Electrocardiographic assessment of repolarization heterogeneity

Bart Hooft van Huysduynen

Electrocardiographic Assessment of Repolarization Heterogeneity Bart Hooft van Huysduynen

Electrocardiographic Assessment of Repolarization Heterogeneity

Electrocardiographic Assessment of Repolarization Heterogeneity

Bart Hooft van Huysduynen

Electrocardiographic Assessment of Repolarization Heterogeneity

PROEFSCHRIFT ter verkrijging van de graad van Doctor aan de Universiteit Leiden op gezag van de Rector Magnificus Dr. D. D. Breimer, hoogleraar in de faculteit der Wiskunde en Natuurwetenschappen en die der Geneeskunde, volgens besluit van het College voor Promoties te verdedigen op donderdag 8 juni 2006 te klokke 16.15 uur

door

Bart Hooft van Huysduynen geboren te Amsterdam in 1974

Promotiecommissie Promotores:

Prof. dr. M.J. Schalij Prof. dr. E.E. van der Wall

Co-promoter: Dr. ir. C.A. Swenne Referent:

Prof. dr. N.M. van Hemel (Hart Long Centrum Utrecht, Nieuwegein)

Overige commissieleden:

Prof. dr. A. van der Laarse Prof. dr. A. van Oosterom (Centre Hospitalier Universitaire Vaudois, Lausanne) Dr. H.W. Vliegen Prof. dr. A.A.M. Wilde (Academisch Medisch Centrum, Amsterdam)

The research described in this thesis was performed at the Department of Cardiology of the Leiden University Medical Center, Leiden, the Netherlands The study described in this thesis was supported by a grant of the Netherlands Heart Foundation ( NHF-2001B177). Financial support by the Netherlands Heart Foundation for the publication of this thesis is gratefully acknowledged.

Aan mijn ouders

© 2006 B. Hooft van Huysduynen, Leiden, the Netherlands All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission from the copyright owner.

ISBN: 978-90-9020616-5 Cover image: Scanning electron microscope picture of Purkinje fibers at the endocardium of the heart (magnification 300x). The electrocardiogram (lead V5) of dr. P.S. Monraats. Printed by: Febodruk B.V. te Enschede Financial contribution to the publication of this thesis was kindly provided by Jacques H. de Jong Stichting, J.E. Jurriaanse Stichting, St Jude Medical, Siemens, Guidant, Novartis, Bayer, Schering-Plough, AstraZeneca, Boehringer Ingelheim, BristolMyers Squibb, Servier, Pfizer, Sankyo and Medtronic.

Contents Chapter 1. Introduction: Electrocardiographic assessment of repolarization heterogeneity. Chapter 2. Validation of ECG indices of ventricular repolarization heterogeneity; A computer simulation study. J Cardiovasc Electrophysiol 2005; 16: 1097-103 Chapter 3. Hypertensive stress increases dispersion of repolarization. Pacing Clin Electrophysiol. 2004; 27: 1603-9 Chapter 4. Increased dispersion of ventricular repolarization during recovery from exercise. submitted Chapter 5. Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients is related to reduction of right ventricular volume after pulmonary valve replacement in Fallot’s tetralogy. Eur Heart J 2005; 26: 928-32 Chapter 6. Pulmonary valve replacement in tetralogy of Fallot improves the repolarization. submitted Chapter 7. Dispersion of the repolarization in cardiac resynchronization therapy. Heart Rhythm 2005; 2: 1286-93 Chapter 8. Summary and conclusions

9

43

65

83

105

117

135

159

Nederlandse samenvatting

164

Dankwoord

168

Curriculum Vitae

169

Chapter 1 Introduction Electrocardiographic assessment of repolarization heterogeneity

9

Chapter 1

Outline of chapter 1 History of the electrocardiogram and the T wave The T wave and action potentials Heterogeneity of the repolarization and arrhythmias Physiological heterogeneity of repolarization -Transmural repolarization heterogeneity -Apico-basal repolarization heterogeneity Electrocardiographic indices of repolarization heterogeneity - QT interval - Tapex-end interval - QT dispersion - T-wave amplitude - T-wave area - QRS-T angle - T-wave complexity - Ventricular gradient Aim and outline of the thesis

10

Introduction: Electrocardiographic assessment of repolarization heterogeneity

History of the electrocardiogram and the T wave The development of electrocardiography has largely taken place in Leiden. Willem Einthoven was one of the founding fathers of electrocardiography, for which he received the Nobel prize in 19241. Einthoven was head of the Leiden University Physiology Laboratory nearby the Academic Hospital2. Initially he improved Lippmann’s electrometer, which Waller had used to record the first human ECG in 18873. In 1895 Einthoven developed a mathematical formula to construct the actual ECG from the signal of the slow responsive electrometer. To discern his calculated ECG from its predecessor, he renamed the ABCD deflections into PQRST (Figure 1a) 4 . These names were universally adopted and are still in use today. He described the T wave more or less as “ein stumpf und aufwärts gerichte Spitze “. In the following years Einthoven developed the world famous string galvanometer5, which allowed recording of high quality, stable electrocardiograms. In 1902 the first so recorded ECGs were published and the actual shape of the T wave was revealed (Figure 1b)6.

Figure 1a. Einthoven calculated the electrocardiogram from the signal of the slowly responsive electrometer and called the derived deflections PQRST, names that are still in use today. Einthoven. Pflügers Arch ges Physiol 1895.

Figure 2a. In 1902 electrocardiograms recorded with the string galvanometer were first published. Einthoven. In: Herinneringsbundel Prof. Rosenstein 1902.

11

Chapter 1

The T wave and action potentials The T wave depends on differences in timing of the repolarization of myocardial cells. Schematically, when two action potentials are subtracted, a T-wave emerges7 (Figure 2). The repolarization time of a given myocardial cell consists of the summation of the activation time and action potential duration (APD).

Figure 2. Schematically, when two action potentials are subtracted, a T-wave emerges. 0 = fast depolarizing upstroke, 1 = initial rapid recovery phase, 2 = plateauphase, 3 = repolarization, 4 = resting potential. Adapted from Franz et al. Prog Cardiovasc Dis 1991.

The primary function of the cardiac electrical system is the coordination of myocardial contraction. After the upstroke of the action potential, myocardial contraction starts, thereafter the plateau phase of the action potential is responsible for the continuation of myocardial contraction. In combination with the specific organization of the myocardial fibers, the contraction of myocardial cells results in a wringing motion8 of the heart that efficiently propels the blood9. Furthermore, action potential durations have the tendency to correct for differences in activation time. In general, the earliest activated regions have the longest action potential duration and the latest activated regions have the shortest action potential duration. These repolarizing properties result in a more homogeneous repolarization10;11 and relaxation12.

12

Introduction: Electrocardiographic assessment of repolarization heterogeneity

Heterogeneity of the repolarization and arrhythmias Besides a direct relation with mechanical function, the shape of the action potential (AP) also has protective electrophysiological properties. The relatively long plateau phase of the cardiac action potential prohibits tetanus in the myocardium, which occurs relatively frequently in skeletal muscle 13. Furthermore, the tendency of APDs to compensate for different activation times diminishes repolarization heterogeneity10;11, which reduces the risk of arrhythmias. Heterogeneous repolarization facilitates the formation of functional barriers surrounded by excitable tissue14;15. Re-entrant arrhythmias may be initiated by an adversely timed stimulus that reaches such a barrier and circles around it16. As a consequence, abrupt, local differences in refractoriness facilitate re-entrant arrhythmias. Repolarization differences between nearby areas are therefore potentially more arrhythmogenic than repolarization differences between areas more distant from each other (Figure 3). Figure 3. Dispersion in repolarization or refractoriness depicted by varying grey scale in two simulated fields of 256 (16×16) electrode sites. Lower panels show histograms of these local inhomogeneity values, corresponding to the two fields. Arrowheads above histograms indicate the values of percentiles P5 ( ► ), P50 ( ▼ )(median) and P95 ( ◄ ). Although the median dispersion of refractory period is the same in both conditions, the left figure shows global dispersion, with smoothly changing differences in refractoriness. Only in the right figure local dispersion exists with possible higher susceptibility to functional barreres and re-entry arrhythmias. Local inhomogeneity values are calculated on the extreme right as the maximum (24 ms, circled) of absolute differences (4, 10, 18 and 24 ms) within a neighbourhood of four electrode sites. Adapted from Burton and Cobbe. Cardiovasc Res 2001. 13

Chapter 1

As can be inferred from the above, repolarization heterogeneity is thus linked to arrhythmogenesis due to the relationship with refractoriness. When the AP of a myocardial cell is still in its plateau phase (phase 2) the cell is absolute refractory, to the contrary, when the cell is fully repolarized (phase 4) the cell is fully excitable. Any phase in between, on the down slope of the APD (phase 3), will result in a partially excitable cell, also named the relatively refractory period, during which a strong stimulus is still able to depolarize the cell17. An exception to these principles is, for example, post-repolarization refractoriness, which can be present in ischemic myocardium18. An ischemic cell may be refractory despite having reached phase 4. Action potentials can be recorded using microelectrodes or monophasic action potential catheters19. The action potential duration is defined as the APD90, which is the time interval from upstroke of the action potential to the moment when action potential amplitude has decreased by 90 % of its maximum amplitude. In vivo, repolarization studies in animals are mostly performed using needle electrodes allowing the measurement of activation recovery intervals (ARIs). The ARI is measured from the negative deflection of the activation complex to the positive deflection of the repolarization wave on the unipolar electrogram. ARIs are a surrogate measure of APD, but with a good correlation20;21 between recorded monophasic action potentials and ARIs 20. As stated before, repolarization heterogeneity may form the substrate for an arrhythmia, but a trigger is also necessary to initiate an arrhythmia. Early after depolarizations may occur in the setting of a disturbed repolarization and may serve as this trigger. The premature stimulus itself also modifies the repolarization heterogeneity22;23. Even in patients without overt structural heart disease, closely coupled, multiple extrastimuli are able to induce ventricular fibrillation. Arrhythmias can also be maintained by continuously firing foci24-26.

Physiological heterogeneity of the repolarization Repolarization heterogeneity is mostly classified in transmural and apico-basal heterogeneity. Repolarization heterogeneity between the left and right ventricle also exists, but data are scarce.

14

Introduction: Electrocardiographic assessment of repolarization heterogeneity

Transmural repolarization heterogeneity The nature of the transmural repolarization differences is not entirely clear; some studies dispute the existence and direction of the transmural repolarization gradient27;28. The presence or absence and direction of the transmural gradient is essential for the understanding of the formation of the normal T wave and will be discussed in detail the following paragraphs. As early as in 1931 Wilson proposed the existence of a ventricular gradient, caused by non-homogeneous action potential durations throughout the heart29. Despite opposite polarities of de- and repolarization currents, human QRS complexes and T waves attain the same polarity in most ECG leads. This concordance between QRS complexes and T waves can be explained by an inverse transmural repolarization order (from epi-to-endocardium) compared to the excitation order (from endo-toepicardium)29;30. Animal studies In canines the polarity of T waves can be varied by changing transmural APD differences by local warming or cooling30. Warming is known to shorten APD and cooling is known to lengthen APD31. Epicardial warming as well as endocardial cooling cause upright, concordant T waves. Endocardial warming and epicardial cooling cause inverse, discordant T waves30. Van Dam and Durrer measured refractory periods in dogs and found the shortest refractory periods in the midwall. Intermediate APDs were recorded from the endocardium and the longest APDs from the epicardium. They reported negative T waves in unipolar leads from the epicardial surface32. On the other hand, Burgess et al. measured longer endocardial than epicardial refractory periods33. Abildskov studied refractoriness and repolarization times (defined as activation time plus refractory period) in 15 anesthetized dogs34. In 5 dogs, transmural excitation and repolarization studies were performed immediately after thoracotomy. Despite an earlier excitation, the endocardium repolarized later than the epicardium, as reflected by longer refractory periods and later repolarization times (figure 4).

15

Chapter 1

Figure 4. Despite an earlier excitation, the endocardium repolarized later than the epicardium, as reflected by longer refractory periods and later repolarization times. Recovery times are used as a surrogate for repolarization time (excitation time + refractory period = repolarization time) Adapted from Abildskov. Circulation 1975.

Spach and Barr used intramural and epicardial electrodes to measure potential distributions during excitation and repolarization35. Beforehand they recorded ECGs to ensure positive (concordant) T waves, and excluded several dogs with negative T waves. Depolarization spread in accordance with the findings of Durrer and coworkers36 from endo- to epicardium, starting at the left midseptum and ending at the base. In general, positive potentials were recorded from the epicardium compared to more negative potentials recorded from the endocardium, implying an earlier epicardial repolarization. El-Sherif et al. performed 3-D mapping of arrhythmias emerging under long QT conditions in an in-vivo canine model37. They found that subendocardial focal activity can maintain arrhythmias but may result in reentrant arrhythmias when the repolarization heterogeneity was large enough. Steep transmural differences in ARI across the wall contributed to this repolarization heterogeneity. Recently, Janse et al. published a study performed in dogs28 that was in line with the findings of Janse’s thesis published in 197138. The epicardial repolarization time was not earlier compared to the endocardial repolarization time. However, the published canine ECG showed discordant QRS complexes and T waves28 as opposed to the T and QRS concordance found in humans. Different species, and more specifically different mammals of different size may show

16

Introduction: Electrocardiographic assessment of repolarization heterogeneity

either concordance or discordance on their ECG39. In dogs concordance may be either present or absent35. In chimpanzees, a species genetically close to humans, concordance is present in most leads40. The ECGs of the giraffe as well as the humpback whale41 show discordant T waves. Several studies disputing the presence of an epi- to endocardial transmural repolarization gradient, depicted surface ECGs with discordant ECGs. Results obtained from these species can therefore not be extrapolated to the human repolarization. Before selecting animals for an invasive repolarization study, electrocardiograms should be recorded to assure concordant T waves. Human studies Franz et al. recorded left ventricular endocardial monophasic action potentials in 7 patients undergoing catheterization (for suspected coronary disease in 5 patients and aortic disease in 2 patients)10. Additionally, they measured epicardial monophasic action potentials during surgery for coronary artery bypass grafting in 3 other patients. To compare endo- and epicardial recovery times in these different patients, and during different interventions, they normalized the repolarization times (RT = activation time + APD) of endo-and epicardium on the individual QT intervals on the surface ECGs. Expressed as percentage of the QT interval, epicardial RTs (71-84 %) were shorter than endocardial RTs (80-98%). Taggart et al. measured left ventricular ARIs in 21 patients during CABG42. Measurements were performed during right ventricular stimulation at different cycle lengths and during spontaneous atrial beats. No statistical differences were found between any of the recording sites. However, when closely observing the transmural ARI graphs, a trend towards a 5 ms shorter subepicardial ARI than subendocardial ARI can be detected. Electrograms provided as example show that the epicardial ARI is 14 ms shorter than the subendocardial ARI. These differences are small but consistent. Possibly, the interindividual variation in ARI is larger than the intraindividual variation in transmural ARI, rendering them undetectable by the used statistical methods. Understandably, Taggart et al. used short needles, the edge of the deepest electrode reaching only 7.15 mm. The authors state that the first 0.5 to 1.0 mm is epicardial fat, this would mean that the center of the deepest electrode reaches only to a depth of 6.5 mm from the epicardial myocardial surface. Therefore the endocardium is virtually left out of these experiments. In conclusion, in animal studies the direction of a transmural gradient determines the 17

Chapter 1

polarity of the T wave. An epi- to endocardial gradient is responsible for concordant T waves. The results of these animal studies combined with the interpretation of the above mentioned human studies suggest that a small transmural epi- (early repolarization) to endocardium (later repolarization) repolarization gradient is likely to be present under physiological conditions in humans. M-cells M-cells may play a pivotal role in transmural repolarization heterogeneity43. Part of the debate on transmural dispersion is the discussion whether M-cells have a significant physiological effect on the repolarization. Yan and Antzelevitch demonstrated the presence of M-cells in a preparation of the left ventricular free wall43. This preparation was made by dissecting a wedge shaped part of the left ventricular wall with its supplying large epicardial artery (which was perfused subsequently). Monophasic action potentials were recorded from epi- and endocardial cells and from the mid-myocardial cells, which were named: M-cells. The M-cells in this preparation had the longest APD and the epicardial cells the shortest APD. The difference in action potential duration and amplitude between these cell layers mainly determined the morphology of the T wave in a pseudo-ECG recorded across the wedge preparation. The shorter epicardial APD and earlier repolarization time resulted in a positive T wave directed towards the epicardium. Drouin et al. confirmed the presence of M-cells in wedge preparations of 4 apparently healthy human hearts44. M-cells were found 1 mm up to 4-5 mm from the epicardial surface, constituting of approximately 30 % of the myocardial mass. M-cells demonstrated an increased rate-dependence of their already longer APD duration during electrical stimulation with 1 to 0.1 Hz. The lower the stimulation frequency, the longer the APD, thereby increasing transmural repolarization heterogeneity. Anyukhovsky et al. performed a comparative study of wedge preparations and in vivo canine hearts45. They measured APDs in transmural wedge preparations and ARIs in in vivo hearts. In the wedge preparations they found M-cells; midmyocardial cells with relatively long APDs that were more sensitive to abrupt changes in cycle length than endo- and epicardial cells. Noteworthy is that the epicardial APD were longer than endocardial APD. However, they did not find any transmural difference in (averaged) ARIs in vivo, supposedly caused by electrotonic interaction between myocar-

18

Introduction: Electrocardiographic assessment of repolarization heterogeneity

dial cells. However, in their example an endocardial (shorter repolarization time) to epicardial (longer repolarization time) gradient was present and accompanied by an ECG with a discordant T wave. Conrath and Opthof used (strand-) simulation models to study the effects of electrical coupling on transmural repolarization differences46. Their conclusion is plausible: in physiological conditions, M-cells do not introduce large transmural repolarization differences; due to intact electrotonic coupling the repolarization differences become smaller. In conclusion, from the wedge preparation studies we know M-cells exist. However, the electrophysiological significance of M-cells in the normal heart is probably small. Large, abrupt repolarization differences due to different repolarization properties of different cells are smoothened by electrotonic interaction with surrounding cells. However, arrhythmias mostly emerge under unphysiological conditions. For example, in heart failure patients connexins are down regulated, which produces uncoupling between transmural muscle layers leading to marked repolarization heterogeneity between epicardial and deeper myocardial layers. Therefore, decreased connexin expression patterns can potentially contribute to an arrhythmic substrate in failing myocardium47. Another argument against the functional significance of Mcells is that APD lengthening appears only at unphysiological slow rates. However, Torsade de pointes arrhythmias are known to be initiated after a short-long(-short) sequences48. Thus, arrhythmias mostly emerge under pathological conditions, with less electrical coupling, greater cycle length changes and adversely timed extrastimuli. These conditions may increase the electrophysiological expression of M-cells resulting in an increase of transmural repolarization heterogeneity to a critical level and an increased susceptibility to arrhythmias.

19

Chapter 1

Apico-basal repolarization heterogeneity Besides a (small) transmural gradient, an apico-basal gradient is probably also present under normal conditions. However, data on apico-basal repolarization heterogeneity is contradicting. The apico-basal gradient is also crucial for the inscription of the normal T wave. The normal T wave forms the basis for further studies on irregular T waves and electrocardiographic indices of repolarization heterogeneity throughout this thesis. An essential aspect of the discussion on the apico-basal gradient is the issue whether action potentials overcompensate or undercompensate for differences in activation times. This would imply difference or similarity between activation and repolarization patterns. In the following studies on the apico-basal gradient, repolarization and activation times were measured parallel to the ventricular walls, on either the epicardium or the endocardium, or both. Subsequently, we will present an analysis based on the characteristics of QRS and T vector loops recorded in healthy subjects. Burgess et al. reported shorter refractory periods at the base than at the apex in dogs33. Restivo et al. however found shorter apical than basal APDs in guinea pigs measured with voltage sensitive dye. Long QT syndrome type 3 was mimicked with anthopleurin-A. Anthopleurin-A exacerbated the normal epicardial uniform apex-base APD gradient, resulting in heterogeneous repolarization gradients, functional blockades, re-entry and ventricular tachycardias49. Gepstein et al. measured endocardial ARI in 13 swine during atrial activation and ventricular pacing50. He found that even after a short period of pacing, ARIs adapted to and compensated for depolarization times. In most pigs ARIs did not overcompensate for the depolarization time, causing repolarization patterns to follow the depolarization patterns. Refractoriness was measured in 3 swine and appeared to be inversely related to the depolarization time, thus overcompensating the depolarization times. The quickness of this ARI adaptation to depolarization times suggests that electrotonic coupling played an important role in the shortening of ARI. Franz et al. showed that differences in activation time were compensated by action potential duration differences so that repolarization was nearly homogeneous measured on the endocardium in some patients as well as measured on the epicardium 20

Introduction: Electrocardiographic assessment of repolarization heterogeneity

in other patients10. Only a trend towards a longer repolarization time of the first activated regions (diaphragmatic and apico-septal) compared to the later activated regions was present. When all activation times (AT) and action potential durations of individual patients were plotted, an inverse relation was found with an average slope greater than negative unity (-1.34), which shows that action potential duration overcompensate for activation time, thereby contributing to concordant T waves. Yuan et al. measured endocardial monophasic APD in eight patients referred for arrhythmia treatment to the electrophysiological laboratory and in 10 swine51. They showed that in most patients and swine, repolarization followed depolarization, despite shorter MAP at later activated sites. In patients the slope between AT and APD was -0.45, showing an incomplete compensation of differences in activation times. Yue et al. measured in 13 patients, mostly referred for idiopathic monomorphic ventricular tachycardia, ARI with a non-contact mapping system11. They showed that on the endocardial surface ARI compensated for AT, but not over-compensated, as the overall regression slope between activation times and ARIs was -0.76. During sinus rhythm, RTs were better compensated (slope - 0.81) than during premature stimulation (slope -0.61). Summarizing, data on human repolarization is limited and contradicting. The discussed older study shows an overcompensation of depolarization time by the action potential durations. The repolarization pattern measured on either the endo- or epicardium (parallel to the ventricular walls) is suggested to be the reverse of the depolarization pattern. However, the more recent studies discussed here found that repolarization followed the depolarization order at the endocardium. The patients in the above studies were not completely healthy, therefore analysis of the QRS and T vector loops of healthy subjects may provide some additional information regarding the normal sequence of repolarization. Typically, the same global orientation and direction of inscription of the QRS and T loops was found in healthy humans52-54. These properties are in agreement with a similar de- and repolarization order measured parallel to the ventricular walls, from midseptum to apex ending at the lateral base, and a reversed transmural repolarization sequence from epi- to endocardium53. Furthermore, the normal T vector loop points to the apex and is smaller and more elongated than the QRS vector loop. These properties implicate a smaller heteroge21

Chapter 1

neity of repolarization times than of activation times, with a larger apico-basal than transmural repolarization gradient (Figure 5). Figure 5. Proposed normal human repolarization order in accordance with vectorloop morphology

and direction.The heart is shown in a horizontal plane. Transventricular repolarization follows activation sequence from sepum to apex to base. Transmural repolarization gradient is small, but inversely directed from epi- to endocardium.

ECG indices of repolarization heterogeneity Repolarization heterogeneity predisposes to arrhythmias14;15;37. Therefore, a non-invasive index of this repolarization heterogeneity potentially would have great clinical value. More than hundred years after its discovery, the ECG is still an easily available, cheap and valuable diagnostic test. At present, the standard 12-lead configuration is most used in the routine clinical setting. Therefore, we used the 12-lead configuration throughout this thesis. QT interval The traditional electrocardiographic repolarization index is the QT interval, defined as the interval from the start of the earliest QRS complex to the latest end of the T-wave in any lead. The QT interval represents the interval from the earliest depolarization to the end of the repolarization anywhere in the heart. The risk of arrhythmias in long QT patients increases with the duration of the QT(c) interval55. A large proportion of drugs with arrhythmogenic side-effects decrease the 22

Introduction: Electrocardiographic assessment of repolarization heterogeneity

rapid delayed rectifier, a repolarizing potassium current, thereby lengthening the QT interval56. Therefore, the American Federal Drug Administration requires QT interval testing for every new drug before market release is authorized. Despite its widespread use, the QT interval has some important limitations as estimator of repolarization heterogeneity. By definition, the QT interval is dependent on the longest action potential durations. However, the duration of the longest action potentials is not related to repolarization heterogeneity per se. For example, amiodarone lengthens the action potential durations homogeneously throughout the ventricular wall and has an anti-arrhythmic effect rather than a pro-arrhythmic effect57. The QT interval varies with heart rate. To estimate the QT interval during varying heart rates, correction factors are needed. The most commonly used formula is Bazett’s58, probably due to its simplicity. QTc = QT / √RR However, the Bazett formula has a tendency to overcorrect the QT interval at fast heart rates and to undercorrect the QT interval at slow heart rates, see Figure 659. Measurement of the end of the T-wave is often difficult due to slowly decreasing slopes at the end of the T-wave, low amplitude T waves and overlap with the P-wave at fast heart rates60. Furthermore, the practical and theoretical disputes to discern the end of the T wave from the U wave have not been settled61;62. A practical solution is often chosen to set the end of T at the crossing of the baseline with the steepest tangent to the descending part of the T wave63 or at the T-U nadir60. In summary, the QT interval represents the end of the repolarization anywhere in the heart and may be useful in conditions which are characterized by a lengthened repolarization. However, the QT interval does not directly assess repolarization heterogeneity.

23

Chapter 1

Figure 6. The Bazett formula has a tendency to overcorrect the QT interval at fast heart rates and to undercorrect the QT interval at slow heart rates. Lecocq et al. Am J Cardiol. 1989.

Tapex-end interval More recently, Yan and Antzelvitch proposed the Tapex-end interval, the interval from the apex of the T-wave to the end of the T-wave, to assess transmural repolarization heterogeneity43. Their proposal is based on sophisticated experiments in a wedge preparation of the left ventricular free wall of canine hearts64. As stated before, their findings appoint the midmyocardial M-cells as the cells with the longest APD. The APD of these cells were reflected in the end of the T wave in the pseudo-ECG they recorded across the preparation (Figure 7). The epicardial cells appeared to have the shortest APDs and the end of the repolarization in these cells coincided with the moment of the apex of the T-wave. They concluded that the Tapex-end interval is therefore a measure of the difference between the epicardial and mid-myocardial repolarization times, i.e., transmural repolarization heterogeneity. Their experiments provide pathophysiological insight in various forms of long QT syndrome. A large number of drugs with arrhythmic side effects are known to inhibit the delayed rectifier current. In wedge experiments these properties were shown to

24

Introduction: Electrocardiographic assessment of repolarization heterogeneity

preferentially lengthen the APD of cells with already the longest APD, the APD of the M cells. This increases the differences in repolarization time between M-cells and surrounding cells, thereby increasing repolarization heterogeneity and facilitating arrhythmias, which could be induced even in a small preparation as the wedge preparation65. Congenital long QT syndrome, types 1, 2 and 3 were mimicked in the wedge with respectively, IKs, IKr and INa inhibition66;67. These experiments showed that transmural repolarization heterogeneity was the culprit for torsade de pointeslike arrhythmias, and that transmural dispersion was adequately reflected in the Tapex-end interval in the pseudoFigure 7. The apex and the end of the T-wave re- ECG of the wedge preparation. corded in the pseudo-ECG were linked to the repo- Although their experiments were ellarization of different cell layers in the ventricular wall. The end of repolarization of the cells with the egant and insightful, the finding that longest APDs, the midmyocardial M-cells, coincided Tapex-end was a measure of transmuwith the end of the T wave. The end of repolariza- ral dispersion of the repolarization in tion of the cells with the shortest APDs, the epicarthe pseudo-ECG, recorded across the dial cells, coincided with apex of the T wave. Yan and Antzelevitch. Circulation 1998. wedge preparation, led to the assumption that these findings could be ex68-70 trapolated to the surface ECG . This, however, is unlikely. First of all, one would need a transmural ECG. This is definitely not the case in ECG leads recorded with electrodes at some distance from the heart, i.e., the limb leads and leads V5 and V6. Laws of physics prescribe that at every moment all leads register the electrical field generated by the heart and that the amplitude of a recorded potential is inversely dependent on the squared distance from the source. Even if in leads V2, V3 and V4, which are recorded from electrodes overlying the heart, would reflect some local information, this would be a composite of information obtained from the right ven25

Chapter 1

tricle and information from the anterior wall of the left ventricle. Therefore also these precordial leads do not reflect pure transmural dispersion of the repolarization. The apex of the T-wave is probably inscribed when, spreaded across the heart, most cells simultaneously repolarize. The end of the T wave is inscribed when the cells with the longest APD, wherever in the heart, are repolarized. So the Tapex-end interval has an indirect link with repolarization heterogeneity generated in the whole heart71. These theoretical deductions were recently evaluated by detailed mapping studies using more than 50 epi- and endocardial monophasic action potential recordings in pigs. In these experiments, the apex of the T wave coincided with the earliest repolarization of cells anywhere in the heart while the end of the T wave was recorded when the last cells repolarized, which were endocardial cells in nine out of ten pigs72. During electrical stimulation the origin of the Tapex-end interval is completely different. The Tapex-end interval is related to the progression of the repolarization wave front spreading from the pacing electrode through the heart73. The slow myocardial cell-to-cell activation has a significant influence on the de- and repolarizing time of the cells, so that not only the depolarization but also the repolarization spreads from the site of stimulation through the heart. The opposite direction of the repolarizing current compared to the depolarizing current causes an opposite orientation of the T-wave compared to the QRS complex observed during pacing. During abnormal, slow activation the Tapex-end interval is related to the interval from the moment the repolarization wave front approximately attains is maximal surface area (Tapex) to the moment when the last parts of heart repolarize (Tend). The areas of the heart that repolarize last are the areas that are geometrically and electrophysiologically farthest away from the pacing electrode. In chapter 5, the behavior of the Tapex-end interval during pacing is further clarified. QT dispersion In 1990, QT dispersion, calculated as the longest minus the shortest QT interval in any of the 12 standard ECG leads, was introduced as a marker of local repolarization heterogeneity74. Promising initial studies, in which QT dispersion was associated with arrhythmic risk74;75, triggered a large number of studies on this new ECG index. More than 850 publications can be currently retrieved (April 2006). However, more recently, QT dispersion as a measure of local heterogeneity came under serious 26

Introduction: Electrocardiographic assessment of repolarization heterogeneity

criticism76;77. QT-interval differences in ECG leads depend on different projections of the (global) heart vector on the 12 lead vectors. The end of the QT interval in an ECG lead is partly determined by the angle of the terminal T vector with the lead vector76. A terminal T vector that is directed perpendicular to a lead vector, is not registered in that lead and shortens the QT interval. Therefore, the end of the T-wave in a certain lead can not be interpreted as the end of repolarization in the myocardial region closest to that lead. Thus, QT dispersion does not represent local repolarization differences. An additional problem of QT dispersion is the low reproducibility78;79, for example due to the subjectivity involved in exclusion of low-amplitude T waves and the difficult measurement of the end of T waves in noisy ECGs. These insights strengthened the opposition against QT dispersion80-82. Nevertheless, QT dispersion may have a weak relation with repolarization disturbances and was associated with arrhythmias in some studies83;84. A large QT dispersion is for example found in patients with low amplitude T waves. Low amplitude T waves can be caused by triangulation of the APD85, which is thought to be related to a decreased repolarization reserve in, for example, patients with the long QT syndrome, i.e., the patient group in which the initial promising results were found74. Articles providing data on the risk-estimating capabilities of QT dispersion are still frequently published. Although QT dispersion is an indirect measure of repolarization heterogeneity, QT dispersion remains in use, probably due to habituation and its accessibility. T-wave amplitude T-wave amplitude, the maximal amplitude of the T-wave or the apex of the T-wave, reflects the maximal potential difference in the heart during the repolarization. The larger the differences in duration and amplitude of action potentials from different parts of the myocardium, the higher the T-wave amplitude becomes86. Syndromes or conditions associated with heterogeneity of the repolarization causing large-amplitude T waves are long QT syndrome type 1 and hyperkalemia. The high amplitude T waves of long QT-1 syndrome were realistically mimicked in wedge preparations of the canine left ventricular wall66. An IKs blocker in combination with isoprenaline caused a relatively long APD in the M-cells, causing a high amplitude T wave in the pseudo-ECG recorded across the ventricular wall. These findings concur with the observed propensity for arrhythmias in long QT 1 patients associated with exercise87;88. Hyperkalemia is well known to cause high amplitude T waves and ar27

Chapter 1

rhythmias, likely to be due to increased repolarization heterogeneity89. Furthermore, in the acute phase of myocardial infarction the AP of the infarcted cells changes; the upstroke velocity falls, maximum amplitude dips and APD shortens90. This causes potential differences between injured and normal cells and a systolic “injury” current directed from normal to ischemic cells. These injury currents cause ST elevation and increased T-wave amplitude. Although T-wave amplitude is an indicator of repolarization heterogeneity in the above conditions, T-wave amplitude is insensitive or even misleading to certain other changes in AP morphology. Triangulation of APs in response to pro-arrhythmogenic drugs is associated with increased repolarization heterogeneity and decreases T-wave amplitude91. Also, long QT syndrome type 2 is typically associated with low amplitude T waves92. Furthermore, a high inter-individual variation in T-wave amplitude exists due to variations in body fat and internal ventricular diameter. Additionally, cancellation has a strong influence on the T-wave amplitude. Based upon animal and modeling studies Abildskov and Klein assessed the amount of cancellation during ventricular depolarization to be approximately two thirds of locally generated potential differences93. Based upon measurements of refactoriness by Durrer32, Burgess and Abildskov assessed the amount of cancellation during repolarization even more than 90 %, due to opposed directions of repolarization vectors within the wall94. The lower T-wave amplitude observed during biventricular pacing compared to single sided pacing is probably caused by a larger cancellation of two repolarization wave fronts instead of one wave front73;94. In conclusion, the use of T-wave amplitude as a measure of repolarization heterogeneity has serious limitations, but T-wave amplitude may be an accurate reflection of repolarization heterogeneity in specific conditions. T-wave area T-wave area reflects magnitude and duration of repolarization differences throughout the heart. Several studies showed a relation between repolarization heterogeneity, assessed from a limited number of action potential recordings, and T-wave area. T-wave area correlated with increased repolarization heterogeneity in rabbit hearts measured by epicardial monophasic action potentials95 In dogs, T-wave96 and QRST97 surface area was related to repolarization heterogeneity and a lowered threshold for ventricular fibrillation.98 Drugs that lengthen APD of specific cell layers, for ex28

Introduction: Electrocardiographic assessment of repolarization heterogeneity

ample, IKr blockers that foremost lengthen midmyocardial APD, cause an increased T-wave area in left ventricular wedge experiments99. Mathematical simulation studies confirmed these experimental findings. Human heart-in-thorax models showed that increased repolarization heterogeneity resulted in increased T-wave area100;101, as presented in chapter 4 of this thesis101. T-wave area can be determined in individual leads or in the vector magnitude constructed by means of the inverse Dower matrix102;103. T-wave area can be objectively and automatically measured. Furthermore, this index is relatively insensitive to noise. Small, peaked oscillations will average out, having little influence on total T-wave area. Its disadvantage is, like T-wave amplitude, a high inter-individual variation which makes individual risk assessment difficult if only one ECG is available. Serial ECG analysis started before a potentially pro-arrhythmic event would therefore be more suited to evaluate repolarization heterogeneity by measurement of T-wave area. QRS-T angle An increased QRS-T angle is predictive for (sudden) death. Kardys et al. showed that a wide QRS-T angle predicted cardiac death in a general population of more than 6000 men and women older than 55 years 104. After adjustment for cardiovascular risk factors, hazard ratios of abnormal QRS-T angles for sudden death were 4.6 (CI 2.5-8.5). Zabel et al. tested five ECG variables of T-wave morphology in patients after myocardial infarction 105. Only the spatial angle between depolarization and repolarization was shown to contribute to the risk stratification of these patients, independent of classical risk factors. Other studies underscored the prognostic value of the spatial QRS-T angle and the orientation of the T axis 106-108 The spatial angle between the QRS and T vectors is normally 78 ± 26º 109. A small QRS-T angle is caused by a similar direction of de- and repolarization. Several pathologies may cause a wide QRS-T angle. An altered activation pattern may cause a similar de- and repolarization sequence as described in the paragraph on the Tapexend interval. Due to the similar de- and repolarization order, but opposite direction of the currents, the QRS and T vectors then attain an opposite direction, and a large QRS-T angle. Furthermore, ischemia may alter the repolarization process and the direction of the T vector. Any other condition that disturbs the normal distribu29

Chapter 1

tion of the action potential durations throughout the heart may increase the QRS-T angle. The multitude of pathologies that may cause an increased QRS-T angle explains the excellent predictive value of this ECG index. Therefore, an increased QRS-T angle is a final common pathway and not very specific for increased repolarization heterogeneity. Nevertheless, this ECG index can be measured accurately (and automatically) and may therefore be useful as a general indicator of the electrophysiological status of the cardiac patient. T-wave complexity Methods used to describe the complexity of the T wave are essentially morphological descriptors of the T-waves in the ECG. We used singular value decomposition to calculate T-wave complexity in this thesis. Singular value decomposition was introduced in cardiology as an algebraic algorithm to distillate non-redundant signals from multiple leads (up to 200) obtained with body surface mapping. We used singular value decomposition to reconstruct the T waves of the eight independent ECG leads (I, II, V1-6) into 8 independent components that are by definition orthogonal to each other (Figure 8). If the T waves can be described by only the first few components, the T waves have a relatively simple shape and are similar to each other in the different leads. The more components are needed to accurately describe the T waves and thus contain a significant amount of information, the more complex the T waves. The energy contained in the eight components is quantified by the corresponding singular values. To calculate T-wave complexity, we divided the higher, more complex singular values 2 to 8 by the first, most simple singular value. Another variation we and others used was the ratio of the second to the first singular value110. Although one influential group used the absolute value of the singular values 4-8111, according to our observations, these highest singular values have a low signal-to-noise ratio. Nevertheless, this method appears to have prognostic capabilities112;113. Furthermore, T-wave complexity has been shown to yield independent prognostic information in patients with cardiovascular disease113. In patients with arrhythmogenic right ventricular dysplasia, higher T-wave complexity is associated with arrhythmias 114. Additionally, T-wave complexity is increased in patients with primary repolarization disturbances and can be used to discriminate these patients from healthy individuals 110 . Van Oosterom mathematically proved that a higher repolarization heterogeneity leads to increased T-wave complexity.115 30

Introduction: Electrocardiographic assessment of repolarization heterogeneity

Figure 8. Singular value decomposition was used to reconstruct the T waves of the eight independent ECG leads (I, II, V1-6) into 8 independent components that are by definition orthogonal to each other.

Our opening statement that T-wave complexity is essentially an index of the morphology of the T wave can now be further refined. T-wave complexity is related to the simplicity of the T-wave form; a smooth T-wave that is similar in different leads can be described by fewer singular values than an irregularly shaped T-wave. Although every clinician knows that irregular T-wave morphology deflects an abnormal repolarization, the advantage of singular value decomposition is the objective quantification of T-morphology aberrancy. Ventricular gradient The concept of the ventricular gradient was originally formulated by Wilson in 193129. The ventricular gradient is calculated by summation of the integral of the spatial depolarization and repolarization vectorloops. This results in the gradients of

31

Chapter 1

AP differences only, while excluding the influence of the depolarization order. When activating the heart from an ectopic focus (as a ventricular extra stimulus), the ventricular gradient was supposed to remain unaltered as the ventricular gradient reflected heterogeneity of action potential (duration and amplitude) and not the activation order. Despite the attractive theoretical background, APD appeared to be influenced by the depolarization order. Adaptation of the APD to activation order that persists after restoration of the original activation order is apparent even after a short time of ectopic activation, a phenomenon which is called T-wave memory116. Furthermore, the direction of the activation wavefront compared to the fiber direction has an effect on APD117. Moreover, the mechanism of arrhythmogenesis is dependent on repolarization differences, which is the resultant of activation and APD; or more specifically refractoriness, which has been shown to facilitate re-entry arrhythmias. Nevertheless, the ventricular gradient still reflects the APD heterogeneity, whether this APD pattern is modified by the activation pattern or not. The ventricular gradient remains an interesting concept in research-oriented ECG analysis. The ventricular gradient can be particularly useful to discern between primary and secondary repolarization changes. Aim and outline of the thesis Repolarization changes due to several interventions were evaluated by measurement of several ECG indices of repolarization heterogeneity in several groups of healthy subjects and patients. Detailed study of different ECG indices in different patient groups may provide insight in their behavior and may guide the appropriate use of these electrocardiographic indices of repolarization heterogeneity. In chapter 2 healthy males are subjected to normotensive stress (modified tilt testing) and hypertensive stress (handgrip). Different tilt angles of the legs are applied to achieve the same heart rate during both stressors to be able to compare the effects of these stressors on electrocardiographic indices of repolarization heterogeneity without the errors introduced by heart rate correction. An increase in sudden cardiac death has been observed during or immediately after exercise. In chapter 3 we measure electrocardiographic indices of repolarization heterogeneity during and after maximal exercise testing. The fitness level of the subjects varied from professional marathon ice skaters to untrained subjects. The response to vigorous exercise is compared in athletes with the largest hearts to the untrained subjects. 32

Introduction: Electrocardiographic assessment of repolarization heterogeneity

Electrocardiographic indices may differ in their reaction on increasing repolarization heterogeneity. In chapter 4 a mathematical ECG simulation model is used to observe whether various ECG indices adequately reflect increasing local repolarization heterogeneity. In chapter 5 the electrocardiographic effects of pulmonary valve replacement in Fallot patients with dilated right ventricles are studied. In these patients QRS duration is known to be predictive of ventricular arrhythmias. We use an interactive ECG analysis program to accurately measure the QRS duration before and a half year after surgery. Changes in right ventricular end-diastolic volumes were previously studied with cardiac magnetic resonance imaging and are incorporated in the present study. In chapter 6 we extend the analysis of the Fallot patients with electrocardiographic indices proposed to measure repolarization heterogeneity. We measure the changes in these indices due to pulmonary valve replacement and study the possible relation with arrhythmias. Previous studies on cardiac resynchronization therapy have suggested a detrimental effect of epicardial pacing on the transmural repolarization heterogeneity, causing arrhythmias in vulnerable patients, who can be identified by electrocardiographic evaluation. In chapter 7 we study the effects of different pacing modes of cardiac resynchronization devices on electrocardiographic indices of repolarization heterogeneity. Subsequently we use a simulation model to interpret the electrocardiographic findings from our heart failure patients.

33

Chapter 1

REFERENCES 1. Schalij MJ, Janse MJ, van Oosterom A, Wellens HJ, Van der Wall EE. Einthoven 2002, 100 years of electrocardiography. 2002. The Einthoven Foundation, Leiden. 2. Snellen FA. Willem Einthoven, Father of electrocardiography. 1995. Kluwer Academic Publishers, Dordrecht. 3. Waller AD. A demonstration on man of electromotive changes accom-panying the heart’s beat. J Physiol. 1887;8:229-234. 4. Einthoven W. Über die Form des menschlichen Electrocardiogramms. Pflügers Arch ges Physiol. 1895;60:101-123. 5. Einthoven W. Un nouveau galvanomètre. Arch néerl des Sciences Exact Nat série 2. 1901;6:625-633. 6. Einthoven W. Galvanometrische registratie van het menschelijk electrocardiogram. In: Herinneringsbundel Prof. Rosenstein. 1902. Eduard IJdo, Leiden. 7. Franz MR, Bargheer K, Costard-Jackle A, Miller DC, Lichtlen PR. Human ventricular repolarization and T wave genesis. Prog Cardiovasc Dis. 1991;33:369384. 8. Torrent-Guasp F, Buckberg GD, Clemente C, Cox JL, Coghlan HC, Gharib M. The structure and function of the helical heart and its buttress wrapping. I. The normal macroscopic structure of the heart. Semin Thorac Cardiovasc Surg. 2001;13:301319. 9. Kilner PJ, Yang GZ, Wilkes AJ, Mohiaddin RH, Firmin DN, Yacoub MH. Asymmetric redirection of flow through the heart. Nature. 2000;404:759-761. 10. Franz MR, Bargheer K, Rafflenbeul W, Haverich A, Lichtlen PR. Monophasic action potential mapping in human subjects with normal electrocardiograms: direct evidence for the genesis of the T wave. Circulation. 1987;75:379-386. 11. Yue AM, Betts TR, Roberts PR, Morgan JM. Global dynamic coupling of activation and repolarization in the human ventricle. Circulation. 2005;112:2592-2601. 12. Nickerson D, Smith N, Hunter P. New developments in a strongly coupled cardiac electromechanical model. Europace. 2005;7 Suppl 2:118-127. 13. Binah O. Tetanus in the mammalian heart: studies in the shrew myocardium. J Mol Cell Cardiol. 1987;19:1247-1252. 14. Han J, Moe GK. Nonuniform recovery of excitability in ventricular muscle. Circ Res. 1964;14:44-60. 15. Kuo CS, Munakata K, Reddy CP, Surawicz B. Characteristics and possible mechanism of ventricular arrhythmia dependent on the dispersion of action potential durations. Circulation. 1983;67:1356-1367. 16. El Sherif N, Caref EB, Yin H, Restivo M. The electrophysiological mechanism of ventricular arrhythmias in the long QT syndrome. Tridimensional mapping of activation and recovery patterns. Circ Res. 1996;79:474-492. 17. Hoffman B.F., Cranefield P.F. Electrophysiology of the heart. 1960. McGraw Hill, New York. 18. Downar E, Janse MJ, Durrer D. The effect of “ischemic” blood on transmembrane potentials of normal porcine ventricular myocardium. Circulation. 1977;55:455462.

34

Introduction: Electrocardiographic assessment of repolarization heterogeneity 19. Franz MR. Long-term recording of monophasic action potentials from human endocardium. Am J Cardiol. 1983;51:1629-1634. 20. Haws CW, Lux RL. Correlation between in vivo transmembrane action potential durations and activation-recovery intervals from electrograms. Effects of interventions that alter repolarization time. Circulation. 1990;81:281-288. 21. Xia Y, Kongstad O, Hertervig E, Li Z, Holm M, Olsson B, Yuan S. Activation recovery time measurements in evaluation of global sequence and dispersion of ventricular repolarization. J Electrocardiol. 2005;38:28-35. 22. Laurita KR, Girouard SD, Rosenbaum DS. Modulation of ventricular repolarization by a premature stimulus. Role of epicardial dispersion of repolarization kinetics demonstrated by optical mapping of the intact guinea pig heart. Circ Res. 1996;79:493-503. 23. Yuan S, Blomstrom-Lundqvist C, Pehrson S, Pripp CM, Wohlfart B, Olsson SB. Dispersion of repolarization following double and triple programmed stimulation. A clinical study using the monophasic action potential recording technique. Eur Heart J. 1996;17:1080-1091. 24. Brachmann J, Scherlag BJ, Rosenshtraukh LV, Lazzara R. Bradycardia-dependent triggered activity: relevance to drug-induced multiform ventricular tachycardia. Circulation. 1983;68:846-856. 25. Cranefield PF, Aronson RS. Torsades de pointes and early afterdepolarizations. Cardiovasc Drugs Ther. 1991;5:531-537. 26. Janse MJ. Electrophysiological changes in heart failure and their relationship to arrhythmogenesis. Cardiovasc Res. 2004;61:208-217. 27. Conrath, C. E. Thesis: Ventricular repolarization and the long QT syndrome. 129151. 2005. Utrecht University. 28. Janse MJ, Sosunov EA, Coronel R, Opthof T, Anyukhovsky EP, de Bakker JM, Plotnikov AN, Shlapakova IN, Danilo P, Jr., Tijssen JG, Rosen MR. Repolarization gradients in the canine left ventricle before and after induction of short-term cardiac memory. Circulation. 2005;112:1711-1718. 29. Wilson F.N., MacLeod A.G., Barker P.S. The T deflection of the electrocardiogram. Trans Assoc Am Physicians. 1931;46:29-38. 30. Hellerstein HK, Liebow IM. Factors influencing the T wave of the electrocardiogram; an experimental study employing intracavitary and extraventricular (epicardial) leads; effects of heating and cooling the endocardium and the epicardium. Am Heart J. 1950;39:35-55. 31. Burnes JE, Ghanem RN, Waldo AL, Rudy Y. Imaging dispersion of myocardial repolarization, I: comparison of body- surface and epicardial measures. Circulation. 2001;104:1299-1305. 32. van Dam RT, Durrer D. Experimental study on the intramural distribution of the excitability cycle and on the form of the epicardial T wave in the dog heart in situ. Am Heart J. 1961;61:537-542. 33. Burgess MJ, Green LS, Millar K, Wyatt R, Abildskov JA. The sequence of normal ventricular recovery. Am Heart J. 1972;84:660-669. 34. Abildskov JA. The sequence of normal recovery of excitability in the dog heart. Circulation. 1975;52:442-446. 35. Spach MS, Barr RC. Ventricular intramural and epicardial potential distributions

35

Chapter 1

36. 37. 38. 39. 40. 41.

42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

36

during ventricular activation and repolarization in the intact dog. Circ Res. 1975;37:243-257. Durrer D, van Dam RT, Freud GE, Janse MJ, Meijler FL, Arzbaecher RC. Total excitation of the isolated human heart. Circulation. 1970;41:899-912. El Sherif N, Caref EB, Yin H, Restivo M. The electrophysiological mechanism of ventricular arrhythmias in the long QT syndrome. Tridimensional mapping of activation and recovery patterns. Circ Res. 1996;79:474-492. Janse, M. J. Thesis: The effect of changes in heart rate on the refractory period of the heart. 1971. Amsterdam, Amsterdam University. Wassenaar, C. Thesis: Comparative electrocardiography in mammals. 1993. Utrecht University. Spach MS, Barr RC, Lanning CF, Tucek PC. Origin of body surface QRS and T wave potentials from epicardial potential distributions in the intact chimpanzee. Circulation. 1977;55:268. Meijler FL, Wittkampf FH, Brennen KR, Baker V, Wassenaar C, Bakken EE. Electrocardiogram of the humpback whale (Megaptera novaeangliae), with specific reference to atrioventricular transmission and ventricular excitation. J Am Coll Cardiol. 1992;20:475-479. Taggart P, Sutton PM, Opthof T, Coronel R, Trimlett R, Pugsley W, Kallis P. Transmural repolarisation in the left ventricle in humans during normoxia and ischaemia. Cardiovasc Res. 2001;50:454-462. Yan GX, Antzelevitch C. Cellular basis for the normal T wave and the electrocardiographic manifestations of the long-QT syndrome. Circulation. 1998;98:1928-1936. Drouin E, Charpentier F, Gauthier C, Laurent K, Le Marec H. Electrophysiologic characteristics of cells spanning the left ventricular wall of human heart: evidence for presence of M cells. J Am Coll Cardiol. 1995;26:185-192. Anyukhovsky EP,Sosunov EA,Rosen MR.Regional differences in electrophysiological properties of epicardium, midmyocardium, and endocardium. In vitro and in vivo correlations. Circulation. 1996;94:1981-1988. Conrath CE, Wilders R, Coronel R, de Bakker JM, Taggart P, de Groot JR, Opthof T. Intercellular coupling through gap junctions masks M cells in the human heart. Cardiovasc Res. 2004;62:407-414. Poelzing S, Rosenbaum DS. Altered connexin43 expression produces arrhythmia substrate in heart failure. Am J Physiol Heart Circ Physiol. 2004;287:H1762H1770. El Sherif N, Caref EB, Chinushi M, Restivo M. Mechanism of arrhythmogenicity of the short-long cardiac sequence that precedes ventricular tachyarrhythmias in the long QT syndrome. J Am Coll Cardiol. 1999;33:1415-1423. Restivo M, Caref EB, Kozhevnikov DO, El Sherif N. Spatial dispersion of repolarization is a key factor in the arrhythmogenicity of long QT syndrome. J Cardiovasc Electrophysiol. 2004;15:323-331. Gepstein L, Hayam G, Ben Haim SA. Activation-repolarization coupling in the normal swine endocardium. Circulation. 1997;96:4036-4043. Yuan S, Kongstad O, Hertervig E, Holm M, Grins E, Olsson B. Global repolarization sequence of the ventricular endocardium: monophasic action potential mapping in

Introduction: Electrocardiographic assessment of repolarization heterogeneity

52. 53. 54. 55.

56. 57. 58. 59. 60. 61. 62. 63.

64. 65. 66.

67. 68.

swine and humans. Pacing Clin Electrophysiol. 2001;24:1479-1488. The normal vectorcardiogram. In: Clinical vectorcardiography. Chou T, Helm RA, Kaplan S, eds. 1974. Grune and Stratton Inc., New York. Abildskov JA, Millar K, Burgess MJ, Green L. Characteristics of ventricular recovery of defined by the vectorcardiographic T loop. Am J Cardiol. 1971;28:670-674. Burger HC, van Herpen G. Clinical application of vectorcardiography. In: heart and vector. Julius HW, ed. 1968. N.V. Philips’ Gloeilampenfabriek, Eindhoven. Conrath, C. E., Wilde, A. A., van Tintelen, J. P., van Langen, I. M., Jongbloed, R. J., Alders, M., Doevendans, P. A., Hauer, R. N., and Opthof, T. Predictors for efficacy of â-adrenoreceptor blockade in symptomatic LQTS1 and LQTS2 patients. 129-151 of Thesis 2005. Utrecht University. Dilaveris PE. Molecular predictors of drug-induced prolongation of the QT interval. Curr Med Chem Cardiovasc Hematol Agents. 2005;3:105-118. Sicouri S, Moro S, Litovsky S, Elizari MV, Antzelevitch C. Chronic amiodarone reduces transmural dispersion of repolarization in the canine heart. J Cardiovasc Electrophysiol. 1997;8:1269-1279. Bazett, H. C. An analysis of the time relation of electrocardiograms. Heart 1920;7:353-367. Lecocq B, Lecocq V, Jaillon P. Physiologic relation between cardiac cycle and QT duration in healthy volunteers. Am J Cardiol. 1989;64:481-486. Lepeschkin E, Surawicz B. The measurement of the Q-T interval of the electrocardiogram. Circulation. 1952;6:378-388. Ritsema van Eck HJ, Kors JA, van Herpen G. The U wave in the electrocardiogram: a solution for a 100-year-old riddle. Cardiovasc Res. 2005;67:256-262. Conrath CE, Opthof T. The patient U wave. Cardiovasc Res. 2005;67:184-186. Medina-Ravell VA, Lankipalli RS, Yan GX, Antzelevitch C, Medina-Malpica NA, Medina-Malpica OA, Droogan C, Kowey PR. Effect of epicardial or biventricular pacing to prolong QT interval and increase transmural dispersion of repolarization: does resynchronization therapy pose a risk for patients predisposed to long QT or torsade de pointes? Circulation. 2003;107:740-746. Yan GX, Shimizu W, Antzelevitch C. Characteristics and distribution of M cells in arterially perfused canine left ventricular wedge preparations. Circulation. 1998;98:1921-1927. Akar FG, Yan GX, Antzelevitch C, Rosenbaum DS. Unique topographical distribution of M cells underlies reentrant mechanism of torsade de pointes in the long-QT syndrome. Circulation. 2002;105:1247-1253. Shimizu W, Antzelevitch C. Cellular basis for the ECG features of the LQT1 form of the long-QT syndrome: effects of beta-adrenergic agonists and antagonists and sodium channel blockers on transmural dispersion of repolarization and torsade de pointes. Circulation. 1998;98:2314-2322. Shimizu W, Antzelevitch C. Sodium channel block with mexiletine is effective in reducing dispersion of repolarization and preventing torsade des pointes in LQT2 and LQT3 models of the long-QT syndrome. Circulation. 1997;96:2038-2047. Lubinski A, Kornacewicz-Jach Z, Wnuk-Wojnar AM, Adamus J, Kempa M, Krolak T, Lewicka-Nowak E, Radomski M, Swiatecka G. The terminal portion of the T wave: a new electrocardiographic marker of risk of ventricular arrhythmias. Pacing

37

Chapter 1 Clin Electrophysiol. 2000;23:1957-1959. 69. Viitasalo M, Oikarinen L, Swan H, Vaananen H, Glatter K, Laitinen PJ, Kontula K, Barron HV, Toivonen L, Scheinman MM. Ambulatory electrocardiographic evidence of transmural dispersion of repolarization in patients with long-QT syndrome type 1 and 2. Circulation. 2002;106:2473-2478. 70. Yamaguchi M, Shimizu M, Ino H, Terai H, Uchiyama K, Oe K, Mabuchi T, Konno T, Kaneda T, Mabuchi H. T wave peak-to-end interval and QT dispersion in acquired long QT syndrome: a new index for arrhythmogenicity. Clin Sci (Lond). 2003;105:671-676. 71. Autenrieth G, Surawicz B, Kuo CS. Sequence of repolarization on the ventricular surface in the dog. Am Heart J. 1975;89:463-469. 72. Xia Y, Liang Y, Kongstad O, Liao Q, Holm M, Olsson B, Yuan S. In vivo validation of the coincidence of the peak and end of the T wave with full repolarization of the epicardium and endocardium in swine. Heart Rhythm. 2005;2:162-169. 73. Hooft van Huysduynen, B and et al. Dispersion of the repolarization in cardiac resynchronization therapy. Heart Rhythm. 2005;2:1286-93 74. Day CP, McComb JM, Campbell RW. QT dispersion: an indication of arrhythmia risk in patients with long QT intervals. Br Heart J. 1990;63:342-344. 75. Buja G, Miorelli M, Turrini P, Melacini P, Nava A. Comparison of QT dispersion in hypertrophic cardiomyopathy between patients with and without ventricular arrhythmias and sudden death. Am J Cardiol. 1993;72:973-976. 76. Kors JA, van Herpen G, van Bemmel JH. QT dispersion as an attribute of T-loop morphology. Circulation. 1999;99:1458-1463. 77. Malik M, Batchvarov VN. Measurement, interpretation and clinical potential of QT dispersion. J Am Coll Cardiol. 2000;36:1749-1766. 78. Kautzner J, Yi G, Camm AJ, Malik M. Short- and long-term reproducibility of QT, QTc, and QT dispersion measurement in healthy subjects. Pacing Clin Electrophysiol. 1994;17:928-937. 79. Glancy JM, Weston PJ, Bhullar HK, Garratt CJ, Woods KL, de Bono DP. Reproducibility and automatic measurement of QT dispersion. Eur Heart J. 1996;17:1035-1039. 80. Coumel P, Maison-Blanche P, Badilini F. Dispersion of ventricular repolarization: reality? Illusion? Significance? Circulation. 1998;97:2491-2493. 81. di Bernardo D, Langley P, Murray A. Dispersion of QT intervals: a measure of dispersion of repolarization or simply a projection effect? Pacing Clin Electrophysiol. 2000;23:1392-1396. 82. Lux RL, Fuller MS, MacLeod RS, Ershler PR, Green LS, Taccardi B. QT interval dispersion: dispersion of ventricular repolarization or dispersion of QT interval? J Electrocardiol. 1998;30 Suppl:176-180. 83. Zaputovic L, Mavric Z, Zaninovic-Jurjevic T, Matana A, Bradic N. Relationship between QT dispersion and the incidence of early ventricular arrhythmias in patients with acute myocardial infarction. Int J Cardiol. 1997;62:211-216. 84. Saadeh A, Evans S, James M, Jones J. QTc dispersion and complex ventricular arrhythmias in untreated newly presenting hypertensive patients. J Hum Hypertens. 1999;13:665-669. 85. Hondeghem LM, Carlsson L, Duker G. Instability and triangulation of the action

38

Introduction: Electrocardiographic assessment of repolarization heterogeneity

86. 87.

88. 89. 90.

91. 92.

93. 94. 95. 96. 97. 98. 99. 100.

potential predict serious proarrhythmia, but action potential duration prolongation is antiarrhythmic. Circulation. 2001;103:2004-2013. van Oosterom A. Genesis of the T wave as based on an equivalent surface source model. J Electrocardiol. 2001;34 Suppl:217-227. Moss AJ, Robinson JL, Gessman L, Gillespie R, Zareba W, Schwartz PJ, Vincent GM, Benhorin J, Heilbron EL, Towbin JA, Priori SG, Napolitano C, Zhang L, Medina A, Andrews ML, Timothy K. Comparison of clinical and genetic variables of cardiac events associated with loud noise versus swimming among subjects with the long QT syndrome. Am J Cardiol. 1999;84:876-879. Wilde AA, Roden DM. Predicting the long-QT genotype from clinical data: from sense to science. Circulation. 2000;102:2796-2798. Wan X, Bryant SM, Hart G. The effects of [K+]o on regional differences in electrical characteristics of ventricular myocytes in guinea-pig. Exp Physiol. 2000;85:769774. Kleber AG, Janse MJ, van Capelle FJ, Durrer D. Mechanism and time course of S-T and T-Q segment changes during acute regional myocardial ischemia in the pig heart determined by extracellular and intracellular recordings. Circ Res. 1978;42:603-613. Shah RR, Hondeghem LM. Refining detection of drug-induced proarrhythmia: QT interval and TRIaD. Heart Rhythm. 2005;2:758-772. Zhang L, Timothy KW, Vincent GM, Lehmann MH, Fox J, Giuli LC, Shen J, Splawski I, Priori SG, Compton SJ, Yanowitz F, Benhorin J, Moss AJ, Schwartz PJ, Robinson JL, Wang Q, Zareba W, Keating MT, Towbin JA, Napolitano C, Medina A. Spectrum of ST-T-wave patterns and repolarization parameters in congenital long-QT syndrome: ECG findings identify genotypes. Circulation. 2000;102:28492855. Abildskov JA, Klein RM. Cancellation of electrocardiographic effects during ventricular excitation. Sogo Rinsho. 1962;11:247-251. Burgess MJ, Millar K, Abildskov JA. Cancellation of electrocardiographic effects during ventricular recovery. J Electrocardiol. 1969;2:101-107. Zabel M, Portnoy S, Franz MR. Electrocardiographic indexes of dispersion of ventricular repolarization: an isolated heart validation study. J Am Coll Cardiol. 1995;25:746-752. van Opstal JM, Verduyn SC, Winckels SK, Leerssen HM, Leunissen JD, Wellens HJ, Vos MA. The JT-area indicates dispersion of repolarization in dogs with atrioventricular block. J Interv Card Electrophysiol. 2002;6:113-120. Abildskov JA, Green LS, Evans AK, Lux RL. The QRST deflection area of electrograms during global alterations of ventricular repolarization. J Electrocardiol. 1982;15:103-107. Kubota I, Lux RL, Burgess MJ, Abildskov JA. Relation of cardiac surface QRST distributions to ventricular fibrillation threshold in dogs. Circulation. 1988;78:171177. Shimizu W, Antzelevitch C. Sodium channel block with mexiletine is effective in reducing dispersion of repolarization and preventing torsade des pointes in LQT2 and LQT3 models of the long-QT syndrome. Circulation. 1997;96:2038-2047. di Bernardo D, Murray A. Explaining the T-wave shape in the ECG. Nature.

39

Chapter 1 2000;403:40. 101. Hooft van Huysduynen, B., Swenne, C. A., Draisma, H. H. M., Antoni, M. L., Van de Vooren, H., Van der Wall, E. E., and Schalij, M. J. Validation of ECG Indices of Ventricular Repolarization Heterogeneity: A Computer Simulation Study. J Cardiovasc Electrophysiol. 2005;16:1097-103. 102. Dower GE, Machado HB, Osborne JA. On deriving the electrocardiogram from vectoradiographic leads. Clin Cardiol. 1980;3:87-95. 103. Edenbrandt L, Pahlm O. Vectorcardiogram synthesized from a 12-lead ECG: superiority of the inverse Dower matrix. J Electrocardiol. 1988;21:361-367. 104. Kardys I, Kors JA, van dM, I, Hofman A, van der Kuip DA, Witteman JC. Spatial QRS-T angle predicts cardiac death in a general population. Eur Heart J. 2003;24:1357-1364. 105. Zabel M, Acar B, Klingenheben T, Franz MR, Hohnloser SH, Malik M. Analysis of 12-lead T-wave morphology for risk stratification after myocardial infarction. Circulation. 2000;102:1252-1257. 106. de Torbal A, Kors JA, van Herpen G, Meij S, Nelwan S, Simoons ML, Boersma E. The electrical T-axis and the spatial QRS-T angle are independent predictors of long-term mortality in patients admitted with acute ischemic chest pain. Cardiology. 2004;101:199-207. 107. Kors JA, de Bruyne MC, Hoes AW, van Herpen G, Hofman A, van Bemmel JH, Grobbee DE. T axis as an indicator of risk of cardiac events in elderly people. Lancet. 1998;352:601-605. 108. Rautaharju PM, Nelson JC, Kronmal RA, Zhang ZM, Robbins J, Gottdiener JS, Furberg CD, Manolio T, Fried L. Usefulness of T-axis deviation as an independent risk indicator for incident cardiac events in older men and women free from coronary heart disease (the Cardiovascular Health Study). Am J Cardiol. 2001;88:118-123. 109. Draper HW, Peffer CJ, Stallmann FW, Littmann D, Pipberger HV. The corrected orthogonal electrocardiogram and vectorcardiogram in 510 normal men (Frank lead system) Circulation. 1964;30:853-864. 110. Priori SG, Mortara DW, Napolitano C, Diehl L, Paganini V, Cantu F, Cantu G, Schwartz PJ. Evaluation of the spatial aspects of T-wave complexity in the long-QT syndrome. Circulation. 1997;96:3006-3012. 111. Acar B, Yi G, Hnatkova K, Malik M. Spatial, temporal and wavefront direction characteristics of 12-lead T-wave morphology. Med Biol Eng Comput. 1999;37:574584. 112. Okin PM, Malik M, Hnatkova K, Lee ET, Galloway JM, Best LG, Howard BV, Devereux RB. Repolarization abnormality for prediction of all-cause and cardiovascular mortality in American Indians: the Strong Heart Study. J Cardiovasc Electrophysiol. 2005;16:945-951. 113. Zabel M, Malik M, Hnatkova K, Papademetriou V, Pittaras A, Fletcher RD, Franz MR. Analysis of T-wave morphology from the 12-lead electrocardiogram for prediction of long-term prognosis in male US veterans. Circulation. 2002;105:10661070. 114. De Ambroggi L, Aime E, Ceriotti C, Rovida M, Negroni S. Mapping of ventricular repolarization potentials in patients with arrhythmogenic right ventricular dysplasia: principal component analysis of the ST-T waves. Circulation.

40

Introduction: Electrocardiographic assessment of repolarization heterogeneity 1997;96:4314-4318. 115. van Oosterom A. Singular value decomposition of the T wave: its link with a biophysical model of repolarization. Int J Bioelectromagnetism. 2002;4:59-60. 116. Rosenbaum MB, Blanco HH, Elizari MV, Lazzari JO, Davidenko JM. Electrotonic modulation of the T wave and cardiac memory. Am J Cardiol. 1982;50:213-222. 117. Osaka T, Kodama I, Tsuboi N, Toyama J, Yamada K. Effects of activation sequence and anisotropic cellular geometry on the repolarization phase of action potential of dog ventricular muscles. Circulation. 1987;76:226-236.

41

42

Chapter 2 Validation of ECG indices of ventricular repolarization heterogeneity A computer simulation study

Bart Hooft van Huysduynen Cees A. Swenne Harmen H.M. Draisma M. Louisa Antoni Hedde van de Vooren, Ernst E. van der Wall Martin J. Schalij

J Cardiovasc Electrophysiol 2005; 16: 1097-103

43

Chapter 2

ABSTRACT Introduction. Repolarization heterogeneity is functionally linked to dispersion in refractoriness and to arrhythmogeneity. In the current study we validate several proposed ECG indices for repolarization heterogeneity: T-wave amplitude, -area, -complexity and -symmetry ratio, QT dispersion, and the Tapex-end interval (the latter being an index of transmural dispersion of the repolarization). Methods and results. We used ECGSIM, a mathematical simulation model of ECG genesis in a human thorax, and varied global repolarization heterogeneity by increasing the standard deviation (SD) of the repolarization instants from 20 (default) to 70 ms in steps of 10 ms. T-wave amplitude, -area, -symmetry and Tapex-end depended linearly on SD. T-wave amplitude increased from 275 ± 173 to 881 ± 456 µV, T-wave area from 34·103 ± 21·103 to 141·103 ± 58·103 µV·ms, T-wave symmetry decreased from 1.55 ± 0.11 to 1.06 ± 0.23 and Tapex-end increased from 84 ± 17 to 171 ± 52 ms. T-wave complexity increased initially but saturated at SD = 50 ms. QT dispersion increased modestly until SD = 40 ms and more rapidly for higher values of SD. Transmural dispersion of the repolarization increased linearly with SD. Tapex-end increased linearly with transmural dispersion of the repolarization, but overestimated it. Conclusion. T-wave complexity did not discriminate between differences in larger repolarization heterogeneity values. QT dispersion had low sensitivity in the transitional zone between normal and abnormal repolarization heterogeneity. In conclusion, T-wave amplitude, -area, -symmetry, and, with some limitations, Tapex-end and T-wave complexity reliably reflect changes in repolarization heterogeneity.

44

INTRODUCTION Cardiac repolarization is more spread in time than cardiac depolarization because of regional differences in action potential duration (APD). Functionally, repolarization heterogeneity (RH) is closely related to dispersion in refractoriness, which in turn increases the vulnerability to reentrant arrhythmias.1;2 RH has been described between the apex and basal areas of the heart,3 between the left and right ventricles4 and between the epicardium, mid-myocardium and endocardium.5 The latter type of RH has been named transmural dispersion of the repolarization (TDR). Primary electrical disease as well as several drugs are known to exaggerate action potential differences in the heart, thus increasing RH and arrhythmia risk. A noninvasive, electrocardiographic index of RH would therefore be of great clinical value.6 Several ECG indices to assess RH have been proposed, like the amplitude of the T-wave7 (Tamplitude), T-wave surface area8 (Tarea), symmetry ratio of the T wave9 (Tsymmetry), complexity of the T wave calculated by singular value decomposition10 (Tcomplexity) and QT dispersion.11 The Tapex-end interval in the left precordial leads (Tapex-end) has been put forward as a measure that directly assesses the duration of TDR.12 Although clinical studies have shown that most of these ECG parameters have prognostic power,13-15 it remains unknown if they really assess RH. Whole heart studies in which endo- and epicardial repolarization as well as surface ECGs are recorded are scarce. Most whole-heart biological models of ECG genesis only measure either epicardial or endocardial dispersion,16,17 thus completely ignoring TDR. Tapex-end, currently the only index that assesses TDR duration, has been validated on the basis of a quasi-ECG obtained from transmural recordings of a wedge preparation of the left ventricular wall,12 but not in a whole heart and torso model. In the current study, we sought to validate the above mentioned electrocardiographic indices of RH by using a mathematical simulation model of ECG genesis of a human heart in a thorax.18 Within this model, we increased the heterogeneity of the action potential durations throughout the heart and measured the consequences of these manipulations for TDR and RH, and for the ECG indices of ventricular repolarization heterogeneity.

45

Chapter 2

METHODS ECGSIM ECGSIM, an interactive computer program conceived and realized by Van Oosterom and Oostendorp,18 is a mathematical model for studying QRST waveform genesis. The model is available in the public domain at www.ecgsim.org. It is a combination of a source model of the heart and a volume conductor model of the torso. The heart is a double layer model in which all electrical activity is represented by equivalent sources on the surface encompassing the ventricular myocardium. This surface, of which the shape has been derived from magnetic resonance imaging data, consists of 257 nodes. Each node has its own electrophysiological properties in the form of a transmembrane action potential, in which the timing of the depolarization, the timing of the maximal negative slope and the magnitude of the transmembrane potential can be changed. The default action potential (AP) onset sequence represents normal conduction of the impulse, and APD differences throughout the heart represent the natural apex-to-base and endo-to-epicardial APD heterogeneity. The heart is placed in a realistic thorax model based on magnetic resonance imaging data and includes conductance inhomogeneities like the lungs. With default parameter settings, ECGSIM generates potentials on the thoracic surface that closely resemble those of a healthy subject. ECGSIM allows for simulations of pathological conditions such as abnormal activation sequences or, by adjusting the magnitude of the transmembrane potential, for simulations of acute ischemia. For the current study, the default, normal settings for activation sequence and source strengths were kept throughout the simulations. Manipulation of repolarization heterogeneity We adopted the standard deviation of the repolarization times of all 257 nodes (SDrep) as a measure of RH (repolarization times in the model are defined as the moments of maximum downslope of the transmembrane action potential). Different levels of SDrep were obtained by increasing the standard deviation of the action potential durations (SDAPD). By setting the model parameter SDAPD at different levels, all 257 APDs are modified without changing the mean APD. Because each node’s repolarization time is calculated by adding its APD to its activation time, an increase of SDAPD increases SDrep as well. In the current version of ECGSIM the model parameter SDAPD can assume a number of discrete values.18 SDAPD values were set 46

Validation of ECG indices of ventricular repolarization heterogeneity

in such a way that SDrep increased in steps of approximately 10 ms. Thus, SDrep was increased from the default value of 20.8 ms to 30.7, 40.6, 51.1, 60.6, and finally to a maximum of 70.7 ms. Sinus rhythm at a rate of 60 beats per minute was maintained during all simulations. As a control experiment, we evaluated the effects of homogeneous APD lengthening on the ECG repolarization indices in the measurable leads. The APDs of all nodes were lengthened to the same extent by increasing the mean APD from 245.1 ms at baseline to 319.0 ms and 395.0 ms, respectively. Calculation of transmural dispersion of the repolarization In the 257 node model, 42 / 61 nodes constitute the endo / epicardium of the free left ventricular wall, 54 / 70 nodes the endo / epicardium of the right ventricular wall and 14 / 16 nodes the left / right septum. Most endocardial nodes were paired with one opposing epicardial node, and TDR was calculated as differences in repolarization time between the endo- and epicardial node. However, as there are more epicardial than endocardial nodes, some endocardial nodes had two opposing epicardial nodes. In these cases, the difference was calculated between the repolarization time of the endocardial node and the mean repolarization time of the two opposing epicardial nodes. Similarly, each left ventricular septal node was paired with one or two right ventricular septal nodes. Subsequently, all paired repolarization time differences were averaged to calculate TDR. ECG analysis The simulated ECGs were analyzed by LEADS (Leiden ECG Analysis and Decomposition Software), our MATLAB (The MathWorks, Natick, USA) program for research oriented ECG analysis. LEADS identifies the apex and end of the T wave in each ECG lead. The end of the T wave was defined as the point where the tangent to the steepest portion of the terminal part of the T wave crosses the isoelectric line. Thereafter, the low amplitude T waves in lead V1 were excluded because these led to erroneous detection of the apex and end of the T wave. LEADS calculated Tapex-end, Tamplitude and Tarea in every measurable lead and in the vector magnitude signal computed from the vectorcardiogram constructed

47

Chapter 2

by using the inverse Dower matrix.19,20 Tsymmetry was calculated in all measurable leads as the ratio of the early T-wave area, from the J point to the apex of the T wave, to the late T wave area, from the T wave apex to the end of the T wave.9 Finally, the values of Tapex-end, Tamplitude, Tarea, Tsymmetry in all measurable leads were averaged. Calculation of repolarization complexity was performed by means of singular value decomposition (SVD) of the T-wave.10;21 SVD was computed over an interval starting 50 ms after the J point until 50 ms after the end of the T-wave. We computed the SVD-based Tcomplexity as the square root of the summed second to eighth singular values divided by the first singular value. QT dispersion was calculated as the longest minus the shortest measurable QT interval in any of the 12 standard ECG leads.

48

Validation of ECG indices of ventricular repolarization heterogeneity

RESULTS Two example ECGs generated with the default, low level of RH (SDrep = 20.8 ms) and with a high level of RH (SDrep = 70.7 ms) are depicted in Figure 1, panels A and B, respectively. Figure 1a. Twelve-lead ECG as generated by ECGSIM with default, low repolarization heterogeneity (standard deviation of the repolarization times = 20.8 ms). As ECGSIM models the ventricular electrical depolarization/repolarization, no P waves are present. The signals were baseline corrected in ESGSIM and one complex is given for each lead.

Figure 1b. Twelve-lead ECG generated by ECGSIM with high repolarization heterogeneity (standard deviation of the repolarization times = 70.7 ms).

49

Chapter 2

SDrep increased linearly with the SDAPD, in an almost 1:1 relationship: SDrep = 0.97 · SDAPD - 7.0 (r2 = 0.99). The average absolute TDR in the left ventricle, septum, right ventricle and whole heart all related linearly to SDrep (Figure 2). The slope of TDR of the right ventricle (0.72) was smaller than the slope of the left ventricle (1.04) and the septum (0.95), while the whole heart slope assumed an intermediate value (0.90). All these relationships had a correlation coefficient of 0.99.

Figure 2. Relation between the standard deviation of the repolarization times (SDrep) and the averaged absolute transmural dispersion of the repolarization (TDR) in the right ventricle, left ventricle, septum and whole heart.

Linear regression plots of Tapex-end in the left precordial leads as a function of TDR in the free wall of the left ventricle are depicted in Figure 3. The relationship of the Tapex-end in leads V4-6 with RH were all linear with correlations ≥ 0.99. Tapexend in leads V2 and V3 showed a discontinuity when T waves became biphasic at increasing RH levels; in lead V2 at an SDrep of 30 ms and in lead V3 at an SDrep of 50 ms. Tapex-end overestimated TDR in all left precordial leads.

50

Validation of ECG indices of ventricular repolarization heterogeneity

Figure 3. Relation between the transmural dispersion of the repolarization (TDR) in the left ventricular free wall and the Tapex-end in the left precordial leads. Discontinuities in the V2 and V3 data are caused by a transition from a monophasic to a biphasic T wave. (dashed line = line of identity)

Tapex-end, Tamplitude, Tarea and Tsymmetry The relationships of Tapex-end, Tamplitude, Tarea and Tsymmetry with RH were close to perfectly linear (Tables 1a and b). Table 1a lists values that were averaged over the ECG leads in which the respective RH index could be measured (lead V1 had to be excluded because of a low amplitude T wave). The RH indices were also measured in the vector magnitude signal (Table 1b). All linear regressions had correlations of at least 0.98. Regressions with the vector magnitude derived indices had generally steeper slopes than those with the average of the ECG leads. All RH indices but Tsymmetry had a positive relation with RH.

51

Chapter 2

Averaged over ECG leads

slope

intercept

r2

Value at baseline RH SDrep = 20.8

Value at maximal RH SDrep = 70.7

Tapex-end (ms)

1.7

48

0.99

84 ± 17

171 ± 52

Tamplitude (µV)

12.1

48

0.99

275 ± 173

881 ± 456

2.1·103

-11·103

0.99

34·103 ± 21·103

141·103 ± 58·103

-0.01

1.72

0.98

1.55 ± 0.11

1.06 ± 0.23

Tarea (µV·ms) Tsymmetry

Table 1a. Slope, intercept and correlation (r2) of the linear regressions of Tapex-end, Tamplitude, Tarea and Tsymmetry on repolarization heterogeneity (RH) and the value of these RH indices at minimal and maximal RH.

In vector magnitude signal

slope

intercept

r2

Value at baseline RH SDrep = 20.8

Value at maximal RH SDrep = 70.7

Tapex-end (ms)

2.7

60

0.99

116

252

Tamplitude (µV)

16.7

86

0.99

403

1253

Tarea (µV·ms)

3.4·103

-19·103

0.99

52·103

221·103

Tsymmetry

-0.012

1.60

0.99

1.39

0.78

Table 1b. Slope, intercept and r2 of the linear relations of Tapex-end, Tamplitude, Tarea and Tsymmetry with RH and the value of these indices at minimal and maximal values of RH.

52

Validation of ECG indices of ventricular repolarization heterogeneity

Tcomplexity Tcomplexity increased initially and saturated at about SDrep ≥ 40 ms, see Figure 4. Tcomplexity increased from 0.13 at default to 0.18 at an SDrep of 40 ms and remained 0.19 while SDrep increased from 50 to 70 ms.

Figure 4. Relation between the standard deviation of the repolarization times (SDrep) and the T-wave complexity as expressed by the ratio of the 2nd-8th components to the first component in the vectorcardiogram. For visual support we fitted a line through the six data points by a 5th order polynomial. The figure shows that the sensitivity to changes in repolarization heterogeneity decreases at higher heterogeneity values.

QT dispersion QT dispersion increased modestly at low and intermediate levels of RH and more rapidly at higher levels of RH, see Figure 5. QT dispersion increased from 58 ms at default to 68 and 85 ms at SDrep = 30 and SDrep = 40 ms, respectively. Thereafter, QT dispersion increased more rapidly to 102, 137 and 171 ms. The slope of the regression line for SDrep ≤ 40 ms was 1.4 and the slope of the regression line for SDrep ≥ 50 ms was 3.5. 53

Chapter 2

Figure 5. Relation between the standard deviation of the repolarization times (SDrep) and QT dispersion. QT dispersion is sensitive to changes in the repolarization heterogeneity in the very pathological zone (slope 3.5), but has a lower sensitivity in the transitional zone between normal and abnormal (slope 1.4). Therefore this parameter is not well suited to discriminate between normal and abnormal repolarization heterogeneity.

Homogeneous APD lengthening Homogeneous APD lengthening resulted in no apparent morphological T-wave changes, only the ST segment lengthened. Changes in Tapex-end, Tamplitude, Tarea and QT dispersion due to homogeneous APD lengthening were less than 3 % of the changes due to SDrep increase. Only Tsymmetry and T complexity decreased considerably, 19 % and 47 % of the changes due SDrep increase, respectively. These changes were in opposite direction to the changes induced by increasing SDrep.

54

DISCUSSION In this modeling study, we evaluated the effects of repolarization heterogeneity on electrocardiographic indices proposed to assess repolarization heterogeneity. The observed changes in the ECG indices were not caused by APD lengthening per se, as homogeneous APD lengthening caused only minor changes in most ECG indices, and counteracted rather than contributed to the changes in the other ECG indices due to RH increase. Tapex-Tend interval Tapex-end is the only measure that potentially estimates the time window during which repolarization is heterogeneous. Tapex-end is believed to represent TDR, the interval from the end of the epicardial APDs to the end of the (sub)endocardial APDs.12 In left ventricular wedge preparations, the apex of the T-wave concurs with the end of the epicardial AP because the end of the epicardial AP is very steep.12,22,23 This steep descent of the epicardial AP caused the largest differences in amplitude with the (sub)endocardial AP. However, when other AP morphologies are present, a different phase of the AP could coincide with the apex of the T wave. For example, a steep descent at the start of phase 3 with a more slowly diminishing tail at the end of the AP could cause the apex of the T wave to coincide with the start of phase 3 of the epicardial AP. In that case, Tapex-end would overestimate TDR. In our study, overall Tapex-end had a good linear relation with RH (Tables 1a and b). However, Tapex-end in the left precordial leads overestimated TDR by several tens of milliseconds, this bias having the same order of magnitude as TDR itself (Figure 3). The abrupt increase in Tapex-end in leads V2 and V3 when biphasic T-waves evolve due to increased RH illustrates an additional problem: the difficult localization of the end of the T wave. The 1:1 relation between the duration of Tapex-end and TDR as found in the wedge preparation did not hold in a three dimensional structure as ECGSIM. The differences between the transmural quasi-ECG in the wedge preparation and the surface ECG may be explained by different AP morphologies and also by the fact that the wedge is but part of a heart. Therefore, the floating endocardial electrode in the wedge

55

Chapter 2

preparation is not the analogue of the Wilson central terminal in electrocardiography. Moreover, in the regular electrocardiogram other structures in the heart than the left ventricular free wall additionally contribute to the cardiac vector. Amongst others, this causes a considerable amount of cancellation, a phenomenon not occurring in the wedge preparation. T-wave complexity by singular value decomposition Singular value decomposition is a method to quantify the complexity of the repolarization. A smooth simple T wave is usually associated with a normal repolarization process, while a notched, irregular morphology is seen with disturbed repolarization processes.24 SVD-calculated complexity is higher in Long-QT patients and can be used to distinguish these patients from healthy subjects.10 In patients with arrhythmogenic right ventricular dysplasia, higher repolarization complexity, measured in body surface maps as a decreased contribution of the first, most simple SVD component, was associated with arrhythmias.25 In U.S. veterans with cardiovascular disease, repolarization complexity calculated with SVD conferred independent prognostic information.15 Van Oosterom mathematically proved that a higher RH leads to increased Tcomplexity.26 Our results in ECGSIM suggest that T-wave complexity reacts to small increases in RH, but fails to increase further with higher levels of RH. SVD may therefore be useful to detect increases of RH, but is unlikely to discriminate between smaller and larger RH values. QT dispersion QT dispersion, defined as the longest minus the shortest QT interval in any of the 12 ECG leads, was initially believed to represent local repolarization differences.11 According to current insight this concept is incorrect; QT-interval differences in ECG leads depends on projections of the (global) heart vector on the different lead vectors and can therefore not represent local repolarization differences.27 Another problem of manual measurement of QT dispersion is the low reproducibility, for example due to the subjectivity involved in exclusion of low-amplitude T waves and the difficult measurement of the end of T waves in noisy ECGs. Nevertheless, QT dispersion may have a weak relation with repolarization disturbances28 and was associated with arrhythmias in some studies.29;30 We measured QT dispersion values up to 171 ms. Although most previous studies 56

Validation of ECG indices of ventricular repolarization heterogeneity

reported smaller values, similar values have been reported in patients shortly before an episode of Torsade de Pointes31 and in Long QT patients who remained symptomatic despite beta-blocker therapy.32 We measured the QT dispersion value of 171 ms at the maximum simulated RH (SDrep = 70 ms), a situation that is likely to be highly arrhythmogenic in reality. In our simulations, QT dispersion increased relatively little with initial RH increases, but it increased more at higher levels of SDrep. QT dispersion is sensitive for changes of RH in the very pathological zone, but has little sensitivity in the transitional zone between normal and abnormal. Therefore, this parameter is not well suited to discriminate between normal and abnormal. The above mentioned theoretical and practical objections in combination with the insensitivity for small increases in RH render QT dispersion unsuitable as an index of RH. T-wave amplitude Tamplitude reflects the net maximal voltage gradients in the whole heart after cancellation. The repolarization gradients in ECGSIM are caused by the voltage difference of opposing endo- and epicardial APs. By increasing RH the already long endocardial APDs were further lengthened and the already short epicardial APDs were further shortened, mainly achieved by a change in the duration of the plateau phase. This caused the endo- and epicardial APs to shift further out of phase such that endocardial APs still had a high amplitude and are less opposed by the already diminished epicardial AP amplitude. Our simulated T waves mimic the high amplitude T waves found in long-QT syndrome type 1.33 In wedge preparations mimicking the long-QT 1 syndrome, application of an IKs current blocker in combination with isoprotenerol caused a relatively longer APD of the mid-myocardium (subendocardium) compared to the epicardial APD. This caused an increased voltage gradient directed toward the epicardium and therefore, high amplitude T-waves on the transmural quasi ECG.23 The mathematical background of the dependence of Tamplitude on RH in the equivalent surface source model was worked out by Van Oosterom.34 For inter-individual comparison, the main disadvantage of the Tamplitude is its individual variability, due to differences in thorax and heart size. For example, athletes will have a higher Tamplitude35 mostly on the basis of a higher cardiac mass and not necessarily caused by an increased RH. This problem may be dissolved by recording a reference ECG, for example before the start of potentially 57

Chapter 2

arrhythmogenic medication. T-wave surface area Several studies showed a relation between RH, assessed from a limited number of action potential recordings, and Tarea. Tarea correlated with increased RH in rabbit hearts in which epicardial monophasic action potentials were recorded simultaneously with a surface ECG.8 In dogs, T-wave36 and QRST37 surface area was related to RH and a lowered threshold for ventricular fibrillation.38 Drugs that lengthen the APDs of specific cell layers, for example IKr blockers that foremost cause a lengthened midmyocardial APD, cause an increased Tarea in left ventricular wedge experiments.22 In our simulations, Tarea showed large increases at increasing RH. Tarea has the practical advantage of a low sensitivity to noise. Tarea may be the most comprehensive measure of RH, as it not only represents the maximum of the summed voltage gradients, like the T-amplitude, but also encompasses the time window during which the repolarization differences exist. T-wave symmetry ratio The T symmetry ratio was brought under attention by di Bernardo and Murray.9 They found that the T wave became more symmetrical with increased apico-basal and transmural dispersion. A normal symmetry ratio is 1.5 and with an increased RH this symmetry ratio approaches 1.0. Ischemia is known to induce high peaked symmetrical T waves, increased RH39 and vulnerability to arrhythmias.40 An advantage of this morphological parameter is that, in contrast to the high individual variability of Tamplitude, Tsymmetry seems to be more stable with different positions of the heart in the thorax.9 In our study, we also found that increased RH is reflected in a decreased Tsymmetry. Limitations As with all electrocardiographic studies, this simulation study addresses the forward problem and not the inverse problem, i.e., the changes in RH cause changes in the repolarization indices in the simulated surface ECG, but such ECG changes could theoretically also be caused by phenomena other than increased RH. Conclusions The well-established concept of RH is likely to play a major role in arrhythmo58

Validation of ECG indices of ventricular repolarization heterogeneity

genesis. In a realistic three dimensional computer model we simulated a number of situations with increased RH. It appeared that transmural dispersion of the repolarization increased linearly with global RH. QT dispersion has a low sensitivity in the transitional zone between normal and abnormal RH and has a weak theoretical basis as an index of RH. Tapex-end in the left precordial leads overestimated TDR. Tcomplexity did not discriminate between higher values of RH. In conclusion, Tsymmetry, Tamplitude, Tarea, and, with some limitations, Tapex-end and Tcomplexity reliably reflect changes in repolarization heterogeneity.

59

Chapter 2

REFERENCES 1. Han J, Moe GK: Nonuniform recovery of excitability in ventricular muscle. Circ Res 1964; 14:44-60. 2. Kuo CS, Munakata K, Reddy CP, Surawicz B: Characteristics and possible mechanism of ventricular arrhythmia dependent on the dispersion of action potential durations. Circulation 1983; 67:1356-1367. 3. Kanai A, Salama G: Optical mapping reveals that repolarization spreads anisotropically and is guided by fiber orientation in guinea pig hearts. Circ Res 1995; 77:784-802. 4. Volders PG, Sipido KR, Carmeliet E, Spatjens RL, Wellens HJ, Vos MA: Repolarizing K+ currents ITO1 and IKs are larger in right than left canine ventricular midmyocardium. Circulation 1999; 99:206-210. 5. Yan GX, Shimizu W, Antzelevitch C: Characteristics and distribution of M cells in arterially perfused canine left ventricular wedge preparations. Circulation 1998; 98:1921-1927. 6. De Ambroggi L: Heterogeneities of ventricular repolarization and vulnerability to arrhythmia. How to detect them with noninvasive methods? Cardiologia 1999; 44:355-360. 7. Anderson KP, Shusterman VBA: Changes in ventricular repolarization preceding the onset of spontaneous sustained ventricular tachycardia [abstract]. Pacing and Clin Electrophysiol 1999; 22:837. 8. Zabel M, Portnoy S, Franz MR: Electrocardiographic indexes of dispersion of ventricular repolarization: an isolated heart validation study. J Am Coll Cardiol 1995; 25:746-752. 9. di Bernardo D, Murray A: Explaining the T-wave shape in the ECG. Nature 2000; 403:40. 10. Priori SG, Mortara DW, Napolitano C, Diehl L, Paganini V, Cantu F, Cantu G, Schwartz PJ: Evaluation of the spatial aspects of T-wave complexity in the long-QT syndrome. Circulation 1997; 96:3006-3012. 11. Day CP, McComb JM, Campbell RW: QT dispersion: an indication of arrhythmia risk in patients with long QT intervals. Br Heart J 1990; 63:342-344. 12. Yan GX, Antzelevitch C: Cellular basis for the normal T wave and the electrocardiographic manifestations of the long-QT syndrome. Circulation 1998; 98:1928-1936. 13. de Bruyne MC, Hoes AW, Kors JA, Hofman A, van Bemmel JH, Grobbee DE: QTc dispersion predicts cardiac mortality in the elderly: the Rotterdam Study. Circulation 1998; 97:467-472. 14. Lubinski A, Kornacewicz-Jach Z, Wnuk-Wojnar AM, Adamus J, Kempa M, Krolak T, Lewicka-Nowak E, Radomski M, Swiatecka G: The terminal portion of the T wave: a new electrocardiographic marker of risk of ventricular arrhythmias. Pacing Clin Electrophysiol 2000; 23:1957-1959. 15. Zabel M, Malik M, Hnatkova K, Papademetriou V, Pittaras A, Fletcher RD, Franz MR: Analysis of T-wave morphology from the 12-lead electrocardiogram for prediction of long-term prognosis in male US veterans. Circulation 2002; 105:1066-1070. 16. Cowan JC, Hilton CJ, Griffiths CJ, Tansuphaswadikul S, Bourke JP, Murray A,

60

Validation of ECG indices of ventricular repolarization heterogeneity

17. 18. 19. 20. 21. 22. 23.

24. 25. 26. 27. 28. 29. 30. 31.

32.

Campbell RW: Sequence of epicardial repolarisation and configuration of the T wave. Br Heart J 1988; 60:424-433. Franz MR, Bargheer K, Rafflenbeul W, Haverich A, Lichtlen PR: Monophasic action potential mapping in human subjects with normal electrocardiograms: direct evidence for the genesis of the T wave. Circulation 1987; 75:379-386. van Oosterom A, Oostendorp TF: ECGSIM: an interactive tool for studying the genesis of QRST waveforms. Heart 2004; 90:165-168. Dower GE, Machado HB, Osborne JA: On deriving the electrocardiogram from vectorcardiographic leads. Clin Cardiol 1980; 3:87-95. Edenbrandt L, Pahlm O: Vectorcardiogram synthesized from a 12-lead ECG: superiority of the inverse Dower matrix. J Electrocardiol 1988; 21:361-367. Lay DC: Linear algebra and its applications. 2nd ed. Reading, MA: Addison-Wesley, 2003, p. 441-486. Shimizu W, Antzelevitch C: Sodium channel block with mexiletine is effective in reducing dispersion of repolarization and preventing torsade des pointes in LQT2 and LQT3 models of the long-QT syndrome. Circulation 1997; 96:2038-2047. Shimizu W, Antzelevitch C: Cellular basis for the ECG features of the LQT1 form of the long-QT syndrome: effects of beta-adrenergic agonists and antagonists and sodium channel blockers on transmural dispersion of repolarization and torsade de pointes. Circulation 1998; 98:2314-2322. Lehmann MH, Suzuki F, Fromm BS, Frankovich D, Elko P, Steinman RT, Fresard J, Baga JJ, Taggart RT: T wave “humps” as a potential electrocardiographic marker of the long QT syndrome. J Am Coll Cardiol 1994; 24:746-754. De Ambroggi L, Aime E, Ceriotti C, Rovida M, Negroni S: Mapping of ventricular repolarization potentials in patients with arrhythmogenic right ventricular dysplasia: principal component analysis of the ST-T waves. Circulation 1997; 96:4314-4318. van Oosterom A: Singular value decomposition of the T wave: its link with a biophysical model of repolarization. Int J Bioelectromagnetism 2002; 4:59-60. Kors JA, van Herpen G, van Bemmel JH: QT dispersion as an attribute of T-loop morphology. Circulation 1999; 99:1458-1463. Krahn AD, Nguyen-Ho P, Klein GJ, Yee R, Skanes AC, Suskin N: QT dispersion: an electrocardiographic derivative of QT prolongation. Am Heart J 2001; 141:111-116. Hii JT, Wyse DG, Gillis AM, Duff HJ, Solylo MA, Mitchell LB: Precordial QT interval dispersion as a marker of torsade de pointes. Disparate effects of class Ia antiarrhythmic drugs and amiodarone. Circulation 1992; 86:1376-1382. Zaidi M, Robert A, Fesler R, Derwael C, Brohet C: Dispersion of ventricular repolarisation: a marker of ventricular arrhythmias in patients with previous myocardial infarction. Heart 1997; 78:371-375. Da Costa A, Chalvidan T, Belounas A, Messier M, Viallet M, Mansour H, Lamaison D, Djiane P, Isaaz K: Predictive factors of ventricular fibrillation triggered by pausedependent torsades de pointes associated with acquired long QT interval: role of QT dispersion and left ventricular function. J Cardiovasc Electrophysiol 2000; 11:990997. Priori SG, Napolitano C, Diehl L, Schwartz PJ: Dispersion of the QT interval. A marker of therapeutic efficacy in the idiopathic long QT syndrome. Circulation 1994;

61

Chapter 2

89:1681-1689. 33. Zhang L, Timothy KW, Vincent GM, Lehmann MH, Fox J, Giuli LC, Shen J, Splawski I, Priori SG, Compton SJ, Yanowitz F, Benhorin J, Moss AJ, Schwartz PJ, Robinson JL, Wang Q, Zareba W, Keating MT, Towbin JA, Napolitano C, Medina A: Spectrum of ST-T-wave patterns and repolarization parameters in congenital longQT syndrome: ECG findings identify genotypes. Circulation 2000; 102:2849-2855. 34. van Oosterom A: Genesis of the T wave as based on an equivalent surface source model. J Electrocardiol 2001; 34 Suppl:217-227. 35. Bjornstad H, Storstein L, Dyre MH, Hals O: Electrocardiographic findings according to level of fitness and sport activity. Cardiology 1993; 83:268-279. 36. van Opstal JM, Verduyn SC, Winckels SK, Leerssen HM, Leunissen JD, Wellens HJ, Vos MA: The JT-area indicates dispersion of repolarization in dogs with atrioventricular block. J Interv Card Electrophysiol 2002; 6:113-120. 37. Abildskov JA, Green LS, Evans AK, Lux RL: The QRST deflection area of electrograms during global alterations of ventricular repolarization. J Electrocardiol 1982; 15:103107. 38. Kubota I, Lux RL, Burgess MJ, Abildskov JA: Relation of cardiac surface QRST distributions to ventricular fibrillation threshold in dogs. Circulation 1988; 78:171177. 39. Qian YW, Sung RJ, Lin SF, Province R, Clusin WT: Spatial heterogeneity of action potential alternans during global ischemia in the rabbit heart. Am J Physiol Heart Circ Physiol 2003; 285:H2722-H2733. 40. Swann MH, Nakagawa H, Vanoli E, Lazzara R, Schwartz PJ, Adamson PB: Heterogeneous regional endocardial repolarization is associated with increased risk for ischemia-dependent ventricular fibrillation after myocardial infarction. J Cardiovasc Electrophysiol 2003; 14:873-879.

62

Validation of ECG indices of ventricular repolarization heterogeneity

63

64

Chapter 3

Hypertensive stress increases dispersion of repolarization Bart Hooft van Huysduynen Cees A. Swenne Henk J. Ritsema van Eck Jan A. Kors Anna L. Schoneveld Hedde van de Vooren Piet Schiereck Martin J. Schalij Ernst E. van der Wall

Pacing Clin Electrophysiol 2004; 27: 1603-9

65

Chapter 3

ABSTRACT Purpose. Several electrocardiographic indices for repolarization heterogeneity have been proposed previously. We studied the behavior of these indices under two different stressors at the same heart rate: i.e., normotensive gravitational stress, and hypertensive isometric stress. Methods. In 56 healthy males ECG and blood pressure were recorded during rest (sitting with horizontal legs), hypertensive stress (performing handgrip) and normotensive stress (sitting with lowered legs). During both stressors, heart rates differed less than 10 % in 41 subjects, who constituted the final study group. Results. Heart rate increased from 63 ± 9 bpm at rest to 71 ± 11 bpm during normotensive and to 71 ± 10 bpm during hypertensive stress (P<0.001). Systolic blood pressure was 122 ± 15 mmHg at rest and 121 ± 15 mmHg during normotensive stress, and increased to 151 ± 17 mmHg during hypertensive stress (P<0.001). The QT interval was larger during hypertensive (405 ± 27) than during normotensive stress (389 ± 26, P<0.001). QT dispersion did not differ significantly between the two stressors. The mean Tapex-Tend interval of the midprecordial leads was larger during hypertensive (121 ± 17 ms) than during normotensive stress (116 ± 15 ms, P<0.001). The singular value decomposition T wave index was larger during hypertensive (0.144 ± 0.071) than during normotensive stress (0.089 ± 0.053, P<0.001). Conclusion. Most indices of repolarization heterogeneity were larger during hypertensive stress than during normotensive stress. Hypertensive stressors are associated with arrhythmogeneity in vulnerable hearts. This may in part be explained by the induction of repolarization heterogeneity by hypertensive stress.

66

Hypertensive stress increases dispersion of repolarization

INTRODUCTION Increased dispersion of the repolarization in the myocardium, also called heterogeneity of repolarization, increases the sensitivity of the substrate to reentry-based arrhythmias such as torsade de pointes. A number of studies involving mathematical simulation models1-4, animal experiments5 and observations in humans6 have provided evidence for the facilitory role of dispersion of the repolarization in arrhythmogenesis. During repolarization (phase 3 of the monophasic action potential), cardiac cells return from the refractory to the excitable state. As cardiac cells have different repolarization times, there is a time window during which a mix of refractory and excitable cells is present, roughly corresponding to the terminal part of the T wave in the electrocardiogram (ECG). Within this dispersion time-window, the conduction pathway of electrical extrastimuli is functionally determined by the distribution of excitable and non-excitable tissue. It is conceivable that hearts with larger dispersion time windows are more vulnerable to reentrant arrhythmias; obviously an additional triggering mechanism is needed to initiate such an arrhythmia. Apart from the patchy, local changes of the electrophysiological properties of the myocardium, resulting from ischemia7;8, infarction9, and inflammation10, which we will not consider here, primary electrical heart disease11, hypertrophy12, medication6, heart rate13, and autonomic nervous system influences14 have been mentioned as causes of increased dispersion of repolarization. All these conditions may amplify the intrinsic differences in action potential durations as normally found within the myocardium, e.g. between the right and the left ventricle, between the apex and the base, and between the epicardium, mid-myocardium and endocardium15. As the concept of dispersion of repolarization evolved, multiple electrocardiographic indices for its quantitative assessment have been proposed, such as the QT interval, QT dispersion, the Tapex-Tend interval in the precordial leads and singular value decomposition of the T wave. These measures are based on different electrophysiological and electrocardiographic concepts and have therefore different theoretical and practical strengths and weaknesses with respect to the nature of the dispersion reflected and the arrhythmogenic risk. To a certain extent they may, however, carry similar information.

67

Chapter 3

In this study we exposed normal subjects to two stressors, gravitational and isometric stress, while taking care that each subject attained the same heart rate under both stressors16. In a previous study14, we have shown that this experimental setup results in intra-individual repolarization differences, despite the identical heart rate. Thus, this experiment facilitates the studying of the dynamic behavior of all indices by intra-individual comparison during normotensive and hypertensive stress with minimal errors introduced by heart rate adjustment. In the current study we sought to answer two questions: 1) Will the above mentioned indices for repolarization heterogeneity all change in the same direction when gravitational and isometric stress are compared? 2) What can be said about differences of repolarization heterogeneity under gravitational and isometric stress?

68

Hypertensive stress increases dispersion of repolarization

SUBJECTS AND METHODS The Leiden University Medical Center Ethics Review Committee approved the protocol of this study. To include subjects with a large range of fitness, healthy male subjects, not intensively taking part in competitive sports, and male marathon skaters, engaged in training on a daily basis for more than 5 years, were recruited by advertisement in a local newspaper and by the Royal Dutch Skating Union, respectively. All participants gave written informed consent. The subjects were instructed to restrict, on the day preceding the measurements, their caffeine and alcohol consumption to respectively 6 and 2 beverages (the alcoholic drinks not later than 8 PM). On the day of the measurements, subjects were instructed not to smoke and not to drink alcohol or caffeine containing beverages. Furthermore, any exertional activity preceding the measurements had to be avoided. None of the participants used any medication. At recruitment, the good health of the participant was ascertained by medical history, physical examination, echocardiography, and a maximal oxygen consumption test was performed after the measurement session described below. Measurement session Instrumentation. Ten electrodes were attached to derive a standard 12-lead ECG. The finger cuff of a noninvasive continuous blood pressure measurement device17 (Finometer Medical systems, Arnhem, The Netherlands) was attached to the middle finger of the non-dominant hand. After the Finometer was switched on, the device was allowed to adapt to the conditions in the finger. If no satisfactory continuous blood pressure signal was attained, the cuff was wrapped around the ring finger of the same hand. The eight ECG signals I, II, V1-V6 were digitally stored together with the blood pressure signal on a ST Surveyor monitoring device (Mortara Rangoni Europe, S. Giorgio di Piano, Italy; sampling rate 500 Hz). Setting. The measurements were done in a quiet air-conditioned room (approximately 22º C). One investigator performed the measurements. No other personnel was allowed to enter the room during the measurement session, and speaking during the measurements was minimized. The subjects were placed on a tilt bed with foot support and an adjustable backrest that was always kept at a 70º angle with respect to the horizontal plane. Thus, the subjects were always “sitting”, with a constant position of the thorax throughout the experiment (Figure 1). 69

Chapter 3

Figure 1. Experimental conditions, from left to right: rest (sitting with horizontal legs), isometric stress (sitting during handgrip), gravitational stress (leg lowering at 4 different angles).

Rest measurement. The rest measurements were performed with the legs in horizontal position as shown in Figure 1. For stabilization purposes, the subjects rested during 20 minutes in this position prior to the actual 5-minute measurement. Isometric stress: handgrip. Still with the legs in horizontal position, the maximal grip force of the dominant hand was measured with a handgrip device ( Jamar, Bolingbrook, IL, USA). Subsequently, the subject performed a standard handgrip maneuver (30% maximal grip force during three minutes). Gravitational stress: leg lowering. After a rest of 5 minutes, a number of recordings were made at various leg lowering angles (tilt bed angles 30º, 45º, 60º, and 70º; the angle of the backrest was always kept at 70° with the horizontal plane) as shown in Figure 1. Each angle was maintained for five minutes. Analysis Rest, gravitational stress, isometric stress. We evaluated the data as obtained under three different loading conditions: rest (sitting with horizontal legs), the state with isometric stress (handgrip) and a selected state of gravitational stress (leg lowering). For each volunteer, a leg-lowering angle was sought at which the heart rate was closest to the heart rate during handgrip (maximal tolerated difference = 10 %). The ECG and blood pressure recorded during the last minute of these measurements were used for further analysis. QT interval. The QT interval is defined as the time between the earliest visible de70

Hypertensive stress increases dispersion of repolarization

flection from the isoelectric line between the P wave and the QRS complex in any of the twelve leads of the ECG until the last visible T wave activity in any of the leads. For the processing of the data the ECG computer program MEANS (Modular ECG Analysis System) was used18. MEANS determines common QRS onset and T offset for all 12 leads on one representative averaged beat, using template matching techniques. Bazett’s formula was used for the computation of the heart rate adjusted QT intervals. QT dispersion. The QT dispersion is defined as the difference between the longest and the shortest QT intervals in any of the leads. Starting from the overall end of T measurement, MEANS also determined lead-dependent T offsets, from which QT dispersion was derived as the difference between the longest and the shortest QT interval in any lead19. Tapex-Tend interval. These ECG measures were determined by means of a recently developed interactive ECG measurement computer program20. The end of T was manually set at the TU nadir of the precordial leads. The apex of the T wave was automatically determined. One observer initially did all analyses after which a second observer reviewed and when necessary adapted the results. Analyst and reviewer were blinded to the subject and state under consideration. Afterwards, the mean Tapex-Tend interval of the precordial leads closest to the heart, V2, V3 and V4, were averaged. Bazett’s formula was used for the computation of the heart rate adjusted Tapex-Tend intervals. Repolarization complexity. Singular value decomposition of the T wave is a method that calculates eight new orthogonal ECG derivations that represent the original ECG. The first derivation contains most energy; the second contains most energy orthogonal to the first derivation, the third most energy orthogonal to the first two derivations, etc. The first three derivations contain the dipolar signal of the ECG; they correspond to the long and the short axis of the spatial T loop and to its nonplanar component, respectively. The next 5 derivations (in which the amplitudes become progressively smaller) reflect non-dipolar components. As an index of dispersion of the repolarization we computed the singular-value-decomposition-based T wave complexity, expressed as the quotient of the second and the first singular values 21 . Singular value decomposition was performed in the terminal 10 seconds of the selected one-minute ECG episodes. Singular value decomposition algorithms were implemented in Matlab 6.5 (The MathWorks, Natick, USA). After computation of an averaged beat, the singular value decomposition was computed22 over an interval 71

Chapter 3

of 250 ms, from –100 ms to +150 ms around the peak of the T wave. Statistics To detect intra-individual differences between the dispersion parameters in the control state, under gravitational stress and under isometric stress, paired two-tailed ttests were done at a significance level of 0.05.

72

Hypertensive stress increases dispersion of repolarization

RESULTS Study group Measurements were initially performed in 56 subjects. In 41 subjects a leg-lowering angle could be found during which the heart rate differed less than 10% from the heart rate during handgrip. These 41 subjects constituted the study group (mean ± SD age 32.6 ± 11.2 years). The 15 subjects excluded from this analysis all had maximal leg lowering heart rates below the heart rate during handgrip. Obviously, prolonged tilting could have increased heart rate further, but we restricted the duration of our gravitational stress protocol to prevent orthostatic complications. Heart rate and systolic blood pressure responses to gravitational and isometric stress. Heart rates during gravitational (71 ± 11 bpm) and isometric stress (71 ± 10 bpm) were significantly higher than at rest (63 ± 9 bpm, P < 0.001). The average individual matching percentage for the heart rates under gravitational and isometric stress was -0.07 ± 3.14 %, without any significant difference between the heart rates (P = 0.948). Systolic blood pressure was 122 ± 15 mmHg at rest, remained 121 ± 15 mmHg during gravitational stress, and increased to 151 ± 17 mmHg during isometric stress (P < 0.001). Compared to rest, the gravitational stress was normotensive and the isometric stressor was hypertensive. The rate-pressure product increased progressively from rest (7684 ± 1438 bpm·mmHg) via gravitational (8638 ± 1605 bpm·mmHg) to isometric stress (10826 ± 2296 bpm·mmHg, P < 0.001 for all conditions). QT interval differences between gravitational and isometric stress The QT interval of the two stress states differed significantly. The QT interval during isometric stress (405 ± 27 ms) was longer than the QT interval during gravitational stress (390 ± 26 ms, P < 0.001). The QT interval during rest (408 ± 27 ms) was only significantly different from the QT interval during gravitational stress (P < 0.001). When corrected for heart rate, the QTc interval during both stressors differed significantly from rest (414 ± 18 ms), and the isometric stress QTc interval (433 ± 17 ms) was larger than the gravitational stress QTc interval (421 ± 18 ms, P < 0.001). Values of the QTc intervals are depicted in Figure 2.

73

Chapter 3

Fig.2. Bazett-corrected QT intervals during rest, gravitational and isometric stress.

QT dispersion differences between gravitational and isometric stress The QT dispersion was 48 ± 22 ms during rest, 45 ± 22 ms during gravitational stress and 51 ± 26 ms during isometric stress, which was not significantly different during any condition. Tapex-Tend interval differences between gravitational and isometric stress. The averaged Tapex-Tend interval from the midprecordial leads V2, V3 and V4 was significantly larger during isometric stress than during gravitational stress, 121 ± 17 ms vs. 116 ± 15 ms, respectively (P < 0.001). The Tapex-Tend interval during rest (123 ± 16 ms) was only significantly different from the interval during gravitational stress (P < 0.001). When corrected for heart rate, Tapex-Tend during isometric stress (129 ± 15 ms) was significantly larger than during rest (126 ± 12 ms, P = 0.003) and during gravitational stress (125 ± 14 ms, P < 0.001). The heart rate corrected TapexTend intervals are depicted in Figure 3. Singular value decomposition differences between gravitational and isometric stress. T wave complexity as computed with the singular value decomposition technique is depictedin Figure 4. All differences were significant; the complexity values under 74

Hypertensive stress increases dispersion of repolarization

isometric stress were larger than under gravitational stress, (0.144 ± 0.071 vs. 0.089 ± 0.053, P < 0.001).

Figure 3. Bazett-corrected mean Tapex-Tend interval of V2,V3 and V4.

Figure 4. Singular value decomposition of the T wave, S2/S1 ratio.

75

Chapter 3

DISCUSSION Our experimental set-up made intra-individual ECG comparisons possible at two different levels of blood pressure with minimal errors introduced by heart rate adjustment. Gravitational stress only increased heart rate, while isometric stress increased blood pressure as well. The rate-pressure product indicated that the stress levels increased from rest via leg-lowering to handgrip. We measured a number of parameters that have been put forward as indices for dispersion of repolarization. Most indices suggested that isometric stress is associated with increased dispersion of repolarization. The QTc interval increased under hypertensive stress, suggesting a lengthened repolarization time. The Tapex-Tend interval in the midprecordial leads was small but significantly longer under isometric stress, denoting increased transmural dispersion of the repolarization. The ratio of the second and first singular values of the singular value decomposition, a measure of T-wave complexity, was also significantly larger during isometric stress, implying increased global dispersion of the repolarization. QT dispersion differences between gravitational and isometric stress remained inconclusive. However, according to current insight, the physiological basis of this index is questionable23;24. Although prolongation of the QT interval has proven to be predictive of sudden cardiac death, arrhythmias and all-cause mortality, not only in populations with diseased hearts25;26 but also in seemingly healthy subjects27;28, the duration of the QT interval is mainly determined by the longest action potentials, rather than by APD differences. QT dispersion was believed to measure the spatial dispersion of the repolarization of the ventricle29. In a healthy population but also in patients with left ventricular dysfunction, QT dispersion is associated with increased susceptibility to ventricular arrhythmias30;31. However, the theoretical foundation of this index has come under severe criticism23;24. If the last part of the T vector is perpendicular to one of the 12 leads, this will result in the shortest QT interval in that lead. Thus, QT dispersion depends not only on the repolarization process itself, but also on the projection of the repolarization vector on the twelve ECG lead vectors24;32;33.

76

Hypertensive stress increases dispersion of repolarization

The interval between the apex and the end of the T wave (Tapex-Tend) in the precordial leads was proposed as a measure reflecting the transmural dispersion of repolarization in the region of the heart immediately under the electrode34. (Sub)endocardial cells have longer action potentials and repolarize later than the epicardial myocytes. In this concept the peak of the T wave corresponds to the end of the epicardial action potential while the end of the T wave corresponds to the end of the longest action potential. The experiments underlying this theory have been done in the laboratory, the ECG being derived from two electrodes placed at a short distance from the epicardium and endocardium in a wedge preparation of the ventricular wall15;35. Mathematical elaboration has shown that with increasing dispersion the relative contributions of derivations two and higher of the singular value decomposition increase36. Sofar, the clinical focus has been on the relative contribution of derivation two37 and on the non-dipolar components, four and higher38;39. The great advantage of singular value decomposition is its objectiveness in contrast to the usual phenomenological approach of the ECG. As mentioned above, all measured parameters have typical flaws and rely on different notions of dispersion. However, they all showed to be predictive for sudden cardiac death. The singular value decomposition and the Tapex-Tend interval provided the strongest physiological link with dispersion of the repolarization, because they are based on mathematical modelling40 and laboratory measurements15;34, respectively. The underlying mechanisms of the altered T-wave morphology under stress are the differences of the electrophysiological properties of the ventricular myocardium due to differences in sympathetic and parasympathetic tone and wall stress. With respect to the normotensive stressor, the hypertensive stressor encompasses increased wall stress due to increased cardiac afterload, elevated sympathetic tone due to chemoreceptor firing from muscles involved in the handgrip maneuver, and baroreflex-mediated elevated para-sympathetic tone14. Our study regards a physiological phenomenon in the general population. The stressors that were applied (leg lowering, that increased heart rate with 8 bpm, and handgrip, that additionally increased blood pressure with 30 mmHg) are comparable with the many physical and mental stressors that are met in daily life41. Our finding that 77

Chapter 3

hypertensive stress increases dispersion of the repolarization has potential clinical implications. In normal subjects, stress is associated with increased ventricular ectopy42. In this way hypertensive stress could provide the trigger (ectopy) plus the substrate (increased dispersion of the repolarization) needed to initiate serious arrhythmias. Although the chances are small, this still might be an important arrhythmogenic scenario as it applies to a large number of individuals (i.e., the general population). In patients with primary electrical disease, like long QT syndrome43, the increase in dispersion of the repolarization and incidence of ectopic beats caused by hypertensive stress may be exaggerated. Similarly, this scenario might explain why anger and mild exercise often precede serious arrhythmias in patients with internal defibrillators44. We conclude that hypertensive stress increases dispersion of the repolarization in the myocardium in healthy male subjects. This conclusion mainly relies on the increases of the Tapex-Tend interval and of the quotient of averaged absolute values of the second and first singular value decompositions of the T-wave. The increase of the QT interval may denote similar changes. Hypertensive stressors (physical and mental) are associated with arrhythmogeneity. These stressors may well increase dispersion of the repolarization, thus setting the stage for serious reentrant arrhythmias in vulnerable subjects.

78

Hypertensive stress increases dispersion of repolarization

REFERENCES 1. Abildskov JA, Lux RL. Cycle-length effects on the initiation of simulated torsade des pointes. J Electrocardiol. 1994;1-9. 2. Okazaki O, Wei D, Harumi K. A simulation study of torsade de pointes with M cells. J Electrocardiol. 1998;31 Suppl:145-151. 3. Viswanathan PC, Shaw RM, Rudy Y. Effects of IKr and IKs heterogeneity on action potential duration and its rate dependence: a simulation study. Circulation. 1999;99:2466-2474. 4. Abildskov JA, Lux RL. Simulated torsade de pointes--the role of conduction defects and mechanism of QRS rotation. J Electrocardiol. 2000;33:55-64. 5. Weissenburger J, Davy JM, Chezalviel F. Experimental models of torsades de pointes. Fundam Clin Pharmacol. 1993;7:29-38. 6. Elming H, Sonne J, Lublin HK. The importance of the QT interval: a review of the literature. Acta Psychiatr Scand. 2003;107:96-101. 7. Carluccio E, Biagioli P, Bentivoglio M et al. Effects of acute myocardial ischemia on QT dispersion by dipyridamole stress echocardiography. Am J Cardiol. 2003;91:385390. 8. Gottwald E, Gottwald M, Dhein S. Enhanced dispersion of epicardial activationrecovery intervals at sites of histological inhomogeneity during regional cardiac ischaemia and reperfusion. Heart. 1998;79:474-480. 9. Kudaiberdieva G, Gorenek B, Goktekin O et al. Combination of QT variability and signal-averaged electrocardiography in association with ventricular tachycardia in postinfarction patients. J Electrocardiol. 2003;36:17-24. 10. Goldeli O, Ural D, Komsuoglu B, Agacdiken A, Dursun E, Cetinarslan B. Abnormal QT dispersion in Behcet’s disease. Int J Cardiol. 1997;61:55-59. 11. Napolitano C, Priori SG, Schwartz PJ. Significance of QT dispersion in the long QT syndrome. Prog Cardiovasc Dis. 2000;42:345-350. 12. Zoghi M, Gurgun C, Yavuzgil O et al. QT dispersion in patients with different etiologies of left ventricular hypertrophy: the significance of QT dispersion in endurance athletes. Int J Cardiol. 2002;84:153-159. 13. Ishida S, Nakagawa M, Fujino T, Yonemochi H, Saikawa T, Ito M. Circadian variation of QT interval dispersion: correlation with heart rate variability. J Electrocardiol. 1997;30:205-210. 14. Frederiks J, Swenne CA, Kors JA et al. Within-subject electrocardiographic differences at equal heart rates: role of the autonomic nervous system. Pflügers Arch. 2001;441:717724. 15. Antzelevitch C, Fish J. Electrical heterogeneity within the ventricular wall. Basic Res Cardiol. 2001;96:517-527. 16. Berntson GG, Cacioppo JT, Quigley KS. Autonomic determinism: the modes of autonomic control, the doctrine of autonomic space, and the laws of autonomic constraint. Psychol Rev. 1991;98:459-487. 17. Imholz BP, Wieling W, Van Montfrans GA et al. Fifteen years experience with finger arterial pressure monitoring: assessment of the technology. Cardiovasc Res. 1998;38:605-616.

79

Chapter 3

18. van Bemmel JH, Kors JA, van Herpen G. Methodology of the modular ECG analysis system MEANS. Methods Inf Med. 1990;29:346-353. 19. Kors JA, Van Herpen G, Van Bemmel JH. QT dispersion as an attribute of T-loop morphology. Circulation. 1999;99:1458-1463. 20. Ritsema van Eck HJ. Fiducial segment averaging to improve cardiac time interval estimates. J Electrocardiol. 2002;35 Suppl:89-93. 21. Priori SG, Mortara DW, Napolitano C et al. Evaluation of the spatial aspects of Twave complexity in the long-QT syndrome. Circulation. 1997;96:3006-3012. 22. Acar B, Yi G, Hnatkova K et al. Spatial, temporal and wavefront direction characteristics of 12-lead T-wave morphology. Med Biol Eng Comput. 1999;37:574-584. 23. Kors JA, Van Herpen G, Van Bemmel JH. QT dispersion as an attribute of T-loop morphology. Circulation. 1999;99:1458-1463. 24. Malik M, Batchvarov VN. Measurement, interpretation and clinical potential of QT dispersion. J Am Coll Cardiol. 2000;36:1749-1766. 25. Padmanabhan S, Silvet H, Amin J et al. Prognostic value of QT interval and QT dispersion in patients with left ventricular systolic dysfunction: results from a cohort of 2265 patients with an ejection fraction of < or =40%. Am Heart J. 2003;145:132-138. 26. Moss AJ. Measurement of the QT interval and the risk associated with QTc interval prolongation: a review. Am J Cardiol. 1993;72:23B-25B. 27. de Bruyne MC, Hoes AW, Kors JA, Hofman A, Van Bemmel JH, Grobbee DE. Prolonged QT interval predicts cardiac and all-cause mortality in the elderly. The Rotterdam Study. Eur Heart J. 1999;4:278-284. 28. Al Khatib SM, LaPointe NM, Kramer JM et al. What clinicians should know about the QT interval. JAMA. 2003;289:2120-2127. 29. Franz MR, Zabel M. Electrophysiological basis of QT dispersion measurements. Prog Cardiovasc Dis. 2000;42:311-324. 30. Okin PM, Devereux RB, Howard BV, Fabsitz RR, Lee ET, Welty TK. Assessment of QT interval and QT dispersion for prediction of all-cause and cardiovascular mortality in American Indians: The strong Heart Study. Circulation. 2000;101:61-66. 31. Padmanabhan S, Silvet H, Amin J et al. Prognostic value of QT interval and QT dispersion in patients with left ventricular systolic dysfunction: Results from a cohort of 2265 patients with an ejection fraction of <= 40%. Am Heart J. 2003;145:132-138. 32. Huikuri H. Dispersion of repolarisation and the autonomic system-can we predict torsade de pointes? Cardiovasc Drugs Ther. 2002;16:93-99. 33. Kors JA, Van Herpen G. Measurement error as a source of QT dispersion: a computerised analysis. Heart. 1998;80:453-458. 34. Yan GX,Antzelevitch C.Cellular basis for the normalT wave and the electrocardiographic manifestations of the long-QT syndrome. Circulation. 1998;98:1928-1936. 35. Medina-Ravell VA, Lankipalli RS, Yan GX et al. Effect of epicardial or biventricular pacing to prolong QT interval and increase transmural dispersion of repolarization: does resynchronization therapy pose a risk for patients predisposed to long QT or torsade de pointes? Circulation. 2003;107:740-746. 36. van Oosterom A. Singular value decomposition of the T wave: its link with a biophysical model of repolarization. Int J Bioelectromagnetism. 2002;4:59-60.

80

Hypertensive stress increases dispersion of repolarization

37. Priori SG, Mortara DW, Napolitano C et al. Evaluation of the spatial aspects of Twave complexety in the long-QT syndrome. Circulation. 1997;96:3006-3012. 38. Malik M, Acar B, Gang Y, Yap YG, Hnatkova K, Camm AJ. QT dispersion does not represent electrocardiographic interlead heterogeneity of ventricular repolarization. J Cardiovasc Electrophysiol. 2000;8:835-843. 39. Zabel M, Malik M, Hnatkova K et al. Analysis of T-wave morphology from the 12lead electrocardiogram for prediction of long-term prognosis in male US veterans. Circulation. 2002;105:1066-1070. 40. van Oosterom A. Genesis of the T wave as based on an equivalent surface source model. J Electrocardiol. 2001;34 Suppl:217-227. 41. Swenne CA, Bootsma M, Van Bolhuis HH. Different autonomic responses to orthostatic and to mental stress in young normals. Homeostasis. 1995;36:287-292. 42. Stamler JS, Goldman ME, Gomes J et al. The effect of stress and fatigue on cardiac rhythm in medical interns. J Electrocardiol. 1992;25:333-338. 43. Wilde AA, Jongbloed RJ, Doevendans PA et al. Auditory stimuli as a trigger for arrhythmic events differentiate HERG- related (LQTS2) patients from KVLQT1related patients (LQTS1). J Am Coll Cardiol. 1999;33:327-332. 44. Lampert R, Joska T, Burg MM et al. Emotional and physical precipitants of ventricular arrhythmia. Circulation. 2002;106:1800-1805.

81

82

Chapter 4 Increased dispersion of ventricular repolarization during recovery from exercise

Harmen H.M. Draisma Bart Hooft van Huysduynen Cees A. Swenne Arie C. Maan Martin J. Schalij Ernst E. van der Wall

Submitted

83

Chapter 4

ABSTRACT Background. T-wave morphology hysteresis between exercise and recovery from exercise has been reported. We investigated the possible impact of this hysteresis for dispersion of the repolarization during exercise and recovery. Methods and Results. We studied 57 trained and untrained men of which the good health was ascertained by medical history, physical examination, echocar• diography and a maximal oxygen consumption ( V O2max) test. Baroreflex sensitivity (BRS) was noninvasively determined. Left ventricular mass (LVM) was assessed from the echocardiogram, and 10 unfit and 8 highly fit subjects were identified on • the basis of their V O2max, LVM, BRS, and resting heart rate (HR). Every ECG • made during the recovery phase of the V O2max tests was individually paired with a HR-matched ECG made during exercise. ECGs were characterized by QRS duration, QTpeak, QTend, T-wave area symmetry ratio, maximal T-wave magnitude and ventricular gradient magnitude. These ECG parameters were binned for exercise and for recovery ECGs according to the corresponding % heart rate reserve (HRR) or recovery time since maximal exercise. QTpeak, QTend, and T-wave area symmetry ratio were smaller, while maximal T-wave vector magnitude and ventricular gradient magnitude were larger during recovery than during exercise. For all parameters, the maximal recovery-exercise hysteresis was observed at 1 or 2 minutes recovery time and at 30-60% HRR. The highly fit subgroup had considerably larger exercise–recovery differences than the unfit subgroup. Conclusions. The observed T-wave hysteresis, accentuated in highly fit subjects, signifies increased dispersion of ventricular repolarization due to increased action potential duration heterogeneity during recovery from exercise.

84

Increased dispersion of ventricular repolarization during recovery from exercise

INTRODUCTION There is a long-standing debate about the safety of exercise. Albert and associates found a slightly increased risk of sudden death immediately following strenuous exercise in male American physicians reportedly free of cardiovascular disease but concluded that the benefits of exercise outweigh this disadvantage.1 However, based on an observed increased risk for sudden cardiac death during or shortly after exercise in athletes,2 the European Society of Cardiology recently proposed ECG screening of young athletes before taking part in competitive sports.3 In athletes, arrhythmogenic right-ventricular dysplasia (ARVD) is a frequent post-mortem finding. Increased vulnerability of the substrate, in combination with ARVD-based triggering ectopic activity might, at least partly, provide an explanation for the increased risk. The recovery phase after exercise deserves special attention. In a recent study, Frolkis and colleagues4 reported that ventricular ectopy occurring in the recovery phase of a maximal exercise test in a regular hospital population bears an independent mortality risk. Though the cause of death in this study remains unknown (all-cause mortality was the primary end point), it is conceivable that a number of deaths occurred after exercise, when abnormal automaticity (ectopy) met a more vulnerable substrate (increased dispersion of repolarization) to induce a lethal reentrant arrhythmia.5-7 In the last decade, assessment of dispersion of the repolarization from the ECG has gained increasing interest.8;9 In the light of this development and the possible increased susceptibility to arrhythmias during recovery from exercise, we sought, in this study, to further establish the concept of the electrophysiological hysteresis10 in the setting of maximal exercise tests.

85

Chapter 4

METHODS The local Ethics Review Committee approved the protocol of this study. To include subjects spanning a large range of fitness, healthy male persons not intensively taking part in competitive sports, and male marathon skaters engaged in training on a daily basis for more than 5 years were recruited by advertisement in a local newspaper and by the Royal Dutch Skating Federation, respectively. All participants gave written informed consent. None of them used any medication. Good health of the participants was ascertained by medical history, physical examination, echocardiog• raphy and a maximal oxygen consumption ( V O2max) test. The study population (see Table 1) consisted of 57 individuals of an original cohort of 70 respondents who fulfilled the previously listed criteria. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Left ventricular mass (LVM) was estimated from the echocardiographic data following the area-length method according to the American Society of Echocardiography Committee on Standards,11 and normalized to body surface area as calculated using the formula of Mosteller.12 Baroreflex sensitivity (BRS) was assessed according to the protocol described by Frederiks et al.13 Bicycle ergometry began with a load of 40 watts (W). This load was increased by 20W per minute until maximal exercise was reached. Ten-second ECG recordings (leads I, II, V1–V6; sampling rate 500 Hz) were obtained twice a minute with a Marquette Case 8000 exercise electrocardiograph (GE Healthcare, Milwaukee, WI, USA) during supine rest, exercise and recovery, up to 5 minutes after maximal exercise. Later, these ECGs were downloaded to a PC for analysis.

86

Increased dispersion of ventricular repolarization during recovery from exercise

All subjects

Unfit

Highly fit

(N = 57)

(N = 10)

(N = 8)

P (H-U)

Mean ± SD

Range

Mean ± SD

Range

Mean ± SD

Range

33.1 ± 11.3

17.9 – 55.8

35.9 ± 11.4

17.9 – 55.8

31.0 ± 10.5

20.4 – 54.0

NS

183 ± 6

172 – 196

181 ± 6

172 – 190

188 ± 5

178 – 193

.018

Weight (kg)

78.5 ± 10.9

61 – 106

80.6 ± 14.1

63 – 101

74.1 ± 5.8

64 – 83

NS

BMI (kg·m–2)

23.4 ± 3.2

18.6 – 33.8

24.8 ± 4.7

19.2 – 33.8

21.0 ± .78

19.9 – 22.3

NS

VO2max (ml·kg–1·min– 1 )

52.9 ± 13.7

29.4 – 78.7

40.4 ± 8.3

30.6 – 51.3

66.8 ± 6.4

57.8 – 75.8

.000

LVM (g·m–2)

117.4 ± 19.9 64.6 – 160.1 100.9 ± 10.8 82.2 – 114.4 140.7 ± 11.3 123.2 – 155.6 .000

Age (yrs) Height (cm)

BRS (ms·mmHg–1)

10.0 ± 5.0

4.1 – 33.4

6.8 ± 1.3

4.5 – 8.7

13.2 ± 3.8

9.6 – 21.1

.000

Resting HR (bpm)

60 ± 8

39 – 79

65 ± 3

60 – 69

53 ± 6

39 – 59

.000

Table 1. Characteristics for all subjects, and for subgroups of low and high cardiorespiratory fitness. Abbreviations: BMI: body mass index; BRS: baroreflex sensitivity; H: highly fit; HR: heart rate; U: unfit; LVM: body surface area-corrected left ventricular mass; NS: nonsignificant; SD: standar d deviation from the mean.

Exercise-recovery ECG comparison requires the selection of heart rate (HR)matched exercise and recovery ECG pairs. In our healthy study group with a relatively rapid heart rate recovery, the number of ECGs recorded during the recovery phase is far less than the number of ECGs recorded during the exercise phase. Therefore, we selected the HR-matched exercise-recovery ECG pairs by finding for every recorded recovery ECG the best HR-matching exercise ECG. Exercise–recovery ECG pairs with a HR matching error >10 beats per minute (bpm) were excluded from analysis. Each thus selected ECG pair was characterized by its corresponding percentage of the heart rate reserve (HRR, the difference between the maximal HR

87

Chapter 4

and HR during supine rest) and by the time instant after peak exercise at which the recovery ECG was recorded (recovery time, RCT). Figure 1 illustrates this procedure for a typical subject.

Figure 1. Symptom-limited exercise test in a normal subject. Triangles: exercise; squares: recovery. Each recovery ECG is heart-rate-matched with one exercise ECG. Thus selected exercise–recovery pairs are indicated by solid markers and connected by horizontal lines. The highlighted exercise–recovery ECG pair is shown in Figures 2 and 3. Abbreviations: HRmin, HRmax: resting heart rate (0% of the heart rate reserve) and maximal heart rate (100% of the heart rate reserve), in bpm.

We used LEADS14 to analyze the ECGs. After interactive beat selection and calculation of the averaged beat, LEADS sets the default end-of-QRS instant at the minimal heart vector between the QRS complex and the T wave, and the default end-of-T instant at the intercept with the baseline of the steepest tangent to the descending limb of the T wave in the vector magnitude signal. The latter prevents possible influences of fusion of the T wave with a U or a P wave — a problem arising at higher HRs.15 Finally, the operator adjusted the end-of-QRS instant by repositioning a vertical crosshair cursor in a display of the superimposed 12 ECG leads that can be magnified at will. In keeping with the Minnesota ECG coding

88

Increased dispersion of ventricular repolarization during recovery from exercise

protocol,16 end-of-QRS was set at the latest J point in any of the leads. In leads with two candidate J points the earliest J point was taken. All interactive procedures (beat selection for averaging; end-of-QRS adjustment) were done by one of the authors (H.H.M.D.). Each ECG was characterized by six parameters: QRS duration, QTpeak, QTend, T-wave area symmetry ratio (SRarea), maximal T-wave magnitude (maxT) and ventricular gradient magnitude (VG). QTpeak, QTend, SRarea and maxT were computed in the vector magnitude signal. SRarea was calculated as the ratio of the early (from the end of the QRS complex to the apex of the T wave) to the late (from the apex of the T wave to its end) T-wave area.17 VG was computed by vectorially adding the QRST areas in the scalar X-Y-Z leads. Exercise and recovery values of these six parameters were compared in the total study group (N = 57) and in two subgroups of unfit (N=10) and highly fit (N=8) subjects, respectively. Subjects were allocated to one of these subgroups on the basis of their • V O2max• , BRS, LVM and resting HR values. Unfit subjects were those whose measured V O2max, BRS, and LVM were below, and whose supine resting HR was above the median of the total study group. Highly fit subjects were those whose measured • V O2max, BRS, and LVM were above, and whose supine resting HR was below the median of the total study group. See Table 1. Data was pooled in %HRR and in RCT bins, 10% and 1 minute wide, respectively. There was data in six RCT bins (centered around 0, 1, …, 5 minutes post peak-exercise) and in nine %HRR bins (centered around 20, 30, …, 100 %HRR). Presence of exercise–recovery hysteresis was tested in the total study group and in the unfit and highly fit subgroups by comparing the exercise and recovery contents of each %HRR and RCT bin, using paired t-tests at the 5% level. The amount of hysteresis in a given ECG parameter in a given bin was expressed as a fraction (difference between the recovery and exercise values, divided by the exercise value). Differences between the unfit and highly fit subgroups were tested in a similar way, this time using unpaired t-tests at the 5% significance level. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.

89

Chapter 4

RESULTS Exercise tests, including recovery, lasted 18 ± 4 minutes (mean ± SD). Subjects had a resting HR of 60 ± 8 bpm, reached a peak HR of 178 ± 11 bpm; after 5 minutes recovery HR was 101 ± 18 bpm (36 ± 11% HRR). Remarkable within-subject T-wave morphology differences were observed at equal heart rates during exercise and recovery (see Figures 2 and 3). The exercise-recovery ECG differences as depicted in Figures 2 and 3 are typical for what was observed in the whole study group: at similar heart rates, during recovery after maximal exercise, QTpeak, QTend and SRarea were smaller, while maxT and VG were larger. QRS durations during exercise and recovery were not different.

A

B

Figure 2. Ten-second ECGs recorded at similar heart rates during exercise (panel A) and recovery (panel B) in a normal subject. ECGs were recorded at 55% of the heart rate reserve (in this subject 137 bpm, see also Figure 1). T-wave amplitude increases during recovery are most prominent in the left-precordial leads V2 and V3.

90

Increased dispersion of ventricular repolarization during recovery from exercise

A

B

Figure 3. Vector magnitude signal of the averaged QRST complexes obtained from the heart-ratematched ECGs displayed in Figure 2. In panel A the exercise ECG is displayed in black; the recovery ECG is displayed in grey for visual comparison. Panel B displays the recovery ECG in black, and the exercise ECG in grey. All parameter values in panel A relate to the exercise ECG, and in panel B to the recovery ECG. In the descending limb of the T wave the tangent to the steepest slope is drawn; the intercept of this tangent with the baseline is taken as the end of the T wave. Left hatched area: area under the curve between end-of-QRS and peak T. Right hatched area: area under the curve between peak T and end-of-T. Note that the differences between the QRS-T complexes concentrate in the T wave, the morphologies of the QRS complexes during exercise and recovery are strikingly similar. Abbreviations: HR: heart rate, maxT: maximal T-wave magnitude, SRarea: T-wave area symmetry ratio, VG: ventricular gradient magnitude.

A summary of the exercise–recovery comparison in all subjects is given in Figures 4 (QTpeak, QTend) and 5 (maxT, SRarea, VG), and in Table 2. Significant exercise–recovery hysteresis was detectable for each T-wave ECG parameter, irrespective of whether the data were distributed over %HRR bins or over RCT bins. Maximal hysteresis in QTpeak and in QTend occurred after 1 minute RCT or 20–60% HRR, recovery values differed about 5 to 9% from the exercise values. Maximal hysteresis in maxT and in VG occurred after 2 minutes RCT or 60% HRR, recovery values differed about 50 to 90% from the exercise values. Maximal hysteresis in SRarea occurred after 1 minute RCT or at 30% HRR, recovery values differed about 20% from exercise values.

91

Chapter 4

Figure 4. QTpeak (upper panels) and QTend (lower panels) in heart-rate-matched ECGs recorded during exercise (triangles) and recovery (squares) in all subjects, ordered according to recovery time (left panels) and percentage HRR (right panels). Error bars indicate standard deviations from the mean. Significant differences between recovery and exercise are indicated by single (P<0.05) and double (P<0.01) asterisks.

92

Increased dispersion of ventricular repolarization during recovery from exercise

Panel A Parameter

RCT (min)

Exercise

Recovery

% Change

P

QTpeak

1

211 ms

193 ms

– 5.6

.000

QTend

1

287 ms

271 ms

– 8.5

.000

max T

2

541 µV

964 µV

+ 78

.000

SRarea

1

1.18

0.95

– 19

.000

VG

2

63 mV·ms

95 mV·ms

+ 52

.000

Panel B Parameter

%HRR

Exercise

Recovery

% Change

P

QTpeak

20

289 ms

264 ms

– 8.6

.048

QTend

60

296 ms

279 ms

– 5.7

.000

max T

60

553 µV

1046 µV

+ 89

.000

SRarea

30

1.51

1.21

– 20

.000t

VG

60

56 mV·ms

94 mV·ms

+ 67

.000

Table 2. Maximal exercise-recovery hystereses in all participants, ordered according to recovery time (RCT, Panel A) and to percentage heart rate reserve (HRR, Panel B). Exercise-recovery changes are expressed with respect to the exercise values. Abbreviations: maxT: maximal T-wave vector magnitude; SRarea: T-wave area symmetry ratio; VG: ventricular gradient magnitude.

Table 1 shows that age, weight and BMI did not differ in the unfit and highly fit subgroups. The highly fit subjects are a bit taller, though. Table 3 shows that, irrespective of RCT or %HRR binning, the amplitude of the hysteresis was larger in highly fit than in unfit subjects. VG and maxT roughly assumed the double value after 1 minute RCT or at 60% or 40% HRR, respectively.

93

Chapter 4

Panel A Parameter

RCT (min)

Unfit

Highly fit

hysteresis

hysteresis

% Difference

P

QTpeak

2

– 9.3 ms

– 24 ms

+ 160

.016

QTend

1

– 7.5 ms

– 27 ms

+ 255

.020

maxT

1

+ 290 µV

+ 570 µV

+ 96

.000

SRarea

2

– 0.23

– 0.25

+8

NS

VG

1

+ 24 mV∙ms

+ 41 mV∙ms

+ 71

.007

Unfit

Highly fit

% Difference

hysteresis

hysteresis

Panel B Parameter

%HRR

P

QTpeak

60

– 9.7 ms

– 27 ms

+ 179

NS

QTend

60

+ 0.6 ms

– 29 ms

– 5.2·10

NS

maxT

60

+ 283 µV

+ 669 µV

+ 136

.000

SRarea

80

– 0.15

+ 1.09

– 807

NS

VG

40

+ 15 mV·ms

+ 38 mV·ms

+ 153

.018

3

Table 3. Maximal exercise-recovery hysteresis differences between highly fit and unfit subjects, related to the unfit hysteresis value, and ordered according to recovery time (RCT, Panel A) and percentage heart rate reserve (%HRR, Panel B). Note the remarkably higher maximal T-wave vector (maxT) and ventricular gradient magnitude (VG) in highly fit subjects. Abbreviations: NS: nonsignificant; SRarea: T-wave area symmetry ratio.

94

Increased dispersion of ventricular repolarization during recovery from exercise

DISCUSSION Our study demonstrates hysteresis in heart-rate-matched ECGs made during exercise and recovery, with the largest differences observed at 1–2 minutes after peak exercise or at 60% of the HRR. Highly fit subjects had more QT- and T-wave hysteresis than unfit subjects. In the following we interpret the ECG differences between exercise and recovery in the light of dispersion of the repolarization. Dispersion of repolarization (DOR) DOR in the ventricles is a physiological phenomenon that is caused by the added effect of heterogeneity in activation time and heterogeneity in action potential morphology. T-wave generation rests on DOR: when all myocytes would repolarize at the same time there would be no T wave at all. DOR is considered to facilitate reentrant activity. Tacitly assuming that DOR is closely linked to dispersion of refractoriness,18 an increase of DOR would widen the time window during which an extrasystole could initiate a tachyarrhythmia.5-7 Evolving insight in exercise-recovery ECG hysteresis Interest in ECG changes during recovery from maximal exercise arose about 35 years ago, when computerized exercise-ECG processing became available. Kitchin and Neilson showed that, after exercise, T-wave amplitude increased, and that the peak of the T wave occurred earlier in the cardiac cycle.19 The most striking changes occurred in the first minute after exercise. The study was conducted in untrained subjects who performed submaximal exercise, and only a single ECG lead was explored. Simoons and Hugenholtz were the first to establish, in maximal exercise tests, differences in QTpeak and maxT at similar HR during exercise and during recovery from maximal exercise.20 They found shorter QTpeak and larger maxT values during recovery. Sarma and colleagues confirmed that QTpeak was shorter during recovery and measured, in addition, that QTend intervals were shorter during recovery than during exercise.10 Chauhan and colleagues found higher QTend hysteresis in healthy females than in healthy males.15 Heart-rate-matched exercise–recovery QTend hysteresis is augmented in patients with long-QT syndrome;21 this phenomenon normalizes with beta-blockade.22

95

Chapter 4

Explanations for exercise–recovery hysteresis in T-wave morphology have ranged from sympathetic withdrawal,19 altered hemodynamic performance of the heart,20 residual sympatho-adrenal activity in the early post-exercise period and relatively slow QT-interval adaptation to rapid changes in the RR interval.10 More recently, Langley and colleagues published further quantitative data on altered T-wave symmetry during recovery from submaximal HR, and compared this with T-wave symmetry at rest.23 They attributed the lower symmetry ratio (more symmetric T wave) during recovery to increased DOR. However, this conclusion has a limited significance, as the heart rates in the resting and recovery states differed. ECG measures of dispersion of repolarization (DOR) In the past decades, multiple ECG indexes to quantify DOR have been proposed, like the QT interval,24 QT dispersion,25 the interval between the apex and end of the T wave (Tapex–end),26 T-wave complexity (Tcomplexity) estimated by singular value decomposition (SVD),27;28 T-wave amplitude, T-wave symmetry.17 In the current study, we used two of these parameters, T-wave amplitude and T-wave symmetry, to quantify DOR. We abandoned the other indexes for the following reasons: • •

•

•

96

The QT interval does not specifically measure DOR,29 as a homogeneous APD increase that does not alter DOR can also prolong this interval. Although the dispersion in QT interval among ECG leads indeed is influenced by APD heterogeneity30 and by T-wave complexity,31 QT dispersion in the ECG does not necessarily represent DOR.32 Moreover, manual measurement of e.g. QT dispersion has been demonstrated to be subjective and error-prone.33 SVD is a mathematical method that indeed has been analytically demon strated to quantify DOR from the T wave in the ECG.34 However, T-wave complexity estimated from singular values does not discriminate between higher values of DOR.35 Moreover, it is sensitive to noise (noise adds to the complexity), which renders it less suited for exercise electrocardiogra phy.27;36 Tapex–end aims to measure the time window during which there is trans mural DOR. This index was conceived on the basis of a wedge preparation of the left ventricular wall and two floating electrodes close to the epicar-

Increased dispersion of ventricular repolarization during recovery from exercise

dium and the endocardium,26 which is an incomplete analogue of the intact heart in a torso. Morover, mathematical simulation revealed that Tapex–end, though linearly related to transmural DOR, overestimates it.35 The ventricular gradientThe concept of the ventricular gradient was conceived by Wilson and co-workers.37;38 It has theoretically been demonstrated that the spatial ventricular gradient (the integral of the spatial QRST loop, G = H (t ) ⋅ d t in which H (t ) the heart vector) is non-zero due to action potential morphology differences — most often thought of as differences in APD — and that it is independent of the order in which the ventricles are electrically activated.37-42 We have used the VG in our study as an extra parameter that helps to trace the cause of altered DOR (that necessarily has to be found in changes in the ventricular activation sequence and / or changes in APD heterogeneity).

∫

Augmented DOR during recovery from exercise Our study demonstrates that, in heart-rate-matched ECGs, QTpeak and QTend intervals are briefer, the T wave is more symmetric, and the maximal T vector and the ventricular gradient are larger during recovery from exercise than during actual exercise. These ECG changes are consistent with the hypothesis that during recovery from exercise the repolarization heterogeneity is augmented due to increased APD heterogeneity. This can be concluded from the following four arguments: 1. APD shortening during recovery. The decreases in QTpeak and QTend during recovery are suggestive for a generalized APD shortening shortly after exercise. This APD shortening could be caused by the increased adrenergic influences during recovery from exercise: norepinephrine and epinephrine levels, increasing during exercise, continue to increase further after exercise.43 Admittedly, parasympathetic outflow, which is very little during actual exercise, resumes during recovery,44 which creates a situation of both enhanced adrenergic and cholinergic influences during recovery from exercise in comparison to during actual exercise. However, the study by Inoue and Zipes45 has shown that, at identical heart rates, the ventricular effective refractory period (and, hence, likely, APD) is smaller under combined elevated sympathetic and parasympathetic stimulation.

97

Chapter 4

Independent of autonomic influences on the myocardium, heart rate memory causes shorter APDs during recovery. At a given heart rate in the exercise phase, the recent history of slower heart rates tends to produce longer APDs than at the same heart rate in the recovery phase, where the recent history of higher heart rates tends to produce shorter APDs. Both of the above mentioned effects may have accounted for the observed QTpeak and QTend shortening during recovery. 2. Increased DOR. The more symmetrical T wave in combination with the increased T-wave amplitude during recovery from exercise indicate increased DOR17;32;35 with respect to actual exercise. 3. Similar depolarization. The absence of exercise–recovery hysteresis in QRS duration suggests that ventricular depolarization did not dramatically change (compare panels A and B in Figures 2 and 3 for a visual impression). Therefore, it is not very likely that the observed T-wave changes were secondary changes (i.e., that these changes were caused by altered intraventricular conduction). This leaves primary changes (APD changes) as a plausible explanation of the modified T wave morphology. 4. Increased APD heterogeneity. The increase in VG during recovery signifies increased APD dispersion during recovery. This is likely to occur because APD alterations in response to a combination of sympathetic and parasympathetic stimulation, such as imposed upon the myocardium during recovery from exercise, differ regionally. E.g., at the endocardium, there is no independent parasympathetic effect on APD: parasympathetic stimulation mainly reduces sympathetically induced APD shortening.46 However, stimulation of the epicardium by acetylcholine has an independent effect, and can slightly increase, but, at higher concentrations, also reduce APD.47 Dynamics of exercise, fitness and hysteresis The marked electrocardiographic differences at matched heart rates during exercise and recovery signify a pronounced electrophysiological hysteresis. This hysteresis rests on the highly dynamic nature of a maximal exercise test, causing exercise–recovery contrasts in autonomic state and in recent heart rate history. The more extreme the exercise protocol, the more outspoken the hysteresis effect will be. Our data demonstrate that T-wave symmetry, T-wave amplitude and VG magnitude 98

Increased dispersion of ventricular repolarization during recovery from exercise

have a much larger hysteresis in highly fit than in unfit subjects. Possible cause is the stronger parasympathetic reactivation during recovery in highly fit subjects, as demonstrated by Imai and colleagues,48 in combination with the increased DOR as associated with hypertrophied hearts.49 It has to be realized that the largest T-wave amplitudes and ventricular gradient magnitudes that occurred during recovery did not occur during any exercise stage (see Figure 5). Hence, a surge of intense exercise is possibly one of the most effective stimuli to create a situation with relatively large cardiac electrical heterogeneity. Clinical implications Our study provides suggestive evidence for the hypothesis that DOR is larger during recovery from exercise than during actual exercise, and that this effect is stronger in highly fit than in unfit persons. On itself, the very presence of a hysteresis phenomenon is not sufficient to Figure 5. Maximal T-wave vector magnitude (upper create a situation of potential risk: panels), T-wave area symmetry ratio (middle panels) the increased T-wave symmetry, and ventricular gradient magnitude (lower panels) in larger T-wave amplitudes (markheart-rate-matched ECGs recorded during exercise (triangles) and recovery (squares) in all subjects, ordered ers for heterogeneity of the repoaccording to recovery time (left panels) and percentage larization) and ventricular gradient heart rate reserve (right panels). Error bars indicate magnitudes (a marker for APD standard deviations from the mean. Significant differences between recovery and exercise are indicated by sin- heterogeneity) during recovery as gle (P<0.05) and double (P<0.01) asterisks. compared to exercise might well be physiologically reasonable and, on itself, not arrhythmogenic. In fact, no arrhythmias were seen during the exercise tests 99

Chapter 4

in our study population. It is conceivable, though, that increased DOR during recovery from exercise plays a facilitatory role in the development of reentrant tachyarrhythmias in persons having ambient arrhythmias during recovery from exercise that could serve as a trigger. However, conclusions about this hypothesis cannot be drawn from our study and require further investigation. We acknowledge that in the general population, habitual sports activity improves several outcomes of overall health, including the risk of death caused by coronary artery disease.50 However, for example in those with increased automaticity due to undiagnosed structural heart disease,2 sports activity may set the stage for reentry type arrhythmias in a substrate which is transiently at increased vulnerability.1 Prevention strategies might aim at the detection of this subgroup. Furthermore, it might be useful to scan ECGs made during exercise tests on the changes as reported in our study. It must, however, be stressed that our study results were obtained in healthy individuals and that extrapolation to patients or to athletes with structural heart disease responsible for sudden death is highly speculative at this time. It remains to be proven that similar changes in dispersion of ventricular repolarization occur in such risk groups, and that increased dispersion of ventricular repolarization during recovery from exercise bears increased risk of sudden death in the setting of sports activity. Conclusion In conclusion, our study provides strong suggestive evidence for increased dispersion of repolarization during recovery from maximal exercise in normal male persons of any fitness level. Obviously, such an explicit exercise–recovery hysteresis will not exist to this extent at lower peak exercise intensity levels. We demonstrated also that, after maximal exercise, highly fit persons have a much larger hysteresis than unfit subjects. It should further be investigated if this effect renders individuals more susceptible to arrhythmia induction (e.g., by a triggering ventricular extrasystole) during recovery from exercise.

100

Increased dispersion of ventricular repolarization during recovery from exercise

REFERENCES 1. Albert CM, Mittleman MA, Chae CU, Lee IM, Hennekens CH, Manson JE. Triggering of sudden death from cardiac causes by vigorous exertion. N Engl J Med. 2000;343:1355-1361. 2. Corrado D, Basso C, Rizzoli G, Schiavon M, Thiene G. Does sports activity enhance the risk of sudden death in adolescents and young adults? J Am Coll Cardiol. 2003;42:1959-1963. 3. Corrado D, Pelliccia A, Bjornstad HH, et al. Cardiovascular pre-participation screening of young competitive athletes for prevention of sudden death: proposal for a common European protocol: Consensus Statement of the Study Group of Sport Cardiology of the Working Group of Cardiac Rehabilitation and Exercise Physiology and the Working Group of Myocardial and Pericardial Diseases of the European Society of Cardiology. Eur Heart J. 2005;26:516-524. 4. Frolkis JP, Pothier CE, Blackstone EH, Lauer MS. Frequent ventricular ectopy after exercise as a predictor of death. N Engl J Med. 2003;348:781-790. 5. Abildskov JA, Lux RL. Cycle-length effects on the initiation of simulated torsade des pointes. J Electrocardiol. 1994;27:1-9. 6. Moe GK, Rheinboldt WC, Abildskov JA. A computer model of atrial fibrillation. Am Heart J. 1964;67:200-220. 7. Kuo CS, Munakata K, Reddy CP, Surawicz B. Characteristics and possible mechanism of ventricular arrhythmia dependent on the dispersion of action-potential durations. Circulation. 1983;67:1356-1367. 8. Surawicz B. The electrophysiologic basis of ECG and cardiac arrhythmias. Baltimore, MD, USA: Williams and Wilkins, 1995. 9. Engel G, Beckerman JG, Froelicher VF, Yamazaki T, Chen HA, Richardson K, McAuley RJ, Ashley EA, Chun S, Wang PJ. Electrocardiographic arrhythmia risk testing. Curr Probl Cardiol. 2004;29:365-432. 10. Sarma JS, Venkataraman SK, Samant DR, Gadgil U. Hysteresis in the human RR-QT relationship during exercise and recovery. Pacing Clin Electrophysiol. 1987;10:485491. 11. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum H, Gutgesell H, Reichek N, Sahn D, Schnittger I, . Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of TwoDimensional Echocardiograms. J Am Soc Echocardiogr. 1989;2:358-367. 12. Mosteller RD. Simplified calculation of body-surface area. N Engl J Med. 1987;317:1098. 13. Frederiks J, Swenne CA, TenVoorde BJ, Honzikova N, Levert JV, Maan AC, Schalij MJ, Bruschke AV. The importance of high-frequency paced breathing in spectral baroreflex sensitivity assessment. J Hypertens. 2000;18:1635-1644. 14. Draisma HHM, Swenne CA, Van de Vooren H, Maan CA, Hooft van Huysduynen B, Van der Wall EE, Schalij MJ. LEADS: an interactive research oriented ECG/VCG analysis system. Computers in Cardiology. 2005;32:515-518. 15. Chauhan VS, Krahn AD, Walker BD, Klein GJ, Skanes AC, Yee R. Sex differences

101

Chapter 4

16.

17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

102

in QTc interval and QT dispersion: dynamics during exercise and recovery in healthy subjects. Am Heart J. 2002;144:858-864. Prineas RJ, Crow RS, Blackburn H. Intraventricular Conduction Defects (7-Codes). The Minnesota Code Manual of Electrocardiographic Findings: Standards and Procedures for Measurement and Classification. Bristol, England: John Wright, 1982: 111-130. Di Bernardo D, Murray A. Medical physics - Explaining the T-wave shape in the ECG. Nature. 2000;403:40. Hoffman BF, Kao CY, Suckling EE. Refractoriness in cardiac muscle. Am J Physiol. 1957;190:473-482. Kitchin AH, Neilson JM. The T wave of the electrocardiogram during and after exercise in normal subjects. Cardiovasc Res. 1972;6:143-149. Simoons ML, Hugenholtz PG. Gradual changes of ECG waveform during and after exercise in normal subjects. Circulation. 1975;52:570-577. Krahn AD, Klein GJ, Yee R. Hysteresis of the RT interval with exercise: a new marker for the long-QT syndrome? Circulation. 1997;96:1551-1556. Krahn AD, Yee R, Chauhan V, Skanes AC, Wang J, Hegele RA, Klein GJ. Beta blockers normalize QT hysteresis in long QT syndrome. Am Heart J. 2002;143:528-534. Langley P, Di Bernardo D, Murray A. Quantification of T wave shape changes following exercise. Pacing Clin Electrophysiol. 2002;25:1230-1234. Moore EN. Mechanisms and models to predict a QTc effect. Am J Cardiol. 1993;72:4B9B. Franz MR, Zabel M. Electrophysiological basis of QT dispersion measurements. Prog Cardiovasc Dis. 2000;42:311-324. Yan GX,Antzelevitch C.Cellular basis for the normalT wave and the electrocardiographic manifestations of the long-QT syndrome. Circulation. 1998;98:1928-1936. Acar B, Koymen H. SVD-based on-line exercise ECG signal orthogonalization. IEEE Trans Biomed Eng. 1999;46:311-321. Priori SG, Mortara DW, Napolitano C, Diehl L, Paganini V, Cantu F, Cantu G, Schwartz PJ. Evaluation of the spatial aspects of T-wave complexity in the long-QT syndrome. Circulation. 1997;96:3006-3012. Antzelevitch C. Arrhythmogenic mechanisms of QT prolonging drugs: is QT prolongation really the problem? J Electrocardiol. 2004;37 Suppl:15-24. Zabel M, Portnoy S, Franz MR. Electrocardiographic indexes of dispersion of ventricular repolarization - an isolated heart validation-study. J Am Coll Cardiol. 1995;25:746-752. Kors JA, van Herpen G, Van Bemmel JH. QT dispersion as an attribute of T-loop morphology. Circulation. 1999;99:1458-1463. van Oosterom A. Genesis of the T wave as based on an equivalent surface source model. J Electrocardiol. 2001;34 Suppl:217-227. Glancy JM,Weston PJ, Bhullar HK, Garratt CJ,Woods KL, de Bono DP. Reproducibility and automatic measurement of QT dispersion. Eur Heart J. 1996;17:1035-1039. van Oosterom A. Singular value decomposition of the T wave: its link with a biophysical model of repolarization. International Journal of Bioelectromagnetism. 2002;4:59-60. Hooft van Huysduynen B, Swenne CA, Draisma HHM, Antoni ML, Van de

Increased dispersion of ventricular repolarization during recovery from exercise

36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

49.

50.

Vooren H, Van der Wall EE, Schalij MJ. Validation of ECG indices of ventricular repolarization heterogeneity: a computer simulation study. J Cardiovasc Electrophysiol. 2005;16:1097-1103. Paul JS, Reddy MR, Kumar VJ. A transform domain SVD filter for suppression of muscle noise artefacts in exercise ECG’s. IEEE Trans Biomed Eng. 2000;47:654663. Wilson FN, MacLeod AG, Barker PS, Johnston FD. The determination and significance of the areas of the ventricular deflections of the electrocardiogram. Am Heart J. 1934;10:46-61. Wilson FN, MacLeod AG, Barker PS. The T deflection of the electrocardiogram. Trans Assn Am Physicians. 1931;46. Geselowitz DB. The ventricular gradient revisited: relation to the area under the action potential. IEEE Trans Biomed Eng. 1983;30:76-77. Hiraoka M, Ogawa S, Kodam I et al., editors. Reflections on T waves. 04; Singapore: World Scientific, 2005. Plonsey R. A contemporary view of the ventricular gradient of Wilson. J Electrocardiol. 1979;12:337-341. Burger HC. A theoretical elucidation of the notion ventricular gradient. Am Heart J. 1957;53:240-246. Dimsdale JE, Hartley LH, Guiney T, Ruskin JN, Greenblatt D. Postexercise peril. Plasma catecholamines and exercise. JAMA. 1984;251:630-632. Kannankeril PJ, Goldberger JJ. Parasympathetic effects on cardiac electrophysiology during exercise and recovery. Am J Physiol Heart Circ Physiol. 2002;282:H2091H2098. Inoue H,Zipes DP.Changes in atrial and ventricular refractoriness and in atrioventricular nodal conduction produced by combinations of vagal and sympathetic stimulation that result in a constant spontaneous sinus cycle length. Circ Res. 1987;60:942-951. Malfatto G, Zaza A, Schwartz PJ. Parasympathetic control of cycle length dependence of endocardial ventricular repolarisation in the intact feline heart during steady state conditions. Cardiovasc Res. 1993;27:823-827. Litovsky SH, Antzelevitch C. Differences in the electrophysiological response of canine ventricular subendocardium and subepicardium to acetylcholine and isoproterenol. A direct effect of acetylcholine in ventricular myocardium. Circ Res. 1990;67:615-627. Imai K, Sato H, Hori M, Kusuoka H, Ozaki H, Yokoyama H, Takeda H, Inoue M, Kamada T. Vagally mediated heart rate recovery after exercise is accelerated in athletes but blunted in patients with chronic heart failure. J Am Coll Cardiol. 1994;24:15291535. Kozhevnikov DO,Yamamoto K,Robotis D,Restivo M,El Sherif N.Electrophysiological mechanism of enhanced susceptibility of hypertrophied heart to acquired torsade de pointes arrhythmias: tridimensional mapping of activation and recovery patterns. Circulation. 2002;105:1128-1134. Paffenbarger RS, Jr., Hyde RT, Wing AL, Lee IM, Jung DL, Kampert JB. The association of changes in physical-activity level and other lifestyle characteristics with mortality among men. N Engl J Med. 1993;328:538-545.

103

104

Chapter 5 Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients is related to reduction of right ventricular volume

Bart Hooft van Huysduynen Alexander van Straten Cees A. Swenne Arie C. Maan Henk J. Ritsema van Eck Martin J. Schalij Ernst E. van der Wall Albert de Roos Mark G. Hazekamp Hubert W. Vliegen

European Heart Journal 2005; 26: 928-32

105

Chapter 5

ABSTRACT Background. Late after total correction, Fallot patients with a long QRS duration are prone to serious arrhythmias and sudden cardiac death. Pulmonary regurgitation is a common cause of right ventricular (RV) failure and QRS lengthening. We studied the effects of pulmonary valve replacement (PVR) on QRS duration and RV volume. Methods and Results. 26 consecutive Fallot patients were evaluated both preoperatively and 6-12 months postoperatively by cardiac magnetic resonance (CMR). In this study, we present the computer-assisted analysis of the standard 12 lead ECGs closest in time to the CMR studies. For the whole group, QRS duration shortened by 6 ± 8 ms from 151 ± 30 to 144 ± 29 ms (P = 0.002). QRS duration decreased in 18 of 26 patients by 10 ± 6 ms, from 152 ± 32 to 142 ± 31 ms. QRS duration remained constant or increased slightly in 8 of 26 patients by 3 ± 3 ms, from 148 ± 27 to 151 ± 25 ms. CMR showed a decrease in RV end-diastolic volume from 305 ± 87 to 210 ± 62 ml (P = 0.000004). QRS duration changes correlated with RV end-diastolic volume changes (r = 0.54, P = 0.01). Conclusions. Our study shows that PVR reduces QRS duration. The amount of QRS reduction is related to the success of the operation, as expressed by the reduction in RV end-diastolic volume.

106

Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients

INTRODUCTION Tetralogy of Fallot (TOF) is the most common cyanotic congenital abnormality1. Surgical total correction has resulted in an increasing number of patients that reach adulthood. However, late after total correction, TOF patients with a QRS duration > 180 ms are prone to ventricular tachycardia and sudden cardiac death2. An important causative factor of increased QRS duration is residual pulmonary valve regurgitation, which may lead to severe right ventricular (RV) dilatation and heart failure3;4. Pulmonary valve replacement (PVR) has been reported to stabilize the gradual progression of the QRS duration on the long run5. Additionally, risk stratification by QRS duration may be further refined by analysis of QRS dispersion and QT dispersion(6). The aim of our study was to evaluate the short-term effects of PVR on the QRS duration, QRS- and QT dispersion. We also examined whether changes in QRS duration were related to changes in RV volume as obtained with cardiac magnetic resonance (CMR).

107

Chapter 5

METHODS From 1997 to 2002, twenty-six (15 male / 11 female) consecutive TOF patients were evaluated with CMR preoperatively and 6-12 months postoperatively7. In the present study, we present the retrospective serial analysis of the standard 12 lead ECGs closest in time to the pre- and postoperative CMR studies. All patients were treated according to our routine clinical protocol. Patients Baseline patient characteristics and surgical procedures are summarized in Table 1. The median age at which initial total repair was performed was 5.0 years (interquartile range (IQR) 2.8 to 6.8 years). A transannular patch was applied in 10 patients. Previous to total repair, a palliative procedure had been performed in 11 patients. Major indications for PVR were pulmonary regurgitation in combination with RV dilatation and a reduced validity. Only 2 patients were in NYHA class I, but these patients had severely dilated RVs, defined as an increase in RV end-diastolic volume (EDV) more than twice the left ventricular EDV. Overall, 15 patients had severe pulmonary regurgitation and 11 patients had moderate pulmonary regurgitation. Severe RV dilatation was seen in 13 patients. Residual pulmonary valve regurgitation was corrected by PVR at a median age of 29.2 years (IQR 24.3 to 39.4 years). CMR Cardiac magnetic resonance (CMR) was performed on a 1.5 Tesla system (NT15 Gyroscan, Philips Medical Systems, Best, The Netherlands). The CMR protocol has been described previously7. In summary, a multiphase, ECG-triggered, multishot echoplanar gradient echo technique was used to acquire short axis images. Images were acquired during breath holds. Slice thickness was 10 mm with a 0.8 to 1.0 mm section gap. The flip angle was 30 degrees and echo time was 5 to 10 ms. Eighteen to 25 frames per cycle resulted in a temporal resolution of 22 to 35 ms. ECG For this study we used all ECGs of these patients that had been stored digitally. Such ECGs were recorded at a median of 7.8 months (IQR 12.4 to 2.5 months) before PVR and at a median of 14.3 months (IQR 3.8 to 20.1 months) after PVR. The

108

Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients

preoperative ECGs closest in time to the preoperative CMR studies were recorded at a median of 2.0 months (IQR 3.8 to -1.1 months) before CMR. The postoperative ECGs closest in time to the postoperative CMR studies were recorded at a median of 2.7 months (IQR –2.3 to 13.8 months) after CMR. All ECGs were standard 12 lead recordings with a sample frequency of 500 Hz. The ECGs were analyzed by our MATLAB computer program LEADS (Leiden ECG Analysis and Decomposition Software). LEADS first computed an averagbeat, to minimize noise. In this averaged beat, the beginning and end of the QRS complex were automatically detected. Finally, the observer, blinded to the patient data, corrected this interval if necessary. To facilitate easy identification of the first deflection in any lead (onset QRS) and the last sharp deflection in any lead (offset QRS), the 12 standard ECG leads were superimposed on the screen. By using the zoom function, the ECG could be magnified at will, which allowed for the most accurate crosshair-cursor measurement of the QRS duration. QRS- and QT dispersion were calculated as the longest minus the shortest interval in any of the 12 leads. The end of the T wave was defined as the moment of return to the baseline. If U waves were present, the end of the T wave was set at the T-U nadir. Statistical Analysis Two-sided paired t tests were used to compare pre- and postoperative data. P-values were Bonferroni corrected for multiple testing. Linear regression analysis was performed to assess the relation between the changes in QRS duration and changes in RV EDV. A probability value of P < 0.05 was considered significant.

109

Chapter 5

RESULTS A typical example of a pre- and postoperative ECG is shown in Figure 1.

Fig. 1. Example of the averaged beat of a patient before (panel A) and after (panel B) PVR. To make the difference in QRS duration clearly visible, the ECGs were plotted on a stretched time scale, with all 12 ECG leads superimposed. Lead V1, relevant for the end of the QRS complex, is plotted as a thick line. Start and end of the QRS complexes are indicated by vertical dashed lines. QRS duration shortened by 14 ms (from the right dashed line that marks the end of the QRS complex in the pre-operative ECG to the left dashed line that marks the end of the QRS complex in the post-operative ECG).

110

Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients

Twenty-four of 26 patients had a right bundle branch block pattern before and after PVR. For the whole group, QRS duration shortened by 6 ± 8 ms from 151 ± 30 to 144 ± 29 ms (P = 0.002, Table 1). QRS duration decreased in 18 of 26 patients by 10 ± 6 ms from 152 ± 32 to 142 ± 31 ms and remained constant or increased slightly in 8 of 26 patients by 3 ± 3 ms, from 148 ± 27 to 151 ± 25 ms. QRS- and QT dispersion did not change significantly, from 22 ± 14 to 23 ± 9 ms (P = 0.97) and from 47 ± 21 to 47 ± 20 ms (P = 0.99), respectively. RV EDV could be obtained in 20 patients both before and after PVR. (In 6 patients, CMR could not be obtained due to technical difficulties, the quality of 4 pre-operative and 2 post-operative CMRs appeared unsatisfactory at the moment of analysis.) CMR showed a RV EDV decrease from 305 ± 87 to 210 ± 62 ml (P = 0.000004). In patients with reduced QRS durations, RV EDV reduced from 325 ± 86 to 220 ± 69 ml (P = 0.00004) and in patients with constant or slightly increased QRS duration, RV EDV decreased from 253 ± 72 to 190 ± 42 ml (P = 0.03). These volume reductions, of 105 and 63 ml, respectively, tended to be larger in the group with reduced QRS duration, but this did not reach significance level (P = 0.08). QRS duration changes correlated with RV EDV changes (r = 0.54, P = 0.01, see Figure 2).

Fig.2. Correlation between changes in right ventricular end-diastolic volume and QRS duration.

111

Chapter 5

DISCUSSION Our study demonstrated a decrease in QRS duration after pulmonary valve replacement in patients with TOF. This reduction in QRS duration appeared to be related to the reduction of RV EDV. Until now, only a stabilization of QRS duration after PVR has been reported(5). The use of a computer-assisted ECG measurement technique allowed us to show that PVR actually reduced QRS duration in most patients. In addition, our results demonstrate a relationship between the structural improvement and the improvement of electrical function. To our knowledge, our study is the first that showed a decrease in QRS duration following PVR in TOF patients. Gatzoulis et al.6 showed that QRS- and QT dispersion could be used to refine risk stratification on top of QRS duration. Although according to present insights QT dispersion only indirectly estimates repolarization disturbances8, gross changes in deand repolarization may still be detected by QRS- and QT dispersion. However, in our patient group no changes in QRS- and QT dispersion were induced by PVR. Relation between QRS duration, RV dilatation and arrhythmias Gatzoulis et al. reported a QRS duration > 180 ms as a risk marker for ventricular arrhythmias and sudden cardiac death2. Other studies confirmed a relation between QRS duration and late arrhythmias9;10. This relation may be explained by common factors that contribute to both the increased QRS duration and the vulnerability to arrhythmias. A central role is probably played by ventricular dilatation. Dilatation of the right ventricle increases wall stress, which leads to fibrosis of the right ventricle11. Fibrotic areas form blockades and areas of slow conduction that facilitate re-entry tachycardias12-15. Furthermore, stretch is known to induce premature ventricular excitations, which may serve as an arrhythmogenic trigger16. Additionally, ventricular dilatation may increase QRS duration by increasing the distance that the electrical activation front has to travel in the right ventricle, as most of our patients had a right bundle branch block pattern. In previous studies17-19, a relation between RV EDV and QRS duration was found in a group of Fallot patients. The present study shows that this relation also holds for changes in RV EDV and QRS duration.

112

Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients

Surgery in TOF patients Total repair itself may contribute to the arrhythmogeneity. Scars made during the transventricular approach and applied patches may form anatomical blockades facilitating re-entry15. On the other hand, surgical resection of aneurysms and correction of ventricular septal defects could reduce the amount of potential contributors to arrhythmias. Pulmonary regurgitation is the predominant hemodynamic lesion in Fallot patients with ventricular tachycardias and sudden cardiac death3;20. However, the timing of PVR remains subject to debate: too late may cause irreversible damage to the RV, whereas too early may lead to multiple re-operations. Our study shows that in TOF patients with dilated RVs, PVR leads not only to mechanical but also electrical beneficial effects. Hopefully, re-operations might be prevented in the future by the use of percutaneous implantation of pulmonary valves21. Limitations As we had to restrict ourselves to digitally stored ECGs, there was a time lag between the ECGs and CMR studies (preoperative ECGs 2.0 months (IQR –1.1 to 3.8 months) before CMR and postoperative ECGs 2.7 months (IQR -2.3 to 13.8 months) after CMR). This imposes a limitation upon the conclusions that can be drawn from our study. However, we think that the observed effect of a reduction in QRS duration after PVR was weakened rather than strengthened by these time differences: relatively early preoperative ECGs may have rendered smaller QRS durations before PVR, whereas relatively late postoperative ECGs may have rendered larger QRS duration after PVR. Both effects may have reduced the observed changes in QRS duration after PVR. This study did not directly assess the effects of PVR on arrhythmias. However, as previous studies with longer follow-up of non-operated TOF patients have found a strong relation of QRS duration and arrhythmias, PVR resulting in reduced QRS duration is likely to protect against arrhythmias. Conclusion PVR reduces QRS duration, a risk marker for ventricular tachycardias and sudden death in TOF patients. The structural success of the operation, measured as the reduction of the RV EDV, is related to the reduction of QRS duration.

113

Chapter 5

REFERENCES 1. Shinebourne EA, Anderson RH. Paediatric cardiology, 2nd edn. London: Curchill Livingstone. 2002;1213-1250. 2. Gatzoulis MA, Till JA, Somerville J et al. Mechanoelectrical interaction in tetralogy of Fallot. QRS prolongation relates to right ventricular size and predicts malignant ventricular arrhythmias and sudden death. Circulation 1995;92:231-237. 3. Gatzoulis MA, Balaji S, Webber SA et al. Risk factors for arrhythmia and sudden cardiac death late after repair of tetralogy of Fallot: a multicentre study. Lancet 2000;356:975-981. 4. Davlouros PA, Kilner PJ, Hornung TS et al. Right ventricular function in adults with repaired tetralogy of Fallot assessed with cardiovascular magnetic resonance imaging: detrimental role of right ventricular outflow aneurysms or akinesia and adverse rightto-left ventricular interaction. J Am Coll Cardiol 2002;40:2044-2052. 5. Therrien J, Siu SC, Harris L et al. Impact of pulmonary valve replacement on arrhythmia propensity late after repair of tetralogy of Fallot. Circulation 2001;103:2489-2494. 6. Gatzoulis MA, Till JA, Redington AN. Depolarization-repolarization inhomogeneity after repair of tetralogy of Fallot. The substrate for malignant ventricular tachycardia? Circulation 1997;95:401-404. 7. Vliegen HW, van Straten A, de Roos A et al. Magnetic resonance imaging to assess the hemodynamic effects of pulmonary valve replacement in adults late after repair of tetralogy of fallot. Circulation 2002;106:1703-1707. 8. Kors JA, van Herpen G, van Bemmel JH. QT dispersion as an attribute of T-loop morphology. Circulation 1999;99:1458-1463. 9. Balaji S, Lau YR, Case CL et al. QRS prolongation is associated with inducible ventricular tachycardia after repair of tetralogy of Fallot. Am J Cardiol 1997;80:160163. 10. Berul CI, Hill SL, Geggel RL et al. Electrocardiographic markers of late sudden death risk in postoperative tetralogy of Fallot children. J Cardiovasc Electrophysiol 1997;8:1349-1356. 11. Janicki JS, Brower GL, Gardner JD et al. The dynamic interaction between matrix metalloproteinase activity and adverse myocardial remodeling. Heart Fail Rev 2004;9:33-42. 12. Deanfield J, McKenna W, Rowland E. Local abnormalities of right ventricular depolarization after repair of tetralogy of Fallot: a basis for ventricular arrhythmia. Am J Cardiol 1985;55:522-525. 13. Downar E, Harris L, Kimber S et al. Ventricular tachycardia after surgical repair of tetralogy of Fallot: results of intraoperative mapping studies. J Am Coll Cardiol 1992;20:648-655. 14. Horowitz LN, Vetter VL, Harken AH et al. Electrophysiologic characteristics of sustained ventricular tachycardia occurring after repair of tetralogy of fallot. Am J Cardiol 1980;46:446-452. 15. Misaki T, Tsubota M, Watanabe G et al. Surgical treatment of ventricular tachycardia after surgical repair of tetralogy of Fallot. Relation between intraoperative mapping and histological findings. Circulation 1994;90:264-271.

114

Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients 16. Hansen DE, Craig CS, Hondeghem LM. Stretch-induced arrhythmias in the isolated canine ventricle. Evidence for the importance of mechanoelectrical feedback. Circulation 1990;81:1094-1105. 17. Abd El Rahman MY, Abdul-Khaliq H, Vogel M et al. Relation between right ventricular enlargement, QRS duration, and right ventricular function in patients with tetralogy of Fallot and pulmonary regurgitation after surgical repair. Heart 2000;84:416-420. 18. Neffke JG, Tulevski II, Van der Wall EE et al. ECG determinants in adult patients with chronic right ventricular pressure overload caused by congenital heart disease: relation with plasma neurohormones and MRI parameters. Heart 2002;88:266-270. 19. Daliento L, Rizzoli G, Menti L et al. Accuracy of electrocardiographic and echocardiographic indices in predicting life threatening ventricular arrhythmias in patients operated for tetralogy of Fallot. Heart 1999;81:650-655. 20. Zahka KG, Horneffer PJ, Rowe SA et al. Long-term valvular function after total repair of tetralogy of Fallot. Relation to ventricular arrhythmias. Circulation 1988;78:III14III19. 21. Khambadkone S, Bonhoeffer P. Percutaneous implantation of pulmonary valves. Expert Rev Cardiovasc Ther 2003;1:541-548.

115

116

Chapter 6 Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

Bart Hooft van Huysduynen Ivo R. Henkens Cees A. Swenne Thomas Oosterhof Harmen H.M. Draisma Arie C. Maan Mark G. Hazekamp Albert de Roos Martin J. Schalij Ernst E. van der Wall Hubert W. Vliegen

Submitted

117

Chapter 6

ABSTRACT Background. Pulmonary valve regurgitation may cause right ventricular failure in adult patients with Fallot’s tetralogy. In these patients, prolonged depolarization and disturbed repolarization are associated with ventricular arrhythmias and sudden cardiac death. We assessed the effect of pulmonary valve replacement (PVR) on the repolarization of patients with tetralogy of Fallot. Methods. Thirty Fallot patients (age 32±9 years, 19 male) eligible for PVR were studied with cardiac magnetic resonance imaging (CMR) before and 6 months after PVR. Electrocardiograms obtained during initial and follow-up CMR were analyzed and occurrence of ventricular arrhythmias was studied. Results. Right ventricular end-diastolic volume (RV EDV) decreased from 322 ± 87 to 215 ± 57 ml after PVR (P < 0.0001). The spatial QRS-T angle normalized from 117 ± 34 to 100 ± 35°, P = 0.0004 (normal < 105°). QT dispersion and T-wave complexity did not change significantly. T-wave amplitude decreased from 376 ± 121 to 329 ± 100 µV (P = 0.01). T-wave area decreased from 43 ± 15 to 38 ± 13 µV·s (P = 0.02). Decreases in T-wave amplitude and –area were most prominent in the right precordial leads overlying the RV. Three patients had sustained ventricular arrhythmias and one patient died suddenly. All these patients had a QRS duration > 160 ms. No severe ventricular arrhythmias were found in patients with a RV EDV < 220 ml, QRS-T angle < 100°, QT dispersion < 60 ms or T-wave complexity < 0.30. Conclusion. Normal repolarization indices are associated with the absence of severe ventricular arrhythmias. PVR in Fallot patients with dilated right ventricles has a beneficial effect on electrocardiographic indices of repolarization heterogeneity.

118

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

INTRODUCTION The prognosis of patients with a tetralogy of Fallot has improved dramatically after introduction of complete surgical repair at young age. However, residual pulmonary regurgitation may cause right ventricular failure in Fallot patients in adulthood1;2. These patients are prone to develop ventricular arrhythmias and/or sudden cardiac death. This risk increases significantly when QRS duration is more than 180 ms3. Besides a prolonged depolarization, a disturbed repolarization may play a role in arrhythmogenesis as well. Repolarization disturbances are widely recognized as contributors to arrhythmias4;5 and QT dispersion has been shown to refine risk stratification for arrhythmias in Fallot patients6. The spatial angle between the QRS and T axes is an electrocardiographic index that comprises both depolarization and repolarization and has prognostic value in normal subjects and selected patient groups7-9. T-wave complexity is related to repolarization heterogeneity, which is a pro-arrhythmogenic factor10. T-wave amplitude and T-wave area are also measures of repolarization heterogeneity11,12. We have previously demonstrated that pulmonary valve replacement (PVR) reduces QRS duration and right ventricular end-diastolic volume13. In the present study we tested whether PVR also has beneficial effects on the repolarization and whether electrocardiographic indices of the repolarization are related to ventricular arrhythmias.

119

Chapter 6

METHODS Thirty Fallot patients were evaluated (19 male), the age of the patients at the initial surgical procedure was 5.7 ± 3.1 years. In 15 patients a transannular patch had been applied during the initial procedure. At PVR, their age was 31.8 ± 9.1 years. Indications for PVR were moderate to severe pulmonary regurgitation in combination with right ventricular dilatation. In addition to PVR, tricuspid regurgitation was corrected in 6 patients and residual ventricular septal defects were closed in 4 patients. CMR CMR was performed on a 1.5 Tesla scanner (NT15 Gyroscan, Philips, Best). Briefly, short axis images of the heart were acquired with a multiphase, ECG-triggered, multishot echoplanar gradient echo technique. Images were acquired during breath holds with a slice thickness of 10 mm and a 0.8 to 1.0 section gap. The flip angle was 30 degrees and echo time was 5 to 10 ms. Eighteen to 25 frames per cycle resulted in a temporal resolution of 22 to 35 ms14. ECG analysis ECGs were obtained before the initial CMR and during the follow-up procedure 6 months after surgery. The routinely-made 10-s ECGs, digitally stored (sampling rate 500 Hz, resolution 5 µV/bit) in our hospital ECG database, were imported into LEADS, a MATLAB (The MathWorks, Natick, MA, USA) computer program that was developed for research-oriented ECG analysis15. After QRS detection and fiducial point determination, the QRS-T complexes in the 10-s ECG were coherently averaged in order to minimize noise. Besides the standard 12-lead ECG representation of the averaged beat, a vectorcardiographic X-Y-Z representation and the magnitude of the heart vector were computed using the inverse Dower matrix16. Onset and end of QRS were computed in the vector magnitude signal by a threshold procedure and by determining the minimal vector size in between the QRS complex and the T wave, respectively. The default end-of-QRS instant was then manually adjusted to meet the Minnesota criteria17 for end-of-QRS determination (being the last J-point in any of the ECG leads, while in leads with two candidate J points the earliest J point is taken). End-of-T instant was set automatically in every lead at the intercept of the steepest tangent to the terminal limb of the T wave with the baseline.

120

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

QT dispersion was calculated as the longest minus the shortest QT interval in any lead. The spatial angle between the mean electrical axes of the QRS complex and the T wave was computed from the vectorcardiogram18. T-wave complexity was derived by means of singular value decomposition of the 8 independent ECG leads I, II and V1-V610,19. T-wave complexity was calculated by dividing the square root of the summed squared singular values 2-8 by the first singular value. Finally, the absolute T-wave amplitude and T-wave area were computed and averaged from the ECG leads. Ventricular arrhythmias The occurrence of ventricular arrhythmias and the relation to ECG and CMR measurements was studied before and after PVR. Pre- and postoperative sustained ventricular tachycardias (lasting > 30 seconds or causing symptoms) and sudden cardiac death were categorized as severe ventricular arrhythmias. Coincidentally recorded non-sustained ventricular tachycardias were excluded from the arrhythmia analysis. Statistical analysis All data are reported as mean ± standard deviation. Two-sided paired and unpaired Student’s t-tests were used wherever appropriate. To correct for multiple testing, the significance level of the P-values was determined according to the false discovery rate method20.

121

Chapter 6

RESULTS Changes in right ventricular volume and QRS duration Surgery had a positive effect on right ventricular end-diastolic volume (RV EDV) which decreased from 322 ± 87 ml before surgery to 215 ± 57 ml after surgery (P < 0.0001). QRS duration decreased from 158 ± 34 ms to 153 ± 32 ms (P = 0.002). Changes in repolarization The spatial angle between QRS and T axes decreased significantly from a preoperative value of 117 ± 34º to 100 ± 35º postoperatively (P = 0.0004). QT dispersion did not change significantly, with a preoperative value of 78 ± 27 ms and a postoperative value of 85 ± 30 ms (P = 0.19). Pre- and postoperative values of T-wave complexity were 0.49 ± 0.22 and 0.44 ± 0.21, respectively (P = 0.16). T-wave amplitude decreased significantly from 376 ± 121 µV to 329 ± 100 µV (P = 0.01). T-wave area decreased significantly from 43 ± 15 µV ·s to 38 ± 13 µV ·s (P = 0.02). These results and the cut-off values for significance according to the false discovery rate method20 are summarized in Table 1. Changes in T-wave amplitude and area for all 12 leads are shown in Figures 1 and 2. Note that changes are most pronounced in the leads overlying the right ventricle. Correlations between electrocardiography and right ventricular volume The average of the pre- and postoperative QRS duration related linearly to the average RV EDV (R = 0.58, P = 0.003). Changes in QRS duration correlated with changes in RV EDV (R = 0.45, P = 0.03). Average QT dispersion was also related to RV EDV (R = 0.44, P = 0.03). Correlations between QRS duration and repolarization indices Average QRS durations correlated with the spatial angle (R = 0.70, P < 0.0001), QT dispersion (R = 0.62, P = 0.0003), T-wave complexity (R = 0.52, P = 0.003) and Twave area (R = 0.43, P = 0.02).

122

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

Figure 1. T-wave amplitude pre- and post PVR for all leads. Right precordial leads show the largest changes in T-wave amplitude.

Figure 2. T-wave area pre- and post PVR for all leads. The changes in T-wave area were most significant in leads V2 and V3, leads overlying the right ventricle.

Correlations between electrocardiographic repolarization indices Average QT dispersion and T-wave complexity were both related to the spatial QRST angle (R = 0.48, P = 0.007 and R = 0.37, P = 0.04). Changes in T-wave amplitude were linearly related to the changes in T-wave area (R = 0.68, P < 0.0001). Average T-wave area was strongly related to the T-wave amplitude (R = 0.93, P < 0.0001).

123

Chapter 6

Ventricular arrhythmias Follow-up was available up to 5.5 ± 1.9 years after PVR. Three patients had sustained ventricular tachycardias and one patient died suddenly. This patient died 18 months post PVR. The cause of death was uncertain, but the patient was hemodynamically stable and had no co-morbidity, making arrhythmia the most probable cause of death. Two of the patients had pre- as well as postoperative ventricular tachycardias and (pre)syncope, for which automatic internal cardiac defibrillators were implanted postoperatively. The last patient had preoperative repetitive sustained ventricular tachycardias, requiring cardioversion and hospitalization. After PVR this patient remained arrhythmia free. All patients with severe ventricular arrhythmias had a QRS duration > 160 ms. Among the 23 patients without any ventricular tachycardias, nine patients also had a pre- or postoperative QRS duration > 160 ms. The group size was too small to analyze whether the combination of QRS duration and a repolarization measure or RV EDV could improve the specificity. However, no severe arrhythmias were found in patients with QRS-T angle < 100°, QT dispersion < 60 ms, T-wave complexity < 0.30 or a RV EDV < 220 ml.

pre PVR

post PVR

P-values

RV EDV (ml)

significance cut-off value

322 ± 87

215 ± 57

< 0.0001 *

< 0.007

QRS duration (ms)

158 ± 34

153 ± 32

0.002 *

< 0.021

QRS-T angle (º)

117 ± 34

100 ± 35

0.0004 *

< 0.014

QT dispersion (ms)

78 ± 27

85 ± 30

0.19

< 0.05

T-wave complexity (·)

0.49 ± 0.22

0.44 ± 0.21

0.16

< 0.04

T-wave amplitude (µV)

376 ± 121

329 ± 100

0.01 *

< 0.029

43 ± 15

38 ± 13

0.02 *

< 0.036

N = 30

T-wave area (µV ·s)

Table 1. Values of end diastolic volumes of the right ventricle, QRS duration and electrocardiographic repolarization indices, measured before and after pulmonary valve replacement. * significant P-values after false discovery rate correction i.e., below the cut-off values provided in the last column. PVR = pulmonary valve replacement; RV EDV = right ventricular end-diastolic volume.

124

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

DISCUSSION In this study we assessed the effects of pulmonary valve replacement in Fallot patients with dilated right ventricles on electrocardiographic indices of repolarization heterogeneity. Furthermore, we analyzed the occurrence of ventricular arrhythmias in these patients. We found that PVR not only has an effect on QRS duration but also on repolarization indices. PVR results in normalization of the spatial QRS-T angle and reduction of T-wave amplitude and area. Most electrocardiographic repolarization indices are related to the QRS duration, which in turn is related to RV EDV. The optimal discriminator of patients with severe arrhythmias was a QRS > 160 ms. No severe arrhythmias were found in patients with QRS-T angle < 100°, QT dispersion < 60 ms, T-wave complexity < 0.30 or a RV EDV < 220 ml. In previous studies in Fallot patients late after initial surgical correction, repolarization heterogeneity was implicated as a potential mechanism for arrhythmias (6,21,22). In the current study, we used a dedicated computer program to enhance the reproducibility and accuracy of ECG analysis. This allowed concomitant calculation of electrocardiographic repolarization indices like the QRS-T angle, T-wave complexity, T-wave amplitude and T-wave area. Spatial QRS-T angle The spatial QRS-T angle comprises properties of both depolarization and repolarization and has prognostic capabilities. Kardys et al. showed that a wide QRS-T angle predicted cardiac death in a general population of more than 6000 men and women older than 55 years7. After adjustment for cardiovascular risk factors, hazard ratios of abnormal QRS-T angles for sudden death were 4.6 (CI 2.5-8.5). Zabel et al. showed that the QRS-T angle contributed to the risk stratification of patients after myocardial infarction, independent of classical risk factors8. Other studies underscored the prognostic value of the spatial QRS-T angle and the orientation of the T axis9;23;24. The large QRS-T spatial angle in our study is related to the right bundle branch block present in most Fallot patients. Their right ventricles are mostly activated by relatively slow myocardial cell-to-cell conduction instead of by the specialized conduction system. The differences in action potential duration that normally determine the order of repolarization, are exceeded by larger differences in repolarization time

125

Chapter 6

introduced by the slow cell-to-cell depolarization. Subsequently, the areas that depolarize last also repolarize last. This similar order of de- and repolarization in combination with the opposed direction of the de- and repolarizing currents cause large differences in the orientation of the QRS and T vectors, i.e., a wide QRS-T angle. An increased QRS-T angle may also be caused by a disturbance in the distribution of myocardial action potential durations. Previous studies showed that increased wall stress and hypertrophy have a direct influence on action potential duration25;26. In dogs, volume overload led to eccentric hypertrophy, interventricular differences in action potential durations and an increased sensitivity to arrhythmogenic medication26. Normal values for the QRS-T angle were defined as being smaller than 105°7,27. The QRS-T angle in our Fallot patients decreased from 117 ± 34º to 100 ± 35º after PVR, denoting a transition from a value outside the normal range to a smaller value within the normal range. We observed no severe arrhythmias in patients with a QRST angle < 100°. QT dispersion Previously, Gatzoulis et al. used QT dispersion to refine risk stratification of Fallot patients with a wide QRS complex6. All patients with clinically relevant arrhythmias appeared to have a QRS duration of more than 180 ms and a QT dispersion of more than 60 ms. In our patient group the combination of QRS duration and QT dispersion could not be assessed as only four patients had severe arrhythmias. However, we found no ventricular tachycardias in patients with QT dispersion < 60 ms. Furthermore, our group of Fallot patients with large RVs had relatively high pre- and postoperative QT dispersion values. Surprisingly, we found no change in QT dispersion after PVR, despite the relatively large right ventricular volume reduction in most patients. This finding is in agreement with the study of Helbing et al. (28), who did not find a correlation between right ventricular volume and QT dispersion in a group of Fallot patients and normal subjects. Initially, QT dispersion was proposed as a measure of local repolarization differences29. However, QT dispersion is strongly dependent on the orientation of the T vector, which represents the summed electromotive forces30. The QT interval is shortest in the ECG lead that is perpendicular to the orientation of the last part of the T vector. This shortest QT interval has a large influence on the magnitude of the QT dispersion, calculated as the longest minus the shortest QT interval in any lead. 126

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

Thus, QT dispersion depends on projections of the global T vector on the different lead vectors and does not necessarily represent local repolarization differences30. T-wave complexity T-wave complexity has been shown to yield independent prognostic information in patients with cardiovascular disease31. In patients with arrhythmogenic right ventricular dysplasia, higher T-wave complexity is associated with ventricular arrhythmias32. Additionally, T-wave complexity is increased in patients with primary repolarization disturbances and can be used to discriminate these patients from healthy individuals10. We calculated T-wave complexity by means of singular value decomposition, which is an algebraic algorithm used to reconstruct the T waves of eight independent ECG leads (I, II, V1-6). If the T waves can be described by only the first few singular values, the T waves have a relatively simple shape and are similar to each other in the different leads. The more singular values are needed to accurately describe the T waves and thus contain a significant amount of information, the more complex the T waves. We observed a nonsignificant reduction in T-wave complexity in our relatively small study, which can be interpreted as a trend in the direction of a more normal repolarization. Additionally, patients with a T-wave complexity < 0.30 had no severe arrhythmias. T-wave amplitude and area T-wave amplitude and area were related to repolarization heterogeneity in previous studies. In rabbit hearts, T-wave area was strongly correlated to repolarization heterogeneity as measured by 7 monophasic action potential electrodes33. T-wave amplitude and area were also related to heterogeneity, measured as the difference in repolarization time between the left and right ventricle in canine hearts12. Experiments in preparations of the left ventricular wall mimicked Long QT 1 syndrome and increased repolarization heterogeneity, which was reflected in an increased Twave amplitude and area34. Mathematical simulation studies confirmed these experimental findings11,35. In our Fallot patients we measured a decrease in T-wave amplitude and T-wave area after PVR, suggesting a decreased repolarization heterogeneity. The changes in T-wave area and amplitude were more explicit in leads V2 and V3 than in the other ECG leads (Figures 1 and 2). Leads V2 and V3 may display more electrical activity 127

Chapter 6

from the right ventricle than the other standard ECG leads due to their proximity to the right ventricle36;37 underscoring that the observed changes in T-wave amplitude and area were indeed related to changes in the right ventricle. The observed changes in T-wave amplitude and T-wave area suggest decreased repolarization heterogeneity in the right ventricle due to PVR. Arrhythmias The patients with severe arrhythmias had a QRS duration longer than 160 ms. Gatzoulis et al. found that every Fallot patient with symptomatic ventricular arrhythmias had a QRS duration longer than 180 ms3. Our study suggests that this criterion should be lowered to ascertain identification of patients with ventricular arrhythmias. The patient who died suddenly had a QRS duration of 164 ms. Our study was too small to combine electrocardiographic indices of the repolarization with the QRS duration to improve specificity. However, patients with either a QRS-T angle lower than 100°, a QT dispersion lower than 60 ms, a T-wave complexity lower than 0.30 or a RV EDV lower than 220 ml had no severe arrhythmias. The observed changes in electrocardiographic indices of repolarization heterogeneity suggest a decreased repolarization heterogeneity. Repolarization heterogeneity may form the substrate for ventricular arrhythmias. Irregular repolarization sequences facilitate the formation of functional barriers, so that an adversely timed extrastimulus may initiate a re-entry arrhythmia5. Furthermore, we found that the spatial QRS-T angle, QT dispersion, T-wave complexity and T-wave area were linearly related to the average QRS duration. A previous study in Fallot patients after total surgical correction registered body surface maps that showed a high similarity between de- and repolarization patterns38. Hence, in our Fallot patients the depolarization process probably influenced the repolarization as well. We could ask ourselves what kind of repolarization heterogeneity is introduced by cell-to-cell depolarization of the right ventricle in our Fallot patients. Repolarization differences caused by and similar to a slowly, but smoothly progressing depolarization wavefront do not necessarily increase the susceptibility to arrhythmias. However, in volume overloaded ventricles, fibrosis and mechanically induced changes in conduction velocity may cause patchy, irregular repolarization sequences.

128

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

Conclusions Pulmonary valve replacement in Fallot patients with dilated right ventricles has a beneficial effect on electrocardiographic indices of repolarization heterogeneity. Normal repolarization indices are associated with the absence of severe ventricular arrhythmias.

129

Chapter 6

REFERENCES 1 Carvalho JS, Shinebourne EA, Busst C, Rigby ML, Redington AN. Exercise capacity after complete repair of tetralogy of Fallot: deleterious effects of residual pulmonary regurgitation. Br Heart J 1992;67:470-3. 2 Gatzoulis MA, Balaji S, Webber SA et al. Risk factors for arrhythmia and sudden cardiac death late after repair of tetralogy of Fallot: a multicentre study. Lancet 2000;356:975-81. 3 Gatzoulis MA, Till JA, Somerville J, Redington AN. Mechanoelectrical interaction in tetralogy of Fallot. QRS prolongation relates to right ventricular size and predicts malignant ventricular arrhythmias and sudden death. Circulation 1995;92:231-7. 4 Han J, Moe GK. Nonuniform recovery of excitability in ventricular muscle. Circ Res 1964;14:44-60. 5 Kuo CS, Munakata K, Reddy CP, Surawicz B. Characteristics and possible mechanism of ventricular arrhythmia dependent on the dispersion of action potential durations. Circulation 1983;67:1356-67. 6 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. 7 Kardys I, Kors JA, van der Meer I, Hofman A, van der Kuip DA, Witteman JC. Spatial QRS-T angle predicts cardiac death in a general population. Eur Heart J 2003;24:1357-64. 8 Zabel M, Acar B, Klingenheben T, Franz MR, Hohnloser SH, Malik M. Analysis of 12-lead T-wave morphology for risk stratification after myocardial infarction. Circulation 2000;102:1252-7. 9 de Torbal A, Kors JA, van Herpen G et al. The electrical T-axis and the spatial QRST ngle are independent predictors of long-term mortality in patients admitted with acute ischemic chest pain. Cardiology 2004;101:199-207. 10 Priori SG, Mortara DW, Napolitano C et al. Evaluation of the spatial aspects of Twave complexity in the long-QT syndrome. Circulation 1997;96:3006-12. 11 Hooft van Huysduynen B, Swenne CA, Draisma HH et al. Validation of ECG indices of ventricular repolarization heterogeneity: a computer simulation study. J Cardiovasc Electrophysiol 2005;16:1097-103. 12 van Opstal JM, Verduyn SC, Winckels SK et al. The JT-area indicates dispersion of repolarization in dogs with atrioventricular block. J Interv Card Electrophysiol 2002;6:113-20. 13 Hooft van Huysduynen B, van Straten A, Swenne CA et al. Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients is related to reduction of right ventricular volume. Eur Heart J 2005;26:928-32. 14 Vliegen HW, van Straten A, de Roos A et al. Magnetic resonance imaging to assess the hemodynamic effects of pulmonary valve replacement in adults late after repair of tetralogy of fallot. Circulation 2002;106:1703-7. 15 Draisma HHM, Swenne CA, Van de Vooren H, Maan AC, Hooft van Huysduynen B, Van der Wall EE et al. LEADS: an interactive research oriented ECG/VCG analysis system. Computers in Cardiology In press.

130

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization 16 Edenbrandt L, Pahlm O. Vectorcardiogram synthesized from a 12-lead ECG: superiority of the inverse Dower matrix. J Electrocardiol 1988;21:361-7. 17 Prineas RJ, Crow RS, Blackburn H. Intraventricular conduction defects. In: The Minnesota code manual of electrocardiographic findings: standards and procedures for measurement and classification. Boston, MA: Wright, 1982:111-30 18 Mirvis DM, Goldberger AL. Electrocardiography. In: Zipes DP, editor. Braunwald’s heart disease: a textbook of cadiovascular medicine. 7th ed. Philadelphia, PA: Elsevier Saunders, 2005:117 19 Lay DC. Symmetric matrices and quadratic forms. In: Linear algebra and its applications. 2nd ed. Reading, MA: Addison-Wesley, 2003:441-486. 20 Benjamini and Hochberg. Controlling the false discovery rate-a practical and powerful approach to multiple testing. J R Stat Soc Ser 1995;B57:289-300 21 Berul CI, Hill SL, Geggel RL et al. Electrocardiographic markers of late sudden death risk in postoperative tetralogy of Fallot children. J Cardiovasc Electrophysiol 1997;8:1349-56. 22 Sarubbi B, Pacileo G, Ducceschi V et al. Arrhythmogenic substrate in young patients with repaired tetralogy of Fallot: role of an abnormal ventricular repolarization. Int J Cardiol 1999;72:73-82. 23 Kors JA, de Bruyne MC, Hoes AW et al. T axis as an indicator of risk of cardiac events in elderly people. Lancet 1998;352:601-5. 24 Rautaharju PM, Nelson JC, Kronmal RA et al. Usefulness of T-axis deviation as an independent risk indicator for incident cardiac events in older men and women free from coronary heart disease (the Cardiovascular Health Study). Am J Cardiol 2001;88:118-23. 25 Dean JW, Lab MJ. Arrhythmia in heart failure: role of mechanically induced changes in electrophysiology. Lancet 1989;1:1309-12. 26 Vos MA, De Groot SH, Verduyn SC et al. Enhanced susceptibility for acquired torsade de pointes arrhythmias in the dog with chronic, complete AV block is related to cardiac hypertrophy and electrical remodeling. Circulation 1998;98:1125-35. 27 Draper HW, Peffer CJ, Stallmann FW, Littmann D, Pipberger HV. The corrected orthogonal electrocardiogram and vectorcardiogram in 510 normal men (Frank lead system). Circulation 1964;30:853-64. 28 Helbing WA, Roest AA, Niezen RA et al. ECG predictors of ventricular arrhythmias and biventricular size and wall mass in tetralogy of Fallot with pulmonary regurgitation. Heart 2002;88:515-9. 29 Day CP, McComb JM, Campbell RW. QT dispersion: an indication of arrhythmia risk in patients with long QT intervals. Br Heart J 1990;63:342-4. 30 Kors JA, van Herpen G, van Bemmel JH. QT dispersion as an attribute of T-loop morphology. Circulation 1999;99:1458-63. 31 Zabel M, Malik M, Hnatkova K et al. Analysis of T-wave morphology from the 12lead electrocardiogram for prediction of long-term prognosis in male US veterans. Circulation 2002;105:1066-70. 32 De Ambroggi L, Aime E, Ceriotti C, Rovida M, Negroni S. Mapping of ventricular repolarization potentials in patients with arrhythmogenic right ventricular dysplasia: principal component analysis of the ST-T waves. Circulation 1997;96:4314-8. 33 Zabel M, Portnoy S, Franz MR. Electrocardiographic indexes of dispersion of

131

Chapter 6

34

35 36 37 38

132

ventricular repolarization: an isolated heart validation study. J Am Coll Cardiol 1995;25:746-52. Shimizu W, Antzelevitch C. Cellular basis for the ECG features of the LQT1 form of the long-QT syndrome: effects of beta-adrenergic agonists and antagonists and sodium channel blockers on transmural dispersion of repolarization and torsade de pointes. Circulation 1998;98:2314-22. di Bernardo D, Murray A. Explaining the T-wave shape in the ECG. Nature 2000;403:40. Shimizu W, Aiba T, Kurita T, Kamakura S. Paradoxic abbreviation of repolarization in epicardium of the right ventricular outflow tract during augmentation of Brugadatype ST segment elevation. J Cardiovasc Electrophysiol 2001;12:1418-21. Blomstrom-Lundqvist C, Hirsch I, Olsson SB, Edvardsson N. Quantitative analysis of the signal-averaged QRS in patients with arrhythmogenic right ventricular dysplasia. Eur Heart J 1988;9:301-12. Liebman J, Rudy Y, Diaz P, Thomas CW, Plonsey R. The spectrum of right bundle branch block as manifested in electrocardiographic body surface potential maps. J Electrocardiol 1984;17:329-46.

Pulmonary valve replacement in tetralogy of Fallot improves the repolarization

133

134

Chapter 7 Dispersion of the repolarization in cardiac resynchronization therapy

Bart Hooft van Huysduynen Cees A. Swenne Jeroen J. Bax Gabe B. Bleeker Harmen H.M. Draisma Lieselot van Erven Sander G. Molhoek Hedde van de Vooren Ernst E. van der Wall Martin J. Schalij

Heart Rhythm 2005; 2: 1286-93

135

Chapter 7

ABSTRACT Background. Pro-arrhythmic effects of cardiac resynchronization therapy (CRT) have been described in a subset of vulnerable patients as a result of increased transmural dispersion of repolarization (TDR) induced by left ventricular (LV) epicardial pacing. The possibility of identifying these patients by repolarization indices on the electrocardiogram has been suggested. The purpose of this study was to test whether repolarization indices on the ECG can be used to measure dispersion of the repolarization during pacing. Methods. CRT devices of 28 heart failure patients were switched among BIV, LV and right ventricular (RV) pacing. Electrocardiographic indices proposed to measure dispersion of the repolarization were calculated. The effects of CRT on repolarization were simulated in ECGSIM, a mathematical model of electrocardiogram genesis. TDR was calculated as the difference in repolarization time between the epi- and endocardial nodes of the heart model. Results. Patients. The interval from the apex to the end of the T wave was shorter during BIV pacing (102 ± 18 ms) and LV pacing (106 ± 21 ms) than during RV pacing (117 ± 22 ms, P ≤ 0.005). T-wave amplitude and area were low during BIV pacing (287 ± 125 µV and 56 ± 22 µV·s, respectively, P = 0.0006 vs. RV pacing), Twave complexity was high during BIV pacing (0.42 ± 0.26, P = 0.004 vs. RV pacing). Simulations. Repolarization patterns were highly similar to the preceding depolarization patterns. The repolarization patterns of different pacing modes explained the observed magnitudes of the electrocardiographic repolarization indices. Average and local TDR were not different between pacing modes. Conclusions. In patients treated with CRT, electrocardiographic repolarization indices are related to the pacing-induced activation sequences rather than to transmural dispersion. TDR during BIV and LV pacing is not larger than during conventional RV endocardial pacing.

136

INTRODUCTION Cardiac resynchronization therapy (CRT) restores coordinated mechanical contraction in most heart failure patients with compromised intraventricular conduction1;2. Application of CRT reduces the occurrence of heart failure3 and all cause mortality4. However, CRT may have pro-arrhythmic effects in a subset of vulnerable patients due to increased dispersion of the repolarization5;6. Left ventricular epicardial pacing reverses the natural activation from endo- to epicardium. This advances the repolarization time of the already short epicardial action potentials, thereby increasing repolarization time differences with the longer underlying action potentials of the midmyocardial and endocardial layers. The thus increased transmural dispersion of repolarization (TDR) may enhance the susceptibility to arrhythmias. It would be of great clinical value if vulnerable patients could be identified by a noninvasive repolarization measure on the electrocardiogram (ECG) such as the interval from the apex to the end of the T wave (Tapex-end) 7. The Tapex-end interval as a measure of dispersion of the repolarization was proposed on the basis of experiments in a wedge preparation of the left ventricular wall8. Although being a valuable research tool, the wedge preparation represents only part of the heart and the pseudo-ECG derived from this preparation is probably not representative for clinical ECGs. Other proposed ECG indices of dispersion of the repolarization are also difficult to verify in vivo, as in most studies either epi- or endocardial repolarization is recorded from a limited number of electrodes on the heart9;10. In the current study we used a whole heart model of ECG simulation, ECGSIM11 to study whether epicardial pacing in CRT increases TDR. Moreover, we sought to clarify how the repolarization process during CRT is reflected in ECG indices proposed to measure dispersion of the repolarization.

137

Chapter 7

METHODS Patients Twenty-eight heart failure patients were studied two days after implantation of a biventricular pacemaker at the Leiden University Medical Center. The characteristics of these patients are listed in Table 1. After the patients had rested supine for five minutes, standard 12-lead ECGs were recorded continuously while the pacemaker was programmed in random order in biventricular (BIV), left ventricular (LV), and right ventricular (RV) mode, and turned off, for three minutes per mode. The last minute of these ECG recordings was subsequently analyzed. ECG characteristics

SR

RV pacing

LV pacing

BIV pacing

Heart rate (bpm)

73 ± 11

72 ± 14

72 ± 12

73 ± 11

QRS duration (ms)

175 ± 31

191 ± 31

167 ± 36

146 ± 17

QTc (ms)

503 ± 47

512 ± 45

487 ± 31

477 ± 44

Tapex-endc (ms)

108 ± 27

117 ± 22

106 ± 21

102 ± 18

Tamplitude (µV)

362 ± 219

478 ± 174

335 ± 143

287 ± 125

Tarea (µV.s)

66 ± 31

94 ± 35

62 ± 26

56 ± 22

Tcomplexity (.)

0.30 ± 0.13

0.25 ± 0.10

0.36 ± 0.19

0.42 ± 0.26

QRS-T angle (°)

162 ± 20

167 ± 9

149 ± 28

154 ± 21

Table 1. Patient characteristics.

ECG analysis was performed by LEADS (Leiden ECG Analysis and Decomposition Software), our MATLAB (The MathWorks, Natick, USA) program for research oriented ECG analysis. LEADS first computed an averaged beat in order to minimize noise. In this averaged beat, the beginning and end of the QRS complex were automatically detected. An observer, who was blinded to the patient data, corrected this interval if necessary. LEADS identified the apex and end of the T wave in each ECG lead. The end of the T wave was set at the point where the tangent to the steepest portion of the terminal part of the T wave crossed the isoelectric line. Subsequently, LEADS searched backward for the apex of the T wave, defined as the 138

Dispersion of the repolarization in cardiac resynchronization therapy

point of the T wave with the highest amplitude. Finally, LEADS calculated a number of ECG indices proposed to assess dispersion of the repolarization: the Tapexend interval8, QT interval12, T-wave amplitude13, T-wave surface area14 and T-wave complexity15. QT and Tapex-end intervals were corrected for heart rate by the Bazett formula16. T-wave complexity was calculated by means of singular value decomposition of the T wave15;17. The higher, more complex, singular values 2 to 8 were divided by the first, most simple component to quantify T-wave complexity. Additionally, LEADS constructed the vectorcardiogram by using the inverse Dower matrix18;19 and calculated the spatial angle between the QRS and T axes. Simulations ECGSIM is an ECG simulation program that consists of a mathematical whole heart model placed in a human thorax11. The model geometry is based on magnetic resonance images of heart and thorax, containing conduction inhomogeneities such as the lungs. Surfaces of the left and right ventricle consist of 257 epicardial and endocardial nodes. A transmembrane action potential in each node represents the local electrical activity13. The naturally present TDR is incorporated in the model: the epicardium has shorter action potential durations than the endocardium, so that despite earlier endocardial activation the epicardium repolarizes earlier. The specific electrophysiological properties of all 257 nodes can be found after downloading the model from www.ecgsim.com (free of charge). With default parameter settings, ECGSIM generates an ECG that closely resembles the ECG of a healthy subject. For the current study, the activation sequences associated with pacing were based on previously published depolarization maps20-24. A left posterolateral node, node #79, was chosen as the location of the LV epicardial pacing site. The averaged QRS duration measured in the heart failure patients was adopted as the total time of ventricular depolarization. The moment of activation of each node was calculated by multiplying the total time of depolarization by the ratio of the distance of each node from the pacing site to the total distance the activation front needed to travel to activate the entire heart. The action potential durations of all nodes were left unaltered. The repolarization times of all nodes were calculated by addition of the action potential durations to the moments of activation. RV pacing was simulated likewise. The heart was activated from a RV apical endo139

Chapter 7

cardial node, node #249. BIV pacing was simulated by simultaneous pacing from nodes ## 79 and 249, thus combining the ventricular depolarization and repolarization times of LV and RV pacing. Thereafter, ECGSIM constructed isochrone maps of depolarization and repolarization times of the heart. The simulated ECGs were analyzed with the LEADS program as described above. For comparison with normal ventricular activation with an intact conduction system, we used the default settings of de- and repolarization times of ECGSIM, which generates a normal ECG. Subsequently, we calculated TDR for the different pacing modes and during normal activation. Each endocardial node was paired with an opposing epicardial node, and TDR was calculated as the difference in repolarization time between the endo- and epicardial node25. As there are more epicardial than endocardial nodes, several endocardial nodes had two opposing epicardial nodes. In these cases, the difference was calculated between the repolarization time of the endocardial node and the mean repolarization time of the two opposing epicardial nodes. Similarly, each LV septal node was paired with one or two RV septal nodes. Subsequently, all paired repolarization time differences were averaged to calculate TDR. Additionally, we measured local TDR directly under the LV epicardial pacing site during LV pacing and during normal ventricular activation. Similarly, local TDR in the septum directly under the RV endocardial pacing site was calculated. Statistical Analysis Repeated measures ANOVA was performed before post-hoc t-testing. Two-sided paired t tests were used to compare data between different pacing modes. P-values were Bonferroni corrected for multiple testing. A probability value of P < 0.05 was considered significant.

140

Dispersion of the repolarization in cardiac resynchronization therapy

RESULTS Patients Four ECGs recorded during sinus rhythm, RV pacing, LV pacing and BIV pacing are depicted in Figure 1, panels A-D. All ECGs were highly discordant: QRS- and T-wave polarities were opposite in almost all leads of every rhythm. Consequently, as shown in Table 2a, the spatial angles between the QRS and T axes (a gross qualifier of concordance / discordance) were close to 180°. The largest average spatial QRS-T angle was attained during RV pacing. Correlation coefficients of the spatial orientation of the QRS- and T-vectors were 0.96 for RV pacing, 0.79 for LV pacing and 0.85 for BIV pacing.

Figure 1. Sample ECGs (average beat) during sinus rhythm, right-, left-, and biventricular pacing. Panel A. Subject #08 in sinus rhythm,Panel B. Subject #26 during right ventricular pacing

141

Chapter 7

Figure 1. Sample ECGs (average beat) during sinus rhythm, right-, left-, and biventricular pacing. Panel C. Subject #19 during left ventricular pacing Panel D. Subject #07 during biventricular pacing.

ECG characteristics

SR

RV pacing

LV pacing

BIV pacing

Heart rate (bpm)

73 ± 11

72 ± 14

72 ± 12

73 ± 11

QRS duration (ms)

175 ± 31

191 ± 31

167 ± 36

146 ± 17

QTc (ms)

503 ± 47

512 ± 45

487 ± 31

477 ± 44

Tapex-endc (ms)

108 ± 27

117 ± 22

106 ± 21

102 ± 18

Tamplitude (µV)

362 ± 219

478 ± 174

335 ± 143

287 ± 125

Tarea (µV.s)

66 ± 31

94 ± 35

62 ± 26

56 ± 22

Tcomplexity (.)

0.30 ± 0.13

0.25 ± 0.10

0.36 ± 0.19

0.42 ± 0.26

QRS-T angle (°)

162 ± 20

167 ± 9

149 ± 28

154 ± 21

Table 2a. ECG characteristics of heart failure patients during different rhythms. Data are given in average ± standard deviation. SR = sinus rhythm; RV = right ventricular; LV = left ventricular; BIV = biventricular.

142

Dispersion of the repolarization in cardiac resynchronization therapy

ECG

RV vs LV

characteristics

LV vs BIV

pacing

BIV vs RV pacing

pacing

Heart rate (bpm)

1

1

1

0.002

0.004

< 0.0001

QTc (ms)

0.005

1

0.003

Tapex-endc (ms)

0.0009

1

0.005

Tamplitude (µV)

0.0006

0.6

< 0.0001

Tarea (µV.s)

0.0006

1

< 0.0001

Tcomplexity (.)

QRS-T angle (°)

0.048

1

0.004

0.02

1

0.008

QRS duration (ms)

Table 2b. P-values for all ECG differences between different pacing modes. RV = right ventricular; LV = left ventricular; BIV = biventricular.

Table 2a lists average heart rate, QRS duration and the values of the ECG indices of dispersion of the repolarization. QRS duration was longest during RV pacing and shortest during BIV pacing, LV pacing had intermediate values (see Table 2b for P-values). QTc and Tapex-endc intervals followed the same pattern with decreasing values from RV to LV and BIV pacing. The same applies to T-wave amplitude and area. The pattern present in the above-described ECG indices during different rhythms is depicted in Figure 2. T-wave complexity was large during BIV pacing and small during RV pacing. Simulations Depolarization and repolarization maps are depicted in Figure 3. The maps of the normal situation with intact conduction system are shown in panel A, in which deand repolarization patterns differ significantly. During RV and LV pacing (panels B and C) depolarization spreads from the RV apex and the LV lateral wall, respectively. Here, repolarization roughly follows the depolarization pattern. During BIV pacing (panel D), the heart is activated simultaneously from these two sites. Also here, the repolarization map resembles the depolarization map.

143

Chapter 7

Figure 2. Values of ECG indices during right ventricular (RV), left ventricular (LV) and biventricular (BIV) pacing relative to sinus rhythm. Values of ECG indices decreased in magnitude from RV-, to LV- to BIV pacing.

The corresponding simulated ECGs are depicted in Figure 4, panels A-D. During normal ventricular activation (panel A), the ECG is normal, hence highly concordant: the QRS complex and the T wave had a similar polarity in most leads. In all pacing modes the ECGs were highly discordant: the QRS complex and the T wave had opposite polarities in most leads. The spatial angle between the QRS and T axes was small during normal ventricular activation, and close to 180° during pacing.

144

Dispersion of the repolarization in cardiac resynchronization therapy

Figure 3. Simulated depolarization/repolarization maps. Hearts are depicted as viewed from antero-superior with the left anterior descending artery as landmark. Red indicates the earliest depolarization/repolarization and blue indicates the latest depolarization/repolarization. Panel A. Normal ventricular activation, intact conduction system. Panel B. Right ventricular endocardial pacing. Apparently, repolarization patterns are now very similar to depolarization patterns. Panel C. Left ventricular epicardial pacing. Repolarization patterns closely resemble the depolarization pattern. Panel D. Biventricular pacing. 145

Chapter 7

Figure 4. ECGs generated by ECGSIM. Panel A. Normal ventricular activation, intact conduction system. Panel B. Right ventricular pacing. Panel C. Left ventricular pacing. Panel D. Biventricular pacing

146

Dispersion of the repolarization in cardiac resynchronization therapy

Table 3 depicts the main characteristics of the simulated ECGs. The values of all ECG repolarization indices were smallest during normal ventricular activation. QRS duration was longest for RV pacing, smaller for LV pacing, while BIV pacing had the smallest paced QRS duration. QTc was longest for RV pacing and smaller during LV and BIV pacing. Tapex-endc interval was longest for RV pacing, while LV and BIV pacing had a smaller Tapex-endc interval. T-wave amplitude and area decreased from RV to LV to BIV pacing. T-wave complexity was largest during BIV pacing. QRS-T angle was largest during RV pacing and decreased from LV to BIV pacing; it remained however relatively large. normal

conduction

RV pacing

LV pacing

BIV pacing

HR (bpm)

60

60

60

60

QRS duration (ms)

100

184

172

166

QTc (ms)

376

450

401

400

Tapex-endc (ms)

79

104

93

94

Tamplitude (µV)

294

820

679

606

Tarea (µV.s)

36

115

91

80

0.13

0.23

0.19

0.36

69

168

151

147

Tcomplexity (.) QRS-T angle (°)

Table 3. ECG characteristics - simulations. RV = right ventricular; LV = left ventricular; BIV = biventricular.

In ECGSIM, TDR in the whole heart and in the LV free wall was larger during pacing than during normal ventricular activation with intact conduction system, but differences between pacing modes were small. TDR in the whole heart was largest for RV pacing, 29 ms, with slightly smaller values for LV and BIV pacing, 26 and 27 ms respectively, while TDR was 18 ms during normal conduction. TDR in the LV free wall showed the same pattern with slightly higher values; 36 ms for RV pacing, 31 ms for LV pacing and 32 ms for BIV pacing, with a TDR of 19 ms for a normally conducted beat. Local TDR directly under the LV pacing site increased from 18 ms during normal

147

Chapter 7

ventricular activation to 39 ms during LV epicardial pacing, as a result of the reversal of the naturally existing endo-to epicardial activation order. In the septum directly under the RV pacing site, local TDR increased from 20 to 75 ms during RV endocardial pacing.

148

Dispersion of the repolarization in cardiac resynchronization therapy

DISCUSSION Our simulations indicate that TDR during LV and BIV pacing was not larger than TDR during conventional RV endocardial pacing, neither in the whole heart nor directly under the pacing sites. In patients treated with CRT, ECG indices of dispersion of the repolarization were measured during different pacing modes. The observed magnitudes of these ECG indices can be explained by interpretation of our simulations. Principles of repolarization during pacing In the virtual situation that all myocardial cells would depolarize at the same time instant, only intrinsic differences in action potential duration would determine the repolarization sequence. During normal ventricular activation there is a mix of conduction system mediated rapid depolarization and slower cell-to-cell conduction. In such cases, the repolarization order is partly determined by the order in which the cells depolarize and partly by intrinsic action potential duration differences. In the paced heart, slow cell-to-cell conduction prevails20-24 and the repolarization order starts to closely resemble the depolarization order; a condition associated with secondary T-wave changes26. In our patients as well as in the paced simulations, the QRS complex and the T wave had opposite polarities and the spatial angle between QRS and T axes approximated 180°. Such discordant ECGs emerge when the repolarization sequence roughly equals the depolarization sequence; the opposite polarity is caused by the opposite direction of the repolarizing and depolarizing currents. This view is supported by the simulated ECGs (Figure 4) that strikingly resemble the ECGs that were recorded in patients (Figure 1). These ECGs could be generated by only changing the ventricular depolarization sequence, while leaving the action potential durations unaffected. The discordance between the QRS complex and the T wave during pacing are in contrast to the concordance found in the simulated ECGs during normal activation (Figure 4A). During normal activation, the epicardium depolarizes last but repolarizes first due to the relatively short epicardial action potential durations. In combination with the opposite direction of de- and repolarizing currents, a concordant QRS

149

Chapter 7

complex and T wave emerge. Transmural dispersion of the repolarization In accordance with the findings of Medina-Ravell6, ECGSIM showed an increased transmural repolarization gradient directly under the LV pacing site during LV pacing. However, during RV pacing, TDR in the septum directly under the RV apical pacing site was also increased. This is not entirely unexpected, because the endocardial RV lead is, like the epicardial LV lead, located outside the left ventricle. Average TDR in the whole heart model was approximately the same during all pacing modes and slightly increased compared to normal activation (29, 26, 27 and 18 ms during RV, LV, BIV pacing and normal activation, respectively). This can be explained on the basis of the repolarization maps (Figure 3). During pacing (panels B-D), the depolarization sequence is followed by a similarly shaped repolarization pattern that spreads from the pacing site through the ventricles. With increasing distance from the pacing site, the epi- to endocardial activation order, which accentuates local TDR, is lost. For example, observe the depolarization wavefront at the basal surface of the left ventricle during LV pacing (panel C). Major part of the LV base is depolarized by a wavefront that progresses in a direction approximately parallel to the ventricular walls. Such a depolarization order does not introduce a large TDR, as it does not advance the depolarization time of the epicardium as compared to the endocardium. Both LV epicardial and RV endocardial pacing create such transventricular de- and repolarization patterns. These patterns are, apart from their opposite direction, not essentially different from another. Although BIV pacing is performed from two pacing sites, similar transventricular repolarization patterns are induced in the greater part of the ventricles (panel D). Hence, TDR is similar during all pacing modes. However, the efficiency of an intact conduction system with well-synchronized endo- to epicardial activation throughout the heart is not attained by any pacing mode. Therefore, TDR is moderately larger during all pacing modes than during sinus rhythm. Tapex-end interval The Tapex-end interval is a measure of TDR in the wedge preparation of the LV free wall8. The apex of the T wave emerges when the transmural voltage gradients are maximal; i.e., when the epicardial action potentials have their steepest decline in 150

Dispersion of the repolarization in cardiac resynchronization therapy

amplitude. This was elegantly shown by Antzelevitch and co-workers8. Recently, the Tapex-end interval was reported to be increased in ECGs of heart failure patients during LV pacing. Based on the assumption that the laboratory situation (wedge) and the clinical situation (whole heart) are analogous, it was concluded that LV pacing increased TDR6. Our simulations suggest that other factors determine the Tapex-end interval during pacing of a whole heart. The occurrence of the apex of the T vector can be understood from the repolarization maps in Figure 3. During RV pacing (panel B) the repolarization wavefront increases in size when progressing from apex to base. The wavefront extends to its maximal area when it reaches the right ventricular base of the heart (t = 342 ms, green isochrone). At that moment the apex of the T vector emerges. During LV pacing (panel C), the repolarization wavefront has the largest area at 320 ms (green isochrone). However, at that moment repolarization has already progressed to the right ventricle before completion of the repolarization of the septum, which implies a certain amount of cancellation. This effect causes a relatively early occurrence (t = 300 ms, yellow isochrone) of the maximal T vector during LV pacing. The end of the T vector occurs when the repolarization wavefronts have traveled to the most distant areas of the heart. In case of RV pacing the final repolarization occurs at the left ventricular base. In case of LV pacing the epicardium of the right ventricular free wall is the last area to repolarize. Final repolarization in both RV and LV pacing occurs in parts of the heart that are relatively distant from the area that generates the maximal repolarization vector. Therefore, neither the Tapex-end interval during RV pacing nor the Tapex-end interval during LV pacing express transmural dispersion of the repolarization. Thus, during pacing the Tapex-end interval contains information of the global transventricular repolarization process, whereas local repolarization differences in adjacent regions, like TDR, are considered more arrhythmogenic27,28. T-wave amplitude Increase of T-wave amplitude has been related to increased dispersion of the repolarization by studies in biological and mathematical models. Increased TDR induced by hyperkalemia29 or experimental Long QT 1 conditions30 increases T-wave amplitude. Also, in ECGSIM experiments the T-wave amplitude increased with enhanced dispersion of the repolarization13;25. 151

Chapter 7

Hence, increased dispersion of the repolarization during regular, conduction-system mediated ventricular activation may cause high amplitude T waves. However, in the present study high amplitude T waves occurred during abnormal, cell-to-cell activation as a consequence of pacing. In the setting of pacing, TDR is relatively normal; the higher T-wave amplitude is here caused by less cancellation due to the asymmetric de- and repolarization sequence. Of the different pacing modes, BIV pacing caused the lowest T-wave amplitude in our simulations and patients, as this pacing mode involves more cancellation due to the opposite directions of the two simultaneously started wavefronts (Figure 3D). T-wave area T-wave area has been related to dispersion of the repolarization in several studies. In rabbit hearts, dispersion of epicardial monophasic action potential duration was accompanied by increased T-wave area on a simultaneously recorded surface ECG14. In dogs, T-wave31 and QRST32 surface area were related to dispersion of repolarization and a lowered threshold for ventricular fibrillation33. We measured the smallest T-wave area during BIV pacing as compared to other pacing modes, in the patients as well as in our simulations. This low T-wave area during BIV pacing resulted from cancellation due to the opposite directions of the two simultaneously started wavefronts and was not due to an increased TDR. T-wave complexity Increased T-wave complexity in patients with arrhythmogenic right ventricular dysplasia has been associated with arrhythmias34. Long-QT patients can be distinguished from healthy subjects by an increased T-wave complexity15. In U.S. veterans with cardiovascular disease, repolarization complexity calculated with singular value decomposition conferred independent prognostic information35. Van Oosterom showed mathematically that T-wave complexity was related to dispersion of the repolarization induced by accentuated action potential duration differences36. In our study, the largest T-wave complexity was found during BIV pacing in heart failure patients and in the simulations. This is caused by the more complex repolarization pattern as a result of two colliding activation wavefronts instead of one. However, as stated earlier, the transventricular repolarization patterns and TDR are not essentially different from single-sided LV or RV pacing. Therefore, increased T-wave 152

Dispersion of the repolarization in cardiac resynchronization therapy

complexity in BIV paced patients cannot be interpreted as an indication of increased arrhythmogeneity. Limitations Like every modelling study, the simulations with ECGSIM are a simplification of reality. The used model is not specifically designed as a heart failure model and lacks complex alterations in anatomy and electrophysiology present in varying degrees in heart failure patients. Recent mapping studies in heart failure patients during pacing and sinus rhythm have shown individually variable areas of slow conduction37 and lines of functional block38 resulting in complex depolarization patterns. However, also in these studies the largest areas were chiefly activated by slowly progressing wavefronts, suggesting myocardial cell-to-cell conduction, as we used in our simulations. We were able to simulate realistic ECGs in a model that has not been primarily designed for the current study. ECGSIM has been extensively evaluated and is sufficiently realistic to show the principles of repolarization and its effects on the ECG13;39. Conclusion In patients treated with CRT, electrocardiographic repolarization indices are related to the pacing-induced activation sequences rather than to transmural dispersion. TDR during BIV and LV pacing is not larger than during conventional RV endocardial pacing.

153

Chapter 7

REFERENCES 1. Abraham WT, Fisher WG, Smith AL, DeLurgio DB, Leon AR, Loh E, Kocovic

2.

3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 154

DZ, Packer M, Clavell AL, Hayes DL, Ellestad M, Trupp RJ, Underwood J, Pickering F, Truex C, McAtee P, Messenger J: Cardiac resynchronization in chronic heart failure. N Engl J Med. 2002;346:1845-53. John Sutton MG, Plappert T, Abraham WT, Smith AL, DeLurgio DB, Leon AR, Loh E, Kocovic DZ, Fisher WG, Ellestad M, Messenger J, Kruger K, Hilpisch KE, Hill MR: Effect of cardiac resynchronization therapy on left ventricular size and function in chronic heart failure. Circulation. 2003;107:1985-90. Bradley DJ, Bradley EA, Baughman KL, Berger RD, Calkins H, Goodman SN, Kass DA, Powe NR: Cardiac resynchronization and death from progressive heart failure: a meta-analysis of randomized controlled trials. JAMA. 2003;289:730-40. Cleland JGF, Daubert J, Erdmann E, Freemantle N, Gras D, Kappenberger L, Tavazzi L: The effect of cardiac resynchronization on morbidity and mortality in heart failure. New England Journal of Medicine. 2005;352:1539-49. Fish JM, Di Diego JM, Nesterenko V, Antzelevitch C: Epicardial activation of left ventricular wall prolongs QT interval and transmural dispersion of repolarization: implications for biventricular pacing. Circulation. 2004;109:2136-42. Medina-Ravell VA, Lankipalli RS, Yan GX, Antzelevitch C, Medina-Malpica NA, Medina-Malpica OA, Droogan C, Kowey PR: Effect of epicardial or biventricular pacing to prolong QT interval and increase transmural dispersion of repolarization: does resynchronization therapy pose a risk for patients predisposed to long QT or torsade de pointes? Circulation. 2003;107:740-6. Yan GX, Lankipalli RS, Burke JF, Musco S, Kowey PR: Ventricular repolarization components on the electrocardiogram: cellular basis and clinical significance. J Am Coll Cardiol. 2003;42:401-9. Yan GX, Antzelevitch C: Cellular basis for the normal T wave and the electrocardiographic manifestations of the long-QT syndrome. Circulation. 1998;98:1928-36. Cowan JC, Hilton CJ, Griffiths CJ, Tansuphaswadikul S, Bourke JP, Murray A, Campbell RW: Sequence of epicardial repolarisation and configuration of the T wave. Br Heart J. 1988;60:424-33. Franz MR, Bargheer K, Rafflenbeul W, Haverich A, Lichtlen PR: Monophasic action potential mapping in human subjects with normal electrocardiograms: direct evidence for the genesis of the T wave. Circulation. 1987;75:379-86. van Oosterom A, Oostendorp TF: ECGSIM: an interactive tool for studying the genesis of QRST waveforms. Heart. 2004;90:165-8. Moss AJ: Measurement of the QT interval and the risk associated with QTc interval prolongation: a review. Am J Cardiol. 1993;72:23-5. van Oosterom A: Genesis of the T wave as based on an equivalent surface source model. J Electrocardiol. 2001;34:S217-27. Zabel M, Portnoy S, Franz MR: Electrocardiographic indexes of dispersion of ventricular repolarization: an isolated heart validation study. J Am Coll Cardiol. 1995;25:746-52. Priori SG, Mortara DW, Napolitano C, Diehl L, Paganini V, Cantu F, Cantu G,

Dispersion of the repolarization in cardiac resynchronization therapy

16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

26. 27. 28. 29. 30.

31. 32.

Schwartz PJ: Evaluation of the spatial aspects of T-wave complexity in the long-QT syndrome. Circulation. 1997;96:3006-12. Bazett HC: An analysis of the time relation of electrocardiograms. Heart. 1920;7:353-67. Lay DC: Linear algebra and its applications. Second Edition. Reading, MA: Addison-Wesley, 2003, pp. 441-486. Dower GE, Machado HB, Osborne JA: On deriving the electrocardiogram from vectoradiographic leads. Clin Cardiol. 1980;3:87-95. Edenbrandt L, Pahlm O: Vectorcardiogram synthesized from a 12-lead ECG: superiority of the inverse Dower matrix. J Electrocardiol. 1988;21:361-7. Faris OP, Evans FJ, Dick AJ, Raman VK, Ennis DB, Kass DA, McVeigh ER: Endocardial versus epicardial electrical synchrony during LV free-wall pacing. Am J Physiol Heart Circ Physiol. 2003;285:H1864-70. Reddy VY, Neuzil P, Taborsky M, Kralovec S, Sediva L, Ruskin JN: Images in cardiovascular medicine. Electroanatomic mapping of cardiac resynchronization therapy. Circulation. 2003;107:2761-3. Scher AM, Young AC: Spread of excitation during premature ventricular systoles. Circ Res. 1955;3:535-42. Spach MS, Barr RC: Analysis of ventricular activation and repolarization from intramural and epicardial potential distributions for ectopic beats in the intact dog. Circ Res. 1975;37:830-43. Vassallo JA, Cassidy DM, Miller JM, Buxton AE, Marchlinski FE, Josephson ME: Left ventricular endocardial activation during right ventricular pacing: effect of underlying heart disease. J Am Coll Cardiol. 1986;7:1228-33. Hooft van Huysduynen B, Swenne CA, Draisma HH, Antoni ML, Van de Vooren H, Van der Wall EE, Schalij MJ: Validation of ECG indices of ventricular repolarization heterogeneity: a computer simulation study. J Cardiovasc Electrophysiol. 2005; 16: 1097-103 Wilson FN, MacLeod AG, Barker PS: The T deflection of the electrocardiogram. Trans Assoc Am Physicians. 1931;46:29-38. Akar FG, Yan GX, Antzelevitch C, Rosenbaum DS: Unique topographical distribution of M cells underlies reentrant mechanism of torsade de pointes in the long-QT syndrome. Circulation. 2002;105:1247-1253. El Sherif N, Caref EB, Chinushi M, Restivo M: Mechanism of arrhythmogenicity of the short-long cardiac sequence that precedes ventricular tachyarrhythmias in the long QT syndrome. J Am Coll Cardiol. 1999;33:1415-23. Litovsky SH, Antzelevitch C: Transient outward current prominent in canine ventricular epicardium but not endocardium. Circ Res. 1988;62:116-26. Shimizu W, Antzelevitch C: Cellular basis for the ECG features of the LQT1 form of the long-QT syndrome: effects of beta-adrenergic agonists and antagonists and sodium channel blockers on transmural dispersion of repolarization and torsade de pointes. Circulation. 1998;98:2314-22. van Opstal JM, Verduyn SC, Winckels SK, Leerssen HM, Leunissen JD, Wellens HJ, Vos MA: The JT-area indicates dispersion of repolarization in dogs with atrioventricular block. J Interv Card Electrophysiol. 2002;6:113-20. Abildskov JA, Green LS, Evans AK, Lux RL: The QRST deflection area of electrograms

155

Chapter 7

33. 34. 35. 36. 37. 38. 39.

156

during global alterations of ventricular repolarization. J Electrocardiol. 1982;15:1037. Kubota I, Lux RL, Burgess MJ, Abildskov JA: Relation of cardiac surface QRST distributions to ventricular fibrillation threshold in dogs. Circulation. 1988;78:171-7. De Ambroggi L, Aime E, Ceriotti C, Rovida M, Negroni S: Mapping of ventricular repolarization potentials in patients with arrhythmogenic right ventricular dysplasia: principal component analysis of the ST-T waves. Circulation. 1997;96:4314-8. Zabel M, Malik M, Hnatkova K, Papademetriou V, Pittaras A, Fletcher RD, Franz MR: Analysis of T-wave morphology from the 12-lead electrocardiogram for prediction of long-term prognosis in male US veterans. Circulation. 2002;105:1066-70. van Oosterom A: Singular value decomposition of the T wave: its link with a biophysical model of repolarization. Int J Bioelectromagnetism. 2002;4:59-60. Lambiase PD, Rinaldi A, Hauck J, Mobb M, Elliott D, Mohammad S, Gill JS, Bucknall CA: Non-contact left ventricular endocardial mapping in cardiac resynchronisation therapy. Heart. 2004;90:44-51. Auricchio A, Fantoni C, Regoli F, Carbucicchio C, Goette A, Geller C, Kloss M, Klein H: Characterization of left ventricular activation in patients with heart failure and left bundle-branch block. Circulation. 2004;109:1133-9. van Oosterom A: The dominant T wave and its significance. J Cardiovasc Electrophysiol. 2003;14:S180-7.

Dispersion of the repolarization in cardiac resynchronization therapy

157

158

Chapter 8 Summary and conclusions

159

Chapter 8

Chapter 1 is a general introduction into the central theme of this thesis; the electrocardiographic assessment of repolarization heterogeneity. The term repolarization heterogeneity refers to differences in repolarization instants in the heart. Normally, repolarization in the human heart is a relatively smooth, continuous process, during which adjacent areas repolarize almost simultaneously. Several drugs or cardiac diseases may disturb the repolarization and thus increase the repolarization heterogeneity, which predisposes to arrhythmias. A non-invasive index able to assess this repolarization heterogeneity would have great clinical value. The standard 12-lead electrocardiogram (ECG) is attractive for this purpose as it is widely used and reflects repolarization heterogeneity: the morphology and duration of the T wave in the ECG depend on differences in the repolarization instants of the ventricular myocardial cells. The exact pattern of the normal repolarization in the human heart is still not precisely known. Several mapping and vectorcardiographic studies suggest that transmural (perpendicular to the myocardial wall) and paramural (along the ventricular walls) repolarization differences determine the normal morphology of the T-wave. Different electrocardiographic measures of T-wave morphology reflect different characteristics of cardiac repolarization, which defines their suitability to assess repolarization heterogeneity, as extensively discussed in this chapter. In Chapter 2 a mathematical ECG model is used to simulate the behavior of various ECG indices of ventricular repolarization with increasing repolarization heterogeneity. A stepwise increase in repolarization heterogeneity caused the T-wave amplitude and -area to increase and led to a more symmetrical T wave. Tapex-end interval and T-wave complexity had some limitations but also adequately reflected the repolarization heterogeneity. However, QT dispersion had low sensitivity in the transitional zone between normal and abnormal repolarization heterogeneity. There is ample evidence that manipulations of parasympathetic and sympathetic outflow result in non-uniform changes in action potential duration (APD). Hence, such manipulations can be used to alter repolarization heterogeneity. Chapter 3 describes how healthy males were subjected to a normotensive stressor (leg lowering) and to a hypertensive stressor (handgrip). During leg lowering, the heart rate could be influenced by varying the angle to which the legs were lowered with respect to the horizontal position. In each individual, the electrocardiogram during handgrip was compared to the electrocardiogram with a leg lowering angle that yielded the same 160

Summary and conclusions

heart rate. Compared with the normotensive stress of leg lowering, the hypertensive stress of the handgrip manoeuver caused a larger QT interval, no significantly different QT dispersion, a larger Tapex-Tend interval and a larger T wave complexity. These results suggest that a hypertensive stressor increases repolarization heterogeneity, which could explain the increased arrhythmogeneity associated with hypertensive stress as observed by others. During exercise the heart is influenced by strongly varying amounts of parasympathetic and sympathetic outflow, possibly leading to changing amounts of repolarization heterogeneity. In Chapter 4 we describe how we studied the electrocardiograms of healthy males of varying fitness levels from professional marathon speedskaters to untrained subjects. Within individuals we compared electrocardiograms with equal heart rates obtained during exercise and during recovery. During recovery, QT was smaller, while T-wave amplitude was larger, and the T-wave became more symmetric. Also, the ventricular gradient, which is a measure of APD heterogenity, was larger during recovery, while the QRS complex remained unchanged. The changes in T-wave amplitude, T-wave symmetry and ventricular gradient, together with an identical QRS complex, are all consistent with increased repolarization heterogeneity during recovery, caused by increased APD heterogeneity. The fittest among the study group, with the largest maximal oxygen consumption, the largest hearts, and the highest baroreflex sensitivity had much more outspoken electrocardiographic exercise-recovery differences than the low fit subjects. Possibly, our study may contribute to the explanation of sudden death after exercise. In Chapter 5 the electrocardiographic effects of pulmonary valve replacement in Fallot patients with dilated right ventricles are described. In these patients QRS duration is known to be predictive of ventricular arrhythmias. Pulmonary regurgitation is a common cause of right ventricular failure and QRS lengthening. By means of LEADS (an ECG analysis program specifically developed for this thesis work) we were able to demonstrate that pulmonary valve replacement actually decreased QRS duration. The structural success of the operation, measured as the reduction of the right ventricular end-diastolic volume, is related to the reduction of QRS duration. In Chapter 6 we extend the analysis of the Fallot patients to the electrocardiographic repolarization indices. After pulmonary valve replacement the QRS-T spatial angle, 161

Chapter 8

T-wave amplitude and T-wave area decreased, which could by potentially anti-arrhythmic. When dividing the group in patients with and without severe ventricular arrhythmias, the optimal discriminator of patients with arrhythmias was a QRS > 160 ms. No severe arrhythmias were found in patients with QRS-T angle < 100°, QT dispersion < 60 ms, T-wave complexity < 0.30 or a RV EDV < 220 ml. In Chapter 7 we describe the effects of different pacing modes of cardiac resynchronization devices on electrocardiographic indices of the repolarization. Previous studies had suggested a detrimental effect of epicardial left and biventricular pacing on the transmural heterogeneity of the repolarization, causing arrhythmias in vulnerable patients. Our patient and simulation study showed that, except for a small region close to the stimulation electrode, transmural heterogeneity of the repolarization was not larger during epicardial left and biventricular pacing than during conventional endocardial right ventricular pacing. The different behavior of the electrocardiographic repolarization indices during the different pacing modes could be explained by the repolarization patterns caused by the slowly progressing activation wave front spreading from the different pacing sites. Thus, electrocardiographic indices of the repolarization were related to the global pacing-induced repolarization patterns rather than to local transmural heterogeneity of the repolarization. In conclusion, mathematical simulations demonstrate that increased dispersion of the repolarization caused by increased dispersion of APD heterogeneity in normally excited hearts is reflected in the T-wave amplitude, T-wave area, T-wave symmmetry, and, with restrictions, in T-wave complexity and the Tapex-Tend interval. Our measurements in normal subjects suggest that hypertensive stress and recovery from exercise are conditions in normal hearts during which repolarization heterogeneity is increased. Immediately after vigorous exercise repolarization heterogeneity is largest. Pulmonary valve replacement in Fallot patients is beneficial for electrical heart function, as it decreases QRS duration and electrocardiographic repolarization heterogeneity indexes. Left ventricular pacing in heart failure patients leads to similar effects on the repolarization heterogeneity as traditional right ventricular pacing. Transmural repolarization heterogeneity cannot be assessed from the electrocardiogram during pacing, as T wave morphology in paced hearts is predominated by the global repolarization pattern induced by pacing.

162

163

Nederlandse samenvatting

Nederlandse samenvatting Hoofdstuk 1 is een algemene introductie van het centrale thema van de thesis; de electrocardiografische inschatting van repolarisatie heterogeniteit. Repolarisatie heterogeniteit weerspiegelt de verschillen in de repolarisatie-tijdtippen in het hart. In het gezonde menselijke hart is de repolarisatie een gelijkmatig proces dat vrijwel gelijktijdig plaatsvindt in naast elkaar gelegen delen. Verschillende medicamenten en ziektes kunnen de repolarisatie echter verstoren en verschillen in repolarisatietijd vergroten. Een toegenomen repolarisatie heterogeniteit predisponeert tot ventriculaire aritmieën. Een niet-invasieve test op repolarisatie heterogeniteit heeft potentieel een grote waarde om patiënten op te sporen met een verhoogd risico op aritmieën. Het electrocardiogram (ECG) wordt veel gebruikt en weerspiegelt de repolarisatie heterogeniteit; de morfologie en duur van de T golf in het ECG worden bepaald door verschillen in repolarisatie van de myocardcellen. Het exacte patroon van de normale menselijke repolarisatie is nog niet helemaal bekend. Verschillende mapping studies en beschouwing van de vectorcardiografie suggereren dat transmurale (haaks op het oppervlakte van de ventrikelwand) en paramurale (parallel aan de ventrikelwand) repolarisatieverschillen de normale morfologie van de T golf bepalen. Verschillende electrocardiografische maten van de T golf reflecteren verschillende karakteristieken van de cardiale repolarisatie, hetgeen hun geschiktheid bepaalt om repolarisatie heterogeniteit in te schatten, zoals uitgebreid wordt besproken in dit hoofdstuk. In Hoofdstuk 2 wordt een mathematisch ECG simulatie model gebruikt om de gevolgen van een vergroting van de repolarisatie heterogeniteit op de verschillende ECG indices van de repolarisatie te meten. Een stapsgewijze vergroting van de repolarisatie heterogeniteit veroorzaakte een vergroting van de T-golf amplitude en oppervlakte en een meer symmetrische T-golf. Behoudens enkele tekortkomingen, werd de toegenomen repolarisatie heterogeniteit ook tot uiting in een verlengd Tapex-Teind interval en een vergrote T-golf complexiteit. QT dispersie daarentegen, had een lage sensitiviteit in de overgangszone tussen een normale en abnormale repolarisatie heterogeniteit. Manipulatie van het parasympathische en sympathische zenuwstelsel veroorzaakt

164

Nederlandse samenvatting

niet-uniforme veranderingen in actie potentiaal duur (APD). Dergelijke manipulaties kunnen worden gebruikt om de repolarisatie heterogeniteit te beïnvloeden. Hoofdstuk 3 beschrijft de effecten van een normotensieve stressor (leg-lowering) en een hypertensieve stressor (handgrip) op de repolarisatie van gezonde mannen. Tijdens leg-lowering neemt de hartfrequentie toe met de hoek waarin de benen van de proefpersoon worden gepositioneerd. Van elke proefpersoon werd het ECG tijdens leglowering geselecteerd waarvan de hartfrequentie het meest overeenkwam met de hartfrequentie tijdens de handgrip procedure. Vergeleken met de normotensieve stress van leglowering, resulteerde de hypertensieve stress van handgrip in een langer QT interval, geen significant verschil in QT dispersie, een langer Tapex-Teind interval en een grotere T-golf complexiteit. Deze resultaten suggereren dat hypertensieve stress de repolarisatie heterogeniteit vergroot, hetgeen de eerder geobserveerde associatie tussen hypertensieve stress en arrhythmogeniciteit zou kunen verklaren. Tijdens lichamelijke inspanning staat het hart onder invloed van variërende hoeveelheden sympatische en parasympatische invloeden, hetgeen kan leiden tot een variatie in repolarisatie heterogeniteit. In Hoofdstuk 4 wordt de invloed van inspanning op de electrocardiografische repolarisatie beschreven. De mannelijke proefpersonen varieerden in fitheid van ongetrainde proefpersonen tot professionele marathon schaatsers. De repolarisatie tijdens tijdens inspanning en herstel werd intra-individueel met elkaar vergeleken in ECGs met dezelfde hartfrequentie tijdens inspanning en herstel. Tijdens de vroege herstelfase was het QT interval korter, de T-golf amplitude groter en werd de T golf symmetrischer. Tevens was de ventriculaire gradient, een maat voor de APD heterogeniteit, groter tijdens de herstelfase. De veranderingen in de T-golf amplitude, T-golf symmetrie en de ventriculaire gradient duiden op een toegenomen repolarisatie heterogeniteit tijdens de herstelfase. De fitste proefpersonen, diegenen met het hoogste maximale zuurstof verbruik, de laagste rust-hartfrequentie, de grootste harten en de grootste baroreflex sensititviteit, hadden grotere inspanning-herstel verschillen in de repolarisatie dan de minst fitte proefpersonen. Wellicht zouden deze resultaten bij kunnen dragen aan de verklaring van een toegenomen risico op plotse dood na inspanning. In Hoofdstuk 5 worden de electrocardiografische effecten beschreven van pulmonaal klep vervanging in patiënten met tetralogie van Fallot met gedilateerde rechter ventrikels. Pulmonalisklep insufficiëntie kan in Fallot patiënten dilatatie van rechter 165

Nederlandse samenvatting

ventrikel en verlenging van de QRS duur veroorzaken. In deze patiëntengroep is de QRS duur voorspellend voor ventriculaire aritmieën en plotse dood. Wij ontwikkelden en gebruikten een interactief ECG analyse programma dat de QRS duur heel nauwkeurig kan meten en vonden dat pulmonaalklepvervanging resulteerde in een verkorting van de QRS duur. Het structurele succes van de operatie, gemeten als een afname van het rechter ventrikel eind-diastolische volume, was gerelateerd aan de verkorting van de QRS duur. In Hoofdstuk 6 breidden we deze analyse uit met een studie naar electrocardiografische repolarisatie. Door pulmonaalklep vervanging verkleinde de QRS-T ruimtehoek, de T-golf amplitude en de T-golf oppervlakte, wat een afgenomen ventriculaire repolarisatie heterogeniteit suggereert. De QRS duur bleek de optimale maat om patiënten met aritmieën te onderscheiden van patiënten zonder aritmieën. Echter, patiënten met een QRS-T hoek < 100°, QT dispersie < 60 ms, een T-golf complexiteit < 0.30 of een rechter ventrikel eind-diastolisch volume < 220 ml hadden geen ernstige ventriculaire aritmieën. In Hoofdstuk 7 worden de effecten van verschillende pacing modaliteiten van cardiale resynchronisatie therapie op de electrocardiografische repolarisatie beschreven. Eerdere studies suggereerden een electrocardiografisch meetbare vergroting van de transmurale repolarisatie heterogeniteit door epicardiaal links- en biventriculair pacen, waardoor aritmieën zouden ontstaan in kwetsbare patiënten. Onze patiënten simulatie studie liet zien dat, behalve in een klein gebied direct onder de pacing electrode, de transmural repolarisatie heterogeniteit niet groter was tijdens epicardiaal links- en biventriculair pacen dan tijdens conventioneel endocardiaal rechts ventriculair pacen. De waargenomen verschillen in de electrocardiografische T-golf tijdens verschillende pacing modi konden worden verklaard door de repolarisatie patronen die grotendeels bepaald werden door de activatie patronen die zich relatief traag verspreiden vanaf de pacing electrodes. De electrocardiografische repolarisatie tijdens pacen bleek niet gerelateerd te zijn aan de lokale transmurale repolarisatie heterogeniteit maar aan de pacing-geïnduceerde globale repolarisatie patronen. Concluderend laten mathematische simulaties zien dat een toegenomen repolarisatie heterogeniteit in normaal geëxciteerde harten wordt weerspiegeld in veranderingen in de T-golf amplitude, T-golf oppervlakte, T-golf symmetrie en, met enkele 166

Nederlandse samenvatting

restricties, T-golf complexiteit en Tapex-eind interval. Onze metingen in gezonde proefpersonen suggereren dat hypertensieve stress de repolarisatie heterogeniteit verhoogt. Onmiddellijk na inspanning is de repolarisatie heterogeniteit het grootst. Pulmonaalklep vervanging in Fallot patiënten verkort de QRS duur en vermindert de repolarisatie heterogeniteit. Links ventriculair pacen in hartfalen patiënten leidt tot vergelijkbare effecten op de repolarisatie heterogeniteit als tijdens conventioneel rechts ventriculair pacen. Lokale transmurale repolarisatie heterogeniteit kan niet worden afgemeten aan het ECG omdat de T-golf morfologie in deze patiënten wordt bepaald door pacing-geïnduceerde globale repolarisatie patronen.

167

Dankwoord Tijdens mijn promotie heb ik het merendeel van mijn tijd in “De Tuin” gezeten. In deze tuin is geen plant te bekennen, maar wel zo’n 15 collega’s. Behalve dat deze groep maakt dat je met plezier naar je werk gaat, leer je ook van elkaar zodat je niet het wiel opnieuw hoeft uit te vinden. Hedde en Harmen, bedankt voor het ‘managen’ van de data en de ontwikkeling van LEADS. Arie, bedankt voor je hulp bij het omzetten van ECGs en Holters. Annelies en Louisa voor metingen en analyses. Henk Ritsema van Eck voor inspiratie en wetenschappelijke discussie. Jan Kors en Ge van Herpen voor commentaar en lering over het electrocardiogram. De Leiding van de afdeling cardiologie voor het mogelijk maken van congresbezoeken, waar ik plezier heb gehad en veel internationale ervaring heb opgedaan. Linda, Lya, Talitha, Saskia en Esther, voor jullie hulp in de wereld van formulieren, statussen en afspraken. Renee voor de hulp bij de inspanningstesten van de proefpersonen. Marike, voor je adviezen, hulp en luisterend oor. Sander en Sven, paranimfen, met zijn drieën hebben we er een mooie tijd van gemaakt en dat zullen we nog lang blijven doen, “Eén voor allen, allen voor één!”. David en PartyBarty voor de gezelligheid buiten de afdeling cardiologie. Lieve pa en ma en Thijs, bedankt voor jullie liefde en ondersteuning die niemand anders kan geven. Pascalle, jij bent het mooiste wat ik aan mijn promotie heb over gehouden.

168

Curriculum Vitae Bart Hooft van Huysduynen werd geboren op 20 oktober 1974 in Amsterdam. Hij is opgegroeid in het pittoreske Loenen aan de Vecht en in 1993 behaalde hij het eindexamen van het Gemeentelijk Gymnasium te Hilversum. Na een jaar psychologie werd hij ingeloot voor geneeskunde in het Hoge Noorden in 1994. In Groningen werkte hij als student mee aan onderzoek naar de interactie tussen atrium fibrilleren en hartfalen in geiten onder leiding van dr. I.C. van Gelder en Bas Schoonderwoerd. Na een jaar op stal gestaan te hebben vertrok hij naar Curaçao voor zijn co-schappen in 1999 en 2000. Deze co-schappen werden bekroond met een keuze-coschap electrofysiologie in het Utrecht onder leiding van professor Hauer, waarna hij zijn artsexamen behaalde in 2001 aan de Rijksuniversiteit Groningen. In Leiden startte hij januari 2002 met zijn promotie-onderzoek onder leiding van dr. ir . Swenne, professor Schalij en professor Van der Wall waarvan u het resultaat nu in handen heeft. Sinds januari 2006 volgt hij de vooropleiding tot cardioloog in het Rijnland Ziekenhuis te Leiderdorp.

169