Demonstration of the Proarrhythmic Preconditioning of Single Premature Extrastimuli by Use of the Magnitude, Phase, and Distribution of Repolarization Alternans
Background—We hypothesized that single premature extrastimuli (S2) insufficient to induce reentry produce proarrhythmic effects (proarrhythmic preconditioning) that are measurable by use of the magnitude, phase, and temporal distribution of repolarization alternans (RPA; alternate-beat fluctuations in ECG repolarization).
Methods and Results—Before programmed electrical stimulation (PES), surface ECG leads I, aVF, and V1 were recorded in 30 patients during simultaneous atrial and ventricular pacing at 500 ms with S2 coupling intervals (CIs) decreasing from 400 to 240 ms in 20-ms steps. We determined RPA magnitude (Valt) as the 0.5-cycle/beat peak after spectral decomposition of consecutive STU intervals over 64 beats immediately preceding and following each S2, RPA phase reversals as discontinuities in the even/odd phase of STU alternation, and RPA distribution as the time point of median RPA magnitude within repolarization. Eighteen patients were induced into ventricular tachycardia (VT), whereas 12 were not. Extrastimuli dynamically modulated each characteristic of RPA. S2 augmented Valt in inducible (8.2±2.3 versus 6.2±1.6 μV; P=0.003) but not noninducible patients. S2 reversed RPA phase more in inducible than in noninducible patients (56.7% versus 45.3%; P=0.02 by χ2), particularly when CI was ≤300 ms (66.3% versus 46.5%; P=0.006). Finally, S2 redistributed RPA significantly later within repolarization in inducible patients. Each effect was more marked for CI ≤300 ms.
Conclusions—A single S2 increases RPA magnitude, reverses its phase, and redistributes it later in repolarization in patients with the substrates for VT. These effects become more pronounced with shorter coupling intervals. These results suggest that it is possible to track the dynamic proarrhythmic preconditioning of single premature depolarizations.
Sudden cardiac death resulting from ventricular tachycardia (VT) or fibrillation (VF)1 occurs with an annual incidence of 1 to 2/1000 in the United States.2 However, the manner in which “experiments of nature” such as premature ventricular depolarizations dynamically modulate electrical instability to facilitate these arrhythmias remains unclear and unmeasured. Although a series of extrastimuli may induce reentrant arrhythmias during programmed electrical stimulation (PES) of the ventricle, the proarrhythmic effects of single premature depolarizations that do not induce echo beats or initiate reentry remain poorly defined and unmeasurable.
We set out to determine whether a single ventricular premature depolarization may alter repolarization dynamics and increase the propensity for ventricular arrhythmias in susceptible individuals (proarrhythmic preconditioning) using repolarization alternans (RPA). RPA measures microvolt-level fluctuations in the ECG STU segment occurring on an alternate-beat basis, reflects spatial3 and temporal4 5 6 dispersion of repolarization, and has been linked with ventricular arrhythmias experimentally7 8 and clinically.7 9 However, work to date has focused on the static relationship between clinical arrhythmic outcome and RPA magnitude at steady-state heart rates.10 11
Three lines of experimental evidence suggest that distinct characteristics of RPA may reflect dynamic changes in myocardial electrical instability. First, ventricular arrhythmias become more likely with increasing magnitude of action potential duration (APD) alternans (in feline myocytes12 and intact guinea pig5 and canine3 13 hearts) and RPA (in hypothermic7 and ischemic8 14 canine hearts). Second, the phase of RPA is not static; phase reversal in the alternation of isolated myocytes12 and opposite phase of alternation between adjacent regions of Langendorff-perfused guinea pig heart15 portend ventricular arrhythmias. Third, we recently reported6 that RPA has a nonuniform temporal distribution within a beat. RPA late in repolarization better reflects reentrant substrates than early RPA, and furthermore, RPA temporally redistributes later with heart rate acceleration.
We therefore hypothesized that the proarrhythmic preconditioning of a single premature depolarization may be reflected by an increased magnitude, reversed phase, and later redistribution of RPA within the STU segment. We further hypothesized that such dynamics should become more marked with increasing stimulus prematurity and in individuals with demonstrable substrates for reentrant arrhythmias. We tested these hypotheses in patients undergoing PES.
Patient Recruitment and Clinical ECG Collection
Studies were approved by the Human Subjects Committee of the Washington University School of Medicine, and all subjects provided written informed consent. We recruited 30 patients referred for PES for the evaluation of unexplained syncope or VT. A subset of these patients has been studied previously.6 Patients were prepared for PES in the standard fashion, and antiarrhythmic medications were withheld for ≥2 days. After light sedation with intravenous midazolam, standard 6F quadripolar catheters were positioned in the right atrium and right ventricle via transvenous sheaths as part of the prescribed PES and used for simultaneous atrial and ventricular pacing.
To provide a uniform heart rate between patients, baseline pacing was applied at a cycle length of 600 ms for 2 to 5 minutes. Extrastimulus experiments then commenced. Ninety-nine beats were delivered at a cycle length of 500 ms, followed by 1 premature extrastimulus (S2) at varying coupling intervals (CIs), then a 500-ms delay (Figure 1⇓). The cycle was repeated with the CI shortening from 400 to 240 ms in progressive 20-ms steps. S2 were delivered twice at each CI. Surface ECG leads I, aVF, and V1 were recorded with a 16-channel analogue amplifier (Bloom & Associates) with bandpass of 0.04 to 100 Hz, sampled at 1 kHz to 12-bit resolution, then transferred to a UNIX workstation (Sun Computer) for offline analysis.
Spectral Computation of RPA
RPA was analyzed with interactive graphic software written by the authors in Labview (National Instruments). For each ECG lead, 64 contiguous-beat sequences avoiding ectopic or fusion complexes were selected (1) before S2 (ending with the beat preceding S2) and (2) after S2 (starting with and including S2) (Figure 1⇑). Beats were baseline corrected, then aligned to their maximal normalized QRS dot-product with a template beat, and then these 64 aligned beats were used to calculate a mean QRS complex that served as the template for a second alignment.16
RPA was computed on arrays of aligned beats from each ECG lead by use of multidimensional spectral analysis over the STU segment, which was identified by an investigator who was blinded to clinical data (Figure 2⇓).7 16 STU series were represented as 2D repolarization matrices R (n, s), where n indicates the sequence beat (0≤n≤63) and s the millisecond time sample within STU. A fast Fourier transform was used to compute power spectra (power= voltage2) across all beats n at each successive time sample s (along each vertical arrow in Figure 2⇓, left).
Power spectra were summed into a composite in which the magnitude at 0.5 cycle/beat indicates raw alternans, ΣT (in μV2) (Figure 2a⇑). The dimensionless T-wave alternans ratio (TWAR) represents the difference between alternans and nonalternating periodicity (spectral noise mean, μnoise, prospectively defined as the 10 spectral point bandwidth, 0.33 to 0.48 cycle/beat), measured in units of the SD of noise (ςnoise) (all in μV2): where TWAR >0 indicates RPA detectable above noise and TWAR ≥3 may predict ventricular arrhythmias.9 The mean absolute voltage difference of alternation was also estimated (Figure 2⇑, left)16 :
RPA magnitudes were analyzed for each ECG lead and combined into the resultant vector by the Pythagorean theorem. Vector magnitudes were subsequently analyzed because of the spatial nonuniformity of RPA between leads.
Determining the Phase of RPA
RPA phase was determined relative to the position of S2 in each 64-beat sequence. The voltage alternation of 1 STU time sample is shown in Figure 3⇓ (top). A premature S2 may leave the phase either unaltered (Figure 3a⇓) or reversed (Figure 3b⇓) in subsequent beats. Phase reversal was therefore defined by interrupted RPA oscillation (Figure 3b⇓) in post-S2 compared with pre-S2 sequences in ≥1 surface ECG lead, providing that TWAR was >0 in both sequences in that lead.
Determining the Temporal Distribution of RPA Within Repolarization
The temporal distribution of RPA within repolarization was analyzed in pre-S2 and post-S2 sequences as described previously.6 Briefly, RPA was reconstructed in the time domain to represent RPA magnitude for each STU time sample (Figure 2b⇑). We used the time point of median RPA magnitude (the abscissa of the center-of-area), x, normalized to STU duration (d), to indicate RPA distribution as T=x/d (0<T<1). Figure 2b⇑ illustrates centrally distributed RPA (T ≈0.5); earlier or later distributions would result in T<0.5 and T>0.5, respectively.
PES Protocol and Outcome Comparisons
PES was performed in standard fashion by pacing at the right ventricular apex. A train of 8 S1 stimuli were followed progressively by single (S2), double (S2S3), or triple (S2S3S4) premature extrastimuli as required to induce sustained monomorphic VT. If unsuccessful, this process was repeated at a second ventricular site. Each patient (n=30) was considered inducible or noninducible accordingly. Similarly, patients were RPA-positive if TWAR was ≥3 in any of the 3 surface leads I, aVF, or V1, or RPA-negative if not. These data were compared prospectively with inducibility at PES.
Continuous data are presented as mean±SD. The 2-tailed t test was used to compare RPA magnitude and centers-of-area between groups, with Bonferroni correction applied for multiple comparisons. The χ2 test was applied to contingency tables of phase reversal. A probability level of 5% (P<0.05) was considered statistically significant.
Sustained monomorphic VT was induced in 18 patients at the time of PES, whereas 12 patients were noninducible. All inducible patients presented with clinical VT. Of the noninducible patients, 3 presented with VT and the remainder with syncope of uncertain cause. Their demographics are shown in Table 1⇓. None of our patients exhibited polymorphic VT or VF, and single extrastimuli did not induce echo beats or VT in any of them.
Relationship Between RPA Magnitudes and VT
RPA magnitude was higher in patients with than in those without VT. In pre-S2 sequences, inducible patients had mean vector Valt=6.15±1.57 versus 3.98±1.65 μV in noninducible patients (P=0.019 by t test). In post-S2 sequences, corresponding magnitudes were Valt=8.22±2.30 versus 3.94±2.18 μV (P=0.003). Similar results were obtained for TWAR (Figure 4⇓).
Effect of S2 on RPA Magnitude
In each patient group, TWAR and Valt were higher in post-S2 than pre-S2 sequences for pooled CIs (computing the unweighted arithmetic mean of data after all CIs). However, the magnitude of post-S2 RPA augmentation was significant only in inducible patients (Valt increased from 6.15±1.57 to 8.22±2.30 μV for pooled CIs; P=0.003) and not in noninducible patients (Valt=3.98±1.65 versus 3.94±2.18 μV; P=NS).
Figure 4⇑ (top) shows that TWAR augmentation was especially marked after S2 with a short CI (≤300 ms) in inducible but not in noninducible patients, further widening the statistical difference between patient groups (pre-S2 P<0.01, post-S2 P<0.005). Conversely, longer S2 CIs (≥320 ms) augmented TWAR to a lesser extent and affected both groups similarly (Figure 4⇑, bottom). Similar results were found for Valt.
RPA Phase Modulation by S2
Phase reversal was significantly more common in inducible patients (and followed 56.7% or 127 of 224 presentations) than in noninducible patients (45.3% or 81 of 179 presentations; P=0.02 by χ2) for pooled CIs (Table 2⇓). The effect of CI was again bimodal, with short (≤300 ms) but not longer (≥320 ms) CIs causing significantly more phase reversals in inducible than in noninducible patients (P=0.006; Table 2⇓ and Figure 5⇓). Only sequences lacking ectopy and other spurious ECG events were analyzed.
Figure 5⇑ shows that the proportion of S2 presentations causing phase reversal in patients with VT increased with progressive S2 prematurity, reaching 83.3% after the shortest CI (240 ms). This trend was not observed in noninducible patients. Importantly, there was a trend for inducible patients who showed RPA phase reversal after S2 CI 240 ms to be more readily inducible on subsequent PES (requiring a mean of 1.6 extrastimuli) than those without phase reversals (2.3 extrastimuli, P=0.10 by t test).
RPA Temporal Distribution After S2
RPA was distributed later within repolarization for inducible than for noninducible patients (Figure⇑ 6). In all patients, extrastimuli redistributed RPA later within repolarization for short CIs (≤300 ms) but not CIs ≥320 ms. Furthermore, inducible patients experienced a greater redistribution that further widened the statistical difference between groups (pre-S2 T=0.574 versus 0.524 [P<0.02] and post-S2 T=0.616 versus 0.562 [P<0.005] [t test]).
When individual S2 CI was considered, the magnitude of RPA redistribution in inducible patients was significant after S2 CIs of 380 (P=0.016), 320 (P=0.011), 300 (P=0.049), 280 (P=0.014), and 240 ms (P=0.049) (Bonferroni correction applied).
This study is the first to demonstrate that single premature ventricular depolarizations that do not induce VT may still produce measurable repolarization effects that reflect an enhanced propensity for ventricular arrhythmias (proarrhythmic preconditioning). Single extrastimuli augment the magnitude, reverse the phase, and cause a later distribution of RPA in inducible patients. These effects become more pronounced with increasing stimulus prematurity (CI ≤300 ms), and phase reversal, in particular, correlates with the subsequent ease of induction of VT. These repolarization dynamics were not observed in patients without inducible VT.
Pathophysiology Underlying RPA
RPA may arise through 2 distinct mechanisms. In the first, intrinsic dispersion of refractoriness prevents the complete depolarization of myocytes with the longest recovery times (or alters their phase 2 action potential morphology15 ), allowing depolarization only on alternate cycles. In the second hypothesis, discordance in the generation of myocardial action potentials17 causes a secondary dispersion of repolarization. Extracellular recordings3 18 19 20 21 and voltage-sensitive dye data5 15 suggest that APD variability is primary. The dispersion of recovery in either case produces ephemeral conduction barriers15 22 that predispose to demonstrable wavefront fractionation and reentrant arrhythmias.23
S2 Augment RPA Magnitude
This report is the first to document the effects of premature extrastimuli on clinical RPA and their differential effects in subjects with and without VT. RPA magnitude in inducible patients was augmented by extrastimuli (Figure 4⇑), consistent with studies of APD in nonischemic24 and ischemic3 25 dogs. These effects were bimodal with respect to CI, and RPA was augmented to a greater extent after CIs ≤300 ms than after CIs ≥320 ms (Figure 4⇑). In noninducible patients, the small magnitude of RPA and its relative insensitivity to extrastimuli and heart rate6 may reflect different underlying substrates.
Mechanistically, extrastimuli may augment RPA magnitude either by modulating repolarization dispersion or by altering depolarization with secondary repolarization effects. Dispersion of repolarization has been shown to follow single5 24 26 and double27 28 extrastimuli in several preparations and may represent the mechanism by which long-short extrastimulus sequences induce reentry. Although the depolarization sequence may be altered by extrastimuli (for example, when CI differs from the paced cycle length by ≥250 ms29 ), the primary contribution of this mechanism to RPA requires additional study.
Extrastimuli Reverse the Alternating Phase of RPA
This is the first study of RPA phase and its modulation by extrastimuli in humans, and its results are consistent with observations of alternans phase reversal and discordance in experimental preparations. Post-S2 RPA phase reversals were significantly more common in patients with than without the substrates for reentry and increased in frequency with S2 prematurity (Table 2⇑ and Figure 5⇑).
RPA on the surface ECG likely represents a weighted integral of alternans in competing myocyte subpopulations, each having distinct magnitude and phase. Phase reversal after appropriately timed extrastimuli may therefore reflect an altered predominance of ≥1 subpopulation or direct reversal (“resetting”) of myocyte APD alternans. The range of CIs found to cause RPA phase reversal in the present study agrees with studies in feline myocytes12 and canine ventricle3 and, in general, may be a function of cycle length, the CI-to–cycle length relationship, and stimulus amplitude (in canine myocytes30 ). Our finding that patients with phase reversal more readily experienced ventricular arrhythmias agrees with results of isolated myocyte studies12 and may represent the onset of discordant alternans: the spatial juxtaposition of subpopulations alternating with opposite phase. In cellular and tissue-level studies, discordance may lead to unidirectional conduction block15 22 and set the stage for reentry. Furthermore, spontaneous discordant APD alternans has been noted in quinidine-intoxicated dogs with VT/VF30 and ischemic canine ventricle preceding the onset of ventricular arrhythmias.3 31 Notably, RPA magnitude and phase may be synergistic. Greater alternans magnitudes are associated with phase reversal of APD alternans at less-premature CIs in feline myocytes12 and facilitate the onset of discordant ST-T alternans in dogs.31
A full understanding of the relationship between RPA phase lability and dynamic arrhythmia initiation requires additional study. Although RPA phase reversal was associated with a readier induction of VT, our goal was to study the effects of stimuli insufficient to induce reentry, and we did not examine RPA phase immediately preceding VT. This may explain why even the most premature S2 produced a submaximal (84.3%) incidence of phase reversals. However, these results hint at the exciting future possibility of dynamically tailoring device or pharmacological therapy in response to measured arrhythmic susceptibility.
RPA Redistributes Later Within Repolarization After Extrastimuli
Extrastimuli caused RPA to redistribute later within repolarization in inducible patients. The extent of temporal redistribution was again bimodal for CI and reached significance for S2 CI ≤300 ms (Figure 6⇓). Evidence increasingly supports the notion that the distal T wave reflects a potentially arrhythmogenic transmural gradient of repolarization in normal32 and long-QT syndrome physiology.33 34 This is reflected in the T-wave peak–to-offset interval,32 by dispersion of the late QT interval in Langendorff-perfused rabbit heart,35 and by late RPA in humans.6 Speculatively, a single extrastimulus may therefore create temporal variability in the trailing edge of repolarization, enabling an advancing wavefront of activation to encounter inhomogeneously refractory myocardium and predisposing to wavefront fractionation and reentry. This hypothesis requires additional study.
We paced the atria and ventricles simultaneously to minimize R-R (and hence repolarization) variability and to allow analysis of late repolarization unobscured by superimposed atrial activity. This method differs from other clinical studies10 11 and may preclude a direct comparison of RPA magnitude between studies.
As expected, there was a trend for inducible patients to have worse cardiac systolic function and a higher incidence of previous myocardial infarction than noninducible patients (Table 1⇑). Although 1 measure of RPA has been found to be independent of the presence of structural heart disease,9 10 data do suggest that myocyte slippage36 and neurohumoral tone may modulate repolarization, whereas myocardial stretch may also modulate action potential morphology.37 Larger studies should therefore address whether RPA reflects electrical instability independently of such pathology. Finally, we acknowledge that ours is a small study and that additional work in larger groups of patients is required to validate our results.
In patients with the substrates for VT, increasingly premature single extrastimuli increase the magnitude of RPA, reverse its phase, and cause it to be distributed later within repolarization. These results suggest that a single premature extrastimulus that fails to induce VT may still produce proarrhythmic effects in susceptible individuals. Measuring such proarrhythmic preconditioning may have implications for dynamically tailoring therapy to arrhythmic susceptibility.
We are grateful to Drs Brett M. Baker, Gregory W. Botteron, Van H. De Bruyn, Mitchell N. Faddis, Marye J. Gleva, and Patricia A. Guerrero for their help in facilitating data recordings for this study. We also thank Dr Benico Barzilai for his helpful comments on the manuscript.
Presented in part in the Samuel A. Levine Young Investigator Award Competition of the 71st Scientific Sessions of the American Heart Association, Dallas, Tex, November 8, 1998, and published in abstract form (Circulation. 1998;98[suppl I]:I-M, I-27).
- Received January 29, 1999.
- Revision received June 24, 1999.
- Accepted July 2, 1999.
- Copyright © 1999 by American Heart Association
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.
Smith JM, Clancy E, Valeri C, Ruskin J, Cohen R. Electrical alternans and cardiac electrical instability. Circulation. 1988;77:110–121.
Nearing BD, Huang AH, Verrier RL. Dynamic tracking of cardiac vulnerability by complex demodulation of the T-wave. Science. 1991;252:437–440.
Rubenstein DS, Lipsius SL. Premature beats elicit a phase reversal of mechanoelectrical alternans in cat ventricular myocytes: a possible mechanism for reentrant arrhythmias. Circulation. 1995;91:201–214.
Saitoh H, Bailey JC, Surawicz B. Alternans of action potential duration after abrupt shortening of cycle length: differences between dog Purkinje and ventricular muscle fibers. Circ Res. 1988;62:1027–1040.
Pastore JM, Girouard SD, Laurita KR, Akar FG, Rosenbaum DS. Mechanism linking T-wave alternans to the genesis of cardiac fibrillation. Circulation. 1999;99:1385–1394.
Hirayama Y, Saitoh H, Atarashi H, Hayakawa H. Electrical and mechanical alternans in canine myocardium in vivo: dependence on intracellular calcium cycling. Circulation. 1993;88:2894–2902.
Tachibana H, Kubota I, Yamaki M, Watanabe T, Tomoike H. Discordant S-T alternans contributes to formation of reentry: a possible mechanism of reperfusion arrhythmia. Am J Physiol. 1998;275:H116–H121.
Gotoh M, Uchida R, Mandel WJ, Fishbein MC, Chen P-S, Karagueuzian HS. Cellular graded responses and a ventricular vulnerability to reentry by a premature stimulus in isolated canine ventricle. Circulation. 1997;95:2141–2154.
Saitoh H, Bailey JC, Surawicz B. Action potential duration in dog Purkinje and ventricular muscle fibers: further evidence in support of two different mechanisms. Circulation. 1989;80:1421–1431.
Hashimoto H, Asano M, Nakashima M. Alternation in refractoriness and in conduction delay in the ischemic myocardium associated with the alternation in the ST-T complex during acute coronary occlusion in anesthetized dogs. J Electrocardiol. 1986;9;77–84.
Kuo C-S, 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.
Kuo C-S, Atarashi H, Reddy CP, Surawicz B. Dispersion of ventricular repolarization and arrhythmia: study of two consecutive premature complexes. Circulation. 1985;72:370–376.
Goyal R, Harvey M, Daoud EG, Brinkman K, Knight BP, Bahu M, Weiss R, Bogun F, Man KC, Strickberger SA, Morady R. Effect of coupling interval and pacing cycle length on morphology of paced ventricular complexes: implications for pace mapping. Circulation. 1996;94:2843–2849.
Karagueuzian HS, Khan SS, Hong K, Kobayashi Y, Denton T, Mandel WJ, Diamond GA. Action potential duration alternans and irregular dynamics in quinidine-intoxicated ventricular muscle cells: implications for ventricular proarrhythmia. Circulation. 1993;87:1661–1672.
Konta T, Ikeda K, Yamaki M, Nakamura K, Honma K, Kubota I, Yasui S. Significance of discordant ST alternans in ventricular fibrillation. Circulation. 1990;82:2185–2189.
Yan G-X, Antzelevitch C. Cellular basis for the normal T-wave and the electrocardiographic manifestations of the long-QT syndrome. Circulation. 1998;98:1928–1936.
Shimizu W, Antzelevitch C. Cellular basis for T-wave alternans. Circulation. 1999;99:1499–1507.
Lubinski A, Lewicka-Nowak E, Kempa M, Baczynska AM, Romanowska I, Swiatecka G. New insight into repolarization abnormalities in patients with congenital long QT syndrome: the increased transmural dispersion of repolarization. Pacing Clin Electrophysiol. 1998;21(pt II):172–175.
Zaman AG, Morris JL, Smyllie JH, Cowan JC. Late potentials and ventricular enlargement after myocardial infarction. Circulation. 1993;88:905–914.
Lab MJ, Dean J. Myocardial mechanics and arrhythmia. J Cardiovasc Pharmacol. 1991;18(suppl 2):S72–S79.