Increased Short-Term Variability of Repolarization Predicts d-Sotalol–Induced Torsades de Pointes in Dogs
Background— Identification of patients at risk for drug-induced torsades de pointes arrhythmia (TdP) is difficult. Increased temporal lability of repolarization has been suggested as being valuable to predict proarrhythmia. The predictive value of different repolarization parameters, including beat-to-beat variability of repolarization (BVR), was compared in this serial investigation in dogs with chronic AV block.
Methods and Results— In anesthetized dogs with electrically remodeled hearts, the dose-dependent difference in drug-induced TdP (d-sotalol, 2 and 4 mg/kg IV over 5 minutes, 25% and 75% TdP, respectively) could not be accounted for by prolongation of QTc (410±37 to 475±60 versus 415±47 to 484±52 ms, respectively). BVR was quantified by Poincaré plots at baseline and immediately before onset of d-sotalol–induced extrasystolic activity. TdP occurrence was associated with an increase in short-term variability (STV) of the left ventricular monophasic action potential duration (3.5±1.5 to 5.5±1.6 versus 3.0±0.7 to 8.6±3.8 ms, respectively), which was reversible when TdP was abolished by IK,ATP activation. The absence of TdP despite QTc prolongation after chronic amiodarone treatment could also be explained by an unchanged STV. In experiments with isolated ventricular myocytes, STV increased after IKr block and was highest in cells that subsequently showed early afterdepolarizations.
Conclusions— Proarrhythmia is not related to differences in prolongation of repolarization but corresponds to BVR, here quantified as STV of the left ventricle. STV could be a new parameter to predict drug-induced TdP in patients.
Received January 26, 2004; de novo received April 13, 2004; accepted May 19, 2004.
Torsades de pointes arrhythmias (TdP) are repolarization-dependent polymorphic ventricular tachyarrhythmias that can stop spontaneously or degenerate into ventricular fibrillation, causing sudden death. One of the electrophysiological hallmarks of TdP is the prolonged QT interval, often regarded as an unwanted adverse effect of drugs. However, it has been argued that QT time by itself is not a sensitive parameter to predict TdP, because similar prolonged QT intervals may have distinct arrhythmogenic outcomes.1–7 Numerous parameters and ideas have been suggested to be more valuable,1,2,8–11 including the concept of diminished repolarization reserve.12 Decreased repolarization force at times of additional demand can explain the higher likelihood for TdP. For the proarrhythmic action of class III antiarrhythmic drugs, this can be viewed as the final hit in a predisposed heart, which, when unchallenged, functions adequately. Thus, repolarization reserve can be regarded as the ability of a heart to withstand a challenge on repolarization.
Creation of chronic AV block (CAVB) in the dog results in a high susceptibility for drug-induced TdP, most likely due to electrical remodeling.13,14 In our opinion, this animal model should have a reduced repolarization reserve. In the past, the proarrhythmic potential of several different class III antiarrhythmic drugs was determined, which showed a comparable and prominent QTc prolongation for all drugs but a low TdP incidence after amiodarone (0%) and d-sotalol (0% to 5%) compared with 56% to 67% TdP for other drugs tested.15 This led us to perform the present serial investigation in anesthetized CAVB dogs, in which the expected dose-dependent differential TdP occurrence after d-sotalol was used to compare the predictive value of different repolarization parameters, including beat-to-beat variability of repolarization (BVR).
Experiments were performed in 8 adult mongrel dogs with CAVB (5 males). In a preliminary operation, the AV node was ablated as described previously.14 Experiments reported here were started >5 weeks after ablation, which exceeded completion of ventricular electrical remodeling and proarrhythmia.14 Under complete anesthesia (induced by sodium pentobarbital 20 mg/kg IV and maintained by halothane), 2 monophasic action potential (MAP) catheters were advanced to the endocardium of left (LV) and right ventricle (RV). d-Sotalol was administered in 2 doses (2 and 4 mg/kg IV over 5 minutes) in a serial random crossover design (2±1 weeks between experiments). The natural sequence of events leading to drug-induced TdP in CAVB dogs can be divided into 2 phases: (1) prolongation of repolarization and (2) occurrence of (multiple) extrasystoles, which are often but not always followed by TdP.13 Dogs were considered inducible when 3 or more episodes of TdP occurred after a dose of d-sotalol. The IK,ATP opener levcromakalim (10 μg/kg IV) was administered in experiments with TdP to assess reversibility.
Transmembrane action potentials from enzymatically isolated myocytes from CAVB dogs16 were recorded (15 cells, 9 dogs). Microelectrodes (20 to 60 MΩ) filled with 3 mol/L KCl and pacing frequencies of 0.5 to 1 Hz were used. Action potentials were recorded under conditions of no IKr block versus full IKr block.
Mean RR and QT intervals from lead II and duration of the 2 MAPs to 100% repolarization (MAPD) of 30 consecutive beats were determined. Measurements were done before d-sotalol administration and in a period during which prolongation of repolarization had reached a plateau but the first drug-induced extrasystole had not yet occurred. Interventricular dispersion (ΔMAPD=LV−RV MAPD) and heart rate–corrected QT interval (QTc, van de Water’s formula) were calculated. These temporally averaged electrophysiological measurements will be referred to as EP parameters to discern them from parameters of lability.
To assess BVR, Poincaré plots were drawn by plotting each value against the former value (Figure 1 for LV MAPD2,17,18). It was performed for RR and QT intervals and for LV and RV MAPD from 30 consecutive beats under stable idioventricular focus. The areas of the plots were determined and their dimensions calculated. The mean orthogonal distance from the diagonal to the points of the Poincaré plot was determined and referred to as short-term variability (STV=Σ|Dn+1−Dn|/[30×√2], where D represents the duration of RR, QT, or MAP). The average distance to the mean of the parameter parallel to the diagonal (Σ|Dn+1+Dn−2Dmean|/[30×√2]) was regarded as long-term variability. This nomenclature is adopted from heart rate variability investigations using Holter monitoring of humans over several hours, in which steady changes over time tend to follow the diagonal and sudden changes (eg, an extrasystole) result in a deviation from the diagonal.18
The QT variability index (QTVI30) as proposed by Berger et al11 and instability30 as proposed by Hondeghem et al1,2 were calculated for QT and for LV and RV MAPD. Because of the fast onset of extrasystolic activity, these were calculated for 30 beats only and not over the proposed 256 seconds and 3 minutes, respectively.
In addition to assessing lability of repolarization at the 2 time points in the d-sotalol experiments, we assessed STV in data gathered in a study3 of 4 weeks of amiodarone treatment (40 mg · kg−1 · d−1 PO). Transmembrane action potential durations to 95% repolarization were measured for 30 action potentials from isolated ventricular myocytes to calculate STV at baseline and before early afterdepolarizations (EADs) or at full IKr block.
d-Sotalol–induced extrasystoles, defined as beats initiating before the end of the preceding T wave, were counted. Discrimination was made between couplets (Figure 2A) and multiples (Figure 2B), because the latter are considered more proarrhythmic.15
Pooled data are expressed as mean±SD unless otherwise stated. Comparisons of EP and lability parameters were performed with a 2-way ANOVA followed by a paired Bonferroni comparison. Countables were tested with a Mann-Whitney rank sum test. Correlations were tested with Pearson product moment correlation. The area under the receiver-operator characteristic curve was used to assess predictive power of variables. Statistical difference was acknowledged at P<0.05.
EP and Lability Parameters
Baseline EP and lability parameters were similar at the start of the 2 experiments (Table 1; Figure 1C). There was a dose-dependent difference in the induction of multiple extrasystoles (Table 1) and TdP after d-sotalol (2/8 after 2 mg/kg versus 6/8 after 4 mg/kg). d-Sotalol dose-independently prolonged the QT interval and LV MAPD, which indicates their inability to provide an explanation for the higher TdP incidence at 4 mg/kg d-sotalol (Table 1). RR interval and RV MAPD were only significantly prolonged after 4 mg/kg d-sotalol; however, these values were not different from those seen with 2 mg/kg d-sotalol.
With regard to lability parameters in general, differences between drug treatments were only seen for LV plot area and LV STV (Figure 3; Table 1), whereas QT plot area and instability30 were increased from baseline after the high dose. No changes were seen in the other lability parameters, such as all RR or RV MAPD plot areas or variability indexes (eg, QTVI30; Table 1). No correlations were identified between RR plot areas and QT or LV plot areas (P=0.7 and 0.4, respectively). Table 2 shows that the accuracy of individual predictions in the present study was considerably higher for LV plot area and STV than for temporally averaged QT or LV MAPD. Amiodarone treatment caused prolongation of repolarization (QT; control 340±40 ms, amiodarone 470±75 ms, P<0.053), whereas STV did not increase (2.4±0.2 to 2.4±0.4 ms, P=NS; Figure 3), in agreement with the total absence of TdP.
Poincaré Plot Area or STV
The advantage of STV over plot area is the feasibility to track its development over time by moving the 30-beats window. In Figure 4, this temporal behavior is shown for a representative dog. On average, half-maximal STV was reached 2.9±1.3 minutes earlier than the onset of ectopic beats (7±3 minutes after start of d-sotalol) and 10±6 minutes earlier than the first TdP. All changes in LV STV and plot area were reversible on IK,ATP activation with levcromakalim (Figure 3).
In the cells treated with IKr blockers, 5 of 15 showed EADs, which divided the population. STV in baseline was similar in these 2 groups and increased on IKr block (Figure 5). In the presence of IKr block, STV was significantly larger in the group with EADs, and the increased STV clearly preceded the occurrence of afterdepolarizations.
Parameters Indicative of Proarrhythmia: Lability of Repolarization
Over the years, several EP parameters that set the stage for drug-induced TdP have been suggested. Prolongation of QT interval and increased spatial dispersion of repolarization have been associated with TdP in patients19 and in experimental models.3,7,13,15,20 Still, the predictive power of these parameters is rather low (Table 2), and more sensitive methods have been sought.
As proposed by the groups of Berger et al11,21–23 and Hondeghem et al,1,2 lability of repolarization could bear important information to determine susceptibility to proarrhythmia in patients or to assess proarrhythmic potential of drugs. It is hypothesized that repolarization becomes labile when insufficient repolarization strength can be generated (decreased repolarization reserve12). For us, “lability of repolarization” includes all temporal assessments, regardless of whether they are gathered for strict beat-to-beat analyses (Poincaré plots), for subsequent calculations in which the order of beats are rearranged (eg, QTVI11), or for nonconsecutive determinations (eg, instability1,2). We consider the term BVR to be restricted to the quantifications derived from Poincaré plots, in which direct beat-to-beat information is taken into account.
Increased BVR Heralds TdP
Occurrence of TdP is dependent on the simultaneous presence of a number of different factors. The necessary accumulated magnitude of these factors depends on the underlying substrate. Drugs antagonizing repolarization are examples of such factors. The proarrhythmic substrate consists of various predisposing factors, which may be described as decreases in repolarization reserve.
In the present study, similar values of LV STV and plot area were found at baseline regardless of the proarrhythmic outcome of the drug (Figure 1). Antagonizing repolarization by d-sotalol elevated LV STV and plot area in a pattern that corresponded to TdP incidence. Prevention of further TdP by the IK,ATP opener levcromakalim returned LV STV and area to baseline levels, which indicates that strengthening of repolarization reserve by increasing outward current can decrease STV (Figure 3). The amiodarone data pointed in the same direction, because a prominent and comparable QT prolongation was seen as with other drugs, but LV STV remained low (Figure 3), and TdP was absent.3
When the experiments are arbitrarily grouped either by proarrhythmic outcome or by dose (Figure 6), it can be seen that the area of the LV Poincaré plot, which could be substituted with LV STV, is closely associated with TdP incidence. On the other hand, the QT interval was equally increased in all groups and did not correlate with TdP induction. The relationship between sensitivity and specificity of some parameters is illustrated in Table 2, which shows a higher predictive power for LV plot area and LV STV. As mentioned above, we prefer LV STV over plot area, because relatively minor computational allocations facilitate online calculation of STV (Figure 4).
In the present study, QTVI30 did not provide similar information, whereas instability30 did, to some extent (Table 1). It must be emphasized that we adapted the 2 methodologies for our 30-beats comparison for methodological reasons.
Table 2 shows that BVR of the LV is superior to the regular EP parameters in the present study. Although not statistically different, one could read a trend toward a dose-dependent increase in RR plot area after d-sotalol (Table 1). There was no correlation between the LV plot areas and RR plot areas or regular RR intervals (P=0.34). Thus, any possible influence of heart rate is not evident in the phase before the onset of extrasystoles.
STV Is an Early Indicator of TdP
In addition to EP and lability parameters, extrasystoles that cause short-long-short morphology of RR intervals have often been proposed as a signal of proarrhythmia.13,24 Previously, we demonstrated that TdP can be induced reproducibly by pacing when d-sotalol does not provide the trigger.13 In the present study, we have shown that the multiple extrasystoles appear dose dependently and at similar QT intervals but presumably at higher plasma or tissue levels of d-sotalol. Therefore, there appears to be a safety window (Figure 4) during which maximal repolarization times are reached before TdP ensues. If BVR increases, TdP is likely to follow. The increase in BVR before the onset of extrasystolic activity suggests a sequence of events that starts with prolongation of repolarization followed by increased BVR, and later, extrasystolic activity and TdP. Because prolongation of repolarization appears to be a dose-independent prerequisite, BVR is the earliest identified determinant of arrhythmias in the present study. Furthermore, it is illustrated (Figure 4) that an abrupt increase in STV occurs approximately at the end of the infusion, which leaves ≈10 minutes’ response time to prevent TdP.
Repolarization Lability Is Confined to the LV
In contrast to LV MAPD, the Poincaré plot area of the QT interval was only significantly increased after 4 mg/kg and was not different from that seen with 2 mg/kg d-sotalol (Table 1). The global nature of the QT interval versus the localized subendocardial LV MAP signal could illustrate regional lability of the LV, setting the stage for arrhythmias. Like discordant T-wave alternans,25 anatomically adjacent regions of the LV with oscillating repolarization durations that are out of phase could confer substantial dispersion, setting the stage for TdP. A MAP catheter registering 1 of these regions would show large BVR, whereas a global ECG would reveal the average repolarization time of all of these regions and thus a relatively lower BVR.
Cellular Origin of BVR
In vivo, we opted to measure BVR before ectopic beats were present. To answer the question of whether BVR has a cellular origin and appears before EADs, we performed single-cell experiments with IKr-blocking drugs. We noticed that these drugs increased STV and that STV was increased to higher levels in cells that demonstrated EADs (Figure 5). Hence, we suggest that BVR precedes EADs and that the higher the STV, the greater the likelihood for EADs.
This study indicates that increased BVR precedes the induction of TdP arrhythmias in CAVB dogs. Cardiovascular and noncardiovascular drugs that induce TdP in patients with hidden predisposing factors are a major concern because identification of these patients is cumbersome.19 Assessment of BVR could identify susceptible patients and provide a parameter together with QT intervals. Patients with congenital long-QT syndrome exhibit higher repolarization variability than unaffected family members, despite comparable heart rate variabilities.26 This suggests that lability of repolarization may be present both in acquired (drug induced) and hereditary repolarization-reserve deficiencies.
This study limits BVR to invasive, endocardial MAP recordings in anesthetized dogs with a high susceptibility for drug-induced TdP. Whether it can be applied to patients at risk for drug-induced TdP is unknown. It is also unknown whether other noninvasive signals can be used to quantify BVR. Before extrapolations between in vitro and in vivo BVR and proarrhythmia can be made, we must address numerous factors such as adrenergic drive, cellular coupling, electrolyte levels, and changes in preload.
The presence or absence of proarrhythmia is not related to differences in prolongation of repolarization parameters but corresponds to BVR, such as STV of LV MAPD. The increase in STV occurred before extrasystoles in vivo and EADs in vitro, which indicates that STV is a candidate parameter to predict drug-induced TdP.
M.B. Thomsen and S.C. Verduyn were financially supported by H. Lundbeck, Denmark, and Janssen Research, Belgium, respectively. P.G.A. Volders was supported by The Netherlands Organization for Health and Development (ZonMw 906-02-068).
Hondeghem LM, Dujardin K, De Clerck F. Phase 2 prolongation, in the absence of instability and triangulation, antagonizes class III proarrhythmia. Cardiovasc Res. 2001; 50: 345–353.
Hondeghem LM, Carlsson L, Duker G. Instability and triangulation of the action potential predict serious proarrhythmia, but action potential duration prolongation is antiarrhythmic. Circulation. 2001; 103: 2004–2013.
van Opstal JM, Schoenmakers M, Verduyn SC, et al. Chronic amiodarone evokes no torsade de pointes arrhythmias despite QT lengthening in an animal model of acquired long-QT syndrome. Circulation. 2001; 104: 2722–2727.
Carlsson L, Abrahamsson C, Andersson B, et al. Proarrhythmic effects of the class III agent almokalant: importance of infusion rate, QT dispersion, and early afterdepolarisations. Cardiovasc Res. 1993; 27: 2186–2193.
Thomsen MB, Volders PG, Stengl M, et al. Electrophysiological safety of sertindole in dogs with normal and remodeled hearts. J Pharmacol Exp Ther. 2003; 307: 776–784.
Milberg P, Eckardt L, Bruns HJ, et al. Divergent proarrhythmic potential of macrolide antibiotics despite similar QT prolongation: fast phase 3 repolarization prevents early afterdepolarizations and torsade de pointes. J Pharmacol Exp Ther. 2002; 303: 218–225.
Berger RD, Kasper EK, Baughman KL, et al. Beat-to-beat QT interval variability: novel evidence for repolarization lability in ischemic and nonischemic dilated cardiomyopathy. Circulation. 1997; 96: 1557–1565.
Verduyn SC, Vos MA, van der Zande J, et al. Further observations to elucidate the role of interventricular dispersion of repolarization and early afterdepolarizations in the genesis of acquired torsade de pointes arrhythmias: a comparison between almokalant and d-sotalol using the dog as its own control. J Am Coll Cardiol. 1997; 30: 1575–1584.
Schoenmakers M, Ramakers C, van Opstal JM, et al. Asynchronous development of electrical remodeling and cardiac hypertrophy in the complete AV block dog. Cardiovasc Res. 2003; 59: 351–359.
Volders PG, Sipido KR, Vos MA, et al. Cellular basis of biventricular hypertrophy and arrhythmogenesis in dogs with chronic complete atrioventricular block and acquired torsade de pointes. Circulation. 1998; 98: 1136–1147.
Volders PG, Stengl M, van Opstal JM, et al. Probing the contribution of IKs to canine ventricular repolarization: key role for β-adrenergic receptor stimulation. Circulation. 2003; 107: 2753–2760.
Kaab S, Hinterseer M, Nabauer M, et al. Sotalol testing unmasks altered repolarization in patients with suspected acquired long-QT-syndrome: a case-control pilot study using i.v. sotalol. Eur Heart J. 2003; 24: 649–657.
Atiga WL, Fananapazir L, McAreavey D, et al. Temporal repolarization lability in hypertrophic cardiomyopathy caused by β-myosin heavy-chain gene mutations. Circulation. 2000; 101: 1237–1242.
Walker ML, Rosenbaum DS. Repolarization alternans: implications for the mechanism and prevention of sudden cardiac death. Cardiovasc Res. 2003; 57: 599–614.