290 Westerhout et al

10.

11.

12.

13.

14.

15.

16.

mortality after prehospital fibrinolysis or primary angioplasty. Data from the CAPTIM randomized clinical trial. Circulation 2003;108:2851-6. Armstrong PW, and the WEST steering committee. A comparison of pharmacologic therapy with/without timely coronary intervention vs primary percutaneous intervention early after ST-elevation myocardial infarction: the WEST (which early ST-elevation myocardial infarction therapy) study. Eur Heart J 2006;27:1530-8. Buller CE, Welsh RC, Westerhout CM, et al. Guideline adjudicated fibrinolytic failure: incidence, findings, and management in a contemporary clinical trial. Am Heart J 2008;155:121-7. Thune JJ, Hoefsten DE, Lindholm MG, et al, for the Danish Multicenter Randomized Study on Fibrinolytic Therapy Versus Acute Coronary Angioplast in Acute Myocardial Infarction (DANAMI)-2 Investigators. Simple risk stratification at admission to identify patients with reduced mortality from primary angioplasty. Circulation 2005;112:2017-21. Morrow DA, Antman EM, Parsons L, et al. Application of the TIMI risk score for ST-elevation MI in the National Registry of Myocardial Infarction 3. JAMA 2001;286:1356-9. Schomig A, Ndrepepa G, Mehilli J, et al. Therapy-dependent influence of time-to-treatment interval on myocardial salvage in patients with acute myocardial infarction treated with coronary artery stenting or thrombolysis. Circulation 2003;108:1084-8. Fernandez-Aviles F, Alonso JJ, Pena G, et al, for the GRACIA-2 Investigators. Primary angioplasty vs early routine post-fibrinolysis angioplasty for acute myocardial infarction with ST-segment elevation: the Gracia 2 non-inferiority, randomized, controlled trial. Eur Heart J 2007;28:949-60. Danchin N, Coste P, Ferrieres J, et al, for the FAST-MI Investigators. Comparison of thrombolysis followed by broad use of percutaneous

17.

18.

19.

20.

21.

22.

coronary intervention with primary percutaneous coronary intervention for ST-segment elevation acute myocardial infarction: data from the French registry in acute ST-elevation myocardial infarction (FASTMI). Circulation 2008;118:268-76. Nielsen PH, Maeng M, Busk M, et al, for the DANAMI-2 Investigators. Primary angioplasty versus fibrinolysis in acute myocardial infarction: long-term follow-up in the Danish acute myocardial infarction 2 trial. Circulation 2010;121:1484-91. Ting HH, Rihal CS, Gersh BJ, et al. Regional systems of care to optimize timeliness of reperfusion therapy for ST-elevation myocardial infarction: the Mayo Clinic STEMI protocol. Circulation 2007; 116:729-36. Kalla K, Christ G, Karnik R, et al, for the Vienna STEMI Registry Group. Implementation of guidelines improves the standard of care: the Viennese Registry on reperfusion strategies in ST-elevation myocardial infarction (Vienna STEMI Registry). Circulation 2006; 113:2398-405. Francone M, Bucciarelli-Ducci C, Carbone I, et al. Impact of primary coronary angioplasty delay on myocardial salvage, infarct size, and microvascular damage in patients with ST-segment elevation myocardial infarction: insight from cardiovascular magnetic resonance. J Am Coll Cardiol 2009;54:2145-53. Goldberg RJ, Spencer FA, Fox KAA, et al. Prehospital delay in patients with acute coronary syndromes (from the Global Registry of Acute Coronary Events [GRACE]). Am J Cardiol 2009;103: 598-603. Bonnefoy E, Steg PG, Boutitie F, et al, for the CAPTIM investigators. Comparison of primary angioplasty and pre-hospital fibrinolysis in acute myocardial infarction (CAPTIM) trial: a 5-year follow-up. Eur Heart J 2009;30:1598-606.

Has the ClOpidogrel and Metoprolol in Myocardial Infarction Trial (COMMIT) of early β-blocker use in acute coronary syndromes impacted on clinical practice in Canada? Insights from the Global Registry of Acute Coronary Events (GRACE) Jeremy Edwards, MD, a,i Shaun G. Goodman, MD, MSc, a,b,i Raymond T. Yan, MD, b,i Robert C. Welsh, MD, c,i Jan M. Kornder, MD, d,i J. Paul DeYoung, MD, e,i Denis Chauret, MD, f,i Jean-Pierre Picard, MD, g,i Kim A. Eagle, MD, h,i and Andrew T. Yan, MD a,b ,i Ontario, Alberta, British Columbia, and Quebec, Canada; and Ann Arbor, MI

Background The COMMIT/CCS-2 trial, published in 2005, demonstrated no net benefit of early β-blocker (BB) therapy in acute coronary syndromes (ACS). We sought to assess the short-term impact of this landmark trial by comparing the use of early BB therapy in patients with a broad spectrum of ACS before and after 2005. Methods Using data from the Global Registry of Acute Coronary Events and Canadian Registry of Acute Coronary Events, we compared the rates of BB use within the first 24 hours of presentation in the periods 1999 to 2005 and 2006 to 2008, after stratifying patients by the type of ACS (ST-segment elevation myocardial infarction [STEMI] and non–ST-segment elevation ACS [NSTEACS]) and clinical presentation. Results Of the 14,231 patients with ACS, 77.7% received BB therapy within 24 hours of presentation (78.5% and 77.4% in the STEMI and NSTEACS groups, respectively). The early use of BB declined in the STEMI group (80.3% to 76.7%, P = .005) but increased in the NSTEACS group (75.4% to 78.9%, P b .001) after 2005. Long-term BB use, higher systolic blood pressure, and higher heart rate were independent predictors of early BB use. Conversely, patients who were female, older, Killip class N1, and had cardiac arrest at presentation were less likely to receive early BB. Multivariable analysis showed a trend toward lower use of BB among patients with STEMI (adjusted odds ratio 0.76, 95% CI 0.57-1.00, P = .055) and a trend toward more frequent BB use among patients with NSTEACS (adjusted odds ratio 1.22, 95% CI 0.96-1.55, P = .11) after 2005. The temporal trends in the early use of BB differed between patients with STEMI and patients with NSTEACS (P for interaction with period b.001). Conclusions

Most patients with STEMI or NSTEACS were treated with early BB therapy. In accordance with the COMMMIT/CCS-2 trial, patients with lower systolic blood pressure and higher Killip class in the “real world” less frequently received early BB therapy. Since the publication of COMMIT/CCS-2, there has been no significant change in the use of BB in patients with STEMI or NSTEACS after controlling for their clinical characteristics. (Am Heart J 2011;161:291-7.)

a

From the Division of Cardiology, Terrence Donnelly Heart Centre, St. Michael's Hospital, University of Toronto, Toronto, Ontario, Canada, bCanadian Heart Research Centre, Toronto, Ontario, Canada, cMazankowksi Alberta Heart Institute, Edmonton, Alberta, Canada, dSurrey Memorial Hospital, Surrey, British Columbia, Canada, eCornwall Community Hospital, Cornwall, Ontario, Canada, fUniversity of Ottawa, Ottawa, Ontario, Canada, gHopital Hotel-Dieu de Sorel, Sorel-Tracy, Quebec, Canada, and hUniversity of Michigan Health System, Ann Arbor, MI. i For the Canadian Global Registry of Acute Coronary Events (GRACE) and Canadian Registry of Acute Coronary Events (CANRACE) investigators. Submitted March 3, 2010; accepted October 18, 2010. Reprint requests: Andrew T. Yan, MD, Division of Cardiology, St. Michael's Hospital, 30 Bond Street, Room 6-030 Queen, Toronto, Ontario, Canada M5B 1W8. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.034

β-Blocker (BB) therapy has been used in the management of acute coronary syndromes (ACS) for decades. The proposed mechanisms of benefit of BBs in the setting of acute myocardial infarction include decreased myocardial oxygen consumption due to effects on heart rate (HR), contractility, and afterload; antiarrhythmic effects; and improved myocardial oxygen supply secondary to a prolonged diastolic filling interval.1 Long-term beneficial effects include protection against adverse remodeling and declining ventricular function.2,3 In late 2005, the COMMIT/CCS-2 trial was published.4 Largely in response to the results of COMMIT/CCS-2, the American College of Cardiology (ACC) and the American Heart Association (AHA) updated their guidelines on the

American Heart Journal February 2011

292 Edwards et al

management of both ST-segment elevation myocardial infarction (STEMI) and non–ST-segment elevation ACS (NSTEACS) in 2007.5,6 The updated guidelines caution against the use of early BB therapy in patients with evidence of or risk factors for hemodynamic compromise. However, there are limited data regarding the impact of COMMIT/CCS-2 on the clinical practice in the real world.7 The primary goal of our study was to assess the shortterm impact of this landmark trial by comparing the patterns of early BB use in Canadian patients with ACS before and after publication of COMMIT/CCS-2 in 2005. The secondary objectives were to investigate (a) what patient factors may be associated with the early use of BB therapy and (b) the relationship between early BB therapy and in-hospital cardiovascular outcomes among unselected patients with ACS in the real world.

Methods Registry design and data collection The full details of the Global Registry of Acute Coronary Events (GRACE) design and methods have been previously published.8 Briefly, GRACE is a multicenter, multinational, prospective registry of a broad spectrum of patients with ACS that was expanded (GRACE2) in 2003.9 The inclusion criteria were age N18 years, alive at hospital presentation, and ACS was the provisional diagnosis with at least one of the following: elevated biomarkers, electrocardiographic changes indicative of ischemia or infarction, and/or known history of coronary artery disease. Patients were excluded if the ACS was precipitated by surgery or other comorbidity. The final ACS diagnosis was classified as STEMI, non–ST-segment elevation myocardial infarction (NSTEMI), or unstable angina. At each site, a trained coordinator collected data on standardized case report forms. Enrollment in GRACE/GRACE2 was completed in December 2007, but Canadian enrollment was extended in 2008 in the Canadian Registry of Acute Coronary Events (CANRACE), which had identical inclusion and exclusion criteria. Across Canada, 53 hospitals participated in GRACE/GRACE2/CANRACE. Of these hospitals, 35% had on-site cardiac catheterization laboratory and 21% had on-site coronary bypass grafting surgery capability.

Temporal trends of early BB therapy in ACS Early BB therapy was defined as (intravenous [IV] or oral) drug administration within 24 hours of ACS presentation. Of the 14,435 patients in GRACE/GRACE2/CANRACE, data on early BB use were available for 14,231 patients (98.6%). We compared BB use in 1999 to 2005 with use in 2006 to 2008, as the COMMIT/ CCS-2 trial was published in late 2005. Recognizing that there could have been a change in BB use over time, independent of the publication of COMMIT/CCS-2, we also repeated the comparison of BB use excluding the patients enrolled from 1999 to 2002. The COMMIT/CCS-2 study and the ACC/AHA guidelines for the management of STEMI and NSTEACS have identified the following as risk factors for hemodynamic and/or respiratory compromise with early BB use4-6: age N70 years, systolic blood pressure (SBP) b120 mm Hg, HR b60 or N110 beat/min, Killip class N1,10 and delayed presentation N13 hours. We specifically

looked at BB use in these subgroups, which were at risk for harm associated with early BB therapy.

Outcomes in relation to early BB use for ACS We compared patients who did and did not receive early BB therapy with respect to these prespecified outcomes: in-hospital mortality, re-infarction, cardiogenic shock, pulmonary edema and/or acute congestive heart failure (CHF), and sustained ventricular tachycardia or fibrillation. The definition of reinfarction was limited to a re-elevation in cardiac biomarkers 24 hours after presentation.11 We also sought to investigate cardiovascular outcomes in high-risk patients by stratifying patients by hemodynamic status, as mentioned above.

Statistical analysis Discrete variables are presented as frequencies or percentages. Continuous variables are presented as medians with interquartile ranges. We compared group differences in discrete and continuous variables using the χ2 and Mann-Whitney U tests, respectively. We developed a multivariable model to examine factors associated with early BB use, based on clinical considerations and prior studies, including the study by Emery et al7 on the use of BB in NSTEACS published in 2006. Generalized estimating equations were used to control for the clustering of patients within hospitals. The variables considered in the multivariable model were age, gender, myocardial infarction, CHF, long-term BB use, SBP, HR, Killip class, cardiac arrest, creatinine, type of ACS, ST-segment deviation, and initial biomarker status.7,12,13 We also included the period (1999-2005 vs 2006-2008) of enrollment in the multivariable model to assess whether the publication of COMMIT/CCS-2 might have impacted upon the use of early BB therapy. Because of the different trends in the use of BB in the NSTEACS and STEMI groups, we tested for their interaction with period of enrollment in the model. To assess for an independent association between early BB use and outcome, we adjusted for other evidence-based therapies and for components of the GRACE risk score. The GRACE score is based upon age, HR, SBP, serum creatinine, Killip class, cardiac arrest at presentation, the presence of STsegment deviation, and elevated cardiac enzymes. Statistical analysis was performed using SPSS v. 15.0 (SPSS, Inc, Chicago, IL), and 2-sided P b .05 was considered to be significant. This research was supported by an unrestricted grant from Sanofi-Aventis, Paris, France, and Laval, Quebec, Canada, and by Bristol Myers Squibb Canada, Montreal, Quebec, Canada. The industrial sponsors had no involvement in the study conception or design; collection, analysis, and interpretation of data; writing, review, or approval of the manuscript; and decision to submit the manuscript for publication. The authors are solely responsible for the design and conduct of this study, all study analyses, and the drafting and editing of the paper and its final contents.

Results Patient characteristics Between 1999 and 2008, there were 14,231 Canadian patients with ACS in GRACE/GRACE2/CANRACE with complete data on early BB use. Overall, 77.7% of patients received early BB therapy. The rates of early BB use were

American Heart Journal Volume 161, Number 2

Edwards et al 293

Table I. Baseline demographic and clinical characteristics

Age, y⁎ Female Medical history Current smoker Diabetes Hypertension Angina Dyslipidemia Peripheral vascular disease TIA/stroke Renal insufficiency MI PCI CABG Preexisting heart failure Atrial fibrillation Long-term BB use Clinical presentation Systolic BP, mm Hg⁎ Diastolic BP, mm Hg⁎ HR, beat/min⁎ Body mass index⁎ Killip class Killip I Killip II Killip III Killip IV Cardiac arrest ST elevation ST depression Significant Q wave Positive initial cardiac biomarker Creatinine, μmol/L⁎ GRACE risk score⁎

Table II. In-hospital management

No early BB therapy (n = 3175)

Early BB therapy (n = 11056)

P

71 (59-80) 38%

66 (56-76) 32%

b.001 b.001

25% 28% 60% 42% 50% 10%

27% 27% 60% 44% 54% 8.4%

.031 .80 .93 .060 b.001 .004

11% 13% 32% 15% 11% 14%

8.6% 10% 33% 18% 13% 10%

b.001 b.001 .20 b.001 .032 b.001

12% 16%

8.6% 40%

b.001 b.001

140 (120-160) 77 (65-89)

144 (126-162) 80 (70-92)

b.001 b.001

78 (65-95) 26.7 (23.7-30.5)

78 (66-92) 27.5 (24.8-31.1)

.39 b.001 b.001

77% 14% 8.6% 0.7% 2.9% 24% 28% 8.7% 46%

86% 9.6% 4.4% 0.3% 1.1% 25% 30% 11% 48%

b.001 .064 .021 .003 .018

89 (69-116) 135 (109-169)

87 (70-107) 125 (101-153)

b.001 b.001

CABG, Coronary artery bypass graft surgery; MI, myocardial infarction; PCI, percutaneous coronary intervention; BP, blood pressure. ⁎ Median (25th and 75th percentiles).

78.5% and 77.4% in the STEMI and NSTEACS groups, respectively (P = .15). Table I summarizes the relevant demographics and clinical features of the study population.

In-hospital management Patients who received early BB therapy were more likely to receive other evidence-based therapies and to undergo cardiac catheterization and percutaneous coronary intervention during index hospitalization (Table II). They were less likely to be treated with an IV inotropic agent, calcium-channel blocker, and diuretic during the first 24 hours of presentation.

Medication use within first 24 h Aspirin Clopidogrel Warfarin Any heparin Glycoprotein IIb/IIIa inhibitor Thrombolytic ACE inhibitor Angiotensin receptor blocker Calcium-channel blocker IV inotropic agent Statin Nitrates Diuretic In-hospital management Cardiac catheterization PCI CABG LVEF assessment LV systolic function Normal Mildly diminished Moderately to severely diminished Length of stay, d⁎

No early BB therapy (n = 3175)

Early BB therapy (n = 11056)

P

83 52 5.6 81 6.9 11 37 12 27 6.6 46 59 33

94 67 5.0 89 10 14 60 11 20 2.1 73 72 28

b.001 b.001 .20 b.001 b.001 b.001 b.001 .092 b.001 b.001 b.001 b.001 b.001

53 31 2.7 59

62 34 3.5 67

b.001 .001 .042 b.001 .015

49 29 22

52 30 19

5 (3-9)

5 (3-8)

.075

Data shown in percentages, unless otherwise indicated. LVEF, Left ventricular ejection fraction; LV, left ventricular. ⁎ Median (25th and 75th percentiles).

Early BB use before and after the publication of COMMIT/CCS-2 Overall, early BB use was slightly higher among patients who were enrolled in 2006 to 2008 than among those enrolled in 1999 to 2005 (78.4% vs 76.9%, P = .039) (Fig. 1). The frequency of early BB use in patients with STEMI decreased in the years 2006 to 2008 compared with 1999 to 2005, whereas the frequency of early BB use rose in patients with NSTEACS (Fig. 1). There were no significant temporal changes in early BB use in subgroups identified in COMMIT/CCS-2 to be at risk for an adverse event associated with early BB therapy: age N70 years, SBP b120 mm Hg, Killip class N1, HR ≥100 beat/min, or time to presentation N13 hours (Fig. 2). There was no significant change in the results when the analysis was repeated excluding the patients from 1999 to 2002. Comparing the period 2003 to 2005 versus 2006 to 2008, there were significant increases in the early use of aspirin (92% to 93%, P = .014), clopidogrel (54% to 74%, P b .001), angiotensin receptor blockers (9% to 12%, P b .001), and statins (64% to 74%, P b .001). There was no significant change in angiotensin-converting enzyme inhibitor use between the 2 periods.

American Heart Journal February 2011

294 Edwards et al

Figure 1

Early BB use by ACS type before and after publication of COMMIT/ CCS-2.

Outcomes according to early BB use The group of patients who did not receive early BB therapy had significantly higher unadjusted in-hospital mortality rate than those who received early BB therapy (8.4% vs 2.2%, P b .001). After adjusting for components of the GRACE score, early BB use was not independently associated with higher in-hospital mortality (odds ratio [OR] 0.54, 95% CI 0.40-0.72, P b .001). Multivariable analysis Long-term BB use, higher HR and SBP, positive biomarkers, and ST-segment deviation at the time of presentation were independent predictors of early BB use (Table III). Conversely, female gender, increasing age, cardiac arrest, and Killip class N1 were independent negative predictors of early BB therapy. After adjusting for patient characteristics in multivariable analysis, there was no significant change in the overall use of early BB therapy in the years 2006 to 2008 compared to 1999 to 2005. However, the temporal trends in the early use of BB differed between patients with STEMI and patients with NSTEACS (P for interaction with period b.001)—there was a trend toward increased use of early BB therapy in patients with NSTEACS and a trend toward decreased use in patients with STEMI. The multivariable analysis was repeated excluding the patients enrolled between 1999 and 2002. For the NSTEACS group, there was no significant change in the use of early BB therapy between the periods 2003 to 2005 and 2006 to 2008 (adjusted OR 1.11, 95% CI 0.87-1.42, P = .40). Similarly, in the STEMI group, there was no significant change either in early BB use between the 2 periods (adjusted OR 0.75, 95% CI 0.56-1.01, P = .057).

Discussion This study demonstrates that most patients with ACS received early BB therapy in the past decade. In accordance with the updated ACC/AHA guidelines for the management of patients with STEMI, patients who were hypotensive or had evidence of heart failure at presentation were less often treated with early BB therapy. Since the publication of COMMIT/CCS-2 in 2005, there has been no significant change in early BB use among patients with STEMI or NSTEACS independent of their clinical characteristics. There have been multiple randomized controlled trials investigating BB use in patients with ACS.14-16 Yet, the optimal timing to initiate BB therapy has been controversial. The effect of early BB therapy on mortality was directly assessed in the COMMIT/CCS-2 trial, which was published in 2005.4 The COMMIT/CCS-2 trial randomized 45,852 patients with suspected ACS and ST-segment deviation or left bundle branch block to receive metoprolol or placebo. Patients in the metoprolol arm received up to 15 mg of IV metoprolol in 3 divided doses at 2- to 3-minute intervals, followed by 200 mg of metoprolol daily. There was no significant difference in mortality between the metoprolol and placebo groups. In the metoprolol arm, there was a trend toward increased mortality in the subgroup of patients who had evidence of hemodynamic instability at presentation, which was opposite to a trend toward decreased mortality in patients who were hemodynamically stable at presentation.4,17 In 2007, the ACC/AHA updated the 2004 STEMI management guidelines to caution against the early use of both oral and IV BB therapy in patients with signs of heart failure, evidence of a low output state, increased risk for cardiogenic shock, or other contraindications to BB therapy.5,18 American College of Cardiology/American Heart Association included a similar caution in their guidelines for the management of NSTEACS.6

Factors associated with early BB therapy Our study found that between 1999 and 2008, 78% of the Canadian patients with ACS enrolled in GRACE/ GRACE2/CANRACE received early BB therapy. This is comparable to the results of Emery et al7 who examined the early use of BB in patients with NSTEMI in the international GRACE up to the year 2004 (before the publication of COMMIT/CCS-2) and found that 76% received early BB therapy. However, Emery et al7 did not specifically investigate the use of early BB therapy in hypotensive patients or in patients with STEMI. In our study, patients who did not receive early BB therapy were older and more likely to be female, similar to other studies investigating the use of evidence-based therapies in patients with ACS.7,19-25 Long-term BB use was much higher in the group that received early BB therapy, suggesting that prior BB use may have

American Heart Journal Volume 161, Number 2

Edwards et al 295

Figure 2

Early BB use according to baseline characteristics, before and after publication of COMMIT/CCS-2.

Table III. Independent predictors of early BB use

Age, per decade increase Female HR 60-109 beat/min b60 beat/min ≥110 beat/min SBP b120 mm Hg Positive initial biomarker History of MI History of CHF Killip class I II III or IV Creatinine, per 10 μmol/L increase Cardiac arrest Long-term BB use ST-segment deviation Long-term BB use and history of MI⁎ Long-term BB use and history of CHF⁎ STEMI† 1999-2005 2006-2008 NSTEACS† 1999-2005 2006-2008

OR

95% CI

P

0.82 0.78

0.78-0.86 0.67-0.91

b.001 .002

reference 0.62 0.93 0.69 1.36 0.72 0.65

0.55-0.71 0.79-1.10 0.60-0.79 1.20-1.54 0.61-0.85 0.52-0.82

b.001 .41 b.001 b.001 b.001 b.001

reference 0.68 0.54 0.98 0.46 4.17 1.15 1.71 1.37

0.57-0.82 0.40-0.71 0.97-0.99 0.30-0.70 2.58-6.74 1.00-1.31 1.29-2.26 1.06-1.77

b.001 b.001 b.001 b.001 b.001 .047 b.001 .017

reference 0.76

0.57-1.00

.055

reference 1.22

0.96-1.55

.11

⁎ Interaction terms. † P for interaction with period b.001.

influenced physicians' decision to administer early BB therapy. Possibly, physicians felt reassured that the patient was tolerating BB therapy or concerned that discontinuing therapy may lead to worse outcomes, with

higher reinfarction rates and larger infarct sizes. Increasing HR and SBP were both found to be independent predictors of early BB use, a result that may be intuitive given that BBs lower both HR and SBP and may result in hemodynamic compromise in bradycardic and hypotensive patients. Conversely, cardiogenic shock and increasing Killip class were both negative predictors of early BB use.

Impact of the publication of COMMIT/CCS-2 on the use of early BB therapy We demonstrated that, overall, there was no significant change in the use of early BB therapy after the publication of COMMIT/CCS-2. In COMMIT/CCS-2, the investigators found no net benefit of early BB therapy. Although the trial included both patients with STEMI and NSTEMI with ST-segment deviation, most (93%) of the patients in COMMIT/CCS-2 had STEMI.4 Thus, the publication of COMMIT/CCS-2 may affect early BB use in patients with STEMI differently compared to patients with NSTEACS. Indeed, we found that there has been a nonsignificant increasing trend in early BB use among patients with NSTEACS, which is opposite to the nonsignificant decreasing trend in BB use among patients with STEMI. This temporal difference between the 2 groups was significant (P for interaction b.001). In COMMIT/CCS-2, the authors suggest that patients with hemodynamic compromise were most susceptible to adverse outcomes with early BB therapy. They found that patients who had an SBP b120 mm Hg or Killip class III had an increased risk of developing cardiogenic shock when allocated to the metoprolol group.4 In our study, we stratified patients based upon age N70 years, SBP b120 mm

American Heart Journal February 2011

296 Edwards et al

Hg, HR b60 or N110 beat/min, Killip class N1,10 and delayed presentation N13 hours. We found that, despite the publication of COMMIT/CCS-2 in 2005, there was no significant change in the use of early BB therapy in these subgroups who may be at increased risk for an adverse outcome associated with early BB therapy. Somewhat surprisingly, there may have been a trend toward an increased use of early BB therapy in these subgroups. Whether this trend to increase use of early BB therapy in high-risk patients is inappropriate deserves further study.

Outcomes of early BB use One of the criticisms about COMMIT/CCS-2 is that the dose of the metoprolol may have been excessive, and all patients in the BB arm received IV metoprolol in addition to oral metoprolol. In contrast, in GRACE/GRACE2/ CANRACE, physicians were not constrained by predefined medication route or dosage, as might be in a randomized controlled trial. As such, if physicians elected to use BB within the first 24 hours, they would have been able to tailor the dosage and route to individual patients. This may have attenuated any significant increase in the risk of adverse events associated with the early and aggressive use of BB in COMMIT/CCS-2. Our observational study was not designed to assess the treatment efficacy or risk of harm of early BB therapy. Such assessment is best accomplished via randomized controlled trials such as COMMIT/CCS-2. Nevertheless, randomized controlled trials may be less generalizable to unselected patients and are less useful for assessing treatment effectiveness (eg, with a titrated dose and different route of administration) in the real world. In our study, we found that patients who received early BB therapy had a significantly lower in-hospital mortality rate, even in patients who had an SBP b120 mm Hg or Killip class N1. There was also a significantly lower incidence of cardiogenic shock, sustained ventricular arrhythmia, and acute CHF in patients who received early BB therapy. Miller et al12 also reported an independent association between BB use and better clinical outcomes in the large CRUSADE database, and similar results have been shown in other observational studies.26-29 Based upon these observational data, early BB therapy did not appear to be associated with worse outcomes, suggesting that physicians may have carefully selected patients for early BB therapy use and tailored treatment appropriately. Importantly, our observational study serves as an illustrative example of how even large and wellconducted pragmatic trials may have limited impact on real-world practice. Several large pragmatic trials have revolutionized patient care in ACS.30,31 However, occasionally, when pragmatic trials (such as COMMIT) fail to demonstrate treatment efficacy, it may merely reflect failure to appropriately tailor therapy, which mandates careful patient selection and dose titration to

be effective. It is plausible that physicians consider the findings of the COMMIT inapplicable to their selected patients with ACS who are treated with different regimens of BB.

Study limitations Several limitations in this study should be mentioned. First, our study was a retrospective analysis using data from large registries. As such, this observational study cannot assess treatment efficacy and establish any causeand-effect relationship. However, it does allow for accurate assessment of real-world practice patterns. Second, we did not collect data on the route of administration and the dose of early BB therapy, which might vary according to the physician's practice and the patient's clinical presentation. Third, we did not specifically explore the reasons for administering or withholding BB therapy (eg, history of asthma or chronic obstructive pulmonary disease). Fourth, there may be a longer time lag beyond 2008 from the publication of COMMIT/CCS-2 to a potential change in clinical practice. Nevertheless, our study may serve as an important benchmark for future analysis of usage trends. Finally, although the registries aimed to recruit unselected consecutive patients, we could not determine how successful the registries were at attaining this goal. For example, patients dying early (before or shortly after admission) were more likely to be excluded because of insufficient time to obtain informed consent, leading to a selection bias toward less sick patients.

Conclusion Most Canadian patients with ACS were treated with early BB therapy over the past decade, irrespective of the type of ACS. In accordance with the ACC/AHA NSTEACS and STEMI management guidelines, we found that patients with lower SBP and higher Killip class in the real world less frequently received early BB therapy. Patients who were already on BB therapy at the time of presentation were more likely to receive early BB therapy. Since the publication of COMMIT/CCS-2 in 2005, there has been no significant change in the use of BB in patients with STEMI or NSTEACS after controlling for clinical characteristics.

Acknowledgements We thank Sue Francis for her assistance with manuscript preparation and all the study investigators, coordinators, and patients who participated in GRACE/ GRACE-2 and CANRACE. Dr. Andrew Yan is supported by a New Investigator Award from the Heart and Stroke Foundation of Canada.

American Heart Journal Volume 161, Number 2

Edwards et al 297

References 1. Lopez-Sendon J, Swedberg K, McMurray J, et al. Expert consensus document on beta-adrenergic receptor blockers. Eur Heart J 2004; 25:1341-62. 2. Doughty RN, Whalley GA, Walsh HA, et al. Effects of carvedilol on left ventricular remodeling after acute myocardial infarction: the CAPRICORN echo substudy. Circulation 2004; 109:201-6. 3. Galcerá-Tomás J, Castillo-Soria FJ, Villegas-García MM, et al. Effects of early use of atenolol or captopril on infarct size and ventricular volume: a double-blind comparison in patients with anterior acute myocardial infarction. Circulation 2001;103:813-9. 4. Chen ZM, Pan HC, Chen YP, et al. COMMIT (ClOpidogrel and Metoprolol in Myocardial Infarction Trial) collaborative group. Early intravenous then oral metoprolol in 45,852 patients with acute myocardial infarction: randomised placebo-controlled trial. Lancet 2005;366:1622-32. 5. Antman EM, Hand M, Armstrong PW, et al. 2007 focused update of the ACC/AHA 2004 guidelines for the management of patients with ST-elevation myocardial infarction. Circulation 2008; 117:296-329. 6. Anderson JL, Adams CD, Antman EM, et al. ACC/AHA 2007 guidelines for the management of patients with unstable angina/ non–ST-elevation myocardial infarction. J Am Coll Cardiol 2007;50: e1-e157. 7. Emery M, López-Sendón J, Steg PG, et al. Patterns of use and potential impact of early beta-blocker therapy in non–ST-elevation myocardial infarction with and without heart failure: the Global Registry of Acute Coronary Events. Am Heart J 2006;152: 1015-21. 8. The GRACE Investigators. Rationale and design of the GRACE (global registry of acute coronary events) project: a multinational registry of patients hospitalized with acute coronary syndromes. Am Heart J 2001;141:190-9. 9. Goodman SG, Huang W, Yan AT, et al. The expanded Global Registry of Acute Coronary Events: baseline characteristics, management practices, and hospital outcomes of patients with acute coronary syndromes. Am Heart J 2009;158:193-201. 10. Killip T, Kimball JT. Treatment of myocardial infarction in a coronary care unit. A two-year experience with 250 patients. Am J Cardiol 1967;20:457-64. 11. Yan AT, Steg PG, FitzGerald G, et al. Recurrent ischemia across the spectrum of acute coronary syndromes: prevalence and prognostic significance of (re-)infarction and ST-segment changes in a large contemporary registry. Int J Cardiol 2010;145:15-20. 12. Miller CD, Roe MT, Mulgund J, et al. Impact of acute beta-blocker therapy for patients with non–ST-segment elevation myocardial infarction. Am J Med 2007;120:685-92. 13. Silvet H, Spencer F, Yarzebski J, et al. Community-wide trends in the use and outcomes associated with beta-blockers in patients with acute myocardial infarction: the Worcester Heart Attack Study. Arch Intern Med 2003;163:2175-83. 14. Freemantle N, Cleland J, Young P, et al. Beta-blockade after myocardial infarction: systematic review and meta regression analysis. BMJ 1999;318:1730-7. 15. Roberts R, Rogers WJ, Mueller HS, et al. Immediate versus deferred beta-blockade following thrombolytic therapy in patients with acute myocardial infarction. Results of the Thrombolysis In

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Myocardial Infarction (TIMI) II-B study. Circulation 1991;83: 422-37. Randomised trial of intravenous atenolol among 16027 cases of suspected acute myocardial infarction: ISIS-1. First International Study of Infarct Survival collaborative group. Lancet 1986;2: 57-66. Roe MT, Chen AY, Riba AL, et al. Impact of congestive heart failure in patients with non–ST-segment elevation acute coronary syndromes. Am J Cardiol 2006;97:1707-12. Salpeter SR, Ormiston TM, Salpeter EE. Cardioselective beta-blockers for chronic obstructive pulmonary disease. Cochrane Database of Syst Rev 2005:CD003566, doi:10.1002/14651858.CD003566. pub2. Yan AT, Yan RT, Tan M, et al. Management patterns in relation to risk stratification among patients with non–ST-elevation acute coronary syndromes. Arch Intern Med 2007;167:1009-16. Mehta RH, Roe MT, Chen AY, et al. Recent trends in the care of patients with non–ST-segment elevation acute coronary syndromes: insights from the CRUSADE initiative. Arch Intern Med 2006;166: 2027-34. Spencer F, Scleparis G, Goldberg RJ, et al. Decade-long trends (1986 to 1997) in the medical treatment of patients with acute myocardial infarction: a community-wide perspective. Am Heart J 2001;142: 594-603. Eagle KA, Kline-Rogers E, Goodman SG, et al. Adherence to evidence-based therapies after discharge for acute coronary syndromes: an ongoing prospective, observational study. Am J Med 2004;117:73-81. Yan AT, Yan RT, Tan M, et al. Optimal medical therapy at discharge in patients with acute coronary syndromes: temporal changes, characteristics, and 1-year outcome. Am Heart J 2007;154: 1108-15. Gurwitz JH, Goldberg RJ, Chen Z, et al. Beta-blocker therapy in acute myocardial infarction: evidence for underutilization in the elderly. Am J Med 1992;93:605-10. Rathore SS, Mehta RH, Wang Y, et al. Effects of age on the quality of care provided to older patients with acute myocardial infarction. Am J Med 2003;114:307-15. Cuculi F, Radovanovic D, Pedrazzini G, et al. Is pretreatment with beta-blockers beneficial in patients with acute coronary syndrome? Cardiology 2009;115:91-7. Granger CB, Steg PG, Peterson E, et al. Medication performance measures and mortality following acute coronary syndromes. Am J Med 2005;118:858-65. Krumholz HM, Radford MJ, Wang Y, et al. Early beta-blocker therapy for acute myocardial infarction in elderly patients. Ann Intern Med 1999;131:648-54. Allen LaPointe NM, Chen AY, Roe MT, et al. Relation of patient age and mortality to reported contraindications to early beta-blocker use for non–ST-elevation acute coronary syndrome. Am J Cardiol 2009; 104:1324-9. Tunis SR, Stryer DB, Clancy CM. Practical clinical trials: increasing the value of clinical research for decision-making in clinical and health policy. JAMA 2003;290:1624-32. Baigent C, Collins R, Appleby P, et al. ISIS-2: 10-year survival among patients with suspected acute myocardial infarction in randomised comparison of intravenous streptokinase, oral aspirin, both, or neither. The ISIS-2 (Second International Study of Infarct Survival) collaborative group. BMJ 1998;316:1337-43.

Incidence and clinical consequences of acquired thrombocytopenia after antithrombotic therapies in patients with acute coronary syndromes: Results from the Acute Catheterization and Urgent Intervention Triage Strategy (ACUITY) trial Adriano Caixeta, MD, PhD, a,c,j George D. Dangas, MD, PhD, a,b,j Roxana Mehran, MD, a,b,j Frederick Feit, MD, e,j Eugenia Nikolsky, MD, PhD, a,c,j Alexandra J. Lansky, MD, d,j Jiro Aoki, MD, PhD, a,c,j Jeffrey W. Moses, MD, a,c,j Steven R. Steinhubl, MD, f,j Harvey D. White, DSc, g,j E. Magnus Ohman, MD, h,j Steven V. Manoukian, MD, i,j Martin Fahy, MSc, c,j and Gregg W. Stone, MD a,c,j New York, NY; New Haven, CT; Lexington, KY; Auckland, New Zealand; Durham, NC; and Nashville, TN

Background The aim of the study was to investigate the incidence and clinical consequences of acquired thrombocytopenia in patients with acute coronary syndromes (ACS) in the ACUITY trial. Methods We examined 10,836 patients with ACS randomized to receive heparin plus glycoprotein (GP) IIb/IIIa inhibitor, bivalirudin plus GP IIb/IIIa inhibitor, or bivalirudin monotherapy. Results

Acquired thrombocytopenia developed in 740 (6.8%) patients; mild (100,000-150,000 platelets/mm3), moderate (50,000-100,000 platelets/mm3), and severe (b50,000 platelets/mm3) developed in 656 (6%), 51 (0.5%), and 33 (0.3%) patients, respectively. Patients with acquired thrombocytopenia, compared with those without, were more likely to develop major bleeding (14% vs 4.3%, P b .0001) at 30 days and had higher rates of mortality (6.5% vs 3.4%, P b .0001) at 1 year. By multivariate analysis, acquired thrombocytopenia was an independent predictor of major bleeding at 30 days (hazard ratio [HR] 1.68, 95% CI 1.04-2.72, P = .03). Moderate and severe acquired thrombocytopenia were predictors of mortality at 1 year (HR 2.89, 95% CI 0.92-9.06, P = .06, and HR 3.41, 95% CI 1.09-10.68, P = .03, respectively). Compared to heparin plus GP IIb/IIIa inhibitor, bivalirudin monotherapy was associated with less declines in platelet count by N25% (7.6% vs 5.6%, P = .0009) and N50% (1.4% vs 0.7%, P = .004) from baseline.

Conclusions Acquired thrombocytopenia occurs in approximately 1 in 14 patients with ACS treated with antithrombin and antiplatelet medications and is strongly associated with hemorrhagic and ischemic complications. Compared to an anticoagulant regimen including a GP IIb/IIIa inhibitor, administration of bivalirudin monotherapy appears to be associated with less frequent declines in platelet count. (Am Heart J 2011;161:298-306.e1.)

From the aCardiovascular Research Foundation, New York, NY, bMount Sinai Medical Center, New York, NY, cColumbia University Medical Center, New York, NY, dYale University Medical Center, New Haven, CT, eNew York University School of Medicine, New York, NY, fUniversity of Kentucky, Lexington, KY, gAuckland City Hospital, Auckland, New Zealand, hDuke University Medical Center, Durham, NC, and iSarah Cannon Research Institute and the Hospital Corporation of America, Nashville, TN. j For the ACUITY trial investigators. RCT reg #NCT00093158. Submitted March 8, 2010; accepted October 29, 2010. Reprint requests: George D. Dangas, MD, PhD, Cardiovascular Institute (Box 1030), Mount Sinai Medical Center, One Gustave L. Levy Place, New York, NY 10029. E-mails: [email protected], [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.035

Current guidelines for patients with moderate- or high-risk acute coronary syndromes (ACS) recommend early invasive management with concomitant antithrombotic therapy, including aspirin, clopidogrel, heparin plus glycoprotein (GP) IIb/IIIa inhibitor, or, as an alternative, bivalirudin.1,2 However, the antithrombotic therapy combination of heparin and GP IIb/IIIa inhibitor use may cause acquired thrombocytopenia and has been strongly associated with increased risks of hemorrhagic and ischemic complications, as well as early and late mortality.3-9 The direct thrombin inhibitor bivalirudin (Angiomax; The Medicines Company, Parsippany, NJ) is an antithrombotic drug, indicated for

American Heart Journal Volume 161, Number 2

Caixeta et al 299

Table I. Baseline and procedural characteristics of patients with and without thrombocytopenia

Age (y), median (IQR) Male Renal insufficiency⁎ Diabetes mellitus Current smoking Previous MI Previous PCI Previous CABG Hypertension Hyperlipidemia Platelet count (×103 platelets/mm3), median (IQR) Creatinine clearance (mL/min), median (IQR) Cardiac biomarker elevation† TIMI risk score Low (0-2) Intermediate (3-4) High (5-7) Medications before hospitalization Aspirin Thienopyridines Antithrombin medications Prerandomization UFH LMWH Antiplatelet medication – Preintervention Aspirin Thienopyridine Randomization Heparin plus GP IIb/IIIa inhibitor Bivalirudin plus GP IIb/IIIa inhibitor Bivalirudin monotherapy Duration from first study drug to first actual PCI (h), median (IQR) Treatment strategy triage PCI Medical therapy

Thrombocytopenia (n = 740)

No thrombocytopenia (n = 10096)

P

67.00 (57.50-74.00) 76.5% 25.1% 27.5% 27.4% 36.9% 43.5% 25.9% 71.1% 60.3% 173.00 (160.00-192.50) 80.11 (59.95-106.48) 59.5%

62.00 (53.00-71.00) 67.6% 18.5% 26.6% 29.8% 31.1% 40.5% 18.4% 66.4% 57.2% 236.00 (202.00-278.00) 88.25 (66.30-114.10) 57.5%

b.0001 b.0001 b.0001 .57 .18 .001 .11 b.0001 .009 .11 b.0001 b.0001 .31

10.1% 54.5% 35.4%

17.2% 54.9% 27.9%

b.0001 .87 b.0001

72.3% 24.2%

69.7% 25.1%

.13 .63

62.0% 39.9% 24.3%

64.2% 40.3% 25.9%

.25 .81 .34

97.7% 67.7%

98.0% 65.6%

.49 .26

34.7% 34.5% 30.8% 4.93 (1.35-20.47)

32.9% 33.4% 33.8% 4.00 (1.33-19.23)

.29 .54 .11 .07

75.9% 24.1%

62.8% 37.2%

b.0001 b.0001

IQR indicates interquartile range; heparin, unfractionated heparin (UFH) or low molecular weight heparin (LMWH) at site discretion. ⁎ Renal insufficiency was defined as a calculated creatinine clearance rate of b60 mL/min as determined by the Cockcroft-Gault equation. † Creatine kinase–MB/troponin I or T elevated.

use in patients with heparin-induced thrombocytopenia undergoing angioplasty,1,2 that has been recently used in patients with ACS and those undergoing percutaneous coronary intervention (PCI).10 The incidence of thrombocytopenia after bivalirudin treatment was significantly lower than heparin plus GP IIb/IIIa inhibitor treatment in REPLACE-2 trial (0.7% vs 1.7%), which included relatively low-risk patients undergoing PCI.10 It is unknown whether bivalirudin reduces the incidence of acquired thrombocytopenia in moderateto high-risk patients with ACS, for whom early invasive therapy is planned, and whether the severity of acquired thrombocytopenia impacts on clinical outcomes. Furthermore, in addition to possible direct effects, declines in platelet count during ACS may also have implications for antiplatelet medication, prescription, and adherence.

The present study evaluated the incidence and clinical consequences of acquired thrombocytopenia in the largescale, contemporary, randomized ACUITY trial.

Methods ACUITY was a prospective, open-label, randomized, multicenter trial. The study design and principal results have been previously described in detail.11 In brief, 13,819 patients with moderate- and high-risk ACS were assigned by a primary randomization to 1 of 3 antithrombin regimens started before angiography: heparin (unfractionated heparin or enoxaparin) plus GP IIb/IIIa inhibitor, bivalirudin plus GP IIb/IIIa inhibitor, or bivalirudin monotherapy. In the bivalirudin monotherapy group, provisional GP IIb/IIIa inhibitor was used in 7% of the patients.11 Angiography was performed in all patients within 72 hours after randomization. Patients were then triaged to PCI, coronary artery bypass grafting (CABG), or medical

American Heart Journal February 2011

300 Caixeta et al

Table II. Thirty-day and 1-year adverse clinical events Thrombocytopenia (n = 740) 30 Days NACE Composite ischemia Death/MI Death Cardiac death Noncardiac death MI Q wave MI Non–q wave MI Unplanned revascularization Non-CABG major bleeding Non-CABG major bleeding (excluding hematoma ≥ 5 cm) Non-CABG related blood transfusion Non-CABG minor bleeding TIMI non-CABG major bleeding TIMI non-CABG minor bleeding Definite/probable stent thrombosis 1 Year Composite ischemia Death/MI Death Cardiac death Noncardiac death MI Q-wave MI Non–Q-wave MI Unplanned revascularization Definite/probable stent thrombosis

No thrombocytopenia (n = 10096)

P

21.7% 12.5% 9.3% 3.1% 2.2% 0.7% 7.5% 1.9% 5.6% 5.3% 14.0% 12.4% 8.9% 30.2% 5.6% 15.2% 1.6%

(160) (92) (69) (23) (16) (5) (55) (14) (41) (39) (103) (91) (66) (223) (41) (112) (12)

9.7% 6.3% 4.9% 1.1% 0.8% 0.1% 4.1% 0.6% 3.5% 2.4% 4.3% 3.4% 1.3% 18.7% 1.3% 4.6% 0.8%

(975) (632) (494) (106) (83) (14) (408) (63) (347) (239) (429) (339) (135) (1886) (127) (463) (78)

b.0001 b.0001 b.0001 b.0001 .0002 .0007 b.0001 b.0001 .0025 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 .01

22.8% 14.2% 6.5% 3.5% 1.9% 10.0% 2.5% 7.6% 13.8% 2.4%

(162) (100) (43) (23) (13) (72) (17) (55) (96) (17)

15.1% 9.1% 3.4% 1.9% 1.0% 6.4% 1.2% 5.3% 9.1% 1.2%

(1446) (872) (309) (177) (95) (622) (108) (521) (860) (112)

b.0001 b.0001 b.0001 .0070 .02 b.0001 .002 .006 b.0001 .003

NACE, Net adverse clinical events, indicating composite ischemia or major bleeding; composite ischemia, death from any cause, MI, or unplanned revascularization for ischemia. Results are given as Kaplan Meier rates (n). Stent thrombosis was defined according to Academic Research Consortium criteria.

management at the discretion of the physician. All antithrombotic agents were discontinued at the completion of angiography or PCI according to the protocol. Aspirin 300 to 325 mg/d orally or 250 to 500 mg/d intravenously was administered during the index hospitalization, and 75 to 325 mg/d was prescribed indefinitely after discharge. Clopidogrel 75 mg/d was recommended for 1 year in all patients with coronary artery disease. The study was approved by the institutional review board or ethics committee at each participating center, and all patients provided written informed consent.

Acquired thrombocytopenia Patients with baseline thrombocytopenia (platelet count b150,000 platelets/mm3) and patients triaged to CABG, for having a high prevalence of acquired thrombocytopenia secondary to the use of extracorporeal circulation, were excluded. Acquired thrombocytopenia was defined as an inhospital nadir platelet count of b150,000 platelets/mm3 (referenced lower limit of normal).12 Thrombocytopenia was classified as mild (100,000-150,000 platelets/mm3), moderate (50,000-100,000 platelets/mm3), or severe (b50,000 platelets/ mm3). A second definition of acquired thrombocytopenia, using a reduction platelet count from baseline (either N25% or N50% reduction)13 was also reported in patients after assignment to antithrombin therapy. Platelet count was analyzed before

angiography and every 24 hours up to 24 hours after PCI or after angiography in patients triaged to medical therapy.

Clinical end points Clinical end points were evaluated at 30 days and 1 year as previously described.11 Major bleeding was defined as the cumulative occurrence within 1 year after randomization of intracranial or intraocular bleeding, hemorrhage at the access site requiring intervention, hematoma with a diameter of at least 5 cm, a reduction in hemoglobin level of at least 4 g/dL without an overt bleeding source or at least 3 g/dL with such a source, reoperation for bleeding, or transfusion of a blood product. Bleeding was also classified according to the criteria of the TIMI study group.14 A clinical events committee blinded to treatment assignment adjudicated all 30-day and 1-year primary end point events using original source documents.

Statistical analysis Continuous variables were expressed as mean and SD and compared using the Student t test or Wilcoxon rank sum test if applicable. Discrete variables were presented as numbers and percentages and compared with the χ2 test, unless the observation in any cell was b5, in which case Fisher exact test was used. Univariable and multivariable analyses by logistic regression were performed to identify the predictors of absolute (nadir platelet count of b150,000 platelets/mm3)

American Heart Journal Volume 161, Number 2

Caixeta et al 301

Figure 1

Kaplan-Meier event curves for patients with and without thrombocytopenia. One-year cumulative event curves for death (A), myocardial infarction (B), and composite ischemic events (C); 30-day cumulative event curves for non–CABG-related major bleeding (D).

and relative acquired thrombocytopenia (drop in platelet count N25% or N50%). Time-dependent Cox regression models, where thrombocytopenia is included in the model as a time-dependent covariate, were performed to identify the predictors of non-CABG major bleeding and death. The multivariable model was built by stepwise variable selection with entry and exit criteria set at the P = .2 and P = .1 levels, respectively. All variables in Table I were considered for the

univariate selection. Time-to-event distributions were displayed according to the Kaplan-Meier method and were compared with the use of the log-rank test. All statistical tests were 2-tailed. Probability was considered significant at a level of b.05. No extramural funding was used to support this work. The ACUITY trial was sponsored by the Medicines Company and

American Heart Journal February 2011

302 Caixeta et al

Figure 1

(continued).

Nycomed. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the manuscript, and its final contents.

Results After exclusion of 1,625 patients with thrombocytopenia at baseline (platelet count b150,000 platelets/mm3) and 1,358 patients triaged to CABG, 10,836 patients from

the ACUITY trial were included in the present analysis. During hospitalization, acquired thrombocytopenia developed in 740 patients (6.8%). Mild, moderate, and severe acquired thrombocytopenia developed in 656 (6%), 51 (0.5%), and 33 (0.3%) patients, respectively. Patients who developed thrombocytopenia were more likely to be older; men; and have impaired creatinine clearance, lower platelet count at baseline, prior myocardial infarction (MI), prior CABG, and higher TIMI risk

American Heart Journal Volume 161, Number 2

Caixeta et al 303

Table III. Incidence of acquired thrombocytopenia (a) Heparin plus (b) Bivalirudin plus GP IIb/IIIa GP IIb/IIIa inhibitor inhibitor In-hospital Any thrombocytopenia 7.2% Mild 6.2% Moderate 0.7% Severe 0.4% 30-Day follow-up Any thrombocytopenia 7.3% Mild 6.1% Moderate 0.8% Severe 0.4% Reduction in platelet count from baseline N25% 7.6% (271/3574) N50% 1.4% (50/3574)

(c) Bivalirudin monotherapy

P (all groups) P (a) vs (b) P (a) vs (c) P (b) vs (c)

7.0% 6.2% 0.4% 0.4%

6.3% 5.8% 0.3% 0.2%

.25 .74 .07 .17

.79 .93 .19 .88

.12 .53 .02 .1

.19 .47 .33 .07

7.2% 6.4% 0.5% 0.4%

6.4% 5.9% 0.4% 0.2%

.26 .71 .06 .12

.89 .68 .12 .83

.13 .67 .02 .04

.17 .40 .46 .07

7.4% (270/3626) 1.1% (41/3626)

5.6% (205/3636) 0.7% (26/3636)

.001 .018

.83 .31

.0009 .004

.002 .06

In-hospital outcomes are summarized as % and compared between groups using χ2 or Fisher exact tests when appropriate. “30 Days” is summarized as Kaplan-Meier % and compared between groups with log-rank tests.

score. The use of aspirin before randomization was similar between the 2 groups. Patients undergoing PCI were more likely to develop thrombocytopenia than those triaged to medical therapy (Table I).

Clinical outcomes stratified by acquired thrombocytopenia Compared to patients without acquired thrombocytopenia, those with this complication had higher rates of mortality, MI, unplanned revascularization for ischemia, stent thrombosis, and composite ischemic events at 30 days and 1 year (Table II). In addition, the severity of acquired thrombocytopenia strongly correlated with adverse clinical outcomes at 30-day and 1-year followup (Figure 1). The 30-day mortality rate in patients without acquired thrombocytopenia was 1.1% versus 2.5%, 7.8%, and 9.1% in those who developed mild, moderate, and severe thrombocytopenia, respectively (log-rank P b .001). The differences in mortality among the groups were maintained at 1-year follow-up: 3.4% in patients without acquired thrombocytopenia versus 6.1%, 9.9%, and 9.1% in those who developed mild, moderate, and severe thrombocytopenia, respectively (log-rank P b .001) (Figure 1A). Similarly, compared with patients without acquired thrombocytopenia, the rates of MI and composite events were approximately 2-, 3-, and 4-fold higher in patients who developed mild, moderate, and severe thrombocytopenia, respectively (Figures 1B and 1C). Patients who had acquired thrombocytopenia were more likely to develop protocol-defined and TIMI-criteria major bleeding complications and to receive blood product transfusions (Table II). In addition, the severity of acquired thrombocytopenia strongly correlated with non-CABG major bleeding at 30 days (Figure 1D). Patients with mild or moderate acquired thrombocytopenia had a 3-fold higher incidence of non-CABG major bleeding, and

those who developed severe acquired thrombocytopenia had a 7-fold higher rate of non-CABG major bleeding at 30-day follow-up.

Incidence of acquired thrombocytopenia by randomized antithrombin therapy As shown in Table III, the incidence of acquired thrombocytopenia was comparable between bivalirudin plus GP IIb/IIIa inhibitor therapy and heparin plus GP IIb/IIIa inhibitor. Compared to heparin plus GP IIb/ IIIa inhibitor, any acquired thrombocytopenia after bivalirudin monotherapy was numerically reduced, with significant reductions in moderate acquired thrombocytopenia (0.7 vs 0.3, P = .02). Furthermore, compared to heparin plus GP IIb/IIIa inhibitors, the reduction of platelet count from baseline (either N25% or N50%) occurred significantly less often after bivalirudin monotherapy. By multivariable analysis, development of mild thrombocytopenia was an independent predictor of major bleeding at 30 days. Moderate and severe acquired thrombocytopenia were significantly associated with 1-year mortality (Table IV). Bivalirudin monotherapy was significantly associated with less frequent declines in platelet by N25% or N50% from baseline (Table V).

Discussion In this large-scale prospective, randomized trial of patients with moderate- or high-risk ACS, the principal findings are (1) occurring in approximately 1 in 14 patients, acquired thrombocytopenia develops frequently with an early invasive strategy and antithrombotic treatment; (2) the 30-day incidence of death, MI, and major bleeding doubled in patients with even mild acquired thrombocytopenia compared to patients without thrombocytopenia; (3) rates of mortality (9.1%) and

American Heart Journal February 2011

304 Caixeta et al

Table IV. Multivariable predictors of non-CABG major bleeding and death Hazard ratio (95% CI) Non-CABG major bleeding at 30 days Male 0.49 (0.40-0.60) PCI triage (vs medical treatment) 2.30 (1.81-2.93) Randomized to bivalirudin alone 0.49 (0.39-0.63) Age 1.02 (1.01-1.03) Renal insufficiency 1.60 (1.26-2.04) Prior PCI 0.68 (0.53-0.87) Diabetes 1.35 (1.09-1.66) ST-segment deviation ≥ 1 mm 1.45 (1.12-1.88) Mild thrombocytopenia (vs none) 1.68 (1.04-2.72) Moderate thrombocytopenia 1.09 (0.15-7.75) (vs none) Death at 1 year Age 1.07 (1.06-1.09) Diabetes 2.00 (1.55-2.60) PCI triage (vs medical treatment) 0.63 (0.49-0.81) Male 1.53 (1.16-2.01) Current smoker 1.64 (1.19-2.26) Severe thrombocytopenia (vs none) 3.41 (1.09-10.68) ST-segment deviation ≥1 mm 1.57 (0.99-2.51) Moderate thrombocytopenia 2.89 (0.92-9.06) (vs none) TIMI risk score high 1.28 (0.98-1.67) Elevated biomarkers 1.41 (0.96-2.08) Renal insufficiency 1.29 (0.96-1.74) Mild thrombocytopenia (vs none) 1.30 (0.84-2.02)

χ2

P

47.88 b.0001 45.96 b.0001 34.07 b.0001 16.36 .0001 15.07 .0001 9.37 .006 7.78 .005 7.75 .005 4.42 .03 0.01 .93

76.24 b.0001 27.60 b.0001 12.8 .0003 9.21 .002 9.19 .002 4.42 .03 3.60 .06 3.31 .06 3.26 2.99 2.79 1.37

.07 .08 .09 .24

Table V. Multivariable predictors of absolute and relative acquired thrombocytopenia during hospitalization Odds ratio (95% CI)

χ2

P

Predictors of acquired thrombocytopenia (<150,000 platelets/mm3) Age 1.03 (1.02-1.04) 57.40 b.0001 PCI triage (vs medical treatment) 1.74 (1.43-2.12) 30.40 b.0001 Male 1.53 (1.25-1.87) 17.13 b.0001 Prior CABG 1.41 (1.14-1.74) 10.32 .001 Prior MI 1.19 (0.98-1.43) 3.21 .07 Prior CABG 1.41 (1.14-1.74) 10.32 .001 Predictors of relative acquired thrombocytopenia (drop N25% in platelet count) PCI triage 2.18 (1.77-2.68) 53.92 b.0001 Renal insufficiency 1.64 (1.35-2.00) 24.24 b.0001 Male 0.67 (0.56-0.80) 19.85 b.0001 Prior PCI 0.74 (0.60-0.91] 8.06 .004 Bivalirudin monotherapy (vs heparin 0.75 (0.61-0.93) 6.93 .008 plus GP IIb/IIIa inhibitors) ST-segment deviation ≥1 mm 1.28 (1.02-1.62) 4.37 .04 TIMI risk score high 1.23 (1.01-1.50) 4.29 .04 ST-segment deviation ≥1 mm 1.28 (1.02-1.62) 4.37 .04 Predictors of relative acquired thrombocytopenia (drop N50% in platelet count) PCI triage 2.56 (1.51-4.34) 12.28 .0005 Bivalirudin monotherapy (vs heparin 0.47 (0.28-0.80) 7.80 .005 plus GP IIb/IIIa inhibitors) TIMI risk score high 1.65 (1.07-2.55) 5.11 .02 Prior PCI 0.61 (0.39-0.96) 4.59 .03 Renal insufficiency 1.60 (1.01-2.52) 4.06 .05

The hazard ratio of severe thrombocytopenia (vs none) is 0 because all patients have thrombocytopenia events happening after non-CABG major bleeding.

major bleeding (30%) at 30 days were extremely high in patients with severe acquired thrombocytopenia; (4) by multivariate analysis, acquired thrombocytopenia was an independent predictor of death at 1-year follow-up; and (5) bivalirudin monotherapy was associated with a lower incidence in platelet count drop from baseline compared with heparin plus GP IIb/IIIa inhibitor.

Incidence of thrombocytopenia Similar to REPLACE-210 and HORIZONS-AMI15 trials, bivalirudin treatment in the present study was associated with a lower incidence of acquired thrombocytopenia compared to use of heparin plus GP IIb/IIIa inhibitor. Supplemental Table 1 (online Appendix) shows the incidence of acquired thrombocytopenia according to different patient populations of a variety of trials and registries. When compared to other trials that have investigated thrombocytopenia after heparin plus GP IIb/ IIIa inhibitor therapy in patients with ACS, the incidence of acquired thrombocytopenia after heparin plus GP IIb/IIIa inhibitor therapy in the present study is lower, possibly due to the exclusion of patients with CABG as well as those with baseline thrombocytopenia (Table V).3-6,15-17 The incidence of thrombocytopenia in the present study is lower than that observed in the REPLACE-2 trial.10 This

might be explained by the use of PCI in all patients for revascularization in that trial, whereas only one third of patients in ACUITY underwent PCI. The incidence of thrombocytopenia in HORIZONS-AMI,12 which evaluated patients with ST-elevation myocardial infarction (STEMI), was higher than that observed in the present study. In the present study, performed in patients with ACS undergoing an invasive strategy, we report acquired thrombocytopenia in a broad range of patients, including those with mild acquired thrombocytopenia (N100,000 platelets/mm3 but b150,000 platelets/mm3). Although various studies apply a threshold of 100,000 platelets/ mm3 to define thrombocytopenia,3-6,10,15-17 150,000 platelets/mm3 is referenced as the lower limit of normal.12 Indeed, some studies further include a relative reduction of ≥50% from baseline in the definition to capture patients as a significant reduction of platelet count despite a nadir within the reference range.13,18 Accordingly, after adjusting for multiple confounding, randomization to bivalirudin monotherapy in the current analysis was associated with lower rates of acquired thrombocytopenia by relative (25% or 50%) reduction in platelet counts from baseline but not by absolute (b150,000 platelets/mm3 ) criteria. Using a relative decrease in platelet count may be more appropriate as this definition is independent of baseline nadir platelet count,13 but further validation is required.

American Heart Journal Volume 161, Number 2

Thirty-day and 1-year clinical outcomes The clinical significance of acquired thrombocytopenia after antithrombin and/or GP IIb/IIIa inhibitor therapy in patients with ACS is well recognized.4-6,13,17,18 Although reversible, thrombocytopenia in this setting is associated with bleeding and ischemic complications, blood product transfusion, and prolonged hospitalization.16,19 In particular, almost 1 in every 10 patients in the present analysis, who developed in-hospital severe thrombocytopenia, died within 30 days, similar to findings reported from the large CRUSADE registry13 and in other randomized trials.5,17 Notably, even mild thrombocytopenia was associated with a 2-fold higher mortality risk. In addition, the rates of MI and repeat revascularization at 30 days and 1 year were significantly higher in patients who developed thrombocytopenia than those who did not. Although these unfavorable outcomes were also observed in other trials,4-6,17 it remains unclear why thrombocytopenia is associated with higher rates of MI and repeat revascularization. It might be explained by enhanced platelet activity and possible cessation of antiplatelet or antithrombin therapies after bleeding, which is strongly related to thrombocytopenia.3,16,20 Indeed, the incidence of major bleeding tripled even in patients with mild thrombocytopenia when compared to patients without thrombocytopenia, and one third of patients with severe thrombocytopenia experienced major bleeding within 30 days in the present study. The high rate of major bleeding in patients with thrombocytopenia likely contributes to unfavorable clinical outcomes.7-9 Previous trials have reported the independent relationship between major bleeding and subsequent mortality in patients with ACS and in those undergoing PCI.7-9,21 Major bleeding was a more powerful predictor of mortality than periprocedural MI after PCI in the REPLACE-2 trial.21 Moreover, in the ACUITY trial, the occurrence of bleeding within 30 days was found to be an independent predictor of subsequent mortality occurring between 30 days and 1 year.22 In the HORIZONS-AMI15 trial, bivalirudin monotherapy, compared to heparin plus GP IIb/IIIa inhibitor, reduced the rates of major bleeding and blood transfusion. Bivalirudin monotherapy also reduced the occurrence of severe thrombocytopenia, which has been strongly associated with death among patients with STEMI undergoing PCI.4,19 A strength of the present study is that the severity of thrombocytopenia affected meaningful clinical outcomes, such as mortality, MI, and major bleeding. These results, now reproduced in 3 consecutive large-scale randomized trials involving N20,000 patients (REPLACE-2, ACUITY, and HORIZONS-AMI), demonstrate that bivalirudin monotherapy, compared with an anticoagulation regimen including a GP IIb/IIIa inhibitor, is associated with a significantly lower risk of acquired thrombocytopenia (relative acquired thrombocytopenia in the present

Caixeta et al 305

study) and suggest that this may reduce the occurrence of both early and late mortality.

Limitations There are several limitations to the current study. First, this is a post hoc analysis, and results should be considered hypothesis-generating and require confirmation in additional studies. Second, the number of patients who developed thrombocytopenia was relatively small, although the ACUITY trial is one of the largest trials ever performed in patients with ACS. The present study was underpowered to compare hard clinical events between the groups. Third, the time course of the development and resolution of thrombocytopenia was not available. Fourth, multivariable analysis might not adequately account for all relevant factors. Finally, a significant proportion of patients were pretreated with either unfractionated or low-molecular-weight heparin before randomization. More than 60% of patients assigned to bivalirudin were exposed to heparin before randomization. Nonetheless, the effects of bivalirudin monotherapy as compared with heparin plus GP IIb/IIIa inhibitor were consistent in the original trial,11 regardless of whether heparin was administered before randomization.

Conclusions Occurring in approximately 1 in 14 patients, acquired thrombocytopenia develops frequently in patients with ACS receiving antithrombin and antiplatelet medications and is strongly associated with ischemic and hemorrhagic complications. Compared to an anticoagulation regimen including a GP IIb/IIIa inhibitor, bivalirudin monotherapy was protective against drops in platelet count, which may contribute to the reduced rates of adverse events reported with bivalirudin in patients with ACS undergoing invasive management.

References 1. Anderson JL, Adams CD, Antman EM, et al. ACC/AHA 2007 Guidelines for the Management of Patients With Unstable Angina/ Non–ST-Elevation Myocardial Infarction—Executive Summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (writing committee to revise the 2002 guidelines for the management of patients with unstable angina/non–ST-elevation myocardial infarction) developed in collaboration with the American College of Emergency Physicians, the Society for Cardiovascular Angiography and Interventions, and the Society of Thoracic Surgeons endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation and the Society for Academic Emergency Medicine. J Am Coll Cardiol 2007;50:652-726. 2. Bassand JP, Hamm CW, Ardissino D, et al. Guidelines for the diagnosis and treatment of non–ST-segment elevation acute coronary syndromes. Eur Heart J 2007;28:1598-660. 3. Merlini PA, Rossi M, Menozzi A, et al. Thrombocytopenia caused by abciximab or tirofiban and its association with clinical outcome in patients undergoing coronary stenting. Circulation 2004;109: 2203-6.

306 Caixeta et al

4. Nikolsky E, Sadeghi HM, Effron MB, et al. Impact of in-hospital acquired thrombocytopenia in patients undergoing primary angioplasty for acute myocardial infarction. Am J Cardiol 2005;96:474-81. 5. McClure MW, Berkowitz SD, Sparapani R, et al. Clinical significance of thrombocytopenia during a non–ST-elevation acute coronary syndrome. The platelet glycoprotein IIb/IIIa in unstable angina: receptor suppression using integrilin therapy (PURSUIT) trial experience. Circulation 1999;99:2892-900. 6. Eikelboom JW, Anand SS, Mehta SR, et al. Prognostic significance of thrombocytopenia during hirudin and heparin therapy in acute coronary syndrome without ST elevation: Organization to Assess Strategies for Ischemic Syndromes (OASIS-2) study. Circulation 2001;103:643-50. 7. Eikelboom JW, Mehta SR, Anand SS, et al. Adverse impact of bleeding on prognosis in patients with acute coronary syndromes. Circulation 2006;114:774-82. 8. Rao SV, O'Grady K, Pieper KS, et al. Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes. Am J Cardiol 2005;96:1200-6. 9. Manoukian SV, Feit F, Mehran R, et al. Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY trial. J Am Coll Cardiol 2007;49:1362-8. 10. Lincoff AM, Bittl JA, Harrington RA, et al. Bivalirudin and provisional glycoprotein IIb/IIIa blockade compared with heparin and planned glycoprotein IIb/IIIa blockade during percutaneous coronary intervention: REPLACE-2 randomized trial. JAMA 2003;289:853-63. 11. Stone GW, McLaurin BT, Cox DA, et al. Bivalirudin for patients with acute coronary syndromes. N Engl J Med 2006;355:2203-16. 12. Kratz A, Ferraro M, Sluss PM, et al. Case records of the Massachusetts General Hospital. Weekly clinicopathological exercises. Laboratory reference values. N Engl J Med 2004;351:1548-63. 13. Wang TY, Ou FS, Roe MT, et al. Incidence and prognostic significance of thrombocytopenia developed during acute coronary syndrome in contemporary clinical practice. Circulation 2009;119: 2454-62.

American Heart Journal February 2011

14. Chesebro JH, Knatterud G, Roberts R, et al. Thrombolysis in Myocardial Infarction (TIMI) trial, phase I: a comparison between intravenous tissue plasminogen activator and intravenous streptokinase. Clinical findings through hospital discharge. Circulation 1987; 76:142-54. 15. Stone GW, Witzenbichler B, Guagliumi G, et al. Bivalirudin during primary PCI in acute myocardial infarction. N Engl J Med 2008;358: 2218-30. 16. Berkowitz SD, Sane DC, Sigmon KN, et al. Occurrence and clinical significance of thrombocytopenia in a population undergoing highrisk percutaneous coronary revascularization. Evaluation of c7E3 for the Prevention of Ischemic Complications (EPIC) study group. J Am Coll Cardiol 1998;32:311-9. 17. Yeh RW, Wiviott SD, Giugliano RP, et al. Effect of thrombocytopenia on outcomes following treatment with either enoxaparin or unfractionated heparin in patients presenting with acute coronary syndromes. Am J Cardiol 2007;100:1734-8. 18. Oliveira GB, Crespo EM, Becker RC, et al. Incidence and prognostic significance of thrombocytopenia in patients treated with prolonged heparin therapy. Arch Intern Med 2008;168:94-102. 19. Kereiakes DJ, Berkowitz SD, Lincoff AM, et al. Clinical correlates and course of thrombocytopenia during percutaneous coronary intervention in the era of abciximab platelet glycoprotein IIb/IIIa blockade. Am Heart J 2000;140:74-80. 20. Spencer FA, Moscucci M, Granger CB, et al. Does comorbidity account for the excess mortality in patients with major bleeding in acute myocardial infarction? Circulation 2007;116:2793-801. 21. Feit F, Voeltz MD, Attubato MJ, et al. Predictors and impact of major hemorrhage on mortality following percutaneous coronary intervention from the REPLACE-2 trial. Am J Cardiol 2007;100: 1364-9. 22. Mehran R, Pocock SJ, Stone GW, et al. Associations of major bleeding and myocardial infarction with the incidence and timing of mortality in patients presenting with non–\ST-elevation acute coronary syndromes: a risk model from the ACUITY trial. Eur Heart J 2009;30:1457-66.

American Heart Journal Volume 161, Number 2

Caixeta et al 306.e1

Appendix Supplemental Table 1. Incidence of acquired thrombocytopenia and 30-day mortality rate according to study population and antithrombotic regimen Trial ACUITY

REPLACE-210 HORIZONS15 TIMI 11B17 OASIS-26 EPIC16 PURSUIT5 CADILLAC4 TARGET3 CRUSADE13

Participants non-STEMI ACS

Treatment

Heparin plus GP IIb/IIIa inhibitor Bivalirudin plus GP IIb/IIIa inhibitor Bivalirudin monotherapy Patients undergoing PCI Heparin plus GP IIb/IIIa inhibitor Bivalirudin monotherapy Patients with STEMI undergoing PCI Heparin plus GP IIb/IIIa inhibitor Bivalirudin monotherapy non-STEMI ACS Unfractionated heparin Enoxaparin non-STEMI ACS Unfractionated heparin Hirudin non-STEMI ACS Heparin Heparin plus GP IIb/IIIa inhibitor non-STEMI ACS Heparin Heparin plus GP IIb/IIIa inhibitor Patients with STEMI undergoing PCI Heparin plus GP IIb/IIIa inhibitor Heparin Patients undergoing PCI Heparin plus GP IIb/IIIa inhibitor non-STEMI ACS Heparin plus GP IIb/IIIa inhibitor†

⁎ Platelet count b100,000 platelets/mm3. † Only 50% of the patients were treated with a GP IIb/IIIa inhibitor. ‡ Death from all causes. § Mortality refers to 14 days. ‖ Cardiovascular death at 7 days. ¶ In-hospital mortality.

Incidence of thrombocytopenia⁎ 30-day mortality‡ 0.7% 0.4% 0.3% 1.7% 0.7% 2.9% 1.1% 1.7% 1.4% 1.1% 0.9% 3.3% 5.2% 4.9% 4.9% 3.2% 1.7% 2.9% 1.2%

1.3% 1.5% 1.6% 0.4% 0.2% 3.1% 2.1% 2.8%§ 2.2%§ 1.5%‖ 1.4%‖ 1.7% 1.7% 3.7% 3.5% 1.9% 2.3% 0.4% 5.2%¶

Valvular and Congenital Heart Disease

Pregnancy in women with corrected tetralogy of Fallot: Occurrence and predictors of adverse events Ali Balci, MD, MSc, a,b,i Willem Drenthen, MD, PhD, a,i Barbara J. M. Mulder, MD, PhD, c,i Jolien W. Roos-Hesselink, MD, PhD, d,i Adriaan A. Voors, MD, PhD, a,i,j Hubert W. Vliegen, MD, PhD, e,i Philip Moons, RN, PhD, f,i Krystyna M. Sollie, MD, g,i Arie P. J. van Dijk, MD, PhD, h,i Dirk J. van Veldhuisen, MD, PhD, a,i,j and Petronella G. Pieper, MD, PhD a ,i Groningen, Utrecht, Amsterdam, Rotterdam, Leiden, and Nijmegen, The Netherlands; and Leuven, Belgium

Background In women with corrected tetralogy of Fallot (ToF), pregnancy is associated with maternal cardiac, obstetric, and offspring complications. Our aim is to investigate the magnitude and determinants of pregnancy outcome in women with corrected ToF. Methods In this retrospective international multicenter study using 2 congenital heart disease registries, 204 women with corrected ToF were identified. Within this group, 74 women had 157 pregnancies, including 30 miscarriages and 4 terminations of pregnancy. Detailed information on each completed pregnancy (n = 123) was obtained using medical records and supplementary interviews. Results Cardiovascular events occurred during 10 (8.1%) pregnancies, mainly (supra)ventricular arrhythmias. Obstetric and offspring events occurred in 73 (58.9%) and 42 (33.9%) pregnancies, respectively, including offspring mortality in 8 (6.4%). The most important predictor was use of cardiac medication before pregnancy (odds ratio for cardiac events 11.7, 95% CI 2.2-62.7; odds ratio for offspring events 8.4, 95% CI 1.4-48.6). In pregnancies with cardiovascular events, significantly more small-for-gestational-age children were born (P value b .01). Conclusions Cardiovascular, obstetric, and offspring events occur frequently during pregnancies in women with ToF. Maternal use of cardiovascular medication is associated with pregnancy outcome, and maternal cardiovascular events during pregnancy are highly associated with offspring events. (Am Heart J 2011;161:307-13.)

Long-term survival after correction of tetralogy of Fallot (ToF) is excellent.1 Pregnancy is generally well tolerated in women with corrected ToF.2 Nevertheless, cardiac, obstetric, and offspring events occurred, respectively, during 4.5% to 18%, 11% to 20% and 16% to 27% of

From the aDepartment of Cardiology, University Medical Center Groningen, Groningen, The Netherlands, bInteruniversity Cardiology Institute of the Netherlands, Utrecht, The Netherlands, cDepartment of Cardiology, Academic Medical Center, Amsterdam, The Netherlands, dDepartment of Cardiology, Erasmus Medical Center, Rotterdam, The Netherlands, eDepartment of Cardiology, Leiden University Medical Center, Leiden, The Netherlands, fDepartment of Cardiology, University Hospitals of Leuven, Leuven, Belgium, gDepartment of Obstetrics and Gynecology, University Medical Center Groningen, Groningen, The Netherlands, and hDepartment of Cardiology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands. i On behalf of the ZAHARA investigators. j Established investigators of the Netherlands Heart Foundation (D97.017 and 2006T037). Submitted August 3, 2010; accepted October 18, 2010. Reprint requests: Petronella G. Pieper, MD, PhD, Department of Cardiology, University Medical Center Groningen, University of Groningen, PO Box 30.001, 9700 RB Groningen, The Netherlands. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.027

completed gestations.3-7 Predictors of adverse outcome in a cohort with ToF are not known; but in a mixed congenital heart disease (CHD) population, severe pulmonary regurgitation (PR) and right ventricular dysfunction were independent predictors of cardiac events during pregnancy.5 Whether or not pulmonary valve replacement (PVR) before pregnancy influences pregnancy outcome is unknown. The primary objective of the present study is to determine the magnitude and the nature of events encountered during pregnancy in women with corrected ToF. Secondary objectives are to identify predictors of adverse pregnancy outcome as well as to investigate whether or not cardiac events could influence offspring outcome during the same pregnancy.

Patients and methods For the present study, all female patients with ToF aged 18 to 58 years enrolled in the nationwide CONgenital CORvitia (CONCOR) registry and a Belgian tertiary medical center's adult CHD database were identified and asked to participate in the study.8,9 The institutional review board or ethics committee

American Heart Journal February 2011

308 Balci et al

at each of the participating tertiary centers approved the protocol. Data were obtained from medical records, and supplementary data were retrieved by a questionnaire. Baseline data included basic cardiac anatomy, history, and electrophysiology; prior cardiac events using the European Pediatric Cardiac Coding; noncardiac comorbidity; maternal age at inclusion; prepregnancy New York Heart Association (NYHA) functional class and cyanosis (oxygen saturation b90%); the use of medication; use of cigarettes, drugs, and/or alcohol; and obstetric history including miscarriages (spontaneous fetal loss b20 weeks of gestation) and/or elective abortions.10-12 Detailed information concerning each completed pregnancy (N20 weeks of gestation) between 1980 and 2007 was recorded: maternal age at conception; mode of delivery; parity; use of cigarettes, drugs, and/or alcohol during pregnancy; use of medication; change in NYHA functional class; 12-lead electrocardiogram; transthoracic echocardiograms; and/or 24-hour electrocardiogram (Holter) registrations. Documented pregnancy-related events were divided into cardiovascular, obstetric, and offspring events (composite end points).

Cardiovascular events Cardiovascular events were defined as follows: documented symptomatic arrhythmia or heart failure requiring treatment (according to attending cardiologist), myocardial infarction, endocarditis, aortic dissection, thromboembolic events, and/or stroke.

Obstetric events Obstetric events were defined as follows: pregnancy-induced hypertension (PIH, ≥20 weeks of gestation, ≥140 mm Hg systolic and/or ≥90 mm Hg diastolic blood pressure, without proteinuria); preeclampsia (PIH with ≥0.3 g of proteinuria in a 24-hour urine sample); eclampsia (preeclampsia with grand mal seizures); hemolysis elevated liver enzymes low platelets syndrome; gestational diabetes13; hyperemesis gravidarum (severe, intractable nausea and vomiting leading to dehydration, loss of weight, metabolic disorders, and/or hospitalization); preterm prelabor rupture of membranes (PPROM); assisted delivery (use of forceps or vacuum extraction); cesarean delivery on medical indication; prolongation of cervix ripening (omitted dilatation of the portio vaginalis for ≥20 hours (nullipara) or ≥14 hours (multipara), despite adequate and regular uterus contractions); prolongation of second stage of delivery (primipara N2 hours or multipara N1 hour); premature labor (PL; spontaneous onset of labor at b37 weeks of gestation); postpartum hemorrhage (PPH; blood loss at vaginal delivery ≥500 mL or cesarean delivery ≥1,000 mL).

Offspring events Offspring events were defined as follows: premature birth (birth b37 weeks of gestation); small-for-gestational-age (SGA; birth weight b10th percentile); fetal mortality (intrauterine death ≥20 weeks of gestation); offspring death (within the first year after birth); and/or iteration of CHD.

Statistical analysis A Clintrial data entry program was used to record information and was converted to SPSS (version 16.0; SPSS

Inc, Chicago, IL) for statistical analysis. Descriptive statistics for nominal data are expressed in absolute numbers and percentages. Mean values and SD are presented for normally distributed continuous variables. For nonnormally distributed continuous variables, median and interquartile ranges were computed. Comparison of continuous variables between groups was made by unpaired Student's t tests or MannWhitney U test depending on distribution. For the comparison of dichotomous variables, we used the χ2 test or Fisher exact test, where applicable. All P values presented are 2-sided. Univariable logistic regression analysis was performed to identify predictors of adverse pregnancy outcome, divided into 3 composite end points (as defined above): cardiac, obstetric, and offspring events. The following prepregnancy baseline variables were assessed: palliative surgery before ToF correction, history of arrhythmias, PVR, use of cardiac prescription medication, NYHA functional class, the presence of a patent shunt, pulmonary atrioventricular valve regurgitation (moderate/severe), pulmonary valve regurgitation (PR; moderate/severe) and/or right ventricular outflow tract obstruction (peak gradient N50 mm Hg), and reduced systemic ventricular function. For the classification of valve regurgitation and valve stenosis, we used the classification as recommended by the European Society of Cardiology and the American Society of Echocardiography in their guidelines on valvular heart disease.14,15 Variables that were associated with an increased incidence of the studied end points (P b .10) entered the multivariable stage, which at least contained the variables age and parity. The final multivariable model was then constructed by backward deletion of the least significant characteristic until all remaining variables were significantly (P b .05) associated with the end point. Because some women went through ≥1 pregnancy, the validity of treating each pregnancy as an independent event was confirmed by general estimating equation analysis.

Funding This work was supported by Netherlands Heart Foundation grant 2002B125. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.

Results Written informed consent was provided by 204 women (82% of all women who were asked to participate) with corrected ToF. We observed 157 pregnancies in 74 of these patients. One third of the ToF women were enrolled in a pilot study published previously.3 Of the 157 pregnancies, 30 ended in a miscarriage (19%) and 4 in an elective abortion (2.5%), leaving 123 completed pregnancies (≥20 weeks of gestation, including 1 twin pregnancy) in 69 different patients. From the 16 women who had 30 miscarriages, 13 women later had completed pregnancies. Table I shows the baseline characteristics of women with and without completed pregnancies. Women without completed pregnancies were younger and more often had a mental disability. Most ToF women had a desire for children (168 of 204; 82%). In the

American Heart Journal Volume 161, Number 2

Balci et al 309

Table I. Patient characteristics for the total cohort and for patients with and without completed pregnancies

Mean age at inclusion, y (±SD) Mean age at repair, y (±SD) Surgeries before or after repair (%) Blalock-Taussig shunt Waterston shunt Potts shunt RVOT procedure PVR Valvular dysfunction (%) Tricuspid regurgitation‡ RVOT obstruction Pulmonary valve regurgitation‡ Ventricular septal defect (%) Mental disability (%)

All patients (N = 204)

Patients without completed pregnancies⁎ (n = 135)

Patients with completed pregnancies⁎ (n = 69)

P value†

31 ± 7.5 5.1 ± 2.7

28 ± 6.7 4.7 ± 2.5

36 ± 5.8 6.1 ± 2.8

b.001 b.001

22 29 2 42 44

(10.8) (14.2) (1.0) (20.6) (21.6)

15 (11.1) 14 (10.4) 0 (0.0) 24 (17.8) 26 (19.3)

7 (10.1) 15 (21.7) 2 (2.9) 18 (26.1) 18 (26.1)

NS b.05 NS NS NS

27 (18.5) 46 (31.5) 77 (52.7) 13 (8.9) 12 (5.9)

15 (18.1) 21 (25.3) 39 (47.0) 6 (4.4) 12 (8.9)

12 (19.0) 25 (39.7) 38 (60.3) 7 (11.1) 0 (0.0)

NS NS NS NS b.01

RVOT, Right ventricular outflow tract; NS, not significant (P N .05). ⁎ Completed pregnancy, lasting N20 weeks. † Comparison of women without completed pregnancies versus women with completed pregnancies. ‡ Moderate or severe.15

Table II. Baseline characteristics of the 123 completed pregnancies⁎ in women with corrected ToF n = 123 Mean age at inclusion, y (±SD) Mean age at repair, y (±SD) Surgery before/after repair, n (%) Blalock-Taussig shunt Waterston shunt Potts shunt RVOT procedure PVR Valvular dysfunction, n (%) Tricuspid regurgitation (moderate/severe)15 RVOT obstruction Pulmonary valve regurgitation (moderate/severe)15 Ventricular septal defect, n (%) History of arrhythmias, n (%) History of heart failure, n (%) NYHA class ≥II prepregnancy, n (%) Cardiac medication used prepregnancy, n (%) β-Blockers† Calcium-channel blockers Digoxin Amiodarone Vitamin K antagonists, n (%) Mean age at pregnancy, y (±SD) Primipara, n (%) Multipara, n (%) Mean pregnancy duration, wk (±SD)

36.4 ± 5.6 6.5 ± 4.2 12 (9.8) 22 (17.9) 2 (1.6) 47 (38.2) 33 (26.8) 25 (20.3) 54 (43.9) 69 (56.1) 20 (16.3) 5 (4.1) 3 (2.4) 8 (6.5) 7 (5.7) 1 (0.8) 1 (0.8) 3 (2.4) 2 (1.6) 2 (1.6) 26.8 ± 4.1 32 (26.0) 91 (74.0) 37.8 ± 4.5

⁎ Completed pregnancy, lasting N20 weeks. † Stopped using because of pregnancy.

childless cohort, the most common reason for being childless at inclusion was that women felt too young (not ready) to have children (n = 51; 38%). Thirteen patients reported that anticipated cardiovascular events were the reason they did not pursue pregnancy (9.6%), and 6 of

them were advised against pregnancy by their cardiologist. Three women (2.2%) did not have children because of concerns regarding heredity of ToF. Because the available data concerning the incomplete pregnancies were very limited, we further focused on the 123 completed pregnancies. In Table II the baseline characteristics of the 123 completed pregnancies are presented. Of the pregnancies in our cohort, 19.5% was observed before 1990. Between 1990 and 2000, 49.6% and, after 2000, 30.9% of the pregnancies took place. There was no relation between pregnancy period and cardiovascular, obstetric, or offspring outcome. Thirtythree women had undergone PVR at a median age of 23.5 years (range 7-33 years). None of the patients in our cohort had cyanosis. In 30 (24%) of 123 pregnancies, the mother smoked before pregnancy; and in 20 (16%) of 123 pregnancies, the mother continued smoking during pregnancy. Details regarding events that were encountered during completed pregnancies are depicted in Table III. No difference in events was seen between the 2 registries.

Cardiovascular outcome The most frequent cardiovascular events were clinically significant arrhythmias needing treatment, which occurred during 8 pregnancies (6.5%); in 4 pregnancies, the mother had atrial fibrillation or flutter; and in the other 4 pregnancies, the mother had symptomatic ventricular tachycardias. Two of these patients had a history of arrhythmias. Heart failure developed during 2 pregnancies (1.6%) both during the third trimester and in the context of severe PR; one of these women had undergone PVR. Pulmonary embolism was diagnosed and subsequently treated with low–molecular weight heparin in

American Heart Journal February 2011

310 Balci et al

Table III. Overview of events that occurred during 123 completed⁎ pregnancies in women with corrected ToF n = 123 ZAHARA n (%) Cardiovascular events Arrhythmias Heart failure Thromboembolic events Obstetric events Cesarean delivery Assisted† vaginal delivery PPH Prolongation of 2nd stage of delivery PPROM PIH Prolongation of cervix ripening Preeclampsia Hyperemesis Solutio placentae Uterus rupture Offspring events SGA PB OM CHD Other offspring events‡

Healthy population§ (%)

10 (8.1) 8 (6.5) 2 (1.6) 1 (0.8) 73 (58.9) 25 (20.3) 16 (13.0) 12 (9.7) 10 (8.1)

6.5 17 2.9 b2.7

8 (6.5) 6 (4.8) 5 (4.1)

1.5 10 –

4 2 1 1 42 23 22 8 3 3

(3.2) (1.6) (0.8) (0.8) (33.9) (18.5) (17.7) (6.4) (2.4) (2.4)

b1 b1 b.3

1.4 0.6 b1 b0.001 10 10 0.9 0.6 –

PB, Preterm birth; OM, offspring mortality. ⁎ Completed pregnancy, lasting N20 weeks. † Delivery assisted using forceps or vacuum extraction. ‡ Other offspring events were fetal asphyxia, trisomy 13, and hydrocephalus. § Normal population, based on literature.4,24,25,27-29,32,33

one pregnancy. Deterioration of NYHA functional class (≥1 class, between prepregnancy and the third trimester) was documented in 22 completed pregnancies (17.7%) and persisted postpregnancy in 3 pregnancies (2%). Myocardial infarction, stroke, or aortic dissection did not occur in this cohort.

Obstetric outcome Cesarean delivery was the most frequently observed obstetric event and was necessary in 20.3% of completed pregnancies (n = 25). A “primary” (planned) cesarean delivery was carried out during 16 gestations, 5 of which were executed based on maternal cardiac indication (moderate/severe pulmonary valve stenosis or regurgitation combined with tricuspid regurgitation and high risk of arrhythmia). Nine “secondary” (unscheduled) cesarean deliveries were performed, all on obstetric indication. Postpartum hemorrhage, hypertensive disorders of pregnancy, and PPROM complicated, respectively, 9.7%, 8%, and 6.5% of completed pregnancies. Forceps were used in 3 pregnancies; and in 13 pregnancies, a vacuum extractor was used to assist vaginal delivery. No significant time effects were found for obstetric outcome.

Table IV. Results of univariable and multivariable logistic regression models and corresponding risk scores for cardiac, obstetric, and offspring events Univariable analysis Palliative surgery History of arrhythmias Prior PVR RVOT obstruction Pulmonary valve regurgitation§ Pulmonary AV valve regurgitation Patent shunt Smoking during pregnancy Use of cardiac medication prepregnancy NYHA class NII

Cardiovascular events⁎

Obstetric events⁎

Offspring events⁎

1.3 (0.3-5.1)

0.8 (0.4-1.7)

2.7 (1.1-6.2)‡

9.3 (1.8-46.9)‡

5.2 (0.6-43.6)

2.0 (0.5-8.5)

3.1 (0.8-11.4)† 0.9 (0.2-3.2)

2.2 (0.9-5.3)† 1.2 (0.6-2.4)

1.4 (0.6-3.2) 1.3 (0.6-2.7)

0.5 (0.1-1.9)

0.5 (0.23-0.99)‡ 1.3 (0.6-2.7)

0.6 (0.1-2.3)

1.4 (0.6-3.4)

2.1 (0.9-5.2)†

0.8 (0.1-3.8) 0.8 (0.1-3.8)

2.5 (0.9-6.7) 0.9 (0.4-2.5)

1.1 (0.4-2.9) 1.4 (0.5-3.7)

11.8 (2.2-63.3)‡

–

5.5 (1.0-29.2)‡

0.6 (0.1-5.3)

–

0.3 (0.0-2.7)

Multivariable analysis for endpoints cardiovascular, obstetric and offspring events Cardiovascular events Use of cardiac medication pre pregnancy Obstetric events Pulmonary valve regurgitation Offspring events Palliative surgery Use of cardiac medication pre pregnancy

Odds ratio (95% CI)

11.7 (2.2 – 62.7)‡ 0.5 (0.2 – 0.99)‡ 3.3 (1.3 – 8.2)‡ 8.1 (1.4 – 48.6)‡

AV, Atrioventricular.

⁎ Expressed as OR (95% CI). † P b .1. ‡ P b .05. § Moderate or severe.14,15

Offspring outcome The median birth weight was 3,100 g (25th percentile 2,755 g; 75th percentile 3,400 g). Twenty-three children (19%) were SGA. Delivery was preterm in 22 (18%) of the completed pregnancies. Iteration of CHD was diagnosed in 3 (2.4%) children. In one child, the cardiac defect (atrial septal defect) was associated with a trisomy 13; the other children had tetralogy of Fallot and atrioventricular septal defect respectively. The overall offspring mortality was 6.4%. Intrauterine demise occurred during 5 gestations, and 3 children died within the first year postpartum. The main reasons for offspring death were immaturity and/or prematurity at birth, intrauterine growth restriction, and recurrence of CHD. Also for offspring outcome, no time effect could be detected. The occurrence of cardiovascular events was associated with a significantly higher number of SGA children (in 5 of 10 pregnancies with cardiovascular events, the offspring

American Heart Journal Volume 161, Number 2

was SGA versus 18 of 114 pregnancies without cardiovascular events [P b .01]).

Predictors The univariable and multivariable predictors for the composite end points are shown in Table IV. Prior PVR was strongly associated with cardiovascular events. Mainly arrhythmias (n = 4) complicated pregnancies in the PVR group. In one pregnancy, pulmonary embolism was observed; and one pregnancy in the PVR group was complicated with arrhythmia as well as with heart failure. Arrhythmias before pregnancy were highly associated with cardiovascular events. The most important independent predictor for cardiovascular as well as offspring events was the use of cardiac medication before pregnancy. In pregnancies of women who used any cardiac medication before pregnancy, the risk for SGA babies was almost 7-fold higher compared with pregnancies of women without history of cardiac medication use (odds ratio [OR] 6.8, 95% CI 1.4-32.9, P = .02). None of the specific cardiac medication used before pregnancy stood out as a culprit. Use of any cardiac medication during pregnancy had a high association with SGA, but this was not statistically significant (OR 4.9, 95% CI 0.9-25.8, P = .06). No association was found between SGA and other known causes, such as PIH or PE, in our cohort. The use of cardiac medication before pregnancy was associated with PVR: 5 of 7 women who used cardiac medication before pregnancy had previous PVR. Another independent predictor of offspring events was palliative surgical intervention before correction. Women with moderate to severe pulmonary valve regurgitation seem to encounter less obstetric events (35 of 69 vs 38 of 54, P = .04).

Discussion Women with corrected ToF are at increased risk of cardiovascular and obstetric events during pregnancy. Most events are well treatable. The incidence of offspring events is markedly increased. Cardiovascular and offspring outcomes are strongly related with the use of cardiac medication before pregnancy. Surgical status before pregnancy also appears to predict pregnancy outcome. Offspring events are related to maternal cardiovascular events. In our cohort, 19% of the pregnancies ended in a spontaneous abortion, which is slightly higher than in healthy women (10%-16%) and slightly higher than reported previously in pregnancies of women with ToF.4,16

Cardiovascular outcome The occurrence of cardiovascular events in the present study is comparable to that reported in a recent literature review, which described the outcome of 200

Balci et al 311

pregnancies in patients with corrected ToF.4 Arrhythmias were the major cardiac event. Arrhythmias before pregnancy predicted cardiovascular events, as described previously. 6,17-19 Supraventricular and ventricular arrhythmias are well-known long-term sequelae of intracardiac repair of ToF.20,21 The use of cardiac medication before pregnancy strongly predicted cardiovascular events. This predictor probably reflects a less favorable cardiac condition in terms of ventricular function and arrhythmias. In contrast to others, we did not observe a relationship between severity of PR and more adverse events.5,7 In previous publications, it was likely that PR was associated with worse pregnancy outcome especially when it was associated with right ventricular dysfunction.7,22 Heart failure occurred only twice in our population, both times in women with severe PR. The most prevalent complications were arrhythmias, which are related not only to hemodynamics but also to surgical scars and therefore may be encountered also in women with an adequate hemodynamic situation but with a history of surgery, as in women with corrected ToF. Pulmonary valve replacement was associated with worse cardiovascular outcome, which is counterintuitive. The use of cardiac medication before pregnancy was strongly associated with PVR (71% of all women who used cardiac medication had PVR), indicating a less favorable cardiac condition in the PVR group. This might be related to late timing of PVR, as all PVRs were performed between 1990 and 2000, when early timing of PVR was not routinely applied. Therefore, in most of these patients, PVR was likely performed when right ventricular function was already compromised. A recent study showed progression of RV dilatation that persisted postpregnancy in women with corrected ToF, which could not be demonstrated in comparable women who did not have pregnancies.23 In our study, the main events in women after PVR were arrhythmias. These patients may have been vulnerable to arrhythmias because of right ventricular dysfunction, but surgical scars may have also played a role in the occurrence of arrhythmias. In addition, several women had arrhythmias prepregnancy, which is a known predictor for arrhythmias during pregnancy.6 Unfortunately, our study does not answer the question if timely prepregnancy PVR would have prevented cardiovascular pregnancy events in women with ToF.

Obstetric outcome The most important obstetric events were PPH and PL based on rupture of membranes, which occurred more frequently than described previously in population-based studies (2.9% and 1.5%, respectively).24,25 No direct explanation was found; more specifically, no relationship with anticoagulation therapy was detected.26 Cesarean deliveries were performed more often in our study (20%) than in the general Dutch population (6.5%), although

American Heart Journal February 2011

312 Balci et al

only 1 in 5 was performed on maternal cardiac indications.27 Possibly, the threshold for performing cesarean delivery for obstetric/offspring reasons is lower than usual because of caution on the part of obstetricians in these more vulnerable mothers and babies. The other events were in concordance with expected frequencies observed in the general population and mentioned in the literature review.4

Offspring outcome The high offspring mortality (6.4%) by far exceeds expected mortality rates in the general population (0.9%) and also the mortality rate in ToF pregnancies mentioned in the literature (2.0%).28 Recurrence of CHD, but also premature births with their known associated events, explains, at least in part, this high mortality. The premature birth rate (18%) was substantially higher than expected.4,29 This was in part attributable to the higher frequency of PL due to rupture of membranes. The incidence of SGA (19%) is also higher than in the general population (by definition 10%), although it is lower than the 35% recently mentioned by Gelson et al.30 The use of maternal cardiac medication before pregnancy was the most important predictor of offspring outcome. Maternal hemodynamic abnormalities as well as direct effects of maternal cardiovascular medication may undermine placental blood flow and induce placental insufficiency with subsequent intrauterine growth restriction resulting in children born SGA as well as in premature birth. The strong association between maternal cardiovascular events and SGA points in this direction. Palliative surgery before correction appears to influence offspring outcome negatively. Longstanding right ventricular pressure loading and hypoxia in women with a later age at correction may have resulted in more hemodynamic compromise and endothelial dysfunction, compromising placental perfusion and fetal well-being.

Limitations The retrospective study design required very strict definitions for events: all mentioned events had to be documented by medically qualified personnel in the records according to preset definitions before data entry. Heart failure and arrhythmia were only recorded if therapy had been administered. This may have caused underestimation of cardiovascular event rate. In addition, selection bias is introduced by the fact that some patients remained childless because of anticipated/expected risks during pregnancy, which may lead to an underestimation of risks. This study was performed in a survival cohort; thus, no conclusions regarding maternal mortality risk can be made. However, from prospective research, it is known that repaired CHD has a very low risk of maternal death.6,31 Finally, together with the given sample size and the possible effects of multitesting, for example, inflation

of type I error, all conclusions of the present study must be drawn with caution.

Disclosures Conflicts of interest: None.

References 1. Nollert G, Fischlein T, Bouterwek S, et al. Long-term survival in patients with repair of tetralogy of Fallot: 36-year follow-up of 490 survivors of the first year after surgical repair. J Am Coll Cardiol 1997;30:1374-83. 2. Loup O, Von WC, Gahl B, et al. Quality of life of grown-up congenital heart disease patients after congenital cardiac surgery. Eur J Cardiothorac Surg 2009;36:105-11. 3. Meijer JM, Pieper PG, Drenthen W, et al. Pregnancy, fertility, and recurrence risk in corrected tetralogy of Fallot. Heart 2005;91:801-5. 4. Drenthen W, Pieper PG, Roos-Hesselink JW, et al. Outcome of pregnancy in women with congenital heart disease: a literature review. J Am Coll Cardiol 2007;49:2303-11. 5. Khairy P, Ouyang DW, Fernandes SM, et al. Pregnancy outcomes in women with congenital heart disease. Circulation 2006;113:517-24. 6. Siu SC, Sermer M, Colman JM, et al. Prospective multicenter study of pregnancy outcomes in women with heart disease. Circulation 2001; 104:515-21. 7. Veldtman GR, Connolly HM, Grogan M, et al. Outcomes of pregnancy in women with tetralogy of Fallot. J Am Coll Cardiol 2004;44:174-80. 8. Drenthen W, Pieper PG, Ploeg M, et al. Risk of complications during pregnancy after Senning or Mustard (atrial) repair of complete transposition of the great arteries. Eur Heart J 2005;26:2588-95. 9. Drenthen W, Pieper PG, van der Tuuk K, et al. Cardiac complications relating to pregnancy and recurrence of disease in the offspring of women with atrioventricular septal defects. Eur Heart J 2005;26:2581-7. 10. New York Heart Association. The criteria committee of the New York Heart Association, functional capacity and objective assessment. In: Dolgin M, editor. Nomenclature and criteria for diagnosis of diseases of the heart and great vessels. 9th ed. Boston, MA: Little Brown and Company; 1994. p. 253-5. 11. Coding Committee of the Association for European Paediatric Cardiology. The European paediatric cardiac code: the first revision. Cardiol Young 2002;12(Suppl 2):1-211. 12. Franklin RC, Anderson RH, Daniels O, et al. Report of the coding committee of the Association for European Paediatric Cardiology. Cardiol Young 2002;12:611-8. 13. American Diabetes Association. Gestational diabetes mellitus. Diabetes Care 2004;27(Suppl 1):S88-90. 14. Baumgartner H, Hung J, Bermejo J, et al. Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice. J Am Soc Echocardiogr 2009;22:1-23. 15. Vahanian A, Baumgartner H, Bax J, et al. Guidelines on the management of valvular heart disease: the task force on the management of valvular heart disease of the European Society of Cardiology. Eur Heart J 2007;28:230-68. 16. Nielsen S, Hahlin M. Expectant management of first-trimester spontaneous abortion. Lancet 1995;345:84-6. 17. Lee SH, Chen SA, Wu TJ, et al. Effects of pregnancy on first onset and symptoms of paroxysmal supraventricular tachycardia. Am J Cardiol 1995;76:675-8. 18. Silversides CK, Harris L, Haberer K, et al. Recurrence rates of arrhythmias during pregnancy in women with previous

American Heart Journal Volume 161, Number 2

19. 20.

21.

22.

23.

24.

25.

tachyarrhythmia and impact on fetal and neonatal outcomes. Am J Cardiol 2006;97:1206-12. Tawam M, Levine J, Mendelson M, et al. Effect of pregnancy on paroxysmal supraventricular tachycardia. Am J Cardiol 1993;72:838-40. Roos-Hesselink J, Perlroth MG, McGhie J, et al. Atrial arrhythmias in adults after repair of tetralogy of Fallot. Correlations with clinical, exercise, and echocardiographic findings. Circulation 1995;91: 2214-9. Gatzoulis MA, Till JA, Redington AN. Depolarization-repolarization inhomogeneity after repair of tetralogy of Fallot. The substrate for malignant ventricular tachycardia? Circulation 1997;95:401-4. Greutmann M, Von KK, Brooks R, et al. Pregnancy outcome in women with congenital heart disease and residual haemodynamic lesions of the right ventricular outflow tract. Eur Heart J 2010;31:1764-70. Uebing A, Arvanitis P, Li W, et al. Effect of pregnancy on clinical status and ventricular function in women with heart disease. Int J Cardiol 2010;139:50-9. Callaghan WM, Kuklina EV, Berg CJ. Trends in postpartum hemorrhage: United States, 1994-2006. Am J Obstet Gynecol 2010; 202:353-6. van der Ham DP, Nijhuis JG, Mol BW, et al. Induction of labour versus expectant management in women with preterm prelabour rupture of membranes between 34 and 37 weeks (the PPROMEXIL-trial). BMC Pregnancy Childbirth 2007;7:11.

Balci et al 313

26. Pieper PG, Balci A, van Dijk AP. Pregnancy in women with prosthetic heart valves. Neth Heart J 2008;16:406-11. 27. Bais JM, van der Borden DM, Pel M, et al. Vaginal birth after caesarean section in a population with a low overall caesarean section rate. Eur J Obstet Gynecol Reprod Biol 2001;96: 158-62. 28. Ravelli AC, Tromp M, van Huis M, et al. Decreasing perinatal mortality in the Netherlands, 2000-2006: a record linkage study. J Epidemiol Community Health 2009;63:761-5. 29. Iams JD, Goldenberg RL, Mercer BM, et al. The preterm prediction study: can low-risk women destined for spontaneous preterm birth be identified? Am J Obstet Gynecol 2001;184:652-5. 30. Gelson E, Gatzoulis M, Steer PJ, et al. Tetralogy of Fallot: maternal and neonatal outcomes. BJOG 2008;115:398-402. 31. CEMACH. In: Lewis G, editor. Saving mothers' lives: reviewing maternal deaths to make motherhood safer—2003-2005. The seventh report of the confidential enquiries into maternal deaths in the United Kingdom. London: CEMACH; 2007; 117-130. 32. Ferrero S, Colombo BM, Ragni N. Maternal arrhythmias during pregnancy. Arch Gynecol Obstet 2004;269:244-53. 33. Li JM, Nguyen C, Joglar JA, et al. Frequency and outcome of arrhythmias complicating admission during pregnancy: experience from a high-volume and ethnically-diverse obstetric service. Clin Cardiol 2008;31:538-41.

Imaging and Diagnostic Testing

Left atrial reverse remodeling and functional improvement after mitral valve repair in degenerative mitral regurgitation: A real-time 3-dimensional echocardiography study Nina Ajmone Marsan, MD, a Francesco Maffessanti, MS, b Gloria Tamborini, MD, a Paola Gripari, MD, a Enrico Caiani, PhD, b Laura Fusini, MD, b Manuela Muratori, MD, a Marco Zanobini, PhD, a Francesco Alamanni, MD, a and Mauro Pepi, MD a Milan, Italy

Background Severe mitral regurgitation is often associated with left atrium (LA) enlargement, which is a well-known predictor of adverse cardiovascular outcomes. However, only few data are available on the effect of mitral valve (MV) repair on LA size. The aim of this study was to evaluate, using real-time 3-dimensional echocardiography, the changes in LA volumes after MV repair. Methods A total of 65 patients with severe mitral regurgitation due to MV prolapse and scheduled for repair at an early stage were enrolled. Before the procedure, real-time 3-dimensional echocardiography was performed to assess LA volumes (maximum, before atrial active contraction [preA], and minimum). The same evaluation was repeated 6 months and 1 year after MV repair. Twenty healthy subjects matched for age and gender were enrolled as a control group. Results

Before MV repair, patients showed significantly higher values of LA volumes (maximum 43 ± 14 mL/m2, preA 33 ± 12 mL/m2, minimum 23 ± 11 mL/m2) as compared to controls (maximum 22 ± 6 mL/m2, preA 13 ± 4 mL/m2, minimum 8 ± 3 mL/m2). Six months after the operation, LA volumes significantly decreased (maximum 25 ± 8 mL/m2, preA 18 ± 8 mL/m2, minimum 13 ± 5 mL/m2), with a further reduction at 1-year follow-up (maximum 23 ± 7 mL/m2, preA 15 ± 7 mL/m2, minimum 11 ± 5 mL/m2), resulting in values similar to controls. The extent of LA reverse remodeling was inversely correlated with age (r = −0.42) and postoperative transmitral mean pressure gradient (r = −0.32), whereas a positive correlation was found with the reduction in left ventricular volume after MV repair (r = 0.35).

Discussion In patients with severe mitral regurgitation due to MV prolapse, MV repair, when performed at an early stage, results in a significant LA reverse remodeling. (Am Heart J 2011;161:314-21.)

Mitral valve (MV) repair is currently the preferred technique for the treatment of severe mitral regurgitation, secondary to either MV prolapse or flail.1,2 This approach is, in fact, associated with better survival as compared to valve replacement and allows preserving both subvalvular apparatus and left ventricular (LV) geometry and function.3 However, little is known about the effect of MV repair on left atrium (LA) size and function,4-7 which is of additional clinical importance. In fact, chronic mitral regurgitation is often associated with LA enlargement, which is a well-known predictor of adverse cardiovascuFrom the aCentro Cardiologico Monzino IRCCS, Milan, Italy, and Biomedical Engineering, Politecnico di Milano, Milan, Italy. Submitted June 25, 2010; accepted October 18, 2010.

b

Department of

Reprint requests: Nina Ajmone Marsan, MD, Department of Cardiovascular Sciences, Centro Cardiologico Monzino IRCCS, Via Parea 4, 20138 Milan, Italy. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.029

lar events such as stroke, atrial fibrillation, heart failure, and death. 8-10 Accurate assessment of LA size is, therefore, crucial and, as recommended,11 should be based on LA volume measurement, taking into account that LA dilatation may lead to significantly asymmetric geometry. Recently, real-time 3-dimensional (3D) echocardiography (RT3DE) has been demonstrated, in a headto-head comparison with magnetic resonance (MR) imaging, to be more accurate and reproducible than 2dimensional (2D) echocardiography for the quantification of LA volumes.12,13 Furthermore, RT3DE may be a novel reliable technique for the assessment of LA function, providing information about phasic changes of LA volumes during the cardiac cycle.14-17 The aim of this study was to evaluate, using RT3DE, the changes in LA size and function at mid- and long-term follow-up after MV repair. In particular, the study focused on patients with degenerative severe mitral regurgitation referred to MV repair at an early stage when no symptoms or significant LV alterations have

American Heart Journal Volume 161, Number 2

occurred. In addition, potential clinical and echocardiographic factors correlated with the changes in LA size after the procedure were explored.

Methods The study population consisted of 70 consecutive patients with severe mitral regurgitation due to degenerative MV prolapse or flail and scheduled for repair at the Centro Cardiologico Monzino, Milan, Italy. Mitral regurgitation was defined as severe in the presence of effective regurgitant orifice area ≥0.4 cm2 estimated by proximal isovelocity surface area technique, according to recent recommendations.18 All patients were referred to surgery in an early stage when no symptoms and no significant LV dilation or dysfunction occurred (LV ejection fraction [LVEF] N60%, LV end-systolic [ES] diameter b45 mm according to recent guidelines1,2). Patients with persistent atrial fibrillation in the preoperative period or during the follow-up were excluded to allow a complete evaluation of LA function. Additional exclusion criteria were (1) presence of MV stenosis or aortic valve disease, (2) history of endocarditis, and (3) coronary artery disease (previous myocardial infarction, percutaneous revascularization, or surgical coronary revascularization). The surgical procedure for MV repair varied according to valvular morphology and the preference of the surgeons. In all cases, a ring annuloplasty was performed to stabilize the annulus and the suture line. The procedure was considered successful in the absence of any significant residual regurgitation (more than mild), stenosis (maximum mean gradient N6 mm Hg), or systolic anterior motion of the anterior leaflet. Before the procedure, severity of mitral regurgitation, transmitral flow pattern, and pulmonary pressures were evaluated by conventional 2D color Doppler echocardiography. Real-time 3D echocardiography was performed to assess LV and LA size and function. The same evaluation was repeated 6 months and 1 year after MV repair to assess the changes in LV and LA size and function. Twenty healthy subjects (without structural heart disease or history of systemic hypertension) matched for age and gender were enrolled as a control group. The protocol was approved by the Institutional Review Board, and informed consent was obtained in all patients and subjects.

Real-time 3D echocardiography Data acquisition. Transthoracic RT3DE was performed using a Philips iE33 system (Philips Medical Systems, Andover, MA) equipped with an X3-1 fully sampled matrix transducer. Apical full-volume data sets were obtained, combining 7 electrocardiogram-triggered wedge-shaped subvolumes within 1 breath-hold (frame rate: 16-37 frames per second) to provide a larger pyramidal volume. Care was taken to include both LV and LA cavities in the 3D data set. Real-time 3D echocardiography data sets were stored digitally, and quantitative analysis was performed offline using a semiautomated contour-tracing algorithm (Q-Lab version 7.0; Philips Medical Systems); postprocessing of the images required 5 to 7 minutes. A 3D data set was considered unsuitable for analysis if N2 segments of the LV or LA walls could not be visualized or if it contained visible translation artifacts.

Marsan et al 315

Assessment of LV volumes and ejection fraction. During postprocessing of the 3D data set, the software automatically displays the apical 4- and 2-chamber views and the parasternal short-axis view. After initial identification of the apex and the mitral annulus with 5 reference points on the enddiastolic (ED) and ES frames, a preconfigured ellipse is fitted to the endocardial border for each frame. A manual adjustment of the endocardial border was performed if required. An LV 3D model is automatically generated, and LV volumes and ejection fraction are calculated. Left ventricular volumes were indexed to the body surface area (BSA). Assessment of LA volumes and function. Quantification of LA size and function was performed using the same semiautomated contour-tracing algorithm, as for the LV, marking 5 atrial reference points (Figure 1). Manual corrections of the automatic trace were made to exclude LA appendage and pulmonary vein entrance. Left atrium volume was measured at 3 phases of the cardiac cycle (Figure 1): (1) maximum volume (Vmax) at end-systole, just before MV opening; (2) minimum volume (Vmin) at end-diastole, just before MV closure; and (3) volume before atrial active contraction (VpreA), obtained from the last frame before MV reopening. Left atrial volumes were indexed to the BSA.11 As previously described,14,17,19 LA function was explored assessing the following indices: 1. Total atrial emptying fraction (EF) = [(Vmax − Vmin)/ Vmax] × 100; 2. Active atrial emptying fraction (EFactive) = [(VpreA − Vmin)/ VpreA] × 100, as an index of LA booster pump function; and 3. Passive atrial emptying fraction (EFpassive) = [(Vmax − VpreA) / Vmax] × 100, as an index of LA conduit function.

Two-dimensional echocardiography Before the operation, severity of mitral regurgitation was defined according to the current guidelines.18 Immediately after MV repair, the same approach was applied to assess the residual regurgitation. In addition, postoperative mean diastolic pressure gradient across the repaired valve was calculated from the apical 4-chamber view using continuous wave Doppler and according to the simplified Bernoulli equation. Before the procedure and during follow-up, systolic pulmonary artery pressure was obtained by the noninvasive Doppler echocardiograph method from the systolic right ventricular— right atrial gradient calculated from the peak velocity of systolic transtricuspid regurgitant flow signal by the modified Bernoulli equation, and the right atrial pressure was derived by means of the inferior vena cava collapsibility index measured in the subcostal view.20

Statistical analysis Continuous data are presented as mean ± SD, whereas categorical variables, as absolute numbers or percentages. Normal distribution of continuous variables was assessed using the Kolmogorov-Smirnov test. General linear model (GLM) for repeated measures was applied to investigate changes in echocardiographic variables during the follow-up. Significant differences were evaluated using the Tukey honestly significant difference post hoc test. Moreover, results from the MV group were compared to controls using the unpaired Student t test corrected for multiple comparisons. Correlation

American Heart Journal February 2011

316 Marsan et al

Figure 1

Assessment of LA volumes using RT3DE. A, During postprocessing, the endocardial border (yellow line) is automatically detected for the entire 3D data set over the whole cardiac cycle and is visualized in the apical 4-chamber, 2-chamber, and short-axis views (from left to right). B, The software automatically provides the LA 3D model (yellow shell); the changes of LA volumes over time (during the cardiac cycle) are plotted as a curve from which 3 main values can be obtained: (1) LA Vmax at end-systole, (2) LA Vmin at end-diastole, and (3) LA VpreA obtained from the last frame before MV reopening.

between postoperative reduction in LA Vmax and clinical and echocardiographic variables was evaluated with the Pearson correlation coefficient. For all tests, P b .05 was considered significant. A statistical software program, SPSS 16.0 (SPSS Inc, Chicago, IL), was used for statistical analysis. In addition, in a randomly chosen subgroup of 20 patients, image analysis was repeated by an additional investigator as well as by the same primary reader at least 1 week later. During these repeated analyses, the investigators were blinded to the results of all prior measurements. Interobserver and intraobserver variability was computed as the coefficient of variation, defined as the absolute difference of the corresponding pair of repeated measurements in percent of their mean in each patient and then averaged over the 20 patients.

Funding This study is part of the SurgAid project (http://www.surgaid. org), cofunded by the Italian Ministry of University and Research, Rome, Italy (PRIN2007).

The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the manuscript, and its final contents.

Results None of the 70 patients had intraoperative complications. Residual MR was trivial or mild in all cases either immediately after the operation or during follow-up. Baseline LV diameters were 59 ± 8 mm (range 43-80 mm) and 33 ± 7 mm (range 20-43 mm), evaluated at ED and ES frames, respectively. Five patients (7%) were excluded from further analysis because of suboptimal 3D images due to stitching artifacts or inadequate definition of the LA endocardial blood/tissue interface, hampering the accuracy of the 3D endocardial surface reconstruction. Clinical characteristics of the remaining 65 patients are summarized in Table I.

American Heart Journal Volume 161, Number 2

Marsan et al 317

Table I. Clinical characteristics of the study population: healthy subjects (controls) in comparison with patients undergoing MV repair Controls

Figure 2

MV repair

n 20 65 Age (y) 59 ± 6 61 ± 12 Gender (male/female) 12/8 (60%/40%) 43/22 (66%/34%) BSA (m2) 1.8 ± 0.2 1.8 ± 0.2 HR (beat/min) 71 ± 15 70 ± 10 Ring size (mm) – 30 ± 2 Medications ACE inhibitors/ATII antagonists – 27 (42%) Diuretics – 58 (89%) Amiodarone – 29 (45%) β-Blockers – 14 (22%) Ca2+ antagonist – 5 (8%) HR, Heart rate; ACE, angiotensin-converting enzyme; ATII, angiotensin II receptor.

Table II. Changes in LV volumes and function in mean transmitral pressure gradient and in systolic pulmonary artery pressure at baseline and during follow-up after MV repair, in a direct comparison with the values obtained in healthy subjects (controls) MV repair Controls Baseline LV EDV/BSA (mL/m2) LV ESV/BSA (mL/m2) LVEF (%) SPAP (mm Hg) Mean transmitral pressure gradient (mm Hg)

54 21 62 26 1.8

± ± ± ± ±

8 5 7 3 0.3

73 ± 22‡ 28 ± 10‡ 61 ± 6 34 ± 4‡ –

6m 54 24 56 25 4.7

± ± ± ± ±

13⁎ 9⁎ 7⁎,‡ 4⁎

1.6‡

12 m 56 23 59 27 4.3

± ± ± ± ±

13⁎ 8⁎ 8† 5⁎ 1.5‡

EDV, ED volume; ESV, ES volume; SPAP, systolic pulmonary artery pressure. ⁎ P b .05 versus baseline, GLM for repeated measures. † P b .05 versus 6 months, GLM for repeated measures. ‡ P b .05 MV repair versus controls, unpaired t test, Bonferroni correction for multiple comparison.

Left ventricular volumes Table II shows the values of LV volumes and function at baseline and during follow-up after MV repair, in a direct comparison with the reference values obtained in healthy subjects. Before surgery, a mild LV dilatation was noted in the patients compared to controls, but normal values of LVEF were observed. Six months post surgery, LV volumes significantly decreased compared to the presurgical values, whereas LVEF depicted an absolute reduction of approximately 5%. The changes in LV volumes persisted 1 year after surgery, whereas LVEF showed a significant improvement toward presurgical values. Changes in LA volumes after MV repair As displayed in Figure 2, before MV repair, patients showed significantly higher values of LA volumes (Vmax = 43 ± 14 mL/m2, VPreA = 33 ± 12 mL/m2, Vmin = 23 ± 11

Changes in LA volumes (Vmax, Vmin, and VpreA) from baseline to 6- and 12-month follow-up after MV repair. The values are displayed in comparison to the corresponding normality range (mean value and 95% CI), computed from the normal group (red bar). ⁎P b .05 versus baseline, GLM for repeated measures; §P b .05 MV repair versus controls, unpaired t test, Bonferroni correction for multiple comparison.

318 Marsan et al

American Heart Journal February 2011

Figure 3

Example of a patient with a significant reduction of LA Vmax at 6-month and 1-year follow-up after MV repair. In the upper panels, the apical 4-chamber views are at an ES frame, showing the LA endocardial border detection. In the bottom panels, the corresponding 3D LA models are displayed and show a clear reduction in LA size.

mL/m2) as compared to healthy subjects (Vmax = 22 ± 6 mL/m2, VPreA = 13 ± 4 mL/m2, Vmin = 8 ± 3 mL/m2). Six months after the operation (Figure 2), LA volumes significantly decreased (Vmax = 25 ± 8 mL/m2, VPreA = 18 ± 8 mL/m2, Vmin = 13 ± 5 mL/m2), with a further reduction at 12-month follow-up (Vmax = 23 ± 7 mL/m2, VPreA = 15 ± 7 mL/m2, Vmin = 11 ± 5 mL/m2), resulting in values similar to controls. Figure 3 shows an example of a patient with a significant reduction of LA Vmax at 6-month and 1-year follow-up after MV repair. As previously reported,4 a significant LA reverse remodeling was defined as a decrease of LA Vmax ≥15%. Accordingly, 6 months after the operation, 53 patients (82%) showed a significant LA reverse remodeling, whereas at 12-month follow-up, the number of patients with significant LA reverse remodeling increased to 55 (85%). Furthermore, according to the definition of the last recommendations for chamber quantification,11 the number of subjects with severely dilated LA (LA Vmax ≥ 40 mL/m2) decreased from 36 (55%) before surgery to 6 (9%) and 2 (3%) at 6- and 12-month follow-up, respectively. Conversely, the number of subjects fulfilling the reference range values (LA Vmax ≤ 28 mL/m2)

increased from 10 (15%) to 43 (66%) and 46 (72%) during follow-up.

Correlates of LA volumes reduction Several clinical and echocardiographic variables were correlated to LA Vmax reduction after MV repair (Table III). Because significant LA reverse remodeling occurred at 6 months with negligible improvements at 12 months, correlation analysis was performed only between preoperative parameters and those at 6 months after surgery. Age was found negatively correlated with the degree of LA reverse remodeling. The extent of postoperative LA Vmax reduction was related to the LV ED volume reduction 6 months after MV repair and inversely to postoperative transmitral mean pressure gradient. Ring size did not show any significant correlation with the entity of LA reverse remodeling. Changes in LA function after MV repair Before surgery, LA emptying fractions, global, active, and passive, were significantly lower than in healthy subjects (Table IV). At 6-month follow-up, after MV repair, LA EF showed an absolute increase of approximately 5% from the baseline value. This improvement was maintained at 1-year follow-up (Table IV).

American Heart Journal Volume 161, Number 2

Marsan et al 319

Table III. Correlates of LA percentage decrease in maximum volume 6 months after MV repair (expressed as Pearson coefficient and P value)

Variable Age (y) Gender Ring size Postoperative mean transmitral pressure gradient LV EF LA Vmax LV dEDV (%)

LA dVmax (%)

P

−0.42 −0.03 0.21 −0.32

b.001 .84 .13 .02

−0.04 0.63 0.53

.81 b.001 b.001

LV dEDV, Percentage decrease in LV ED volume 6 months after MV repair; LA dVmax, LA percent decrease in Vmax.

A similar trend was observed for LA EFpassive, with no significant difference compared to normal values after the surgical procedure (Table IV). In turn, LA EFactive was significantly impaired at baseline and did not change significantly during followup (Table IV).

Reproducibility Reproducibility analysis showed an interoperator variability in the computation of LA volumes equal to 10.5%, 10.5%, and 11.4% for Vmax, VpreA, and Vmin, respectively. Intraoperator variability was also good with coefficient of variation of 5.8%, 5.9%, and 7.1% for Vmax, VpreA, and Vmin, respectively.

Discussion The main findings of the current study can be summarized as follows: (1) both at 6 months and 1 year after MV repair, a significant LA reverse remodeling was observed, in parallel with a significant reduction in LV volumes; (2) the extent of LA reverse remodeling was inversely correlated with age and postoperative transmitral mean pressure gradient, whereas a positive correlation was found with the reduction in LV volume after the procedure; and (3) MV repair also resulted in a significant improvement in global LA function.

Changes in LA volumes after MV repair The assessment of LA size is of particular importance in patients with mitral regurgitation because the disease directly impacts on the LA,9 which plays an important role compensating the volume overload and preventing pulmonary congestion.21 Furthermore, LA enlargement, specifically defined as LA Vmax ≥ 40 mL/m2, showed an important prognostic value being a predictor of heart failure and stroke occurrence, new onset of atrial fibrillation, and cardiovascular death.8-10 Consequently, whether MV surgery results in a significant reduction of LA volume is also of clinical interest, taking into consideration the potential prognostic implications.

Table IV. Left atrial total EF, EFpassive, and EFactive obtained in healthy subjects (controls) and for patients undergoing MV repair, at baseline, 6-month, and 12- month follow-up after surgery MV repair

LA EF (%) LA EFpassive (%) LA EFactive (%)

Controls

Baseline

6m

12 m

61 ± 6 39 ± 8 37 ± 10

47 ± 10† 29 ± 8† 26 ± 9†

51 ± 10⁎,† 32 ± 12⁎ 27 ± 10†

52 ± 9⁎,† 34 ± 11⁎ 27 ± 10†

⁎ P b .05 versus baseline, GLM for repeated measures. † P b .05 MV repair versus controls, unpaired t test, Bonferroni correction for multiple comparison.

However, little is known on the effect of MV surgery, and in particular of MV repair, on LA size.4-6 The most recent recommendations for cardiac chamber quantification indicate LA volume as the preferred measurement of LA size.11 In fact, LA shape is often irregular, and 2D measurements, such as LA anteroposterior or superoinferior diameters, are therefore less reliable. Furthermore, LA volume assessment by 2D echocardiography is limited by significant geometric assumptions and low reproducibility because of diverging position and orientation of imaging planes. In the present study, LA volume assessment was performed using RT3DE that, in a direct comparison with MR imaging, demonstrated to be far more accurate than 2D echocardiography and with the most favorable test-retest variation.12,13 In particular, the current study evaluated the changes in LA volume after MV repair in patients with degenerative severe mitral regurgitation who were referred to surgery at an early stage when no symptoms and no significant LV dilation or dysfunction occurred. A severely dilated LA (LA Vmax ≥ 40 mL/m2) was noted before the operation in N50% of the patients. Six months and 1 year after the procedure, a significant LA reverse remodeling occurred, with a normalization of the LA volumes at long-term follow-up. These results are in line with initial studies that applied only conventional 2D echocardiography.4,6 However, the extent of LA reverse remodeling after MV repair was greater than in previous reports, with only 3% of patients with significant LA enlargement after the procedure. These findings support the importance of an early surgical approach in degenerative severe mitral regurgitation, when most of the cardiac alterations are still reversible. In fact, LA reverse remodeling was found to be correlated with the reduction in LV ED volume after the procedure, underlying the important interrelationship between these 2 cardiac chambers. In previous studies,4,5 this relationship could not be reported, probably because the patients were mainly operated on when symptoms and significant LV dilatation/dysfunction already occurred. However, our findings could not be univocally attributed to the early surgical approach

American Heart Journal February 2011

320 Marsan et al

because of lack of a “delayed surgical approach” counterpart. To this respect, future randomized studies could give new insights on the relation between surgical timing and LA remodeling. In the present study, the value of LA volume before the operation was found to be correlated also with the extent of LA reverse remodeling, probably suggesting that when operated on at an early stage, patients with LA enlargement can still experience a normalization of LA size. Obviously, patients with already normal or nearly normal LA dimensions may not show significant changes after MV repair. Age-induced histologic changes in LA myocardium (mainly fibrosis) may explain the inverse correlation noted between age and LA reverse remodeling after the operation.4,22 However, an advanced age at operation may also suggest a longer history of MR with a lower likelihood of reversible LA morphological changes. In the reference ranges, though, the postoperative mean transmitral pressure gradient was found inversely correlated with the reduction in LA volume after repair, suggesting that even a mild pressure overload may affect LA remodeling process. Consequently, the evaluation of MV repair results should carefully take into account this parameter. In turn, no relation was found between LA reverse remodeling and the ring size.

Changes in LA function after MV repair Although the prognostic value is not clearly demonstrated, clinical interest has been growing in the analysis of LA function. However, accurate noninvasive measurements for LA function are still lacking. Recently, RT3DE has been proposed as a novel technique for the assessment of the different LA functions (global, passive, and active), providing an automatic detection of the phasic changes in LA volume during the cardiac cycle.14-17 In the current study, a significant LA reverse remodeling was observed after MV repair, together with a significant improvement in LA global emptying fraction (LA ejection fraction). In particular, LA ejection fraction was significantly impaired before the operation as compared to healthy subjects and showed a significant improvement already 3 months after the surgical operation. This improvement was mainly the result of a normalization of the LA passive emptying fraction, which is largely dependent on the loading conditions and is strongly related to LV function.14,23 Consequently, the significant reverse remodeling and the improved hemodynamic conditions of the LV after MV repair were reflected on the LA passive function. In turn, LA active emptying fraction, which was significantly reduced before the implantation, did not show significant changes. The LA active function is mainly determined by the intrinsic contractility of LA myocardium and partially by the LA VPreA, according to the Frank-Starling

effect.14,23 Consequently, it can be hypothesized that, in these patients, no improvement in LA active function could be observed because of 2 reasons: (1) irreversible structural changes (such as fibrosis or eccentric hypertrophy) occurred even at an early stage and affected LA myocardial contractility and (2) the compensatory FrankStarling effect could not be activated because a significant reduction of LA VPreA occurred after the procedure.

Study limitations The main limitations of this study are the following: 1. The frame rate of RT3DE is still lower than that of 2D echocardiography. In our study, this issue may limit the accuracy in the evaluation of the LA VPreA and, therefore, may interfere on accurate assessment of LA passive and active function. However, this does not affect the calculation of LA Vmax, LA Vmin, and LA EF because these measurements are not significantly influenced by a low frame rate. 2. Exclusion criteria included the presence of persistent atrial fibrillation, which is often associated with severe MR because full-volume acquisition mode is based on electrocardiogram triggering and because arrhythmias would result in artifacts. The upcoming single-beat real-time 3D imaging promises to allow reliable LV and LA volume analysis also in this patient population. 3. The effects of MV repair on LA volume and function in patients with MR associated with LV dysfunction are yet to be determined, although this goal would be difficult to achieve because patients are usually referred to surgery before severe dysfunction occurs. 4. Moreover, 45% of patients were treated with amiodarone. Despite these considerations, the marked LA reverse remodeling demonstrated in our series (toward a normalization of LA volumes) further reinforces the importance of an early surgical repair in patients with severe mitral regurgitation. 5. Finally, larger studies are needed to confirm the results of the present study, and the prognostic implications of a LA reverse remodeling after MV repair should be demonstrated with a longer clinical follow-up.

Conclusions In patients with severe degenerative mitral regurgitation, MV repair, when performed at an early stage, results in a significant LA reverse remodeling and improvement of function as demonstrated by RT3DE.

References 1. Bonow RO, Carabello BA, Chatterjee K, et al. Focused update incorporated into the ACC/AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American

American Heart Journal Volume 161, Number 2

2.

3.

4.

5.

6.

7.

8.

9.

10.

11. 12.

College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1998 Guidelines for the Management of Patients With Valvular Heart Disease): endorsed by the Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. Circulation 2008;118:e523-661. Vahanian A, Baumgartner H, Bax J, et al. Guidelines on the management of valvular heart disease: The Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology. Eur Heart J 2007;28:230-68. Enriquez-Sarano M, Schaff HV, Orszulak TA, et al. Valve repair improves the outcome of surgery for mitral regurgitation. A multivariate analysis. Circulation 1995;91:1022-8. Antonini-Canterin F, Beladan CC, Popescu BA, et al. Left atrial remodelling early after mitral valve repair for degenerative mitral regurgitation. Heart 2008;94:759-64. Cho DK, Ha JW, Chang BC, et al. Factors determining early left atrial reverse remodeling after mitral valve surgery. Am J Cardiol 2008; 101:374-7. Dardas PS, Pitsis AA, Tsikaderis DD, et al. Left atrial volumes, function and work before and after mitral valve repair in chronic mitral regurgitation. J Heart Valve Dis 2004;13:27-32. Geidel S, Lass M, Schneider C, et al. Downsizing of the mitral valve and coronary revascularization in severe ischemic mitral regurgitation results in reverse left ventricular and left atrial remodeling. Eur J Cardiothorac Surg 2005;27:1011-6. Benjamin EJ, D'Agostino RB, Belanger AJ, et al. Left atrial size and the risk of stroke and death. The Framingham Heart Study. Circulation 1995;92:835-41. Messika-Zeitoun D, Bellamy M, Avierinos JF, et al. Left atrial remodelling in mitral regurgitation—methodologic approach, physiological determinants, and outcome implications: a prospective quantitative Doppler-echocardiographic and electron beamcomputed tomographic study. Eur Heart J 2007;28:1773-81. Reed D, Abbott RD, Smucker ML, et al. Prediction of outcome after mitral valve replacement in patients with symptomatic chronic mitral regurgitation. The importance of left atrial size. Circulation 1991;84: 23-34. Lang RM, Bierig M, Devereux RB, et al. Recommendations for chamber quantification. Eur J Echocardiogr 2006;7:79-108. Khankirawatana B, Khankirawatana S, Lof J, et al. Left atrial volume determination by three-dimensional echocardiography reconstruc-

Marsan et al 321

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

tion: validation and application of a simplified technique. J Am Soc Echocardiogr 2002;15(10 Pt 1):1051-6. Poutanen T, Ikonen A, Vainio P, et al. Left atrial volume assessed by transthoracic three dimensional echocardiography and magnetic resonance imaging: dynamic changes during the heart cycle in children. Heart 2000;83:537-42. Anwar AM, Geleijnse ML, Soliman OI, et al. Left atrial Frank-Starling law assessed by real-time, three-dimensional echocardiographic left atrial volume changes. Heart 2007;93:1393-7. Anwar AM, Soliman OI, Geleijnse ML, et al. Assessment of left atrial volume and function by real-time three-dimensional echocardiography. Int J Cardiol 2008;123:155-61. Marsan NA, Bleeker GB, Ypenburg C, et al. Real-time threedimensional echocardiography as a novel approach to assess left ventricular and left atrium reverse remodeling and to predict response to cardiac resynchronization therapy. Heart Rhythm 2008;5:1257-64. Marsan NA, Tops LF, Holman ER, et al. Comparison of left atrial volumes and function by real-time three-dimensional echocardiography in patients having catheter ablation for atrial fibrillation with persistence of sinus rhythm versus recurrent atrial fibrillation three months later. Am J Cardiol 2008;102:847-53. Zoghbi WA, Enriquez-Sarano M, Foster E, et al. Recommendations for evaluation of the severity of native valvular regurgitation with twodimensional and Doppler echocardiography. J Am Soc Echocardiogr 2003;16:777-802. Leung DY, Boyd A, Ng AA, et al. Echocardiographic evaluation of left atrial size and function: current understanding, pathophysiologic correlates, and prognostic implications. Am Heart J 2008;156: 1056-64. Pepi M, Tamborini G, Galli C, et al. A new formula for echo-Doppler estimation of right ventricular systolic pressure. J Am Soc Echocardiogr 1994;7:20-6. Braunwald E, Awe WC. The syndrome of severe mitral regurgitation with normal left atrial pressure. Circulation 1963; 27:29-35. Thomas L, Levett K, Boyd A, et al. Changes in regional left atrial function with aging: evaluation by Doppler tissue imaging. Eur J Echocardiogr 2003;4:92-100. Spencer KT, Mor-Avi V, Gorcsan III J, et al. Effects of aging on left atrial reservoir, conduit, and booster pump function: a multiinstitution acoustic quantification study. Heart 2001;85:272-7.

Congestive Heart Failure

Certoparin versus unfractionated heparin to prevent venous thromboembolic events in patients hospitalized because of heart failure: A subgroup analysis of the randomized, controlled CERTIFY study Ulrich Tebbe, MD, a Sebastian M. Schellong, MD, b Sylvia Haas, MD, c Horst Eberhard Gerlach, MD, d Claudia Abletshauser, PhD, e Christian Sieder, MSc, e Peter Bramlage, MD, f and Hanno Riess, MD g Detmold, Dresden, München, Mannheim, Nürnberg, Mahlow, and Berlin, Germany

Background Despite the elevated risk for developing venous thromboembolic events in patients with heart failure, there are no randomized, double-blind, controlled trial data on the comparison of low-molecular-weight heparin with unfractionated heparin (UFH) in this patient population. Methods This was a subgroup analysis of the CERTIFY trial, which included 3,239 nonsurgical, acutely ill medical patients 70 years or older. Patients were randomized to receive 3,000-U anti-Xa certoparin once daily or 5,000-IU UFH 3 times a day. The analysis was performed on a subgroup of 542 patients diagnosed with heart failure at hospital admission. Results

Patients with heart failure differed from patients without heart failure in that they were more likely using antiplatelets (67.2% vs 48.9%; P b .0001) and had a lower glomerular filtration rate (8.0% vs 5.5%; ≤30 mL/min per 1.73 m2; P = .0232). Thromboembolic risk was comparable except for a higher incidence of distal deep venous thrombosis (DVT) in patients with heart failure (10.80% vs 7.26%; P = .0144). Within the heart failure population, patient characteristics were comparable between randomized treatment groups. The incidence of the primary end point (proximal DVT, symptomatic nonfatal pulmonary embolism, and venous thromboembolism–related death combined) was numerically, slightly smaller with certoparin (3.78% vs 4.74% with UFH; odds ratio 0.79, 95% CI 0.32-1.94), and the incidence of major bleeding was 0.72% with certoparin versus 0.38% with UFH.

Conclusions Patients hospitalized for heart failure are at high risk for developing distal DVT and bleeding complications compared with acutely ill medical patients without heart failure. Within the heart failure population, the observed differences in prophylactic efficacy between 3,000-U anti-Xa certoparin once daily and 5,000-IU UFH 3 times a day were similar to those observed in the overall study population; this suggests that certoparin might be at least as effective as UFH also in this subgroup. There were no relevant differences in bleeding risk or frequency of adverse events. (Am Heart J 2011;161:322-8.)

Among the hospitalized medical patients, those with heart failure are at a particularly high risk for venous thromboembolism (VTE).1-4 Proposed additional thromboembolic risk factors in patients with heart failure are

From the aKlinikum Lippe-Detmold, Detmold, Germany, bKlinikum Friedrichstadt, Dresden, Germany, cInstitut für experimentelle Onkologie und Therapieforschung, Technische Universität München, München, Germany, dT6, 25, 68161 Mannheim, Germany, Novartis Pharma GmbH, Nürnberg, Germany, fInstitute for Cardiovascular Pharmacology and Epidemiology, Mahlow, Germany, and gCharité, Campus Virchow Klinikum, Berlin, Germany. RCT reg no. NCT00451412. Submitted July 2, 2010; accepted October 1, 2010. e

Reprint requests: Ulrich Tebbe, MD, Klinikum Lippe GmbH, Fachbereich Herz-Kreislauf, Röntgenstrasse 18, 32756 Detmold, Germany. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.005

aberrant blood flow dynamics (dilated cardiac chambers, wall motion abnormalities, poor contractility, and in some cases atrial fibrillation), a hypercoagulable state, and an abnormal endocardial surface. In addition to this, more general risk factors for VTE such as advanced age, placement of a central venous or pulmonary artery catheter, obesity, and reduced mobility are frequent comorbid conditions. Both lowmolecular-weight heparins (LMWHs) and unfractionated heparins (UFHs) are recommended for thromboprophylaxis in these patients by the recent American College of Chest Physicians guidelines.5 In deciding whether to use LMWH or UFH, LMWHs are frequently preferred over UFH because of their improved bioavailability, reduced risk of heparin-induced thrombocytopenia (HIT), a prolonged predictable anticoagulant response, and once-daily subcutaneous administration.6,7

American Heart Journal Volume 161, Number 2

There is, however, no specific double-blind, randomized, head-to-head comparison of LMWH and UFH in this patient population, although heart failure is a frequent medical condition in placebo-controlled trials such as MEDENOX, PREVENT, and ARTEMIS (25%-52%) and in trials versus UFH (13%-50%).8-13 The only specific analysis on patients with heart failure has been conducted in the open PRINCE study, which demonstrated that 40-mg enoxaparin once daily was at least as effective as UFH.11 Prompted by the need for more specific data on the efficacy and safety of LMWH versus UFH in this quantitatively important high-risk patient group, we prespecified a subgroup analysis of the randomized, double-blind CERTIFY study.14 CERTIFY has demonstrated noninferiority of the LMWH 3,000-U anti-Xa certoparin once daily versus 5,000-IU UFH 3 times a day for the prophylaxis of VTE in acutely ill, nonsurgical patients 70 years or older. More specifically, we aimed to compare patient groups with or without heart failure with respect to thromboembolic risk and bleeding complications during heparin use and to compare the efficacy and safety of certoparin and UFH in the subgroup of patients with heart failure.

Patients and methods CERTIFY (the randomized, double-blind, multicenter comparison of the efficacy and safety of certoparin [3000 U of antiFXa once daily] with UFH [5000 IU t.i.d.] in the prophylaxis of thromboembolic events in acutely ill medical patients) was a study of 3,239 hospitalized medical patients aged 70 years or older and an expected significant decrease in mobility for at least 4 days, which had been randomized to receive either 3,000 U anti-Xa of certoparin (Mono-Embolex; Novartis Pharma GmbH, Nürnberg, Germany) or 5,000 IU of UFH (Liquemin N 5000; Hoffmann-LaRoche AG, Grenzach-Wyhlen, Germany) in a double-blind design.14 Patients assigned to the UFH treatment group received 5,000 IU of subcutaneous UFH TID. Patients assigned to the certoparin treatment group received a single daily dose of 3,000 U anti-Xa certoparin administered subcutaneously, followed by 2 placebo injections to match the total number of injections with UFH. Exclusion criteria for CERTIFY were immobilization longer than 3 days before randomization, immobilization due to cast or fracture, expected major surgical or invasive procedure within 3 weeks after randomization, patients with severe sepsis or need for ventilatory support (permitted were continuous positive airway pressure, oxygen mask, etc), LMWH/heparin administration longer than 48 hours in the 5 days before randomization, indication for anticoagulation or thrombolysis, life expectancy b6 months or illness with very high acute mortality (N30%), acute symptomatic deep venous thrombosis (DVT)/pulmonary embolism (PE), acute or history of HIT type II, acute or history of nonhemorrhagic stroke (b3 months), hemorrhagic stroke or intracranial bleeding (b12 months), acute or ongoing intracranial disease, high risk of gastrointestinal bleeding, spinal or epidural anesthesia, lumbar punction within the last 12 hours, uncontrolled hypertension, severe liver or renal disease, acute endocarditis, known active retinopathy, and intravitreal or other intraocular bleeding.

Tebbe et al 323

This subgroup analysis of the CERTIFY trial included patients with a diagnosis of heart failure on hospital admission (physicians' diagnosis).

End points The primary efficacy end point for CERTIFY was the combined incidence of proximal DVT, symptomatic nonfatal PE, and VTE-related death occurring during the core study (covering the treatment period of 8-20 days, after which compression ultrasound sonography was performed). Secondary end points included each of the components of the primary efficacy measure; the incidence of distal DVT (alone and in combination with proximal DVT); the incidence of symptomatic DVT; the incidence of a combination of proximal DVT, nonfatal PE, and death from all causes including PE; the incidence of death from all causes; and the incidence of documented symptomatic VTE (PE and/or DVTs). Safety end points included major and minor bleeding complications and HIT. Heparin-induced thrombocytopenia was diagnosed and treated according to local standards, and the drug was discontinued in case of suspected HIT.

Bleeding complications Major bleeding was defined as fatal bleeding, clinically overt bleeding associated with a fall of the hemoglobin level concentration N20 g/L compared with the baseline hemoglobin level concentration, clinically overt bleeding that required transfusion of ≥2 U of packed red cells or whole blood, and symptomatic bleeding in a critical area or organ (intracranial, intraspinal, retroperitoneal, and pericardial).

Statistical analysis All patients who received at least 1 dose of study drug were included in the safety analysis (safety population). Not all patients had a complete evaluation including ultrasound on follow-up. Therefore, the number of patients from the total (safety) population evaluable for each of the end points is indicated in the respective figures and tables. Point estimates and respective 95% CI were calculated. For details of the statistical analysis of the overall trial, see Riess et al.14 P values were determined from 2-sample t tests for continuous or from asymptotic odds ratio (OR) tests (logistic regression) for binary variables. P values were determined from a univariate logistic regression model or from the interaction term of a logistic regression model with factors treatment, subgroup, and treatment subgroup, as appropriate.

Results Patient with or without heart failure From a total of 3,239 patients in CERTIFY, 542 were classified by the treating physician to have heart failure at hospital admission (2,697 controls without heart failure). Patients with heart failure were more likely to be using antiplatelets (67.2% vs 48.9%; P b .0001) and had a lower glomerular filtration rate (GFR) (8.0% vs 5.5%; ≤30 mL/ min per 1.73 m2; P = .0232) than patients without heart failure at hospital admission. Other variables were

American Heart Journal February 2011

324 Tebbe et al

Table I. Baseline demographic characteristics for patients with or without heart failure at hospital admission (safety population) Patients with heart failure at hospital admission

Female (%) Mean age ± SD (y) Mean body weight ± SD (kg) Antiplatelet use Yes No GFR (mL/min per 1.73 m2)⁎ ≤30 30 b GFR ≤ 60 N60 Hospitalization, mean ± SD (d) Immobilization, mean ± SD (d) Mean exposure ± SD (d)

Patients without heart failure

Certoparin (n = 277)

UFH (n = 265)

Total (n = 542)

Total (n = 2697)

P value with vs without heart failure

62.1 79.2 ± 6.2 73.0 ± 16.9

164 (61.9) 78.9 ± 6.0 73.6 ± 16.9

336 (62.0) 79.0 ± 6.1 73.3 ± 16.9

1579 (58.5) 78.8 ± 6.3 71.8 ± 15.5

.1366 .4978 .0430

67.1% 32.9%

67.2% 32.8%

67.2% 32.8%

48.9% 51.1%

b.0001 b.0001

8.4% 54.6% 37.0% 12.0 ± 5.3 9.4 ± 3.0 9.0 ± 2.8

7.6% 56.3% 36.1% 12.4 ± 5.0 9.8 ± 4.3 9.1 ± 2.9

8.0% 55.4% 36.6% 12.2 ± 5.1 9.6 ± 3.7 9.0 ± 2.9

5.5% 51.5% 43.0% 12.4 ± 6.3 10.0 ± 4.4 9.1 ± 3.5

.0232 .1085 .0071 .4873 .0477 .5330

⁎ Six missing values for patients with heart failure and 37 missing values for patients without heart failure (P values from 2-sample t tests for continuous or from asymptotic OR tests [logistic regression] for binary variables).

Figure 1

Risk for the combined primary end point in patients with or without heart failure, irrespective of prophylaxis. P values from univariate logistic regression model. HF, heart failure.

comparable between both patient groups as well as between the certoparin and UFH groups (Table I). The risk for the combined end point was comparable in patients with or without heart failure (4.26% vs 4.22%; P, nonsignificant) (Figure 1). There was, however, an increased risk for the incidence of distal DVT (10.80% in patients with heart failure and 7.26% in patients without heart failure; P = .0144), which translated to an overall risk increase for the combination of proximal or distal DVT (P = .03331).

Adverse events were more frequent in patients with heart failure (66.2%) than without (60.4%; P = .0115), and risk of major bleeding was comparable between groups. All-cause mortality was comparable in both patient groups (Table II).

Patients with heart failure receiving certoparin or UFH Of the 542 patients with heart failure, 277 had been randomized to receive certoparin, and 265, to receive UFH. Both treatment groups were comparable with respect to baseline characteristics (Table I). Patients were hospitalized for a mean duration of 12.2 ± 5.1 days and received thromboprophylaxis for a mean of 9.0 ± 2.8 days (certoparin) and 9.1 ± 2.9 days (UFH). Four hundred ninety patients completed the core study. The reasons for discontinuation are displayed in Table III. A total of 238 patients in the certoparin group and 232 patients in the UFH group were evaluated for the combined end point of proximal DVT, symptomatic nonfatal PE, and VTE-related death (Figure 2). The risk was 3.78% with certoparin and 4.74% with UFH, which corresponded to a difference in proportions of −0.96 (95% CI −4.61 to 2.69) and an OR of 0.79 (95% CI 0.32-1.94). There is no evidence that the relative efficacy of certoparin in the heart failure population might be different from that in patients without heart failure, as indicated by the overlapping CIs and the lack of significance in the heterogeneity test. Figure 2 also illustrates the ORs between certoparin and UFH in the single components of the combined end point and all-cause death in patients with or without heart failure. There was no difference in efficacy between certoparin and UFH in the rates of proximal and/or distal DVT, symptomatic DVT, and symptomatic nonfatal PE. There were, however, more cases of all-cause death in patients with heart failure

American Heart Journal Volume 161, Number 2

Tebbe et al 325

Table II. Safety data for patients with or without heart failure Patients with heart failure at hospital admission

N Patients with any bleeding complications, n (%) With minor bleeding With major bleeding HIT type II⁎ Patients with AEs, n (%) With suspected drug relation Leading to dose adjustment or study drug interruption Leading to study drug discontinuation Requiring concomitant medication/nondrug therapy Patients with SAEs, n (%) Deaths With suspected drug relation Leading to study drug discontinuation

Patients without heart failure

Certoparin

UFH

Total

Total

P value with vs without heart failure

277 11 (3.97) 9 (3.25) 2 (0.72) 1 (0.36) 185 (66.8) 7 (2.5) 1 (0.4) 6 (2.2) 121 (43.7) 16 (5.8) 6 (2.2) 1 (0.4) 5 (1.8)

265 10 (3.77) 9 (3.40) 1 (0.38) 2 (0.75) 174 (65.7) 14 (5.3) 4 (1.5) 13 (4.9) 115 (43.4) 15 (5.7) 0 (0.0) 3 (1.1) 7 (2.6)

542 21 (3.87) 18 (3.32) 3 (0.55) 3 (0.55) 359 (66.2) 21 (3.87) 5 (0.92) 19 (3.51) 236 (43.5) 31 (5.72) 6 (1.11) 4 (0.74) 12 (2.21)

2697 105 (3.89) 92 (3.41) 14 (0.52) 0 (0) 1630 (60.4) 72 (2.67) 15 (0.56) 109 (4.04) 1161 (43.0) 169 (6.27) 35 (1.30) 7 (0.26) 54 (0.20)

.9836 .9158 .9194 – .0115 .1275 .3256 .5593 .8320 .6296 .7173 .0949 .7502

AE, Adverse event; SAE, serious adverse event. ⁎ One case was rated as unclear in each group (P values from asymptotic OR tests).

Table III. Reasons for premature discontinuation in patients with heart failure by treatment group Certoparin

Treated Completed core study Discontinued core study Adverse event(s) Subject withdrew consent Death Subject's condition no longer requires study drug Administrative problems Abnormal laboratory value(s) Protocol violation

UFH

Total

n

%

n

%

n

%

277 248 29 6 10 6 3 3 0 1

100.0 89.5 10.5 2.2 3.6 2.2 1.1 1.1 0.0 0.4

265 242 23 13 7 0 2 0 1 0

100.0 91.3 8.7 4.9 2.6 0.0 0.8 0.0 0.4 0.0

542 490 52 19 17 6 5 3 1 1

100.0 90.4 9.6 3.5 3.1 1.1 0.9 0.6 0.2 0.2

(certoparin 6/268, UFH 0/260), none of which was related to VTE (no patients in either group). The 6 patients instead died of sepsis, pulmonary edema, stroke, heart failure, and twice of cardiac arrest. Event rates for symptomatic DVT, nonfatal PE, and all-cause death were, however, very low in the heart failure population. All 542 patients with heart failure receiving trial medication were included in the safety evaluation (Table II). There were no noticeable differences in the rate of patients experiencing any bleeding complications (3.97% certoparin vs 3.77% UFH; OR 1.05, 95% CI 0.44-2.53). Adverse (66.8% vs 65.7%) and serious adverse events (5.8% vs 5.7%) were equally frequent in both treatment groups.

Discussion Heart failure is the second most common risk factor for VTE in hospitalized patients, as shown in ENDORSE.7

CERTIFY, which has demonstrated noninferiority of certoparin versus UFH in acutely ill medical patients in general, had included 542 patients with heart failure at hospital admission. Our analyses demonstrate a slight increase of thromboembolic risk (eg, distal DVT) and a higher propensity for (major) bleeding complications in patients with heart failure. It further shows that the effect size and relative efficacy of certoparin versus UFH documented in CERTIFY are well preserved in the subgroup of patients with heart failure and are not different from patients with admission diagnoses other than heart failure.

Comparison of study populations Patients with heart failure in CERTIFY were older (79.0 ± 6.1 vs 75 ± 11 years), were more often women (62.0% vs 57.1%), and had a similar body weight (73 vs 72 kg) in comparison with patients in PRINCE. However,

326 Tebbe et al

American Heart Journal February 2011

Figure 2

Combined end point of proximal DVT, symptomatic nonfatal PE, and VTE-related death in the CERTIFY study. The total number of patients displayed for every end point reflects that not all patients were evaluable for every single end point. P values from the interaction term of a logistic regression model with factors treatment, subgroup, and treatment ⁎ subgroup. HF, hear failure; n, events; N, evaluable.

only patients with New York Heart Association III/IV heart failure were included in PRINCE, which could have been causative for the higher mortality rates in PRINCE despite the lower mean age in the latter study (all-cause death in CERTIFY 1.1 % vs 5.8% in PRINCE). Thus, PRINCE and CERTIFY might represent different subsets of patients with heart failure (advanced disease in younger patients vs less advanced disease in the elderly). There is, however, a limitation to this CERTIFY subgroup analysis: patients were selected based on their admission diagnosis, and no further description of the severity of the disease or objective verification of the diagnosis has been available. On the other hand, PRINCE had an open-trial design, making the results prone to bias.

Incidence rates of VTE and efficacy of prophylaxis Placebo-controlled trials for the efficacy of LMWHs are available from the MEDENOX,15 PREVENT,16 and ARTEMIS studies.17 MEDENOX has shown that 40-mg enoxaparin (but not 20 mg) is beneficial in acutely ill medical patients with a reduction of symptomatic VTE and asymptomatic DVT from 14.9% with placebo to 5.5% with enoxaparin without increasing the risk of adverse

events.15 In PREVENT, the incidence of symptomatic VTE and asymptomatic proximal DVT was significantly reduced (5.0%-2.8%) with the use of 5,000-U dalteparin daily.16 Finally, in ARTEMIS, 2.5-mg fondaparinux was more effective than placebo in medical inpatients older than 60 years (5.6% with fondaparinux vs 10.5% with placebo).17 Although heart failure was one of the main risk factors leading to eligibility for inclusion into the trials (MEDENOX 34%, PREVENT 52%, and ARTEMIS 25% of patients), none of these trials presented subgroup analyses for patients with heart failure. On the other hand, there was a predefined subgroup of patients with heart failure in the open PRINCE study (50% of patients), which compared the efficacy and safety of enoxaparin with UFH. The incidence rate of proximal and distal DVT in CERTIFY (certoparin 12.4%, UFH 12.3%) in patients with heart failure was somewhat different compared with patients with advanced heart failure in PRINCE study11 (enoxaparin 9.7%, UFH 16.2%). It was, however, substantially higher than in medically ill patients hospitalized for other reasons than heart failure (certoparin 8.1%, UFH 9.9%). This might point at the increased thromboembolic risk of patients with heart failure. None of these differences were significant,

American Heart Journal Volume 161, Number 2

however (333 patients with heart failure were included in PRINCE, and 542 patients, in CERTIFY), resulting in no evidence for a differential efficacy of both treatment options for the prophylaxis of VTE. Further differences in treatment populations, as previously outlined, have to be considered when aiming to compare these study results. Proximal DVT (3.78% with certoparin, 4.74% with UFH) and, even more so, distal DVT (11.1% and 10.5%, respectively) were the most frequent thromboembolic events observed in patients with heart failure in CERTIFY, but the differences between both treatment options were not significant. For the PRINCE population, no respective results were reported.11 For the total population, including those with respiratory disease, the nominally lower incidence of thromboembolic events in patients treated with enoxaparin versus patients treated with UFH was primarily caused by a reduced frequency of distal DVT (5.9% with enoxparin, 8.5% with UFH) but not proximal DVT (2.1% and 1.9%, respectively, in the total study, results not available for patients with heart failure).

Safety Patients with heart failure were characterized by a more frequent use of antiplatelet agents and a lower mean GFR. The frequency of minor and major bleeding complications as well as both measures combined was comparable in both treatment groups, although total bleeding (on the account of minor bleeding events) was more frequent with UFH. It appears to be still warranted, however, to carefully consider concomitant pharmacotherapy (antiplatelet agents) when treating patients with heart failure. The incidence of all-cause mortality in certoparintreated patients with heart failure appears to be noteworthy. As to our perception, this observation is likely due to chance, based on the age and morbidity burden of patients included in CERTIFY. None of the reported patients died of VTE-related events but instead of sepsis, pulmonary edema, stroke, heart failure, and cardiac arrest. Limitations Despite the randomized, controlled, double-blind trial design of CERTIFY and the large sample size demonstrating noninferiority of certoparin versus UFH in acutely ill medical patients in the overall trial, the present analysis is limited by the following statistical and quantitative aspects: 1. Because of the lower sample size within this subgroup, the noninferiority documented for the overall study cannot be claimed for the subgroup of patients with heart failure. 2. Event rates for symptomatic DVT, nonfatal PE, and all-cause death were very low in the heart failure population. This is also illustrated by wide CIs in

Tebbe et al 327

specific subgroups of patients. Therefore, clinically relevant differences may have not been significant, and only a specific study with a large number of patients is able to specifically address this. 3. Further interpretation may be perceived to be limited by the validity of the complete compression ultrasound.18 This strategy has, however, been reliable in other trials in internal medicine such as PREVENT16 and EXCLAIM.19 Beyond that, there is a growing acceptance of ultrasound in phase III trials.20 4. Finally, there are countries in which there is limited access to UFH in prophylactic doses or to certoparin. Although LMWH is usually perceived to be more expensive than UFH, this assumption does not apply in major markets such as Germany, in which the costs of a prophylaxis per day is €5.08 for certoparin (Mono-Embolex 3000 IU; Novartis Pharma GmbH) and €6.74 for sodium heparin.

Conclusions Patients hospitalized for heart failure are at high risk for developing distal DVT and bleeding complications compared with acutely ill medical patients without heart failure. Within the heart failure population, 3,000U anti-Xa certoparin once daily was equally as effective as 5,000-IU UFH 3 times a day with no difference in bleeding risk or adverse events.

Disclosures This study was funded by Novartis Pharma GmbH, Nürnberg, Germany.

References 1. Howell MD, Geraci JM, Knowlton AA. Congestive heart failure and outpatient risk of venous thromboembolism: a retrospective, case-control study. J Clin Epidemiol 2001;54:810-6. 2. Heit JA, Silverstein MD, Mohr DN, et al. Risk factors for deep vein thrombosis and pulmonary embolism: a population-based case-control study. Arch Intern Med 2000;160:809-15. 3. Samama MM. An epidemiologic study of risk factors for deep vein thrombosis in medical outpatients: the Sirius study. Arch Intern Med 2000;160:3415-20. 4. Douketis JD, Foster GA, Crowther MA, et al. Clinical risk factors and timing of recurrent venous thromboembolism during the initial 3 months of anticoagulant therapy. Arch Intern Med 2000;160: 3431-6. 5. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous thromboembolism: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133:381S-453S. 6. Weitz JI. Low-molecular-weight heparins. N Engl J Med 1997;337: 688-98. 7. Cohen AT, Tapson VF, Bergmann JF, et al. Venous thromboembolism risk and prophylaxis in the acute hospital care setting (ENDORSE

328 Tebbe et al

8.

9.

10.

11.

12.

13.

study): a multinational cross-sectional study. Lancet 2008;371: 387-94. Lechler E, Schramm W, Flosbach CW. The venous thrombotic risk in non-surgical patients: epidemiological data and efficacy/safety profile of a low-molecular-weight heparin (enoxaparin). The Prime Study Group. Haemostasis 1996;26(Suppl 2):49-56. Harenberg J, Roebruck P, Heene DL. Subcutaneous low-molecularweight heparin versus standard heparin and the prevention of thromboembolism in medical inpatients. The Heparin Study in Internal Medicine Group. Haemostasis 1996;26:127-39. Bergmann JF, Neuhart E. A multicenter randomized double-blind study of enoxaparin compared with unfractionated heparin in the prevention of venous thromboembolic disease in elderly in-patients bedridden for an acute medical illness. The Enoxaparin in Medicine Study Group. Thromb Haemost 1996;76:529-34. Kleber FX, Witt C, Vogel G, et al. Randomized comparison of enoxaparin with unfractionated heparin for the prevention of venous thromboembolism in medical patients with heart failure or severe respiratory disease. Am Heart J 2003;145:614-21. Harenberg J, Kallenbach B, Martin U, et al. Randomized controlled study of heparin and low molecular weight heparin for prevention of deep-vein thrombosis in medical patients. Thromb Res 1990;59: 639-50. Poniewierski M, Barthels M, Kuhn M, Poliwoda H. [Effectiveness of low molecular weight heparin (Fragmin) in the prevention of thromboembolism in internal medicine patients. A randomized double-blind study]. Med Klin (Munich) 1988;83:241–245, 278.

American Heart Journal February 2011

14. Riess H, Haas S, Tebbe U, et al. A randomized, double-blind, multicenter study of certoparin versus UFH to prevent venous thromboembolic events in acutely ill, non-surgical patients: the CERTIFY Study. J Thromb Haemost 2010;8:1209-15. 15. Samama MM, Cohen AT, Darmon JY, et al. A comparison of enoxaparin with placebo for the prevention of venous thromboembolism in acutely ill medical patients. Prophylaxis in Medical Patients with Enoxaparin Study Group. N Engl J Med 1999;341: 793-800. 16. Leizorovicz A, Cohen AT, Turpie AG, et al. Randomized, placebocontrolled trial of dalteparin for the prevention of venous thromboembolism in acutely ill medical patients. Circulation 2004;110: 874-9. 17. Cohen AT, Davidson BL, Gallus AS, et al. Efficacy and safety of fondaparinux for the prevention of venous thromboembolism in older acute medical patients: randomised placebo controlled trial. BMJ 2006;332:325-9. 18. Schellong SM, Beyer J, Kakkar AK, et al. Ultrasound screening for asymptomatic deep vein thrombosis after major orthopaedic surgery: the VENUS study. J Thromb Haemost 2007;5:1431-7. 19. Hull RD, Schellong SM, Tapson VF, et al. Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recently reduced mobility: a randomized trial. Ann Intern Med 2010; 153:8-18. 20. Schellong SM. Venous ultrasonography in symptomatic and asymptomatic patients: an updated review. Curr Opin Pulm Med 2008;14: 374-80.

A randomized controlled trial evaluating the safety and efficacy of cardiac contractility modulation in advanced heart failure Alan Kadish, MD, a ,s Koonlawee Nademanee, MD, b,s Kent Volosin, MD, c,s Steven Krueger, MD, d,s Suresh Neelagaru, MD, e,s Nirav Raval, MD, f,s Owen Obel, MD, g,s Stanislav Weiner, MD, h,s Marc Wish, MD, i,s Peter Carson, MD, j,s Kenneth Ellenbogen, MD, k,s Robert Bourge, MD, l,s Michael Parides, PhD, m,s Richard P. Chiacchierini, PhD, n,s Rochelle Goldsmith, PhD, o,s Sidney Goldstein, MD, p,s Yuval Mika, PhD, q,s Daniel Burkhoff, MD PhD, o,q,s and William T. Abraham, MD r,s Chicago, IL; Inglewood, CA; Philadelphia, PA; Lincoln, NE; Amarillo, Dallas, and Tyler, TX; Atlanta, GA; Fairfax, and Richmond, VA; Birmingham, AL; New York, and Orangeburg, NY; Detroit, MI; and Columbus, OH

Background Cardiac contractility modulation (CCM) delivers nonexcitatory electrical signals to the heart during the absolute refractory period intended to improve contraction. Methods We tested CCM in 428 New York Heart Association class III or IV, narrow QRS heart failure patients with ejection fraction (EF) ≤35% randomized to optimal medical therapy (OMT) plus CCM (n = 215) versus OMT alone (n = 213). Efficacy was assessed by ventilatory anaerobic threshold (VAT), primary end point, peak VO2 (pVO2), and Minnesota Living with Heart Failure Questionnaire (MLWFQ) at 6 months. The primary safety end point was a test of noninferiority between groups at 12 months for the composite of all-cause mortality and hospitalizations (12.5% allowable delta). Results The groups were comparable for age (58 ± 13 vs 59 ± 12 years), EF (26% ± 7% vs 26% ± 7%), pVO2 (14.7 ± 2.9 vs 14.8 ± 3.2 mL kg−1 min−1), and other characteristics. While VAT did not improve at 6 months, CCM significantly improved pVO2 and MLWHFQ (by 0.65 mL kg−1 min−1 [P = .024] and −9.7 points [P b .0001], respectively) over OMT. Fortyeight percent of OMT and 52% of CCM patients experienced a safety end point, which satisfied the noniferiority criterion (P = .03). Post hoc, hypothesis-generating analysis identified a subgroup (characterized by baseline EF ≥25% and New York Heart Association class III symptoms) in which all parameters were improved by CCM. Conclusions

In the overall target population, CCM did not improve VAT (the primary end point) but did improve pVO2 and MLWHFQ. Cardiac contractility modulation did not have an adverse affect on hospitalizations or mortality within the prespecified boundaries. Further study is required to clarify the role of CCM as a treatment for medically refractory heart failure. (Am Heart J 2011;161:329-337.e2.)

From the aNorthwestern University, Chicago IL, bPacific Rim EP, Inglewood, CA, cUniversity of Pennsylvania, Philadelphia, PA, dBryan LGH, Lincoln, NE, eLone Star Arrhythmia, Amarillo, TX, fSt. Joseph's Research Institute, Atlanta, GA, gUT Southwestern, Dallas, TX, h Tyler Cardiovascular Consultants, Tyler, TX, iInova Arrhythmia Associates, Fairfax, VA, j Washington VA Hospital, Washington DC, kVirginia Commonwealth University School of Medicine, Richmond, VA, lThe University of Alabama, Birmingham, AL, mMount Sinai, New York, NY, nR.P. Chiacchierini and Associates, LLC, oColumbia University, New York, NY, pHenry Ford Hospital, Detroit, MI, qIMPULSE Dynamics, Orangeburg, NY, and rThe Ohio State University Heart Center, Columbus, OH. s On behalf of the FIX-HF-5 Investigators and Coordinators. See online Appendix for complete listing. Submitted August 9, 2010; accepted October 18, 2010. Reprint requests: Alan Kadish, MD, Division of Cardiology, Department of Medicine, Northwestern University Medical School, Chicago, IL 60611-2908. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.025

Cardiac resynchronization therapy (CRT) enhances pump function, improves quality of life, improves exercise tolerance, and reduces hospitalizations and mortality in the population of chronic heart failure (CHF) patients with ejection fraction (EF) ≤35% and New York Heart Association (NYHA) class III or IV symptoms with QRS duration N120 to 130 milliseconds.1-3 Nevertheless, b50% of CHF patients with decreased EF meet QRS duration criteria for CRT and approximately 30% of patients receiving CRT are considered nonresponders because their symptoms do not improve.1 Thus, there is a large unmet need for new therapies that can improve CHF symptoms, especially for medically refractory patients with normal QRS duration. Cardiac contractility modulation (CCM) is an electrical device–based approach developed for the treatment of CHF.4,5 Cardiac contractility modulation signals are nonexcitatory electrical signals applied during the cardiac

330 Kadish et al

absolute ventricular refractory period that enhance the strength of cardiac muscular contraction.6 Cardiac contractility modulation signal application is associated with normalization of phosphorylation of key proteins and expression of genes coding for proteins involved in regulation of calcium cycling and contraction.4,7,8 The results of prior clinical studies of CCM (delivered by the OPTIMIZER, Impulse Dynamics, Orangeburg, NY) have supported its safety and efficacy5 including a recent double-blind, double crossover study in 164 subjects.9 This latter study showed that 3 months of CCM treatment improved quality of life and exercise tolerance as judged by peak VO2 in patients with NYHA class II and III symptoms. The purpose of the present study was to test the longer-term safety and efficacy of CCM treatment.

Methods The FIX-HF-5 study was a prospective, randomized, parallelgroup, controlled trial of optimal medical therapy (OMT group) versus OMT plus CCM (CCM group) conducted at 50 centers in the United States. The details of the protocol, device implantation procedure, primary and secondary end points, and statistical analysis plan have been provided previously.10 In brief, the study included subjects ≥18 years old with EF ≤35%, with NYHA class III or IV symptoms despite medical treatment with angiotensin-converting enzyme inhibitor and/or angiotensin receptor blocker and β-blockers for at least 3 months with a baseline peak VO2 on cardiopulmonary stress testing (CPX) ≥9 mL O2 kg−1 min−1 who were in normal sinus rhythm and not indicated for a CRT device (ie, QRS duration b130 milliseconds). Unless there were extenuating circumstances, subjects were required to have an implantable cardioverter defibrillator (ICD). Subjects were excluded if they were hospitalized within 30 days of enrollment, were inotrope dependent, had N8,900 premature ventricular contractions per 24 hours on a baseline Holter monitor, had permanent atrial fibrillation, had a myocardial infarction within 90 days, had percutaneous coronary intervention within 30 days, or had coronary artery bypass surgery within 90 days of enrollment. After informed consent, all subjects underwent baseline evaluation that included CPX, Minnesota Living with Heart Failure Questionnaire (MLWHFQ), 6-minute hall walk test, NYHA class determination by a clinician blinded to therapy assessment, an echocardiogram, and a 24-hour Holter monitor. After meeting inclusion criteria, a device implant date was scheduled. This scheduled implant date served as the study start date (SSD) for all subjects. Subjects were then randomized (1:1) to either the OMT group or to the CCM group. Subjects randomized to the CCM group underwent OPTIMIZER device implantation on the SSD. The implant procedure and electrical characteristics of CCM signals have been detailed previously.10 Major follow-up visits were at 3 and 6 months.

Statistical considerations and analysis plan The primary effectiveness end point, as required by the US Food and Drug Administration (FDA), was the change from baseline in the ventilatory anaerobic threshold (VAT) measured on CPX. The primary analysis is based on a comparison of

American Heart Journal February 2011

“responder” rates between the CCM and OMT groups at the 24-week follow-up visit. An individual subject is considered a responder if VAT increases by ≥20% at 24 weeks compared to their respective baseline value. Comparison of responder rates between groups was by a 1-sided Fisher exact test with an α of .025. The primary analysis was based on the intent to treat population; imputation was used to account for missing data as detailed previously.10 Ten different imputations were performed and the data combined to arrive at an overall test of significance according to the methods of Rubin.11 Secondary efficacy end points were peak VO2 and quality of life assessed by MLWHFQ. Each of these parameters was also assessed by a responder analysis with a 20% increase in peak VO2 and a 10point reduction in MLWHFQ used to define responders. The type I error rate was maintained across multiple tests of efficacy by using a closed-form hierarchical testing procedure. Additional end points included changes in NYHA functional class (with a one-class change considered a response) and 6-minute walk (6MW) test (with a 40-m increase considered a response). In addition to the responders analyses, treatment effects were also assessed by applying traditional methods using comparison of mean changes from baseline in each parameter. These comparisons were made using 1-sided Student t tests (with equal or unequal variances as appropriate). Baseline characteristics were compared with 2-sided Wilcoxon rank sum test, Fisher exact test, Pearson χ2 test and 2-sample t tests as appropriate and as specified in the text. P values ≤.025 for 1-sided tests and ≤.05 for 2-sided tests were considered to be statistically significant. All statistical tests were performed using SAS Version 9.13 (SAS Institute, Cary, NC). The primary safety end point was the composite event rate of all-cause mortality and all-cause hospitalization through 50 weeks. The primary safety analysis was a test of the noninferiority of CCM therapy compared to OMT with respect to the proportion of subjects experiencing death or hospitalization within 50 weeks using the Blackwelder non-inferiority test12 with a prespecified noninferiority margin of 0.125. The noninferiority margin was selected to be 12.5% and α was set at .05, which resulted in a sample size of 198 subjects per group. A percentage of subjects (∼7%) were expected to be lost to followup, so that a total sample size of 428 subjects (214 per group) was selected. While the basis for sample size was the safety end point, power was computed for the expected difference between the control and test populations. It was anticipated that the control success rate in the responder analysis would be about 20% and that the rate in the test group would be N40%. Thus, the power of this difference is in excess of 95% with a sample size of N198 in each group.

Core laboratories and oversight committees Because of the upfront known difficulties in assessing VAT, significant measures were taken to optimize CPX quality.10 All CPX tests were sent to a single core laboratory where a detailed procedure was followed for objective determination of VAT (using the V-slope method13) by 2 independent readers blinded to treatment group. Ventilatory anaerobic threshold could not be assigned in tests without clear changes in slopes; these were classified as indeterminate. When discrepancies (amounting to N10% differences) arose between the 2 readers, a third reader was used and the final VAT was determined by the 2 closest values. When concordance could not be achieved,

American Heart Journal Volume 161, Number 2

Kadish et al 331

Table I. Baseline characteristics Variable Age (y) Male Ethnicity White Black Other Weight (kg) BMI (kg/m2) Resting HR (beat/min) SBP (mm Hg) CHF etiology Ischemic Idiopathic Other NYHA Class I Class II Class III Class IV Prior MI Prior CABG Prior PCI Diabetes QRS duration (ms) PVCs/24 h (Holter) LVEF (%) LVEDD (mm) MLWHFQ 6MW (m) CPX (core laboratory) Duration (min) Peak SBP (mm Hg) Peak HR (beat/min) Peak RER Peak VO2 (mL kg−1 min−1) VAT (mL kg−1 min−1)

OMT group (n = 213) Mean (SD) or n (%)

CCM group (n = 215) Mean (SD) or n (%)

P

58.55 (12.23) 151 (70.9%)

58.09 (12.79) 158 (73.5%)

.5109⁎ .5901†

142 45 26 93.30 30.95 73.74 115.61

(66.7%) (21.1%) (12.2%) (22.16) (6.53) (12.19) (17.61)

142 (66.7%) 48 (22.5%) 23 (10.8%) 0 1 183 29 126 86 83 102 101.51 1365 26.09 63.01 57.38 323.99 11.50 138.8 121.2 1.13 14.71 10.97

(0%) (0.47%) (85.92%) (13.62%) (59.15%) (40.38%) (38.97%) (47.89%) (12.81) (2001) 457 (1499)‖ (6.54) (8.56) (22.62) (92.44) (3.46) (24.6) (20.5) (0.09) (2.92) (2.18)

154 36 25 91.17 30.44 73.98 116.65

(71.6%) (16.7%) (11.7%) (23.27) (7.04) (13.13) (19.48)

139 (64.7%) 58 (27.0%) 18 (8.3%) 0 0 196 19 125 82 86 91 101.63 1323 25.74 62.41 60.49 326.38 11.34 139.7 122.1 1.14 14.74 10.95

(0%) (0%) (91.16%) (8.84%) (58.14%) (38.14%) (40%) (42.33%) (15.30) (1931) 339 (2136)‖ (6.60) (9.22) (23.00) (82.10) (3.20) (27.1) (20.2) (0.10) (3.06) (2.24)

.5026‡ .1632⁎ .2179⁎ .9681⁎ .8695⁎ .6465‡

.1720‡

.8449† .6923† .8437† .2853† .5968§ .5113⁎ .5641⁎ .7715⁎ .1109⁎ .5971⁎ .4814⁎ .9714⁎ .5223⁎ .5189⁎ .8575⁎

.9719§

HR, Heart rate; RER, respiratory exchange ratio; OMT, optimal medical therapy; CCM, cardiac contractility modulation; BMI, body mass index; SBP, systolic blood pressure; CHF, chronic heart failure; NYHA, New York Heart Association symptom class; MI, myocardial infarction; CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention; PVC, premature ventricular contractions; LVEF, left ventricular ejection fraction; LVEDD, left ventricular end diastolic dimension; MLWHFQ, Minnesota Living with Heart Failure Questionnaire; 6MW, 6-minute walk test; CPX, cardiopulmonary stress testing; VAT, ventilatory anaerobic threshold. ⁎ Two-sided Wilcoxon rank sum test. † Two-sided Fisher exact test. ‡ Two-sided Pearson χ2 test. § Two-sided unequal variance 2-sample t test. ‖ PVCs/24 h is nonnormally distributed; values provided are both mean (SD) and median (interquartile range).

tests were also classified as indeterminate. The core laboratory procedures have been detailed previously.10 Despite these efforts, it was anticipated that some tests would be classified as indeterminate because of poor test quality, inability of subjects to reach VAT, or because of subject noncompliance with scheduled follow-up visits. In an exploratory analysis, the impact of specific baseline characteristics (heart failure etiology, NYHA, and EF) on treatment effectiveness was evaluated using regression analysis. Prespecified cut points for the subgroup analysis included NYHA class III vs IV symptoms and LVEF dichotomized at 25%, which was the median value for the overall population. After an exploratory analysis revealed that CCM tended to be more effective in patients with less severe heart failure, a post hoc

analysis was performed on patients with LVEF N25% and NYHA class III heart failure. To ensure accuracy of the primary safety end point, an independent Events Adjudication Committee evaluated original records of every hospitalization and death. Protocol-specified hospitalizations included any admission that results in a calendar date change or was related to an adverse event that caused a prolongation of the index hospitalization for device implantation. An independent Data and Safety Monitoring Board was established to review aggregate safety data and monitor for the emergence of any significant safety concerns. The study was supported by IMPULSE Dynamics. The authors are solely responsible for the design and conduct of this study,

American Heart Journal February 2011

332 Kadish et al

Figure 1

Flow of patients through the study. W/D, Withdrawn.

all study analyses, the drafting and editing of the paper and its final contents.

Results Baseline characteristics of enrolled study subjects Between March 2005 and June 2007, 773 potential study subjects provided informed consent to participate in this study. From among these subjects, 428 subjects passed baseline screening and were randomized to either the OMT group (n = 213) or the CCM group (n = 215). The baseline characteristics of these subjects are summarized in Table I. These characteristics are similar between groups and are consistent with the study inclusion and exclusion criteria. Eightytwo percent of subjects had an ICD before entry into the study. Another 11% had an ICD placed at the start of the study. Another 2% of patients had devices implanted during the follow-up period so that overall, 95% of study subjects had an ICD. Implantable cardioverter defibrillator use was balanced between groups (202/213, 95% of the OMT group; 207/215, 96% of the CCM group). The reason why some patients did not have an ICD was because of patient refusal for a device. Patients in both groups were well medicated with angiotensin-converting enzyme inhibi-

tors or angiotensin receptor blockers (91%) and βblockers (93%) as detailed previously.10

Screening and randomization The flow of subjects through the course of the study is summarized in Figure 1. Of 774 subjects who signed informed consent, 429 subjects passed initial baseline testing and agreed to participate. One subject died before randomization. Four hundred twenty-eight subjects were randomized, 213 to the OMT group and 215 to the CCM group. As detailed in the figure, 17 subjects withdrew and 7 subjects died in the OMT group, so that a total of 189 (88.7%) subjects completed the 50-week follow-up period. In the CCM group, 3 subjects died before the implant and 7 subjects elected not to undergo device implantation. In 2 subjects, the implant was aborted; one because of right ventricular perforation that led to cardiac tamponade and one because of a substantially prolonged PR interval (∼300 milliseconds) that precluded CCM signal delivery for technical reasons. (After this experience, subjects with PR interval N275 milliseconds were excluded). Of the 203 subjects with a successful implant, 5 withdrew and 10 died so that 188 (92.6%) completed the 50 follow-up period. The implant procedure took 180 ± 91 minutes (median 165 minutes) and involved 2.4 ± 2.2 (median 1.5,

American Heart Journal Volume 161, Number 2

Kadish et al 333

Table II. Serious adverse events Randomization to SSD AE category General cardiopulmonary event Arrhythmias Worsening heart failure ICD/pacemaker system related Bleeding Localized infection Sepsis Neurologic dysfunction Thromboembolism (nonneurologic) General medical Total

SSD to 1 y

OMT (n = 213)

CCM (n = 215)

OMT (n = 212)

CCM (n = 210)

2 (2) 2 (2) 3 (3) 0 0 0 0 0 0 2 (2) 9 (8)

2 (2) 5 (4) 5 (5) 1 1 2 (2) 1 0 0 5 (5) 22 (13)

58 (46) 30 (25) 85 (50) 7 (6) 8 (8) 36 (29) 2 (2) 14 (12) 5 (5) 81 (54) 326 (115)

60 (42) 40 (29) 72 (50) 13 (11) 8 (6) 33 (27) 11 (10) 3 (3) 3 (3) 98 (63) 341 (129)

interquartile range 2.0) different electrode positions to reach an 8.1% ± 3.7% (median 7%) increase in dP/dtmax in response to acute CCM application.

Safety end points and adverse events For the composite safety end point of all-cause hospitalizations and all-cause mortality, 4 subjects in the CCM group and 14 subjects in the OMT group were withdrawn from the study before experiencing a safety end point and therefore lost to follow-up. Based on best efforts to confirm vital status (including a search of the death registries), none of these subjects died within the 50-week follow-up period. For the intent-to-treat population (assuming subjects lost to follow-up did not have any events), there were 103 events in the 213 subjects randomized to the OMT group (48.4%) and 112 events in the 215 subjects randomized to the CCM group (52.1%). Based on the Blackwelder test, the difference of 3.7% had an upper 1-sided 95% confidence limit of 11.7%, which was below the prespecified allowable 12.5% (P = .035). Thus, the primary safety end point of the study was met. As noted above, 7 (3.3%) of the 213 OMT subjects and 10 (4.9%) of the 203 subjects who received and OPTIMIZER system died during the 50-week follow-up period (P = .47, Fisher exact test). With an intent-to-treat analysis, 13 (6.0%) of the 215 subjects randomized to the CCM group died during the 50-week follow-up period (P = .25 vs OMT by Fisher exact test). A summary of serious adverse events (defined as any event that was considered life-threatening, required a hospitalization, or required invasive treatment) is provided in Table II. Several events were reported between the time of randomization and the SSD, slightly more in the CCM group (22 events in 13 patients) than in the OMT group (9 events in 8 patients, P = .027, Fisher exact test). Overall, serious adverse events were balanced between the groups, with 326 events reported

Table III. Device-related serious adverse events OPTIMIZER system related OPTIMIZER lead fracture OPTIMIZER RV lead dislodgement IPG problem/change OPTIMIZER RA lead dislodgement OPTIMIZER pocket dehiscence/erosion OPTIMIZER pocket infection OPTIMIZER pocket stimulation Lead perforation OPTIMIZER pocket bleeding Sensation due to CCM Extracardiac stimulation

30 (27) 3 6 2 6 3 2 2 2

(3) (6) (1) (5) (3) (2) (2) (2) 1 2 (2) 1

Number of events (number of patients). RV, Right ventricular; RA, right atrium; IPG, implanted pulse generator.

in 115 OMT patients versus 341 events in 129 CCM subjects (P = .66). Device-related serious adverse events are summarized in Table III. The most common adverse events were lead fracture or displacement. The total incidence of lead complications was 14 (7%).

Efficacy end points Results of primary and secondary efficacy variables and analyses are summarized in Figure 2. For each variable, the figure indicates the number of completed cases for each group, the mean changes from baseline, and the difference (and P value) between the changes. Results of the responders analysis are summarized in Table IV. Ventilatory anaerobic threshold (the primary efficacy parameter) decreased by 0.14 mL kg−1 min−1 in both groups at 24 weeks. For 17.6% of subjects in the CCM group and 11.7% of subjects in the OMT group, VAT increased by ≥20%; the difference in responder rates at 24 weeks (5.9%) was not statistically significant (P = .093). Data were missing for VAT from 59 subjects (27.7%) in the

334 Kadish et al

American Heart Journal February 2011

Figure 2

Efficacy results in the completed cases population. OMT, Optimal medical therapy; CCM, group receiving CCM signals.

OMT group and 56 subjects (26.0%) in the CCM group. For the responders analysis of the intent-to-treat population (which was the primary efficacy end point), an overall P value was obtained by combining information from 10 separate imputations11 with a final P value of .31. At 50 weeks, 14.4% of patients in the OMT group versus 23.7% of patients in the CCM group were responders, a difference of 9.3% (P = .027, completers analysis). Peak VO2 increased in the CCM group and decreased in the OMT group; the difference (0.65 mL kg−1 min−1) was statistically significant (P = .024). The responders analysis, however, did not show a difference in the percent of patients in which peak VO2 improved by ≥20%. The MLWHFQ and NYHA improved significantly more in the CCM group when analyzed either as differences in changes of mean values from baseline or with a

responders analysis. There were also nonsignificant (∼10 m) increases in 6MW distances. There was no significant difference between groups in ejection fraction or left ventricular end-diastolic dimensions.

Subgroup analyses The etiology of heart failure (ischemic versus nonischemic) was not associated with improvement. A multivariate model of the continuous variables done with Proc Mixed detected a statistically significant interaction between the treatment and composite variable of ejection fraction and NYHA class with a P value of .0219. Patients with an EF ≥25% in the CCM group had a 12.2% greater responder rate than those in the OMT group. Patients with NYHA class III in the CCM

American Heart Journal Volume 161, Number 2

Kadish et al 335

Table IV. Results of responders analyses for primary and secondary study end points at 24-week follow-up in the entire study cohort Parameter VAT (mL kg−1 min−1) VAT (mL kg−1 min−1) (ITT⁎) Peak VO2 (mL kg−1 min−1) MLWHFQ NYHA class 6MW (m)

OMT group (n = 213) n/N (%) LCL, UCL

CCM group (n = 215) n/N (%) LCL, UCL

CCM-OMT LCL, UCL (%)

18/154 (11.7) 7.1, 17.8 28/213 (13.2%) 8.9, 18.4 23/168 (13.7) 8.9, 19.8 77/184 (41.8) 34.6, 49.3 63/183 (34.4) 27.6, 41.8 51/173 (29.5) 22.8, 36.9

28/159 (17.6) 12.0, 24.4 38/215 (17.7%) 12.8, 23.4 31/179 (17.3) 12.1, 23.7 110/196 (56.1) 48.9, 63.1 94/191 (49.2) 41.9, 56.5 65/190 (34.2) 27.5, 41.4

5.9 −2.0, 13.9 4.5 −2.4, 11.5 3.6 −4.1, 11.3 14.3 4.2, 24.1 14.8 4.8, 24.5 4.7 −4.9, 14.2

P .093 .314 .233 .0037 .0026 .197

All data are based on completed cases population except for VAT, for which both completed cases and intent-to-treat populations are included. P values by 1-sided Fisher exact test. LCL, Lower confidence limit; UCL, upper confidence limit; ITT, intention to treat. ⁎ ITT population based on imputation of missing data.

group had a response rate that was 6.9% greater than those in the OMT group. Patients with NYHA class IV who were in the CCM group had a 7.3% lower response rate. Thus, an additional analysis was performed in those patients with LVEF ≥25% and NYHA class III. This subgroup was composed of 97 OMT and 109 CCM subjects, 48% of the overall population. In this subgroup, there were statistically and clinically significantly greater improvements in VAT (0.64 mL kg−1 min−1, P = .03 for the completed cases; P = .024 for the intention-to-treat population with imputed missing data), increased peak VO2 (1.31 mL kg−1 min−1, P = .001), improved MLWHFQ (10.8 points, P = .003), and improved NYHA (−0.29, P = .001) at 24 weeks. With regard to the primary safety end point, there were 42 events in the 97 OMT subjects (43.3%) compared to 52 events in the 109 CCM subjects (47.7%, Blackwelder test P = .12). From among these subjects, there were 2 deaths in the OMT group (0.9%) and 4 deaths in the CCM group from the SSD to 1 year (2.0%, P = .69, Fisher exact test).

Discussion Prior studies have provided evidence of safety and efficacy of 3 months CCM treatment in subjects with symptomatic heart failure with EF ≤35% and normal QRS duration.9 The FIX-HF-4 study9 enrolled 168 patients in a randomized, double-blind, double crossover study of patients with NYHA II or III symptoms and EF ≤35% showed an average increase of peak VO2 of ∼0.6 mL kg−1 min−1 and a reduction in MLWHFQ of ∼3 points with just 3 months of treatment. The present study was designed to test the longer-term effects of CCM treatment. The study demonstrated that CCM was safe within prespecified boundaries but did not meet the primary end point of an improvement in VAT.

The primary safety end point of the study, which was a noninferiority assessment of the composite of all-cause mortality and all-cause hospitalizations, was satisfied. The primary efficacy end point of the study, that is, the proportion of subjects whose VAT increased by N20% at 24 weeks, was not different between groups; nor was there any difference in the mean change of VAT from baseline. However, mean peak VO2 increased more in the CCM than OMT group at 24 weeks. We found that subjects in the CCM group exercised for longer durations, but there was no difference in respiratory exchange ratio at peak exercise between groups, indicating equal degrees of subject effort during exercise. In addition, there was no difference between groups at the earlier 12-week follow-up (data not shown), a time point at which the placebo effect was expected to be greatest. Thus, these supporting data argue against, although do not exclude, a placebo effect on peak VO2 as a cause for the difference between the groups at 24 weeks. The MLWHFQ improved by an average of 10 points more in the CCM group and 20% more subjects experienced a 10-point or greater reduction in the CCM group. However, this parameter is subject to placebo effect in the context of the present unblinded study. Nevertheless, the magnitude of this point reduction is similar to what has been reported previously for CRT.1 Similar effects were noted in NYHA, although the magnitudes were slightly less than reported for CRT.1

Limitations The results of the present study need to be interpreted within the context of several important and, in some respects, unique aspects of the study design that were less than ideal. It is important to note that the study design was developed under restrictions imposed by the

American Heart Journal February 2011

336 Kadish et al

FDA and study details were arrived at largely on the need to obtain safety data through 1 year of treatment. Because of difficulties ensuring blinding with a device that has a weekly recharging requirement, prior blinded study designs used in the evaluation of CRT devices were not applicable. It was also felt that implanting a device and leaving it “off” for 12 months was not ethical or practical. Thus, an unblinded design was used. Because most measures of quality of life and exercise tolerance used in heart failure studies are subjective, FDA required that VAT be the primary end point of the study because it is considered to be objective and not subject to placebo effect.14,15 However, VAT, although appealing from a theoretical perspective, has not been validated as an end point in heart failure trials and, when evaluated in a reallife application, there were extensive missing data due to inability to designate a value even when the test was conducted properly. Another unique aspect of the study is that the primary end point was analyzed through a responders analysis.16 The goal of using this approach, in contrast to the traditional comparison of mean changes, is to be able to more clearly define the population of subjects who exhibit a clinically meaningful benefit from the therapy. This may have certain advantages from a regulatory perspective. However, such an approach has not been used for primary and secondary end point analyses in prior heart failure studies. A dilemma in interpretation of the results is created by the fact that peak VO2 increased by a statistically significant amount (just slightly less than in prior studies of CRT) but failed to show an increase in the rate of “responders” (when defined as a 20% improvement from baseline). These unusual aspects of study design complicate interpretation of the results of the present study. Other study limitations should be considered. As is the case with most multicenter randomized studies of device therapies, study recruitment practices may differ among centers so that study subjects may not be consecutive that could result in selection bias compared to the overall population of eligible patients. Issues related to the relatively large number of imputations required for primary intent-to-treat efficacy analysis of anaerobic threshold parameter have been discussed. Although the primary effectiveness end point was not achieved in the overall study population, signs of efficacy were noted in the less severely ill subgroup in subjects with baseline ejection fraction ≥25% and NYHA class III symptoms. The explanation for this finding cannot be determined with certainty. One possible explanation is that the effects of CCM delivered only to the right ventricular septum are not substantial enough to overcome severe contractile dysfunction with severely reduced EF, which also correlates with a more severely enlarged heart. It could be that in such cases, CCM delivered to multiple sites (although less practical to implement) could be more effective. Another possibility

is that the molecular effects of CCM identified in prior studies8 are not as effective when the degree of dysfunction is too severe. Independent of the underlying reason, because this subgroup was identified based on a post hoc analysis, we consider these findings to be hypothesis generating. Patients enrolled in the present study were required to have an ICD. There was no increase in reports of ventricular arrhythmias, ICD shocks, or antitachycardia pacing. It is important to note that the risk of potential interference between the ICD and the OPTIMIZER device is eliminated at the time of implantation and by specifically designed testing procedure and recommendations for ICD device programming. There have been no reports of either inappropriate ICD firings or failure to detect an arrhythmia and deliver therapy when this testing and programming are performed. Cardiac resynchronization therapy is approved by the US FDA for subjects with QRS duration N120 to 130 milliseconds, EF ≤35%, and NYHA class III or IV symptoms despite appropriate medical therapy. However, NYHA functional class does not improve in approximately 30% of subjects receiving a CRT device.17,18 In addition, CRT has only been shown to be effective in patients with a prolonged QRS duration.19,20 Cardiac contractility modulation was developed several years ago to treat underserved populations.4,5 Prior short-term (3-month), double-blind studies showed CCM to be safe and effective.10 The results of the present study show that over a 1-year follow up period, CCM was safe within the prespecified boundaries. However, based on the prespecified primary end point, CCM efficacy was not demonstrated. Further studies will be required to determine the role of CCM in the treatment of patients with medically refractory heart failure.

Acknowledgements This study was supported by IMPULSE Dynamics. Y Mika is an employee and shareholder in IMPULSE Dynamics. D Burkhoff, A Kadish, and WT Abraham are consultants to IMPULSE Dynamics.

Disclosures The study was supported but a grant from Impulse Dynamics (New York, NY), manufacturer of the CCM device.

References 1. Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization in chronic heart failure. N Engl J Med 2002;346:1845-53. 2. Bristow MR, Saxon LA, Boehmer J, et al. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med 2004;350:2140-50.

American Heart Journal Volume 161, Number 2

3. Cleland JG, Daubert JC, Erdmann E, et al. Longer-term effects of cardiac resynchronization therapy on mortality in heart failure [the CArdiac REsynchronization-Heart Failure (CARE-HF) trial extension phase]. Eur Heart J 2006;27:1928-32. 4. Burkhoff D, Ben Haim SA. Nonexcitatory electrical signals for enhancing ventricular contractility: rationale and initial investigations of an experimental treatment for heart failure. Am J Physiol Heart Circ Physiol 2005;288:H2550-6. 5. Lawo T, Borggrefe M, Butter C, et al. Electrical signals applied during the absolute refractory period: an investigational treatment for advanced heart failure in patients with normal QRS duration. J Am Coll Cardiol 2005;46:2229-36. 6. Brunckhorst CB, Shemer I, Mika Y, et al. Cardiac contractility modulation by non-excitatory currents: studies in isolated cardiac muscle. Eur J Heart Fail 2006;8:7-15. 7. Imai M, Rastogi S, Gupta RC, et al. Therapy with cardiac contractility modulation electrical signals improves left ventricular function and remodeling in dogs with chronic heart failure. J Am Coll Cardiol 2007;49:2120-8. 8. Butter C, Rastogi S, Minden HH, et al. Cardiac contractility modulation electrical signals improve myocardial gene expression in patients with heart failure. J Am Coll Cardiol 2008;51:1784-9. 9. Borggrefe MM, Lawo T, Butter C, et al. Randomized, double blind study of non-excitatory, cardiac contractility modulation electrical impulses for symptomatic heart failure. Eur Heart J 2008;29:1019-28. 10. Abraham WT, Burkhoff D, Nademanee K, et al. A randomized controlled trial to evaluate the safety and efficacy of cardiac contractility modulation in patients with systolic heart failure: rationale, design, and baseline patient characteristics. Am Heart J 2008;156:641-8.

Kadish et al 337

11. Rudin D. Multiple imputation for nonresponse in surveys. New York: John Wiley and Sons; 1987. 12. Blackwelder WC. “Proving the null hypothesis” in clinical trials. Control Clin Trials 1982;3:345-53. 13. Sue DY, Wasserman K, Moricca RB, et al. Metabolic acidosis during exercise in patients with chronic obstructive pulmonary disease. Use of the V-slope method for anaerobic threshold determination. Chest 1988;94:931-8. 14. Milani RV, Lavie CJ, Mehra MR, et al. Understanding the basics of cardiopulmonary exercise testing. Mayo Clin Proc 2006;81: 1603-11. 15. Pina IL, Karalis DG. Comparison of four exercise protocols using anaerobic threshold measurement of functional capacity in congestive heart failure. Am J Cardiol 1990;65:1269-71. 16. Burkhoff D, Parides M, Borggrefe M, et al. “Responder Analysis” for assessing effectiveness of heart failure therapies based on measures of exercise tolerance. J Cardiac Fail 2009 In press. 17. Adamson PB, Abraham WT. Cardiac resynchronization therapy for advanced heart failure. Curr Treat Options Cardiovasc Med 2003;5: 301-9. 18. Birnie DH, Tang AS. The problem of non-response to cardiac resynchronization therapy. Curr Opin Cardiol 2006;21:20-6. 19. Beshai JF, Grimm RA, Nagueh SF, et al. Cardiac-resynchronization therapy in heart failure with narrow QRS complexes. N Engl J Med 2007. 20. Linde C, Abraham WT, Gold MR, et al. Randomized trial of cardiac resynchronization in mildly symptomatic heart failure patients and in asymptomatic patients with left ventricular dysfunction and previous heart failure symptoms. J Am Coll Cardiol 2008;52:1834-43.

American Heart Journal Volume 161, Number 2

Appendix A Arizona Arrhythmia Research Center, Scottsdale, AZ: Thomas Mattioni, MD, Vijay Swarup, MD, Sara Scrivano, Claudia Williams, RN; Arrhythmia Center for Southern Wisconsin, Ltd./St. Luke's Medical Center, Milwaukee, WI: Imran Niazi, MD, Nguyen Phan, MD, Rebecca Dahme, RN, Jo Ann Kiemen; Aurora Denver Cardiology Associates, Aurora, CO: Andrew I. Cohen, MD, Susan M. Polizzi, MD, Karen Bickett; Bryan LGH Heart Institute, Lincoln, NE: Andrew Merliss, MD, Steven K. Krueger, MD, June Christy, RN; California Pacific Medical Center, San Francisco, CA: Steven C. Hao, MD, Richard H. Hongo, MD, Eric J. Bernier, RN, Gina Im; Cardiovascular Associates, Kingsport, TN: Greg Jones, MD, Arun Rao, MD, Tammy Dicken; Cardiovascular Medical Group of Southern California, Beverly Hills, CA: Eli S. Gang, MD, Ronald P. Karlsberg, MD, Maria M.Thottam, Tracey S. Gerez; Center at St. Francis Hospital, Roslyn, NY: Steven M. Greenberg, MD, Rebecca Seeman, RN, Nedda Easterling; Center for Cardiac Arrhythmias, Houston, TX: Hue-Teh Shih, MD, Candace Pourciau; Comprehensive Cardiovascular Care, Milwaukee, WI: Masood Akhtar, MD, Anthony Chambers, RN; Deborah Heart & Lung Center, Trenton, NJ: Raffaele Corbisiero, MD, Linda Dewey, RN; Emory University Hospital, Atlanta GA: Jonathan Langberg, MD, Andrew Smith, MD, Sheila Heeke, RN, Jerilyn Steinberg, RN; Forsyth Medical Center, Winston-Salem, NC: David Smull, DO, Mark Mitchell, MD, Janice Dickson, RN; Harper University Hospital, Detroit, MI: Randy A. Lieberman, MD, Anne B. Mick; Heart & Vascular Institute of Texas, San Antonio, TX: Gregory A. Buser, MD, Armistead Lanford Wellford IV, MD, Edwin L. Whitney, MD, Steven W. Farris, RN; Henry Ford Hospital, Detroit, MI: Barbara Czerska, MD, Karen Leszczynski, RN; Inova Heart and Vascular Institute/Inova Fairfax Hospital: Marc Wish, MD, Ted Friehling, MD, Jessica Wolfe, RN, Marie Blake, RN; Lahey Clinic Medical Center, Burlington, MA: Roy M. John, MD, David T. Martin, MD, Bruce G. Hook, MD, Jean M. Byrne, RN; Lancaster Heart and Stroke Foundation, Lancaster, PA: Seth J. Worley, MD, Douglas C. Gohn, MD, Diane Noll, RN; Lone Star Arrhythmia and Heart Failure Center, Amarillo, TX: Suresh B. Neelagaru, MD, Tanya Welch, RN; Mayo Clinic, Rochester, MN: David L. Hayes, MD, Robert F. Rea, MD, Jane Trusty, RN, Mary (Libby) Hagen, RN; Midwest Heart Foundation, Lombard, IL: Maria Rosa Costanzo, MD, Lea Elder, RN; Moses Cone Hospital and Lebauer Cardiovascular Research Foundation, Greensboro, NC: Steve Klein, MD, Daniel Bensimhon, MD, Paul Chase; Mount Sinai Medical Center, Miami, FL: Gervasio A. Lamas, MD, Todd J. Florin, MD, Beatriz E. Restrepo, MD, MPH; Newark Beth Israel Medical Center, Newark, NJ; David A. Baran, MD, Laura Adams, RN; Northwestern University, Chicago, IL: Jeffrey Goldberger, MD, Dinita Galvez, RN, Katherine Small; NYU Medical Center,

Kadish et al 337.e1

New York, NY: Jill Kalman, MD, Cristina Surach, RN; Ochsner Health Systems, New Orleans, LA: Freddy Abi-Samra, MD, Timothy Donahue, MD, Melanie Lunn, Christine Hardy; Ohio State University, Columbus, Ohio: Charles C. Love, MD, Philip E. Binkley, MD, Garrie J. Haas, MD, Leah Sanuk, RN, Laura Yamokoski, RN; Hope Heart Institute, Bellevue, WA: J. Alan Heywood, MD, Amy Payne, RN; Pacific Rim EP, Inglewood, CA: Koonlawee Nademanee, MD, Carla Drew; Penn Presbyterian Medical Center, Philadelphia, PA: Kent Volosin, MD, Janet Riggs, MSN, RN; Riverside Regional Medical Center, Newport News, VA: Allan L. Murphy, MD, Virginia M. Oehmann, RN; Southern California Heart Centers, Stanley K. Lau, MD, Nita Cheng, RN, Peter Yiu; Spokane Cardiology/Deaconess Medical Center, Spokane, WA: Harold R. Goldberg, MD, Vickie Shumaker, RN; Stern Cardiovascular Center, Germantown, TN: Frank McGrew lll, MD, Barbara Hamilton, RN; St. Joseph's Research Institute, Atlanta, GA: Nirav Raval, MD, Nicolas Chronos, MD, Stephen P. Prater, MD, Sarah Conley; St. Lukes-Roosevelt Hospital Center, New York, NY: Jonathan S. Steinberg, MD, Marrick L. Kukin, MD, Robin Knox, RN, Cathleen B. Varley, RN; St. Paul Heart Clinic, St. Paul, MN: Alan Bank, MD, Stuart Adler, MD, R. Dent Underwood, MD, Lisa Tindell, RN; Texas Cardiac Arrhythmia Research, Austin, TX: Javier E. Sanchez, MD, G. Joseph Gallinghouse, MD, Deb S. Cardinal, RN, Chantel M. Scallon, RN; Tyler Cardiovascular Consultants, Tyler, TX: Stanislav Weiner, MD, Linda Holt; University of Alabama at Birmingham, Birmingham, AL: Jose Tallaj, MD, Tom McElderry Jr, MD, Karen Rohrer, RN; University of South Florida Heart Health, Tampa, FL: Bengt Herweg, MD, Robyn Aydelott-Nuce, RN, Mary Ann K. Yarborough, RN; University of Texas Southwestern Medical Center, Dallas, TX: Jose Joglar, MD, Owen Obel, MD, Carol Nguyen, RN, Dana Red, RN; University of Wisconsin, Madison, WI: Nancy Sweitzer, MD; Vanderbilt Heart and Vascular Institute, Nashville, TN: Mark Wathen, MD, Darwood Darber, MD, Nancy M. McDonough, RN, Lindee D. Dye, RN; Virginia Commonwealth University Health System/MCV Hospitals, Richmond, VA: Mark Wood, MD, Kenneth Ellenbogen, MD, Michael Hess, MD, Kim Hall, RN.

Appendix B Committees Steering Committee: William T. Abraham (Co-Chairman), Alan Kadish (Co-Chairman), Koonlawee Nademanee, Peter Carson, Robert Bourge, Kenneth A Ellenbogen and Michael Parides. Events Adjudication Committee: Peter Carson (Chairman), Christopher O'Connor, Inder Anand. Data Safety and Monitoring Board: Sidney Goldstein (Chairman), Stephen Gottlieb, Andrea Natale, David Naftel, David Callans.

337.e2 Kadish et al

Appendix C Core Laboratories Cardiopulmonary Stress Test: Rochelle Goldsmith, Columbia University.

American Heart Journal February 2011

Echocardiography: Marco DiTullio, Columbia University. NYHA Blinded Core Lab: Steven P. Schulman, The Johns Hopkins University.

Effects of n-3 polyunsaturated fatty acids on malignant ventricular arrhythmias in patients with chronic heart failure and implantable cardioverter-defibrillators: A substudy of the Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca (GISSI-HF) trial Andrea A. Finzi, MD, a Roberto Latini, MD, b Simona Barlera, MSc, b Maria G. Rossi, MD, c,k Albarosa Ruggeri, MD, d,k Alessandro Mezzani, MD, e,k Chiara Favero, BSc, b Maria G. Franzosi, BiolD, b Domenico Serra, MD, b Donata Lucci, MSc, f Francesca Bianchini, BSc, f Roberto Bernasconi,b Aldo P. Maggioni, MD, f Gianluigi Nicolosi, MD, g Maurizio Porcu, MD, h Gianni Tognoni, MD, i Luigi Tavazzi, MD, j and Roberto Marchioli, MD i Milano, Reggio Calabria, Veruno, Florence, Pordenone, Cagliari, S Maria Imbaro, and Cotignola, Italy; and Lugano, Switzerland

Background The antiarrhythmic effects of n-3 polyunsaturated fatty acids (n-3PUFA) in ischemic heart disease have been demonstrated; however, studies in patients surviving malignant ventricular arrhythmias of different etiologies treated with an implantable cardioverter-defibrillator (ICD) have given conflicting results. The purpose of this study was to assess the antiarrhythmic effect of n-3PUFA versus placebo in 566 patients with heart failure enrolled in the GISSI-HF trial who received an ICD for secondary or primary prevention of ventricular fibrillation (VF) or tachycardia (VT). Methods Clinical data and arrhythmic event recordings extracted from the device memory were obtained. We tested the treatment effect by a multivariate Cox model adjusting for all clinical parameters associated with the primary end point defined as time to first appropriate ICD discharge for VT/VF. Results In the 566 patients with at least one recorded follow-up visit, 1363 VT and 316 VF episodes were terminated by ICD pacing or shock over a median follow-up of 928 days. The incidence of the primary end point event was 27.3% in the n-3PUFA group and 34.0% in the placebo group (adjusted hazard rate = 0.80, 95% CI 0.59-1.09, P = .152). Patients who received 1, 2 to 3, or N3 ICD discharges were 8.9%, 7.1%, and 11.1% in the n-3PUFA group, compared with slightly higher rates of 11.1%, 10.7%, and 12.1% in the placebo group (overall P = .30). Patients with the highest 3-month increase in plasma n-3PUFA had a somewhat lower incidence of arrhythmic events. Conclusions

The results of this study, though not statistically significant, support prior evidences of an antiarrhythmic effect of n-3PUFA in patients with ICD, although they leave open the issue of whether this effect leads to a survival benefit. (Am Heart J 2011;161:338-343.e1.)

Patients with chronic heart failure (HF) and low left ventricular ejection fraction (LVEF) are at a significantly elevated risk of sudden cardiac death. Recent guidelines

strongly support a primary prevention strategy based on implantable cardiac defibrillators (ICDs).1 The introduction of cardiac resynchronization therapy (CRT) with

From the aUnit of Cardiovascular Medicine, Cardiopulmonary Dept, Fondazione Ca' Granda/Policlinico, Milano, Italy, bIstituto Mario Negri, Milano, Italy, cCardiocentro Ticino, Lugano, Switzerland, dPoliclinico Madonna della Consolazione, Reggio Calabria, Italy, e Fondazione Salvatore Maugeri, Veruno, Italy, fANMCO Research Center, Florence, Italy, g AO S Maria Angeli, Cardiology Unit, Pordenone, Italy, hAO Brotzu-S Michele, Cagliari, Italy, iConsorzio Mario Negri Sud, S Maria Imbaro, Italy, and jMaria Cecilia Hospital,

The GISSI Studies are endorsed by the Associazione Nazionale Medici Cardiologi Ospedalieri (ANMCO), by the Istituto di Ricerche Farmacologiche "Mario Negri", and by the Consorzio Mario Negri Sud, Italy. Submitted May 14, 2010; accepted October 15, 2010. Reprint requests: Roberto Latini, MD, Department of Cardiovascular Research, Istituto di Ricerche Farmacologiche Mario Negri, Via La Masa 19, 20156 MILANO, Italy.

GVM Care and Research, Ettore Sansavini Health Science Foundation, Cotignola, Italy. On behalf of the Investigators of the ICD substudy of GISSI-HF. See the online Appendix for complete listing. ClinicalTrials.gov Identifier: NCT00336336 (for the main study).

E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.032

k

American Heart Journal Volume 161, Number 2

biventricular pacemakers in HF patients nonresponsive to optimal medical therapy has expanded the number of subjects who receive an ICD as backup for primary prevention, as both treatments can be combined in the same pacing system.2 Epidemiological studies indicate that intake of marine long-chain n-3 polyunsaturated fatty acids (n-3PUFA) is associated with a reduction in cardiovascular (CV) mortality. This is supported by a number of basic studies providing evidence that n-3PUFA reduce surrogate markers of arrhythmia driven by Ca2+ overload.3 The Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto miocardico (GISSI)–Prevenzione trial found a lower overall mortality, mostly attributable to fewer sudden cardiac deaths among patients treated with n-3PUFA after myocardial infarction (AMI) compared with controls.4 GISSI-HF was the first large-scale, randomized, placebo-controlled trial showing that n-3PUFA can reduce mortality and CV hospitalization in patients with symptomatic HF of any etiology and with any level of LVEF.5 A recent reanalysis of GISSI-HF showed that fatal and nonfatal ventricular arrhythmic events were significantly reduced by n-3PUFA treatment.6 However, 3 trials in patients with an ICD implanted after malignant ventricular tachycardia (VT) or ventricular fibrillation (VF) of different etiology failed to demonstrate a significant reduction of ICD interventions or of mortality with n-3PUFA, even after pooled analysis.7-11 As a sizable proportion of GISSI-HF patients had an ICD implanted before or after entering the main trial, we report here the results of a substudy, designed and conducted in parallel to the main one, investigating the antiarrhythmic effect of n-3PUFA in HF patients with implanted ICD.

Methods Eighty-nine of the 325 centers participating in the GISSI-HF trial accepted to collect an additional set of data on their clinical history with respect to ICD management and reasons for ICD discharge. Patients who had an ICD implanted before entering the main trial or after randomization, either for a history of cardiac arrest, sustained VT, or syncope of suspected tachyarrhythmic origin or for primary prevention, were eligible for this substudy. Data were prospectively collected over the last period of the study, from September 2004 to December 2007.

Study end points The primary end points were time to an appropriate ICD intervention for major arrhythmias, that is, shock at high or low intensity or overdrive pacing for spontaneous ventricular tachyarrhythmia, either VT or VF, and the number of VT and VF episodes. Other end points were total and CV mortality, total hospitalizations, and hospitalizations for CV events (ie, worsening of HF, arrhythmias, and other CV causes). The ICD data were reviewed by a central ICD Core Laboratory for validation of arrhythmic events to rule out inappropriate ICD electrical treatment and to characterize the detected and treated pathologic electrograms. The clinical events recorded

Finzi et al 339

in the study were adjudicated blindly by an ad hoc committee on the basis of agreed definitions and procedures. All reports included a narrative summary with supporting documentation for every event (eg, clinical records, death certificates, and other relevant documentation).12

Procedures Patients were randomly assigned to receive 1 capsule per day of 1 g n-3PUFA (850 mg eicosapentaenoic acid and docosahexaenoic acid as ethyl esters in the average ratio of 1.2:1) or matching placebo. Patients eligible for the main GISSI-HF trial were randomized from August 6, 2002, to February 28, 2005. Patients were scheduled for clinical visits at 1, 3, 6, and then every 6 months until the end of the trial. The present substudy was conducted from September 2004 to the end of the main study, December 2007, in 89 centers. As patients could be admitted to the ICD substudy at any time, hospitalizations before ICD implant as well as those scheduled for ICD implant or maintenance were discarded from the analysis. The ICD programming was left to the cardiologist's decision. At each follow-up visit, the ICD was interrogated by telemetry; and the data were stored. Data from unscheduled visits for clinical events were also collected to ensure complete follow-up of ICD data. Blood samples for plasma levels of n-3 PUFA were obtained at baseline and after 3 months in 159 of the 566 patients enrolled in the ICD substudy.13

Statistical methods Baseline characteristics for randomized treatments, expressed as proportions, were compared by the χ2 test for categorical variables and by the t test for continuous variables. To estimate the effect of n-3PUFA on the primary end point, univariate and multivariable Cox proportional hazards models were fitted. Multivariable analysis was adjusted for baseline variables that were significantly related with the outcome at univariate analysis (P value b .05) (eg, sex, smoking, history of atrial fibrillation, hemoglobin ≤12 g/dL, previous pacemaker, previous ICD, previous percutaneous coronary intervention, bundle-branch block, pulmonary rales, hematocrit, use of direct-acting vasodilating agents). Kaplan-Meier estimates for the time to first appropriate ICD intervention by treatment were presented and compared by the Wilcoxon test. Time to outcome event was computed either from the time of randomization to the main trial, for patients implanted before entering GISSI-HF, or from time of ICD implant, for patients implanted during follow-up. Treatment effect was also analyzed for secondary end points by univariate and multivariable Cox models, adjusting for covariates associated with the outcome of interest. Treatment effects on changes of n-3 and n-3/n-6 fatty acids from baseline were analyzed by analysis of variance, adjusting for their baseline value. Analyses were done with SAS software (version 9.1; SAS, Cary, NC).

Sources of funding The GISSI-HF main study was supported by AstraZeneca, SPA, Pfizer, Sigma-Tau, and Pharmacia. The present substudy was planned, conducted, and analyzed without specific financial support by the GISSI group that has full ownership of the data, in complete independence from the sponsors. The authors are

American Heart Journal February 2011

340 Finzi et al

Table I. Baseline characteristics and treatment in the ICD substudy population and in the overall GISSI-HF study

Age (y), mean ± SD Age N70 y, n (%) Women, n (%) Smokers, n (%) History of hypertension, n (%) NYHA class, n (%) II III IV LVEF (%), mean ± SD LVEF N40%, n (%) Admission for HF in previous year, n (%) Previous AMI, n (%) Stroke, n (%) Diabetes mellitus, n (%) CABG, n (%) PCI, n (%) ICD, n (%) Pacemaker, n (%) History of atrial fibrillation, n (%) Peripheral vascular disease, n (%) Neoplasia, n (%) Medical treatment ACE inhibitors, n (%) ARBs, n (%) ACE inhibitors/ARBs, n (%) β-Blockers, n (%) Spironolactone, n (%) Diuretics, n (%) Digitalis, n (%) Oral anticoagulants, n (%) Aspirin, n (%) Other antiplatelet agents, n (%) Nitrates, n (%) Amiodarone, n (%) Calcium-channel blockers, n (%) Statin (open), n (%)

n-3PUFA (n = 278)

Placebo (n = 288)

P value⁎

Overall study (N = 6975)

64.9 ± 9.5 91 (32.7) 28 (10.1) 39 (14.0) 113 (40.7)

64.8 ± 9.8 86 (30.0) 38 (13.2) 50 (17.4) 143 (49.7)

.9149 .4611 .2472 .2763 .0314

67 ± 11 2947 (42.3) 1516 (21.7) 987 (14.1) 3809 (54.6)

176 (63.3) 101 (36.3) 1 (0.4) 28.1 ± 6.5 5 (1.8) 120 (43.2)

186 (64.6) 100 (34.7) 2 (0.7) 28.7 ± 6.9 5 (1.7) 111 (38.5)

.8033

.2258 .9550 .2632

4425 (63.4) 2365 (33.9) 187 (2.7) 33 ± 8.5 653 (9.4) 3384 (48.5)

161 (57.9) 13 (4.7) 65 (23.4) 74 (26.6) 58 (20.9) 129 (46.4) 62 (22.3) 45 (16.2) 17 (6.1) 15 (5.4)

162 (56.2) 22 (7.6) 77 (26.7) 82 (28.5) 63 (21.9) 129 (44.8) 68 (23.6) 44 (15.3) 30 (10.4) 8 (2.8)

.6893 .1435 .3574 .6218 .7691 .7004 .7113 .7664 .0637 .1148

2909 (41.7) 346 (5.0) 1974 (28.3) 1271 (18.2) 866 (12.4) 497 (7.1) 892 (12.8) 1325 (19.0) 610 (8.8) 265 (3.7)

209 (75.2) 64 (23.0) 264 (94.7) 214 (77.0) 125 (45.0) 254 (91.4) 97 (34.9) 110 (39.6) 130 (46.8) 24 (8.6) 85 (30.6) 106 (38.1) 12 (4.3) 95 (34.17)

223 (77.4) 55 (19.1) 273 (94.8) 223 (77.4) 136 (47.2) 270 (93.8) 90 (31.3) 102 (35.4) 136 (47.2) 43 (14.9) 96 (33.3) 98 (34.0) 17 (6.0) 83 (28.8)

.5289 .2520 .9259 .8980 .5900 .2795 .3571 .3076 .9128 .0204 .4819 .3096 .3922 .1703

5374 (77.1) 1321 (18.9) 6520 (93.5) 4522 (64.8) 2740 (39.3) 6260 (89.8) 2588 (37.1) 2009 (28.8) 3358 (48.1) 716 (10.3) 2472 (35.4) 1358 (19.5) 709 (10.2) 1579 (22.6)

NYHA, New York Heart Association; AMI, acute myocardial infarction; CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker. ⁎ Comparisons are between n-3PUFA and placebo groups only in the ICD substudy.

solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.

Results Patient population Of the 566 patients with at least one recorded follow-up visit and therefore analyzable within the GISSI-HF ICD study, 278 were assigned to receive n-3PUFA and 288 to placebo. Follow-up was concluded on December 31, 2007. The median follow-up duration was 928 days (quartiles 1-3 521-1287 days). Table I shows the baseline demographic characteristics including medical treatment of the n-3PUFA and placebo arms of the ICD study. Whereas no differences were observed between patients in the treatment and placebo arms of the ICD patient

population, with the exception of history of hypertension and antiplatelet use, this subgroup was younger than the main study population and more severely ill. At randomization, 233 patients had already an ICD (41%); and it was implanted during the follow-up in the other 333 (172 in placebo and 161 in n-3PUFA group). Indications for ICD implant were secondary prevention in 27.5%, primary prevention in 56.6%, and syncope of suspected arrhythmic etiology in 15.9% (Table II). Primary prevention increasingly became the main indication for implant during the study. Similarly, an increasing number of biventricular devices with ICD backup were implanted in 219 patients (39.0%). Baseline characteristics as well as the effect of n-3PUFA on main outcome were similar for patients who had a first ICD implant and those with an ICD already present at randomization.

American Heart Journal Volume 161, Number 2

Finzi et al 341

Table II. Indications for ICD implant by time of implant Postrandomization implant

n-3PUFA (161)

Placebo (172)

Syncope, n (%) Primary prevention, n (%) Secondary prevention, n (%) CRT, n (%)

21 (13.6) 102 (65.8) 32 (20.7) 68 (42.2)

21 (12.6) 118 (70.7) 28 (16.8) 78 (46.2)

.47

Prerandomization implant

n-3PUFA (117)

Placebo (116)

P value

Syncope, n (%) Primary prevention, n (%) Secondary prevention, n (%) CRT, n (%)

27 40 47 32

18 51 44 41

(23.7) (35.1) (41.2) (27.6)

(15.9) (45.1) (38.9) (35.3)

P value .61

.20

.20

ICD interventions During follow-up, 174 (30.7%) patients had at least one episode of appropriate ICD intervention. As shown in the Kaplan-Meier curves, the incidence of ICD interventions in the n-3PUFA group was somewhat lower than that in the placebo group, respectively (76 [27.3%] and 98 patients [34.0%] in the placebo group; Wilcoxon test P = .146) (Figure 1). Time to first ICDtreated arrhythmic event showed a tendency, not statistically significant, toward lower risk of ICD discharge in patients treated with n-3PUFA (unadjusted hazard rate [HR] 0.82, 95% CI 0.61-1.11, P = .210; adjusted HR 0.80, 95% CI 0.59-1.09, P = .152). A total of 1693 ICD-treated episodes of VT/VF were recorded during the follow-up: 1363 (80.5%) VT, 316 (18.7%) VF, and 14 (0.8%) not defined. ICD-driven therapy for VT episodes was recorded in 143 patients, 61 (22.0%) in the treatment group and 82 (28.5%) in the placebo group (unadjusted HR 0.80, 95% CI 0.57-1.12, P = .184). Patients who received 1, 2 to 3, or N3 ICD discharges amounted to 8.9%, 7.1%, and 11.1%, respectively, in n-3PUFA, as compared with 11.1%, 10.7%, and 12.1% in the placebo group (overall P = .30). ICD–driven therapy for VF was recorded in 71 patients (12.5%), 34 (12.2%) in the treatment group and 37 (12.9%) in the placebo group (P = .825). When a post hoc analysis was done on the effects of n-3PUFA by type of device implanted, treatment significantly reduced VT/VF only in patients without CRT (unadjusted HR 0.68, 95% CI 0.48-0.98, P = .0372), although the interaction between treatment and type of device was not significant (P = .32). Secondary clinical end points In this relatively small subgroup of patients, total mortality was 26.6% in the n-3PUFA group and 24.3% in the placebo group (adjusted HR 1.25, 95% CI 0.89-1.75, P = .19). Among the other end points, the rate of CV hospitalization tended to be lower in the treatment group

(63.0%) than the placebo group (71.5%) (adjusted HR 0.87, 95% CI 0.71-1.07, P = .18), in line with the results of the main study, although not statistically significant (Table III).

n-3PUFA plasma concentrations Changes in n-3PUFA plasma levels from baseline to 3 months were obtained for 159 patients, 77 in the treatment group and 82 in the placebo group. Mean ± SD n-3PUFA plasma levels expressed as the percentage of plasma phospholipids at baseline were 4.98 ± 1.8 in the n3PUFA group and 5.23 ± 1.73 in the placebo group. Mean changes expressed as least square means were 2.23 for the treatment group and −0.38 for the placebo group (P b .0001). The rate of appropriate ICD interventions tended to be lower in patients with larger 3-month increases (ie, N0.66, median 3-month change) in plasma n-3PUFA (26% above the median change vs 32% below, P = .405).

Discussion In this study, we observed a nonsignificant 20% reduction in the number of patients with appropriate ICD interventions for VT/VF in the n-3PUFA compared with the placebo group. This was consistent with (a) the smaller number of patients with N1 ICD discharge, (b) the smaller number of ICD-treated episodes, and (c) the smaller number of arrhythmic events in subjects with the highest plasma levels of n-3PUFA. However, results indicate a minimal excess in total mortality in treated patients (26.6% vs 24.3%), opposite to the results of the main trial; on the other side, a 13% reduction in CV hospitalization was observed. These discrepancies may well be due to the small size of this higher-risk subpopulation of 566 patients as compared with the almost 7,000 patients in the main trial. These findings are in line with experimental,14 epidemiological, and clinical studies showing a protective effect of n-3PUFA on arrhythmias and sudden cardiac death in the general population and in patients recovering from AMI.15-18 In GISSI-HF, survival curves for total mortality started to diverge after 2 years of treatment because of the lack of effect of n-3PUFA on most of the competing causes of death in HF patients, whereas the onset of benefit of n-3PUFA on fatal and nonfatal ventricular arrhythmias was early in post-AMI patients.4 Accordingly, the level of n-3PUFA in plasma and blood membranes rises steeply after the start of therapy13; and after 3 months, the plasma concentrations of n-3PUFA had risen N45% in GISSI-HF. To date, this is the largest study assessing the effects of n-3PUFA in ICD patients, though with mixed indications that may have confounded the proper assessment of n3PUFA antiarrhythmic effect. Accordingly, even this relatively large study was underpowered to assess the potential benefit. A posteriori, the statistical power to detect a 20% reduction of appropriate ICD discharge was 40% in the present trial, whereas 2,000 patients are

American Heart Journal February 2011

342 Finzi et al

Figure 1

Kaplan-Meier curves for the time to first appropriate ICD intervention (VT or VF) by randomized treatment. ⁎For ICD implanted during the study, the start of follow-up date is the date of ICD implant, whereas for ICD implanted before entering GISSI-HF, the starting time corresponds to randomization to the main trial.

Table III. Secondary end points and treatment Secondary endpoints Mortality Mortality for worsening of HF Mortality for arrhythmias Hospitalization for any cause Hospitalization for CV causes Hospitalization for worsening of HF Hospitalization for arrhythmias Hospitalization for other CV causes

n-3PUFA (278) 74 42 10 193 175 110 48 107

(26.6%) (15.1%) (3.6%) (69.4%) (63.0%) (39.6%) (17.3%) (38.5%)

Placebo (288) 70 45 6 218 206 127 46 127

Unadjusted HR (95% CI)

(24.3%) (15.6%) (2.1%) (75.7%) (71.5%) (44.1%) (16.0%) (44.1%)

1.18 1.04 1.84 0.92 0.88 0.95 1.12 0.88

(0.85-1.64) (0.68-1.59) (0.67-5.05) (0.76-1.11) (0.72-1.08) (0.74-1.23) (0.75-1.68) (0.68-1.14)

P value .32 .85 .24 .38 .21 .71 .58 .32

Adjusted HR (95% CI) 1.25 1.15 1.81 0.92 0.87 0.96 0.98 0.87

(0.89-1.75) (0.75-1.78) (0.63-5.15) (0.75-1.13) (0.71-1.07) (0.74-1.24) (0.65-1.48) (0.67-.13)

P value .19 .51 .27 .43 .18 .72 .92 .30

Hazard ratios were calculated by using Cox univariate and multivariable models. Data are for patients who had at least one event. All analyses were based on the time to the first event. Hospitalizations for ICD implant/maintenance and those that occurred before the ICD implant are not included.

needed to reach at least a 75% power in case a prospective randomized trial were to be planned. So far, 3 clinical trials have been conducted in different patient populations and using different dosages of n3PUFA,7-9 with contrasting results when compared.10,11 The clinical characteristics of GISSI-HF patients are different from those of these studies. First, patients were included on the basis of the diagnosis of HF with or without a history of malignant ventricular arrhythmias. Consequently, a substantial proportion of patients had an ICD implanted for primary prevention, as recommended by recent scientific statements and guidelines; and the proportion rose rapidly during the trial. In contrast, all patients in previous trials received an ICD as secondary prevention.7-9 Primary prevention has a lower risk profile

than secondary prevention after a major ventricular arrhythmia and/or aborted sudden death. Patients with an indication for CRT, 39.5% of all implants, have a higher arrhythmic risk profile than those in a better functional status who do not need this treatment and only receive a simple ICD device for primary prevention. In a post hoc analysis on our study, treatment with n-3PUFA significantly reduced time to first appropriate ICD discharge only in patients without CRT. Although the interaction is not statistically significant, the finding may suggest that the more advanced disease in patients with CRT and its proarrhythmic potential outweighed the possible protective effects of n-3PUFA. Ischemic etiology of HF accounted only for 61.8% of the patients as compared with 76%; mean LVEF was

American Heart Journal Volume 161, Number 2

28.4% ± 6.7%, whereas in the other 3 studies,7-9 it was around 35%. This could indicate that, parallel to a more compromised functional status, our patient population had more severe ventricular arrhythmias. Overall, the severity of HF, as expected, was higher in implanted patients than in the whole GISSI-HF population (total mortality of 24.3% in the placebo- and 26.6% in the n3PUFA–treated groups in 2 years compared with 6% to 9.4% in the other studies).7-9 The dose of n-3PUFA in GISSI-HF was 850 mg/d, as compared with 900 to 2600 mg/d in the 3 previous studies.7-9 Although no clear dose-effect relationship has been demonstrated, it cannot be excluded that a higher n-3PUFA intake might be more beneficial. This study has several limitations. It is a substudy of the large GISSI-HF trial with different primary end points, whose criteria for eligibility did not include reasons for ICD implant. However, ICD events were prospectively collected and adjudicated by an independent committee12 of experts blind to study treatments. The trial was conducted in 89 hospitals in the Italian National Health system, ensuring the generalizability of its findings, although the insufficient power of the study does not allow definite conclusions. As during the comparatively long period of the study there were considerable shifts in cardiologists' attitudes in favor of the use of ICD for primary prevention, as supported by data in Table II, and a widespread indication for CRT with ICD back-up protection, enrollment in this study of a majority of patients at lower-than-expected risk profile was unavoidable. This could have negatively affected the possibility of demonstrating any significant effect of n-3PUFA. Unbalance between study groups in hypertension and antiplatelet use, though unrelated to study outcome, could have affected the results.

Conclusions Encouraging data, in terms of reduction of sudden cardiac death, have been shown in post-AMI and HF patients treated with n-3PUFA. In GISSI-HF, n-3PUFA decreased nonsignificantly the number of ICD interventions for VT/VF as well as other arrhythmic events, in the absence of a reduction in “painful shocks” and mortality. The present study supports the trend (not statistically significant) toward prevention of life-threatening ventricular arrhythmias found in animal models and in epidemiological and comparative studies.

Acknowledgements GISSI is endorsed by Associazione Nazionale Medici Cardiologi Ospedalieri (ANMCO), Firenze, Italy; by Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy; and by Consorzio Mario Negri Sud, Santa Maria Imbaro, Italy. We are grateful to Valter Torri for his statistical advice and to Francesca Perego for secretarial assistance.

Finzi et al 343

References 1. European Society of Cardiology. ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Eur Heart J 2008;29:2388-442. 2. Cleland JG, Daubert JC, Erdmann E, et al. The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med 2005;352:1539-49. 3. London B, Albert C, Anderson ME, et al. Omega-3 fatty acids and cardiac arrhythmias: prior studies and recommendations for future research: a report from the National Heart, Lung, and Blood Institute and Office of Dietary Supplements Omega-3 Fatty Acids and Their Role in Cardiac Arrhythmogenesis Workshop. Circulation 2007;116:e320-35. 4. Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto miocardico. Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: results of the GISSI Prevenzione Trial. Lancet 1999;354:447-55. 5. Gruppo Italiano per lo Studio della Sopravvivenza nell'Insufficienza cardiaca. GISSI-HF. Effects of n-3 polyunsaturated fatty acids in patients with chronic heart failure (the GISSI-HF trial): a randomized double-blind, placebo-controlled trial. Lancet 2008;372:1223-30. 6. Marchioli R, Aldegheri MP, Borghese L, et al. Time course analysis of the effect of n-3PUFA on fatal and non fatal arrhythmias in heart failure: secondary results of the GISSI-HF trial. Eur Heart J 2009;30 (Abstract Supplement):165. 7. Leaf A, Albert CM, Josephson M, et al. Prevention of fatal arrhythmias in high-risk subjects by fish oil n-3 fatty acid intake. Circulation 2005; 112:2762-8. 8. Raitt MH, Connor WE, Morris C, et al. Fish oil supplementation and risk of ventricular fibrillation in patients with implantable defibrillators: a randomized controlled trial. JAMA 2005;293:2884-91. 9. Brouwer IA, Zock PL, Camm J, et al. Effect of fish oil on ventricular tachyarrhythmia and death in patients with implantable cardioverter defibrillators: the Study on Omega-3 Fatty Acids and Ventricular Arrhythmia (SOFA) randomized trial. JAMA 2006;295:2613-9. 10. Brouwer IA, Raitt MH, Dullemeijer C, et al. Effect of fish oil on ventricular tachyarrhythmia in three studies in patients with implantable cardioverter defibrillators. Eur Heart J 2009;30:820-6. 11. Leon H, Shibata MC, Sivakumaran S, et al. Effect of fish oil on arrhythmias and mortality: systematic review. BMJ 2008;a2931:337. 12. Tavazzi L, Tognoni G, Franzosi MG, et al. Rationale and design of the GISSI heart failure trial: a large trial to assess the effects of n-3 polyunsaturated fatty acids and rosuvastatin in symptomatic congestive heart failure. Eur J Heart Fail 2004;6:635-41. 13. Di Stasi D, Bernasconi R, Marchioli R, et al. Early modifications of fatty acid composition in plasma phospholipids, platelets and mononucleates of healthy volunteers after low doses of n-3 polyunsaturated fatty acids. Eur J Clin Pharmacol 2004;60:183-90. 14. Leaf A, Kang JX, Xiao YF, et al. Clinical prevention of sudden cardiac death by n-3 polyunsaturated fatty acids and mechanism of prevention of arrhythmias by n-3 fish oils. Circulation 2003;107:2646-52. 15. Siscovick DS, Raghunathan TE, King I, et al. Dietary intake and cell membrane levels of long-chain n-3 polyunsaturated fatty acids and the risk of primary cardiac arrest. JAMA 1995;274:1363-7. 16. Kromhout D, Bosschieter EB, de-Lezenne-Coulander C. The inverse relation between fish consumption and 20-year mortality from coronary heart disease. N Engl J Med 1985;312:1205-9. 17. Albert CM, Campos H, Stampfer MJ, et al. Blood levels of long-chain n-3 fatty acids and the riskof sudden death.N EnglJ Med 2002;346:1113-8. 18. Burr ML, Fehily AM, Gilbert JF, et al. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: Diet and Reinfarction Trial (DART). Lancet 1989;2:757-61.

American Heart Journal Volume 161, Number 2

Appendix. Investigators and centers participating to the ICD substudy of GISSI-HF Switzerland Lugano (T. Moccetti, E. Pasotti, M. G. Rossi). Italy Borgomanero (A. M. Paino, U. Parravicini), Orbassano (L. Montagna, A. Previti, G. P. Varalda), Torino, Evangelico Valdese (L. Bo, N. Massobrio), Torino, Maria Vittoria (M. Imazio, R. Trinchero), Veruno (U. Corrà, A. Mezzani, P. Giannuzzi). Aosta (G. Begliuomini). Bergamo (A. Gavazzi, M. Gori), Brescia, Spedali Civili Cardiologia (L. Dei Cas, S. Nodari), Cernusco sul Naviglio (S. Dell'Orto, M. Sfolcini), Como, Valduce (E. Miglierina, M. Santarone, L. Sormani), Como, S. Anna (R. Jemoli, F. Tettamanti), Giussano (K. N. Jones, A. Volpi), Milano, Monzino (P. Agostoni, P. Palermo), Milano, San Carlo Borromeo (L. Squadroni), Montescano (J. Baccheschi, O. C. Febo, F. Olmetti), Monza, San Gerardo (V. Antonazzo, E. Piazzi, A. Vincenzi), Passirana-Rho (A. Frisinghelli, M. Palvarini, M. D. Veniani), Pavia, San Matteo (N. Ajmone Marsan, G. Piccoli, L. Scelsi), Pavia, Salvatore Maugeri (A. Gualco, C. Opasich), Pieve di Coriano (M. A. Iannone), Sondalo, E. Morelli, Cardiologia Riabilitativa (N. Partesana), Tradate (R. Raimondo), Varese, Del Ponte (I. Ghezzi), Bovolone (S. Boni, A. Pasini), Camposampiero (A. A. Zampiero), Este (F. Caneve), Mirano (S. Milan, P. Sarto), Negrar (E. Barbieri, P. Girardi), Portogruaro (D. Milan), Verona (M. Cicoira, L. Zanolla), Gorizia (R. Chiozza, G. Giuliano), San Daniele del Friuli (L. Mos, O. Vriz), Tolmezzo (E. Alberti, M. Werren), Trieste (F. Longaro, G. Sinagra), Udine (M. C. Albanese, P. De Biaggio, D. Miani), GenovaSestri Ponente (D. Caruso), Sarzana-Loc. S. Caterina

Finzi et al 343.e1

(A. Cantarelli), Bentivoglio (R. Vandelli), Cotignola (A. Barbieri), Rimini (F. Bologna, G. Piovaccari), Bagno a Ripoli (M. Nannini), Cecina (F. Mazzinghi), Cortona (D. Cosmi), Empoli (A. Cecchi, A. Zipoli), Firenze, Nuovo S Giovanni di Dio (C. Minneci, G. M. Santoro), Firenze, S. M. Nuova (M. Milli), Fucecchio (A. A. Brandinelli Geri), Livorno (M. Carluccio, E. Magagnini), Massa, S. S. Giacomo e Cristoforo (V. Molendi), Poggibonsi (G. Scopelliti), Città di Castello (G. M. Arcuri, D Severini), Todi (R. Panciarola), Ascoli Piceno (G. Gregori, L. Moretti, L. Partemi), Macerata (G. L. Morgagni), Albano Laziale (P. Midi, G. Pajes), Cassino (V. Di Spirito), Civitavecchia (S. Calcagno), Roma, Santo Spirito (N. Aspromonte, R. Ricci), Roma, Gemelli (C. Ierardi), Velletri (D. Banda, F. Frasca), Teramo (L. L. Piccioni), Castellammare di Stabia (G. De Caro), Napoli, Monaldi, I Div Medica (O. Maiolica, R. Santoro), Napoli, Federico II (M. Chiariello, P. Perrone Filardi), Oliveto Citra (P. Bottiglieri), Piedimonte Matese, Ave Gratia Plena (L. De Risi, A. Vetere), Salerno (P. Predotti, F. Silvestri), Bari, Policlinico (D. Traversa), Bari-Carbonara (C. M. Altamura, G. Scalera), Casarano (S. A. Ciricugno, M. Gualtieri, L. Manca), Fasano (F. Loliva), Putignano (G. Cellamare, D. De Laura), Taranto, Moscati (P. Palmisano, A. Peluso), Tricase (P. Palma), Potenza, A. O. San Carlo (M. Chiaffitelli, A. Lopizzo), Cosenza, S. S. Annunziata (G. Misuraca), Reggio Calabria (G. Cutrupi, G. Pulitanò, A. Ruggeri), Vibo Valentia (G. Maglia), Agrigento (R. Rametta), Caltagirone (A. Alì), Palermo, Civico e Benfratelli (L. Lo Presti, G. Stassi), Palermo, Villa Sofia (M. T. Cinà, V. Cirrincione, F. Ingrillì), Palermo, Cervello, Cardiologia I (R. V. Cappello, G. Geraci, F. Romano), Palermo, Cervello, Cardiologia II (F. Enia, A. M. Floresta), Cagliari, Marino (P. Siddi).

Coronary Artery Disease

Common oral mucosal diseases, systemic inflammation, and cardiovascular diseases in a large cross-sectional US survey Stefano Fedele, DDS, PhD, a Wael Sabbah, BDS, MSc, b Nikos Donos, DDS, MS, PhD, a Stephen Porter, BSc, MD, PhD, a and Francesco D'Aiuto, DMD, PhD a London, United Kingdom

Background Inflammation of the gingivae (periodontitis) has been associated with raised serum biomarkers of inflammation, sub-clinical markers of atherosclerosis, and increased risk of and/or mortality from cardiovascular disease (CVD). There remain little information regarding the association between other common oral inflammatory disease, systemic inflammation, and CVD. The objective of the study was to assess the association between common oral mucosal diseases, circulating markers of inflammation, and increased prevalence of CVD in a cross-sectional survey of a nationally representative sample of the noninstitutionalized civilians in the United States. Methods Data for this study are from 17,223 men and women aged ≥17 years who received oral examination as part of the Third National Health and Nutrition Examination Survey. The primary and secondary outcome measures were the association of oral mucosal diseases with raised serum levels of C-reactive protein/fibrinogen and increased prevalence of CVD, respectively. Adjustment for common confounding factors was performed. Results

Having oral mucosal disease was associated with systemic inflammation (serum levels of C-reactive protein ≥10 mg/dL) (odds ratio 1.41, 95% CI 1.02-1.94). Individuals with oral mucosal disease were 1.36 times (95% CI 1.02-1.80) more likely to have history of myocardial infarction and 1.33 times (95% CI 1.03-1.71) more likely to report angina than unaffected individuals. All associations were independent of common confounding factors.

Conclusions This is the first study to suggest that common oral mucosal diseases are independently associated with raised markers of systemic inflammation and history of CVD. (Am Heart J 2011;161:344-50.)

Cardiovascular disease (CVD) remains a leading cause of death in the United States and worldwide.1 Primary prevention of CVD currently involves targeting interventions to those individuals at high absolute risk, identified using risk-prediction instruments such as the Framingham equation that integrate information on established risk factors such as hypertension, dyslipidemia, smoking, and diabetes.2 Yet these factors do not explain all of the excess risk because a proportion of cardiovascular events occur among individuals with no or near-average levels of traditional risk factors.3,4 Approximately 40% of coronary heart disease deaths occur in persons with cholesterol levels that are lower

From the aUCL Eastman Dental Institute, Oral Medicine, University College London, London, United Kingdom, and bUCL Department of Epidemiology and Public Health, University College London, London, United Kingdom. Submitted January 25, 2010; accepted November 7, 2010. Reprint requests: Stefano Fedele, DDS, PhD, Eastman Dental Institute, University College London, 256 Gray's Inn Rd, London WC1X 8LD, United Kingdom. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.009

than the population average,4 and there remains a large part of the population who is classified as intermediate risk via current criteria.5,6 There is an urgent need for new or emerging factors that could account for some of the unexplained variability in CVD risk and identify those individuals in this group who are actually at high risk and may benefit from more aggressive risk reduction strategies.5,6 Inflammation is frequently discussed as a potential major mechanistic contributor to atherothrombosis, and measurement of inflammatory markers could have the potential of improving risk stratification beyond current global risk assessment.3-8 Indeed, inflammation contributes to all stages in the pathogenesis of atherogenesis from plaque formation, the acute atherothrombotic event, and the myocardial damage following ischemia.7-9 Lowgrade systemic inflammation as assessed by raised serum biomarkers such as C-reactive protein (CRP) has been linked to future risk of coronary events, subclinical measures of atherosclerosis, and stroke,3-9 although there remains controversy regarding the potential improvement in risk stratification or reclassification from addition of CRP to current models.3

American Heart Journal Volume 161, Number 2

Recent evidence supports the notion that extravascular chronic inflammation might also contribute to individuals' CVD risk.9 Examples include chronic bacteria-induced inflammation of the periodontal tissues (periodontitis)10-12 and chronic immunomediated inflammation of the skin (psoriasis)13-16 and joints (arthritis).17-19 These disorders have been associated with systemic inflammation, endothelial dysfunction, and subclinical atherosclerosis, which have all been associated with increased risk/prevalence of and/or mortality from CVD.9-19 Further acute inflammatory episodes from intercurrent infection in the respiratory (eg, influenza) or urinary tracts have also been associated with increased odds of CVD events.20,21 Nevertheless, the evidence remains controversial, as these associations may have noncausal explanations but merely be a result of reverse causality or residual confounding. Periodontitis has received great attention because of its high prevalence worldwide (10%-30% of the adult population)22 and because it is easier to control (via mechanical cleaning of the diseased dentition) than most other risk factors. Indeed, interventional studies have suggested that treatment of periodontitis improves endothelial function and reduces systemic inflammation.9,23 The oral cavity, however, can be affected by a wide range of common disorders other than periodontitis that are characterized by recurrent or chronic local inflammation of the oral mucosa and submucosal stroma. The inflammatory component of these disorders can be primary (autoimmune) or secondary to infections (fungal, viral, or bacterial). Recurrent aphthous stomatitis and recurrent secondary herpes simplex virus–1 infection (herpes labialis), for example, can affect up to 40% of the general population24,25; oropharyngeal candidiasis is by far the most common oral fungal infection in men and can affect up to 90% of individuals in certain disease groups (eg, HIV infected)26; and oral lichen planus has been reported to affect 1% to 2% of the general population, this figure being very similar to psoriasis prevalence.27 Currently, limited evidence exists on the possible association of these oral mucosal disorders with a state of systemic inflammation and CVD. The primary aim of the present study was to investigate whether non-periodontal oral mucosal inflammatory diseases are independently associated with increased circulating markers of inflammation in a large representative sample of the US population, based on the 19881994 Third National Health and Nutrition Examination Survey (NHANES III).28 Secondary aims included association of oral mucosal diseases with increased prevalence of history of CVD.

Methods Data for this study are from the NHANES III conducted in 1988-1994.28 NHANES used a stratified multistage probability

Fedele et al 345

sampling design representative of the noninstitutionalized civilian American population and included a comprehensive medical and oral examination, laboratory investigations, and questionnaires on demographic, socioeconomic, self-reported diagnoses of medical conditions, and health-related behaviors data. Oral examination consisted of assessment of dental and periodontal status and also included clinical diagnosis of disorders of the oral mucosa.29 Data on the adult population aged ≥17 years were used.

Oral mucosal diseases Four variables for oral mucosal disease were created and used in the analysis, namely, (1) lichen planus and 3 aggregate variables indicating the presence of (2) any infectious, (3) any inflammatory, and (4) any infectious or inflammatory oral mucosal disease. Categorization of oral mucosal disorders recorded in NHANES into the study aggregates was based on pathogenetic features (inflammation secondary to infection vs inflammation not related to infectious agents). The infectious disorders included denture stomatitis (5.9%), denture hyperplasia (1.1%), angular cheilitis (0.7%), denture inflammation (0.6%), median rhomboid glossitis (0.2%), candidiasis (0.1%), erythematous candidiasis (0.01%), herpes labialis (1.6%), herpetic gingivostomatitis (0.01%), papilloma/wart (0.5%), and acute necrotizing gingivitis (0.02%). Inflammatory disorders included lichen planus (0.1%), recurrent aphthous ulceration (0.9%), nonspecific ulcers (0.1%), and denturerelated ulcers (0.7%). Lichen planus was also examined as a separate independent variable because of its similarities with psoriasis, which was already reported to be an independent risk factor for CVD.13-16 Details on the diagnosis and definition of oral mucosal disorders in NHANES were previously reported.29

Outcome variables (serum inflammatory markers and history of CVD) Inflammatory measures included 2 binary variables of serum levels of CRP (≥3 and ≥10 mg/L) and 1 for fibrinogen (≥3.25 g/dL) as markers of acute and low-grade inflammation.30,31 Cardiovascular disease prevalence was expressed as self-reported diagnoses of myocardial infarction, congestive heart failure, stroke, clinically measured high blood pressure (systolic N130 mm Hg or diastolic N85 mm Hg), and angina pectoris based on the World Health Organization Rose questionnaire.32 Details on laboratory investigations and diagnosis and definition of CVD in NHANES were previously reported.29

Data analysis (a): characteristics of study population First, we examined the distribution of the 8 indicators of systemic inflammation and CVD (outcome variables) and the 4 indicators of oral mucosal diseases in the database. We also estimated the means of age, body mass index, systolic and diastolic blood pressure, serum CRP, fibrinogen, cholesterol, triglycerides, low-density lipoprotein cholesterol and highdensity lipoprotein cholesterol, sex, ethnicity, education (years), poverty-income ratio, periodontitis, and smoking status (current smokers within groups with and without oral mucosal diseases) to characterize the study population.

American Heart Journal February 2011

346 Fedele et al

Data analysis (b): associations and logistic regression models A series of logistic regression models were conducted to test the relationship between each outcome variable (CRP ≥3 mg/L, CRP ≥10 mg/L, fibrinogen ≥3.25 g/dL, myocardial infarction, congestive heart failure, stroke, high blood pressure, angina) and each of the 4 indicators of oral mucosal diseases (lichen planus; any infectious, any inflammatory, or any infectious and inflammatory mucosal disorder). After the first crude model of association, a second model was adjusted for sex, age, and ethnicity (white, African American, Mexican American, and other ethnicities); and a third and final model was additionally adjusted for the following variables: education (years), povertyincome ratio, body mass index, reported diagnosis of diabetes, triglycerides, cholesterol, and smoking (current smokers, noncurrent smokers, and nonrespondent) and periodontal status (severe, moderate, mild, or no periodontitis). Detailed description and justification of the variables used in this analysis were provided in previous articles.33-35 The analysis was conducted using survey command in Stata (StataCorp, College Station, TX) and using final examination sampling weights. Data reported in results and tables refer to the third final logistic regression model. No dedicated extramural funding was used to support this work. This work was undertaken at University College London/ University College London Hospital, which received a proportion of funding from the Department of Health's National Institute for Health Research Biomedical Research Centre funding scheme. F. D. holds a Clinical Senior Lectureship Award supported by the United Kingdom Clinical Research Collaboration. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.

Results Characteristics of study population Data on oral mucosal examination relevant to 17,235 people were identified in the NHANES database. Among these, the initial descriptive analysis was conducted for adults aged ≥17 years for whom data on oral mucosal examination and the main explanatory variables were available (Table I). Only 0.1% of them had lichen planus, 9.2% had any infectious oral mucosal disease, and 1.7% had any inflammatory oral mucosal disease, for a total of 1,684 individuals (10.63%). Analysis of the distribution of the outcomes variables of CVD and systemic inflammation showed that fibrinogen N3.25 g/L and high blood pressure had the highest prevalence with 33.2% of the sample affected, whereas stroke (1.7%) and congestive heart failure (1.8%) had the lowest prevalence (Table I). Further details of the study group are provided in Table II. In brief, individuals with oral mucosal disease were generally older, had higher systolic and diastolic blood pressure, and had higher levels of triglycerides, cholesterol, fibrinogen, and CRP compared with those with no oral mucosal disorders (Table II).

Table I. Distribution of the 8 indicators of systemic inflammation and CVD and the 4 indicators of oral mucosal diseases in the database (weighted) Condition (n) Lichen planus (15485) Infectious mucosal diseases (15485) Inflammatory mucosal diseases (15485) Any infectious or inflammatory mucosal disease (15485) CRP ≥3 mg/L (14641) CRP ≥10 mg/L (14641) Fibrinogen ≥3.25 g/L (7878) Myocardial infarction⁎ (15270) Congestive heart failure⁎ (15465) Stroke⁎ (15480) High blood pressure† (15453) Angina‡ (15273)

Observation

Percentage (95% CI)

15 1451

0.10% (0.05-0.21) 9.16% (8.40-9.97)

280

1.72% (1.42-2.09)

1684

10.63% (9.78-11.54)

4864 1357 3007 653

27.63% 7.27% 32.00% 3.15%

(25.27-30.12) (6.49-8.14) (29.43-34.68) (2.71-3.67)

501

1.83% (1.62-2.08)

406 6305 1279

1.71% (1.41-2.06) 33.21% (31.37-35.12) 6.65% (5.99-7.38)

⁎ Self-reported physician diagnosis of the respective condition. † High blood pressure: systolic ≥130 mm Hg or diastolic ≥85 mm Hg. ‡ Angina according to Rose questionnaire or physician-diagnosed myocardial infarction.

Infectious oral mucosal diseases, inflammation, and CVD Table III shows the relationship between systemic inflammation and prevalence of CVD and having any infectious oral mucosal disease. Generally, individuals with any infectious oral mucosal disease had higher probabilities of having high levels of CRP and fibrinogen and any history of CVD in the unadjusted or partially adjusted models (data not shown); but the association became not statistically significant after adjusting for all confounders in the third and final logistic regression model (Table III). However, the association with angina remained statistically significant, with individuals with any infectious oral mucosa disease being 1.33 times more likely to have history of angina (95% CI 1.05-1.68) than controls (Table III). Inflammatory oral mucosal diseases, inflammation, and CVD The odds ratios for having CRP ≥10 mg/L for individuals with inflammatory mucosal disease were significant in the third and final adjusted model (Table III). Having any inflammatory oral mucosa disease indicated a higher probability of having myocardial infarction in the unadjusted model (data not shown), but the relationship was not statistically significant in the final adjusted model (Table III) Any oral mucosal disease, inflammation, and CVD Presence of any oral mucosal disease studied was associated with greater odds of high systemic inflammation

American Heart Journal Volume 161, Number 2

Fedele et al 347

Table II. Characteristics of the study group Individuals with oral mucosa disease Age (mean) Sex (%) Male Female Ethnicity (%) White American African American Mexican American Other ethnicities Periodontitis (%) Severe periodontitis Moderate periodontitis Mild or no periodontitis Current smokers (%) Yes No Mean poverty-income ratio Mean years of education Body mass index (mean) Systolic blood pressure (mean) Diastolic blood pressure (mean) Cholesterol, mg/dL (mean) Triglycerides, mg/dL (mean) LDL cholesterol, mg/dL (mean) HDL cholesterol, mg/dL (mean) CRP (mean) Fibrinogen (mean)

Individuals without oral mucosa disease

P value for difference between groups

53.06 (51.2-55.0)

41.53 (40.8-42.3)

b.001

10.5% (9.4-11.7) 10.8% (10.0-12.0)

89.5% (88.3-90.6) 89.2% (88.0-90.2)

NS

11.3% 8.1% 7.1% 10.2%

88.7% 91.9% 92.9% 89.8%

(10.3-12.5) (7.0-9.4) (6.3-8.0) (8.2-12.8)

10.8% (6.5-17.6) 13.9% (10.9-17.5) 7.4% (6.6-8.2) 12.0% 10.1% 2.8 11.4 27.1 128.2 74.9 211.7 162.3 133.8 49.1 4.71 316.15

(10.4-13.7) (9.4-10.9) (2.7-2.9) (11.2-11.7) (26.7-27.5) (126.4-130.0) (74.2-75.6) (208.0-215.5) (153.7-171.0) (128.3-139.2 (47.8-50.3) (4.11-5.32) (307.05-325.25)

(87.5-89.7) (90.7-93.0) (92.0-93.7) (87.2-91.8)

b.001

89.2% (82.4-93.5) 86.1% (82.5-89.1) 92.6% (91.9-93.4)

b.001

88.0% 89.9% 3.1 12.5 26.3 120.7 73.5 200.9 138.9 125.1 50.9 3.51 302.20

(86.3-90.0) (89.1-90.6) (3.0-3.2) (12.3-12.7) (26.0-26.5) (120.0-121.4) (73.1-73.9) (199.3-202.5) (134.5-142.8) (123.3-126.9) (50.2-51.5) (3.35-3.68) (296.97-307.43)

b.05 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.01 b.01 b.001 b.001

NS, Not significant; LDL, low-density lipoprotein; HDL, high-density lipoprotein.

and any history of CVD in the unadjusted model (data not shown). In the final adjusted model, these associations remained statistically significant only for CRP ≥10 mg/L, history of myocardial infarction, and angina (Table III).

Lichen planus, inflammation, and CVD High levels of CRP and fibrinogen were not associated with lichen planus (Table III). Lichen planus however was associated with congestive heart failure in the final fully adjusted model (odds ratio = 7.14, 1.37-37.26). The probabilities of having myocardial infarction, high blood pressure, and angina were also higher among individuals with lichen planus (data not shown); but the association became not statistically significant after adjusting for other confounders (Table III). The relationship between stroke and lichen planus could not be conducted because of a very small number of cases.

Discussion This analysis provides, for the first time, some evidence of a positive association between nonperiodontal oral inflammation (common oral mucosal diseases) and systemic inflammation and history of CVD.

In the 3 last decades, there has been increasing evidence that localized chronic infection/inflammation of the periodontal tissues (periodontitis) is associated with systemic inflammation, dyslipidemia, glucose intolerance, a procoagulant state, and endothelial dysfunction.9-12 Individuals with periodontitis present with an overall 20% increased risk of future CVD.36 Yet, the mouth can be affected by a wide range of infectious and immunomediated disorders other than periodontitis that are well known to cause localized acute, recurrent, and chronic inflammation of oral epithelium and underlying stroma. Interestingly, there remains no study that has investigated the association between these nonperiodontal oral diseases with systemic inflammation and CVD. The rationale behind this potential association is the biological plausible mechanism that links localized organspecific inflammation of nonvascular tissues (such as arthritis, psoriasis, and recurrent respiratory and urinary tract infections) to systemic inflammation and CVD.9-20 Indeed, some of the mucosal diseases studied in the present analysis have been preliminarily reported to be associated with markers of systemic inflammation and/ or CVD risk. Lichen planus and recurrent aphthous stomatitis have been associated with elevated serum levels of interleukin (IL)-637 and IL-838,39; as well, a

American Heart Journal February 2011

348 Fedele et al

Table III. Probabilities of having CVD and raised CRP/fibrinogen for persons diagnosed with any oral mucosal disease (N = 15270)—adjusted model Odds ratio (95% CI)

CRP ≥3 mg/L (14641) CRP ≥10 mg/L (14641) Fibrinogen ≥3.25 g/L (7878) Myocardial infarction (15270) Congestive heart failure (15465) Stroke (15480) High blood pressure (15453) Angina (15273)

Any infectious oral mucosal disease

Any inflammatory oral mucosal disease

Any oral mucosal disease

Oral lichen planus

1.09NS (0.87-1.36) 1.26NS (0.90-1.76) 1.05NS (0.87-1.28) 1.28NS (0.95-1.74)

1.15NS (0.79-1.66) 2.38⁎ (1.34-4.22) 1.05NS (0.67-1.65) 1.58NS (0.86-2.88)

1.10NS (0.88-1.37) 1.41† (1.02-1.94) 1.04NS (0.87-1.23) 1.36† (1.02-1.80)

0.76NS (0.26-2.23) 0.84NS (0.12-5.86) 0.36NS (0.11-1.23) 3.62NS (0.56-23.53)

1.01NS (0.70-1.47)

0.79NS (0.30-2.09)

1.02NS (0.69-1.49)

7.14† (1.37-37.26)

0.86NS (0.56-1.33) 0.88NS (0.71-1.08) 1.33† (1.05-1.68)

0.78NS (0.31-2.00) 0.86NS (0.54-1.35) 1.21NS (0.68-2.15)

0.86NS (0.57-1.30) 0.86NS (0.70-1.05) 1.33† (1.03-1.71)

– 1.74NS (0.29-10.27) 3.14NS (0.82-12.11)

⁎ P b .01. † P b .05.

Figure 1

A

B

C

D

Odd ratios (95% CIs) of CVD–CRP/fibrinogen and the 4 oral mucosal disease groups (A, any infectious mucosal disease; B, any inflammatory mucosal disease; C, any mucosal disease; D, oral lichen panus)—adjusted values.

reduction of their serum concentration has been observed after appropriate treatment.38,39 A series of viral infections, including herpes simplex infection, showed associated serum CRP and serum amyloid A elevation.40,41 Seropositivity status to Herpesviridae has been associated with the future risk of CVD and cardiovascular death,42,43 and oral candidiasis has also

been associated with elevated serum levels of IL-6 and tumour necrosis factor–α.44 The present study is the first to investigate in details a large cohort of individuals for the potential association between nonperiodontal oral inflammatory disease, presence of circulating inflammatory markers, and CVD (Figure 1).

American Heart Journal Volume 161, Number 2

Our findings provide evidence of a possible association between common oral mucosal disease and history of myocardial infarction/angina as well as systemic inflammation as defined by levels of CRP ≥10 mg/L. These associations were independent of common confounding factors including sex, age, ethnicity, body mass index, diabetes, cholesterol and triglycerides, periodontitis, smoking status, as well as income and education. Noteworthy is the finding that also inflammatory noninfectious oral mucosal diseases were associated with high systemic inflammation, although no statistically significant relationship with CVD was found in the final adjusted model. This supports the hypothesis that local inflammation, rather than infection, may cause systemic inflammation as reported in individuals affected by other autoimmune disorders such as rheumatoid arthritis and psoriasis.13-19 Of note, a recent study demonstrated that invasive dental procedures may be associated with a transient increase in the risk for stroke and myocardial infarction and thus provided further evidence to support the link between oral inflammation, systemic inflammation, and the risk for vascular events.45 Lichen planus was studied as an independent variable from the other oral mucosal disease because of similarities with psoriasis. We indeed observed a strong association between diagnosis of lichen planus and congestive heart failure, although the low number of cases reported in the survey suggests that results should be interpreted with caution. The mechanisms behind the association are unclear because there was no evident correlation with increased systemic inflammation measured via CRP and fibrinogen levels. It could be possible that different pathways are involved or, as observed in psoriasis,14 that the severity and extension of the disease could play a major role in determining a detectable increase of inflammatory serum markers. These details were unfortunately not available in NHANES III. Our analysis has a number of limitations worth mentioning. The cross-sectional nature of the NHANES study does not allow for causal interpretations (no information as to the direction of a potential causal pathway), and the high number of outcomes analyzed in the statistical models increases the chance of type I error. It should also be noted that the criteria used to diagnose oral mucosal diseases in the NHANES study were based solely on clinical examination, whereas further investigations (eg, detailed medical history, histopathology) are required to define some oral mucosal disorders (eg, lichen planus). Indeed, individuals with congestive heart failure are often managed with medications known to potentially cause mucocutaneous lesions that resemble lichen planus (eg, lichenoid drug reactions).46 It is noteworthy to mention that the number of individuals presenting with oral mucosal diseases was relatively small; and therefore, results could not be extrapolated to

Fedele et al 349

other populations and should therefore be replicated. Nevertheless, the elaborate sampling and weights characteristics of NHANES would still mean that people with oral mucosal diseases included in the analysis refer to 10.6% of US individuals. For example, the presence of 0.1% of individuals with lichen planus in the survey would correspond, by virtue of its complex design, to 250,000 individuals according to the US population in 1990. In addition, several factors (eg, genetics, alcohol consumption, exercise) that could play a role have not been assessed; and we cannot therefore exclude the possibility of a spurious association due to unmeasured confounding factors potentially linked per se to increased systemic inflammation and history of CVD. Nonetheless, efforts were made to minimize the residual confounding. The present study reports for the first time, with the limitations of a cross-sectional survey-based analysis, a possible association between oral diseases other than periodontitis, systemic inflammation, and CVD. Because some of the oral mucosa disorders considered in this analysis are commonly seen in the population, these findings might have important implications for the health of the public. Further research is needed to confirm these preliminary results.

References 1. World Health Organization. World health statistics 2009. http:// www.who.int/whosis/whostat/EN_WHS09_Full.pdf. Last accessed November 24, 2009. 2. Grundy SM, Cleeman JI, Daniels SR, et al. Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute scientific statement. Curr Opin Cardiol 2006;21:1-6. 3. Shah T, Casas JP, Cooper JA, et al. Critical appraisal of CRP measurement for the prediction of coronary heart disease events: new data and systematic review of 31 prospective cohorts. Int J Epidemiol 2009;38:217-31. 4. Buckley DI, Fu R, Freeman M, et al. C-reactive protein as a risk factor for coronary heart disease: a systematic review and meta-analyses for the U.S. Preventive Services Task Force. Ann Intern Med 2009; 151:483-95. 5. Smith Jr SC. Current and future directions of cardiovascular risk prediction. Am J Cardiol 2006;97:28A-32A. 6. Helfand M, Buckley DI, Freeman M, et al. Emerging risk factors for coronary heart disease: a summary of systematic reviews conducted for the U.S. Preventive Services Task Force. Ann Intern Med 2009; 151:496-507. 7. Casas JP, Shah T, Hingorani AD, et al. C-reactive protein and coronary heart disease: a critical review. J Intern Med 2008;264: 295-314. 8. Hingorani AD, Shah T, Casas JP, et al. C-reactive protein and coronary heart disease: predictive test or therapeutic target? Clin Chem 2009;55:239-55. 9. Hingorani AD, D'Aiuto F. Chronic inflammation, periodontitis and cardiovascular diseases. Oral Dis 2008;14:102-4. 10. Persson GR, Persson RE. Cardiovascular disease and periodontitis: an update on the associations and risk. J Clin Periodontol 2008;35: 362-79.

350 Fedele et al

11. Davé S, Van Dyke T. The link between periodontal disease and cardiovascular disease is probably inflammation. Oral Dis 2008;14: 95-101. 12. D'Aiuto F, Ready D, Tonetti MS. Periodontal disease and C-reactive protein–associated cardiovascular risk. J Periodontal Res 2004;39: 236-41. 13. Federman DG, Shelling M, Prodanovich S, et al. Psoriasis: an opportunity to identify cardiovascular risk. Br J Dermatol 2009;160: 1-7. 14. Mallbris L, Akre O, Granath F, et al. Increased risk for cardiovascular mortality in psoriasis inpatients but not in outpatients. Eur J Epidemiol 2004;19:225-30. 15. Neimann AL, Shin DB, Wang X, et al. Prevalence of cardiovascular risk factors in patients with psoriasis. J Am Acad Dermatol 2006;55: 829-35. 16. Kaye JA, Li L, Jick SS. Incidence of risk factors for myocardial infarction and other vascular diseases in patients with psoriasis. Br J Dermatol 2008;159:895-902. 17. Peters MJ, van Halm VP, Voskuyl AE, et al. Does rheumatoid arthritis equal diabetes mellitus as an independent risk factor for cardiovascular disease? A prospective study. Arthritis Rheum 2009;61: 1571-9. 18. Rho YH, Chung CP, Oeser A, et al. Inflammatory mediators and premature coronary atherosclerosis in rheumatoid arthritis. Arthritis Rheum 2009;61:1580-5. 19. Bergström U, Jacobsson LT, Turesson C. Cardiovascular morbidity and mortality remain similar in two cohorts of patients with long-standing rheumatoid arthritis seen in 1978 and 1995 in Malmo, Sweden. Rheumatology (Oxford) 2009;48:1600-5. 20. Smeeth L, Thomas SL, Hall AJ, et al. Risk of myocardial infarction and stroke after acute infection or vaccination. N Engl J Med 2004;351: 2611-8. 21. Warren-Gash C, Smeeth L, Hayward AC. Influenza as a trigger for acute myocardial infarction or death from cardiovascular disease: a systematic review. Lancet Infect Dis 2009;9:601-10. 22. Cobb CM, Williams KB, Gerkovitch MM. Periodontol 2000 2009;50: 13-24. 23. Tonetti MS, D'Aiuto F, Nibali L, et al. Treatment of periodontitis and endothelial function. N Engl J Med 2007;356:911-20. 24. Scully C. Clinical practice. Aphthous ulceration. N Engl J Med 2006; 355:165-72. 25. Arduino PG, Porter SR. Herpes simplex virus type 1 infection: overview on relevant clinico-pathological features. J Oral Pathol Med 2008;37:107-21. 26. Ellepola AN, Samaranayake LP. Oral candidal infections and antimycotics. Crit Rev Oral Biol Med 2000;11:172-98. 27. McCartan BE, Healy CM. The reported prevalence of oral lichen planus: a review and critique. J Oral Pathol Med 2008;37: 447-53. 28. US Department of Health and Human Services (DHHS). National Center for Health Statistics. Third National Health and Nutrition Examination Survey, 1988-1994, NCHS CD-ROM Series 11, No. 1, Version 1.22a. Public Use Data File Documentation Number 76200. Hyattsville, MD: Centers for Disease Control and Prevention; 1997.

American Heart Journal February 2011

29. Shulman JD, Beach MM, Rivera-Hidalgo F. The prevalence of oral mucosal lesions in U.S. adults: data from the Third National Health and Nutrition Examination Survey, 1988-1994. J Am Dent Assoc 2004;135:1279-86. 30. Slade GD, Offenbacher C, Beck JD, et al. Acute-phase inflammatory response to periodontal disease in the US population. J Dent Res 2000;79:49-57. 31. Ford ES, Giles WH. Serum C-reactive protein and fibrinogen concentrations and self-reported angina pectoris and myocardial infarction: findings from National Health and Nutrition Examination Survey III. J Clin Epidemiol 2000;53:95-102. 32. Rose GA. Cardiovascular survey methods. 2nd ed. Geneva: World Health Organization; 1982. 33. Page RC, Eke PI. Case definitions for use in population-based surveillance of periodontitis. J Periodontol 2007;78:1387-99. 34. Sabbah W, Tsakos G, Chandola T, et al. Social gradients in oral and general health. J Dent Res 2007;86:992-6. 35. Sabbah W, Watt RG, Sheiham A, et al. Effects of allostatic load on the social gradient in ischaemic heart disease and periodontal disease: evidence from the third national health and nutrition examination survey. J Epidemiol Community Health 2008;62: 415-20. 36. Scannapieco FA, Bush RB, Paju S. Ann Periodontol 2003;8:38-53. 37. Zhang Y, Lin M, Zhang S, et al. NF-kappaB–dependent cytokines in saliva and serum from patients with oral lichen planus: a study in an ethnic Chinese population. Cytokine 2008;41:144-9. 38. Sun A, Chang YF, Chia JS, et al. Serum interleukin-8 level is a more sensitive marker than serum interleukin-6 level in monitoring the disease activity of recurrent aphthous ulcerations. J Oral Pathol Med 2004;33:133-9. 39. Sun A, Chia JS, Chang YF, et al. Serum interleukin-6 level is a useful marker in evaluating therapeutic effects of levamisole and Chinese medicinal herbs on patients with oral lichen planus. J Oral Pathol Med 2002;31:196-203. 40. Salonen EM, Vaheri A. C-reactive protein in acute viral infections. J Med Virol 1981;8:161-7. 41. Sarov I, Shainkin-Kestenbaum R, Zimlichman S, et al. Serum amyloid A levels in patients with infections due to cytomegalovirus, varicella-zoster virus, and herpes simplex virus. J Infect Dis 1982; 146:443. 42. Rupprecht HJ, Blankenberg S, Bickel C, et al. Impact of viral and bacterial infectious burden on long-term prognosis in patients with coronary artery disease. Circulation 2001;104:25-31. 43. Danesh J, Collins R, Peto R. Chronic infections and coronary heart disease: is there a link? Lancet 1997;350:430-6. 44. Pietruski JK, Pietruska MD, Jabłońska E, et al. Interleukin 6, tumor necrosis factor alpha and their soluble receptors in the blood serum of patients with denture stomatitis and fungal infection. Arch Immunol Ther Exp (Warsz) 2000;48:101-5. 45. Minassian C, D'Aiuto F, Hingorani AD, et al. Invasive dental treatment and risk for vascular events: a self-controlled case series. Ann Intern Med 2010;153:499-506. 46. Ismail SB, Kumar SK, Zain RB. Oral lichen planus and lichenoid reactions: etiopathogenesis, diagnosis, management and malignant transformation. J Oral Sci 2007;49:89-106.

Prevention and Rehabilitation

Reducing cardiovascular disease risk in medically underserved urban and rural communities Alfred A. Bove, MD, PhD, FACC, a William P. Santamore, PhD, a Carol Homko, RN, PhD, a Abul Kashem, MD, PhD, a Robert Cross, MD, a Timothy R. McConnell, PhD, b Gail Shirk, RN, c and Francis Menapace, MD c Philadelphia, Bloomsburg, and Danville, PA

Objectives The aim of this study is to evaluate methods for lowering cardiovascular disease (CVD) risk in asymptomatic urban and rural underserved subjects. Background Medically underserved populations are at increased CVD risk, and systems to lower CVD risk are needed. Nurse management (NM) and telemedicine (T) systems may provide low-cost solutions for this care. Methods We randomized 465 subjects without overt CVD, with Framingham CVD risk N10% to NM with 4 visits over 1 year, or NM plus T to facilitate weight, blood pressure (BP), and physical activity reporting. The study goal was to reduce CVD risk by 5%. Results Three hundred eighty-eight subjects completed the study. Cardiovascular disease risk fell by ≥5% in 32% of the NM group and 26% of the T group (P, nonsignificant). In hyperlipidemic subjects, total cholesterol decreased (NM −21.9 ± 39.4, T −22.7 ± 41.3 mg/dL) significantly. In subjects with grade II hypertension (systolic BP ≥160 mm Hg, 24% of subjects), both NM and T groups had a similar BP response (average study BP: NM 147.4 ± 17.5, T 145.3. ± 18.4, P is nonsignificant), and for those with grade I hypertension (37% of subjects), T had a lower average study BP compared to NM (NM 140.4 ± 16.9, T 134.6 ± 15.0, P = .058). In subjects at high risk (Framingham score ≥20%), risk fell 6.0% ± 9.9%; in subjects at intermediate risk (Framingham score ≥10, b20), risk fell 1.3% ± 4.5% (P b .001 compared to high-risk subjects). Medication adherence was similar in both high- and intermediate-risk subjects. Conclusions In 2 underserved populations, CVD risk was reduced by a nurse intervention; T did not add to the risk improvement. Reductions in BP and blood lipids occurred in both high- and intermediate-risk subjects with greatest reductions noted in the high-risk subjects. Frequent communication using a nurse intervention contributes to improved CVD risk in asymptomatic, underserved subjects with increased CVD risk. Telemedicine did not change the effectiveness of the nurse intervention. (Am Heart J 2011;161:351-9.)

Recent data have demonstrated a significant reduction in cardiovascular disease (CVD) mortality in part, related to more aggressive management of modifiable CVD risk factors.1 Although mortality from CVD is diminishing, ethnic minorities and medically underserved populations

From the aTemple University Medical Center, Philadelphia, PA, bBloomsburg University, Bloomsburg, PA, and cGeisinger Medical Center, Danville, PA. Clinical Trial no: NCT00778804. This study was supported by Pa State Grant no. RFA-ME02-380, a grant from the Commonwealth of Pennsylvania. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations, or conclusions. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents. Submitted June 1, 2010; accepted November 7, 2010. Reprint requests: Alfred A. Bove, MD, PhD, FACC, Cardiology Section, Temple University Medical Center, Philadelphia, PA 19140. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.008

are at increased CVD risk because of a high prevalence of obesity with accompanying diabetes, hyperlipidemia, and hypertension.2-6 Management of the presymptomatic phase of these disorders is best done by incorporating patient participation, improved health literacy, and monitoring of patient status through frequent communication between patient and health care provider. Nurse management (NM) has proven effective in improving diabetes,7,8 hyperlipidemia,9 and hypertension.10 In this study, we compared an NM CVD risk reduction program to an NM system augmented with telemedicine (T) communication. The T system allowed subjects to report their weight, blood pressure (BP), and physical activity and to receive feedback regarding CVD risk management.

Methods The study was conducted at Temple University Medical Center, which serves an inner city, predominantly African-

American Heart Journal February 2011

352 Bove et al

American population, and at Geisinger Medical Center, which serves a rural, predominantly white population. Both institutions provide health care in areas designated as medically underserved.11 The study was funded by a Pennsylvania State Department of Health grant number RFA-ME02-380. The study was approved by the Institutional Review Boards at Temple University Medical School, the Geisinger Medical Center, and through an independent institutional review board for Insight Telehealth, LLC, which supported the Web site and the information database used in the T arm of the study. Patients of either sex between 18 and 85 years old and with a ≥10% 10-year risk of CVD12,13 were randomized to NM or NM plus T communication. The patients were provided resources for measuring BP, weight, and daily activity at home and were followed up for 1 year with the primary end point being a 5% reduction in their 10-year CVD risk.

Baseline assessment and randomization Subjects were recruited through public advertising, presentations at community centers and churches, and contact with medical clinics at both institutions. All subjects were able to read and had access to a telephone. Exclusion criteria included clinically evident coronary artery disease, class 3 or 4 heart failure, end-stage renal disease, subjects living in nursing homes or boarding homes, and pregnancy. All subjects signed an informed consent. The subjects reported to a clinical research center at Temple or Geisinger medical centers for initial evaluation by our research staff that included a physical examination, a fasting blood sample to determine blood glucose, A1c, serum cholesterol, low-density lipoprotein (LDL), highdensity lipoprotein (HDL), triglycerides, and C-reactive protein (CRP). The subjects underwent assessment of CVD risk using the Framingham risk model, which incorporates age, sex, BP, cholesterol, HDL, smoking status, and diabetes to determine the subject's 10-year risk of CVD.12 All subjects were trained on use of a computer and the Internet and tested for competency using a previously developed protocol.14 Subjects who did not achieve a basic level of Internet competency after a 2-hour training session were excluded from the study. Only 2 individuals failed to achieve adequate skill levels. After the computer training, 465 subjects were randomized to NM or NM plus T groups. Subjects were provided a digital sphygmomanometer, a scale if needed (Taylor Digital LCD Scale-7006, Taylor Precision Products, Las Cruces, NM), and a pedometer (Digi-Walker SW200, Yamax Inc, Tokyo, Japan) to count their steps per day. Twelve percent were randomly selected to use a digital sphygmomanometer with memory (3AC1-AP; Microlife USA Inc, Dunedin, FL). Health knowledge and behavioral surveys were provided to all subjects at the beginning and end of the study. All patients returned to their clinical research center at 4, 8, and 12 months for repeat examination and laboratory studies. Of the 465 subjects who were randomized, 388 (83.4%) completed the 1-year study.

Nurse management Nurse management involved office encounters with our research nurses at 4-month intervals for 1 year. The team of physicians and nurses involved in the research study did not manage the health care of the study patients. Rather, they

provided education, reminders, and surveillance of CVD risk to the subjects and their care providers. Subjects were instructed to record their data (weight, BP, steps per day, and cigarettes per day) at least weekly and enter the data in a logbook, which was reviewed quarterly during a clinic visit with a research nurse. At each visit, subjects were asked to grade their medication adherence and to report the number of primary care visits that they experienced in the previous 4 months. Subjects received counseling regarding healthy lifestyle behaviors at each clinical encounter and were encouraged to communicate their information to their primary care physician. Accuracy of BP reporting was confirmed by comparing reported data with values stored in the digital sphygmomanometer.15

Telemedicine Subjects in the T group were provided with the NM program described above. In addition, each subject was provided with a login name and password to gain access to a secure Internetbased T system and received instructions on how to access and use the T system. Laboratory data and medications were entered into the T system by a research nurse and were accessible to the subject via the Internet. To increase computer access, we provided local churches with computers and Internet access and identified community centers and libraries where Internet access was available. Subjects were instructed to send their data at least weekly via the T system. The study investigators reviewed this information and sent either an automated response indicating that the data were acceptable or a tailored message with advice and instructions aimed at reducing CVD risk. The T subjects also had access to their individual BP data and to educational information about CVD risk factors. Each subject (via T and mail) and their primary care physician (via fax) received summaries of their BP and lipid values, guideline-recommended goals for BP and lipids, and their overall CVD risk assessment at 4-month intervals during the 1 year of surveillance. To improve patient-to-provider communication, the subjects were instructed to bring the paper report to their primary care physician for discussion during their office visits.

Telemedicine system The Telemedicine system (InSight Telehealth Systems; LLC, Valley Forge, PA) is a disease-management interactive health surveillance system composed of a secure Internet server and a database. Details of the Telemedicine system have been described in previous publications.16,17 This system provides Internet access to a personal health record focused on CVD and allows patients to send data directly to the research team via the Internet. Patient health information and the communication transactions are stored in the database. The Web site is divided into a patient domain and a provider domain. Each requires a login ID and password. This system provides Health Information Portability and Accountability Act-compliant 128-bit encryption with a secure socket layer and a public key infrastructure so that medical information is safely transmitted via the Internet.

Statistical analysis The primary end point was a reduction in CVD risk score of 5%. We hypothesized that 37.5% of the T group would achieve this goal compared to 25% of the NM group. We set α = .05 and

American Heart Journal Volume 161, Number 2

Bove et al 353

Table I. Characteristics of the urban and rural subjects who completed the study

n Age (y) Male (%) African American (%) White (%) Hispanic (%) Hypertension (%) Hyperlipidemia (%) Smoker (%) Diabetes (%) BMI ≥30 (%) High CVD risk (%)

Rural

Urban

205 62.5 ± 9.6 55.6 1.0 99.0 0.0 73.7 86.3 10.2 36.6 55.5 24.9

183 57.9 ± 10.1⁎ 52.5 76.5 14.2 7.7 51.4⁎ 84.2 39.3⁎ 57.4⁎ 59.5 32.2

⁎ Significant difference between urban and rural with P b .01.

power at 0.8 and needed a minimum of 450 patients in the study to meet the power requirement with an anticipated dropout rate of 20%. The primary end point was analyzed with a χ2 analysis. We compared continuous clinical variables using analysis of variance and paired and unpaired Student t tests. We tested the influences of demographic factors on outcome using binary logistic regression. Secondary end points compared outcome in intermediate and high CVD risk subjects. Randomization was accomplished by assigning subjects in sequence to a random number list. Even random numbers were assigned to the T group; odd random numbers were assigned to the NM group. Tabular data are expressed as mean ± SD; graphics show SEs. Statistical analysis was performed using SPSS v18.

Results Table I shows the baseline characteristics of the 388 subjects (83.4%) who completed the study. Notable differences in the study populations include a dominance of African-American subjects in the urban group, a higher incidence of diabetes, and a higher incidence of smoking in the urban group, whereas hypertension was more prevalent in the rural subjects. One fourth of the rural subjects had a high risk score (≥20%), whereas nearly one third of the urban subjects were at high risk (P = .068). More than half of both groups were obese (body mass index [BMI] ≥30).

Telemedicine versus nurse management After randomization, 49.3% of the rural subjects and 50.3% of the urban subjects were assigned to the T group. Table II shows the demographic distribution of the T and NM groups. Ethnic background was similar in the 2 groups; annual income for 50% of the NM group and 36% of the T group was near or below the federal poverty level (U.S.$22,050). Table III shows the data for initial and final measures in the 2 groups. The T group had a higher number of persons with diabetes and showed a larger reduction in systolic BP (SBP). Changes in both groups over the 1-year study were similar in the NM and T

Table II. Race/ethnicity, family income, and educational level of the study population NM (n = 195) Race/ethnicity White African American Latino/Hispanic Other Total family income b$15000 $15000-24999 $25000-34999 $35000-44999 N$45000 Not reported Education bHigh school High school graduate NHigh school

NM + T (n = 193)

58% 37% 4% 1%

(114) (72) (7) (2)

59% 36% 4% 1%

(115) (70) (7) (1)

27% 23% 16% 14% 15% 5%

(53) (45) (32) (28) (29) (8)

24% 12% 27% 13% 22% 2%

(46) (23) (52) (25) (43) (4)

11% (21) 42% (83) 47% (91)

7% (13) 37% (72) 56% (108)

Data are expressed as a percentage (actual no. of subjects). Income levels are in U.S. dollars.

groups. In both groups, risk score was reduced early in the study and was sustained throughout the 1-year study (Figure 1). Risk reduction in both groups was accomplished predominantly by reduction in prevalence of hypertension and hyperlipidemia. The rate of BP monitoring was significantly higher (P b .05) in the T group (92% of the subjects) compared to the NM group (48%). In the T group, 75% of patients met or exceeded the suggested weekly rate of monitoring. The overall reporting rate for the T group averaged 6.3 reports per month.

Hypertension In patients with grade I hypertension (SBP b160 and ≥140 mm Hg), the T group had a more rapid and larger reduction in SBP than patients in the NM group, and a slightly higher percentage of subjects achieved an SBP b140 mm Hg (67% vs 53%, nonsignificant [NS]). For grade I hypertension, the earlier, sustained reduction noted in the T group resulted in a lower mean pressure for the year. After the initial fall in BP, the average systolic pressure over the remaining 8 months of the study for subjects with grade I hypertension was 140.4 ± 16.9 in the NM group and 134.6 ± 15.0 in the T group (P = .058). For subjects with grade II hypertension (SBP ≥160 mm Hg) at baseline, SBP decreased significantly over the 1-year study period with no difference in average BP between NM (147.4 ± 17.5) and T groups (145.3 ± 18.4). Hyperlipidemia Over the 1-year study, total and LDL cholesterol decreased in both the NM and T groups. However, the major changes in the lipid profile occurred in the first 4 months of the study. Thereafter, the values for total and LDL cholesterol did not change. Among subjects

American Heart Journal February 2011

354 Bove et al

Table III. Baseline and final values for NM and T subjects NM (n = 195) Baseline Age (y) Male, n (%) African American, n (%) White, n (%) Hispanic, n (%) Urban, n (%) Obese, n (%) Height, in Weight, lb Waist circumference, in BMI, kg/m2 Systolic BP, mm Hg Diastolic BP, mm Hg Total cholesterol, mg/dL HDL cholesterol, mg/dL LDL cholesterol, mg/dL Triglycerides, mg/dL Blood glucose, mg/dL A1c, (%) CRP, mg/L 6-min walk, ft Smoker (%) Diabetes (%) Hyperlipidemia (%) Hypertension (%) CVD risk (%)

61.8 ± 10.5 101 (52) 72 (37) 114 (59) 7 (4) 91 (46.7) 115 (59) 66.6 ± 10.5 200.4 ± 39.5 40.9 ± 5.8 31.9 ± 6.3 145.6 ± 18.7 81.9 ± 11.8 203 ± 46 47 ± 13 121 ± 41 180 ± 170 123 ± 47 6.44 ± 1.40 .473 ± .645 1469 ± 345 28 42 86.2 63.1 17.8 ± 9.6

T (n = 193) Final

Baseline

Final

199.2 ± 40.0 41.0 ± 5.9 31.7 ± 6.5 136.5 ± 18.6† 77.8 ± 9.4† 193 ± 44† 49 ± 14† 114 ± 41† 150 ± 98† 117 ± 39 6.45 ± 1.28 .583 ± .885 1494 ± 399 24 42 57.7† 41.5† 15.1 ± 9.3†

60.9 ± 9.6 109 (56) 70 (36) 115 (60) 7 (4) 92 (47.7) 109 (56.5) 66.9 ± 9.6 201.4 ± 40.0 40.8 ± 6.0 31.8 ± 6.5 146.4 ± 18.0 82.3 ± 10.4 199 ± 44 47 ± 13 120 ± 39 163 ± 91 133 ± 60 6.86 ± 1.74 .461 ± .713 1526 ± 345 20 51⁎ 84.5 63.2 17.5 ± 10.3

200.8 ± 42.2 40.6 ± 6.2 31.6 ± 6.9 134.3 ± 17.9† 77.2 ± 10.5† 190 ± 40† 47 ± 13 114 ± 36† 149 ± 81† 126 ± 50† 6.67 ± 1.43† .437 ± .555 1550 ± 400 17 51⁎ 61.1† 35.8† 15.0 ± 9.3†

Data are expressed as mean ± SD. ⁎ Significant difference compared to NM group at P b .01. † Significant difference compared to baseline at P b .05.

Figure 1

group and 26.9% in the T group reached a goal of ≤100 mg/dL (P, NS).

Time course of reduction in Framingham risk score throughout the 1-year study. Changes were similar in the NM and T groups. The greatest changes occurred in the first 4 months of the study.

Effects of initial risk score Although there were no differences found for the total cohort when comparing the NM group to the T group, we did note important differences in the subjects who entered the study with a high Framingham risk score compared to those who were at intermediate risk. Tables IV and V show the data for T and NM groups in the 2 risk groups. The percent reduction in risk was significantly greater in the high-risk group (19.1% ± 33.6% vs 8.1% ± 34.2%, P = .004). Risk reduction was accomplished by reductions in blood lipids and BP in both groups. Changes in lipids and BP were greater in the high-risk group (Table V) compared to the intermediate-risk group (Table IV).

who had elevated total cholesterol levels at baseline, 37.6% in the NM group and 35.4% in T group reached a goal of ≤200 mg/dL (P, NS). In subjects with elevated LDL cholesterol at baseline, 25.6% in the NM

Medication adherence We recorded the medication adherence by asking the subjects at each 4-month visit to complete a questionnaire on aspirin, hypertension, lipid, and diabetes medication. Adherence was self-reported on a scale from 1 to 4, with 4 being full adherence and 1 being occasional adherence.

American Heart Journal Volume 161, Number 2

Bove et al 355

Table IV. Baseline and final data for subjects with initial risk between 10% and 20% (intermediate risk) NM (n = 137) Baseline Age (y) Male, n (%) African American, n (%) White, n (%) Hispanic, n (%) Urban, n (%) Obese, n (%) Height, in Weight, lb Waist circumference, in BMI, kg/m2 Systolic BP, mm Hg Diastolic BP, mm Hg Total cholesterol, mg/dL HDL cholesterol, mg/dL LDL cholesterol, mg/dL Triglycerides, mg/dL Blood glucose, mg/dL A1c (%) CRP, μg/L 6-min walk, ft Smoker (%) Diabetes (%) Hyperlipidemia (%) Hypertension (%) CVD risk (%)

61.2 ± 10.7 70 (51) 51 (37) 83 (61) 2 (2) 62 (45.3) 75 (54.7) 66.7 ± 3.8 199.6 ± 39.40 40.6 ± 5.8 31.6 ± 6.2 144.3 ± 18.0 81.7 ± 12.1 197.2 ± 44.6 48.4 ± 13.4 116.9 ± 41.0 158.8 ± 108.4 115.2 ± 38.3 6.24 ± 1.27 0.422 ± 0.611 1488 ± 321 21 35 82.5 59.1 12.8 ± 3.8

T (n = 141) Final

Baseline

198.3 ± 40.2 40.8 ± 5.9 31.4 ± 6.4 136.0 ± 19.1† 78.5 ± 9.2† 191.6 ± 46.0 50.8 ± 14.6† 112.3 ± 41.1 144.4 ± 93.8† 110.4 ± 29.2 6.24 ± 1.03 0.531 ± 0.841 1520 ± 401 22 35 53.7 40.9† 11.7 ± 5.5†

59.2 ± 9.4 79 (56) 48 (34) 85 (60) 7 (5) 62 (44.0) 72 (51.1) 67.2 ± 4.4 200.8 ± 42.2 40.3 ± 6.2 31.4 ± 7.0 145.3 ± 18.1 82.4 ± 10.4 199.4 ± 45.2 49.3 ± 13.1 118.6 ± 39.5 159.4 ± 92.1 126.8 ± 55.9⁎ 6.67 ± 1.73 0.422 ± 0.672 1540 ± 360 15 42 82.3 61.7 12.8 ± 3.6

Final

200.7 ± 45.2 40.2 ± 6.8 31.3 ± 7.5 133.1 ± 17.8† 77.7 ± 10.7† 191.6 ± 42.0† 49.2 ± 13.0 114.2 ± 37.9† 148.5 ± 82.1 121.5 ± 43.8⁎ 6.51 ± 1.40 0.371 ± 0.420⁎ 1589 ± 409 20 40 61.7 34.0† 11.4 ± 5.0†

⁎ NM versus T significance at P b .05. † Baseline versus final significance at P b .05.

Figure 2 shows the average scores for the high- and intermediate-risk groups for each medication class. There were no differences for any of the medications when comparing high-risk versus intermediate-risk groups. Telemedicine had no influence on adherence when compared to the NM group.

Primary care visits We also queried the subjects at each visit to establish the number of primary care visits that they had at each 4-month interval. Figure 3 shows the number of visits in each group over the 1-year study. The high-risk subjects showed a higher total number of visits (year total: intermediate 3.9 ± 3.0, high 4.6 ± 3.4, P = .044). To identify factors that characterized those who reached the study goal and those who did not, we used binary logistic regression to examine the effects of age, sex, ethnic group, urban or rural location, use of T, income, education level, and initial risk score. The initial Framingham score was the only factor that identified those who achieved the study goal. The reduction in score was positively correlated with the initial score (r = 0.433, P b .001).

Discussion There are ample data that demonstrate significant differences in health status and health care outcome in low-income underserved populations.2-6 To reduce the incidence of CVD, it is necessary to engage subjects in healthy lifestyle behaviors and to manage treatable CVD risk. Nurse management alone proved to be successful in reducing overall CVD risk. The improved CVD risk occurred primarily by reductions in serum cholesterol and BP. Communication in both groups between subjects, the research nurses, and the care providers created the needed incentive to improve CVD risk.

Hyperlipidemia All subjects in the study were updated at 4-month intervals on their progress and on the importance of lipid lowering. Over the 1-year study, total blood cholesterol and LDL were reduced significantly in both NM and T groups. The major improvements in lipids occurred in the first 4 months. About one third of the subjects with elevated baseline total cholesterol levels reached recommended targets. Telemedicine did not add further to the lowering of blood lipids.

American Heart Journal February 2011

356 Bove et al

Table V. Baseline and final data for subjects with initial risk N20% (high risk) NM (n = 58)

T (n = 52)

Baseline Age, y Male, n (%) African American, n (%) White, n (%) Hispanic, n (%) Urban (%) Obese (%) Height, in Weight, lb Waist circumference, in BMI, kg/m2 Systolic BP, mm Hg Diastolic BP, mm Hg Total cholesterol, mg/dL HDL cholesterol, mg/dL LDL cholesterol, mg/dL Triglycerides, mg/dL Blood glucose, mg/dL A1c (%) CRP, μg/L 6-min walk, ft Smoker (%) Diabetes (%) Hyperlipidemia (%) Hypertension (%) CVD risk (%)

63.2 ± 10.0 31 (53) 21 (36) 31 (53) 5 (9) 29 (50) 40 (69) 66.4 ± 3.8 202.1 ± 40.1 41.5 ± 5.7 32.4 ± 6.6 148.6 ± 20.0 82.3 ± 11.1 215.7 ± 45.4 43.3 ± 9.6 131.4 ± 40.1 231.5 ± 259.9 141.0 ± 58.4 6.95 ± 1.57 0.599 ± 0.702 1355 ± 436 45 59 94.8 72.4 29.7 ± 8.6

Final

201.4 ± 39.7 41.6 ± 5.4 32.3 ± 6.7 137.6 ± 17.5† 76.0 ± 9.9† 194.4 ± 37.5† 44.5 ± 10.5 116.3 ± 39.5† 166.1 ± 108.4† 134.5 ± 51.2 6.96 ± 1.63 0.609 ± 1.016 1421 ± 412 45 59 67.2 43.1 23.2 ± 11.3†

Baseline 65.4 ± 8.8 30 (58) 22 (42) 30 (58) 0 30 (57.7) 37 (71.2) 66.1 ± 3.2 203.1 ± 33.6 42.3 ± 5.0 32.7 ± 5.1 149.2 ± 17.6 82.2 ± 10.7 198.8 ± 41.0 40.3 ± 7.8⁎ 125.1 ± 35.4 176.1 ± 88.1 150.0 ± 66.4 7.38 ± 1.69⁎

0.612 ± 0.803⁎ 1427 ± 334 33 75⁎ 90.4 67.3 30.0 ± 12.0

Final

201.1 ± 33.3 41.7 ± 4.3 32.4 ± 5.2 137.4 ± 17.9† 75.6 ± 9.7† 184.0 ± 34.8† 40.2 ± 9.0⁎ 113.9 ± 30.6† 151.0 ± 76.9 138.1 ± 61.3 7.12 ± 1.41 0.600 ± 0.717 1436 ± 350 33 75⁎ 59.6 40.4 24.5 ± 11.5†

⁎ NM versus T significance at P b .05. † Baseline versus final significance at P b .05.

Figure 2

Figure 3

Average medication compliance scores for patients at high and intermediate initial risk. Data were obtained from questionnaires at 4 visits during the 1-year study. A score of 4 represents full compliance. Error bars show SEM. HTN, Hypertension.

Number of PCP visits over the 4-month intervals in the 1-year study. The high-risk subjects had a greater number of visits throughout the year (P = .044). PCP, Primary care provider.

American Heart Journal Volume 161, Number 2

Robinson et al18 found an increase in statin use with a corresponding reduction in LDL cholesterol with use of postcard reminders. Automated telephone calls have been used in low-income public clinic patients to help manage lipid levels.19 Electronic medical records have also been used to identify patients with hyperlipidemia and to generate automated telephone calls or e-mails.20 Our data also support the concept that frequent reminders improve adherence to a CVD risk reduction program.

Hypertension Significant reductions occurred in BP within 4 months of starting the study and were sustained throughout the 1-year study. Similar improvements in BP were found in NM and T groups. Subjects with high CVD risk showed the greatest improvements in BP. Other studies have shown a benefit of telecommunication systems using home-monitored BP to manage hypertension.21,22,23 Artinian et al24 evaluated an NM program that incorporated telemonitoring using a device that linked the BP monitor to the telephone. Over a 12-month period, the telemonitoring group had a greater reduction in SBP (13.0 mm Hg) than controls (7.5 mm Hg). Green et al25 evaluate an Internet-based BP reporting system, incorporating a pharmacist intervention to manage medications. Fifty-six percent of patients in the Internet plus pharmacist group reached goal BP (b140/90) compared with 36% who self-monitored but did not receive feedback for BP management. In our study, 57% of the T group reached goal BP, whereas 46% of the NM patients reached goal BP. Madsen et al26 demonstrated that home-measured BP combined with a T system for BP management could achieve similar BP control compared with patients who were followed up by frequent office visits. Internet-reported BP can also allow patients with hypertension to self-titer medication based on home-measured BP.27 Implications A better understanding of health disparities in underserved communities has made clear the need for improved health literacy and healthier living in these communities.28 The increased incidence of stroke and heart failure in particular is related to the high prevalence of hypertension and hyperlipidemia.29 Our study shows that improved communication between underserved subjects and a health provider will reduce CVD risk. Subjects with high CVD risk achieved the greatest response and visited with their caregivers more than subjects with intermediate risk. It is of interest that a higher incidence of hypertension was noted in the rural subjects, whereas the urban subjects had a higher incidence of diabetes and cigarette smoking. High risk is achieved in the rural communities based on hypertension

Bove et al 357

and hyperlipidemia, whereas the urban group achieved high risk from diabetes and cigarette smoking and less often from hypertension. Hyperlipidemia incidence was high in both groups, and more than half of urban and rural subjects were obese. These findings suggest that risk reduction in underserved urban communities should focus not only on hypertension and hyperlipidemia but also on diabetes and cigarette smoking. The study by Haskell et al30 highlights the time and effort in an NM approach and the clinical benefits of such an approach. Compared with usual care, NM produced clinically important decreases in SBP, total cholesterol, and LDL cholesterol. These improvements may not be adequate to achieve long-term risk reduction if smoking, obesity, and diabetes are not also addressed, particularly in the urban high-risk population. Home BP monitoring is an essential component of BP control.31 Many studies have used telemonitoring devices to capture home BP measurements and systematically transmit these data via telephone or computer to a data repository.23,24,26,32,33 However, similar BP reductions have been achieved in studies without automated telemonitoring devices.25,32,34 In our study, the subjects measured and entered their BP values into a T Web site and had a similar decrease in SBP (15.2 mm Hg), as reported in the study of Artinian et al24 (13.0 mm Hg) where a telemonitoring device was used. Providing inexpensive or free BP measurements at publically available sites coupled with an Internet-reporting system might achieve similar goals at minimal expense. There appears to be little difference in clinical outcomes between studies with direct provider-patient contact and indirect contact.24,26,34 An advantage of a technology approach is that the communication can be easily documented and can become part of an electronic health record. Automated, computer-generated messages offer another time-saving approach, which has been used successfully to manage hypertension.33 For medication management, nurse, pharmacist, and self-management have been successful.24,25,27 Nonphysician care providers could be assisted in medical management using expert systems.35,36

Limitations of the study The research team provided education, reminders, and surveillance of CVD risk to the subjects and their care providers. In the T group, the research team provided written documentation of the patients' CVD risk status to both patient and primary care physician in addition to appropriate guideline-based goals of therapy. At completion of the 1-year study, most study subjects stated that the periodic reports were reviewed with their primary care provider and made communication more meaningful. Although our data were unable to distinguish the exact cause of the reduced CVD risk, the feedback on health status and guideline-based goals to both patient

American Heart Journal February 2011

358 Bove et al

and primary care physician likely contributed to the lowered BP and blood lipids.

Conclusions In rural and urban medically underserved populations, NM was successful in reducing overall CVD risk. Study subjects who were classified initially as high risk had the greatest response and visited their care provider more often. Medication adherence was similar in subjects who started the study with high risk compared to intermediate risk. Modifying CVD risk in urban communities must focus on diabetes and cigarette smoking in addition to hypertension and hyperlipidemia.

12.

13.

14.

15.

16.

Disclosures Alfred A. Bove, M.D., Ph.D. is a Consultant for InSight Telehealth Inc. William P. Santamore, Ph.D. owns stocks in InSight Telehealth Inc.

17.

18.

References 1. Ford ES, Ajani UA, Croft JB, et al. Explaining the decrease in U.S. deaths from coronary disease, 1980-2000. N Engl J Med 2007;356: 2388-98. 2. Lloyd-Jones D, Adams RJ, Brown TM, et al. Heart disease and stroke statistics—2010 update: a report from the American Heart Association. Circulation 2010;121:e46-e215. 3. Vital and Health Statistics. Summary Health Statistics for U.S. Adults: National Health Interview Survey, 2009. Series 10: Data From the National Health Interview Survey No. 249 U.S. Department of Health and Human Services Centers for Disease Control and Prevention. National Center for Health Statistics. Hyattsville, Maryland. August 2010 DHHS Publication No. (PHS) 2011-1577. 4. Appel SJ, Harrell JS, Deng S. Racial and socioeconomic differences in risk factors for CVD among Southern rural women. Nurs Res 2002; 51:140-7. 5. Cooper R, Cutler J, Desvigne-Nickens P, et al. Trends and disparities in coronary heart disease, stroke, and other CVDs in the United States: findings of the National Conference on CVD Prevention. Circulation 2000;102:3137-47. 6. Yancey AK, Robinson RG, Ross RK, et al. Discovering the full spectrum of cardiovascular disease: Minority Health Summit 2003: report of the Advocacy Writing Group. Circulation 2005;111:e140-9. 7. Taylor CB, Miller NH, Reilly KR, et al. Evaluation of a nurse-care management system to improve outcomes in patients with complicated diabetes. Diabetes Care 2003;26:1058-63. 8. Aubert RE, Herman WH, Waters J, et al. Nurse case management to improve glycemic control in diabetic patients in a health maintenance organization. A randomized, controlled trial. Ann Intern Med 1998; 129:605-12. 9. Becker DM, Allen JK. Improving compliance in your dyslipidemic patient: an evidence-based approach. J Am Acad Nurse Pract 2001; 13:200-7. 10. Rudd P, Miller NH, Kaufman J, et al. Nurse management for hypertension. A systems approach. Am J Hypertens 2004;17: 921-7. 11. http://datawarehouse.hrsa.gov/GeoAdvisor/ShortageDesignationAdvisor.aspx. U.S. Department of Health and Human Services.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

Health resources and services administration. Last Accessed December 2, 2010. Wilson PW, D'Agostino RB, Levy D, et al. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97: 1837-47. D’Agostino RB, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care. The Framingham Heart Study. Circulation 2008;117:743-53. Masucci M, Homko C, Santamore WP, et al. Bridging the digital divide in underserved patients with cardiovascular risk. Telemed J E Health 2006;12:58-65. Santamore WP, Homko CJ, Kashem A, et al. Accuracy of blood pressure measurements transmitted through a telemedicine system in underserved populations. Telemed J E Health 2008; 14:333-8. Homko CJ, Santamore WP, Zamora LC, et al. Cardiovascular disease knowledge and risk perception among underserved individuals at increased risk for cardiovascular disease. J Cardiovasc Nurs 2008; 11:21-6. Kashem A, Droogan MT, Santamore WP, et al. Managing heart failure care using an Internet-based telemedicine system. J Card Fail 2008;14:121-63. Robinson JG, Conroy C, Wickemeyer WJ. A novel telephone-based system for management of secondary prevention to a low-density lipoprotein cholesterol b or = 100 mg/dl. Am J Cardiol 2000;85: 305-8. Hyman DJ, Ho KS, Dunn JK, et al. Dietary intervention for cholesterol reduction in public clinic patients. Am J Prev Med 1998; 15:139-45. Palmieri J, Redline S, Morita R. Goal attainment in patients referred to a telephone-based dyslipidemia program. Am J Health Syst Pharm 2005;62:1586-91. Bondmass MD, Bolger NE, Castro GM, et al. Rapid control of hypertension in African Americans achieved utilizing home monitoring. Circulation 1998;98(Suppl 1):I-517. Friedman RH, Kazis LE, Jette A, et al. Telecommunications system for monitoring and counseling patients with hypertension: impact on medication adherence and blood pressure control. AmJ Hypertens 1996;9:285-92. Rogers MA, Small D, Buchan DA, et al. Home monitoring service improves mean arterial pressure in patients with essential hypertension. A randomized, controlled trial. Ann Intern Med 2001;134: 1024-32. Artinian NT, Flack JM, Nordstrom CK, et al. Effects of nurse-managed telemonitoring on blood pressure at 12-month follow-up among urban African Americans. Nurs Res 2007;56:312-22. Green BB, Cook AJ, Ralston JD, et al. Effectiveness of home blood pressure monitoring, Web communication, and pharmacist care on hypertension control: a randomized controlled trial. JAMA 2008; 299:2857-67. Madsen LB, Kirkegaard P, Pedersen EB. Blood pressure control during telemonitoring of home blood pressure. A randomized controlled trial during 6 months. Blood Press 2008;17:78-86. Bobrie G, Postel-Vinay N, Delonca J, et al. Self-measurement and self-titration in hypertension: a pilot telemedicine study. Am J Hypertens 2007;20:1314-20. Sirovich BE, Gottlieb DJ, Welch HG, et al. Regional variations in health care intensity and physician perceptions of quality of care. Ann Intern Med 2006;144:641-9. Cooper RA. States with more health care spending have better-quality health care: lessons about medicare. Health Affairs 2009;28: w103-15.

American Heart Journal Volume 161, Number 2

Bove et al 359

30. Haskell WL, Berra K, Arias E, et al. Multifactor cardiovascular disease risk reduction in medically underserved, high-risk patients. Am J Cardiol 2006;98:1472-9. 31. Pickering TG, Miller NH, Ogedegbe G, et al. Call to action on use and reimbursement for home blood pressure monitoring. J Cardiovascular Nursing 2008;23:299-323. 32. O'Brien E. A Website for blood pressure monitoring devices: dableducational.com. Blood Press Monit 2003;8:117-80. 33. Logan AG, McIsaac WJ, Tisler A, et al. Mobile phone–based remote patient monitoring system for management of hypertension in diabetic patients. Am J Hypertens 2007;20:942-8.

O

34. Canzanello VJ, Jensen PL, Schwartz LL, et al. Improved blood pressure control with a physician-nurse team and home blood pressure measurement. Mayo Clin Proc 2005;80:31-6. 35. Goldstein MK. Using health information technology to improve hypertension management. Curr Hypertens Rep 2008;10: 201-7. 36. Goldstein MK, Hoffman BB, Coleman RW, et al. Implementing clinical practice guidelines while taking account of changing evidence: ATHENA DSS, an easily modifiable decision-support system for managing hypertension in primary care. Proc AMIA Symp 2000: 300-4.

N THE MOVE? Send us your new address at least 6 weeks ahead

Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least 6 weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed because of late notification. JOURNAL TITLE: Fill in the title of the journal here.

__________________________________________________________

OLD ADDRESS:

NEW ADDRESS:

Affix the address label from a recent issue of the journal here.

Clearly print your new address here. Name ____________________________________________________ Address __________________________________________________ City/State/ZIP _____________________________________________

COPY AND MAIL THIS FORM TO: Elsevier Health Sciences Division Subscription Customer Service 3251 Riverport Lane, Maryland Heights, MO 63043

OR FAX TO:

OR PHONE:

341-447-8029

1-800-654-2452 Outside the U.S., call 341-447-8871

OR E-MAIL:

[email protected]

Interventional Cardiology

Incidence and clinical outcome of minor surgery in the year after drug-eluting stent implantation: Results from the Evaluation of Drug-Eluting Stents and Ischemic Events Registry Emmanouil S. Brilakis, MD, PhD, a,h David J. Cohen, MD, MSc, b,h Neal S. Kleiman, MD, c,h Michael Pencina, PhD, d,h Deborah Nassif, PhD, d,h Jorge Saucedo, MD, e,h Robert N. Piana, MD, f,h Subhash Banerjee, MD, a,h Michelle J. Keyes, PhD, d,h Chen-Hsing Yen, MS, d,h and Peter B. Berger, MD g,h Dallas, and Houston, TX; Kansas City, MO; Boston, MA; Oklahoma City, OK; Nashville, TN; and Danville, PA

Background The aim of the study was to describe the incidence and consequences of minor surgery after drug-eluting stent (DES) implantation. Methods The Evaluation of Drug-Eluting Stents and Ischemic Events (EVENT) Registry prospectively enrolled unselected patients undergoing percutaneous coronary intervention at 47 US centers between July 2004 and December 2007. We examined 8,323 patients who received a DES in EVENT to determine the frequencies of minor surgery and postoperative adverse events. Results Minor surgery (defined as procedures not requiring a major surgical incision) was performed in 164 (2.0%) of 8,323 patients b1 year after stenting, as follows: pacemaker/defibrillator implantation (46%), eye surgery (17%), orthopedic (9%), dermatologic (8%), endovascular (6%), and gastrointestinal procedures (5%). Compared with patients who did not undergo minor surgery, those who did were older, had more comorbidities, had more extensive coronary disease, and were more likely to have received warfarin after stenting. Only 1 (0.6%, 95% CI 0.0%-3.4%) of 164 patients had an event (stent thrombosis causing myocardial infarction) during the first week after minor surgery; this rate was slightly higher than the background rate of ischemic events in the study population (exact mid P = .01). Clopidogrel use at 12 months was similar between patients who did and those who did not undergo minor surgery (65.2% vs 65.5%, P = .95). Conclusions In the EVENT Registry, minor surgery was performed in 2% of patients in the first year after DES implantation. The risk of stent thrombosis during the first week after surgery was increased slightly compared with background rates, but the absolute event rate was low (0.6%). (Am Heart J 2011;161:360-6.)

Noncardiac surgery after coronary stent implantation is associated with an increased risk of complications, including stent thrombosis and myocardial infarction

From the aVA North Texas Healthcare System and University of Texas Southwestern Medical Center at Dallas, Dallas, TX, bSaint Luke's Mid America Heart Institute, Kansas City, MO, cMethodist DeBakey Heart Center, Houston, TX, dHarvard Clinical Research Institute, Boston, MA, eUniversity of Oklahoma, Oklahoma City, OK, fVanderbilt University, Nashville, TN, and gGeisinger Clinic, Danville, PA. h On behalf of the EVENT Investigators. Funding for the Evaluation of Drug Eluting Stents and Ischemic Events (EVENT) Registry and its analysis was provided by grants from Millennium Pharmaceuticals and Schering Plough Inc. Submitted August 12, 2010; accepted September 28, 2010. Reprint requests: Peter B. Berger, MD, Geisinger Center for Health Research, 100 North Academy Avenue, MC 40-04, Danville, PA 17822. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.09.028

(MI).1 For bare metal stents (BMS), the risk appears to become very low if N4 to 6 weeks have elapsed after stent implantation before surgery.2,3 One of the reasons for the relatively short duration of excess risk is thought to be that BMS are rapidly covered with an endothelial layer; as the circulating blood elements are no longer exposed to metal, dual antiplatelet therapy is no longer considered necessary to prevent stent thrombosis, and clopidogrel can be safely discontinued for a surgical procedure.2,3 For drug-eluting stents (DES), the optimum delay between stenting and noncardiac surgery is unknown but may be considerably longer than for BMS. Delayed endothelialization of DES may place a patient at risk for stent thrombosis for months and possibly years after implantation.4 Although major surgery soon after stent implantation is known to be associated with an increased risk of stent thrombosis and other complications, few studies have

American Heart Journal Volume 161, Number 2

examined patients who underwent minor surgery shortly after stent placement, and none has focused exclusively on such patients.5,6 In the present study, we examined patients enrolled in the Evaluation of Drug-Eluting Stents and Ischemic Events (EVENT) Registry7,8 to determine the frequency of minor noncardiac surgery during the first year after placement of DES and the frequency of postoperative major adverse cardiac events.

Methods Patient population and end points The EVENT is a multicenter registry that prospectively enrolled unselected patients undergoing percutaneous coronary intervention (PCI) at 47 US centers7,8 between July 2004 and December 2007. Enrollment occurred in 4 discrete “waves” of approximately 2,500 patients; each wave was separated by approximately 1 year. The current analysis includes all patients who received a DES and were enrolled in any of the 4 waves. Of the 10,148 patients enrolled in EVENT, the following patients were excluded from this analysis: (a) patients who died (n = 19), had acute stent thrombosis (n = 17), or required coronary artery bypass graft surgery (n = 30) before hospital discharge; (b) patients in whom stent implantation was not successful or who only received BMS (n = 1,400); and (c) patients who had major noncardiac surgery during the followup period (n = 359). All remaining 8,323 patients were included in the current analysis. Indications for PCI included ST-segment elevation MI, acute coronary syndrome, as well as chronic stable angina or a positive functional test. Data on patient characteristics, presentation, and treatment were collected prospectively on standardized case report forms and submitted to a central data coordinating center. It was required by protocol that creatine kinase (CK) and creatine kinase-muscle brain fraction (CK-MB) levels were assessed at baseline (before the PCI) and every 8 hours after; a minimum of 2 samples were required after the procedure. Samples were assayed at each site's clinical laboratory using local reference values. If MI was suspected clinically at a later point, biomarkers were reassessed as indicated. The institutional review board of each hospital approved their participation in EVENT. Written informed consent was obtained from each participating patient. Telephone follow-up was performed at 6 and 12 months after PCI to ascertain the occurrence of major adverse cardiovascular events. Follow-up for N330 days or until patients had an event was available in 7,611 (91.4%) of 8,323 patients. The primary end point of the current study was the incidence of death, MI, or stent thrombosis in the 7 days after the surgical procedure. Myocardial infarction was defined as an elevation of CK-MB (or CK) at least 3 times the local upper limit of normal or persistent ST-segment elevation N1 mm in 2 contiguous electrocardiographic limb leads or N2 mm in 2 contiguous precordial leads. For patients whose CK-MB (or CK) level was elevated at baseline, an increase of at least 2-fold above the baseline value was required. Suspected MIs were adjudicated by 2 independent cardiologists. Of note, cardiac biomarkers were not routinely sampled in patients undergoing surgical procedures.

Brilakis et al 361

Definitions Minor surgical procedures were defined as invasive procedures not requiring a large surgical incision that would be expected to result in significant bleeding.

Statistical analysis Continuous parameters are reported as mean ± SD; discrete parameters are reported as percentages. Continuous parameters were compared using the t test or the Wilcoxon rank sum test, and discrete parameters were compared using the χ2 or the Fisher exact test. Event rates within 7 days of minor surgery were compared against the background rate among patients not receiving minor surgery, using the exact mid P value.9-11 All analyses were performed using SAS version 8.0 (SAS Institute, Cary, NC). The authors are solely responsible for the design and conduct of this study, all study analyses, and the drafting and editing of the paper and its final contents.

Results Study population and types of surgery During the first year after implantation of DES, minor surgery was performed in 164 (2.0%) of 8,323 patients. The timing of these procedures after the index PCI is displayed in Figure 1. The most commonly performed procedures were pacemaker/defibrillator implantation (46%) followed by eye surgery (17%), orthopedic (9%), dermatologic (8%), endovascular (6%), and gastrointestinal (5%) procedures. Each patient had a single procedure, except for 2 patients who underwent 2 minor surgical procedures. Baseline clinical characteristics The baseline clinical characteristics of the study population are shown in Table I. Compared with patients who did not undergo minor surgery, those who did were older, more likely to have risk factors for atherosclerosis, and were more likely to have both cardiac conditions (congestive heart failure, prior coronary artery bypass graft surgery, and a lower left ventricular ejection fraction) and noncardiac conditions (peripheral arterial disease, stroke, and renal failure). Angiographic and procedural characteristics The baseline angiographic characteristics and index PCI results are shown in Table II. Patients who went on to undergo a minor surgical procedure had fewer coronary lesions and fewer stents implanted but were more likely to undergo stenting of an unprotected left main coronary artery or a saphenous vein graft. Medications at discharge The medications prescribed at the time of discharge after implantation of DES are shown in Table III. Patients who subsequently required minor surgery were more

American Heart Journal February 2011

362 Brilakis et al

Figure 1

Frequency and timing of minor surgical procedures in the year after implantation of DES in the EVENT Registry.

Table I. Clinical characteristics of the study patients

Demographics Age (y)⁎ Male Medical history Diabetes Hypertension Hyperlipidemia Current smoking Renal dialysis Cerebrovascular accident Congestive heart failure Peripheral arterial disease MI PCI Coronary artery bypass graft Clinical characteristics No. of diseased vessels⁎ MI within 7 d before the PCI Indication for PCI Chronic stable angina Positive stress test Acute coronary syndrome Other Ejection fraction b25% 25%-35% 36%-50% N50% Unknown Baseline laboratory data Platelet count b100 (103/μL) Creatinine N2.0 mg/dL Creatinine clearance b30 mL/min ⁎ Data are presented as mean ± SD.

No minor surgery (n = 8159)

Minor surgery (n = 164)

P

64 ± 12 (8159) 69.2% (5648/8159)

68 ± 11 (164) 65.2% (107/164)

b.001 .275

33.9% (2759/8142) 77.8% (6326/8133) 74.0% (5984/8089) 24.3% (1970/8100) 1.5% (121/8141) 8.4% (683/8135) 9.2% (746/8114) 10.2% (820/8069) 37.0% (2982/8063) 37.5% (3043/8108) 20.8% (1691/8135)

43.9% (72/164) 86.5% (141/163) 77.6% (125/161) 19.0% (31/163) 4.3% (7/164) 14.1% (23/163) 22.7% (37/163) 19.8% (33/167) 43.3% (71/164) 42.7% (70/164) 32.3% (53/164)

.008 .008 .294 .118 .004 .010 b.001 b.001 .098 .178 b.001

1.7 ± 0.8 (7891) 14.5% (1170/8063)

1.9 ± 0.8 (157) 11.0% (18/164)

.041 .202

21.4% 39.4% 35.1% 4.1%

(1747/8159) (3215/8159) (2862/8159) (335/8159)

14.6% 32.3% 37.2% 15.8%

.036 .066 .574 .005 .002

2.0% 6.2% 21.0% 53.1% 17.7%

(166/8159) (503/8159) (1716/8159) (4331/8159) (1443/8159)

12.2% (20/164) 14.0% (23/164) 22.6% (37/164) 41.5% (68/164) 9.8% (16/164)

1.2% (93/7798) 3.0% (237/7786) 2.8% (217/7785)

(24/164) (53/164) (61/164) (26/164)

1.3% (2/158) 10.9% (17/156) 7.7% (12/156)

.880 b.001 b.001

American Heart Journal Volume 161, Number 2

Brilakis et al 363

Table II. Angiographic and procedural characteristics and procedural outcomes of the study population

Lesions per patient⁎ Most severe lesion classification A B1 B2 C Target vessel Left anterior descending artery Circumflex Right coronary artery Left main coronary artery Saphenous vein graft Unprotected left main Preprocedure Thrombolysis In Myocardial Infarction flow 0 1 2 3 Bifurcation Thrombus Postprocedure Thrombolysis In Myocardial Infarction flow 0 1 2 3 Embolic protection No. of lesions treated 1 2 3 4 Stents per patient⁎ DES type Sirolimus-eluting Paclitaxel-eluting No. of vessels treated per patient⁎ Total stent length (mm)⁎

No minor surgery (n = 8159)

Minor surgery (n = 164)

P

1.40 ± 0.66 (8159)

1.29 ± 0.53 (164)

.006 .824

11.1% (904/8108) 30.2% (2446/8108) 34.5% (2800/8108) 24.1% (1958/8108)

16.8% 24.2% 32.9% 26.1%

(27/161) (39/161) (53/161) (42/161)

42.7% (3485/8159) 28.5% (2328/8159) 36.5% (2974/8159) 2.2% (183/8159) 7.3% (599/8159) 0.7% (60/8158)

45.1% 22.6% 31.7% 3.7% 11.6% 3.0%

(74/164) (37/164) (52/164) (6/164) (19/164) (5/164)

6.8% 3.8% 8.5% 80.8% 12.3% 6.8%

(555/8104) (312/8104) (692/8104) (6545/8104) (1001/8159) (555/8158)

5.0% (8/160) 4.4% (7/160) 9.4% (15/160) 81.3% (130/160) 11.0% (18/164) 6.7% (11/164)

0.2% (14/8119) 0.0% (3/8119) 0.4% (33/8119) 99.4% (8069/8119) 3.4% (277/8154)

0.0% (0/161) 0.0% (0/161) 1.9% (3/161) 98.1% (158/161) 3.6% (6/168)

68.1% (5555/8159) 24.9% (2029/8159) 5.8% (476/8159) 1.2% (96/8159) 1.65 ± 0.92 (8159)

75.0% (123/164) 21.3% (35/164) 3.7% (6/164) 0.0% (0/164) 1.52 ± 0.79 (164)

.037

59.5% (4851/8159) 42.5% (3464/8159) 1.16 ± 0.38 (8159) 30.77 ± 20.12 (8159)

56.1% (92/164) 45.7% (75/164) 1.12 ± 0.33 (164) 28.79 ± 20.75 (164)

.386 .401 .182 .211

.577 .097 .220 .278 .049 b.001 .801

.617 .962 .050

.802 .043

⁎ Data are presented as mean ± SD.

likely to receive warfarin and nitrates at discharge from the index PCI procedure. Nearly all patients in both groups received a thienopyridine and aspirin at discharge from the index PCI.

postoperative death, MI, or stent thrombosis (0.6%, 95% CI 0.0%-3.4%) was significantly higher than the risk in any average 7-day period in patients who did not undergo minor noncardiac surgery (exact mid P = .044).

Adverse events after surgery The cumulative incidence of adverse events throughout the 12-month follow-up period was similar between patients who did and did not undergo minor surgery (Table IV). Only 1 (0.6%, 95% CI 0.0%-3.4%, exact mid P = .01) of 164 patients had an event (stent thrombosis resulting in MI) during the first week after minor surgery performed 8 months after implantation of DES. The risk for postoperative stent thrombosis in the 7 days after minor surgery, although low, was significantly higher than the background risk for a comparable 7-day period among patients who did not undergo minor noncardiac surgery (exact mid P = .01). Similarly, the risk for

Antiplatelet medication The details of the management of antithrombotic therapy before and after minor surgery were not recorded in the first 3 waves of the EVENT Registry; the case report forms were adjusted to collect such information for the fourth wave of enrollment (Table V). Aspirin and clopidogrel were discontinued in the perioperative period in approximately one third of patients who underwent surgery for 1 to 2 weeks and were restarted postoperatively in nearly all patients. Similar proportions of patients who did and did not undergo minor noncardiac surgery were receiving clopidogrel at 12 months (64.9% vs 65.5%, P = .63).

American Heart Journal February 2011

364 Brilakis et al

Table III. Medications prescribed at discharge after coronary stenting No minor surgery (n = 8159) Aspirin Warfarin Statin Nonstatin lipid lowering β-Blocker Calcium-channel blocker Angiotensin-converting enzyme inhibitor Angiotensin receptor blocker Nitrates Ticlopidine⁎ Clopidogrel⁎

97.5% 5.0% 82.4% 17.2%

Minor surgery (n = 164)

(7875/8081) 95.1% (154/162) (405/8081) 9.9% (16/162) (6657/8081) 83.3% (135/162) (1388/8081) 22.2% (36/162)

.058 .005 .752 .093

47.8% (3866/8081) 51.2% (83/162)

.392

12.2% (986/8081)

.087

15.2% (1232/8081) 24.1% (39/162) .002 1.0% (61/5942) 0.8% (1/126) .797 96.9% (5757/5942) 98.4% (124/126) .327

⁎ The frequency with which ticlopidine and clopidogrel were prescribed at discharge was only assessed in waves 2-4 of EVENT.

Table IV. Frequency of adverse events in the year after PCI among study patients who did and did not undergo minor surgery No minor surgery Minor surgery (n = 8159) (n = 164) Cardiac death Cardiac death/MI Stent thrombosis Cardiac death/MI/stent thrombosis

0.8% (64/8159) 2.5% (200/8159) 0.6% (49/8159) 2.6% (215/8159)

0.0% 2.4% 1.2% 2.4%

(0/164) (4/164) (2/164) (4/164)

Minor noncardiac surgery patients (n = 39)

P

78.2% (6316/8081) 74.1% (120/162) .213 17.5% (1416/8081) 22.8% (37/162) .079

16.7% (27/162)

Table V. Management of antiplatelet therapy in wave 4 of the EVENT Registry

P .64 1.00 .27 1.00

Discussion The 2 most important findings of this study are that minor noncardiac surgery was performed in approximately 2% of patients in the first year after implantation of DES and that the incidence of stent thrombosis and other cardiac events in the week after surgery was low (0.6%).

Background Most previously published studies assessing the safety of surgery in patients who have undergone PCI have described the outcomes of patients undergoing major noncardiac surgery.5,6,12-19 The current study is the first that we are aware of to specifically describe the frequency and outcomes of a large cohort of patients undergoing minor surgery; in it, we included only patients who had received DES. Nearly half of the surgical procedures involved implantation of a pacemaker or defibrillator.

Aspirin discontinued before minor surgery Days off aspirin Mean ± SD (n) Range (minimum-maximum) Clopidogrel discontinued before minor surgery Days off clopidogrel Mean ± SD (n) Range (minimum-maximum) Aspirin and clopidogrel both discontinued before minor surgery

33.3% (13/39) 13.38 ± 16.87 (13) 2.00-60.00 28.2% (11/39) 7.18 ± 7.95 (11) 2.00-30.00 23.1% (9/39)

Patients in whom both aspirin and clopidogrel were discontinued before minor surgery were included in all 3 categories described above.

Although the risk of a postoperative adverse event is believed to be very low when even major surgery is performed N4 to 6 weeks after implantation of BMS,2,3 the reported risk associated with major noncardiac surgery after implantation of DES varies widely. Reported risks range between 0% and 22%, with most studies reporting a risk ≤6%.5,6,12-16 Schouten et al12 reported a 22% postoperative event rate for major noncardiac surgery performed in the 3 to 6 months after implantation of a sirolimus- or paclitaxel-eluting stent when antiplatetelet therapy was discontinued (2/9 patients); none of 14 patients had an event if dual antiplatelet therapy was continued. In that series, a major perioperative event occurred in only 1 (1%) of 76 patients who underwent major noncardiac surgery N3 to 6 months after implantation of DES. Brotman et al13 reported only one postoperative MI (2.2%) among 45 patients undergoing noncardiac surgery within 6 months of implantation of DES. Compton et al5 reported no stent thromboses among 38 patients who had DES; Kim et al14 reported stent thrombosis in 3 (2.2%) of 138 patients who had DES; Godet et al,15 in 2 (2%) of 96 patients; and Rhee et al,16 in 7 (5%) of 141 patients. Rabbitts et al6 reported an adverse cardiac event rate of 5.4% in 520 patients who had DES; Anwaruddin et al,17 in 56 (9%) of 606 surgeries; and Assali et al,19 in 6 (7.7%) of 78 patients. Berger et al20 reported that, in the EVENT Registry, major noncardiac surgery was performed in 204 (4.4%) of 4,637 patients within the first year (median time to surgery 179 days [25th quartile 112 days, 75th quartile 266 days, range 13-360 days]) and that the incidence of death, MI, or stent thrombosis in the first 7 days post surgery was 1.9%. In most12,17,18 (although not all) studies, a longer time interval between DES implantation and surgery has been associated with a lower risk of perioperative complications. Accordingly, the current American College of Cardiology/American Heart

American Heart Journal Volume 161, Number 2

Association guidelines recommend delaying noncardiac surgery for 12 months after implantation of DES.21,22

Minor surgery Most of these studies excluded patients undergoing minor surgical procedures; studies that did include a few patients who underwent minor surgery did not present their outcomes separately.6 The only one that did, reported no events among 18 patients who underwent a minor surgical procedure a median of 297 days after implantation of DES.5 In the current study of 164 patients, the postoperative risk after minor surgery was low, 0.6%, although it was higher than the background event rate among patients not undergoing minor surgery. Compared with the event rate after major surgery, the lower event rate after minor surgery could be because of several reasons, including less activation of the coagulation cascade during the immediate postoperative period,1 fewer hemodynamic perturbations during the operation and postoperative period, continuation of antiplatelet and anticoagulation regimens throughout the perioperative period,22 and a shorter duration of discontinuation of antiplatelet therapy before the surgical procedure and more rapid reinitiation of antiplatelet therapy after the surgical procedure. Additional reasons include a fairly long delay to minor surgery (performed after 3 months in most patients, as shown in the figure) and good medical therapy with β-blockers, statins, and when feasible, antiplatelet agents. The lower risk of perioperative complications after minor surgery is reassuring, given the higher risk profile of such patients. Our findings should provide reassurance that minor surgical procedures that cannot be delayed can be performed after implantation of DES with a low risk of major complications. Limitations Our study has several important limitations. The periprocedural management of antithrombotic therapy was not recorded in the first 3 waves of EVENT; whether dual antiplatelet therapy was managed in a similar fashion in waves 1 to 3 of EVENT as in wave 4 (when such data were collected) is unknown. Little is known about the timing of reinitiation of antiplatelet therapy after the procedure and whether a loading dose was administered. Only major postoperative complications (MI, stent thrombosis, and death) were recorded; minor complications, such as hematomas or bleeding at the surgical site, were not recorded in EVENT. It is possible that an adverse event might have occurred before minor surgery because of discontinuation of antiplatelet therapy that resulted in cancellation of the planned minor surgical procedure; this appears to be a rare event, however, and this potential bias exists in all retrospective studies examining the risk of surgery after

Brilakis et al 365

a PCI procedure. Finally, although our study is the largest of its kind, it remains underpowered for defining the incidence of infrequent events, such as stent thrombosis, with precision; as a result, the 95% CIs of stent thrombosis in the first week after minor surgery are wide (0.0%-3.4%).

Conclusions In contemporary practice, approximately 2% of patients undergo ≥1 minor surgical procedures during the year after implantation of DES. Among such patients, there is a small but definite increase in the risk of thrombotic events during the early postoperative period. Nonetheless, the overall perioperative complication rate appears to be acceptable.

Disclosures Dr. Brilakis: Speaker honoraria from St Jude Medical (Minneapolis, MN); consulting fees from Medicure (Winnipeg, Manitoba, CANADA); research support from Abbott Vascular (Santa Clara, CA); salary support from Medtronic (Minneapolis, MN) (spouse). Dr. Cohen: Research grant support from Boston Scientific (Natick, MA), Abbott Vascular (Santa Clara, CA), Edwards Lifesciences (Irvine, CA), Eli Lilly (Indianapolis, IN), Daiichi-Sankyo (Parsippany, NJ), Merck/ Schering Plough (Whitehouse Station, NJ). Consultant to Cordis (Warren, NJ), Medtronic (Santa Rosa, CA), Eli Lilly (Indianapolis, IN), Merck/Schering Plough (Whitehouse Station, NJ), and the Medicines Company (Parsippany, NJ). Dr. Kleiman: research grants from Merck-Schering Plough (Whitehouse Station, NJ), Bristol Myers Squibb (New York, NY), Eli Lilly (Indianapolis, IN); consultant to Schering Plough (Whitehouse Station, NJ) and Bristol Myers Squibb (New York, NY). Dr. Pencina: none Dr. Nassif: none Dr. Saucedo: research grants and advisory board for Merck (Whitehouse Station, NJ), Eli Lilly (Indianapolis, IN), and Daichi Sankyo (Parsippany, NJ). Dr. Piana: none Dr. Banerjee: Speaker honoraria from St. Jude Medical (Minneapolis, MN), Medtronic (Santa Rosa, CA), and Johnson & Johnson (Warren New Jersey) and research support from Boston Scientific (Natick, MA) and The Medicines Company (Parsippany, NJ). Ms Keyes : none Dr. Yen : none Dr. Berger: Consultant to AstraZeneca (Wilmington, DE), Boehringer Ingelheim (Ridgefield, CT), Eli Lilly (Indianapolis, IN), Daiichi-Sankyo (Parsippany, NJ), Medicure (Winnipeg, Manitoba, CANADA), and Ortho McNeil (Raritan, NJ) (each for less than $10,000). Equity in

366 Brilakis et al

Lumen, Inc. (Maple Grove, MN) (greater than $10,000). Research funding to Geisinger Clinic (Danville, PA) for studies on which Dr. Berger is the PI: Thrombovision (Houston, TX), Helena (Beaumont, TX), Accumetrics (San Diego, CA), AstraZeneca (Wilmington, DE), Haemoscope (Niles, IL), The Medicines Company (Parsippany, NJ), Corgenix/Aspirinworks (Broomfield, CO), and Eli Lilly (Indianapolis, IN), Daiichi-Sankyo (Parsippany, NJ) (all for more than $10,000).

References 1. Brilakis E, Banerjee S, Berger P. Perioperative management of coronary stents. J Am Coll Cardiol 2007;49:2145-50. 2. Kaluza GL, Joseph J, Lee JR, et al. Catastrophic outcomes of noncardiac surgery soon after coronary stenting. J Am Coll Cardiol 2000;35:1288-94. 3. Wilson SH, Fasseas P, Orford JL, et al. Clinical outcome of patients undergoing non-cardiac surgery in the two months following coronary stenting. J Am Coll Cardiol 2003;42:234-40. 4. Brilakis ES, Banerjee S, Berger PB. The risk of drug-eluting stent thrombosis with noncardiac surgery. Curr Cardiol Rep 2007;9:406-11. 5. Compton PA, Zankar AA, Adesanya AO, et al. Risk of noncardiac surgery after coronary drug-eluting stent implantation. Am J Cardiol 2006;98:1212-3. 6. Rabbitts JA, Nuttall GA, Brown MJ, et al. Cardiac risk of noncardiac surgery after percutaneous coronary intervention with drug-eluting stents. Anesthesiology 2008;109:596-604. 7. Jacob S, Cohen DJ, Massaro J, et al. Design of a registry to characterize “real-world” outcomes of percutaneous coronary revascularization in the drug-eluting stent era. Am Heart J 2005;150: 887-92. 8. Win HK, Caldera AE, Maresh K, et al. Clinical outcomes and stent thrombosis following off-label use of drug-eluting stents. JAMA 2007; 297:2001-9. 9. Lancaster HO. Significance test in discrete distributions (corrections 57:919). J Amer Statistical Assoc 1961;56:223-34. 10. Lydersen S, Fagerland MW, Laake P. Recommended tests for association in 2 × 2 tables. Stat Med 2009;28:1159-75. 11. Agresti A. Exact inference for categorical data: recent advances and continuing controversies. Stat Med 2001;20:2709-22. 12. Schouten O, van Domburg RT, Bax JJ, et al. Noncardiac surgery after coronary stenting: early surgery and interruption of antiplatelet therapy are associated with an increase in major adverse cardiac events. J Am Coll Cardiol 2007;49:122-4.

American Heart Journal February 2011

13. Brotman DJ, Bakhru M, Saber W, et al. Discontinuation of antiplatelet therapy prior to low-risk noncardiac surgery in patients with drugeluting stents: a retrospective cohort study. J Hosp Med 2007;2: 378-84. 14. Kim HL, Park KW, Kwak JJ, et al. Stent-related cardiac events after non-cardiac surgery: drug-eluting stent vs. bare metal stent. Int J Cardiol 2008;123:353-4. 15. Godet G, Le Manach Y, Lesache F, et al. Drug-eluting stent thrombosis in patients undergoing non-cardiac surgery: is it always a problem? Br J Anaesth 2008;100:472-7. 16. Rhee SJ, Yun KH, Lee SR, et al. Drug-eluting stent thrombosis during perioperative period. Int Heart J 2008;49:135-42. 17. Anwaruddin S, Askari AT, Saudye H, et al. Characterization of postoperative risk associated with prior drug-eluting stent use. JACC Cardiovasc Interv 2009;2:542-9. 18. van Kuijk JP, Flu WJ, Schouten O, et al. Timing of noncardiac surgery after coronary artery stenting with bare metal or drug-eluting stents. Am J Cardiol 2009;104:1229-34. 19. Assali A, Vaknin-Assa H, Lev E, et al. The risk of cardiac complications following noncardiac surgery in patients with drug eluting stents implanted at least six months before surgery. Catheter Cardiovasc Interv 2009;74:837-43. 20. Berger PB, Kleiman NS, Pencina MJ, et al. Frequency of major noncardiac surgery and subsequent adverse events in the year after drugeluting stent placement: results from the EVENT Registry. JACC Cardiovasc Interv 2010;3:920-7. 21. Fleisher LA, Beckman JA, Brown KA, et al. ACC/AHA 2007 guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) developed in collaboration with the American Society of Echocardiography, American Society of Nuclear Cardiology, Heart Rhythm Society, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, and Society for Vascular Surgery. J Am Coll Cardiol 2007;50:e159-241. 22. Grines CL, Bonow RO, Casey Jr DE, et al. Prevention of premature discontinuation of dual antiplatelet therapy in patients with coronary artery stents: a science advisory from the American Heart Association, American College of Cardiology, Society for Cardiovascular Angiography and Interventions, American College of Surgeons, and American Dental Association, with representation from the American College of Physicians. J Am Coll Cardiol 2007; 49:734-9.

Qualitative assessment of neointimal tissue after drug-eluting stent implantation: Comparison between follow-up optical coherence tomography and intravascular ultrasound Sung Woo Kwon, MD, a Byeong-Keuk Kim, MD, a Tae-Hoon Kim, MD, a Jung-Sun Kim, MD, a Young-Guk Ko, MD, a Donghoon Choi, MD, a Yangsoo Jang, MD, a,b and Myeong-Ki Hong, MD a,b Seoul, Korea

Background Characterization of neointimal tissue is essential to understand the pathophysiology of in-stent restenosis after drug-eluting stent (DES) implantation. We compared the morphological characteristics of neointimal tissue as assessed by optical coherence tomography (OCT) and intravascular ultrasound (IVUS) in patients treated with DES. Methods A total of 243 patients (250 lesions) underwent follow-up OCT and IVUS after DES implantation. Results Mean time interval from DES implantation to follow-up OCT/IVUS was 12.0 ± 9.3 (range 2.8-68.5)

months. Percent neointimal hyperplasia (NIH) cross-sectional area (CSA) was calculated as (NIH CSA/stent CSA) × 100 for receiveroperating characteristic analysis of NIH detection by IVUS; the optimal cutoff value of percent NIH CSA was 14.7%, as determined by OCT (sensitivity 0.887, specificity 0.790). Neointimal hyperplasia was detected by both OCT and IVUS in 121 of 250 lesions and categorized as homogenous (OCT n = 74, IVUS n = 107), heterogeneous (OCT n = 34, IVUS n = 4), or layered (OCT n = 13, IVUS n = 10). Of the 121 NIH lesions, nonhomogenous NIH was detected in 14 (11.6%) by IVUS and 47 (38.8%) by OCT. Optical coherence tomography and IVUS assessments of NIH morphology showed a moderate correlation (P < .001, r = 0.455); however, assessments differed in 37 (30.6%) of 121 lesions.

Conclusion Optical coherence tomography–assessed NIH morphology might be different from that by IVUS in about 30% of the lesions that were treated with DES implantation. (Am Heart J 2011;161:367-72.)

Neointimal growth after percutaneous coronary intervention creates a double-edged sword; restenosis requires repeated coronary intervention, but incomplete reendothelialization is prone to stent thrombosis.1-4 Although both features are considered to have equivalent clinical implications, the former has been the major therapeutic challenge with bare-metal stents, whereas the latter has become a concern with drug-eluting stents (DESs). Intravascular ultrasound (IVUS) has been used extensively to assess neointimal hyperplasia (NIH) after percutaneous coronary intervention.5 However, the evaluation of

From the aDivision of Cardiology, Severance Cardiovascular Hospital, Yonsei University College of Medicine, Seoul, Korea, and bSeverance Biomedical Science Institute, Yonsei University College of Medicine, Seoul, Korea. The first 2 authors contributed equally to this article. Submitted August 10, 2010; accepted October 18, 2010. Reprint requests: Myeong-Ki Hong, MD, PhD, Division of Cardiology, Yonsei Cardiovascular Center and Severance Biomedical Science Institute, Yonsei University College of Medicine, 250 Seongsanno, Seodaemun-gu, Seoul 120-752, Korea. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.026

NIH has been confined to quantitative assessment; the possibility of qualitative NIH assessment has not been sufficiently recognized, probably because most NIH lesions detected by IVUS appear homogeneous.6,7 Intracoronary optical coherence tomography (OCT) technology provides a higher axial resolution than IVUS.8,9 A recent OCT study revealed the diverse characteristics of NIH after stent implantation.10 The aim of the present study was to compare the tissue characteristics of NIH as assessed by OCT and IVUS in patients who underwent DES implantation.

Methods The Yonsei OCT Registry lists patients who have undergone follow-up OCT examination after stent implantation. General exclusion criteria for the follow-up OCT study were as follows: (1) untreated significant left main coronary artery disease, (2) apparent congestive heart failure, (3) renal insufficiency (baseline creatinine ≥2.0 mg/dL), and (4) lesions unsuitable for OCT procedure (vessel size ≥3.5 mm or lesions within 10 mm of the ostium of each major epicardial artery). A total of 243 nonconsecutive patients with 250 lesions who underwent both follow-up OCT and IVUS examinations in the same lesions were

368 Kwon et al

American Heart Journal February 2011

Figure 1

Neointimal tissue characteristics were assessed by OCT and IVUS. The upper panels show IVUS images, and the lower panels show corresponding OCT images. A and D show homogenous neointimal coverage, B and E show heterogeneous neointimal coverage, and C and F show layered neointimal coverage.

selected from the Yonsei OCT Registry database. Mean time interval from DES implantation to follow-up OCT/IVUS was 12.0 ± 9.3 (range 2.8-68.5) months. The DESs were coated with sirolimus (Cypher, Cordis, Miami Lakes, FL) (n = 98), paclitaxel (Taxus, Boston Scientific, Natick, MA) (n = 51), zotarolimus (Endeavor Splint; Medtronic, Santa Rosa, CA) (n = 74), or everolimus (Xience, Abbott Vascular, Abbott Park, IL) (n = 27). Inclusion criteria for the present study were (1) de novo lesion with N50% diameter stenosis at index procedure and (2) both IVUS and OCT imaging at follow-up coronary angiography. Exclusion criteria were (1) poor OCT image quality, (2) baremetal stent implantation, and (3) stented segments overlapped with bare-metal stent or other types of DES. Stents were implanted using current conventional techniques without other complications. Unfractionated heparin was administered as an initial bolus of 100 IU/kg, and additional boluses were administered during the procedure to achieve an activated clotting time of 250 to 300 seconds. Dual antiplatelet therapy (aspirin and clopidogrel) was administered to all patients until the follow-up angiogram and OCT were performed. All patients provided written informed consent, and the institutional review boards at our center approved this study. The OCT imaging was performed with a conventional OCT system (Model M2 Cardiology Imaging System, LightLab Imaging Inc, Westford, MA) with a motorized pullback rate of 1.0 mm/s. The occlusion catheter was positioned proximal to the stent, and a 0.014-in wire-type imaging catheter (ImageWire; LightLab Imaging Inc) was positioned distal to the stent. During image acquisition, the occlusion balloon (Helios, Avantec Vascular Corp, Sunnyvale, CA) was inflated from 0.4 to 0.6 atm, and Ringer's lactate solution was infused at 0.5 to 1.0 mL/s.

The imaging wire was pulled from distal to proximal, and continuous images were acquired and stored digitally for subsequent analysis.11 The segment with minimal lumen cross-sectional area (CSA) and largest percent NIH CSA was selected for qualitative and quantitative analyses. The stent and luminal CSA were measured, and NIH CSA was calculated as the difference between stent CSA and luminal CSA. Percent NIH CSA was then calculated as (NIH CSA/stent CSA) × 100. Neointima tissue characteristics were categorized as homogenous, heterogeneous, or layered; homogeneous neointima exhibits uniform optical properties without focal variation in the backscattering pattern, heterogeneous neointima exhibits focally changing optical properties and various backscattering patterns, and layered neointima consists of concentric layers with different optical properties (Figure 1).10 All OCT images were independently analyzed by 2 investigators who were unaware of the patients' clinical information. Differences between the observer's analyses were resolved by consensus. After intracoronary administration of 0.2 mg nitroglycerin, follow-up IVUS imaging was performed with a motorized transducer pullback system (0.5 mm/s) and with a commercial scanner (SCIMED, Minneapolis, MN) consisting of a rotating transducer (30 or 40 MHz) within a 3.2- or a 2.6F catheter imaging sheath. The imaging probe was positioned distal to the target lesions and pulled from distal to proximal; continuous images were acquired and stored digitally for subsequent off-line analysis with echoPlaque software (INDEC Medical Systems, Santa Clara, CA). Follow-up IVUS study was initially performed, and then, OCT was performed with additional intracoronary administration of 0.2 mg nitroglycerin after the end of the IVUS procedure. Both IVUS and OCT procedures were performed

American Heart Journal Volume 161, Number 2

Kwon et al 369

Table I. Baseline, angiographic, and OCT characteristics of patients according to detection of NIH by IVUS Variables Baseline characteristics Age, y Male gender Diabetes mellitus Hypertension Dyslipidemia Current smoker Previous myocardial infarction Clinical diagnosis Stable angina Unstable angina or non-ST elevation myocardial infarction ST-elevation myocardial infarction Angiographic characteristics Extent of coronary artery disease 1-vessel disease 2-vessel disease 3-vessel disease Target vessel Left anterior descending artery Left circumflex artery Right coronary artery Stent diameter, mm Stent length, mm Stent type Sirolimus-eluting stent Paclitaxel-eluting stent Zotarolimus-eluting stent Everolimus-eluting stent OCT Characteristics Time to OCT, m Percent NIH CSA, %

NIH not detected by IVUS (123 patients, 129 lesions)

NIH detected by IVUS (120 patients, 121 lesions)

60.8 ± 9.5 76/123 (61.8%) 45/123 (36.6%) 67/123 (54.5%) 40/123 (32.5%) 34/123 (28.3%) 5/123 (4.1%)

61.7 ± 9.1 76/120 (63.3%) 37/120 (30.8%) 72/120 (60.0%) 42/120 (35.0%) 44/120 (36.7%) 5/120 (4.2%)

.5 .9 .4 .4 .7 .17 1.0

42/123 (34.1%) 56/123 (45.5%)

52/120 (43.3%) 46/120 (38.3%)

.3

25/123 (20.3%)

22/120 (18.3%)

36/123 (29.3%) 62/123 (50.4%) 25/123 (20.3%)

44/120 (36.7%) 48/120 (40.0%) 28/120 (23.3%)

.24

70/129 (54.3%) 34/129 (26.4%) 25/129 (19.4%) 2.9 ± 0.3 25.4 ± 5.7

62/121 (51.2%) 22/121 (18.2%) 37/121 (30.6%) 3.0 ± 0.3 25.4 ± 6.4

.110

77/129 13/129 17/129 22/129

21/121 38/121 57/121 5/121

(59.7%) (10.1%) (13.2%) (17.1%)

11.5 ± 7.9 (2.8-58.6) 11.2 ± 7.6

(17.4%) (27.3%) (47.1%) (4.1%)

12.5 ± 10.6 (2.8-68.5) 27.1 ± 10.7

P

.096 .8 <.0001

.4 <.0001

Data are expressed as mean ± SD or n (%).

without any complications and required additional fluoroscopic time and radiation exposure for <3 minutes. Similar to the OCT assessment, the neointimal tissue characteristics revealed by IVUS imaging were also categorized as homogenous, heterogeneous, or layered; homogeneous neointima exhibits uniform hyperechogenic properties, heterogeneous neointima exhibits focal hypodensity within the hyperechogenic regions, and layered neointima consists of concentric layers with different echogenic properties (Figure 1). All IVUS images were also independently analyzed by 2 investigators who were unaware of the patients' clinical information. Both IVUS and OCT images were acquired with a motorized automatic pullback system, and all OCT cross sections were analyzed. Matching of OCT frames to corresponding IVUS images was performed using stent edges as anatomical landmarks, measuring the distance from the edge of the stent. All analyses were performed with SPSS software (SPSS Inc, Chicago, IL). Continuous data were analyzed with Student t test and expressed as mean ± SD. Categorical data were compared with the χ2 test and expressed as number and percentage. The κ test was used to assess interobserver and intraobserver variability for the qualitative OCT and IVUS assessments. For

IVUS detection of NIH, the optimal cutoff value of OCTmeasured percent NIH CSA was determined by receiveroperating characteristic (ROC) analysis. The relationship between NIH characteristics assessed by OCT and IVUS was evaluated by Cramer's V measure of nominal correlation. All statistical tests were 2 sided, and differences were considered statistically significant at P < .05. This study was partly supported by grants from the Korea Healthcare Technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (No. A085012 and A102064); the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea (No. A085136); and the Cardiovascular Research Center, Seoul, Korea. The authors are solely responsible for the design and conduct of this study, all study analyses, and the drafting and editing of the paper and its final contents.

Results Baseline, angiographic, and OCT characteristics of patients are listed in Table I. Neointimal hyperplasia was

American Heart Journal February 2011

370 Kwon et al

Figure 2

and in 12 (33.3%) of 36 lesions by OCT in group 2 (Cramer's V = 0.564), and in 9 (20.0%) of 45 lesions by IVUS and in 26 (57.8%) of 45 lesions by OCT in group 3 (Cramer's V = 0.418). The relationship between percent NIH CSA and NIH morphology in OCT and IVUS is shown in Table IV. Interobserver variability was κ = 0.90 for OCT assessment and κ = 0.86 for IVUS. Intraobserver variability was κ = 0.92 for OCT and κ = 0.88 for IVUS.

Discussion

The optimal cutoff value of OCT-measured percent NIH CSA for detection of NIH by IVUS was determined by ROC analysis. AUC, area under the curve.

detected by both OCT and IVUS in 120 patients (121 lesions) but detected by OCT only in the remaining 123 patients (129 lesions). The types of DES used differed between these 2 groups; sirolimus-eluting stents were more frequently used in the group in which NIH was not detected by IVUS (59.7%), and zotarolimus-eluting stents were more frequently used in the group in which NIH was detected by IVUS (47.1%, P < .0001). For detection of NIH by IVUS, ROC analysis determined the optimal cutoff value of percent NIH CSA to be 14.7% as measured by OCT (sensitivity 0.887, specificity 0.790, area under the curve 0.905, P < .001) (Figure 2). Figure 3 shows a representative assessment of NIH detectable by OCT alone. Optical coherence tomography and IVUS assessments of NIH characteristics are shown in Table II. Neointimal hyperplasia was categorized as nonhomogeneous in 14 (11.6%) of 121 lesions by IVUS and in 47 (38.8%) of 121 lesions by OCT. Optical coherence tomography and IVUS assessments differed in 37 lesions (30.6%). Cramer's V nominal correlation test showed a moderate correlation between OCT and IVUS assessments of NIH morphology (P < .001, r = 0.455). The 121 lesions were arbitrarily divided into 3 subgroups by OCT-measured percent NIH CSA (Table III; group 1, percent NIH CSA <20%; group 2, 20% ≤ percent NIH CSA <30%; group 3, percent NIH CSA ≥30%). Neointimal hyperplasia appearance was categorized as nonhomogenous in 1 (2.5%) of 40 lesions by IVUS and in 9 (22.5%) of 40 lesions by OCT in group 1 (Cramer's V = 0.698), in 4 (11.1%) of 36 lesions by IVUS

This present study comparing OCT and IVUS showed that qualitative assessment of NIH by IVUS is possible when percent NIH CSA after DES implantation is at least 14.7% (as determined by OCT). Nonhomogenous NIH was detected in 11.6% of the lesions by IVUS and in 38.8% of the lesions by OCT. Although OCT and IVUS assessments of NIH morphology are moderately correlated, OCTassessed NIH morphology differed from IVUS assessment in approximately 30% of the lesions treated by DES. Intravascular ultrasound and OCT are both invasive diagnostic modalities that assess NIH better than invasive coronary angiography. Intravascular ultrasound has been considered the criterion standard for NIH detection after coronary stent implantation.5 Although IVUS is valuable for evaluating in-stent restenosis and for determining the need for coronary revascularization, its relatively low resolution (100-150 μm) is unable to detect small degrees of neointimal stent coverage.7 The axial resolution of OCT is about 10 times higher than that of IVUS, enabling more sensitive detection of stent strut coverage.8 Because of its higher spatial resolution and better visualization of intracoronary features, a recent OCT study reported that NIH has diverse tissue characteristics that can be categorized as homogenous, heterogeneous, or layered.10 Quantitative and qualitative evaluation of minimal NIH by IVUS is difficult; therefore, it would be useful to know the detection sensitivity of IVUS for NIH. In the present study, we showed that the cutoff value for percent NIH CSA was 14.7% for detecting NIH by IVUS, indicating that IVUS is unable to detect early stages of reendothelialization over stent struts. Although an unusual NIH appearance (ie, black holes) is sometimes detected by IVUS after DES implantation,12 IVUS assessment categorizes most NIH as homogenous.6,7 In contrast, OCT allows more diverse tissue characterization. In the present study, a nonhomogenous NIH appearance was visualized in 38.8% of lesions by OCT but in only 11.6% of lesions by IVUS. Thus, OCT and IVUS assessments differed in 30.6% of lesions, possibly because of resolution differences. There was a moderate correlation between OCT and IVUS assessments of NIH morphology in overall 121 lesions; however, subgroup analysis based on percent NIH CSA suggested that the correlation between OCT and IVUS assessments decreased as the percent NIH CSA increased. In addition,

American Heart Journal Volume 161, Number 2

Kwon et al 371

Figure 3

The representative example of NIH that was not detected by IVUS (A) but was visualized by OCT (B and C are identical images).

Table II. Optical coherence tomography and IVUS assessment of neointimal tissue characteristics OCT IVUS Homogenous Heterogeneous Layered Total

Homogenous

Heterogeneous

Layered

Total

74 0 0 74

28 3 3 34

5 1 7 13

107 4 10 121

P < .0001, r = 0.455

Table III. Subgroup analysis of OCT and IVUS assessments of neointimal tissue characteristics according to degree of NIH OCT IVUS

Homogenous

Heterogeneous

Layered

Total

(a) Group 1 (percent NIH CSA <20%) Homogenous 31 Heterogeneous 0 Layered 0 Total 31

7 0 0 7

1 0 1 2

39 0 1 40

P < .0001, r = 0.698

(b) Group 2 (20% ≤ percent NIH CSA <30%) Homogenous 24 Heterogeneous 0 Layered 0 Total 24

7 1 0 8

1 1 2 4

32 2 2 36

P < .0001, r = 0.564

14 2 3 19

3 0 4 7

36 2 7 45

P = .003, r = 0.418

(c) Group 3 (percent NIH CSA ≥30%) Homogenous 19 Heterogeneous 0 Layered 0 Total 19

the proportion of a nonhomogenous NIH as determined by both IVUS and OCT increased as percent NIH CSA increased. These results are similar to those of previous OCT studies.10,13 A recent pathologic study reported that DES therapy may increase atherosclerosis while reducing

restenosis. Atherosclerotic changes (eg, foamy macrophage infiltration and early necrotic core formation) were observed in N40% of patients 9 months after sirolimuseluting stent implantation. In contrast, atherosclerotic changes do not appear until 2 years after bare-metal stent

American Heart Journal February 2011

372 Kwon et al

Table IV. Relationship between percent NIH CSA and NIH morphology Percent NIH CSA <20% (n = 40)

20% ≤ percent NIH CSA <30% (n = 36)

Percent NIH CSA ≥30% (n = 45)

39 0 1

32 2 2

36 2 7

.107

31 7 2

24 8 4

19 19 7

.017

IVUS Homogenous Heterogeneous Layered OCT Homogenous Heterogeneous Layered

P

implantation and remains a rare finding until 4 years after implantation.14 Another postmortem pathologic study demonstrated that prominent infiltration by lipid-laden macrophages into neointima and adherent thrombus is found on disrupted lumen ≥4 years after bare-metal stent implantation.13,15 These pathologic findings may account for the OCT results in the present study. A recent IVUS study reported that disease progression with neointimal rupture was related with occurrence of very late stent thrombosis in patients who underwent bare-metal stent implantation.16 Therefore, different NIH appearance may be involved in the future development of stent thrombosis after DES implantation. The higher resolution of OCT, which allows visualization of early reendothelialization, may lead to adequate guidance for determination of more effective treatment strategies. This study had some limitations. First, this was a nonrandomized study that was retrospectively analyzed; therefore, there is a possibility of selection bias that may have influenced the results. Second, the NIH tissue characteristics were not validated by histopathologic examination. Finally, long-term clinical data from the study participants were not available; therefore, the effect of NIH tissue characteristics on long-term prognosis could not be evaluated. In conclusion, OCT and IVUS assessments of NIH morphology were moderately correlated in the present study; however, nonhomogenous NIH was more frequently detected by OCT than by IVUS.

Disclosures The authors do not have any potential conflicts of interest associated with this paper.

References 1. Bavry AA, Kumbhani DJ, Helton TJ, et al. Late thrombosis of drugeluting stents. Am J Med 2006;119:1056-61. 2. Mauri L, Hsieh WH, Massaro JM, et al. Stent thrombosis in randomized clinical trials of drug-eluting stents. N Engl J Med 2007; 356:1020-9. 3. Joner M, Finn AV, Farb A, et al. Pathology of drug-eluting stents in humans: delayed healing and late thrombotic risk. J Am Coll Cardiol 2006;48:193-202. 4. Finn AV, Joner M, Nakazawa G, et al. Pathological correlates of late drug-eluting stent thrombosis: strut coverage as a marker of endothelialization. Circulation 2007;115:2435-41. 5. Mintz GS, Weissman NJ. Intravascular ultrasound in the drug-eluting stent era. J Am Coll Cardiol 2006;48:421-9. 6. Capodanno D, Prati F, Pawlowsky T, et al. Comparison of optical coherence tomography and intravascular ultrasound for the assessment of in-stent tissue coverage after stent implantation. Eurointervention 2009;5:538-43. 7. Suzuki Y, Ikeno F, Koizumi T, et al. In vivo comparison between optical coherence tomography and intravascular ultrasound for detecting small degrees of in-stent neointima after stent implantation. J Am Coll Cardiol Intv 2008;1:168-73. 8. Jang IK, Bouma BE, Kang DH, et al. Visualization of coronary atherosclerotic plaques in patients using optical coherence tomography: comparison with intravascular ultrasound. J Am Coll Cardiol 2002;39:604-9. 9. Bouma BE, Tearney GJ, Yabushita H, et al. Evaluation of intracoronary stenting by intravascular optical coherence tomography. Heart 2003;89:317-20. 10. Gonzalo N, Serruys PW, Okamura T, et al. Optical coherence tomography patterns of stent restenosis. Am Heart J 2009;158: 284-93. 11. Kim U, Kim JS, Kim JS, et al. The initial extent of malapposition in STelevation myocardial infarction treated with drug-eluting stent: the usefulness of optical coherence tomography. Yonsei Med J 2010;51: 332-8. 12. Costa MA, Sabate M, Angiolillo DJ, et al. Intravascular ultrasound characterization of the “black hole” phenomenon after drug-eluting stent implantation. Am J Cardiol 2006;97:203-6. 13. Takano M, Yamamoto M, Inami S, et al. Appearance of lipid-laden intima and neovascularization after implantation of bare-metal stents: extended late-phase observation by intracoronary optical coherence tomography. J Am Coll Cardiol 2010;55:26-32. 14. Nakazawa G, Vorpahl M, Finn AV, et al. One step forward and two steps back with drug-eluting-stents: from preventing restenosis to causing late thrombosis and nouveau atherosclerosis. JACC: Cardiovascular imaging 2009;2:625-8. 15. Inoue K, Abe K, Ando K, et al. Pathological analyses of long-term intracoronary Palmaz-Schatz stenting: is its efficacy permanent? Cardiovasc Pathol 2004;13:109-15. 16. Lee CW, Kang SJ, Park DW, et al. Intravascular ultrasound findings in patients with very late stent thrombosis after either drug-eluting or bare-metal stent implantation. J Am Coll Cardiol 2010;55: 1936-42.

Standard versus high loading doses of clopidogrel in Asian ST-segment elevation myocardial infarction patients undergoing percutaneous coronary intervention: Insights from the Korea Acute Myocardial Infarction Registry Cheol Ung Choi, MD, a Seung-Woon Rha, MD, a Dong Joo Oh, MD, a Kanhaiya L. Poddar, MBBS, a Jin Oh Na, MD, a Jin Won Kim, MD, a Hong Euy Lim, MD, a Eung Ju Kim, MD, a Chang Gyu Park, MD, a Hong Seog Seo, MD, a Taek Jong Hong, MD, b,t Jong-Seon Park, MD, c,t Young Jo Kim, MD, c,t Seung Ho Hur, MD, d,t In Whan Seong, MD, e,t Jei Keon Chae, MD, f,t Myeong Chan Cho, MD, g,t Jang Ho Bae, MD, h,t Dong Hoon Choi, MD, i,t Yang Soo Jang, MD, i,t In Ho Chae, MD, j,t Hyo Soo Kim, MD, k,t Chong Jin Kim, MD, l,t Jung Han Yoon, MD, m,t Tae Hoon Ahn, MD, n,t Seung-Jea Tahk, MD, o,t Wook Sung Chung, MD, p,t Ki Bae Seung, MD, p,t Shung Chall Chae, MD, q,t Seung Jung Park, MD, r,t Young Keun Ahn, MD, s,t and Myung Ho Jeong, MD s,t Seoul, Pusan, Daegu, Daejeon, Jeonju, Chongju, Bundang, Wonju, and Gwangju, South Korea

Background

The optimal loading dose of clopidogrel in Asian patients with ST-segment elevation myocardial infarction (STEMI) has not been fully investigated. We compared bleeding, vascular complications, and midterm outcomes of a 300-mg versus a 600-mg loading dose of clopidogrel in a large series of Korean patients with STEMI undergoing primary percutaneous coronary intervention (PCI).

Methods A total of 2,664 STEMI patients (age 61.96 ± 11.91 years, men 70.4%) who underwent primary PCI were enrolled in this study. The patients were divided into a standard loading dose group (300 mg; n = 1,447 patients) and a high loading dose group (600 mg; n = 1,217 patients). Bleeding and vascular complications, and in-hospital and clinical outcomes up to 12 months were compared between the 2 groups. Results In-hospital bleeding and vascular complications were similar between the 2 groups. There were no differences in bleeding and vascular complications and in 1- and 12-month clinical outcomes, including mortality, myocardial infarction, repeated PCI, and major adverse cardiac events, between the 2 groups. These findings were consistent even after the propensity score–matched analysis. Conclusions

The standard loading dose of clopidogrel may be as safe and similarly effective as the high loading dose in Asian STEMI patients undergoing primary PCI. (Am Heart J 2011;161:373-382.e3.)

From the aCardiovascular Center, Korea University Guro Hospital, Seoul, South Korea, b Cardiovascular Center, Pusan National University Hospital, Pusan, South Korea,

Hospital, Seoul, South Korea, pCardiovascular Center, Seoul St. Mary's Hospital, Seoul, South Korea, qCardiovascular Center, Kyungpook National University Hospital, Daegu, Korea,

c Cardiovascular Center, Yeungnam University Hospital, Daegu, South Korea, dCardiovase cular Center, Keimyung University Dongsan Hospital, Daegu, South Korea, Cardiovascular f Center, Chungnam National University Hospital, Daejeon, South Korea, Cardiovascular Center, Chonbuk National University Hospital, Jeonju, South Korea, gCardiovascular Center, Chungbuk National University Hospital, Chongju, South Korea, hCardiovascular Center, Konyang University Hospital, Daejeon, South Korea, iCardiovascular Center, Yonsei University Severance Hospital, Seoul, South Korea, jCardiovascular Center, Seoul

r

National University Bundang Hospital, Bundang, South Korea, kCardiovascular Center, Seoul National University Hospital, Seoul, South Korea, lCardiovascular Center, Kyung Hee University East-West Neo Medical Center, Seoul, South Korea, mCardiovascular Center, Yonsei University Wonju Christian Hospital, Wonju, South Korea, nCardiovascular Center, Gacheon University Gil Hospital, Seoul, South Korea, oCardiovascular Center, Ajou University

Cardiovascular Center, Ulsan University Asan Medical Center, Seoul, South Korea, and Cardiovascular Center, Chonnam National University Hospital, Gwangju, Korea. t And other Korea Acute Myocardial Infarction Registry Investigators. See online Appendix for complete listing. The first two authors (Drs. C.U. Choi and S.W. Rha) contributed equally to this article. Submitted April 29, 2010; accepted October 18, 2010. s

Reprint requests: Dong Joo Oh, MD, Cardiovascular Center, Korea University, Guro Hospital, 97 Guro Dong, Guro Gu, Seoul 152-703, South Korea. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.031

American Heart Journal February 2011

374 Choi et al

Lowered platelet inhibition is known to be a strong independent predictor of ischemic events in patients undergoing percutaneous coronary intervention (PCI) with drug-eluting stents (DESs).1 A 600-mg clopidogrel double bolus achieved greater platelet inhibition than a conventional single loading dose.2 Several studies have shown that a 600-mg loading dose of clopidogrel consistently leads to greater platelet inhibition than a 300-mg loading dose in patients with stable angina and non–ST-segment elevation myocardial infarction (NSTEMI) undergoing PCI.3,4 Van de Werf et al5 reported that it would be beneficial to administer additional clopidogrel in primary PCI and stenting in ST-segment elevation myocardial infarction (STEMI) patients, preferably with a high loading dose of 600 mg. The HORIZONSAMI trial reported that a 600-mg loading dose of clopidogrel may safely reduce 30-day ischemic adverse event rates compared with a 300-mg loading dose in patients with STEMI undergoing primary PCI.6 However, most of the previous data came from the Western people; and currently, there are limited data comparing the effects of a standard (300 mg) and a high (600 mg) loading dose of clopidogrel in Asian STEMI patients undergoing primary PCI. Because of relatively smaller body weight, different degree of platelet aggregation, and possible ethnic differences in clopidogrel response, the optimal clopidogrel loading dose in Asian patients may not be same as that in Western people. In addition, there are no data in the literature comparing midterm clinical outcomes of a 300- with a 600-mg loading dose of clopidogrel in Asian STEMI patients undergoing primary PCI. In the present study, we compared the efficacy and safety between a 300- and a 600-mg loading dose of clopidogrel in a series of Korean STEMI patients undergoing primary PCI.

Methods Korea Acute Myocardial Infarction Registry The Korea Acute Myocardial Infarction Registry (KAMIR) is an online Korean prospective multicenter registry that has been investigating the risk factors of mortality in patients with acute myocardial infarction (AMI) since November 2005 with the aim of establishing universal management guidelines for the prevention of AMI, with the support of the Korean Circulation Society. Online registry of AMI (found at www. kamir.or.kr) has been carried out in 41 primary PCI centers, hospitals capable of primary PCI with sufficient experience and volume. The study protocol was approved by the ethics committee at each participating institution. Data were registered and submitted from individual institutions via passwordprotected Internet-based electronic case report forms. We enrolled patients if they had an AMI, including STEMI or NSTEMI. Patients were diagnosed with STEMI when they had new or presumed new ST-segment elevation of ≥1 mm seen in any location or a new left bundle-branch block on the index or subsequent electrocardiogram with ≥1 positive cardiac bio-

chemical marker of necrosis (including creatine kinase–MB and troponin I and T).

Study sample A total of 6,381 patients diagnosed with AMI were enrolled from December 2006 to March 2008. In the present study, we retrospectively analyzed the patients with acute STEMI who underwent primary PCI (PCI performed ≤24 hours after the symptom onset). Initial treatment strategy for patients with STEMI could be PCI, thrombolysis, or conservative treatment, at the discretion of the attending physician. The criteria for exclusion included NSTEMI, STEMI undergoing primary PCI with balloon angioplasty only and conservative treatment without PCI, contraindication to antithrombotic agents, known bleeding disorders, thrombocytopenia (b100 × 109/ L), administration of thrombolytic or fibrinolytic medications for STEMI, infarction related to the grafted vessel, severe hepatic or renal dysfunction (serum creatinine N2 mg/dL), and estimated life expectancy of b12 months. Loading doses of aspirin and clopidogrel were administered immediately after the patients agreed to receive PCI. The loading doses of aspirin were 200 or 300 mg, and maintenance doses were 100 mg QD. The loading doses of clopidogrel were 300 to 600 mg, and maintenance doses were 75 mg QD. The clopidogrel loading dose was decided by individual physician's discretion. Patients were divided into 2 groups as follows: a standard loading dose of clopidogrel group (300 mg; n = 1,447 patients) and a high loading dose of clopidogrel group (600 mg; n = 1,217 patients). Dual-antiplatelet therapy was strongly recommended to all patients for ≥12 months, as per existing guidelines.

PCI procedure and medical treatment Diagnostic angiography and PCI were performed after unfractionated heparin (50-70 U/kg) was administered. Coronary angiography was performed through the femoral or radial artery. During the procedure, patients received unfractionated heparin to maintain an activated clotting time of N250 seconds. Stents were chosen and deployed at the physician's discretion during index PCI after the following lesion preparation: glycoprotein IIb/IIIa receptor blockers, thrombectomy devices, thrombus aspiration devices, and/or predilatation balloons. A successful PCI procedure was defined as the achievement of an angiographic minimum stenosis diameter reduction to b30% in the presence of Thrombolysis in Myocardial Infarction (TIMI) grade III flow. During the in-hospital period, patients received essential medical treatment that included β-blockers, angiotensin-converting enzyme inhibitors, or angiotensin II receptor blockers; calcium-channel blockers; and statins. After discharge, the patients continued to receive the same kinds of optimal medications that they received in-hospital, except for some intravenous or temporary medications.

Clinical follow-up Patients were required to visit the outpatient clinic of the cardiology department at the end of the first month, every 6 months after the PCI procedure, and whenever angina-like symptoms occurred. At 12 months after the index PCI, follow-up data were obtained by reviewing medical records and/or

American Heart Journal Volume 161, Number 2

Choi et al 375

Figure 1

Flowchart of the study. KAMIR, Korea Acute Myocardial Infarction Registry; AMI, acute myocardial infarction; NSTEMI, non-ST-segment elevation myocardial infarction; STEMI, non-ST-segment elevation myocardial infarction; PCI, percutaneous coronary intervention.

telephone interviews with patients. All data were entered into an electronic Web-based case report form.

Study definitions and clinical end points The records of cardiovascular risk factors and history (age, gender, hypertension, dyslipidemia, smoking, diabetes mellitus, family history of coronary heart disease, prior myocardial infarction, chronic heart failure, prior cerebrovascular disease, and peripheral arterial disease) were dependent mainly on the patient's self-report. The final records were left to the physician's discretion after he or she comprehensively considered the patient's self-report and the in-hospital examination results. After exclusion of noncardiac deaths, every death was considered a cardiac death. Recurrent myocardial infarction was defined as recurrent symptoms with new ST-segment elevation or reelevation of cardiac markers to at least twice the upper limit of normal. Target lesion revascularization was defined as ischemia-induced PCI of the target lesion resulting from restenosis or reocclusion within the stent or in the adjacent 5-mm area of the distal or proximal segment. Total major adverse cardiac events (MACEs) were defined as the composite of all-cause death, nonfatal myocardial infarction, and repeated PCI or coronary artery bypass grafting (CABG).

Cardiac death was defined as death from pump failure, arrhythmia, or mechanical complications including ventricular septal rupture and free wall rupture. Major bleeding was defined as any intracranial bleeding, significant gastrointestinal bleeding, retroperitoneal bleeding, bleeding associated with the need for blood transfusion, or any other clinically relevant bleeding, as judged by the investigator. Vascular complications were defined as a major hematoma needing transfusion or arteriovenous fistula at the puncture site. The incidences of major bleeding events and major clinical outcomes including MACEs, in-hospital and at 12 months, were evaluated between the groups of 300- and 600-mg loading dose of clopidogrel.

Statistical analysis For continuous variables, differences between groups were evaluated by Student t test. All continuous variables were described as mean ± SD. For discrete variables, differences were expressed as counts and percentages and were analyzed with the χ2 (or Fisher exact) test between the groups, as appropriate. Propensity score matching is a method of adjusting for observed characteristics of patients nonrandomly assigned to differing treatments.7 To adjust for potential selection biases

American Heart Journal February 2011

376 Choi et al

Table I. Baseline characteristics according to clopidogrel loading dose

Age (y) Male Weight (kg) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) History Hypertension Diabetes Dyslipidemia Current smoking Ischemic heart disease Myocardial infarction CABG Heart failure Ejection fraction b40% Prior medication Aspirin Clopidogrel Glucose (mg/dL) Serum creatinine (mg/dL) Maximum CK-MB (ng/mL) Maximum troponin I (mg/mL) Total cholesterol (mg/dL) Triglyceride (mg/dL) HDLc (mg/dL) LDLc (mg/dL) hs-CRP (mg/dL) NT-proBNP (pg/mL)

300 mg (n = 1447)

600 mg (n = 1217)

P

62.83 ± 12.93 69.1% 64.83 ± 17.68 24.07 ± 3.30 129.43 ± 27.88 79.13 ± 25.98

62.77 ± 12.41 70.7% 64.36 ± 11.71 23.99 ± 3.24 129.77 ± 28.18 78.79 ± 16.80

.899 .374 .454 .627 .760 .702

49.0% 21.7% 11.7% 43.7% 15.7% 6.6% 1.0% 2.2% 15.6%

45.7% 23.2% 10.9% 54.8% 15.9% 7.4% 0.6% 2.3% 15.7%

.106 .375 .389 .656 .915 .446 .280 .896 .955

10.2% 3.4% 175.92 ± 85.39 1.22 ± 1.42 170.04 ± 193.77 70.02 ± 283.89 181.10 ± 43.12 119.07 ± 85.44 44.25 ± 12.50 114.38 ± 37.22 8.60 ± 27.84 2289.28 ± 5928.25

9.3% 3.2% 174.04 ± 76.02 1.14 ± 0.95 201.42 ± 270.78 60.99 ± 185.23 182.73 ± 42.74 121.43 ± 88.87 44.15 ± 11.78 116.86 ± 36.07 12.24 ± 51.58 2031.58 ± 5028.69

.433 .828 .555 .106 .001 .402 .340 .503 .846 .107 .028 .391

CABG, Coronary artery bypass graft; CK-MB, creatine kinase-MB; HDLc, high density lipoprotein-cholesterol; LDLc, low density lipoprotein-cholesterol; hs-CRP, high sensitivity C-reactive protein; NT-proBNP, N-terminal pro-B-type natriuretic peptide.

Table II. Procedural characteristics according to clopidogrel loading dose

Multivessel disease Target vessel Left main LAD LCX RCA TIMI flow grade pre-PCI (TIMI 0/I) Final TIMI flow grade after PCI (TIMI III) Sirolimus-eluting stent Paclitaxel-eluting stent Zotarolimus-eluting stent Bare metal stent Other DES Stent number Stent length Stent diameter

300 mg (n = 1447)

600 mg (n = 1217)

P

53.3%

50.3%

.129

1.9% 51.6% 9.7% 36.8% 69.2% 93.3% 35.6% 28.1% 21.1% 12.7% 2.4% 1.44 ± 0.74 24.71 ± 5.90 3.20 ± 0.44

2.1% 48.6% 9.1% 40.3% 69.0% 92.5% 35.1% 28.1% 27.9% 5.7% 3.1% 1.39 ± 0.69 25.46 ± 6.23 3.22 ± 0.44

.890 .120 .642 .066 .930 .479 .831 1 b.001 b.001 .381 .146 .003 .290

LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; RCA, right coronary artery; TIMI, thrombolysis in myocardial infarction; PCI, percutaneous coronary intervention; DES, drug eluting stent.

and potential confounding, patients in the 300- and the 600-mg clopidogrel groups were matched using propensity scores. A logistic regression model was fitted relating the clopidogrel groups (300 vs 600 mg) to pretreatment patient characteristics. We tested variables that could be of potential relevance: age,

gender, prior cerebrovascular disease, maximum creatine kinase– MB (CK-MB) level, high-sensitivity C-reactive protein (hs-CRP), low–molecular weight heparin (LMWH), glycoprotein IIb/IIIa receptor blockers, pre-TIMI flow grade, post-TIMI flow grade, stent type (DES or not), and stent length. Multiple logistic regression

American Heart Journal Volume 161, Number 2

Choi et al 377

Table III. In-hospital medications 300 mg (n = 1447)

600 mg (n = 1217)

P

34.9% 65.1% 25.0%

31.5% 68.5% 18.1%

.063 .063 b.001

98.5% 97.6% 29.3% 66.9% 11.4% 70.1% 76% 1.5%

97.6% 96.4% 29.5% 68.0% 12.3% 71.1% 75.2% 1.0%

.119 .085 .932 .532 .429 .599 .617 .279

Low molecular weight heparin Unfractionated heparin Glycoprotein IIb/IIIa receptor inhibitor In-hospital medication Aspirin Clopidogrel Cilostazol Angiotensin-converting enzyme inhibitor Angiotensin II receptor blockers β-Blockers Statin Warfarin

Figure 2

In-hospital and 1-month MACEs according to clopidogrel loading dose. MI, myocardial infarction; PCI, percutaneous coronary intervention; MACEs, major adverse adverse cardiac events.

analysis was used to identify independent predictors of 1- and 12-month MACEs. We calculated the 95% CI for each odds ratio (OR), and all P values were 2-tailed. A P b .05 was considered statistically significant. Statistical analysis was performed using the SPSS 17.0 software package (SPSS Inc, Chicago, IL). This program was supported by a Korea University Grant.

Results Baseline patient characteristics Of the 6,381 patients enrolled in KAMIR, 3,717 patients were excluded according to the exclusion criteria. Therefore, the present study was composed of 2,644 patients, including 1,447 patients who received a standard loading dose (300 mg) of clopidogrel and 1,217 patients who received a high loading dose (600 mg) of clopidogrel before cardiac catheterization (Figure 1).

Baseline clinical characteristics were similar between the 2 groups except for a higher probability of a history of cerebrovascular disease, maximum CK-MB, and hs-CRP in the 600-mg loading dose group compared with the 300-mg loading dose group (Table I). The procedural characteristics are listed in Table II. The 2 groups had nearly similar procedural characteristics, except that thepatientsin the300-mgloadingdosegrouphadhigher TIMI flow grade before PCI and received more bare metal stents than the patients in the 600-mg loading dose group. In addition, the patients in the 600-mg loading dose group received longer stents and more zotarolimus-eluting stent than the patients in 300-mg loading dose group. The in-hospital medications administered to the patients are listed in Table III. Patients in the 300-mg loading dose group received more glycoprotein IIb/IIIa receptor inhibitors during the procedure than patients in the 600-mg loading dose group.

378 Choi et al

American Heart Journal February 2011

Figure 3

In-hospital and other clinical outcomes at 1 month according to clopidogrel loading dose. ARF, Acute renal failure; HF, heart failure; CVA, cerebrovascular accident; VT, ventricular tachycardia; VF, ventricular fibrillation.

Figure 4

Twelve-month cumulative clinical outcomes according to clopidogrel loading dose. MI, myocardial infarction; PCI, percutaneous coronary intervention; MACEs, major adverse adverse cardiac events.

Clinical outcomes In-hospital and 1-month clinical outcomes. There were no differences in in-hospital and 1-month cardiac death, noncardiac death, recurrent MI, repeated PCI, total MACEs, and bleeding complications between the 300and the 600-mg loading dose groups (Figures 2 and 3). Twelve-month cumulative clinical outcomes. There were no differences in the 12-month cumulative clinical outcomes including cardiac death, noncardiac death, recurrent MI, repeated PCI, and total MACEs

between the 300- and the 600-mg loading dose groups (Figure 4). Independent predictors of 1- and 12-month MACEs. Independent predictors of MACEs at 1 month included left main target vessel, TIMI grade III after PCI, and N-terminal pro–B-type natriuretic peptide (NT-proBNP) level. Loading dose of clopidogrel was not an independent predictor of MACEs at 1 month (Table IV). Independent predictors of MACEs at 12 months were diabetes mellitus, left main target vessel,

American Heart Journal Volume 161, Number 2

Choi et al 379

Table IV. Independent predictors of 1-month MACEs by multivariate analysis Propensity-matched patients (n = 1542)

All patients (N = 2664)

Age Female Ejection fraction b40% Previous IHD Diabetes Multivessel disease Target vessel (left main) Lesion type C TIMI flow grade pre-PCI (TIMI 0/I) Final TIMI flow grade after PCI (TIMI III) Maximum CK-MB (ng/mL) Maximum troponin I (ng/mL) NT-proBNP (pg/mL) Cardiogenic shock In-hospital medication (statin) Clopidogrel 600-mg loading dose

OR

95% CI

P

OR

95% CI

P

0.99 1.37 2.85 0.49 2.51 0.92 5.85 1.36 0.89 0.31 1 1.00 1 1.42 0.90 1.18

0.96-1.01 0.71-2.64 1.48-5.48 0.19-1.24 1.19-3.89 0.51-1.65 1.37-25.06 0.74-2.48 0.46-1.75 0.11-0.85 0.99-1.00 1.00-1.01 1.00-1.00 0.44-4.59 0.46-1.77 0.66-2.21

.324 .349 .002 .132 .039 .769 .017 .325 .754 .023 .897 .074 .007 .556 .762 .570

0.97 1.57 4.07 0.74 1.55 0.70 6.09 1.00 0.76 0.14 1.00 1.00 1 2.35 0.76 0.77

0.94-1.01 0.65-3.81 1.71-9.67 0.24-2.25 0.68-3.58 0.32-1.57 0.93-39.99 0.44-2.27 0.32-1.80 0.02-0.87 0.99-1.00 0.99-1.00 1.00-1.00 0.63-8.74 0.31-1.88 0.35-1.72

.106 .315 .001 .187 .300 .390 .060 .995 .527 .035 .460 .448 .005 .204 .552 .523

Adjusted risk factors were age, gender, ejection fraction, previous ischemic heart disease, diabetes, multivessel lesion, target vessel location (left main), lesion type, TIMI flow grade pre-PCI, final TIMI flow grade after PCI, maximum CK-MB, maximum troponin I, NT-proBNP, cardiogenic shock, and in-hospital statin medication. IHD, ischemic heart disease; TIMI, thrombolysis in myocardial infarction; PCI, percutaneous coronary intervention; CK-MB, creatine kinase; NT-proBNP, N-terminal pro-B-type natriuretic peptide.

Table V. Independent predictors of 12-month MACEs by multivariate analysis Propensity-matched patients (n = 1542)

All patients (N = 2664)

Age Female Ejection fraction b40% Previous IHD Diabetes Multivessel disease Target vessel (left main) Lesion type C TIMI flow grade pre-PCI (TIMI 0/I) Final TIMI flow grade after PCI (TIMI III) NT-proBNP (pg/mL) Cardiogenic shock Discharge medication (statin) Clopidogrel 600-mg loading dose

OR

95% CI

P

OR

95% CI

P

0.99 1.15 1.55 1.22 2.01 1.64 6.18 1.08 1.47 0.91 1.00 0.76 1.01 1.02

0.97-1.01 0.76-1.74 0.98-2.44 0.77-1.96 1.38-2.93 1.13-2.37 2.04-18.71 0.75-1.56 0.99-2.21 0.40-2.06 1.00-1.00 0.32-1.83 0.66-1.55 0.72-1.47

.347 .514 .059 .398 b.001 .009 .001 .695 .066 .826 .001 .763 .969 .899

0.99 1.19 1.95 0.89 1.44 1.44 2.59 1.07 1.47 0.66 1.00 0.90 0.46 0.96

0.97-1.01 0.70-2.02 1.09-3.48 0.47-1.70 0.86-2.42 0.90-2.33 0.60-11.17 0.67-1.74 0.87-2.47 0.14-3.23 1.00-1.00 0.31-2.57 0.29-0.75 0.60-1.53

.214 .528 .024 .730 .165 .130 .201 .770 .149 .610 .002 .840 .002 .863

Adjusted risk factors were age, gender, ejection fraction, previous IHD, diabetes, multivessel lesion, target vessel location (left main), lesion type, TIMI flow grade pre-PCI, final TIMI flow grade after PCI, NT-proBNP, cardiogenic shock, and discharge statin medication. IHD, ischemic heart disease; TIMI, thrombolysis in myocardial infarction; PCI, percutaneous coronary intervention; NT-proBNP, N-terminal pro-B-type natriuretic peptide.

and NT-proBNP level. Loading dose of clopidogrel was not an independent predictor of MACEs at 12 months, too (Table V). Results from the propensity score–matched patients. A total of 1,542 matched cases, 771 in each of the clopidogrel (300 and 600 mg) groups, were found. Propensity score–matched analysis results were similar to those of the main analysis in the full cohort. There were no differences in 1- and 12-month clinical outcomes between 300- and 600-mg loading

dose of clopidogrel among propensity score–matched patients (Figure 5). Loading dose of clopidogrel was not an independent predictor of MACEs at 1 and 12 months even after propensity score matching, too (Tables IV and V). Intention-to-treat analysis in all patients with STEMI who got the loading dose of clopidogrel. We analyzed all patients with STEMI who got the loading dose of clopidogrel, irrespective of the treatment strategy used (conservative or invasive) or intention-to-

380 Choi et al

American Heart Journal February 2011

Figure 5

In-hospital, 1-month, and 12-month MACEs among propensity-matched patients. A, Clinical outcomes at 1 month. B, Other clinical outcomes including bleeding complications at 1 month. C, Clinical outcomes at 12 months.

American Heart Journal Volume 161, Number 2

treat method, to reduce the interference of potentially biasing factors. Of the 6,381 patients enrolled in KAMIR, 2,862 STEMI patients got the loading dose of clopidogrel; and the patients were divided into a standard loading dose group (300 mg; n = 1,555 patients) and a high loading dose group (600 mg; n = 1,307 patients). The baseline and procedural characteristics in all STEMI patients who got the loading dose of clopidogrel (including patients with and without receiving PCI) were similar to those in STEMI patients undergoing PCI. In-hospital medication and 1- and 12-month clinical outcomes were also similar to those of the analysis in STEMI patients undergoing primary PCI (data are shown in supplements I-IV).

Discussion The major findings of the present study are as follows: (1) there were no differences in 1- and 12-month major clinical outcomes including the incidence of stent thrombosis, mortality, reinfarction, and total MACEs between a 300- and a 600-mg loading dose of clopidogrel; and (2) there were no differences in 1-month major bleeding complications between a 300- and a 600-mg loading dose of clopidogrel in Korean patients. Several studies showed that a 600-mg loading dose of clopidogrel consistently led to greater platelet inhibition when compared with a 300-mg loading dose in patients with stable angina, NSTEMI, and STEMI undergoing PCI.1-4,6,8,9 However, it is possible that there may be differences in platelet aggregation and clopidogrel responsiveness between patients of Asian and Western descent. Although variability in patient response to antiplatelet treatment has been observed previously,10 speculation remains with regard to populational differences and antiplatelet responsiveness. To the best of our knowledge, there are no data regarding the optimal loading dose of clopidogrel in Asian AMI patients undergoing primary PCI. However, there are some data showing that the optimal loading dose of ticlopidine in Japanese patients was 200 mg, which differs from American and European data.11 In addition, a previous study reported that patients of Asian descent showed greater platelet inhibition in response to prasugrel, a novel member of the thienopyridine class of oral antiplatelet agents, when compared with patients of European descent.12 The most frequent and clinically relevant safety issues in antiplatelet therapy are lethal bleeding complications, including intracranial hemorrhage and gastrointestinal and retroperitoneal bleeding. The currently available data also suggest that there are differences in adverse effects of antiplatelet agents in Asian patients when compared with European patients, with a high prevalence and incidence of bleeding complication.13

Choi et al 381

Clopidogrel exhibits wide interindividual variation in its antiplatelet effects.10,14 In Western population, the rates of patients with clopidogrel resistance ranged between 5% and 44%.15 Hoshino et al16 showed that the rate of clopidogrel nonresponders was 14% among Japanese patients. We assume that the proportion may be similar in Korean patients. Importantly, several studies have revealed that cardiovascular risk is increased in patients with clopidogrel resistance17; and these particular subset of patients may need higher loading and/or maintenance dose to reduce the risk of ischemic complications. According to previous data,6 body mass index was about 27 kg/m2 in Western patients with STEMI undergoing primary PCI. It is larger than that of the present data (about 24 kg/m2). This difference in body mass index between Asian and Western could influence the results of present study. In the European Society of Cardiology 2009, investigators reported the initial results of CURRENT-OASIS 7 trial that double dose of clopidogrel reduced stent thrombosis and MACEs in PCI.18 To the best of our knowledge, of the 25,087 patients enrolled CURRENT-OASIS 7, there were 640 Korean patients (about 2.6%). We think that this small number could not influence the whole results of CURRENT-OASIS 7 trial. In addition, enrolled patients included not only those with STEMI, but also other acute coronary syndrome patients. Therefore, there could be a different result from the present study. From these results, we can conclude that, in Asian patients with STEMI undergoing primary PCI, a 600-mg loading dose of clopidogrel could not afford additional benefit when compared with a 300-mg loading dose of clopidogrel. In addition, even in South Korea only, between 3,000 and 3,500 new patients are presenting with STEMI during 1 year according to KAMIR data. Therefore, in terms of cost-effectiveness, if a 300-mg loading dose of clopidogrel is sufficient for STEMI patients undergoing primary PCI, the economic benefit would be considerable without increasing major cardiovascular events. The present study is the first one that compares the safety and efficacy of 300- and 600-mg loading doses of clopidogrel in Asian STEMI patients undergoing primary PCI. It would be necessary to clarify a detailed difference between patients of Asian and Western descent in regard to the optimal loading dose of clopidogrel in STEMI patients undergoing primary PCI. In the future, we expect more research supporting the present clinical data to be carried out. The limitations of our study are as follows. First, this is a retrospective analysis and is subject to certain inherent limitations and potential biases. Second, because the choice of loading dose of clopidogrel was decided by the discretion of physician, there could be a selection bias. However, a relatively larger sample was enrolled; and propensity score match analysis was performed to minimize many confounding

382 Choi et al

factors. This study may have enough statistical power to prove important results. In conclusion, we suggest that a standard loading dose of clopidogrel (300 mg) would be enough for Asian STEMI patients undergoing primary PCI when considering the safety and clinical efficacy as well as the cost-effectiveness.

Acknowledgements This program was supported by a Korea University Grant, and we appreciated the help of all KAMIR investigators. The authors are solely responsible for the design and conduct of the present study, all study analyses, and the drafting and editing of the article and its final contents.

Disclosures The authors have no conflicts of interest to disclose with regard to the work reported herein.

References 1. Buonamici P, Marcucci R, Migliorini A, et al. Impact of platelet reactivity after clopidogrel administration on drug-eluting stent thrombosis. J Am Coll Cardiol 2007;49:2312-7. 2. L'Allier PL, Ducrocq G, Pranno N, et al. Clopidogrel 600-mg double loading dose achieves stronger platelet inhibition than conventional regimens: results from the PREPAIR randomized study. J Am Coll Cardiol 2008;51:1066-72. 3. Cuisset T, Frere C, Quilici J, et al. Benefit of a 600-mg loading dose of clopidogrel on platelet reactivity and clinical outcomes in patients with non–ST-segment elevation acute coronary syndrome undergoing coronary stenting. J Am Coll Cardiol 2006;48:1339-45. 4. Patti G, Colonna G, Pasceri V, et al. Randomized trial of high loading dose of clopidogrel for reduction of periprocedural myocardial infarction in patients undergoing coronary intervention: results from the ARMYDA-2 (Antiplatelet therapy for Reduction of MYocardial Damage during Angioplasty) study. Circulation 2005;111: 2099-106. 5. Van de Werf F, Bax J, Betriu A, et al. Management of acute myocardial infarction in patients presenting with persistent STsegment elevation: the Task Force on the Management of ST-Segment Elevation Acute Myocardial Infarction of the European Society of Cardiology. Eur Heart J 2008;29:2909-45.

American Heart Journal February 2011

6. Dangas G, Mehran R, Guagliumi G, et al. Role of clopidogrel loading dose in patients with ST-segment elevation myocardial infarction undergoing primary angioplasty: results from the HORIZONS-AMI (harmonizing outcomes with revascularization and stents in acute myocardial infarction) trial. J Am Coll Cardiol 2009;54:1438-46. 7. D'Agostino Jr RB. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265-81. 8. Gurbel PA, Bliden KP, Samara W, et al. Clopidogrel effect on platelet reactivity in patients with stent thrombosis: results of the CREST study. J Am Coll Cardiol 2005;46:1827-32. 9. Lotrionte M, Biondi-Zoccai GG, Agostoni P, et al. Meta-analysis appraising high clopidogrel loading in patients undergoing percutaneous coronary intervention. Am J Cardiol 2007;100:1199-206. 10. Serebruany VL, Steinhubl SR, Berger PB, et al. Variability in platelet responsiveness to clopidogrel among 544 individuals. J Am Coll Cardiol 2005;45:246-51. 11. Gawaz M, Seyfarth M, Müller I, et al. Comparison of effects of clopidogrel versus ticlopidine on platelet function in patients undergoing coronary stent placement. Am J Cardiol 2001;87: 332-6. 12. Small DS, Kothare P, Yuen E, et al. The pharmacokinetics and pharmacodynamics of prasugrel in healthy Chinese, Japanese, and Korean subjects compared with healthy Caucasian subjects. Eur J Clin Pharmacol 2010;66:127-35. 13. Kuo PI, Severino R, Pashkow FJ. Mortality rates and hemorrhagic complications in Asian-Pacific islanders during treatment of acute myocardial infarction. Am J Cardiol 2004;94:644-6. 14. Gurbel PA, Bliden KP, Hiatt BL, et al. Clopidogrel for coronary stenting: response variability, drug resistance, and the effect of pretreatment platelet reactivity. Circulation 2003;107:2908-13. 15. Gurbel PA, Becker RC, Mann KG, et al. Platelet function monitoring in patients with coronary artery disease. J Am Coll Cardiol 2007;50: 1822-34. 16. Hoshino K, Horiuchi H, Tada T, et al. Clopidogrel resistance in Japanese patients scheduled for percutaneous coronary intervention. Circ J 2009;73:336-42. 17. Geisler T, Langer H, Wydymus M, et al. Low response to clopidogrel is associated with cardiovascular outcome after coronary stent implantation. Eur Heart J 2006;27:2420-5. 18. Mehta SR. CURRENT OASIS 7 trial results: a randomized comparison of a clopidogrel high loading and maintenance dose regimen versus standard dose and high versus low dose aspirin in 25,000 patients with acute coronary syndromes. Available at: http://www.escardio. org/congresses/esc-2009/congress-reports/Documents/706003mehta-slides.pdf.

American Heart Journal Volume 161, Number 2

Choi et al 382.e1

Appendix A. Supplementary data Supplement I. Baseline characteristics according to clopidogrel loading dose in all patients with and without PCI

Age (y) Male Weight (kg) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Heart rate History Hypertension Diabetes mellitus Dyslipidemia Current smoking IHD Myocardial infarction Peripheral vascular disease CABG Heart failure Cerebrovascular disease Family history of IHD Killip class 2-4 Ejection fraction b40% Prior medication Aspirin Clopidogrel Lipid-lowering drug Glucose (mg/dL) Serum creatinine (mg/dL) Maximum CK-MB Maximum troponin I Total cholesterol (mg/dL) Triglyceride (mg/dL) HDLc (mg/dL) LDLc (mg/dL) hs-CRP (mg/dL) NT-proBNP (pg/mL)

300 mg (n = 1555)

600 mg (n = 1307)

P

62.44 ± 12.95 70.00% 64.78 ±11.61 24.12 ± 3.30 127.92 ± 29.99 78.61 ± 26.11 78.20 ± 20.18

62.95 ± 12.50 70.80% 64.17 ± 11.69 23.92 ± 3.26 129.35 ± 38.26 79.11 ± 31.27 77.43 ± 32.53

.750 .666 .188 .124 .271 .65 .449

48.50% 24.40% 12.00% 44.20% 15.70% 18.90% 2.70%

45.50% 25.20% 11.10% 45.50% 15.40% 20.20% 1.90%

.114 .198 .337 .774 .836 .624 .421

6.10% 13.80% 7.40%

5.90% 17.90% 7.20%

.905 .080 .738

15.40%

15.70%

.835

9.90% 3.40% 5.30% 175.16 ± 83.80 1.20 ± 1.38 176.81 ± 196.46 74.49 ± 279.12 182.11 ± 44.36 125.50 ± 120.92 44.41 ± 12.95 115.59 ± 39.08 15.07 ± 85.55 2050.69 ± 5496.76

8.70% 2.80% 4.70% 173.41 ± 76.75 1.13 ± 0.93 202.55 ± 265.67 62.70 ± 180.66 183.38 ± 49.10 123.63 ± 96.20 46.04 ± 36.94 118.26 ± 47.28 20.96 ± 105.27 1935.17 ± 4837.45

.280 .378 .458 .569 .127 .003 .247 .480 .663 .118 .123 .151 .669

IHD, ischemic heart disease; CABG, Coronary artery bypass graft; CK-MB, creatine kinase-MB; HDLc, high density lipoprotein-cholesterol; LDLc, low density lipoprotein-cholesterol; hs-CRP, high sensitivity C-reactive protein; NT-proBNP, N-terminal pro-B-type natriuretic peptide.

American Heart Journal February 2011

382.e2 Choi et al

Supplement II. Procedural characteristics according to clopidogrel loading dose in all patients with and without PCI

Multivessel disease Target vessel Left main LAD LCX RCA TIMI flow grade pre-PCI TIMI 0/I TIMI II TIMI III Final TIMI flow grade after PCI TIMI 0/I TIMI II TIMI III Stent type Sirolimus-eluting stent Paclitaxel-eluting stent Zotarolimus-eluting stent Bare metal stent Other DES Stent number Stent length Stent diameter

300 mg (n = 1555)

600 mg (n = 1307)

P

54.00%

51.10%

.122

1.70% 50.40% 9.50% 38.40%

1.60% 48.50% 9.00% 40.90%

.899 .323 .643 .187

66.30% 15.30% 18.40%

67.10% 17.40% 15.50%

.670 .150 .055

2.20% 4.50% 93.30%

2.90% 4.60% 92.40%

.205 .884 .372

34.00% 28.50% 21.40% 13.40% 2.80% 1.45 ± 0.76 24.16 ± 6.05 3.23 ± 0.47

35.10% 28.00% 28.10% 5.40% 3.40% 1.41 ± 0.67 25.03 ± 6.24 3.22 ± 0.45

.542 .773 b.001 b.001 .360 .094 b.001 .683

LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; RCA, right coronary artery; TIMI, thrombolysis in myocardial infarction; PCI, percutaneous coronary intervention; DES, drug eluting stent.

Supplement III. In-hospital medications in all patients with and without PCI

LMWH, low molecular weight heparin Unfractionated heparin Glycoprotein IIb/IIIa receptor inhibitor In-hospital medication Aspirin Clopidogrel Cilostazol

300 mg (n = 1555)

600 mg (n = 1307)

P

34.10% 70.70% 14.30%

32.20% 51.20% 8.70%

.299 b.001 b.001

99.50% 98.70% 29.60%

99.40% 98.10% 30.70%

.540 .223 .535

Supplement IV. One- and 12-month cumulative clinical outcomes according to clopidogrel loading dose in all patients with and without PCI

1-month clinical outcomes Stent thrombosis (acute and subacute) Cardiac death Noncardiac death CABG Recurrent myocardial infarction Repeated PCI Total MACEs 12-m clinical outcomes Stent thrombosis (late) Cardiac death Noncardiac death CABG Recurrent myocardial infarction Repeated PCI Total MACEs

300 mg (n = 1555)

600 mg (n = 1307)

P

1.00% 1.60% 0.60% 0.10% 0.30% 1.40% 4.00%

1.80% 0.80% 0.90% 0% 0.40% 1.40% 3.60%

.195 .1 .468

0.60% 1.50% 0.30% 0.00% 0.50% 3.80% 6.10%

0.30% 0.40% 0.60% 0.00% 0.00% 4.60% 5.90%

CABG, Coronary artery bypass graft; PCI, percutaneous coronary intervention; MACEs, major adverse cardiac events.

1 .725 .955 .682 .496 .072 .672 .354 .304 .470 .867

American Heart Journal Volume 161, Number 2

Appendix B. Kamir Investigators Korea Acute Myocardial Infarction Registry (KAMIR) Investigators: Myung Ho Jeong, Young Jo Kim, Chong Jin Kim, Myeong Chan Cho, Young Keun Ahn, Jong Hyun Kim, Shung Chull Chae, Seung Ho Hur, In Whan Seong, Taek Jong Hong, Dong Hoon Choi, Jei Keon Chae, Jae Young Rhew, Doo Il Kim, In Ho Chae, Jung Han Yoon, Bon Kwon Koo, Byung Ok Kim, Myoung Yong Lee, Kee Sik Kim, Jin Yong Hwang, Seok Kyu Oh, Nae Hee Lee, Kyoung Tae Jeong, Seung Jea Tahk, Jang Ho Bae, Seung Woon Rha, Keum Soo Park, Kyoo Rok Han, Tae Hoon Ahn, Moo Hyun Kim, Ju Young Yang, Chong Yun Rhim, Hyeon Cheol Gwon, Seong Wook Park, Young Youp Koh, Seung Jae Joo, Soo Joong Kim, Dong Kyu Jin, Jin Man Cho, Jeong Gwan Cho, Wook Sung Chung, Yang Soo Jang, Ki Bae Seung, and Seung Jung Park.

Choi et al 382.e3

Comparison of 2 point-of-care platelet function tests, VerifyNow Assay and Multiple Electrode Platelet Aggregometry, for predicting early clinical outcomes in patients undergoing percutaneous coronary intervention Young-Guk Ko, MD, a Jung-Won Suh, MD, PhD, c Bo Hyun Kim, BA, a Chan Joo Lee, MD, a Jung-Sun Kim, MD, PhD, a Donghoon Choi, MD, PhD, a Myeong-Ki Hong, MD, PhD, a,b Myung-Ki Seo, MD, c Tae-Jin Youn, MD, PhD, c In-Ho Chae, MD, PhD, c Dong Joo Choi, MD, PhD, c and Yangsoo Jang, MD, PhD a,b Seoul, and Gyeonggi-do, Korea

Background Various platelet function tests are currently used to measure responsiveness to antiplatelet therapy. We sought to compare 2 point-of-care platelet function tests, VerifyNow Assay (Accumetrics, San Diego, CA) and Multiple Electrode Platelet Aggregometry (MEA) (Dynabyte, Munich, Germany), for predicting early clinical outcomes after percutaneous coronary intervention. Methods Platelet reactivity in the arachidonic acid–induced and adenosine diphosphate (ADP)–induced platelet aggregation was measured simultaneously with the VerifyNow Assay and MEA in 222 patients undergoing percutaneous coronary intervention between August and October 2009. We investigated the correlations between the 2 tests and performed receiver operating characteristic curve analysis for major adverse cardiovascular events (MACE), a composite of death, myocardial infarction (MI), stroke, and target vessel revascularization, at 30 days. Results Major adverse cardiovascular events occurred in 19 patients (8.6%), including 14 patients with periprocedural MI and 5 patients with stroke. Correlations were weak between the 2 tests in the arachidonic acid–induced (Spearman r = 0.189, P = .006) and ADP-induced platelet reactivity (Spearman r = 0.390, P b .001). Although the VerifyNow P2Y12 Assay (Accumetrics) was able to predict periprocedural MI (area under the aggregation curve 0.680, P = .024) and 30-day MACE (area under the aggregation curve 0.649, P = .032), VerifyNow Aspirin Assay (Accumetrics), MEA ASPI test, and MEA ADP test failed to predict such clinical events. Hyporesponsiveness to clopidogrel based on the VerifyNow Assay was associated with about a 6-fold increased risk of MACE at 30 days. Conclusions Hyporesponsiveness to clopidogrel measured by VerifyNow Assay was able to identify patients with dual antiplatelet therapy who were at higher risk for periprocedural MI and MACE at 30 days. Further randomized studies are required to validate the effectiveness of different platelet function tests for predicting long-term clinical outcomes. (Am Heart J 2011;161:383-90.)

Dual antiplatelet therapy with aspirin and clopidogrel effectively reduces atherothrombotic events and improves short- and long-term clinical outcomes in patients undergoing percutaneous coronary intervention (PCI). However, individual variability exists in response to aspirin or clopidogrel, and hyporesponsiveness to

antiplatelet therapy is associated with increased immediate and late clinical events after PCI.1-3 Therefore, it seems important to identify patients who are hyporesponsive to antiplatelet therapy early and to provide appropriate additional treatment to prevent such adverse cardiac events. Currently, 2 different point-of-care tests,

From the aDivision of Cardiology & Cardiovascular Institute, Severance Cardiovascular Hospital, Yonsei University Health System, Seoul, Korea, bSeverance Biomedical Science Institute, Yonsei University Health System, Seoul, Korea, and cDivision of Cardiology, Seoul National University Bundang Hospital, Seongnam City, Gyeonggi-do, Korea. Drs Young-Guk Ko and Jung-Won Suh contributed equally to the preparation of this manuscript. Submitted April 18, 2010; accepted October 29, 2010.

Reprint requests: Yangsoo Jang, MD, PhD, Division of Cardiology & Cardiovascular Institute, Severance Cardiovascular Hospital, Yonsei University Health System, 250 Seongsanno, Seodaemun-gu, Seoul, Korea (120-752). E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.036

American Heart Journal February 2011

384 Ko et al

the VerifyNow Assay (Accumetrics, San Diego, CA) and Multiple Electrode Platelet Aggregometry (MEA) (Dynabyte, Munich, Germany), are commonly used to evaluate platelet function.3,4 Resistance to clopidogrel diagnosed by these 2 systems has shown to have prognostic value for future cardiovascular events after PCI.5-7 However, there has been no head-to-head comparison between the 2 systems for their ability to predict cardiovascular events. Thus, in the present study, we directly compared the ability of the 2 tests to predict early clinical outcomes in patients undergoing PCI.

Methods Study population Between August and October 2009, consecutive patients with coronary artery disease undergoing PCI in Severance Cardiovascular Hospital, Yonsei University Health System, Seoul, Korea, and in Seoul National University Bundang Hospital, Korea, were enrolled. Inclusion criteria were symptomatic coronary artery disease or documented myocardial ischemia by noninvasive stress tests and angiographic evidence of N50% diameter stenosis. Exclusion criteria were contraindication to antiplatelet agents, previous allergy to or intolerance of aspirin or clopidogrel, concomitant use of warfarin or other antiplatelet agents other than aspirin or clopidogrel, active bleeding, known platelet dysfunction, platelet count ≤100,000/mm3, and hematocrit ≤30%. All patients were pretreated with aspirin (100 mg/d) and clopidogrel (75 mg/d) at least 5 days before the procedure or received oral loading doses of 250 mg aspirin and 300 mg clopidogrel 12 to 24 hours before the procedure. The PCI procedure was performed on the following day according to standard practice, after the patients received additional maintenance doses of 100 mg aspirin and 75 mg clopidogrel. Before PCI, patients received 100 to 140 IU/kg unfractionated heparin with a target activated clotting time ≥250 seconds. All patients were treated with implantation of drug-eluting stents. After PCI, dual antiplatelet therapy with 100 mg/d aspirin and 75 mg/d clopidogrel was continued for at least 6 months. Informed written consent was obtained from all patients, and the study was approved by the local ethical review board in accordance with the Declaration of Helsinki. No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the manuscript, and its final contents.

Blood sampling Blood samples for platelet function tests were obtained from patients in the catheterization room through the indwelling femoral vessel sheath before administration of heparin and PCI. Blood samples were collected in Greiner Bio-One 3.2% citrate Vacuette tubes (Greiner Bio-One, Kremsmünster, Austria) for the VerifyNow Assay and in 4.5-mL plastic tubes containing the anticoagulant lepirudin (25 μg/mL, Refludan, hirudin blood collection tubes; Dynabyte) for MEA. Blood samples were kept at room temperature for at least 30 minutes before platelet function testing and used for testing within 3 hours of blood collection. The platelet function tests were performed by experienced laboratory personnel blinded to clinical data in

each participating investigator's institutions according to the instructions of the device companies. Further blood samples were taken in all patients before and at 8 and 24 hours after PCI to assay creatine kinase-MB (CK-MB) fraction and troponin T. Additional measurements were performed if patients developed postprocedural symptoms suggestive of myocardial ischemia.

VerifyNow Assay The VerifyNow system (Accumetrics) is a turbidimetric-based optical detection system that measures platelet-induced aggregation as an increase in light transmittance in whole blood. VerifyNow Aspirin Assay (Accumetrics) is a cartridge-based assay designed to measure platelet aggregation in the presence of human fibrinogen–coated beads and arachidonic acid (AA). The AA-induced platelet aggregation is quantified with aspirin reaction units (ARU). VerifyNow P2Y12 Assay (Accumetrics) uses adenosine diphosphate (ADP) to activate platelets and then measures platelet aggregation. Platelet reactivity to ADP is quantified as P2Y12 reaction units (PRU). Multiple Electrode Platelet AggregometryMultiple Electrode Platelet Aggregometry was performed with the Multiplate analyzer (Dynabyte), a whole blood impedance aggregometer. One multiplate test cell contains 2 independent sensor units, and 1 unit consists of 2 silver-coated highly conductive copper wires with a length of 3.2 mm. After diluting 1:2 with 0.9% NaCl solution and hirudin-anticoagulated whole blood and stirring in the test cuvettes for 3 minutes at 37°C, AA (ASPI test) or ADP (ADP test) was added, and aggregation was continuously recorded for 5 minutes. The adhesion of activated platelets to the electrodes led to increased impedance, which was detected for each sensor unit separately and transformed to aggregation units (AU) that were plotted against time. Aggregation was quantified as the area under the aggregation curve (AUC [AUd min]).

Definitions Major adverse cardiovascular events (MACE) were defined as a composite of death, myocardial infarction (MI), stroke, and target vessel revascularization from the time of the procedure to 30 days. Periprocedural MI was defined as postprocedural increases of cardiac biomarkers (troponin or CK-MB) N3 times the 99th percentile of the upper limit of normal in patients with normal baseline levels and as a subsequent elevation N3 times in patients with elevated baseline levels.8 Bleeding was defined by the GUSTO bleeding classification;9 bleeding complications were classified as severe or lifethreatening if they were intracerebral or if they resulted in substantial hemodynamic compromise requiring treatment. Moderate bleeding was defined by the need for transfusion. Minor bleeding referred to other bleeding, not requiring transfusion or causing hemodynamic compromise. Bleeding of percutaneous entry site was defined as the size of external hematoma N10 cm for femoral and N2 cm for radial access during or after catheterization laboratory visit until discharge.10

Statistical analysis Continuous variables were reported as mean ± SD, and categorical variables were presented as frequencies and percentages. Continuous variables were compared by unpaired Student t test for normally distributed values; otherwise, the

American Heart Journal Volume 161, Number 2

Mann-Whitney U test was used. Categorical variables were compared by χ2 or Fisher exact test when the expected frequency was b5. Correlations between results obtained by the 2 platelet function tests were evaluated using Spearman correlation coefficient (r). Bland-Altman analysis was performed using MedCalc version 9.3.5.0 (MedCalc Software, Mariakerke, Belgium) to evaluate the agreement between the 2 methods and to determine the limits of the agreement (mean ± 1.96 times the SD of the differences). Ability of the assays to discriminate between patients with and without periprocedural MI or MACE at 30 days was evaluated by receiver operating characteristic (ROC) curve analysis. The optimal cutoff value was calculated by determining the value providing the greatest sum of sensitivity and specificity. Multivariate logistic regression analysis was performed using the Enter method to identify predictors of 30-day MACE, and variables were chosen if the univariate P value was b.15. Odds ratios (OR) were presented with 95% CI. Analyses were performed using SPSS version 13.0 statistical software (SPSS Inc, Chicago, IL); 2-sided P b .05 was considered statistically significant.

Results Initially, 231 consecutive patients were evaluated for platelet reactivity using VerifyNow and MEA assays. However, 9 patients were excluded from the analysis because the 2 assays could not be performed simultaneously in these patients. Baseline clinical and procedural characteristics for 222 patients are presented in Table I. Procedural success was achieved in all patients. No vessel or side branch (≥2 mm) closure or emergent surgical reintervention occurred. Within 30 days, MACE occurred in 19 patients (8.6%), including 14 patients (6.3%) with periprocedural MI and 5 patients (2.3%) with stroke. Death, stent thrombosis, or target vessel revascularization did not occur within 30 days of the procedure. Moderate or severe bleeding was found in 5 patients (2.3%). There were 1 case of each retroperitoneal hemorrhage and hemopericardium and 3 cases of access site bleeding.

Arachidonic acid–induced platelet aggregation Median values for AA-induced platelet aggregation were 420.0 ARU (interquartile range [IQR] 405.0-454.5 ARU) by the VerifyNow Aspirin Assay and 60.0 AUd min (IQR 22.0-100.0 AUd min) by the MEA ASPI test. The AAinduced platelet aggregation by the 2 tests showed a statistically significant but weak correlation (Spearman r = 0.189, P = .006) (Figure 1, A). In addition, limits of agreement between ARU by the VerifyNow Aspirin Assay and AU d min by the MEA ASPI test for the AA-induced platelet aggregation were assessed. The mean value of differences ± SD was 354.5 ± 98.2 with 95% limits of agreement ranging from 161.9 to 547.0 (Figure 1, B). Receiver operating characteristic curve analysis demonstrated that the VerifyNow Aspirin Assay was not able to distinguish between patients with and without 30-day MACE (AUC 0.605, P = .140) or between patients with and without periprocedural MI (AUC 0.602, P = .220)

Ko et al 385

Table I. Baseline clinical and procedural data Total (N = 222) Age (y) Male Body mass index (kg/m2) Diabetes mellitus Insulin-requiring Hypertension Hypercholesterolemia Chronic renal failure Current smoker Previous CVA Previous MI Previous PCI Previous CABG ACS/NSTEMI LVEF b45% Hemoglobin level (g/L) Platelet count (×103/mm3) hsCRP (mg/L) Target vessel Left anterior descending Left circumflex Right coronary artery Left main artery Lesion type B2/C Restenotic lesions Stent diameter (mm) Stents per patient DES per patient Total stent length (mm) Medication (≥5 days) before PCI Statin Aspirin Clopidogrel

63.3 ± 10.0 152 (68.5%) 24.8 ± 3.1 78 (32.0%) 7 (3.2%) 160 (72.1%) 104 (46.8%) 20 (9.0%) 27 (12.2%) 33 (14.9%) 13 (5.9%) 37 (16.7%) 1 (0.5%) 98 (44.2%)/21 (10.1%) 22 (9.9%) 13.3 ± 1.7 255.0 ± 71.3 3.4 ± 11.5 134 (60.4%) 77 (34.7%) 81 (35.6%) 8 (3.6%) 167 (75.2%) 10 (4.6%) 3.2 ± 0.4 1.8 ± 1.0 1.8 ± 1.0 40.4 ± 24.4 123 (55.4%) 199 (89.6%) 160 (72.1%)

Values are expressed as number of patients (%) or mean ± SD. CVA, Cerebrovascular attack; CABG, coronary artery bypass graft; ACS, acute coronary syndrome; NSTEMI, non–ST-segment elevation MI; LVEF, left ventricular ejection fraction; hsCRP, high-sensitivity C-reactive protein; DES, drug-eluting stent.

(Figure 2, A, C). The MEA ASPI test also failed to predict 30-day MACE (AUC 0.602, P = .152) or periprocedural MI (AUC 0.541, P = .622) (Figure 2, A, C).

Adenosine diphosphate–induced platelet aggregation Median values for ADP-induced platelet aggregation were 290.0 PRU (IQR 244.5-365.0 PRU) by the VerifyNow P2Y12 Assay and 249.0 AUd min (IQR 142.5409.5 AUd min) by the MEA ADP test. The ADP-induced platelet aggregation by the 2 tests showed a significant but weak correlation (Spearman r = 0.390, P b .001) (Figure 1, C). Limits of agreement between PRU by the VerifyNow P2Y12 Assay and AU d min by the MEA ADP test for the ADP-induced platelet aggregation were also assessed. The mean value of differences ± SD was −17.1 ± 232.1 with 95% limits of agreement ranging from −472.0 to 437.8 (Figure 1, D). Receiver operating characteristic curve analysis revealed that the VerifyNow P2Y12 Assay was able to predict both MACE (AUC 0.649, P = .032) and

American Heart Journal February 2011

386 Ko et al

Figure 1

periprocedural MI (AUC 0.680, P = .024) (Figure 2, B, D). However, the MEA ADP test did not distinguish either between patients with and without MACE (AUC 0.443, P = .415) or between patients with and without periprocedural MI (AUC 0.419, P = .310) (Figure 2, B, D). A PRU value ≥274 by the VerifyNow P2Y12 Assay was the optimal cutoff point to predict 30-day MACE with sensitivity of 83.3% and specificity of 48.1%, and periprocedural MI with sensitivity of 92.3% and specificity of 47.9%.

Hyporesponsiveness to clopidogrel Hyporesponsiveness to clopidorel according to the cutoff value (PRU ≥274) by VerifyNow P2Y12 Assay was found in 54.5% of the enrolled patients and was associated with increased risk of periprocedural myonecrosis (Figure 3). On univariate analysis, total stent length, hyporesponsiveness to clopidogrel, and no previous use of statin were related to 30-day MACE, whereas gender, age, diabetes, hypertension, hypercholesterolemia, chronic renal failure, previous stroke, acute coronary syndrome, multivessel disease, complex lesions, and non–long-term use of clopidogrel had no impact on early clinical outcomes. Multivariate logistic analysis revealed that total stent length and hyporesponsiveness to clopidogrel were independent predictors of MACE at 30 days (Table II). Patients with hyporesponsiveness to clopidogrel had about a 6-fold increased risk of MACE at 30 days after PCI. Incidence of moderate or severe bleeding was similar between patients with and without hyporesponsiveness to clopidogrel (Figure 3). The numbers of patients from the present study population identified to have clopidogrel resistance according to different definitions were presented in Table III.

Discussion We compared the VerifyNow Assay and MEA for their ability to predict clinical outcomes after PCI. Both platelet function tests showed wide individual variability in the response to aspirin and clopidogrel. However, there was only weak correlation between the 2 tests in AA- and ADP-induced platelet reactivity. Although the VerifyNow P2Y12 Assay discriminated between patients with and without periprocedural MI or MACE at 30 days after PCI,

A, Correlations between platelet reactivities measured by the VerifyNow Aspirin Assay and MEA ASPI test for AA-induced platelet aggregation. B, Bland-Altman plot of platelet reactivities by VerifyNow Aspirin Assay and MEA ASPI test. C, Correlations between platelet reactivities measured by the VerifyNow P2Y12 Assay and MEA ADP test for ADP-induced platelet aggregation. D, Bland-Altman plot of platelet reactivities by VerifyNow P2Y12 Assay and MEA ADP test.

American Heart Journal Volume 161, Number 2

Ko et al 387

Figure 2

Receiver operating characteristic curve for MACE at 30 days (A, VerifyNow Aspirin; B, MEA ASPI test) and periprocedural MI (C, VerifyNow P2Y12 Assay; D, MEA ADP test).

VerifyNow Aspirin Assay, MEA ASPI test, and MEA ADP test failed to predict such clinical events. Hyporesponsiveness to clopidogrel based on the VerifyNow Assay was associated with about a 6-fold increased risk of MACE at 30 days. There are different methods to evaluate platelet function. Light transmittance aggregometry had been the standard test, but it is time intensive, labor intensive, and, therefore, not practical.3,11 The VerifyNow Assay and MEA are commonly used point-of-care test systems. Although results of both test methods correlated with those of light transmittance aggregometry, the VerifyNow Assay showed stronger correlation (r = 0.61) with light transmittance aggregometry than MEA (r = 0.35).4 However, to date, there has been no study directly comparing both test systems for clinical outcomes after PCI.

Similar to our study, Patti et al5 evaluated ADP-induced platelet reactivity by the VerifyNow Assay in patients undergoing PCI and found that residual platelet reactivity despite dual antiplatelet therapy was associated with increased periprocedural MI and MACE by 30 days. Their cutoff value for MACE by 30 days was lower than ours (240 vs 274 PRU); however, its sensitivity (81% vs 83.3%) and specificity (53% vs 48.1%) were similar to those of our study. In an observational study, Price et al6 found that residual platelet reactivity by the VerifyNow Assay was associated with 6-month thrombotic events including stent thrombosis after PCI with drug-eluting stents. A PRU value ≥235 showed a sensitivity of 78% and a specificity of 68% for clinical outcomes at 6 months. Recently, Sibbing et al7 reported that a low response to clopidogrel assessed

388 Ko et al

Figure 3

Incidence of CK-MB elevation and moderate or severe bleeding according to responsiveness (PRU b274 vs PRU ≥274) to clopidogrel by VerifyNow P2Y12 Assay.

with MEA was significantly associated with an increased risk of stent thrombosis. However, the authors did not present data for adverse clinical events other than stent thrombosis. Why the 2 platelet function tests showed different results in our study is not clear. Both test systems use whole blood for testing. However, the working mechanisms of the test devices for the measurement of platelet reactivity are different. Although VerifyNow measures the change in light transmission because of platelet aggregation in a liquid phase, MEA quantifies the change in the impedance of electrodes because of adhesion of activated platelets.2,4 Whether this difference in the working mechanisms of the test devices contributes to the discrepancies of the results is not known. Previous studies have shown that platelet function tests correlate poorly among themselves and are not equally efficient in predicting clinical outcomes.4,12,13 In the present study, the range of values for the MEA assay seemed to be rather limited compared to that of VerifyNow Assay. It is possible that insufficient discrimination among different degrees of platelet aggregation limits the predictive ability of MEA. Several studies also evaluated the aspirin assay to assess the association of aspirin resistance with early and late clinical outcomes in patients treated with PCI.14-16 Chen et al14 reported that aspirin resistance measured by VeryfyNow Assay was associated with a 2.9-fold increased incidence of myonecrosis related to PCI. Various studies have demonstrated a relation between aspirin resistance and increased risk of cardiovascular events.15,16 Furthermore, combined aspirin and clopidogrel resistance has been reported to be associated with higher risk of cardiovascular events.17,18 In the present study, VerifyNow Aspirin Assay and MEA ASPI test failed to predict early clinical outcomes with statistical significance;

American Heart Journal February 2011

however, high residual platelet reactivity measured by the 2 tests showed a trend toward increased incidence of MACE at 30 days. Our data suggest that the presence of hyporesponsiveness to clopidogrel may have a greater impact on early clinical outcomes in patients undergoing PCI compared to hyporesponsiveness to aspirin. The optimal cutoff value (PRU ≥274) of residual platelet reactivity to clopidogrel for clinical outcomes after PCI in our study was higher than those of previous studies under similar clinical conditions (Table III). Patti et al5 suggested PRU ≥240 as an optimal cutoff value of ADP-induced aggregation by VerifyNow for predicting 30-day clinical outcomes. Similarly, Price et al6 estimated PRU ≥235 as an optimal cutoff value for discriminating patients with thrombotic events at 6 months. In previous studies, a higher dose (600 mg) of clopidogrel loading was used, compared with 300 mg in the present study. This might have at least partially contributed to the discrepancy. Furthermore, the ethnic background of the study subjects also plays an important role. Several studies reported that the frequency of CYP2C19⁎2 and CYP2C19⁎3 polymorphisms of hepatic cytochrome p450 isoenzymes in the Asian population is significantly higher than in whites.19,20 These polymorphisms are associated with decreased response to clopidogrel.21,22 Despite higher frequency of these polymorphisms among Asians, stent thrombosis in Asian cohort studies does not seem to exceed that in whites.23,24 The cause of this paradoxical phenomenon is not unknown. Based on the results of our study, the VerifyNow P2Y12 Assay is an efficient point-of-care platelet function test to identify patients with hyporesponsiveness to clopidogrel who are at increased risk for periprocedural MI and early adverse cardiovascular events. Therefore, it seems plausible to screen those patients with poor response to clopidogrel before PCI and to provide additional treatment to prevent periprocedural adverse cardiovascular events and to improve early clinical outcomes. However, in the present study, the total stent length was a more important independent predictor of MACE at 30 days. That the average number of stents per patient was 1.8 and that the mean total stent length was 40 mm indicate that our study included a group of patients with fairly complex lesions and high risk of periprocedural MI and stent thrombosis. Therefore, it appears that, in a patient population with complex lesions, the total length of stents required for covering the lesions may be more strongly associated with risk of periprocedural MI than the hyporesponsiveness to clopidogrel.

Limitations First, this is an observational study. We excluded patients treated with cilostazol or glycoprotein IIb/IIIa inhibitors, in addition to aspirin and clopidogrel before or

American Heart Journal Volume 161, Number 2

Ko et al 389

Table II. Multivariate logistic regression analysis for MACE at 30 days Crude OR (95% CI)

P

Adjusted OR (95% CI)

P

1.03 (1.01-1.04) 5.50 (1.24-24.48) 0.44 (0.17-1.16)

.001 .025 .096

1.03 (1.01-1.05) 5.95 (1.26-28.1) 0.60 (0.24-1.50)

.001 .024 .493

Total stent length (per mm) Hyporesponsiveness to clopidogrel by VerifyNow Statin before PCI

Table III. The numbers of patients from the present study population who were identified to have clopidogrel resistance according to different definitions

Study (y) Present study Price et al6 (2008) Patti et al5 (2008) Jeong et al25 (2008)

Definition of clopidogrel resistance PRU ≥274 PRU ≥235 PRU ≥240 Percent platelet inhibition b20%

No. of patients with clopidogrel resistance 121 174 172 138

(54.5%) (78.4%) (77.5%) (62.2%)

during PCI, because of possible interference of these agents with platelet reactivity by point-of-care platelet function tests. However, not all patients had the same baseline conditions regarding previous antiplatelet therapy. Pretreatment with clopidogrel before admission was noted in 72% of the enrolled patients. Second, the study population included a group of patients at higher risk for periprocedural MI and early cardiovascular adverse events as mentioned above. Therefore, the study subjects in our study may not represent the whole population undergoing PCI. Third, we did not perform conventional light transmission aggregometry as a comparator, which is currently considered the criterion standard method for assessing platelet reactivity to different antiplatelet drugs.3,4 However, the light transmission aggregometry has also disadvantages such as requirement for immediate processing, variable reproducibility, large required sample volumes, and lengthy processing time.3,4,11

Conclusions Hyporesponsiveness to clopidogrel measured by VerifyNow Assay was able to identify patients with dual antiplatelet therapy who were at higher risk for periprocedural MI and early adverse cardiovascular events. Further studies are required to validate the effectiveness of different point-of-care platelet function tests for predicting long-term clinical outcomes.

Disclosures This study was funded by the Healthcare Technology R&D Project, Ministry for Health, Welfare, & Family

Affairs, Republic of Korea (no A085012 and A000385), by a grant from the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea (no A085136), and by the Cardiovascular Research Center, Seoul, Korea.

References 1. Ben-Dor I, Kleiman NS, Lev E. Assessment, mechanisms, and clinical implication of variability in platelet response to aspirin and clopidogrel therapy. Am J Cardiol 2009;104:227-33. 2. Kereiakes DJ, Gurbel PA. Peri-procedural platelet function and platelet inhibition in percutaneous coronary intervention. J Am Coll Cardiol Interv 2008;1:111-21. 3. Gurbel PA, Becker RC, Mann KG, et al. Platelet function monitoring in patients with coronary artery disease. J Am Coll Cardiol 2007;50: 1822-34. 4. Gremmel T, Steiner S, Seidinger D, et al. Comparison of methods to evaluate clopidogrel-mediated platelet inhibition after percutaneous intervention with stent implantation. Thromb Haemost 2009;101: 333-9. 5. Patti G, Nusca A, Mangiacapra F, et al. Point-of-care measurement of clopidogrel responsiveness predicts clinical outcome in patients undergoing percutaneous coronary intervention results of the ARMYDA-PRO (Antiplatelet therapy for Reduction of MYocardial Damage during Angioplasty-Platelet Reactivity Predicts Outcome) study. J Am Coll Cardiol 2008;52:1128-33. 6. Price MJ, Endemann S, Gollapudi RR, et al. Prognostic significance of post-clopidogrel platelet reactivity assessed by a point-of-care assay on thrombotic events after drug-eluting stent implantation. Eur Heart J 2008;29:992-1000. 7. Sibbing D, Braun S, Morath T, et al. Platelet reactivity after clopidogrel treatment assessed with point-of-care analysis and early drug-eluting stent thrombosis. J Am Coll Cardiol 2009;53:849-56. 8. Thygesen K, Alpert JS, White HD. On behalf of the Joint ESC/ACCF/ AHA/WHF task force for the redefinition of myocardial infarction. Universal definition of myocardial infarction. J Am Coll Cardiol 2007;50:2173-95. 9. The GUSTO investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med 1993;329:673-82. 10. Mehta SK, Frutkin AD, Lindsey JB, et al. National Cardiovascular Data Registry. Bleeding in patients undergoing percutaneous coronary intervention: the development of a clinical risk algorithm from the National Cardiovascular Data Registry. Circ Cardiovasc Interv 2009;2:222-9. 11. Michelson AD. Platelet function testing in cardiovascular diseases. Circulation 2004;110:489-93. 12. Lordkipanidzé M, Pharand C, Schampaert E, et al. A comparison of six major platelet function tests to determine the prevalence of aspirin resistance in patients with stable coronary artery disease. Eur Heart J 2007;28:1702-8.

American Heart Journal February 2011

390 Ko et al

13. Breet NJ, van Werkum JW, Bouman HJ, et al. Comparison of platelet function tests in predicting clinical outcome in patients undergoing coronary stent implantation. JAMA 2010;303:754-62. 14. Chen WH, Lee PY, Ng W, et al. Aspirin resistance is associated with a high incidence of myonecrosis after non-urgent percutaneous coronary intervention despite clopidogrel pretreatment. J Am Coll Cardiol 2004;43:1122-6. 15. Snoep JD, Hovens MM, Eikenboom JC, et al. Association of laboratory-defined aspirin resistance with a higher risk of recurrent cardiovascular events: a systematic review and metaanalysis. Arch Intern Med 2007;167:1593-9. 16. Krasopoulos G, Brister SJ, Beattie WS, et al. Aspirin “resistance” and risk of cardiovascular morbidity: systematic review and metaanalysis. BMJ 2008;336:195-8. 17. Lev EI, Patel RT, Maresh KJ, et al. Aspirin and clopidogrel drug response in patients undergoing percutaneous coronary intervention. The role of dual drug resistance. J Am Coll Cardiol 2006;47: 27-33. 18. Ivandic BT, Sausemuth M, Ibrahim H, et al. Dual antiplatelet drug resistance is a risk factor for cardiovascular events after percutaneous coronary intervention. Clin Chem 2009;55:1171-6. 19. Lee JM, Park S, Shin DJ, et al. Relation of genetic polymorphisms in the cytochrome p450 gene with clopidogrel resistance after

20.

21.

22.

23.

24.

25.

drug-eluting stent implantation in Koreans. Am J Cardiol 2009;104: 46-51. Kim IS, Choi BR, Jeong YH, et al. The CYP2C19⁎2 and CYP2C19⁎3 polymorphisms are associated with high post-treatment platelet reactivity in Asian patients with acute coronary syndrome. J Thromb Haemost 2009;7:897-9. Trenk D, Hochholzer W, Fromm MF. Cytochrome P450 2C19 681GNA polymorphism and high on clopidogrel platelet reactivity associated with adverse 1-year clinical outcome of elective percutaneous coronary intervention with drug eluting or bare-metal stents. J Am Coll Cardiol 2008;51:1925-34. Frere C, Cuisset T, Morange PE, et al. Effect of cytochrome p450 polymorphisms on platelet reactivity after treatment with clopidogrel in acute coronary syndrome. Am J Cardiol 2008;101:1088-93. Kimura T, Morimoto T, Nakagawa Y, et al. Antiplatelet therapy and stent thrombosis after sirolimus-eluting stent implantation. Circulation 2009;119:987-95. Park DW, Yun SC, Lee JY, et al. C-reactive protein and the risk of stent thrombosis and cardiovascular events after drug-eluting stent implantation. Circulation 2009;120:1987-95. Jeong YH, Kim IS, Choi BR, et al. The optimal threshold of high post-treatment platelet reactivity could be defined by a point-of-care VerifyNow P2Y12 assay. Eur Heart J 2008;29:2186-7.

AVAILABILITY OF JOURNAL BACK ISSUES As a service to our subscribers, copies of back issues of the American Heart Journal for the preceding 2 years are maintained and are available for purchase from Elsevier until inventory is depleted. Please write to Elsevier Health Sciences Division, Subscription Customer Service, 3251 Riverport Lane, Maryland Heights, MO 63043, or call 800-654-2452 or 314-447-8871 for information on availability and prices of particular issues.

Impact of baseline thrombocytopenia on the early and late outcomes after ST-elevation myocardial infarction treated with primary angioplasty: Analysis from the Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI) trial Diaa A. Hakim, MD, PhD, a,g George D. Dangas, MD, PhD, b,g Adriano Caixeta, MD, PhD, a,g Eugenia Nikolsky, MD, PhD, a,g Alexandra J. Lansky, MD, c,g Jeffrey W. Moses, MD, a,g Bimmer Claessen, MD, a,g Elias Sanidas, MD, a,g Harvey D. White, DSc, d,g E. Magnus Ohman, MD, e,g Steven V. Manoukian, MD, f,g Martin Fahy, MSc, a,g Roxana Mehran, MD, b,g and Gregg W. Stone, MD a,g New York, NY; New Haven, CT; Auckland, New Zealand; Durham, NC; and Atlanta, GA

Background

Thrombocytopenia (TP) is a common abnormality in patients presenting with acute coronary syndrome. Whether baseline TP has any influence on the outcome of patients treated with primary angioplasty for acute myocardial infarction is unknown.

Methods We sought to detect the impact of baseline TP on the early and late outcomes of patients with ST-elevation myocardial infarction in the HORIZONS-AMI trial that included a protocol of immediate angiography and primary percutaneous coronary intervention. Results Baseline TP was found in 4.2% of patients and was associated with a higher incidence of cardiovascular mortality, major bleeding, and major cardiovascular events at short- and long-term follow-up. The 30-day rates of death, major bleeding, major cardiac events, and major cardiac events plus major bleeding were 6.2%, 11.9%, 9.6%, and 18.5% in the TP group, respectively, compared with 2.1%, 7%, 5.2%, and 10.8% in those without TP (P b .05 for all). Similarly, event rates at 2 years were 11.3%, 12.7%, 24.7%, and 30.8% compared with 5.1%, 7.9%, 18.5%, and 23.3% (P b .05). By multivariate analysis, baseline TP was an independent predictor of 30-day net adverse clinical events but not of any 2-year events. Conclusions We found that baseline TP in patients with ST-elevation myocardial infarction undergoing routine angiography and primary percutaneous coronary intervention is strongly associated with early adverse events and is a maker of late events, related to both ischemia and bleeding. (Am Heart J 2011;161:391-6.)

From the aColumbia University Medical Center and the Cardiovascular Research Foundation, New York, NY, bMount Sinai Medical Center and Cardiovascular Research Foundation, New York, NY, cYale University Medical Center, New Haven, CT, dAuckland City Hospital, Auckland, New Zealand, eDuke University Medical Center, Durham, NC, and fEmory University School of Medicine, Atlanta, GA. g For the HORIZONS trial investigators. RCT reg # NCT00433966. J. Michael DiMaio, MD served as guest editor for this article. Submitted July 19, 2010; accepted November 3, 2010. Reprint requests: George D. Dangas, MD, PhD, Cardiovascular Institute (Box 1030), Mount Sinai Medical Center, One Gustave L. Levy Place, New York, NY 10029. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.001

Thrombocytopenia (TP) is considered one of the risk factors for bleeding and other adverse cardiac events in patients presenting with acute coronary syndrome treated with antithrombotic therapies.1-13 This laboratory abnormality is often overlooked because many pharmacologic trials involving antithrombotic or antiplatelet agents have been routinely excluding patients with varying degrees of TP. The incidence of acquired TP was reduced with bivalirudin treatment as compared with heparin plus glycoprotein (GP) IIb/IIIa inhibitor use in several trials.14-16 Baseline TP has been shown to be a predictor of in-hospital mortality in patients undergoing percutaneous coronary intervention (PCI)12; however,

American Heart Journal February 2011

392 Hakim et al

there are insufficient data on the effect of baseline TP on patients with ST-elevation myocardial infarction (STEMI) undergoing a primary PCI strategy. We aimed to evaluate the impact of baseline TP on early and late outcomes of primary PCI of patients with STEMI in the patient population of the HORIZONS-AMI trial.16

Methods The HORIZONS-AMI trial was a prospective, open-label, randomized, multicenter study comparing bivalirudin with heparin plus GP IIb/IIIa inhibitor in patients with STEMI undergoing immediate angiography and primary PCI. The study protocol and the results have been previously described in detail.16 Per study protocol, patients with bleeding diathesis or known platelet count of b100,000/mm3 were excluded from HORIZONS; however, the acuity of STEMI clinical presentation and the necessity to expedite emergency door-to-balloon time allowed enrollment of patients before the complete blood count results would be available. As part of this analysis, baseline TP was defined as platelet count b150,000 platelets/mm3 (reference lower limit of normal) at blood specimen obtained in the hospital before the time of randomization; 146 patients were identified in the TP group (4.2%), and the rest served as control (n = 3,300).

Clinical end points Two primary end points were prespecified at 30 days and 2 years: major bleeding (not related to coronary artery bypass graft [CABG]) and the net adverse clinical events (NACE) defined as the combination of major bleeding or the composite of major adverse cardiovascular events (MACE), including death, reinfarction, target vessel revascularization (TVR) for ischemia, and stroke. Major bleeding was defined as intracranial or intraocular hemorrhage; retroperitoneal bleeding; bleeding at the access site, with hematoma ≥5 cm in diameter or required intervention; reduction of hemoglobin level of 4 g/dL without an overt source of bleeding or ≥3 g/dL with an overt source of bleeding; reoperation for bleeding; or blood transfusion. Cardiac death was defined as death due to acute myocardial infarction (MI), cardiac perforation, cardiac tamponade, arrhythmia or conduction abnormalities, stroke, procedural complications, or any death for which a cardiac cause can not be excluded. Stent thrombosis was defined as definite or probable by the Academic Research Consortium criteria.17

Statistical analysis All categorical variables were expressed as percentages and compared with Fisher exact test; continuous variables were expressed as median (interquartile range) and compared with Student t test. Multivariate analyses with Cox proportional hazards methods derived the independent predictors of adverse events; variables included in these analyses were baseline creatinine clearance, history of congestive heart failure, baseline platelet count b150,000 platelets/mm3, baseline TIMI 0/1, number of diseased vessels, age, gender, clopidogrel loading 600 mg versus 300 mg, bivalirudin, Killip class 1 versus higher, anemia, hypertension, white blood cells 1,000-U increase, medications at discharge, and current smoking status.

Table I. Clinical characteristics TP (n = 146)

Control (n = 3330)

P

Age (y), 67.6 (58.3-74.3) 59.8 (52.3-69.4) b.0001 median (range) Male 82.9% 76.4% .07 Hypertension 54.8% 53.3% .73 Hyperlipidemia 40.3% 43.4% .46 Chronic renal 5.5% 2.9% .07 insufficiency Diabetes mellitus 22.6% 16.3% .04 Creatinine clearance 80.3 (60.6-105.3) 89.3 (68.9-114.3) .0009 (mL/min) Previous MI 18.5% 10.3% .001 Previous CABG 6.2% 2.7% .03 History of CHF 6.8% 2.7% .009 Treatment strategy Primary PCI 88.4% 93.2% .02 CABG 1.8% 1.8% .75 Medical treatment 9.6% 4.8% .01 alone MI, Myocardial Infarction; CABG, coronary Artery Bypass Graft; CHF, congestive heart failure; PCI, Percutaneous Coronary Intervention.

No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this subanalysis, all study analyses, the drafting and editing of the manuscript, and its final content.

Results The overall rate of baseline TP was 4.2% (n = 146). Patients in the TP group were more likely to be older; smokers; persons with diabetes; and have lower creatinine clearance, previous infarction, PCI, coronary bypass surgery (CABG), and a history of congestive heart failure than those in the control group (Table I). With respect to prerandomization medications, the use of aspirin was more frequent in the TP group than in the control group, whereas the use of thienopyridine, heparin, or GP IIb/IIIa before randomization was similar between the 2 groups. There was no significant difference regarding randomization to neither bivalirudin nor heparin preprocedure or during catheterization between the 2 groups, whereas clopidogrel use at discharge was lower in the TP group than in the control group (Table II). Patients in the TP group underwent primary PCI less frequently than control patients, and the reverse was evident with respect to assignment to medical therapy alone after diagnostic angiography. Both groups were similar in the frequency of deferred PCI and CABG.

Clinical outcomes at 30 days Patients in the TP group had higher 30-day NACE because of both significantly higher MACE rate as well as higher major bleeding compared to patients without TP (Table III). The higher MACE rate in the TP group included significantly higher mortality, both

American Heart Journal Volume 161, Number 2

Hakim et al 393

Table II. Antiplatelets and antithrombotic medications TP (n = 146) Aspirin Before hospitalization During hospitalization At discharge At 30 days Clopidogrel Before hospitalization During hospitalization At discharge At 30 days Ticlopidine Before hospitalization During hospitalization At discharge Heparin Prerandomization Preprocedure In catheterization laboratory Bivalirudin In catheterization laboratory GP IIb/IIIa inhibitor In emergency department In catheterization laboratory

Control (n = 3330)

P

34.2% 99.3% 97.1% 97.0%

(50/146) (145/146) (132/136) (129/133)

26.8% 99.8% 97.7% 97.1%

(929/3472) (3321/3328) (3184/3259) (3046/3138)

.03 .29 .55 .7961

2.7% 93.2% 85.4% 86.5%

(4/146) (136/146) (117/137) (115/133)

3.5% 98.6% 91.6% 91.15%

(117/3326) (3280/3326) (2985/3258) (2861/3271)

.61 .0001 .01 .06

0.7% (1/146) 2.1% (3/146) 1.5% (2/137)

0.3% (10/3326) 1.35% (44/3326) 2.0% (46/3258)

.37 .44 1.00

66.7% (97/146) 73.6% (107/146) 55.6% (81/146)

66.3% (2207/3330) 71.8% (2390/3330) 50.6% (1519/3330)

.93 .63 .25

44.1% (64/146)

49.1% (1474/3330)

.24

8.3% (12/146) 56.3% (82/146)

6.4% (213/3330) 50.9% (1694/3330)

.35 .21

Table III. Clinical outcomes at 30-day and 2-year follow-up (Kaplan-Meier estimates)

30-d adverse events (% [n]) NACE, MACE, or major bleeding (non-CABG) MACE (death, MI, ischemic TVR, stroke) Death Cardiac death Ischemic TVR Stent thrombosis (definite/probable) Major bleeding (non–CABG-related) Drop in mercury ≥3 g/dL with overt source Blood product transfusion Major bleeding (including CABG-related) 2-y clinical events (% [n]) MACE (death, MI, ischemic TVR, stroke) Death Cardiac death Ischemic TVR Stent thrombosis (definite/probable)

TP (n = 146)

Non-TP (n = 3330)

P

18.5 (27) 9.6 (14) 6.2 (9) 6.2 (9) 4.2 (6) 4.9 (6) 11.9 (17) 4.2 (6) 8.4 (12) 15.4 (22)

10.8 (359) 5.2 (173) 2.3 (78) 2.1 (69) 2.2 (79) 2.4 (70) 7 (232) 1.9 (62) 2.8 (94) 9.1 (302)

.003 .02 .003 .002 .14 .07 .02 .05 .002 .01

24.7 (35) 11.3 (16) 7.7 (11) 15 (20) 5.8

18.5 (595) 5.1 (165) 3.1 (101) 11.8 (366) 4.4

.04 b.001 .002 .2 .42

all-cause and cardiac. The rate of stent thrombosis was comparable between the TP and control groups. The higher bleeding rate in the TP group included significantly more blood product transfusion and higher CABG-related major bleeding than in the control group (Table III). Most bleeding episodes occurred in the first 30 days, especially the first 10 days, of followup (Figure 3). Multivariate analysis showed that baseline TP was an independent predictor of NACE at 30 days (Table IV).

Clinical outcomes at 2 years At 2-year follow-up, patients with baseline TP continued to have a higher rate of NACE than control patients because of significantly higher MACE (Figure 1), as well as higher bleeding rate (Table III). All-cause mortality was higher in the TP group than in the control group (11.3% vs 5.1%, P = .001) (Figure 2) and included a higher rate of cardiac mortality (7.7% vs 3.1%, P = .002). Table V summarizes the list of independent predictors of mortality and MACE at 2-year follow-up,

American Heart Journal February 2011

394 Hakim et al

Baseline creatinine clearance b60 History of CHF Baseline platelet count b150 000 platelets/mm3 Baseline TIMI 0/1 No. of diseased vessels Age Clopidogrel 600 mg vs 300 mg Bivalirudin Male Killip class 1

Hazard ratio (95% CI)

P

2.41 (1.86-3.12) 1.77 (1.03-3.04) 1.74 (1.04-2.71)

b.0004 .04 .03

1.51 1.24 1.01 0.75

(1.41-2.01) (1.07-1.43) (1.00-1.03) (0.58-0.96)

.004 .005 .03 .02

0.70 (0.55-0.89) 0.65 (0.50-0.85) 0.50 (0.35-0.70)

Figure 2 Thrombocytopenia No Thrombocytopenia 15

11.3%

Death (%)

Table IV. Multivariate analysis. Independent predictors of 30-day NACE

10

5.1%

5

.004 .002 b.0001

HR: 2.31 [95% CI: 1.38, 3.86] P = <.001 0

0

2

4

6

8

10

12

14

16

18

20

22

24

Time in Months Number at risk 146 Thrombo No Thrombo 3330

Figure I Thrombocytopenia No Thrombocytopenia

128 3134

127 3100

126 3063

122 2995

121 2974

104 2495

Kaplan-Meier event curves for patients with and without baseline TP: 2-year cumulative all-cause mortality.

40 35

MACE (%)

30 25

24.7%

20

Table V. Multivariate analyses. Independent predictors of 2-year mortality and MACE

18.5%

Variable

15 10 5

HR: 1.43 [95% CI: 1.02, 2.01] P = .039

0 0

2

4

6

8

10

12

14

16

18

20

22

24

Time in Months Number at risk Thrombo 146 No Thrombo 3330

120 2993

114 2904

113 2822

108 2633

105 2583

88 2141

Kaplan-Meier event graphs for patients with and without baseline TP: 2-year cumulative MACE rates.

which did not include baseline TP. The bleeding events (Figure 3) were essentially unchanged beyond the 30-day time point.

Discussion The present analysis of a large multicenter, randomized trial of patients undergoing PCI for acute MI documented that baseline TP was present in a smaller number of patients (4.2%) compared with 11% in the ACUITY trial15

P

Predictors of 2-y mortality (all-cause) Anemia .002 Age (10-y increase) b.0001 History of hypertension .0327 LVEF (10% decrease) .0004 WBC (1000-U increase) .0006 Aspirin at discharge .0022 Predictors of 2-y MACE Insulin treatment .0027 Anemia .0004 Current smoker status .1229 Aspirin at discharge .0014

Hazard ratio (CI 95%)

2.54 2.36 1.91 1.38 1.12 0.23

(1.40-4.58) (1.82-3.07) (1.05-3.46) (1.15-1.65) (1.05-1.19) (0.09-0.59)

1.68 1.56 1.15 0.41

(1.20-2.37) (1.22-1.99) (0.96-1.37) (0.24-0.71)

LVEF, Left ventricular ejection fraction; WBC, white blood cells.

and 7% in the CRUSADE trial.18 The lower incidence of baseline TP in this study in comparison to the above 2 studies might be attributed in part to the fact that the previous studies were conducted in patients with unstable angina treated with different medications and also to differences in baseline patient characteristics. Patients in this study had less incidence of comorbid conditions such as hypertension, diabetes, renal insufficiency, and history of MI and/or CABG than these previous studies; incidence of hypertension, diabetes, previous MI, and previous CABG in CRUSADE was 72%,

American Heart Journal Volume 161, Number 2

Hakim et al 395

Figure 3 Thrombocytopenia No Thrombocytopenia

Major Bleed (%)

20

15 12.7% 10 7.9% 5 HR: 1.65 [95% CI: 1.03, 2.67] P = .036 0 0

2

4

6

8

10

12

14

16

18

20

22

24

Time in Months Number at risk Thrombo 146 No Thrombo 3330

113 2932

112 2899

111 2864

108 2798

107 2778

94 2335

Kaplan-Meier event curves for patients with and without baseline TP: 2-year cumulative major bleeding.

36%, 32%, and 22%, respectively, whereas it was 67%, 70%, 30%, and 18%, respectively, in ACUITY. In addition, a significant portion of patients in ACUITY and CRUSADE were treated with oral antiplatelet agents and heparin infusion before randomization compared with patients in HORIZONS. Thrombocytopenia was associated with many baseline risk features, including older age, male gender, and diabetes mellitus, previous evidence of MI or revascularization, and preadmission aspirin therapy. The same has been observed in other studies.7,15,18 Therefore, it appears that this hematologic parameter can be affected by several systemic conditions and, at the same time, may confer a readily available risk-stratification parameter. In the present study, both the 30-day and the 2-year follow-up results indicated higher ischemic bleeding and composite end points by univariate analysis in relation to baseline TP, supporting the use of this parameter as a readily available marker of risk. The higher rate of MACE and NACE that occurred in the first 30 days in the TP versus control groups may be due to the higher incidence of major bleeding during the first 10 days. Baseline TP appeared to have altered the rate of prescription of antiplatelet therapy, particularly clopidogrel, at discharge (Table II), which may have in turn contributed to increased adverse cardiac events in the TP group. This high rate of adverse events was comparable with other trials that studied the effect of acquired TP on the outcome in different patient subsets and follow-up time frames. In the ACUITY trial,15 patients with acquired TP compared with patients without acquired TP had

significantly higher rates of mortality, MI, and composite ischemic events at both 30-day and 1-year followups. In addition, the severity of acquired TP strongly correlated with unfavorable ischemic outcomes at follow-up. In the PURSUIT trial,7 the primary end point composed of death and nonfatal MI at 30-day follow-up was higher in baseline TP compared with that in the control group, which was attributed to both higher mortality and MI rates. Patients with TP showed a significantly higher rate of bleeding compared to patients with a normal platelet count. The multivariate regression models for predictors of bleeding and mortality in the previous studies demonstrated that TP was independently correlated with an increased risk of bleeding and mortality at 30 days after unstable angina. Regarding timing of risk stratification, a risk conferred by baseline TP is easily accessible on admission, whereas acquired TP assessment would be available only at discharge. In summary, we found that although baseline TP was an independent predictor of NACE at 30 days, it was not an independent factor of mortality or MACE at 2-year follow-up. Therefore, this laboratory abnormality appears to be a very useful, readily available marker of adverse events, particularly within the early time frame.

Study limitations A larger sample size of patients with TP might have afforded greater statistical power. Because this study included patients with relatively mild TP, the fact that significant correlation with clinical adverse events was documented poses the question of an even greater possible relationship between more severe TP and adverse clinical events. Because of the exclusion of patients with severe TP from the HORIZONS-AMI study (similar to the exclusion criteria in any pharmacologic trial), we were not able to address this question. A larger sample size in the TP group might have also allowed an assessment of the time course of TP according to the various therapies administered. Finally, there are limitations on our understanding of platelet function in relation to various forms of TP; such analyses were not possible because of the absence of dedicated blood specimen withdrawal and were beyond the scope of the present article.

Disclosures Drs Dangas and Mehran have received speaker honoraria from Astra Zeneca, Sanofi Aventis, Cordis, and The Medicines Company and research grant from Sanofi–Bristol Myers Squibb. Dr Lansky has received research grants from The Medicines Company, Cordis, Boston Scientific, Medtronic, and Abbott. Dr Stone has received research support from The Medicines Company, Abbott Vascular, and Boston Scientific.

396 Hakim et al

References 1. Sideris SK, Bonios MJ, Eftihiadis EE, et al. Severe thrombocytopenia after heparin therapy in a patient with unstable angina and recent stent implantation. Hellenic J Cardiol 2005;46:242-6. 2. Bovill EG, Terrin ML, Stump DC, et al. Hemorrhagic events during therapy with recombinant tissue-type plasminogen activator, heparin, and aspirin for acute myocardial infarction. Results of the Thrombolysis in Myocardial Infarction (TIMI), Phase II Trial. Ann Intern Med 1991;115:256-65. 3. Merlini PA, Rossi M, Menozzi A, et al. Thrombocytopenia caused by abciximab or tirofiban and its association with clinical outcome in patients undergoing coronary stenting. Circulation 2004;109:2203-6. 4. Salengro E, Mulvihill NT, Farah B. Acute profound thrombocytopenia after use of eptifibatide for coronary stenting. Catheter Cardiovasc Interv 2003;58:73-5. 5. Kereiakes DJ, Berkowitz SD, Lincoff AM, et al. Clinical correlates and course of thrombocytopenia during percutaneous coronary intervention in the era of abciximab platelet glycoprotein IIb/IIIa blockade. Am Heart J 2000;140:74-80. 6. Makoni SN. Acute profound thrombocytopenia following angioplasty: the dilemma in the management and a review of the literature. Heart 2001;86:E18. 7. McClure MW, Berkowitz SD, Sparapani RR, et al. Clinical significance of thrombocytopenia during a non–ST-elevation acute coronary syndrome. The platelet glycoprotein IIb/IIIa in unstable angina: receptor suppression using integrilin therapy (PURSUIT) trial experience. Circulation 1999;99:2892-900. 8. Nagge J, Jackevicius C, Dzavik VD, et al. Acute profound thrombocytopenia associated with eptifibatide therapy. Pharmacotherapy 2003;23:374-9. 9. Harrington RA, Sane DC, Califf RM, et al. Clinical importance of thrombocytopenia occurring in the hospital phase after administration of thrombolytic therapy for acute myocardial infarction. The Thrombolysis and Angioplasty in Myocardial Infarction study group. J Am Coll Cardiol 1994;23:891-8.

American Heart Journal February 2011

10. Manor SM, Guillory GS, Jain SP. Clopidogrel-induced thrombotic thrombocytopenic purpura-hemolytic uremic syndrome after coronary artery stenting. Pharmacotherapy 2004;24:664-7. 11. Best PJ, Mathew V, Markovic SN. Clopidogrel-associated autoimmune thrombocytopenic purpura. Catheter Cardiovasc Interv 2004; 62:339-40. 12. Nikolsky E, Sadeghi HM, Effron MB, et al. Impact of in-hospital acquired thrombocytopenia in patients undergoing primary angioplasty for acute myocardial infarction. Am J Cardiol 2005;96: 474-81. 13. Berkowitz SD, Sane DC, Sigmon KN, et al. Occurrence and clinical significance of thrombocytopenia in a population undergoing high risk percutaneous coronary revascularization. Evaluation of c7E3 for the Prevention of Ischemic Complications (EPIC) study group. J Am Coll Cardiol 1998;32:311-9. 14. Lincoff AM, Bittl JA, Harrington RA, et al. Bivalirudin and provisional glycoprotein IIb/IIIa blockade compared with heparin and planned glycoprotein IIb/IIIa blockade during percutaneous coronary intervention: REPLACE-2 randomized trial. JAMA 2003; 289:853-63. 15. Stone GW, Bertrand ME, Moses JW, et al. ACUITY Investigators. Routine upstream initiation vs deferred selective use of glycoprotein IIb/IIIa inhibitors in acute coronary syndromes: the ACUITY trial. JAMA 2007;297:591-602. 16. Stone GW, McLaurin BT, Cox DA, et al. Bivalirudin for patients with acute coronary syndromes. The HORIZONS-AMI. N Engl J Med 2006;355:2203-16. 17. Cutlip DE, Windecker S, Mehran R, et al. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 2007;115:2344-51. 18. Wang TY, Ou FS, Roe MT, et al. Incidence and prognostic significance of thrombocytopenia developed during acute coronary syndrome in contemporary clinical practice (subanalysis of the CRUSADE registry). Circulation 2009;119: 2454-62.

Temporal changes in the outcomes of patients with diabetes mellitus undergoing percutaneous coronary intervention in the National Heart, Lung, and Blood Institute dynamic registry Elizabeth M. Holper, MD, MPH, a J. Dawn Abbott, MD, b Suresh Mulukutla, MD, c Helen Vlachos, MSc, c Faith Selzer, PhD, c Darren McGuire, MD, MHSc, a David P. Faxon, MD, d Warren Laskey, MD, e Vankeepuram S. Srinivas, MD, f Oscar C. Marroquin, MD, c and Alice K. Jacobs, MD g Dallas, TX; Providence, RI; Pittsburgh, PA; Boston, MA; Albuquerque, NM; and New York, NY

Background Patients with diabetes mellitus (DM) are at higher risk for adverse outcomes following percutaneous coronary intervention (PCI). Methods To determine whether outcomes have improved over time, we analyzed data from 2,838 consecutive patients with medically treated DM, including 1,066 patients (37.6%) treated with insulin, in the National Heart, Lung, and Blood Institute Dynamic Registry undergoing PCI registered in waves 1 (1997-1998), 2 (1999), 3 (2001-2002), 4 (2004), and 5 (2006). We compared baseline demographics and 1-year outcomes in the overall cohort and in analyses stratified by recruitment wave and insulin use. Results Crude mortality rates by chronological wave were 9.5%, 12.5%, 8.9%, 11.6%, and 6.6% (P valuetrend = .33) among those treated with insulin and, respectively, 9.7%, 6.5%, 4.1%, 5.4%, and 4.7% (P valuetrend = .006) among patients treated with oral agents,. The adjusted hazard ratios of death, myocardial infarction (MI), and overall major adverse cardiovascular events (death, MI, revascularization) in insulin-treated patients with DM in waves 2 to 5 as compared with wave 1 were either higher or the same. In contrast, the similar adjusted hazard ratios for oral agent–treated patients with DM were either similar or lower. Conclusions Significant improvements over time in adverse events by 1 year were detected in patients with DM treated with oral agents. In insulin-treated diabetic patients, despite lower rates of repeat revascularization over time, death and MI following PCI have not significantly improved. These findings underscore the need for continued efforts at optimizing outcomes among patients with DM undergoing PCI, especially those requiring insulin treatment. (Am Heart J 2011;161:397-403.e1.)

The number of patients with diabetes mellitus (DM) who undergo invasive cardiac evaluation has been increasing with the rising number of Americans diagnosed with the disorder. The total prevalence of diabetes

From the aDivision of Cardiology, University of Texas Southwestern Medical Center, Dallas, TX, bDivision of Cardiology, Rhode Island Hospital, Providence, RI, cGraduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, dDepartment of Internal Medicine, Brigham and Women's Hospital, Boston, MA, eDivision of Cardiology, University of New Mexico, Albuquerque, NM, fDivision of Cardiology, Montefiore Medical Center, New York, NY, and gDivision of Cardiology, Boston University Medical Center, Boston, MA. Submitted November 3, 2009; accepted November 7, 2010. Reprint requests: Elizabeth M. Holper, MD, MPH, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8837. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.004

in the United States is projected to more than double (from 5.6% to 12.0%) from 2005 to 2050.1 Patients with DM are at higher risk of both short- and long-term adverse events following percutaneous coronary intervention (PCI).2-5 Specifically, patients with DM are at higher risk for restenosis, target vessel revascularization, late myocardial infarction (MI), and death. Furthermore, the subset of patients with DM treated with insulin is a particularly high-risk group.6-8 An evaluation of temporal trends in outcomes of patients with DM undergoing PCI has not been reported. Also lacking are data assessing the association between treatment regimens for DM and clinical outcomes over time. We evaluated the outcome of patients with DM who underwent PCI in multiple sequential waves of the National Heart, Lung, and Blood Institute (NHLBI) Dynamic Registry to evaluate the temporal outcome of such patients.

American Heart Journal February 2011

398 Holper et al

Methods Study population The specific methodologies and characteristics of the NHLBI Dynamic Registry have been reported previously.9 In brief, data were collected on approximately 2,000 consecutive patients undergoing PCI during 5 recruitment “waves” across 27 clinical centers and incorporate an enriched sample of women and minorities (wave 1: July 1997-February 1998; wave 2: FebruaryJune 1999; wave 3: October 2001-March 2002; wave 4: February-May 2004; wave 5: February-August 2006). Patients were contacted via telephone interview at 1 year by trained nurse coordinators to assess vital status, symptoms, coronary events or cardiac-related hospitalizations, and medication status. Informed consent was obtained for all patients, and the study protocol was approved by Institutional Review Boards at the respective clinical sites and at the University of Pittsburgh data coordinating center. The present analyses comprise data from patients with medically treated DM undergoing PCI in waves 1 to 5.

Definitions Major in-hospital complications (death from any cause, MI, or coronary artery bypass graft [CABG]) were recorded. Myocardial infarction was defined as evidence of 2 or more of the following: (1) typical chest pain N20 minutes' duration not relieved by nitroglycerin, (2) serial electrocardiogram recordings showing changes from baseline or serially in ST-T and/or Q waves in ≥2 contiguous leads, (3) serum enzyme elevation of creatine kinase–MB N5% (total creatine kinase N2× normal, lactate dehydrogenase subtype 1 N lactate dehydrogenase subtype 2, or troponin N0.2 μg/mL), or (4) new wall motion abnormalities. A composite outcome of major adverse coronary event (MACE) rate was defined as a composite of death, MI, and repeat revascularization. Successful lesion dilation was defined as an absolute 20% reduction in lesion severity with a final stenosis ≤50%. Angiographic success was defined as successful treatment of all lesions.

Statistical analysis A stratified analysis was performed because (1) interaction terms are often difficult to interpret by clinical personnel and (2) direct interpretation of the point estimates could be made for each of our diabetes groups. Trends in baseline clinical, angiographic, and procedural characteristics between the waves were compared by Mantel-Haenszel test for trend for discrete variables and the Jonckheere-Terpstra test for continuous variables. Clinical event rates at 1 year were calculated via the Kaplan-Meier approach and compared using the log-rank test for trend. Patients who did not experience the outcome of interest were censored at the last known date of contact or at 1 year if contact extended beyond 1 year. The association between recruitment wave and 1-year safety (death, MI, death/ MI, and death/MI/CABG) and efficacy (CABG, repeat PCI) events were evaluated using Cox proportional hazards regression modeling. The reference group for all models was wave 1. Important covariate variables considered for all models were initially screened for univariable association with each outcome of interest where P b .20. For the final model, the recruitment wave variable was forced to stay in the model; and the remaining

Table I. Demographic features of diabetic patients by enrollment wave and medical therapy Wave Wave Wave Wave Wave Characteristic 1 2 3 4 5 No. of patients Total 553 505 Insulin 215 193 Oral agent 338 312 Mean age Total 63.6 63.1 Insulin 63.9 62.2 Oral agent 63.5 63.7 Female (%) Total 45.9 48.1 Insulin 56.3 53.9 Oral agent 39.3 44.6 Hypertension (%) Total 73.1 78.4 Insulin 78.5 79.2 Oral agent 69.6 77.9 Congestive heart failure (%) Total 18.4 15.8 Insulin 23.1 21.6 Oral agent 15.3 12.2 Prior PCI (%) Total 33.2 36.3 Insulin 36.0 38.3 Oral agent 31.5 35.0 Prior CABG (%) Total 21.4 22.6 Insulin 27.9 29.5 Oral agent 17.2 18.2 Prior MI (%) Total 39.8 39.6 Insulin 40.9 47.4 Oral agent 39.1 34.9 Renal disease (%) Total 7.3 8.7 Insulin 11.3 12.6 Oral agent 4.8 6.4 Revascularization indication (%) Asymptomatic CAD Total 1.8 5.0 Insulin 2.3 6.2 Oral agent 1.5 4.2 Stable angina Total 23.0 19.8 Insulin 20.1 19.2 Oral agent 24.9 20.2 Unstable angina Total 49.6 49.5 Insulin 53.3 52.8 Oral agent 47.3 47.4 Acute MI Total 19.4 22.4 Insulin 19.6 18.1 Oral agent 19.2 25.0

527 194 333

611 229 382

P value trend

642 235 407

65.2 64.2 65.7

64.3 63.5 64.8

63.6 62.5 64.2

.53 .65 .26

43.5 49.0 40.2

40.9 47.6 36.9

37.5 45.1 33.2

.0002 .007 .01

86.3 88.7 84.9

87.6 88.9 86.8

88.5 88.7 88.4

b.0001 b.0001 b.0001

22.9 26.8 20.7

13.1 18.9 9.7

15.9 22.4 12.2

.11 .64 .11

41.3 42.3 40.8

37.9 34.9 39.8

43.9 46.9 42.3

b.0001 .006 .0003

25.6 29.9 23.1

23.8 28.5 44.1

25.6 33.6 21.0

.21 .42 .28

30.7 31.2 30.4

29.1 28.1 29.7

27.8 31.5 25.8

b.0001 .0003 b.0001

14.3 22.2 9.6

17.0 27.3 10.8

17.6 29.3 10.8

b.0001 b.0001 .0006

9.9 7.2 11.4

9.2 8.7 9.4

14.8 14.5 15.0

b.0001 b.0001 b.0001

20.1 19.6 20.4

24.7 21.8 26.4

17.0 15.3 18.0

.15 .39 .24

45.2 42.3 46.8

37.6 39.3 36.6

37.3 37.4 37.2

b.0001 b.0001 .0001

21.8 28.9 17.7

22.6 23.6 22.0

24.3 23.8 24.6

.06 .14 .21

CAD, Coronary artery disease.

covariates (Pentry b .15, Pstay b .05) identified from the initial screening were assessed using backward stepwise process. The proportionality assumption was assessed for all Cox proportional-hazards models graphically, and the hazard ratios represent

American Heart Journal Volume 161, Number 2

Holper et al 399

Table II. Procedural characteristics of diabetic patients by enrollment wave and medical therapy Wave 1

Wave 2

Patient level (n) Total 553 505 Insulin 215 193 Oral agent 338 312 Mean ejection fraction Total 52.4 51.6 Insulin 50.5 48.3 Oral agent 53.7 53.8 Vessel disease (%) Total Single 34.0 33.3 Double 32.4 35.2 Triple 33.5 30.9 Insulin Single 32.1 28.5 Double 28.8 38.3 Triple 39.2 32.6 Oral agent Single 35.2 36.2 Double 34.6 33.3 Triple 29.9 29.8 Mean significant lesions Total 3.4 3.2 Insulin 3.5 3.4 Oral agent 3.3 3.0 No. of lesions attempted Total 1 63.9 70.4 2 27.2 22.5 N2 8.8 7.2 Insulin 1 65.6 70.3 2 30.2 21.9 N2 7.9 3.7 Oral agent 1 62.9 70.4 2 25.2 22.8 N2 11.9 6.7 GP IIb/IIIa inhibitor (%) Total 23.9 30.1 Insulin 20.5 30.6 Oral agent 26.0 29.8 Lesion level (n) Total 823 705 Insulin 300 274 Oral agent 523 431 Mean lesion length Total 12.7 13.6 Insulin 12.9 13.2 Oral agent 12.6 13.8 Evidence of thrombus (%) Total 16.7 13.5 Insulin 16.5 10.7 Oral agent 16.8 15.4 Calcified lesion (%) Total 28.3 25.8 Insulin 27.9 31.2 Oral agent 28.6 22.4

Wave 3

527 194 333 49.4 48.4 50.0

Wave 4

611 229 382 50.8 51.8 50.2

Wave 5

.56 .45 .14 .006

29.8 27.7 42.1

27.7 34.4 37.3

Wave 1

P value trend

642 235 407 51.2 49.3 52.2

Table II (continued )

28.7 32.1 38.6

Stent use (%) Total 55.5 Insulin 52.3 Oral agent 57.4 Drug-eluting stent use (%) Total 0.0 Insulin 0.0 Oral agent 0.0 Angiographic success (%) Total 94.5 Insulin 93.3 Oral agent 95.2

Wave 2

Wave 3

Wave 4

Wave 5

P value trend

68.8 74.5 65.2

74.3 71.7 75.8

89.8 88.7 90.5

91.7 91.2 92.0

b.0001 b.0001 b.0001

0.0 0.0 0.0

0.0 0.0 0.0

67.2 68.3 66.5

83.6 83.4 83.7

b.0001 b.0001 b.0001

95.7 96.0 95.6

96.8 96.2 97.2

96.0 96.1 95.9

96.6 96.1 96.6

.05 .13 .27

GP, Glycoprotein.

.23 25.3 25.3 49.0

27.1 34.9 37.6

25.5 32.8 41.3

32.4 29.1 38.1

28.0 34.0 37.2

30.5 31.7 37.1

3.6 3.9 3.5

3.4 3.7 3.3

3.4 3.8 3.3

64.9 25.4 9.7

72.5 21.8 5.7

69.3 22.9 7.8

64.9 24.2 1.5

71.2 24.5 1.3

68.9 23.0 1.7

64.9 26.1 9.0

73.3 20.2 6.5

69.5 22.9 2.0

51.4 48.2 53.3

33.1 38.4 29.8

31.2 29.8 31.9

.009

.05 .16 .11 .02

.72

.008

778 286 492

823 309 514

.01 .01 .23

903 331 572

13.6 13.9 13.5

16.0 15.7 16.1

17.2 17.6 17.1

b.0001 b.0001 b.0001

10.3 15.4 7.6

10.0 10.6 9.7

11.9 13.9 10.8

.0005 .40 .0001

20.5 22.4 19.4

23.7 29.3 20.4

34.9 33.2 35.9

.01 .25 .02

the average risk during the year after the index procedure. The complete list of variables that were included in the regression models is in the online Appendix. This study was supported by grant number HL-33292 from the NHLBI of the National Institutes of Health. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.

Results We compared baseline demographics and 1-year outcomes in the overall cohort and in analyses stratified by (a) recruitment wave and (b) insulin use. The study cohort comprised 2,838 patients with medically treated DM from waves 1 (insulin treated n = 215, oral agent n = 338), 2 (insulin n = 193, oral agent n = 312), 3 (insulin n = 194, oral agent n = 333), 4 (insulin n = 229, oral agent n = 382), and 5 (insulin n = 235, oral agent n = 407). Baseline characteristics stratified by wave of enrollment and diabetic treatment are presented in Table I. In the overall cohort and both insulin- and oral agent–treated patients with DM, there was a lower prevalence of women and of a prior MI in more recent cohorts, whereas the prevalence of prior percutaneous procedures and renal disease increased over time (P valuetrend b .01 for each). The procedural characteristics of the overall cohort and diabetic patients stratified by treatment with insulin or oral agents are in Table II. In the overall cohort and in the oral agent–treated diabetic patients, there was an increase in the percentage of 3-vessel disease in later cohorts. In the overall cohort and both treatment groups, the lesion length was significantly longer in more recent cohorts. The frequency of attempted lesions containing thrombus decreased with time in the overall cohort as well as in oral agent–treated patients, but this observation was not noted in the insulin-treated patients with diabetes.

American Heart Journal February 2011

400 Holper et al

The in-hospital and overall cumulative 1-year event rates in patients with medically treated DM are presented in Table III. The overall cohort of medically treated diabetic patients had a significant reduction in in-hospital mortality over time. Diabetic patients treated with oral agents had a significant reduction in inhospital mortality and in the combined end point of death/any MI/any CABG in later cohorts. Insulin-treated patients experienced no significant change in any of the in-hospital adverse event rates over time. The percentage of patients with medically treated DM in the overall cohort and in the subgroups treated with oral agents and insulin who were discharged on angiotensin-converting enzyme inhibitors, β-blockers, and statins increased significantly over time (P valuetrend for each b .001). The overall cumulative 1-year mortality rate in patients with medically treated DM was 7.3%, including 9.7% in insulin-treated patients and 5.9% among those treated with oral agents. At 1 year, the overall cohort of diabetic patients had a significant reduction in mortality over time with no change in MI rates (Table III). Whereas there was no change in cumulative 1-year mortality rates over time in insulin-treated patients, oral agent–treated patients died less frequently in more recent recruitment waves. The cumulative MI rates by wave and diabetes treatment demonstrated a significantly higher 1-year rate of MI in the insulin-treated diabetic patients in later waves, which was not seen in the oral agent–treated patients. There was a significant reduction in the cumulative 1-year need for repeat PCI after discharge in the overall cohort and the treatment subgroups. There was a significant reduction in the combined 1-year event rate by recruitment wave of repeat PCI after discharge and CABG in the overall cohort, the insulin-treated patients, and the oral agent–treated patients. The cumulative 1-year MACE rates were significantly lower in later recruitment waves in the overall cohort, in patients treated with insulin, and in patients treated with oral agents. The adjusted proportional hazard models for 1-year adverse events in diabetic patients stratified by treatment with insulin and oral agents are presented in Figures 1 and 2, respectively. After adjustment for clinical and angiographic predictors of risk, diabetic patients treated with insulin demonstrated a significant increase in mortality in waves 2 and 4 compared with wave 1. There were also a significant increase in the hazard ratio of MI in waves 3 and 4 compared with wave 1, but a decrease in the hazard ratio of repeat PCI/CABG in wave 3, 4, and 5 compared with wave 1. This resulted in no significant difference in the insulin-treated patients for the hazard ratio of MACE by enrollment wave. Diabetic patients treated with oral agents had no significant difference in the adjusted hazard ratio of mortality by wave, a significant decrease in MI in wave 5 compared with wave 1, and a significantly reduced

Table III. In-hospital and 1-year cumulative event rates of diabetic patients by enrollment wave and medical therapy Wave 1

Wave 2

In-hospital cumulative events No. of patients Total 553 505 Insulin 215 193 Oral agent 338 312 Death Total 3.1 1.4 Insulin 1.4 2.6 Oral agent 4.1 0.6 MI Total 1.6 1.4 Insulin 0.5 1.0 Oral agent 2.4 1.6 CABG Total 0.5 1.4 Insulin 0.5 0.5 Oral agent 0.6 1.9 Death/any MI/any CABG Total 4.9 3.8 Insulin 2.3 4.1 Oral agent 6.5 3.5 1-y cumulative events Death Total 9.6 8.8 Insulin 9.5 12.5 Oral agent 9.7 6.5 MI Total 5.8 5.3 Insulin 4.6 5.9 Oral agent 6.6 5.0 CABG Total 8.0 8.7 Insulin 7.1 6.2 Oral agent 8.6 10.2 Repeat PCI after discharge Total 19.3 16.3 Insulin 21.6 18.2 Oral agent 17.8 15.2 CABG/repeat PCI after discharge Total 25.2 22.3 Insulin 27.5 23.8 Oral agent 23.6 21.6 Death/MI/CABG/repeat PCI Total 33.7 31.0 Insulin 34.6 35.5 Oral agent 33.1 28.2

Wave 3

Wave 4

Wave 5

527 194 333

611 229 382

642 235 407

P value trend

1.1 1.5 0.9

2.3 3.9 1.3

0.8 0.9 0.7

.03 .95 .003

1.5 2.1 1.2

2.8 3.1 2.6

1.1 0.9 1.2

.90 .31 .55

0.2 0.5 0.0

1.0 1.3 0.8

0.3 0.9 0.0

.43 .40 .08

2.7 3.6 2.1

5.6 7.4 4.5

1.9 2.1 1.7

.08 .51 .005

5.9 8.9 4.1

7.7 11.6 5.4

5.2 6.6 4.7

.005 .33 .006

7.7 11.5 5.5

7.1 9.6 5.6

5.8 9.5 3.9

.7 .05 .19

5.3 3.9 6.1

3.6 3.3 3.8

3.2 2.7 3.6

.0001 .02 .0002

16.3 16.1 16.5

12.2 14.2 11.0

12.4 13.8 13.0

.0002 .02 .02

19.7 18.0 20.6

14.7 16.3 13.7

15.0 15.0 16.3

.0001 .0006 .0005

27.9 32.0 25.5

23.1 27.8 20.3

21.5 23.5 21.9

.0001 .005 .00001

adjusted hazard ratio of revascularization with PCI or CABG in waves 4 and 5 compared with wave 1. This resulted in a significant reduction in the combined adjusted hazard ratio of MACE in waves 4 and 5 compared with wave 1.

Discussion In this evaluation of temporal outcomes of diabetic patients in a contemporary PCI registry, we found that

American Heart Journal Volume 161, Number 2

Holper et al 401

Figure 1 Adjusted Hazard Ratio and 95% Confidence Interval Death Wave 2 Wave 3 Wave 4 Wave 5 MI Wave 2 Wave 3 Wave 4 Wave 5 Repeat PCI Wave 2 Wave 3 Wave 4 Wave 5 CABG/Repeat PCI Wave 2 Wave 3 Wave 4 Wave 5 MACE Wave 2 Wave 3 Wave 4 Wave 5

0 Improved Outcomes

1

2

3

4

5

6

Worsened Outcomes

Adjusted 1-year adverse events in insulin-treated diabetic patients.

the adjusted risk of major adverse cardiac events among diabetic patients treated with oral agents decreased over time. However, among insulin-treated diabetic patients, the adjusted risk for major adverse cardiac events remained steady, despite a reduction in repeat revascularization with PCI and CABG. In addition, the risk of MI grew in these patients despite advancements in medical therapy. There are several hypotheses for the higher MI rate in diabetic patients treated with insulin compared with patients managed with oral agents. It is most likely that this represents confounding by disease severity, with insulin treatment a marker of more severe disease in this group. In addition, endothelial dysfunction as an effect of insulin therapy might be responsible for accelerated atherosclerosis in untreated lesions or enhanced restenosis in treated lesions.10,11 Hyperinsulinemia, which also occurs in insulin-requiring diabetic patients treated with insulin, may contribute to the atherogenic process. Although we did not collect information on the etiology of diabetes in insulin-treated patients in the Dynamic Registry, it is likely that the vast majority of patients were likely to have type 2, insulin-resistant diabetes. The BARI 2D trial randomized 2,368 patients in a 2 × 2 factorial design to medical therapy versus revascularization (with CABG or PCI at the discretion of the clinician) and also randomized to a glucose management strategy with an insulin-sensitizing strategy versus an insulin-providing one.12 The rates of death and freedom from major

cardiovascular events at 5 years were not significantly different between the revascularization versus medical therapy group or between the insulin-sensitizing versus insulin-providing strategy. An evaluation of outcomes based on treatment strategies demonstrated no significant difference in MI rates in the PCI stratum between the insulin sensitization and insulin provision groups.13 Only a few studies in the bare-metal stent era have focused on the outcomes after PCI of insulin-treated versus oral agent–treated diabetic patients.14,15 Our data are similar to the reported experience at Emory in the angioplasty era, in which insulin-requiring diabetic patients had reduced 5-year survival and infarction-free survival compared with non–insulin-requiring diabetic patients.14 Despite the fact that these data come from patients treated from 1980 to 1990, the similarity in regard to higher adverse outcomes in insulin-treated diabetic patients is striking. This highlights the fact that our advancements in medical therapy post-PCI as well as improved procedural techniques have not mitigated some of the underlying risks in insulin-treated diabetic patients undergoing PCI. Several studies have described temporal improvements in outcomes post-PCI in patients with DM, despite an increase in the risk profile of these patients. However, these analyses focused on all diabetic patients, regardless of treatment. It is clear from our analysis that diabetic patients undergoing PCI who are treated with oral agents as compared with insulin have very different risk profiles.

American Heart Journal February 2011

402 Holper et al

Figure 2 Adjusted Hazard Ratio and 95% Confidence Interval Death Wave 2 Wave 3 Wave 4 Wave 5 MI Wave 2 Wave 3 Wave 4 Wave 5 Repeat PCI Wave 2 Wave 3 Wave 4 Wave 5 CABG/Repeat PCI Wave 2 Wave 3 Wave 4 Wave 5 MACE Wave 2 Wave 3 Wave 4 Wave 5

0 Improved Outcomes

1

2

3

4

5

6

Worsened Outcomes

Adjusted 1-year adverse events in oral agent–treated diabetic patients.

Our data show significant improvement in the outcomes of the oral agent–treated diabetic patients that was not seen in the patients treated with insulin. The Mid America Heart Institute investigators reported improvements in outcomes from 1980 to 1999 in diabetic patients undergoing elective PCI that were not seen in the diabetic patients undergoing urgent PCI.16 Given the dates evaluated in this analysis, the majority of these patients would not have undergone intracoronary stenting. Furthermore, our group has reported improvements of outcomes in PCI in patients with DM when stent era patients are compared with angioplasty era patients.17 Our study represents a unique, contemporary analysis of temporal trends in outcomes in the stent era of diabetic patients undergoing PCI. The Dynamic Registry has published data focusing on the outcomes of drug-eluting stents versus bare-metal stents in insulin-treated versus oral agent–treated diabetic patients that describe improved efficacy and similar safety of drug-eluting stents in insulin-treated diabetic patients.18 These data support the findings seen in other clinical trials of drug-eluting stents in diabetic patients.19,20 In line with these findings, we observed a lower risk of repeat PCI and revascularization in insulintreated diabetic patients in waves 4 and 5 that occurred after the approval of drug-eluting stents. This may therefore represent one important strategy to reduce the repeat PCI rates in all diabetic patients. It is possible that a component of the subsequent MI rate will be

reduced in the setting of longer drug-eluting stents placed in insulin-treated diabetic patients, which will require long-term follow up. However, reduction of the overall mortality of insulin-treated patients who require PCI will likely require focus on alternative therapeutic strategies outside of the catheterization laboratory. The limitations of this study are that it uses registry data as opposed to a randomized trial, such that differences in the demographic characteristics of the groups exist, especially differences between those treated with insulin compared with those treated with oral agents. This analysis was designed to evaluate the temporal trends of the 2 groups, such that this variability would be expected. Although a multivariable analysis was performed to adjust for differences in demographic and angiographic characteristics, such statistical adjustments may not have fully accounted for the patient differences. However, because many diabetic patients have features of risk that would preclude their entry into randomized clinical trials of PCI, we feel that these data offer a critical, real-world experience with diabetic patients undergoing PCI. Diabetic phenotyping is somewhat limited in the present data, with no information collected on duration of diabetes, decision making regarding insulin therapy, concomitance of microvascular disease complications, and differentiation of type 1 versus type 2 diabetes; and no assessments were made of glycemic control. Therefore, these factors could not be factored into the analyses. In addition, although the number of insulin-treated

American Heart Journal Volume 161, Number 2

patients in each wave is representative of current catheterization laboratory volume, the numbers are smaller than those of oral agent–treated diabetic patients, leading to less power to detect temporal differences in this group. Lastly, important differences exist between the various oral agents presently available for the treatment of diabetes; and the data collection does not allow for analyses according to specific oral agents by drug or by class. Our study highlights the need for further investigation into therapeutic strategies to reduce intermediate-term outcomes in insulin-requiring diabetic patients undergoing PCI.

References 1. Wild S, Roglic G, Green A, et al. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 2004;27:1047-53. 2. Abizaid A, Kornowski R, Mintz G, et al. The influence of diabetes mellitus on acute and late clinical outcomes following coronary stent implantation. J Am Coll Cardiol 1998;32:584-9. 3. Matthew V, Gersh B, Williams B, et al. Outcomes in patients with diabetes mellitus undergoing percutaneous coronary intervention in the current era: a report from the Prevention of REStenosis with Tranilast and its outcomes (PRESTO) trial. Circulation 2004;109:476-80. 4. Kip KE, Faxon DP, Detre KM, et al. Coronary angioplasty in diabetic patients: the National Heart, Lung, and Blood Institute Percutaneous Transluminal Coronary Angioplasty Registry. Circulation 1996;94: 1818-25. 5. Barsness GW, Peterson E, Ohman E. Relationship between diabetes mellitus and long-term survival after coronary bypass and angioplasty. Circulation 1997;96:2551-6. 6. Rozenman Y, Sapoznikov D, Gotsman M. Restenosis and progression of coronary disease after balloon angioplasty in patients with diabetes mellitus. Clin Cardiol 2000:890-4. 7. Elezi S, Kastrati A, Pache J. Diabetes mellitus and the clinical and angiographic outcome after coronary stent placement. J Am Coll Cardiol 1998;32:1866-73. 8. Joseph T, Fajadet J, Jordan C. Coronary stenting in diabetics: immediate and mid-term clinical outcome. Catheter Cardiovasc Interv 1999;47:279-84. 9. Williams DO, Holubkov R, Yeh W, et al. Percutaneous coronary intervention in the current era compared with 1985-1986. The National Heart, Lung, and Blood Institute Registries. Circulation 2000;102:2945-51.

Holper et al 403

10. Arcaro G, Cretti A, Balzano S, et al. Insulin causes endothelial dysfunction in humans: sites and mechanisms. Circulation 2002;105: 576-82. 11. Nishimoto Y, Miyazaki Y, Toki Y, et al. Enhanced secretion of insulin plays a role in the development of atherosclerosis and restenosis of coronary arteries: elective percutaneous transluminal coronary angioplasty in patients with effort angina. J Am Coll Cardiol 1998;1998:1624-9. 12. Brooks MM, Frye RL, Genuth S, et al. Hypotheses, design, and methods for the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial. Am J Cardiol 2006;97:9-19. 13. Chaitman BR, Hardison RM, Adler D, et al. Bypass Angioplasty Revascularization Investigation 2 Diabetes Study Group. The Bypass Angioplasty Revascularization Investigation 2 Diabetes randomized trial of different treatment strategies in type 2 diabetes mellitus with stable ischemic heart disease: impact of treatment strategy on cardiac mortality and myocardial infarction. Circulation 2009;120: 2529-40. 14. Stein B, Weintraub WS, Gebhart SSP, et al. Influence of diabetes mellitus on early and late outcome after percutaneous transluminal coronary angioplasty. Circulation 1995;91:979-89. 15. Mathew V, Frye RL, Lennon R, et al. Comparison of survival after successful percutaneous coronary intervention of patients with diabetes mellitus receiving insulin versus those receiving only diet and/or oral hypoglycemic agents. Am J Cardiol 2004;93: 399-403. 16. Marso S, Giorgi L, Johnson W, et al. Diabetes mellitus is associated with a shift in the temporal risk profile of inhospital death after percutaneous coronary intervention: an analysis of 25,223 patients over 20 years. Am Heart J 2003;145:270-7. 17. Freeman A, Abbott J, Jacobs A, et al. Marked improvements in outcomes of contemporary percutaneous coronary intervention in patients with diabetes mellitus. J Interv Cardiol 2006;19:475-82. 18. Mulukutla SR, Vlachos HA, Marroquin OC, et al. Impact of drugeluting stents among insulin-treated diabetic patients: a report from the National Heart, Lung, and Blood Institute Dynamic Registry. J Am Coll Cardiol Intv 2008;1:139-47. 19. Sabate M, Jimenez-Quevedo P, Angiolillo DJ, et al, for the DI. Randomized comparison of sirolimus-eluting stent versus standard stent for percutaneous coronary revascularization in diabetic patients: the Diabetes and Sirolimus-Eluting Stent (DIABETES) trial. Circulation 2005;112:2175-83. 20. Hermiller JB, Raizner A, Cannon L, et al. Outcomes with the polymer-based paclitaxel-eluting TAXUS stent in patients with diabetes mellitus: the TAXUS-IV trial. J Am Coll Cardiol 2005;45: 1172-9.

American Heart Journal Volume 161, Number 2

Appendix The complete list of variables that were included in the regression models is as follows:

Insulin-treated patients Death: age; chronic kidney disease; cardiogenic shock; circumstances/acuity of procedure; attempted lesion supplies collaterals; reason for revascularization; number of significant lesions; and discharged on thienopyridines, angiotensin-converting enzyme inhibitors, statins, and β-blockers. Myocardial infarction: history of congestive heart failure, vessel disease, attempted lesion in ostial location, circumstances/acuity of procedure, and attempted lesion supplies collaterals. Coronary artery bypass graft: cardiogenic shock, discharged on thienopyridines and angiotensin-converting enzyme inhibitors. Repeat PCI: history of prior PCI and number of significant lesions. Repeat revascularization: circumstances/acuity of procedure and any total occlusion. Major adverse coronary event: age, cardiogenic shock, number of significant lesions, any total occlusion, circumstances/acuity of procedure, chronic kidney disease, and discharged on thienopyridines and statins.

Holper et al 403.e1

Oral agent–treated patients Death: age; history of congestive heart failure; cardiogenic shock; chronic kidney disease; reason for revascularization; attempted tortuous lesion; attempted lesion supplies collaterals; circumstances/acuity of procedure; number of significant lesions; attempted calcified lesion; cancer; and discharged on thienopyridines, β-blockers, and statins. Myocardial infarction: prior PCI, attempted calcified lesion, any total occlusion, vessel disease, and peripheral vascular disease. Coronary artery bypass graft: sex, prior PCI, attempted calcified lesion, number of significant lesions, hypertension, and discharged on thienopyridines. Repeat PCI: age, sex, number of significant lesions, prior PCI, reason for revascularization, peripheral vascular disease, and presence of luminal irregularities. Repeat revascularization: age, sex, prior PCI, reason for revascularization, number of significant lesions, attempted calcified lesion, vessel disease, chronic kidney disease, and peripheral vascular disease. Major adverse coronary event: sex; number of significant lesions; circumstances/acuity of procedure; prior PCI; history of congestive heart failure; attempted tortuous lesion; attempted lesion with thrombus; attempted calcified lesion; attempted class C lesion; peripheral vascular disease; and discharged on thienopyridines, statins, and β-blockers.

Surgery

Clopidogrel loading dose and bleeding outcomes in patients undergoing urgent coronary artery bypass grafting Nicholas L.M. Cruden, PhD, MBChB, MRCP, a,b,c Kristin Morch, a Daniel R. Wong, MD, MPH, FRCSC, b W. Peter Klinke, MD, FRCPC, a,b John Ofiesh, MD, FRCSC, b and J. David Hilton, MD, FRCPC a,b British Columbia, Canada; and Edinburgh, United Kingdom

Background Coronary artery bypass grafting (CABG) performed within 5 days of clopidogrel administration is associated with increased bleeding. The impact of clopidogrel loading dose is unknown. We examined the effect of clopidogrel loading dose on bleeding outcomes in patients undergoing urgent CABG. Methods Clinical outcomes were examined retrospectively for 196 consecutive patients undergoing urgent CABG within 5 days of a clopidogrel loading dose between January 2003 and June 2009. Major bleeding was defined as a fall in hemoglobin N5g/dL, fatal or intracranial bleeding, or cardiac tamponade. Results One hundred forty-eight patients received 300 mg and 48 patients received ≥600 mg clopidogrel loading. Patients were predominantly male (78%) with a mean age of 66 ± 10 years. Mean duration from clopidogrel loading to CABG was 3.0 ± 1.5 and 3.0 ± 1.6 days for the 300 and 600 mg loading doses, respectively. Major bleeding occurred in 47% of patients receiving 300 mg and 73% of patients receiving ≥600 mg clopidogrel loading (P = .002). Compared with 300 mg, patients receiving ≥600 mg had greater 24-hour chest tube output (391 ± 251 vs 536 ± 354 mL, P = .01), stayed longer in surgical intensive care (4.3 ± 4.1 vs 5.0 ± 3.1 days, P = .0001), and trended toward greater reoperation for bleeding (5% vs 12%, P = .09). Following multivariate analysis, clopidogrel loading dose ≥600 mg (odds ratio 2.8, CI 1.2-6.6), preoperative hemoglobin (3.4, 2.7-5.0 per 1 g/dL increase), and female gender (2.9, 1.1-7.4) predicted major bleeding. Conclusions Higher clopidogrel loading doses are associated with increased bleeding when administered within 5 days of CABG. The development of shorter-acting, reversible, oral antiplatelet agents may reduce perioperative bleeding in this population. (Am Heart J 2011;161:404-10.)

Oral administration of clopidogrel, in combination with aspirin, significantly reduces ischemic vascular events in patients presenting with an acute coronary syndrome (ACS)1 or undergoing percutaneous coronary intervention (PCI).2,3 Recent attention has focused on the increased bleeding potential associated with clopidogrel

From the aVictoria Heart Institute Foundation, Victoria, British Columbia, Canada, b Department of Cardiac Services, Royal Jubilee Hospital, Victoria, British Columbia, Canada, and cDepartment of Cardiology, Royal Infirmary of Edinburgh, Edinburgh, United Kingdom. Relationship with industry: Dr Cruden recently undertook an unrestricted Interventional Fellowship supported by Boston Scientific. Dr Cruden was supported by an unrestricted Fellowship from Boston Scientific. Ms Morch was supported by a Vacation Research Fellowship from the Victoria Heart Institute Foundation. Submitted April 22, 2010; accepted October 29, 2010. Reprint requests: Nicholas L. M. Cruden, PhD, MBChB, MRCP, Department of Cardiology, Royal Infirmary of Edinburgh, Edinburgh, United Kingdom, EH16 5SB. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.10.037

use and the impact of bleeding on clinical outcomes in these high-risk patient groups.4 Excess bleeding is of particular concern in patients undergoing coronary artery bypass grafting (CABG), with significant implications for patient morbidity and mortality, length of hospital stay, and health care costs.5,6 In the United States, approximately 1 in 7 patients presenting with an ACS is referred for CABG during their index admission, with a median duration from hospital admission to CABG of 69 hours.7,8 In the vast majority (∼80%), CABG is performed within 5 days of admission.7,8 Although not a universal finding,9 a number of studies have demonstrated that CABG performed within 5 days of clopidogrel administration is associated with increased bleeding and bleeding-related complications, including cardiac tamponade and reoperation for bleeding.10-12 As a result, current guidelines recommend that, where possible, CABG should be delayed for at least 5 days following the last dose of clopidogrel.13 However, only in a minority of patients does this occur.11

American Heart Journal Volume 161, Number 2

Cruden et al 405

Table I. Patient demographics

Age, y Male BMI, kg/m2 Preoperative creatinine, mg/dL Preoperative hemoglobin, g/dL Risk factors Diabetes mellitus Hypertension Smoking history Congestive heart failure Prior myocardial infarction Prior PCI Previous CABG Admission diagnosis Dilated cardiomyopathy Stable angina Unstable angina NSTEMI STEMI

All (N = 196)

Clopidogrel 300 mg (n = 148)

Clopidogrel ≥600 mg (n = 48)

P⁎

66 ± 10 153 (78%) 28 ± 6 1.03 ± 0.26 13.8 ± 1.4

66 ± 10 113 (76%) 28 ± 4 1.02 ± 0.27 13.7 ± 1.4

64 ± 10 40 (84%) 27 ± 5 1.05 ± 0.24 14.2 ± 1.3

.10 .31 .68 .32 .02

49 (25%) 127 (65%) 116 (59%) 11 (6%) 38 (19%) 29 (15%) 1 (0.5%)

40 (27%) 98 (66%) 84 (57%) 7 (5%) 31 (21%) 26 (18%) 1 (0.7%)

9 (18%) 29 (60%) 32 (67%) 4 (8%) 7 (14%) 3 (6%) 0 (0%)

.25 .46 .22 .47 .33 .06 1.0

2 (1%) 6 (3%) 32 (16%) 149 (76%) 9 (6%)

2 (1%) 6 (4%) 24 (16%) 112 (76%) 4 (3%)

0 (0%) 0 (0%) 8 (17%) 35 (73%) 5 (10%)

.12

Values are expressed as mean ± SD or number (percentage). BMI, Body mass index; NSTEMI, non–ST-elevation myocardial infarction; STEMI, ST-elevation myocardial infarction. ⁎ Clopidogrel 300 mg versus ≥600 mg; binary variables compared by v2 or Fisher exact tests; preoperative hemoglobin compared by Student t test; all other continuous variables compared by Wilcoxon rank sum test.

Although a substantial body of evidence exists to support the use of a 300 mg loading dose of clopidogrel administered at least 6 hours before PCI,14 platelet inhibition studies demonstrate more rapid and greater platelet inhibition with higher loading doses (≥600 mg).15,16 Moreover, these pharmacodynamic benefits appear to translate to the clinical arena. Evidence of a reduction in ischemic vascular events with higher loading doses of clopidogrel in patients undergoing PCI is now emerging,17-19 with the result that an increasing number of patients are likely to undergo CABG within 5 days of a 600 mg loading dose of clopidogrel. Despite recent concerns,20 the impact of higher loading doses of clopidogrel on bleeding outcomes in patients undergoing urgent CABG is unknown. The aim of this study was to examine the effect of clopidogrel loading dose, administered within 5 days of CABG, on bleeding and bleeding-related outcomes.

Methods Patients One hundred ninety-six consecutive patients undergoing urgent CABG at the Royal Jubilee Hospital, Victoria, within 5 days of receiving a loading dose of clopidogrel between January 2003 and June 2009 were identified in a retrospective case record review. Coronary artery bypass grafting was defined as urgent if there was a clinical indication for surgery to be performed during the index admission. Patients undergoing elective CABG, or CABG combined with valve replacement or other complex cardiac surgical procedures, with the exception of mitral valve repair, were excluded from the study. This study

was performed with the approval of the local institutional review board.

Data collection Data collection was performed using a standardized proforma. Baseline demographics, surgical procedure details, pre- and perioperative medications, and clinical outcomes were obtained from individual chart review. Blood product use was obtained from centralized hospital transfusion records.

Outcomes The primary outcome was major bleeding, defined previously as N5 g/dL drop in hemoglobin, intracranial bleed, fatal bleed, or cardiac tamponade.10 Secondary end points included reoperation for bleeding, minor bleeding (documented bleeding with a fall in hemoglobin of N3 but ≤5 g/dL), chest drain output in the 24 hours immediately following CABG, length of postoperative stay in cardiac intensive care unit, number of red cell concentrate units transfused, length of hospital stay, inhospital mortality, postoperative myocardial infarction (defined as a clinical diagnosis made by the supervising physician and recorded in the patient records), and stroke.

Statistical analysis Statistical analysis was performed using SAS software, version 8.0 (SAS Institute Inc, Cary, NC). Patients were analyzed according to clopidogrel loading dose (300 vs ≥600 mg). Only hemoglobin exhibited a normal distribution. Univariate comparison of hemoglobin between treatment groups was performed using an unpaired Student t test. Univariate analysis of all other continuous variables was performed using nonparametric tests (Wilcoxon rank sum). Categorical variables were examined using χ2 or Fisher exact test. Data are presented as mean ± SD or

American Heart Journal February 2011

406 Cruden et al

Table II. Perioperative variables

Duration from loading dose of clopidogrel to CABG, d Duration from last dose of clopidogrel to CABG, d Left main stem stenosis N50% Cross clamp duration, min Operative time, min Off pump Mitral valve repair No. of grafts Graft used Saphenous vein LIMA RIMA Right radial artery Left radial artery Perioperative medications Aprotinin Aspirin Heparin Fondaparinux Warfarin ACE inhibitor/ARB Glycoprotein IIb/IIIa inhibitor Statin β-Blocker Proton pump inhibitor

All (N = 196)

Clopidogrel 300 mg (n = 148)

Clopidogrel ≥600 mg (n = 48)

P⁎

3.0 ± 1.5

3.0 ± 1.5

3.0 ± 1.6

.86

1.6 ± 1.3

1.6 ± 1.2

1.5 ± 1.4

.39

95 (48%) 86 ± 30 220 ± 53 2 (1%) 3 (2%) 3.0 ± 0.6

67 (46%) 85 ± 30 220 ± 54 2 (1%) 3 (2%) 3.1 ± 0.7

28 (60%) 86 ± 30 219 ± 50 0 (0%) 0 (0%) 3.0 ± 0.5

.12 .82 .97 1.0 1.0 .72

184 (94%) 165 (84%) 3 (2%) 7 (4%) 17 (9%)

137 (93%) 122 (83%) 2 (1%) 7 (5%) 15 (10%)

47 (98%) 43 (92%) 1 (2%) 0 (0%) 2 (4%)

.73 .47 .57 .20 .25

13 (7%) 175 (89%) 153 (78%) 14 (7%) 4 (2%) 111 (57%) 15 (8%) 140 (71%) 176 (90%) 15 (8%)

13 (9%) 132 (90%) 114 (77%) 12 (8%) 3 (2%) 81 (55%) 12 (8%) 105 (72%) 130 (88%) 11 (8%)

0 (0%) 43 (92%) 39 (81%) 2 (4%) 1 (2%) 30 (63%) 3 (6%) 35 (74%) 46 (96%) 4 (9%)

.04 1.0 .54 .52 1.0 .49 1.0 .92 .25 1.0

Values are expressed as mean ± SD or number (percentage). LIMA, Left internal mammary artery; RIMA, right internal mammary artery; ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker. ⁎ Clopidogrel 300 mg versus ≥600 mg; binary variables compared by v2 or Fisher exact tests, and continuous variables compared by Wilcoxon rank sum test.

number (percentage). Univariate and multivariate logistic regression analysis were performed to identify variables predictive of the primary end point of major bleeding. Variables demonstrating an association with major bleeding or imbalance between study groups with a significance of P b .1 on univariate analysis were considered for stepwise selection in the multivariate model. Statistical significance was assumed at the b5% level. No extramural funding was used specifically to support this work, but Dr Cruden's Fellowship was supported by an unrestricted grant from Boston Scientific. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.

Results One hundred forty-eight patients received a 300 mg loading dose of clopidogrel and 48 patients received a loading dose of clopidogrel of ≥600 mg (46 patients received 600 mg, 1 patient received 675 mg, and 1 patient received 900 mg) within 5 days of undergoing CABG. Patient demographics are shown in Table I. Patients were predominantly male (78%) with a mean age of 66 ± 10 years and were generally matched between groups (Table I). Mean duration from loading to CABG was 3.0 ± 1.5 and 3.0 ± 1.6 days for the 300 mg

and ≥600 mg groups, respectively (P = .86). Preoperative hemoglobin was lower (13.7 ± 1.4 vs 14.2 ± 1.3 g/dL, P = .02) and aprotinin was administered more frequently (13 [9%] vs 0 [0%], P = .04) in patients receiving 300 mg compared with ≥600 mg of clopidogrel (Table II). Three patients receiving a 300 mg clopidogrel loading dose underwent CABG plus mitral valve repair compared with no patients in the ≥600 mg group (P = 1.0) (Table II). Major bleeding occurred in 47% of patients receiving a 300 mg loading dose of clopidogrel and 73% of patients receiving ≥600 mg (P = .002) (Table III). Compared with patients receiving 300 mg, patients receiving ≥600 mg had greater 24-hour chest tube output (391 ± 251 vs 536 ± 354 mL, P = .01), stayed longer in surgical intensive care (4.3 ± 4.1 vs 5.0 ± 3.1 days, P = .0001), and trended toward greater reoperation for bleeding (5% vs 12%, P = .09) (Table III). There were no differences in rates of blood product use between treatment groups (Table III). Following univariate analysis, body mass index (P = .04), preoperative diastolic blood pressure (P = .08), the presence of diabetes mellitus (P = .008), preoperative hemoglobin (P b .0001), perioperative aspirin use (P = .02) and dose (P = .03), clopidogrel loading dose (P = .002), perioperative warfarin administration (P = .045), and operation time (P = .04) were identified as

American Heart Journal Volume 161, Number 2

Cruden et al 407

Table III. Clinical outcomes All (N = 196) Bleeding outcomes Major bleeding Fall in hemoglobin N5 g/dL Fatal bleed Cardiac tamponade Intracranial bleed Reoperation for bleeding Minor bleeding Patients receiving blood products RCC Platelets Cryoprecipitate Plasma Units of RCC per patient 24-h chest tube output, mL Major adverse cardiac events In-hospital mortality Death/reinfarction/stroke Length of stay in cardiac ICU, d Length of hospital stay, d

105 (53%) 102 (52%) 0 (0%) 6 (3%) 2 (1%) 13 (7%) 10 (5%)

Clopidogrel 300 mg (n = 148)

70 67 0 4 2 7 7

(47%) (45%) (0%) (3%) (1%) (5%) (5%)

Clopidogrel ≥600 mg (n = 48)

P⁎

35 (73%) 35 (73%) 0 (0%) 2 (3%) 0 (0%) 6 (12%) 3 (6%)

.002 .0009 1.0 .82 1.00 .09 .71

92 (47%) 15 (8%) 3 (2%) 25 (13%) 1.5 ± 2.2 427 ± 286

71 (48%) 11 (7%) 1 (1%) 17 (11%) 1.5 ± 2.1 391 ± 251

21 (43%) 4 (8%) 2 (4%) 8 (17%) 1.4 ± 2.3 536 ± 354

.62 .76 .14 .33 .55 .01

8 (4%) 28 (14%) 4.5 ± 3.9 11.5 ± 10.9

7 (5%) 21 (14%) 4.3 ± 4.1 11.6 ± 11.6

1 (2%) 7 (15%) 5.0 ± 3.1 11.3 ± 8.4

.68 1.0 .0001 .18

Values are expressed as mean ± SD or number (percentage). RCC, Red cell concentrate; ICU, intensive care unit. ⁎ Clopidogrel 300 mg versus ≥600 mg; binary variables compared by v2 or Fisher exact tests, and continuous variables compared by Wilcoxon rank sum test.

potential predictors of major bleeding (Tables IV and V) and were included in the multivariate logistic regression model. Preoperative characteristics that differed between treatment groups (Table I) were also considered as candidate variables in the model. On multivariate analysis, a clopidogrel loading dose of ≥600 mg (odds ratio 2.8, CI 1.2-6.6), preoperative hemoglobin (3.4, 2.7-5.0 per 1-g/dL increase), and female gender (2.9, 1.1-7.4) predicted major bleeding.

Discussion To our knowledge, this is the first study to examine the effect of clopidogrel loading dose on bleeding outcomes in patients with symptomatic coronary artery disease undergoing CABG. Importantly, we have demonstrated that major bleeding is significantly increased in patients receiving a 600 mg loading dose of clopidogrel within 5 days of CABG when compared with those receiving 300 mg. Patients receiving a 600 mg loading dose of clopidogrel spent a longer time in cardiac intensive care postoperatively, and there was a trend toward greater reoperation for bleeding when compared with a loading dose of 300 mg. A significant proportion of patients in this study (53%) met the primary end point of major bleeding defined as N5 g/dL drop in hemoglobin, intracranial hemorrhage, fatal bleed, or cardiac tamponade. Driven largely by a significant fall in hemoglobin, this end point has been used widely as a reliable indicator of bleeding both in patients with ACS1,21,22 and in the setting of surgical revascularization.10 Our findings are comparable to rates

of major bleeding observed in contemporary study populations undergoing CABG following recent clopidogrel exposure (35%-50%)9,10 and are consistent with the risk profile of the population examined. Importantly, the rate of major bleeding in patients undergoing CABG within 5 days of a clopidogrel loading dose observed in this study (53%) was more than double that reported for clopidogrel-naive patients with ACS undergoing CABG (26%) in a multicenter US study.10 We identified female gender as an independent predictor of major bleeding in patients undergoing CABG. Our findings are consistent with previous data reporting increased bleeding rates in female patients presenting with ACS,23-25 following PCI,26 or undergoing CABG.27 The mechanism accounting for the observed gender-based differences in bleeding is unclear but may reflect differences in body mass index,26 baseline hemoglobin, and triggers for transfusion,28 as well as excess dosing of antiplatelet and antithrombin agents.24 We also identified baseline hemoglobin as an independent predictor of the primary outcome. This may, in part, reflect the choice of composite primary end point, as well as a small but significant difference in preoperative hemoglobin between the 2 groups. Although processes other than bleeding, such as hemodilution or hemolysis, may have contributed to a fall in hemoglobin in patients undergoing CABG, it is reassuring that the increase in chest tube output and greater need for reoperation due to bleeding observed with a clopidogrel loading dose of ≥600 mg are consistent with the principal finding of increased major bleeding. Although we cannot exclude the potential for interaction between

American Heart Journal February 2011

408 Cruden et al

Table IV. Baseline demographic predictors of major bleeding on univariate analysis

Age, y Male Weight, kg Height, m BMI, kg/m2 Preoperative heart rate, beat/min Preoperative systolic blood pressure, mm Hg Preoperative diastolic blood pressure, mm Hg Preoperative creatinine, mg/dL Risk factors Diabetes mellitus Hypertension Congestive heart failure Hypercholesterolemia Chronic renal impairment Smoking history Prior myocardial infarction Prior PCI Previous CABG Prior bleeding disorder STEMI at presentation Preoperative hemoglobin, g/dL Preoperative platelet count, ×106/L

No bleed (n = 91)

Bleed (n = 105)

P⁎

66 ± 10 68 (74.7) 80.2 ± 13.6 1.72 ± 0.11 27 ± 4 73 ± 13

65 ± 9 85 (81.0) 83.3 ± 15.0 1.72 ± 0.10 28 ± 4 75 ± 15

.29 .30 .11 .80 .04 .27

131 ± 27

135 ± 24

.39

72 ± 13

76 ± 15

.08

1.02 ± 0.27

1.04 ± 0.25

.32

31 (34.1) 62 (68.1) 5 (5.5) 65 (71.4) 2 (2.2) 54 (59.3) 22 (24.2) 15 (16.5) 1 (1.1) 1 (1.1) 3 (3.3) 13.1 ± 1.2

18 (17.1) 65 (61.9) 6 (5.7) 75 (71.4) 5 (4.8) 62 (59.0) 16 (15.2) 14 (13.3) 0 (0) 1 (1.0) 5 (4.8) 14.5 ± 1.1

.008 .36 .95 1.0 .45 .97 .11 .54 .46 1.0 .73 b.0001

250.3 ± 83.9

253.1 ± 83.0

.82

Values are expressed as mean ± SD or number (percentage). ⁎ No bleed versus bleed; binary variables compared by v2 or Fisher exact tests; preoperative hemoglobin compared by Student t test; all other continuous variables compared by Wilcoxon rank sum test.

transfusion and the primary end point of major bleeding, this would seem unlikely given the absence of a significant difference in transfusion rates between the 2 treatment groups. Large variation exists in rates of transfusion, with a lack of consensus on triggers for transfusion, fears over viral transfection, and the potential impact of transfusion on mortality.28,29 These factors, coupled with the higher baseline hemoglobin values in the ≥600 mg loading group, may explain the absence of an association between clopidogrel loading dose and transfusion rates in the current study. The perioperative use of aprotinin has been shown to reduce bleeding in patients undergoing CABG. In the current study, the rate of perioperative aprotinin use was low (b10%), although it was used more frequently in patients receiving a 300 mg loading dose of clopidogrel compared with those treated with 600 mg. This may, in part, reflect temporal changes in the use of aprotinin relative to clopidogrel loading strategies. Despite this, we found no evidence of a major interaction between aprotinin use and major bleeding following univariate

Table V. Perioperative predictors of major bleeding on univariate analysis

Perioperative medications Clopidogrel loading dose, mg 300 ≥600 Aspirin Aspirin dose, mg 0 81 160 325 Heparin Warfarin Fondaparinux ACE inhibitor/ARB Glycoprotein IIb/IIIa inhibitor Statin Aprotinin β-Blocker Proton pump inhibitor Left main stem stenosis N50% Left ventricular dysfunction Off pump No. of grafts used 1 2 3 4 5 Graft used Saphenous vein LIMA RIMA Right radial artery Left radial artery Duration from clopidogrel loading to CABG, d Duration from last clopidogrel dose to CABG, d Cross clamp duration, min Operative time, min

No bleed (n = 91)

Bleed (n = 105)

78 (85.7) 13 (14.3) 87 (95.6)

70 (66.7) 35 (33.3) 90 (85.7)

4 (4.4) 69 (75.8) 7 (7.7) 11 (12.1) 72 (79.1) 4 (4.4) 7 (7.7) 50 (55.0) 8 (8.8) 62 (68.1) 8 (8.8) 82 (90.1) 8 (8.8) 48 (52.7) 40 (44.0) 0

15 (14.3) 67 (63.8) 15 (14.3) 8 (7.6) 81 (77.1) 0 7 (6.7) 60 (57.1) 7 (6.7) 78 (74.3) 5 (4.8) 93 (88.6) 8 (7.6) 47 (44.8) 39 (37.1) 1 (1.0)

P⁎

.002

2 12 62 15

(2.2) (13.2) (68.1) (16.5) 0

16 67 20 2

.02 .03

.74 .045 .78 .76 .58 .34 .26 .73 .76 .26 .33 1.00 .45

0 (15.2) (63.8) (19.0) (1.9)

86 (94.5) 75 (82.4) 1 (1.1) 5 (5.5) 5 (5.5) 3.0 ± 1.4

98 (93.3) 90 (85.7) 2 (1.9) 2 (1.9) 12 (11.4) 3.0 ± 1.6

.73 .53 1.00 .25 .14 .92

1.7 ± 1.2

1.6 ± 1.3

.50

85 ± 31 213 ± 56

86 ± 29 227 ± 50

.67 .04

Values are expressed as mean ± SD or number (percentage). ⁎ No bleed versus bleed; binary variables compared by v2 or Fisher exact tests, and continuous variables compared by Wilcoxon rank sum test.

analysis or when aprotinin was included in the multivariate logistic regression model.

Study limitations We included all patients undergoing urgent CABG in our institution over a 6-year period in the current study, the study populations were generally well matched, and multivariable logistic regression was used to adjust for any differences in baseline variables. However, we must acknowledge the limitations associated with a retrospective, nonrandomized study design. Although we are not aware of any evidence to suggest that patient management differed significantly according to clopidogrel

American Heart Journal Volume 161, Number 2

loading dose within our institution, we cannot exclude the possibility that differences in clinical practice, both between individuals within our institution and over time, may have confounded our results.

Conclusions In summary, we have demonstrated that a loading dose of clopidogrel of ≥600 mg, administered within 5 days of CABG, is associated with significantly greater major bleeding when compared with 300 mg. Data are emerging to support the widespread use of a 600 mg loading dose of clopidogrel in patients with ACS undergoing PCI. A substantial proportion of these patients will undergo urgent CABG within 5 days of admission. Although this study did not specifically examine the risk-benefit ratio of a 600 mg loading dose in this population, our findings highlight the potential hazard of major bleeding associated with higher loading doses of clopidogrel. The development of novel, shorter-acting antiplatelet agents with more rapid onset-offset of action30 may limit the excess bleeding hazard in those patients requiring urgent CABG while maintaining the benefits of early, intensive antiplatelet activity in those patients who require PCI. Further work is required to determine the optimal antiplatelet strategy in patients with ACS who require urgent CABG.

Acknowledgements We would like to thank the staff of Health Records and the Transfusion Service at the Royal Jubilee Hospital for their help and assistance.

Disclosures Funding source: Victoria Heart Institute Foundation.

References 1. Yusuf S, Zhao F, Mehta SR, et al. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494-502. 2. Mehta SR, Yusuf S, Peters RJ, et al. Effects of pretreatment with clopidogrel and aspirin followed by long-term therapy in patients undergoing percutaneous coronary intervention: the PCI-CURE study. Lancet 2001;358:527-33. 3. Steinhubl SR, Berger PB, Mann JT, et al. Early and sustained dual oral antiplatelet therapy following percutaneous coronary intervention: a randomized controlled trial. JAMA 2002;288:2411-20. 4. Holmes Jr DR, Kereiakes DJ, Kleiman NS, et al. Combining antiplatelet and anticoagulant therapies. J Am Coll Cardiol 2009;54: 95-109. 5. Burdess A, Cruden NL, Fox KA. Antiplatelet therapy and coronary bypass surgery: risks and benefits. In: Wiviott SD, editor. Antiplatelet therapy in ischemic heart disease. Chichester, West Sussex, UK: Wiley-Blackwell; 2009. p. 233-50. 6. Berger JS. Platelet-directed therapies and coronary artery bypass grafting. Am J Cardiol 2009;104:44C-8C.

Cruden et al 409

7. Mehta RH, Chen AY, Pollack CV, et al. Challenges in predicting the need for coronary artery bypass grafting at presentation in patients with non–ST-segment elevation acute coronary syndromes. Am J Cardiol 2006;98:624-7. 8. Bae JP, Dobesh PP, McCollam PL, et al. Potential unrecognised costs of clopidogrel pretreatment in acute coronary syndrome. J Med Econ 2009;12:325-30. 9. Ebrahimi R, Dyke C, Mehran R, et al. Outcomes following preoperative clopidogrel administration in patients with acute coronary syndromes undergoing coronary artery bypass surgery: the ACUITY (Acute Catheterization and Urgent Intervention Triage strategY) trial. J Am Coll Cardiol 2009;53:1965-72. 10. Berger JS, Frye CB, Harshaw Q, et al. Impact of clopidogrel in patients with acute coronary syndromes requiring coronary artery bypass surgery: a multicenter analysis. J Am Coll Cardiol 2008;52: 1693-701. 11. Mehta RH, Roe MT, Mulgund J, et al. Acute clopidogrel use and outcomes in patients with non–ST-segment elevation acute coronary syndromes undergoing coronary artery bypass surgery. J Am Coll Cardiol 2006;48:281-6. 12. Fox KA, Mehta SR, Peters R, et al. Benefits and risks of the combination of clopidogrel and aspirin in patients undergoing surgical revascularization for non–ST-elevation acute coronary syndrome: the Clopidogrel in Unstable angina to prevent Recurrent ischemic Events (CURE) trial. Circulation 2004;110: 1202-8. 13. Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHA 2004 guideline update for coronary artery bypass graft surgery: summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1999 Guidelines for Coronary Artery Bypass Graft Surgery). Circulation 2004;110:1168-76. 14. Smith SC, Feldman TE, Hirshfeld JW, et al. ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/ AHA/SCAI Writing Committee to Update 2001 Guidelines for Percutaneous Coronary Intervention). Circulation 2006;113: e166-286. 15. Montalescot G, Sideris G, Meuleman C, et al. A randomized comparison of high clopidogrel loading doses in patients with non– ST-segment elevation acute coronary syndromes: the ALBION (Assessment of the Best Loading Dose of Clopidogrel to Blunt Platelet Activation, Inflammation and Ongoing Necrosis) trial. J Am Coll Cardiol 2006;48:931-8. 16. Gurbel PA, Bliden KP, Zaman KA, et al. Clopidogrel loading with eptifibatide to arrest the reactivity of platelets: results of the Clopidogrel Loading With Eptifibatide to Arrest the Reactivity of Platelets (CLEAR PLATELETS) study. Circulation 2005;111: 1153-9. 17. Mehta SR, Tanguay JF, Eikelboom JW, et al. Double-dose versus standard-dose clopidogrel and high-dose versus low-dose aspirin in individuals undergoing percutaneous coronary intervention for acute coronary syndromes (CURRENT-OASIS 7): a randomised factorial trial. Lancet 2010;376:1233-43. 18. Patti G, Colonna G, Pasceri V, et al. Randomized trial of high loading dose of clopidogrel for reduction of periprocedural myocardial infarction in patients undergoing coronary intervention: results from the ARMYDA-2 (Antiplatelet therapy for Reduction of MYocardial Damage during Angioplasty) study. Circulation 2005;111: 2099-106.

American Heart Journal February 2011

410 Cruden et al

19. Dangas G, Mehran R, Guagliumi G, et al. Role of clopidogrel loading dose in patients with ST-segment elevation myocardial infarction undergoing primary angioplasty: results from the HORIZONS-AMI (harmonizing outcomes with revascularization and stents in acute myocardial infarction) trial. J Am Coll Cardiol 2009;54:1438-46. 20. Patel SS, Mascarenhas D. Rethinking loading dose clopidogrel in light of increased bleeding complications in bypass patients. J Am Coll Cardiol 2009;54:90. 21. Wallentin L, Becker RC, Budaj A, et al. Ticagrelor versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 2009;361: 1045-57. 22. Chesebro JH, Knatterud G, Roberts R, et al. Thrombolysis in Myocardial Infarction (TIMI) Trial, Phase I: a comparison between intravenous tissue plasminogen activator and intravenous streptokinase. Clinical findings through hospital discharge. Circulation 1987; 76:142-54. 23. Moscucci M, Fox KA, Cannon CP, et al. Predictors of major bleeding in acute coronary syndromes: the Global Registry of Acute Coronary Events (GRACE). Eur Heart J 2003;24:1815-23. 24. Alexander KP, Chen AY, Roe MT, et al. Excess dosing of antiplatelet and antithrombin agents in the treatment of non–ST-

25.

26.

27. 28. 29.

30.

segment elevation acute coronary syndromes. JAMA 2005;294: 3108-16. Alexander KP, Chen AY, Newby LK, et al. Sex differences in major bleeding with glycoprotein IIb/IIIa inhibitors: results from the CRUSADE (Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/ AHA guidelines) initiative. Circulation 2006;114:1380-7. Byrne J, Spence MS, Fretz E, et al. Body mass index, periprocedural bleeding, and outcome following percutaneous coronary intervention (from the British Columbia Cardiac Registry). Am J Cardiol 2009; 103:507-11. Kim JH, Newby LK, Clare RM, et al. Clopidogrel use and bleeding after coronary artery bypass graft surgery. Am Heart J 2008;156:886-92. Goodnough LT. Transfusion triggers. Surgery 2007;142:S67-70. Bennett-Guerrero E, Zhao Y, O'Brien SM, et al. Variation in use of blood transfusion in coronary artery bypass graft surgery. JAMA 2010;304:1568-75. Gurbel PA, Bliden KP, Butler K, et al. Randomized double-blind assessment of the ONSET and OFFSET of the antiplatelet effects of ticagrelor versus clopidogrel in patients with stable coronary artery disease: the ONSET/OFFSET study. Circulation 2009;120: 2577-85.

Pediatrics

Factors associated with the physical activity level of children who have the Fontan procedure Patricia E. Longmuir, PhD, a,b Jennifer L. Russell, MD, FRCP(C), a,c Mary Corey, PhD, d Guy Faulkner, PhD, e and Brian W. McCrindle, MD, FRCP(C) a,c Toronto, Canada

Background Children with complex heart defects are sedentary, with activity level unrelated to exercise capacity. We sought to identify factors associated with physical activity level for children who have the Fontan procedure. Methods

We used a cross-sectional study, 64 children (25 female, 5-11 years) after Fontan. Measurements were weekly minutes of moderate-to-vigorous physical activity, cardiac status, resting/exercise cardiopulmonary capacity, gross motor skill, health-related endurance/strength/body composition, and parent/child activity perceptions.

Results Participants performed 361 ± 137 minutes per week of moderate-to-vigorous physical activity. Increased activity related to antithrombotic medication use (86 min/wk), lower resting heart rate (3 min/wk), higher weekday outdoor time (0.7 minutes per outside minute), lower family income (13 minutes per $10,000), and higher parent rating of child's activity relative to peers (36 min/wk). Factors related to decreased activity were winter season (−84 min/wk), history of arrhythmia (−96 min/wk), and greater child confidence in own ability to be active (−113 min/wk). Conclusions Physical activity after the Fontan procedure is primarily associated with factors unrelated to cardiac status. Interventions that impact these modifiable factors would be expected to enable these children to achieve the recommended activity levels associated with optimal health. (Am Heart J 2011;161:411-7.)

Children should do at least 60 minutes of physical activity daily,1 although most do not achieve that level.2 Activity is important for childhood growth, socialization, and quality of life.3,4 Children with functional single ventricle after a Fontan procedure are more sedentary than healthy peers.5 These sedentary lifestyles immediately impact peer socialization and hinder motor skill development,6 which may lead to decreased activity selfefficacy7 and increase the risk of sedentary lifestyle morbidities (obesity, diabetes, and atherosclerosis8). Understanding the factors associated with activity participation would enable effective interventions to increase daily activity. Exercise capacity can be increased through training,9,10 but it is daily physical activity that is most important for peer interactions and decreasing

From the aLabatt Family Heart Centre, The Hospital for Sick Children, Toronto, Canada, b Department of Physical Therapy, Faculty of Medicine, University of Toronto, Toronto, Canada, cDepartment of Paediatrics, Faculty of Medicine, University of Toronto, Toronto, Canada, dChild Health Evaluative Sciences, The Hospital for Sick Children, Dalla Lana School of Public Health, University of Toronto, Toronto, Canada, and eFaculty of Physical Education and Health, University of Toronto, Toronto, Canada. Submitted August 30, 2010; accepted November 13, 2010.

Reprint requests: Brian McCrindle, MD, FRCP(C), Labatt Family Heart Centre, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.11.019

sedentary lifestyle morbidities.8 We examined relationships between moderate-to-vigorous physical activity (MVPA) and exercise capacity, gross motor skill, healthrelated fitness (strength, flexibility, endurance, and body composition needed for a healthy lifestyle), and parent and child perceptions of physical activity among families with a child who had the Fontan procedure.

Methods The protocol was approved by the Research Ethics Board of The Hospital for Sick Children and conducted in accordance with the Tri-Council Policy Statement.11 Written informed parent consent and informed child assent were obtained. The Heart and Stroke Foundation of Ontario (Grant #NA 5950) funded this research. Dr Longmuir was supported by a Doctoral Research Award, Canadian Institutes for Health Research. The authors are solely responsible for the design and conduct of this study, all analyses, and the drafting and editing of the manuscript, and its final contents.

Subject selection All 117 children (47 girls) known to The Hospital for Sick Children who were at least 1 year post–Fontan procedure, were 6 to 11 years old, and had no other physical activity contraindications were eligible for this study. Six children were not living in Canada, and 1 refused to be approached for research. Study invitations were sent to all remaining families (n = 110). Families of 71 children (26 girls) consented to participate, 31

American Heart Journal February 2011

412 Longmuir et al

families refused, 7 families could not be contacted, and 1 child was found to be ineligible. Seven families withdrew because the child either became ineligible (n = 6) or was unable to cooperate with testing (n = 1). Overall participation was 67% of eligible children (n = 64/95).

Measurements Demographics and cardiac history were abstracted from medical records. Children completed standardized measures of gross motor skill,12 cardiopulmonary capacity,13 and healthrelated fitness.14 The child's attitude, predilection, enjoyment of physical activity,15 perception of activity benefits/barriers, social support, expectations for16 and confidence in being active,17 and activity preferences18 were obtained via interview. Parent questionnaire19 assessed child’s activity level, coordination and enjoyment relative to peers, involvement in structured/ unstructured activity, and neighborhood characteristics. Children refrained from eating for 2 hours before the cardiopulmonary exercise test; otherwise, snacks and breaks were provided as required. Cardiopulmonary capacity was assessed at rest (5-10 minutes, semisupine, quiet room) and during continuous, graded exercise (General Electric T-2000 treadmill). Oxygen consumption (Physiodyne MAX-2 metabolic cart; Quogue, NY) and 12-lead electrocardiogram (GE Medical Systems CASE 8000, Milwaukee, WI) were continuously monitored. Blood pressure was measured during each stage. Two 2-minute warm-up stages (6% grade at 1.0 km/h and 8% grade at 2.0 km/h) preceded the Bruce protocol to enable younger children to become accustomed to the treadmill and monitoring equipment. Pulse count and blood pressure after walking up and down 2 steps for 3 minutes at a predetermined pace assessed healthrelated endurance.14 Faster bouts were completed until the pulse count exceeded 80% of peak heart rate; there was no increase in pulse count with an increase of stepping speed; or the child stopped because of fatigue or being unable/unwilling to maintain the required pace. Grip and trunk strength were assessed as maximum right/left hand (Smedlays dynamometer) and number of partial curl-ups at a pace of 25 per minute, respectively. Flexibility was assessed as the distance from outstretched fingers to soles of the feet while seated with legs extended and ankles dorsiflexed to 90°. Height and weight were measured, and body mass index was calculated.

Measurement of weekly MVPA An omnidirectional accelerometer (Mini Mitter Respironics, Actical 2.1) was worn above the iliac crest at the midaxillary line for 5 school days (weekday) and 2 nonschool days (weekend). Groups of accelerometers worn simultaneously established monitor comparability. Data were stored in 15-second epochs. Log sheets recorded unusual events. A log review indicated events were rare and did not influence the data. Participants required at least 8 hours of valid accelerometer measurements per day on 3 weekdays and 1 weekend day. Sensory malfunction removed data for one child; all others (63/64 or 98%) achieved at least 4 monitoring days.

Data analyses Moderate-to-vigorous activity cutpoint was 1,600 counts per minute.20 Epochs were summed per day, and total minutes per

week calculated ([5 × weekday average] + [2 × weekend average]). Normative data for healthy children converted gross motor skill12 and body composition21 result to age- and gendermatched percentile scores. Cardiopulmonary test Z scores were based on data for children with innocent heart murmurs.22 Flexibility and grip strength Z scores were based on data for healthy children.23 Summary statistics are described with the median (first [Q1] and third [Q3] quartiles) or frequencies as appropriate. All analyses were performed using SAS software, version 9.2 (SAS Institute, Cary, NC). Multivariable linear regression with backward variable selection identified demographic, physical performance, and activity perception factors significantly associated with MVPA. A bootstrap-bagging algorithm (1,000 samples excluding collinear/ill-conditioned/substantially missing variables) identified factors for modeling because of the large number of measures relative to sample size. Factors selected in at least 40% of the bootstrap samples were included. No transformation of MVPA was required because scores were normally distributed. The final multivariable model was not obtained through bootstrapping. Reported is the percent reliability (percentage of bootstrap samples in which factor was selected), estimate effect size and standard error, and P value in the multivariable model. Regression models via maximum likelihood algorithm determined parameter estimates. Statistical significance was P b .05.

Results Patient characteristics Sixty-three patients (25 female [40%]), 9.1 years old (Q1 7.7, Q3 10.5), completed the study. They were 5.5 years (Q1 4.2, Q3 7.3) post-Fontan, which was performed at 2.9 years (Q1 2.3, Q3 3.6). Participants had a functional single left (33/64 or 52%) or right ventricle and status as summarized in Table I. Postal code–estimated family income was diverse ($75,741, Q1 $56,044, Q3 $87,607). Those who refused participation (n = 31) were similar to participants in gender (20 male, 11 female) and age (9.8 years, Q1 7.5, Q3 10.7). Level of MVPA After Fontan, children perform 341 minutes (Q1 271, Q3 422) or 5.7 hours of weekly MVPA. Two children (3%) achieved 90 minutes of daily MVPA.24 The most sedentary children completed b15 minutes of MVPA per day. Boys and girls at all ages were significantly (P b .001) more active on weekdays (55 minutes) than on weekends (34 minutes). Physical performance measures Physical performance measures varied substantially. Gross motor scores (42%, Q1 16%, Q3 79%) were similar to healthy children but with a broader than normative distribution. Participants were shorter (128 cm, Q1 119, Q3 138) and lighter (26.0 kg, Q1 22, Q3 33) than expected for age (23rd and 24th percentiles,21 respectively). Body mass index was the 53rd percentile. Peak heart rate during cardiopulmonary exercise was 164

American Heart Journal Volume 161, Number 2

Longmuir et al 413

Table I. Cardiac status of the study participants Variable Cardiac anatomy

Status

Hypoplastic left heart syndrome Tricuspid atresia Double-inlet left ventricle Pulmonary atresia Double-outlet right ventricle Type of Fontan Extracardiac Fontan connection Lateral tunnel Fontan Bjork procedure Current ASA medications Warfarin β-Blocker Calcium-channel blocker Angiotensin-converting enzyme inhibitor Glycoside Pacemaker Fenestration Open status Not seen Closed Never created Complication Stroke history Thrombus Arrhythmia Ventricular dysfunction Seizure Protein-losing enteropathy

No. of % of participants participants 20

31

12 14

19 22

4 13

6 21

55 7 1 34 6 5 0

87 11 2 54 10 8 0

29

46

7 4 6 17 33 7 6 24 29 10

11 6 10 27 52 11 10 38 46 16

6 5

10 8

beats/min (Q1 150, Q3 179), and peak oxygen consumption was 28.1 mL/kg per minute (Q1 24.6, Q3 32.1). That 40% (25/63) of participants had a peak respiratory exchange ratio b1.0 and 32% (13/41) had a peak VO2 Z score N−1.0 indicate the heterogeneity of response. Flexibility (−0.36, Q1 −1.6, Q3 8.4) and handgrip (0.19, Q1 −0.56, Q3 0.94) Z scores were similar to healthy peers.

Child and parent activity perceptions Most children described positive feelings about physical activity (43/61 or 70%), enjoyment of physical activity (11, Q1 9, Q3 12, maximum possible score 12), having adequate skills (21, Q1 17, Q3 23, maximum score 28), and being strongly influenced by peers (7, Q1 5, Q3 8, maximum score 8). Half of the children (33/63) reported they were as active as their peers, one third (20/63) stated they were less active. Running perceptions were neutral (2, Q1 1, Q3 4, maximum score 4), whereas “being sweaty” was perceived as negative (1, Q1 1, Q3 3, maximum score 4). The children were moderately confident in their own ability to be active daily (2, Q1 2, Q3 3, maximum score 4) and perceived an equal

number of activity benefits and barriers (3, Q1 −2, Q3 11, theoretical range −41 to +35). Parents reported19 their children perform enough physical activity (3, Q1 2, Q3 4, maximum score 4) and can safely play outside (3, Q1 2, Q3 4, maximum score 4). They indicated their children were similar to peers in activity level (2, Q1 1, Q3 2, maximum score 4) and coordination (2, Q1 1, Q3 2, maximum score 4). Their children really enjoy physical activity (4, Q1 3, Q3 4, maximum score 4) but choose both sedentary and active leisure pursuits (2, Q1 1, Q3 3, maximum score 4). Parents reported moderate family and peer support for their child's activity (10.2, Q1 7, Q3 13.3, maximum score 20), although 33 parents (52%) did not know whether peers were supportive of their child's activity.

Factors associated with physical activity participation Factors significantly related to increased MVPA (Table II) were spring season, current use of antithrombotic medication, increased weekday time outdoors, lower resting heart rate, and higher parent rating of child's activity relative to peers. The spring season was associated with 1.6 to 1.7 additional hours of MVPA per week relative to the fall and winter seasons, respectively. Children currently taking antithrombotic medication performed 1.5 additional hours of activity per week. Factors significantly related to decreased MVPA were a history of arrhythmia, increased family income, the presence of a learning problem, double-inlet left ventricle anatomy, child's report of increased confidence in own ability to be active every day, and more perceived barriers to activity. Weekly activity decreased by 2.0 h/wk for children with a history of arrhythmia, 0.2 h/wk for each $10,000 increase in family income, 1.25 h/wk for children with a learning problem, and 1.0 h/wk for children born with double-inlet left ventricle anatomy. Only 4 children (2 female) with a history of arrhythmias relied on a pacemaker. Age, gender, type of Fontan, age at Fontan procedure, history of ventricular dysfunction, parent/child reports of the family's activity, parent reports of activity importance, and child's perception that activity was expected did not improve our multivariable model that explains 63% of the variance in weekly MVPA.

Discussion Cardiac function correlates with exercise capacity.25 Therefore, published training studies have assumed that cardiac function would also be strongly associated with MVPA.10 However, direct measurements of MVPA after Fontan5 found MVPA was not related to exercise capacity, medical history, or functional health status. This study demonstrates that children after Fontan typically have low levels of physical activity. Their activity is associated with some factors unique to this

American Heart Journal February 2011

414 Longmuir et al

Table II. Factors associated with increased MVPA Multivariable model⁎

Bootstrap Factor Age at study (y)║ Gender║ Spring║ vs winter Spring║ vs fall Current use of antithrombotics No arrhythmia history Lower family income¶ (per $10,000) Higher weekday time outside (per min) No learning problem# Lower resting heart rate (per beat/min) Child is less confident about being active17 Parent rating of child's activity vs peers (less/same/more) Child perceives fewer barriers to activity16 Not DILV⁎⁎

% Reliability† Obligatory Obligatory Obligatory Obligatory 66 55 46 52 51 51 48 41 42 45

Estimate (SE)‡

H of MVPA per wk§

P

(7.18) (23.05) (29.85) (31.14) (24.00) (23.35) (4.07) (0.24) (23.71) (0.86) (11.63) (14.41)

1.6 1.7 1.5 2.0 0.2 0.7/1-h outside 1.25 0.25/5-beat/min decrease 0.2 if not confident 0.5

.93 .74 b.001 b.001 b.001 b.001 .003 .005 .002 b.001 .003 .02

3.09 (1.39) 64.83 (30.70)

0.5 if no perceived barriers 1.1

.03 .04

0.63 7.60 98.73 104.72 90.41 121.73 12.28 0.67 75.79 2.95 35.31 32.79

Variables with reliability N40% in the bootstrap algorithm that did not contribute significantly to the multivariable model were ASA (62%) or warfarin use (57%), age at Fontan procedure (48%), status of fenestration (48%), activity restriction by MD (45%) or parent (43%), tricuspid atresia anatomy (42%), resting oxygen consumption (42%), flexibility (41%), Bruce treadmill Z score (40%), and thrombosis history (40%). ⁎ Our multivariable model explains 63% of the variance in weekly MVPA among our participants. † Percentage of bootstrap samples in which the factor was selected. ‡ Parameter estimate. § Increase in MVPA that would result from 1-U change in factor (h/wk). ║ Age, gender, and season of year were forced into the model because they are known to significantly influence the physical activity of healthy children. ¶ Income estimated from census data for income by postal code. # Parent reports that the child has difficulty in school (with or without formal diagnosis of learning disability). ⁎⁎ Child's unrepaired anatomy was not double-inlet left ventricle.

population, but most are similar to the activity-influencing factors for healthy children.26

Physical activity level of children after a Fontan procedure Children in this study performed 341 minutes (5.7 hours) of MVPA per week. Moderate-to-vigorous physical activity level was comparable to the only previous report of children after the Fontan procedure.5 More activity was performed on weekdays than on weekends (55 and 36 min/d, respectively). Moderate-to-vigorous physical activity level was below the 90 min/d recommended for optimal health24 and 4 to 5 times lower than the activity level of healthy children.27 The MVPA of children after Fontan is highly variable. Two children achieved the recommended 90 min/d, and 2 performed b15 min/d. Age- and gender-matched gross motor skill scores ranged from the 1st to the 95th percentile. These results establish that it is possible for children, after Fontan, to achieve excellent motor skills and live heart-healthy, active lifestyles.24 However, it appears equally likely that some children will adopt the sedentary lifestyles associated with atherosclerosis, diabetes, and obesity.8 Young children after the Fontan procedure are often small in stature (as shown in this study) but come to have a weight for height that is higher than their peers during adolescence28 when MVPA is typically lower.27 This

pattern demonstrates the importance of developing active lifestyles early, so they can be maintained lifelong. Because our participants performed 16 additional minutes of activity on weekdays, interventions targeting sedentary weekend lifestyles may be particularly effective.

Increasing physical activity Current use of antithrombotic medication, no history of arrhythmia, and the spring season of the year increase MVPA by 1.5 to 2.0 more hours per week (Figure 1). Although 10 of the 13 factors in our multivariable model are factors that are common to all children, the use of antithrombotic medication, a history of arrhythmia, and double-inlet left ventricle anatomy are unique to children with heart problems. These factors could not be changed by an intervention, suggesting that some children will find it more difficult to be active in lifestyle after the Fontan procedure. Children after a Fontan procedure are significantly more active during warmer weather. Activity participation decreases during fall and is lowest during the winter months. Seasonal variations in MVPA participation are widely recognized.13 Activity levels are highest in summer because of amenable weather and children not being in school. Moderate-to-vigorous physical activity measurements were not obtained during school holidays. Nevertheless, a significant relationship between season and weekly MVPA remained. Children who spend more

American Heart Journal Volume 161, Number 2

Figure 1

Factors most strongly associated with weekly minutes of MVPA are season of the year, arrhythmia history, and antithrombotic use. Boxes represent 25th and 75th percentiles from mean (midline of box). Whiskers represent 5th and 95th percentiles.

time outside are also known to be more active.29 Given the lower levels of physical activity during the winter months, providing specific advice and recommendations about winter activity options should be a high priority for health professionals. Children taking antithrombotic medication perform 1.5 more hours per week of MVPA, with those taking warfarin (n = 6) being the most active (424 min/wk, Q1 386, Q3 436). Children taking acetylsalicylic acid (ASA) (n = 34) were slightly less active (384 minutes, Q1 297, Q3 476). Children not taking antithrombotic medication (n = 23) were much less active (317 minutes, Q1 300, Q3 475). Children on antithrombotic medication are restricted from body contact activities; therefore, cardiologists explicitly counsel activities such as bicycling and walking that can easily occur on a daily basis. The counseling may also enable the child/family to be more confident about appropriate activities for the child leading to increased activity participation. The 29 children with a positive history of arrhythmia performed 2.0 h/wk less of MVPA than children with no history of rhythm disorders. Four children (2 female) had a pacemaker or implanted defibrillator, which typically results in body contact restrictions similar to children taking antithrombotic medication. Three of 4 children with implanted devices performed over 420 minutes (7 h/wk) of MVPA, which again suggests that specific counseling for activity restrictions encourages increased activity participation. Future research with a larger number of children with implanted devices is needed to verify the trends suggested by this research. Children from families with a lower income were more active. Higher income is associated with decreased

Longmuir et al 415

MVPA30 because children do less walking for transportation and have more sedentary screen time. The association between lower resting heart rate and higher MVPA suggests that children who have a more efficient cardiorespiratory system are more active.31 Children who perceive more barriers to physical activity are less active.16 Parents of children with learning problems indicated that activity was strongly limited by the additional time needed for academic success. These families indicated that their children would spend 2 to 4 hours per night completing school work or in tutoring. Overall, these results suggest that physical activity counseling after Fontan should be based on many of the well-established risk factors that are known to influence childhood physical activity.

Factors not associated with physical activity In contrast to the research of others,26 there was no relationship in this study between the child's activity and parent/child reports of family activity, activity expectations, or activity importance. Anecdotal parent/child reports during the study indicated that almost all families had extremely low expectations for the child's physical activity. They commented on early and frequent counseling to expect the child to have limited exercise tolerance and often expressed surprise and satisfaction with the activity level that the child had obtained. Further research would be helpful in elucidating whether the low activity expectations established for these children by health care professionals are sufficient to negate the impact of family physical activity on the child's activity participation. Surprisingly, the gender- and age-related changes in MVPA that are well documented for healthy children13 were also not observed among this group of children. Large accelerometry studies have reported ≥600 to 800 minutes of MVPA per week for children,32-34 with teens performing 350 to 500 min/wk.32-34 The 341 minutes of MVPA per week performed by children in this study indicates that these young children have the more sedentary lifestyles typical of healthy teens. These much lower physical activity levels in early childhood suggest that these children do not perform the physical activity needed for physical, motor, emotional, psychosocial, and cognitive development.35 Other factors unique to children who have had the Fontan procedure were not related to their level of MVPA. Consistent with other reports, age of Fontan repair, previous medical/surgical procedures, history of thrombosis, and ventricular dysfunction5,36 were not related to the children's current level of physical activity. Study limitations These results are directly relevant only to children with functional single ventricle after a Fontan procedure. Factors associated with MVPA among those with other

416 Longmuir et al

heart defects37 remain unknown. All eligible children known to our clinic were approached for this study. Age and gender were similar among participants and nonparticipants, but we could not assess the activity levels of nonparticipants. One might hypothesize that those who are inactive would be more likely to refuse participation, suggesting the population as a whole may be less active and more in need of physical activity counseling than the group of children who completed this study. The issue of multiple comparisons must also be considered, given the large number of variables assessed. Therefore, the results presented are from a multivariable analysis completed only on variables identified as having a reliability of N40% through a bootstrap bagging algorithm.

Conclusions and implications Children after Fontan can engage in MVPA at levels meeting, or close to, recommended guidelines. Use of antithrombotic medication (1.5 h/wk), arrhythmia history (2.0 h/wk), and the spring season (1.6-1.7 h/wk) had the biggest impact on MVPA level. Factors influencing MVPA are similar among healthy children and children after Fontan. Higher activity was associated with increased outdoor time, lower family income, lower resting heart rate, no learning problem, and fewer perceived barriers to activity. Interventions to increase MVPA among children after Fontan should examine strategies effective with healthy children because many factors amenable to change are similar. Despite often sedentary lifestyles, our patients can achieve the active lifestyles associated with optimal growth, peer interactions, and health benefits, independent of their cardiopulmonary exercise capacity. Physicians should specifically counsel their patients about physical activity after the Fontan procedure, particularly in the winter months and when activity is unrestricted. Counseling should be based on the factors associated with MVPA among healthy children and the patient's antithrombotic and arrhythmia history.

Acknowledgements The support of the participating families and the contributions of Laura Banks, Stephanie Wong, Gareth Smith, Susan Iori, Laura De Souza, Laura Fenwick, and Faith Bangawan for study assessments are greatly appreciated.

References 1. Centers for Disease Control and Prevention. How much physical activity do children need? Atlanta: Centers for Disease Control and Prevention; 2008. Available at: http://www.cdc.gov/physicalactivity/everyone/ guidelines/children.html. 2. Active Healthy Kids Canada. It's Time to Unplug our Kids. Toronto, Ontario: Active Healthy Kids Canada; 2008. 3. Centers for Disease Control. Physical activity and health: the benefits of physical activity. Atlanta: Centers for Disease Control; 2008.

American Heart Journal February 2011

4. Ginsburg KR. The importance of play in promoting healthy child development and maintaining strong parent-child bonds. Pediatrics 2007;119:182-91. 5. McCrindle BW, Williams RV, Mital S, et al. Physical activity levels in children and adolescents are reduced after the Fontan procedure, independent of exercise capacity, and are associated with lower perceived general health. Arch Dis Child 2007;92:509-14. 6. Klavora P. Foundations of Exercise Science. Toronto, Ontario: Sport Books Publisher; 2004. 7. Moola F, Faulkner GEJ, Kirsh JA, et al. Physical activity and sport participation in youth with congenital heart disease: perceptions of children and parents. Adapt Phys Activ Q 2007;25:49-70. 8. Haskell WL, Lee IM, Pate RR, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation 2007;116:1081-93. 9. Longmuir PE, Tremblay MS, Goode RC. Postoperative exercise training develops normal levels of physical activity in a group of children following cardiac surgery. Pediatr Cardiol 1990;11:126-30. 10. Rhodes J, Curran TJ, Camil L, et al. Sustained effects of cardiac rehabilitation in children with serious congenital heart disease. Pediatrics 2006;118:e586-93. 11. Medical Research Council of Canada, Natural Sciences and Engineering Research Council of Canada, Sciences and Humanities Research Council of Canada. Tri-council policy statement: ethical conduct for research involving humans. Ottawa, Ontario: Medical Research Council of Canada, Natural Sciences and Engineering Research Council of Canada, Social Sciences and Humanities Research Council of Canada; 1998. 12. Ulrich DA. Test of Gross Motor Development (TGMD-2). Austin (Tex): PRO-ED; 2000. 13. Bar-Or O, Rowland TW. Habitual activity and energy expenditure in the healthy child. Pediatric exercise medicine: from physiologic principles to health care application. Champaign (Ill): Human Kinetics; 2004. p. 64-7. 14. Tremblay MS, Langlois R, Bryan S, et al. Canadian Health Measures Survey pre-test: design, methods and results. Statistics Canada. Ottawa, Ontario: Statistics Canada; 2007. p. 24. 15. Hay JA. Adequacy in and predilection for physical activity in children. Clin J Sport Med 1992;2:192-201. 16. Garcia AW, Broda MA, Frenn M, et al. Gender and developmental differences in exercise beliefs among youth and prediction of their exercise behavior. J Sch Health 1995;65:213-9. 17. Saunders RP, Pate RR, Felton G, et al. Development of questionnaires to measure psychosocial influences on children's physical activity. Prev Med 1997;26:241-7. 18. Sallis JF, Prochaska JJ, Taylor WC, et al. Correlates of physical activity in a national sample of girls and boys in grades 4 through 12. Health Psychol 1999;18:410-5. 19. Sallis JF, Taylor WC, Dowda M, et al. Correlates of vigorous physical activity for children in grades 1 through 12: comparing parentreported and objectively measured physical activity. Pediatr Exerc Sci 2002;14:30-44. 20. Puyau MR, Adolph AL, Vohra FA, et al. Prediction of activity energy expenditure using accelerometers in children. Med Sci Sports Exerc 2004;36:1625-31. 21. Centers for Disease Control. CDC Growth Charts for the United States: methods and development. Department of Health and Human Services. Hyattsville (Md): Department of Health and Human Services; 2000. 22. Cumming GR, Everatt D, Hastman L. Bruce treadmill test in children: normal values in a clinic population. Am J Cardiol 1978;41:69-75.

American Heart Journal Volume 161, Number 2

23. Tremblay MS, Shields M, Laviolette M, et al. Fitness of Canadian children and youth: results from the 2007-2009 Canadian Health Measures Survey. Ottawa: Statistics Canada; 2010. 24. Public Health Agency of Canada. Canada's physical activity guide for children. Ottawa, Ontario: Public Health Agency of Canada; 2005. p. 1-4. 25. Takken T, Hulzebos HJ, Blank AC, et al. Exercise prescription for patients with a Fontan circulation: current evidence and future directions. Neth Heart J 2007;15:142-7. 26. Sallis JF, Prochaska JJ, Taylor WC. A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc 2000;32: 963-75. 27. Pate RR, Freedson PS, Sallis JF, et al. Compliance with physical activity guidelines: prevalence in a population of children and youth. Ann Epidemiol 2002;12:303-8. 28. Vogt KN, Manlhiot C, Arsdell GV, et al. Somatic growth in children with single ventricle physiology. J Am Coll Cardiol 2007;50: 1876-83. 29. Cooper AR, Page AS, Wheeler BW, et al. Patterns of GPS measured time outdoors after school and objective physical activity in English children: the PEACH project. Int J Behav Nutr Phys Activ 2010;7:31. 30. Kemper HCGM, Spekreuse M, Slooten J, et al. Physical activity in prepubescent children: relationship with residential

Longmuir et al 417

31.

32.

33.

34.

35.

36.

37.

altitude and socioeconomic status. Pediatr Exerc Sci 1996;8: 57-68. Ruiz JR, Rizzo NS, Hurtig-Wennlof A, et al. Relations of total physical activity and intensity to fitness and fatness in children: the European Youth Heart Study. Am J Clin Nutr 2006;84:299-303. Troiano RP, Berrigan D, Dodd KW, et al. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc 2008;40:181-8. Lopes VP, Vasques CMS, Maia JAR, et al. Habitual physical activity levels in childhood and adolescence assessed with accelerometry. J Sports Med Phys Fitness 2007;47:217-22. Nilsson A, Anderssen SA, Andersen LB, et al. Between- and withinday variability in physical activity and inactivity in 9- and 15-year-old European children. Scand J Med Sci Sports 2009;19:10-8. Bjarnason-Wehrens B, Dordel S, Schickendantz S, et al. Motor development in children with congenital cardiac diseases compared to their healthy peers. Cardiol Young 2007;17:487-98. Paridon SM, Mitchell PD, Colan SD, et al. A cross-sectional study of exercise performance during the first 2 decades of life after the Fontan operation. J Am Coll Cardiol 2008;52:99-107. Reybrouck T, Mertens L. Physical performance and physical activity in grown-up congenital heart disease. Eur J Cardiovasc Prev Rehabil 2005;12:498-502.

Correction Some abbreviations were confusing in the published article [Am Heart J 2010;160:1130-1136.e3]. These have been expanded for clarity in this updated version.

Use of intensive lipid-lowering therapy in patients hospitalized with acute coronary syndrome: An analysis of 65,396 hospitalizations from 344 hospitals participating in Get With The Guidelines (GWTG) Usman Javed, MD, a Prakash C. Deedwania, MD, a Deepak L. Bhatt, MD, MPH, b Christopher P. Cannon, MD, b David Dai, PhD, c Adrian F. Hernandez, MD, MHS, c Eric D. Peterson, MD, MPH, c and Gregg C. Fonarow, MD d Fresno, and Los Angeles, CA; Boston, MA; and Durham, NC

Objectives The study aimed to analyze the use of intensive lipid-lowering therapy (I-LLT) at discharge in a broad population of patients hospitalized with acute coronary syndrome (ACS). Background Early and intensive statin therapy in ACS has been shown to reduce cardiovascular morbidity and mortality. Utilization and predictors of I-LLT among hospitalized ACS patients are not known. Methods The GWTG database was analyzed for ACS-related hospitalizations from 2005 to 2009. The use of I-LLT (defined as dose of statin or combination therapy likely to produce N50% reductions in low-density lipoprotein [LDL]) and less intensive lipid-lowering therapy (LI-LLT) at discharge was assessed. Baseline characteristics and temporal trends in LLT were compared in these 2 treatment groups. Results

Of 65,396 patients receiving LLT, only 25,036 (38.3%) were treated with an I-LLT regimen. Mean total cholesterol, LDL, and triglycerides were significantly higher in the I-LLT group. Even among those with LDL N130 mg/dL, 50% or less received I-LLT. Predictors of I-LLT at discharge included LLT before admission, hyperlipidemia, prior coronary artery disease, increasing body mass index, and in-hospital percutaneous coronary intervention. Although there was some temporal improvement in the rate of I-LLT from 2005 to 2007, a decline in use of I-LLT was noted in 2008 and 2009. This was attributed to a sharp reduction in use of ezetimibe in combination with statin, without corresponding increases in intensive statin monotherapy.

Conclusions In a large cohort of patients admitted with ACS, most of the eligible patients were not discharged on I-LLT. These data suggest the need for better implementation of guideline-recommended intensive statin therapy in patients with ACS. (Am Heart J 2011;161:418-424.e3.) Several large studies have consistently demonstrated that lipid-lowering therapy (LLT) with 3-hydroxy-3methylglutaryl coenzyme A reductase inhibitors (statins) reduce cardiovascular risk irrespective of underlying coronary artery disease (CAD).1 In patients with stable From the aUniversity of California, San Francisco-Fresno Medical Education Program, Fresno, CA, bVA Boston Healthcare System and Brigham and Women's Hospital, Harvard Medical School, Boston, MA, cDuke University Medical Center, Durham, NC, and d University of California, Los Angeles School of Medicine, Los Angeles, CA. Vera Bittner, MD, MSPH served as guest editor for this article. Reprint requests: Prakash C. Deedwania, MD, Cardiology Section, UCSF Program at Fresno, 2615 E. Clinton Avenue (111), Fresno, CA, 93703. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2010.12.014

CAD and acute coronary syndrome (ACS), statin therapy has shown a reduction in mortality and recurrent cardiac events.2-6 These data have established a very early clinical benefit that persisted on long-term follow-up. The PROVE IT-TIMI 225 and MIRACL6 trials have shown even better clinical outcomes with early and intensive statin therapy in ACS. It is also well established that the adherence to the use of statin therapy in the post-ACS patient is directly related to statin initiation during the index admission.7 In light of above the evidence, the recent National Cholesterol Education Program Adult Treatment Panel guideline update recommended an optional low-density lipoprotein (LDL) treatment goal of b70 mg/dL for patients with ACS.8,9 Moreover, the current guidelines of the American College of Cardiology/American Heart

American Heart Journal Volume 161, Number 2

Association (ACC/AHA) recommend measurement of lipid levels on admission and instituting LLT before hospital discharge in patients with ACS.10,11 The objective of our study was to assess the use of intensive lipid lowering therapy (I-LLT) at time of discharge in patients admitted with ACS along with patient and hospital characteristics associated with use of I-LLT. This study analyzed data from the hospitals participating in AHA's GWTG-CAD program from 2005 to 2009. Temporal trends in use of I-LLT were also assessed. In patients admitted with ACS, prescription of various LLTs (various agents and their prescribed doses) at time of the hospital discharge was also assessed in relation to the patients' admission lipid profile with the probability of achieving LDL goal of b100 mg/dL and LDL b70 mg/dL.

Methods GWTG-CAD is a national initiative of the AHA to promote guidelines adherence in management of hospitalized patients with coronary artery disease. The data collection process used in this study and quality control features have been previously described.12 All participating institutions were granted waiver of informed consent by their local institutional review boards. The Duke Clinical Research Institute (Durham, NC) serves as the data analysis center and has an agreement to analyze the aggregate de-identified data for research purposes. The GTWG program is supported by the American Heart Association in part through an unrestricted education grant from the Merck Schering Plough Partnership that did not participate in the design, analysis, preparation, review, or approval of the manuscript. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents.

Study population This study was drawn from 159,713 admissions with the diagnosis of ACS (including ST-segment elevation myocardial infarction [STEMI], non-STEMI [NSTEMI], and unstable angina), between July 2005 and December 2009, from 410 participating hospitals across the United States. Patients were excluded if they left against medical advice, discontinued care, died, or were discharged to a federal hospital, hospice, or another acute care hospital. Of the 138,216 patients discharged, 119,387 (86.4%) were receiving LLT and 14,279 (10.3%) were discharged without LLT. Lipid-lowering therapy was contraindicated in 4,550 (3.3%). Of patients discharged on LLT, 53,991 admissions were also excluded because the details describing agent/dose was missing. The data from 65,396 admissions at 344 sites were complete for the purposes of this analysis and formed the final study population. Appendix Table IA (online) shows the characteristics of patients included and excluded from the study population.

Lipid-lowering therapy Intensive lipid-lowering therapy was defined as therapy likely to achieve a N50% reduction in LDL and included atorvastation 40 or 80 mg, rosuvastatin 20 or 40 mg, simvastatin 80 mg, or any

Javed et al 419

statin of any dose used in combination with ezetimibe (statin/ ezetimibe). All other LLTs were considered as less intensive LLT (LI-LLT). A secondary analysis excluding ezetimibe and statin combination was performed to assess use and trends in intensive statin monotherapy. Data collected included patient demographics, pertinent medical history, symptoms on arrival, laboratory results, inhospital treatment and procedures, discharge treatment, risk factor counseling, and patient disposition. The lipid levels obtained within the first 24 hours of hospitalization were measured at the local hospital laboratory. Yearly trends in admission lipid values and I-LLT were assessed from 2005 to 2009.

Statistical analysis Patients were divided into the I-LLT and LI-LLT categories as defined above. In addition, I-LLT rate were noted in various subgroups based on admission LDL and high-density lipoprotein (HDL). In the descriptive analysis, the mean (±SD) and percentages were reported for continuous and categorical variables, respectively. For comparison of baseline characteristics in I-LLT and LI-LLT groups, Wilcoxon rank-sum tests were used for continuous variables and χ2 tests for categorical variables. In examining the association between LDL and I-LLT, a multivariable logistic regression was used. The generalized estimating equation (GEE) method with exchangeable working correlation structure was used to account for within-hospital clustering because patients at the same hospital are more likely to have similar responses relative to patients in other hospitals (ie, within-center correlation for response). The method produces estimates similar to those from ordinary logistic regression, but the estimated variances of the estimates are adjusted for the correlation of outcomes within each hospital. The variables entered into the model are patient age, gender, race, body mass index, cardiovascular risk factors (smoking, hypertension, hyperlipidemia, diabetes mellitus, renal insufficiency, prior MI stroke, heart failure, LLT before admission), and type of ACS. A sensitivity analysis (28,724 subjects at 76 sites), confined to the centers with N70% statin medication dose reporting compliance, was used to exclude any selection bias in the primary analysis. A P value of b.05 was considered significant for the test of each variable. All analyses were performed using SAS software (version 9.2, SAS Institute, Cary, NC) by the Duke Clinical Research Institute (Durham, NC).

Results The clinical characteristics of the patient study population are shown in Table I. Admission diagnosis was MI in 91.7% patients, while the remaining patients had unstable angina. There were 41.7% of patients who were receiving LLT before the index ACS admission. Admission LDL levels were assessed in 54,892 (83.9%) of patients. The characteristics of patients with and without LDL levels assessed are shown in Appendix Table IB (online). Patients without lipid testing during hospitalization were more likely to have been receiving LLT before admission. At hospital discharge, there were 25,036 (38.3%) patients receiving I-LLT and 40,360

American Heart Journal February 2011

420 Javed et al

Table I. Baseline characteristics in intensive and less intensive LLT groups Patient characteristics Age (y) Female Race/ethnicity White Black Hispanic Asian Diagnosis STEMI/non-STEMI Unstable Angina LLT taken before Admission Prior myocardial infarction Prior stroke Peripheral vascular disease Hypertension Diabetes—IDDM Diabetes—NIDDM Hyperlipidemia Smoking (current or prior 1 y) β-Blockers ACE inhibitors ARBs Aspirin Clopidogrel Warfarin Nitrates Calcium channel blockers Aldosterone blockers Total cholesterol (mg/dL) LDL cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL)

Overall (N = 65 396)

Intensive LLT (n = 25 036)

Less intensive LLT (n = 40 360)

P value

64.7 ± 13.9 34.3%

62.6 ± 13.4 32.2%

66.0 ± 14.1 35.6%

b.001 b.001

72.3% 7.0% 6.4% 2.8%

71.4% 7.7% 6.2% 2.9%

72.9% 6.5% 6.5% 2.8%

b.001 b.001 .149 .220

91.7% 8.3% 41.7% 19.9% 8.2% 8.1% 67.8% 9.3% 15.9% 55.7% 33.5% 97.7% 72.9% 12.5% 98.2% 80.8% 10.5% 27.0% 9.0% 3.6% 170.1 ± 103.4 ± 38.1 ± 155.4 ±

92.4% 7.6% 44.4% 21.0% 7.5% 8.0% 67.5% 9.9% 15.9% 58.7% 35.8% 98.1% 75.5% 12.4% 98.6% 84.9% 10.2% 27.8% 8.4% 3.8% 174.6 ± 107.2 ± 37.9 ± 161.0 ±

91.4% 8.6% 40.0% 19.2% 8.6% 8.2% 67.9% 8.9% 15.9% 53.7% 32.0% 97.4% 71.2% 12.5% 98.0% 78.3% 10.6% 26.6% 9.4% 3.5% 167.1 ± 101.0 ± 38.3 ± 151.8 ±

b.001 b.001 b.001 b.001 b.001 .341 .319 b.001 .993 b.001 b.001 b.001 b.001 .724 b.001 b.001 .187 b.001 b.001 .066 b.001 b.001 .254 b.001

48.2 40.0 12.4 124.7

51.0 43.0 11.9 128.1

46.2 37.7 12.7 122.3

ARB, Angiotensin receptor blocker; ACE, angiotensin converting enzyme; IDDM, insulin dependent diabetes mellitus; NIDDM, non-insulin dependent diabetes mellitus.

(61.7%) receiving LI-LLT. Patients receiving I-LLT were younger, less likely to be female, and had higher admission LDL levels (Table I). There were 30.0% of patients who received statin monotherapy, whereas 8.2% received statin/ezetimibe. The characteristics of patients receiving statin monotherapy and those receiving statin/ ezetimibe combination are shown in Appendix Table II (online). Among various statins, the rate of use and dosage of various statins in I-LLT subgroup are shown in Appendix Table III (online). When the analysis was confined to 76 hospitals that collected statin dose in N70% of patients (n = 28,724), the findings were similar (39.3% on I-LLT, of which 33.1% were on statin monotherapy). In comparison, the patients without LLT dose documentation had lower rates of prior LLT, diabetes, hyperlipidemia (lower total choles-terol, LDL cholesterol, and triglycerides), established CAD, prior CABG or percutaneous coronary intervention (PCI), and acute STEMI. These patients however had higher prevalence of hypertension, PVD, prior MI or cerebrovascular accident, and NSTEMI. The excluded sites also had lower rates of revascularization (PCI or CABG) and teaching hospitals.

Factors associated with I-LLT A number of patient characteristics were more frequent in patients discharged with I-LLT (Table I). Diagnosis of STEMI, presence of ST changes/LBBB on admission ECG, and PCI with or without stent, elevated total cholesterol, LDL, and triglyceride values were more likely to be discharged on I-LLT. There was no impact of uninsured status, non–insulin-dependent diabetes mellitus, prior CABG, or HDL between the 2 groups. Table II provides the rates of I-LLT based on admission LDL and HDL levels. Multivariate analysis of these data using the GEE model demonstrated LLT before admission, history of CAD or prior MI, hyperlipidemia, LDL per 10 mg/dL rise, body mass index (BMI) increase by 5 units, PCI with stent placement, and male gender as independent predictors of I-LLT. Patients with increasing age, chronic dialysis, and unstable angina had a lower likelihood of receiving I-LLT (Figure 1). There was a marginal impact of confining the analysis to intensive statin monotherapy and center reporting compliance on GEE model results. In the statin monotherapy model, diabetes mellitus, prior PCI, and prior CABG were additional independent predictors of I-LLT.

American Heart Journal Volume 161, Number 2

Javed et al 421

Table II. Use of intensive LLT at discharge based on admission HDL-C and LDL-C levels LDL (mg/dL)

HDL (mg/dL) <40 (n = 27 762) 40-60 (n = 14 589) ≥ 60 (n = 2 737) Total (n = 45 088)

<70 (n = 9 157)

70-100 (n = 13 603)

100-130 (n = 11 918)

130-160 (n = 6 672)

≥ 160 (n = 3 738)

Total (n = 45 088)

13.29% 38.19% 5.70% 36.95% 1.31% 33.45% 20.31% 37.53%

18.71% 36.03% 9.55% 35.53% 1.92% 30.44% 30.17% 35.51%

16.15% 38.68% 8.73% 37.68% 1.56% 34.33% 26.43% 38.09%

8.76% 44.14% 5.28% 43.32% 0.76% 38.71% 14.80% 43.57%

4.66% 52.38% 3.10% 52.58% 0.53% 44.96% 8.29% 51.98%

61.57% 39.58% 32.36% 39.26% 6.07% 34.38% 100% 39.16%

Frequency missing 20,308. Numbers in bold denote the percentage of ACS hospitalizations in that cell that received intensive LLT at hospital discharge.

Figure 1

2007). However, an insignificant drop in rate of I-LLT was noted with a decline to 35.7% by December 2009. We found this to be primarily due to a significant drop in the use of statin/ezetimibe combination from 11.4% in 2007 to 3.4% in 2009 (Table III). When statin/ezetimibe combination was excluded as I-LLT, less than one third of ACS patients was treated with intensive statin monotherapy. Use of intensive statin monotherapy at discharge was 28.0% in 2005 and 33.1% in 2009, without significant change during the 2007 to 2009 period.

Discussion The present analysis shows that among hospitals participating in GWTG-CAD, most hospitalized ACS patients are not discharged on I-LLT. Even among those with admission LDL N130 mg/dL, 50% or less received I-LLT. During the first 3 years of observation in this study, there was very modest improvement in I-LLT on discharge. This trend did not persist, instead a decline in this therapeutic approach was observed during 2008 to 2009.

Factors associated with I-LLT by multivariate GEE model.

Temporal trends in the use of I-LLT were also examined. I-LLT rates since the publication of updated National Cholesterol Education Program-ATP guidelines in 2004 increased initially from 35.5% to 41.6% (2005 to

Role of I-LLT in ACS Although statins play a pivotal role in LDL reduction, they may also exhibit a pleotropic effect by decreasing extent of myocardial ischemia, remodeling, as well as promoting plaque stabilization and endothelial function.13-15 Based on these mechanistic properties, and as demonstrated in several clinical studies, it is now widely accepted that initiation of an early and intensive statin therapy in ACS is associated with reduced inpatient mortality and morbidity3,4,6,16 as well as improved longterm survival and lower rates of recurrent coronary events. In the MIRACL trial, I-LLT with atorvastatin 80 mg/d (vs placebo) was started within 24 to 96 hours of presentation with ACS.6 It was associated with a lower risk of symptomatic ischemia requiring emergent rehospitalization. This effect was independent of baseline LDL level, although LDL was decreased from 126 to 72 mg/dL in the treatment group. The clinical benefit

American Heart Journal February 2011

422 Javed et al

Table III. Temporal trends of intensive LLT and intensive statin monotherapy Year

Total (N = 65 396)

Intensive LLT overall (n = 25 036)

Yearly trend P

2005 2006 2007 2008 2009

5283 15 520 18 082 17 143 9368

1875 (35.49) 6108 (39.36) 7523 (41.60) 6188 (36.10) 3342 (35.67)

.039 .138 b.001 .221

started to exhibit at 4 weeks and then persisted for the duration of the study. The PROVE-IT TIMI 22 trial has further demonstrated that aggressive LLT in ACS, with even lower targets LDL levels, leads to reduction in revascularization and unstable angina.3,6 In PROVE-IT TIMI 22, the median LDL was decreased to 62 mg/dL on 80 mg of atorvastatin in comparison to 95 mg/dL on 40 mg of pravastatin. Similar to the findings in the MIRACL, the beneficial effects of high-dose statins emerged as early as 30 days and then persisted during the 2 years of follow-up. Based on the available evidence, the revised Adult Treatment Panel III guidelines has recommended early and I-LLT in patients admitted for ACS and has included an optional therapeutic goal of LDL b70 mg/dL in these high-risk patients.8,9 Our analysis shows that despite available evidence and recommendations, in this large cohort of hospitalized patient with ACS, 10.3% (n = 14,279) of patients with ACS were not discharged on LLT. Moreover, only 38.2% of eligible patients were discharged on I-LLT. Although LDL remains the primary goal for therapeutic intervention, the I-LLT prescribed at the time of discharge may also take HDL into consideration. The inverse relationship of HDL and with nonfatal MI and cardiovascular-related death has been demonstrated previously.17 The present study illustrates that 61.6% of patients presenting with ACS have HDL levels b40 mg/dL. To improve secondary prevention of cardiovascular risk, it may be necessary to implement additional lipid-modifying therapy (together with routine statin therapy) targeting HDL N40 mg/dL in males and N50 mg/dL in females.

Lipid measurement in hospitalized patients with ACS Although the current guidelines recommend lipid measurement in ACS, it is measured in less than half of these patients in routine clinical practice.18 This practice has been largely based on the convention that lipid levels are unreliable in ACS settings and usually associated with an initial decrement in total cholesterol and LDL.19 However, more recent data have shown less pronounced changes in lipid profile.20 In this analysis, about half of ACS patients had LDL b100 mg/dL, with I-LLT used in about 36% of such patients. Although rate of I-LLT increased with the rise in LDL, nearly half of patients with

Intensive statin monotherapy (n = 19 645) 1422 4516 5467 5220 3020

(26.92) (29.10) (30.23) (30.45) (32.24)

Yearly trend P

.448 .381 .204 .734

Ezetimibe plus statin (n = 5391) 453 1592 2056 968 322

(8.57) (10.26) (11.37) (5.65) (3.44)

Yearly trend P

.043 .202 b.001 .002

LDL N160 mg/dL were still left untreated with I-LLT. Thus, these patients had a low probability of achieving target LDL in near future. It is interesting to note that in our study, I-LLT was more likely to be used in younger patients, male, smokers, overweight patients, in those with STEMI and otherwise those more likely to undergo PCI, and those with high lipid levels. Ironically, patients with diagnosis of unstable angina, prior stroke, heart failure, and renal insufficiency were treated with less LI-LLT. Moreover, there was no difference in the type of therapy in those with prior CABG and non–insulin-dependent diabetes mellitus. The present study demonstrates the underutilization of I-LLT in the very high risk group, which is prone to recurrent ischemic cardiovascular events.

I-LLT at discharge The available evidence suggests better long-term compliance and higher survival rates in ACS patients initially discharged on statin therapy than those who were not. 7 Subsequently, the CRUSADE Quality Improvement Initiative also showed that the use of LLT at discharge among select ACS patients increased from 78% in 2000 to 88% in 2004. 21 The overall low rate of I-LLT observed in our analyses along with recent declines during 2008 to 2009 are concerning and emphasize the need for implementation of evidence-based and guideline-recommended therapy in most patients with ACS. Although intensive statin monotherapy continued to increase marginally, the drop-off in 2008 and 2009 in intensive LLT was essentially due to the decrease in use of ezetimibe in combination in statin therapy. However, there was no offsetting increased use of intensive statin therapy. These data represent interesting, but potentially unfortunate consequences of the well-publicized ENHANCE trial controversy.22-25 In its aftermath, the plethora of discussion raised further controversy about even the proven benefits of statin therapy. This has had a large impact on the contemporary practice of lipid management without providing a clear alternative to the use of ezetimibe. As a result, starting in 2008, fewer ACS patients were treated with therapy that would allow them to achieve LDL cholesterol goals recommended in national guidelines. Despite the lower use of statin/ ezetimibe combination, there was little to no shift to

American Heart Journal Volume 161, Number 2

highdose statin therapy (at least at time of hospital discharge). These data underscore yet another impact on routine clinical practice heralded by safety or lack of efficacy concerns as raised by some recent controversies.26,27 Nevertheless, ezetimibe alone or in combination with statin therapy has not been proven to change outcomes in ACS, so the full clinical implications of these treatment shifts are not yet known.

Limitations There is a potential for selection bias in this study because discharge LLT dosing data were not available in 50% of patients. There were modest baseline differences between those with and without discharge dosing of lipid therapy recorded. These factors may influence the generalizability of these findings. Furthermore, the GWTG-CAD database is voluntary and therefore may not be representative of the entire US practice. These findings may not reflect care at hospitals that differ substantially from participating hospitals. Registry hospitals tend to be larger than nonparticipating hospitals, are more likely to be affiliated with a medical school, and are more likely to have available facilities for cardiac catheterization, PCI, and cardiac surgery. GWTG-CAD participating hospitals also were provided with feedback on performance that may have also influenced the care patterns. The hospitals participating in the GWTG-CAD program may be more likely to prescribe I-LLT, such that the treatment gaps are even larger than what was observed here. Hence, the data presented here might reflect different and possibly higher rates of I-LLT than actually prescribed among patients and hospitals that differ from those participating in GWTG-CAD. Although the lipid levels obtained in this study were measured in the first 24 hours of admission, they may or may not be entirely reflective of the baseline steady-state lipid levels. Furthermore, we do not have data as to whether patients were in the fasting state. This real-world study used results of various commercially available lipid panel assays rather than results from a single central core laboratory. Although this methodology may introduce great variability to lipid testing results, this approach makes these findings more applicable to clinical practice. This study only assessed LLT dosing at time of hospital discharge. Some patients may have had subsequent change in dosing or modification of their lipid therapy regimen as an outpatient. As the current guidelines do not specify a dose of statin for ACS patients but only a target of therapy (ie, optional target of LDL b70 mg/dL in high risk patients), many clinicians may believe that the titration to I-LLT can occur post discharge, and this may explain the reason for the treatment gap. The utilization of I-LLT at hospital discharge does not necessarily indicate that patients remained adherent to their discharge regimen. The extent of dietary and exercise counseling were not available in this study.

Javed et al 423

Conclusions During the period of 2005 to 2009, only about one third of patients hospitalized with ACS were discharged on I-LLT. Even among patients with documented admission LDL, which would require N50% reduction to achieve an optional goal of LDL b70 mg/dL, only about 50% were discharged on I-LLT. Independent predictors of I-LLT at discharge included LLT before admission, history of hyperlipidemia or coronary artery disease, increasing BMI and lipid level, and in-hospital percutaneous coronary intervention. In addition, the rate of adopting I-LLT in ACS decreased significantly over the last 2 years because of a marked decline in the use of ezetimibe in combination with statin therapy without an offsetting increase in intensive statin monotherapy. These findings underscore the importance of ongoing emphasis regarding implementation of current guidelines for measuring lipids and intensive statin therapy in all ACS patients.

Disclosures Dr Bhatt: research grants from Astra Zeneca, BristolMyers Squibb, Eisai, Ethicon, Heartscape, Sanofi Aventis and the Medicines Company. Dr Deedwania: consultant/ advisory board of AstraZeneca and Pfizer. Dr Peterson: research funding from Bristol Myers Squibb, Sanofi Aventis partnership. Dr Cannon: research grants from Accumetrics, AstraZeneca, Bristol-Myers Squibb/Sanofi Partnership, Glaxo Smith Kline Intekrin Therapeutics, Novartis, Takeda, clinical advisor and equity in Automedics Medical Systems. Dr Hernandez: research grant from Johnson and Johnson, Merck, and honorarium from AstraZeneca and Medtronic. Dr Fonarow: consultant/ advisory board to Merck Schering Plough and honorarium from Abbott, AstraZeneca, Merck Schering Plough, and Pfizer. Other authors have no disclosures.

References 1. Baigent C, Keech A, Kearney PM, et al. Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet 2005;366:1267-78. 2. Blazing MA, De Lemos JA, Dyke CK, et al. The A-to-Z Trial: methods and rationale for a single trial investigating combined use of lowmolecular-weight heparin with the glycoprotein IIb/IIIa inhibitor tirofiban and defining the efficacy of early aggressive simvastatin therapy. Am Heart J 2001;142:211-7. 3. Cannon CP, McCabe CH, Belder R, et al. Design of the Pravastatin or Atorvastatin Evaluation and Infection Therapy (PROVE IT)-TIMI 22 trial. Am J Cardiol 2002;89:860-1. 4. de Lemos JA, Blazing MA, Wiviott SD, et al. Early intensive vs a delayed conservative simvastatin strategy in patients with acute coronary syndromes: phase Z of the A to Z trial. JAMA 2004;292: 1307-16.

424 Javed et al

5. Cannon CP, Braunwald E, McCabe CH, et al. Intensive versus moderate lipid lowering with statins after acute coronary syndromes. N Engl J Med 2004;350:1495-504. 6. Kinlay S, Schwartz GG, Olsson AG, et al. High-dose atorvastatin enhances the decline in inflammatory markers in patients with acute coronary syndromes in the MIRACL study. Circulation 2003;108: 1560-6. 7. Muhlestein JB, Horne BD, Bair TL, et al. Usefulness of in-hospital prescription of statin agents after angiographic diagnosis of coronary artery disease in improving continued compliance and reduced mortality. Am J Cardiol 2001;87:257-61. 8. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Final Report. Circulation 2002;106:3143-421. 9. Grundy SM, Cleeman JI, Merz CNB, et al, for the Coordinating Committee of the National Cholesterol Education Program and Endorsed by the National Heart, Lung, and Blood Institute, American College of Cardiology Foundation, and American Heart Association. Implications of Recent Clinical Trials for the National Cholesterol Education Program Adult Treatment Panel III Guidelines. Circulation 2004;110:227-39. 10. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction—executive summary. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (writing committee to revise the 1999 guidelines for the management of patients with acute myocardial infarction). J Am Coll Cardiol 2004;44:671-719. 11. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA 2002 guideline update for the management of patients with unstable angina and non–ST-segment elevation myocardial infarction— summary article. A report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines (Committee on the Management of Patients With Unstable Angina). J Am Coll Cardiol 2002;40:1366-74. 12. Sachdeva A, Cannon CP, Deedwania PC, et al. Lipid levels in patients hospitalized with coronary artery disease: an analysis of 136,905 hospitalizations in Get With The Guidelines. Am Heart J 2009;157: 111-7.e2. 13. Lefer AM, Campbell B, Shin YK, et al. Simvastatin preserves the ischemic-reperfused myocardium in normocholesterolemic rat hearts. Circulation 1999;100:178-84.

American Heart Journal February 2011

14. Every NR, Frederick PD, Robinson M, et al. A comparison of the national registry of myocardial infarction 2 with the cooperative cardiovascular project. J Am Coll Cardiol 1999;33:1886-94. 15. Di Napoli P, Antonio Taccardi A, et al. Simvastatin reduces reperfusion injury by modulating nitric oxide synthase expression: an ex vivo study in isolated working rat hearts. Cardiovasc Res 2001;51: 283-93. 16. Fonarow GC, Wright RS, Spencer FA, et al. Effect of statin use within the first 24 hours of admission for acute myocardial infarction on early morbidity and mortality. Am J Cardiol 2005;96:611-6. 17. Gordon T, Castelli WP, Hjortland MC, et al. High density lipoprotein as a protective factor against coronary heart disease: The Framingham Study. Am J Med 1977;62:707-14. 18. Ko DT, Alter DA, Newman AM, et al. Association between lipid testing and statin therapy in acute myocardial infarction patients. Am Heart J 2005;150:419-25. 19. Fyfe T, Baxter RH, Cochran KM, et al. Plasma-lipid changes after myocardial infarction. Lancet 1971;2:997-1001. 20. Pitt B, Loscalzo J, Ycas J, et al. Lipid levels after acute coronary syndromes. J Am Coll Cardiol 2008;51:1440-5. 21. Mehta RH, Roe MT, Chen AY, et al. Recent trends in the care of patients with non-ST-segment elevation acute coronary syndromes: insights from the CRUSADE initiative. Arch Intern Med 2006;166: 2027-34. 22. Simvastatin with or without ezetimibe in familial hypercholesterolemia. N Engl J Med 2008;358:1431-43. 23. Carey J. Do cholesterol drugs do any good? Bus Week 2008:52-9. 24. Taubes G. What's cholesterol got to do with it? N Y Times 2008. 25. Greenland P, Lloyd-Jones D. Critical Lessons From the ENHANCE Trial. JAMA 2008;299:953-5. 26. Roe MT, Chen AY, Cannon CP, et al. Temporal changes in the use of drug-eluting stents for patients with non–ST-segment–elevation myocardial infarction undergoing percutaneous coronary intervention from 2006 to 2008: results from the Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA guidelines (CRUSADE) and Acute Coronary Treatment and Intervention Outcomes Network–Get With The Guidelines (ACTION–GWTG) Registries. Circ Cardiovasc Qual Outcomes 2009;2:414-20. 27. Atwater BD, Oujiri J, Wolff MR. The immediate impact of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial on the management of stable angina. Clin Cardiol 2009;32:E1-3.

American Heart Journal Volume 161, Number 2

Javed et al 424.e1

Appendix Table IA. Patient characteristics of the study and excluded patients Patient characteristics Age (y) Female Race/ethnicity White Black Hispanic Asian Diagnosis STEMI/non-STEMI Unstable angina LLT taken before admission Prior myocardial infarction Prior stroke Peripheral vascular disease Hypertension Diabetes—IDDM Diabetes—NIDDM Hyperlipidemia Smoking (current or prior 1 y) β-Blockers ACE inhibitors ARBs Aspirin Clopidogrel Warfarin Nitrates Calcium channel blockers Aldosterone blockers Total cholesterol (mg/dL) LDL cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL) PCI with stent CABG

Overall (n = 159 713)

Excluded (n = 94 317)

Study cohort (N = 65 396)

P value

66.2 ± 14.4 35.6%

67.3 ± 14.6 36.5%

64.7 ± 13.9 34.3%

b.001 b.001

70.6% 7.4% 5.6% 3.2%

69.4% 7.7% 5.1% 3.5%

72.3% 7.0% 6.4% 2.8%

b.001 b.001 b.001 b.001

93.6% 6.4%

94.9% 5.1%

91.7% 8.3%

b.001 b.001

20.1% 8.9% 8.7% 68.2% 8.2% 14.1% 51.3% 31.1% 96.9% 70.8% 12.2% 97.4% 75.6% 11.3% 17.8% 6.5% 3.7% 168.7 ± 102.1 ± 38.5 ± 152.9 ± 35.4% 7.5%

20.4% 9.6% 9.2% 68.6% 7.1% 12.3% 47.2% 29.5% 96.3% 68.8% 12.1% 96.7% 69.9% 12.4% 11.4% 4.7% 3.9% 167.5 ± 101.1 ± 38.8 ± 150.6 ± 25.1% 6.5%

19.9% 8.2% 8.1% 67.8% 9.3% 15.9% 55.7% 33.5% 97.7% 72.9% 12.5% 98.2% 80.8% 10.5% 27.0% 9.0% 3.6% 170.1 ± 103.4 ± 38.1 ± 155.4 ± 50.2% 8.7%

.042 b.001 b.001 .003 b.001 b.001 b.001 b.001 b.001 b.001 .043 b.001 b.001 b.001 b.001 b.001 .102 b.001 b.001 b.001 b.001 b.001 b.001

48.2 40.0 12.7 122.3

48.1 40.1 13.1 120.0

48.2 40.0 12.4 124.7

ARB, Angiotensin receptor blocker; ACE, angiotensin converting enzyme; IDDM, insulin dependent diabetes mellitus; NIDDM, non-insulin dependent diabetes mellitus.

American Heart Journal February 2011

424.e2 Javed et al

Table IB. Patient characteristics based on measurement of lipid levels Patient characteristics Age (y) Female Race/ethnicity White Black Hispanic Asian Diagnosis STEMI/non-STEMI Unstable angina LLT taken before Admission Prior myocardial infarction Prior stroke Peripheral vascular disease Hypertension Diabetes—IDDM Diabetes —NIDDM Hyperlipidemia Smoking (current or prior 1 y) β-Blockers ACE inhibitors ARBs Aspirin Clopidogrel Warfarin Nitrates Calcium channel blockers Aldosterone blockers PCI with Stent CABG Intensive LLT Intensive statin monotherapy Statin/ezetimibe

Overall (N = 65 396)

Lipids measured (n = 54 892)

Lipids not measured (n = 10 504)

P value

64.7 ± 13.9 34.3%

64.0 ± 13.8 33.6%

68.4 ± 13.9 38.0%

b.001 b.001

72.3% 7.0% 6.4% 2.8%

72.4% 7.1% 6.4% 2.4%

71.9% 6.3% 6.2% 5.0%

.290 .002 .427 b.001

91.7% 8.3% 41.7% 19.9% 8.2% 8.1% 67.8% 9.3% 15.9% 55.7% 33.5% 97.7% 72.9% 12.5% 98.2% 80.8% 10.5% 27.0% 9.0% 3.6% 50.2% 8.7% 38.3% 30.0% 8.2%

91.9% 8.1% 38.7% 18.9% 7.7% 7.6% 67.1% 8.4% 15.0% 55.2% 34.9% 97.9% 74.0% 12.0% 98.5% 82.2% 10.4% 26.8% 8.5% 3.4% 52.4% 8.6% 38.9% 30.8% 8.1%

91.1% 8.9% 57.2% 24.9% 10.9% 10.7% 71.4% 13.7% 20.8% 58.2% 26.2% 96.3% 66.8% 15.1% 97.0% 73.5% 11.0% 28.1% 11.8% 4.7% 38.9% 9.0% 35.2% 26.2% 8.9%

.008 .008 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 b.001 .081 .008 b.001 b.001 b.001 .147 b.001 b.001 .004

Table II. Patient characteristics in intensive LLT groups: overall, statin monotherapy, and ezetimibe plus any statin Patient characteristics Age (y) Female Race/ethnicity White Black Hispanic Asian Diagnosis STEMI/non-STEMI Unstable angina LLT taken before admission Prior myocardial infarction Prior stroke Peripheral vascular disease Hypertension Diabetes—IDDM Diabetes—NIDDM Hyperlipidemia Smoking (current or prior 1 y) Total cholesterol (mg/dL) LDL cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL)

Overall I-LLT (n = 25 036) Statin monotherapy (n = 19 645) Statin/ezetimibe (n = 5 391) P value 62.6 ± 13.4 32.2%

62.4 ± 13.5 31.7%

63.6 ± 12.8 34.0%

b.001 .001

71.4% 7.7% 6.2% 2.9%

69.6% 8.1% 6.8% 3.1%

77.9% 6.3% 4.2% 2.2%

b.001 b.001 b.001 b.001

92.4% 7.6% 44.4% 21.0% 7.5% 8.0% 67.5% 9.9% 15.9% 58.7% 35.8% 174.6 ± 107.2 ± 37.9 ± 161.0 ±

93.5% 6.5% 41.3% 19.8% 7.3% 7.4% 66.3% 9.6% 15.6% 55.5% 37.3% 175.6 ± 108.5 ± 37.9 ± 158.8 ±

88.3% 11.7% 56.0% 25.2% 8.2% 10.0% 71.7% 11.1% 16.8% 70.0% 30.5% 170.8 ± 102.1 ± 38.1 ± 169.4 ±

b.001 b.001 b.001 b.001 .038 b.001 b.001 .002 .050 b.001 b.001 b.001 b.001 .085 b.001

51.0 43.0 11.9 128.1

49.9 42.5 11.9 125.0

54.5 44.6 11.9 138.4

American Heart Journal Volume 161, Number 2

Javed et al 424.e3

Table III. Specific lipid-lowering agents in treatment groups

Statin Rosuvastatin

Atovastatin

Simvastatin

Simvastatin/ezetimibe Combination

Statin/ezetimibe⁎

Stain dose

Overall (N = 65 396)

Overall (%)

I-LLT (n = 25 036)

I-LLT (%)

LI-LLT (n = 40 360)

LI-LLT (%)

5 mg 10 mg 20 mg 40 mg 10 mg 20 mg 40 mg 80 mg 5 mg 10 mg 20 mg 40 mg 80 mg 10-10 mg

649 2116 1156 417 4122 5683 7331 7919 94 1086 6665 10 887 3589 153

0.99 3.24 1.77 0.64 6.30 8.69 11.21 12.11 0.14 1.66 10.19 16.65 5.49 0.23

31 85 1156 417 52 109 7331 7919 2 24 121 284 3589 153

0.12 0.34 4.62 1.67 0.21 0.44 29.28 31.63 0.01 0.10 0.48 1.13 14.34 0.61

618 2031 0 0 4070 5574 0 0 92 1062 6544 10 603 0 0

1.53 5.03 0.00 0.00 10.08 13.81 0.00 0.00 0.23 2.63 16.21 26.27 0.00 0.00

10-20 mg 10-40 mg 10-80 mg Other

1005 1815 603 131 1690

1.54 2.78 0.92 0.20 2.58

1005 1815 603 131 1690

4.01 7.25 2.41 0.52 6.75

0 0 0 0 0

0.00 0.00 0.00 0.00 0.00

⁎Ezetimibe and any dose of any statin in separate doses.

Correction In the article “Serum 25-hydroxyvitamin D concentration is associated with functional capacity in older adults with heart failure” (Am Heart J 2010:160:893-9), the authors would like to correct two errors in their article. In the abstract methods the “Biodex leg press” should read “Biodex isokinetic dynamometer”. The second error appears in the legend to Figure 1. The correct legend should read as follows: “Univariate relationship between log transformed peak VO2 and 25OHD.” The original legend portion “Peak VO2 is back-transformed for interpretability” should be omitted.

Correction In the article “The new high-sensitivity cardiac troponin T assay improves risk assessment in acute coronary syndromes” (Am Heart J 2010:160; 224-9) an error was published in Figure 1. The correct figure and legend should appear as follows:

Figure 1

Mortality at 1 year, death or AMI at 30 days, and PCI at 30 days in patients with both cTnT and hs-cTnT less than the 99th percentile level; with cTnT less than but hs-cTnT higher than the 99th percentile level; and with both cTnT and hs-cTnT higher than the 99th percentile level, respectively.

Correction In the article “The new high-sensitivity cardiac troponin T assay improves risk assessment in acute coronary syndromes” (Am Heart J 2010:160; 224-9) an error was published in Figure 1. The correct figure and legend should appear as follows:

Figure 1

Mortality at 1 year, death or AMI at 30 days, and PCI at 30 days in patients with both cTnT and hs-cTnT less than the 99th percentile level; with cTnT less than but hs-cTnT higher than the 99th percentile level; and with both cTnT and hs-cTnT higher than the 99th percentile level, respectively.

Letters to the Editor

An important indirect drug interaction between dronedarone and warfarin that may be extrapolated to other drugs that can alter gastrointestinal function To the Editor: I read with interest the article by Shirolkar et al entitled “Dronedarone and vitamin K antagonists: a review of drug-drug interactions” in the American Heart Journal (2010;160:577-82). They did an admirable job in discussing typical drug interactions with warfarin (essentially pharmacokinetic rather than pharmacodynamic)— especially as they apply to amiodarone but not to dronedarone—as judged from dronedarone's basic pharmacology and its clinical trials. However, an indirect interaction of clinical importance was not considered in this review. Vitamin K antagonists work by inhibiting the formation of vitamin K–dependent hepatic protein synthesis, culminating in the reduction of several important clotting factors. Agents such as amiodarone and others that can alter the handling of warfarin (its binding and/or metabolism/elimination) can alter its effects, resulting in increased or decreased anticoagulant effect. These would be considered direct pharmacokinetic drug-drug interactions. However, among the factors that can alter the manifestations of warfarin therapy is the availability of vitamin K from the gut. Drugs or disorders that alter intestinal flora and the bacterial production of vitamin K therein can also affect warfarin therapy, by increasing its apparent action, resulting in a rise in the international normalized ratio (INR). This should also be true for drugs or conditions

that accelerate gastrointestinal transit time, thus reducing the time for vitamin K formation and absorption. I have now encountered 3 patients who had entirely stable INR values on a stable warfarin regimen before the administration of dronedarone. Dronedarone was administered for atrial fibrillation management in each of the 3 and showed clinically acceptable efficacy in 2. Each of the 3, however, developed some degree of diarrhea on dronedarone (one of its more common adverse effects). The diarrhea was reduced by an empirical trial of limiting both lactose products and fiber in the diet (although none of the 3 had any prior history of any food intolerance). However, during the period of diarrhea in each patient, their previously stable INR values rose (by 1½- to 2-fold) and required a reduction in their warfarin dose for correction. Hence, despite any direct pharmacokinetic interaction of dronedarone with warfarin, there can be an indirect effect, as noted in these 3 patients, of clinical significance if gastrointestinal function is altered. Physicians, therefore, need to be cognizant not only of pharmacokinetic and pharmacodynamic drug interactions but also those that can arise indirectly via an effect (from drug and/or disease) on substrates that are essential for drug action.

Am Heart J 2011;161:e5. 0002-8703/$ - see front matter doi:10.1016/j.ahj.2010.11.013

James A. Reiffel, MD Columbia University, New York, NY E-mail: [email protected]

Shirolkar’s reply to Reiffel’s letter to the editor To the Editor: We thank Dr Reiffel for his comments on the possible drug-to-drug interactions between dronedarone and warfarin. Pharmacokinetic, pharmacodynamic, and pharmacogenetic factors can all impact the response to warfarin. The variability of vitamin K in diet, malabsorption, antibiotic usage, and liver disorders are among the many factors that can influence vitamin K availability and can have an impact on response to warfarin therapy. When performing a literature search for our article, we did not encounter studies or case reports regarding the particular drug-to-drug interaction that has been men-

tioned. We would encourage readers to report such interactions that can enhance our understanding, and provide us information that can be invaluable in daily clinical practice. Am Heart J 2011;161:e7 0002-8703/$ - see front matter doi:10.1016/j.ahj.2010.11.014

Shailesh C. Shirolkar, MD Department of Medicine Duke University Medical Center Durham, NC E-mail: [email protected]