Further Observations to Elucidate the Role of Interventricular Dispersion

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

JACC Vol. 30, No. 6 1575 November 15, 1997:1575–84

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

S. CORA VERDUYN, PHD, MARC A. VOS, PHD, JOLANDA VAN DER ZANDE, BS, ATILLA KULCSA` R, PHD, HEIN J. J. WELLENS, MD, FACC Maastricht, The Netherlands

Objectives. We sought to further elucidate the role of early 60 ms; d-sotalol: 80 ؎ 45 ms, p < 0.05). The incidence of EADs (18 ؎ afterdepolarizations (EADs) and interventricular dispersion of of 22 vs. 11 of 24, p < 0.05) and single ectopic beats (EBs) (1.5 repolarization (⌬APD) in the genesis of acquired torsade de 2 vs. 24 ؎ 32, p < 0.01) was more frequently observed after pointes (TdP) arrhythmias. almokalant than after d-sotalol. Moreover, multiple EBs only Background. Administration of class III agents can be associ- occurred after almokalant. These beats interfered with the basic ated with TdP. We developed a dog model in which TdP can be rhythm, leading to dynamic changes in left ventricular APD and to reproducibly induced by pacing after d-sotalol. This model shows additional increases in ⌬APD. Spontaneous TdP was observed in reproducible results over weeks. 9 of 14 dogs after almokalant and could be increased to 12 of 14 Methods. In 14 anesthetized dogs with chronic complete atrio- with programmed electrical stimulation. After d-sotalol, TdP ventricular block, two separate experiments were performed in could only be induced by programmed electrical stimulation (5 of which d-sotalol (2 mg/kg body weight) or almokalant (0.12 mg/kg) 14, p < 0.05). was administered. Monophasic action potentials were simulta- Conclusions. In the same dog, almokalant induced more delay neously recorded from the endocardium of the right and left in repolarization, more EADs, multiple EBs and more ventricular ventricle to register EADs and to measure the action potential inhomogeneity in APD than d-sotalol. These changes were related duration (APD). ⌬APD was defined as the APD of the left to a higher incidence of TdP and thereby confirm a strong ventricle minus that of the right ventricle. association of the occurrence of EADs, multiple EBs and ⌬APD in Results. Baseline conditions were identical in the serially the genesis of TdP. These findings also show the possible value of performed experiments. The cycle length and QT time increased our model for evaluating the proarrhythmic potential of different by 16% and 26% after d-sotalol and by 15% and 31% after drugs. almokalant, respectively. After both drugs the action potential of (J Am Coll Cardiol 1997;30:1575–84) the left ventricle prolonged more than that of the right ventricle, ©1997 by the American College of Cardiology ؎ thereby increasing ⌬APD (almokalant [mean ؎ SD]: 110

Antiarrhythmic drugs that prolong repolarization without af- sade de pointes (TdP) arrhythmias in ϳ1% to 5% of patients fecting conduction have attracted interest because of their (1,2,4–9). possible value in the prevention and suppression of reentrant Many reports have pointed to the relevance of early after- tachycardias (1–3). However, class III drugs do produce tor- depolarizations (EADs) and EAD-dependent ectopic beats (EBs) for the initiation of TdP arrhythmias (9–13). Whether they are solely responsible for the initiation of TdP arrhyth- From the Department of Cardiology, Cardiovascular Research Institute, mias or whether other factors such as ventricular dispersion of University of Limburg, Maastricht, The Netherlands. This study was supported by Grant 91.104 from The Netherlands Heart Foundation, The Hague. ASTRA repolarization also contribute is still a matter of discussion Nederland, Rijswijk, The Netherlands provided the almokalant; Bristol Myers (14,15). Regional differences in the duration of the action Squibb, Woerden, The Netherlands provided the d-sotalol; and the Bakken potential can be present within the ventricular wall (transmu- Research Institute (Medtronic), Maastricht, The Netherlands provided the epicardial electrodes used in this study. ral), within one ventricle (intraventricular) or between ventri- Manuscript received August 17, 1995; revised manuscript received July 18, cles (interventricular). 1997, accepted August 12, 1997. In anesthetized dogs with chronic complete atrioventricular Address for correspondence: Dr. Marc A. Vos, Department of Cardiology, Cardiovascular Research Institute Maastricht, University of Limburg, P.O. Box (AV) block, TdP could be reproducibly initiated using the 5800, 6202 AZ Maastricht, The Netherlands. E-mail: [email protected]. combination of d-sotalol and pacing in 50% of experiments

©1997 by the American College of Cardiology 0735-1097/97/$17.00 Published by Elsevier Science Inc. PII S0735-1097(97)00333-1 1576 VERDUYN ET AL. JACC Vol. 30, No. 6 RELEVANT FACTORS FOR THE INITIATION OF TdP November 15, 1997:1575–84

agents (0.015 mg/kg of intramuscular buprenorfine). A tempo- Abbreviations and Acronyms rary ventricular pacemaker (VVI) was sometimes placed after APD ϭ action potential duration the AV block operation and after the experiments. Pacing was AV ϭ atrioventricular switched off after a maximum of 24 h. EADs ϭ early afterdepolarizations Determination of almokalant dose. In dogs in sinus EBs ϭ ectopic beats LV ϭ left ventricular rhythm, other investigators have shown (19) that 0.35 mg/kg of MAP ϭ monophasic action potential almokalant had approximately two times the effect on repolar- RV ϭ right ventricular ization as 3 mg/kg of d-sotalol. In our previous studies (16–18) TdP ϭ torsade de pointes we used 2 mg/kg d-sotalol to induce TdP by pacing. Because we ⌬APD ϭ interventricular dispersion wanted to obtain a similar effect on action potential duration (APD) prolongation, a dose of 0.12 mg/kg of almokalant (1/3 of 0.35 mg/kg ϭ 0.12) was chosen. In the 14 dogs tested, d-sotalol was given first ([mean Ϯ SD] 4 Ϯ 2.2 weeks after creation of complete AV block) in 11 dogs, (16). We compared inducible versus noninducible TdP in this followed by almokalant 3 Ϯ 2.1 weeks later. In three dogs the canine model and demonstrated (17) that the number of EADs order was reversed, with almokalant given at 5 Ϯ 2.6 weeks and is higher and the amount of interventricular dispersion larger d-sotalol at 9 Ϯ 3.1 weeks of chronic complete AV block. than in inducible dogs. Because the response is maintained Induction of TdP arrhythmias. A detailed description of over weeks (16), we compared the effects of two antiarrhythmic the TdP protocol is described elsewhere (16). In short, at least drugs, d-sotalol and almokalant, in their ability to induce TdP 2 weeks after creation of complete AV block, anesthetized arrhythmias in the same dog. In this way we could assess whether differences in the occurrence of EADs, EAD- animals received two defibrillation patches that were attached dependent EBs or interventricular dispersion or repolarization to both sides of the chest and connected with a defibrillator. At could explain spontaneous initiation of TdP. least 30 min after the onset of anesthesia, programmed electrical stimulation was performed from the epicardial electrode. Stimulation was done with a programmable stimulator capable Methods of pacing synchronously to the QRS complexes. Unipolar stimuli were given using a pulse of 2 ms and a stimulus strength The study protocol was approved by the Committee for of twice diastolic threshold. As an indifferent electrode, a Experiments on Animals of the University of Limburg, Maas- tricht, The Netherlands and was conducted in accordance with needle was placed through the skin. the guidelines of the American Physiological Society. Programmed electrical stimulation consisted of two differ- General protocol. The experiments were performed in ent pacing protocols: 1) a short–long–short sequence (400 (ϩ extrastimulus), and 2 1,200 600ءanesthetized adult male and female mongrel dogs with a body 800 ϩ extrastimulus, or 4 weight between 20 and 31 kg. In a preliminary operation, a eight basic stimuli followed by an extrastimulus. The inter- right thoracotomy was performed to induce a permanent stimulus intervals were 600 or 1200 ms. The extrastimulus complete AV block by injection of 37% formaldehyde into the interval in both pacing protocols was shortened from 500 ms AV junction (18). During the same session, a pacing electrode using steps of 50 ms until 300 ms. After completion of the basic (Bakken Research Center, Medtronic) was inserted at the apex pacing protocol, d-sotalol (2 mg/kg for 5 min) or almokalant of the left ventricle. The wire was exteriorized through the (0.12 mg/kg for 10 min) was administered. back of the neck of the dog. Six surface electrocardiographic Pacing was resumed 10 min after the start of the infusion, leads and two endocardial monophasic action potential (MAP) unless spontaneous TdP had occurred during the observation signals were simultaneously registered and stored on optical period. When TdP occurred, we tried to perform the pacing disc. All drugs were administered through a cannula in the protocol at 15 min. Pacing was always performed in random cephalic vein. order. A TdP arrhythmia was defined as a polymorphic ventric- Anesthesia was induced by 1) intramuscular premedication ular tachycardia consisting of Ն5 beats twisting around the (1 ml/5 kg: 10 mg of oxycodon, 1 mg of acepromazine and baseline in the setting of a prolonged QT(U) duration. TdP 0.5 mg of atropine), and 2) sodium pentobarbital (20 mg/kg was terminated using cardioversion (60 to 70 J) when it lasted body weight intravenously). The dogs were artificially venti- Ͼ10 s. A dog was said to have inducible TdP when TdP was lated through a cuffed endotracheal tube using a mixture of induced three times or more using the same pacing mode or oxygen, nitrous oxide and halothane (vapor concentration when spontaneous initiation of TdP occurred (three times or 0.5% to 1%) by a respirator. Ventilation was controlled by more). In one experiment, problems with cardioversion urged continuous reading of the carbon dioxide concentration in the us to stop prematurely, so that reproducibility could not be expired air. A thermal mattress was used to maintain adequate assessed. body temperature. MAPs. MAPs were recorded to observe the occurrence of Proper care was taken before and after the experiments, EADs and to measure the duration of the action potential of including antibiotic (1,000 mg of ampicillin) and analgesic the left and right ventricles. Quadripolar contact electrodes JACC Vol. 30, No. 6 VERDUYN ET AL. 1577 November 15, 1997:1575–84 RELEVANT FACTORS FOR THE INITIATION OF TdP

Table 1. Electrophysiologic Effects of d-Sotalol and Almokalant d-Sotalol Almokalant Baseline 10 min % Baseline 10 min % (ms) (ms) Inc (ms) (ms) Inc CL IVR 1,620 Ϯ 300 1,880 Ϯ 370* ϩ16 1,670 Ϯ 360 1,910 Ϯ 425* ϩ15 QT 390 Ϯ 50 495 Ϯ 80* ϩ26 415 Ϯ 35 545 Ϯ 115* ϩ31 LV APD 370 Ϯ 45 470 Ϯ 60* ϩ27 375 Ϯ 45 545 Ϯ 95*† ϩ45† RV APD 320 Ϯ 30 390 Ϯ 45 ϩ21 335 Ϯ 35 435 Ϯ 60*† ϩ29† ⌬APD 50 Ϯ 25 80 Ϯ 45* ϩ75 40 Ϯ 20 110 Ϯ 60*† ϩ330 *p Ͻ 0.05, 10 min versus baseline. †p Ͻ 0.05, almokalant versus d-sotalol. Data presented are mean value Ϯ SD. APD ϭ action potential duration; CL IVR ϭ cycle length of idioventricular rhythm; Inc ϭ increase; LV ϭ left ventricular; RV ϭ right ventricular; ⌬APD ϭ LV APD Ϫ RV APD.

(Franz combination catheter, EPT No. 1650) that provide both Results pacing and MAP recording capabilities were placed endocar- Baseline conditions were similar in each of the 14 serially dially in the right and left ventricles through the external tested animals (Table 1). jugular vein and the carotid artery under fluoroscopic guid- Electrophysiologic and proarrhythmic effects of d-sotalol. ance. MAP phases were defined according to the definitions Values for cycle length, QT duration, APD of both ventricles used for transmembrane potentials (20). Amplitude was de- and ⌬APD are summarized in Table 1. At 10 min, d-sotalol fined between phases 4 and 2 of the signal. Besides a minimal had significantly prolonged cycle length by 16% and QT amplitude of 15 mV, the MAP had to have a constant duration by 26% and also significantly increased the APD of configuration and a smooth shape during control circum- both the left and right ventricles. Because prolongation of the stances. The MAP catheters were randomly placed in the action potential of the left ventricle was more pronounced ventricle and accepted for analysis if all these conditions were (27% vs. 21%), ⌬APD increased significantly (Table 1). fulfilled. MAPs of good quality in both ventricles were present In the first 10 min, EADs were seen in 11 of 24 MAPs after in 11 of 14 experiments with almokalant and in 12 of 14 d-sotalol, which tended to appear more frequently in the left experiments with d-sotalol. EADs were defined as an interrup- than in the right ventricle (p Ͻ 0.1) (Table 2). Single EBs tion of the smooth contour of phase 2 or 3 of the action developed in Ͻ50% of the dogs (Table 2), with a maximum of potential (10). The presence of EADs was examined in both 5 during 10 min (range 0 to 5). No dog developed spontaneous MAPs. TdP after d-sotalol. Data analysis. With the use of a custom-made computer The time-dependent electrophysiologic effects (0 to 10 min) program with a resolution of 2 ms and adjustable gain and time for left and right ventricular APDs are illustrated for a specific scale, the following variables were measured every 30 s during dog in Figure 1. It can be clearly seen that the left ventricular the first 10 to 15 min after the start of the class III medication: APD increased more than the right ventricular APD, leading cycle length of the idioventricular rhythm, QT time and APD. to a maximum ⌬APD of 150 ms at 8 min. In this dog, no EBs In case EBs occurred, measurements were made solely in occurred during this period, so there are no dynamic changes single beats. Electrophysiologic data in the text and tables are in the APD. the mean of 5 consecutive (single) beats. In the graphs the Electrophysiologic and proarrhythmic effects of al- individual values are reported. All values were verified by an mokalant. Almokalant also significantly increased the cycle independent observer (A.K.) in blinded manner. Interventricu- length and all repolarization variables. ⌬APD again increased lar dispersion (⌬APD) was defined as the difference between (p Յ 0.05) as a result of a more pronounced increase in the the left and right APD measured after total repolarization APD of the left than in the right ventricle (45% vs. 29%) (APD100). Also, the number of spontaneous EBs after both (Table 1, Fig. 2 to 4). In the first 10 min, almokalant admin- drugs during this period were scored as single and multiple istration resulted in the occurrence of EADs in 18 of 22 MAPs EBs (i.e., doublets, triplets or quadruplets). (Table 2) that occurred in both ventricles and contributed to Statistics. Multiple measures analysis of variance, followed ⌬APD. Single EBs developed in almost all dogs with al- by a Bonferroni t test, was used to compare the data between mokalant. The maximal number was 109 single EBs/10 min the two treatments, and a chi-square test was used for data as (mean 24 Ϯ 32). Moreover, multiple EBs were observed percentages. A p value Յ0.05 was considered significant. (Table 2), with a maximum of 13 doublets (4.2 Ϯ 3.9), 8 triplets Results are presented as mean value Ϯ SD. Furthermore, a (2.3 Ϯ 2.7) and 2 quadruplets (0.4 Ϯ 0.7). In some experiments regression analysis was performed to calculate the correlation these spontaneous beats could be associated with triggering of between QT time and the other repolarization variables (left EADs recorded with the MAP catheter (Fig. 4). and right ventricular APDs and ⌬APD). In the first 10 min, almokalant induced repetitive sponta- 1578 VERDUYN ET AL. JACC Vol. 30, No. 6 RELEVANT FACTORS FOR THE INITIATION OF TdP November 15, 1997:1575–84

Table 2. Incidence of Early Afterdepolarizations and Torsade de Pointes Arrhythmias After Administration of d-Sotalol and Almokalant EADs EBs Indu of TdP Dog No. LV/RV dS A dS A dS A 1LV ϩϩNone Single No No RV ϩϩ 2LV ϩϩSingle Multiple PES Spon* RV Ϫϩ 3LV ϩϩNone Multiple No Spon RV Ϫϩ ϩPES 4LV ϪϩSingle Multiple PES PES RV Ϫϩ 5LV ϪϩNone Multiple No Spon* RV Ϫϩ 6LV ϩϩNone Multiple PES Spon RV ϩϪ ϩPES 7LV ϩϩNone Multiple PES Spon* RV Ϫϩ 8LV ϩϩNone Multiple No Spon RV Ϫϩ ϩPES 9LV ϩϪSingle Single PES PES RV ϪϪ 10 LV ϩϩSingle Multiple No Spon RV ϩϪ ϩPES 11 LV ϪϩNone Multiple No PES RV Ϫϩ 12 LV Ϫ ND None Multiple No Spon RV Ϫ 13 ND Single Multiple No Spon* 14 ND None None No No Total 11/24 18/22† 5/14 13/14 5/14 12/14† *Because of continuous spontaneous (Spon) activity, no programmed electrical stimulation (PES) could be performed. †p Ͻ 0.05 versus d-sotalol (dS). A ϭ almokalant; EADs ϭ early afterdepolarizations; EBs ϭ ectopic beats (multiple ϭ 2 to 4 ectopic beats); Indu ϭ inducibility; LV ϭ left ventricle; ND ϭ not determined; RV ϭ right ventricle; TdP ϭ torsade de pointes. neous TdP in nine dogs (10 Ϯ 15 periods) (Fig. 4 and 5). The However, when the separate conditions were taken into con- time-dependent electrophysiologic effect of almokalant (0 to sideration (i.e., baseline vs. class III), the correlation between 8 min) is illustrated in Figure 2. In this dog, almokalant QT and ⌬APD is absent, whereas the correlation between QT induced single EBs at 200 s (Fig. 3), multiple EBs at 280 s and and LV and RV APD remained present (Table 3). spontaneous TdP at 6 min (Fig. 4). The occurrence of EBs was Comparison of d-sotalol and almokalant. When d-sotalol associated with dynamic changes leading to additional in- and almokalant were compared, no differences were found in creases in the amount of ⌬APD due to the changes in LV APD the amount of prolongation of the cycle length. However, the in the post-extrasystolic beat (Fig. 3). Before TdP, ⌬APD was effect on the repolarization variables was greater after al- ϳ200 ms (Fig. 2 and 4). mokalant, including ⌬APD (Table 1). The number of EADs Compared with the animals given almokalant that did not and single EBs was also higher after almokalant (1.5 Ϯ 2 vs. develop spontaneous TdP, the occurrence of TdP was associ- 24 Ϯ 32 beats/10 min, p Յ 0.01). In addition, multiple EBs ated with 1) a significantly larger ⌬APD (140 Ϯ 55 vs. 75 Ϯ were only seen after almokalant (Table 2). These EBs inter- 40 ms, p Յ 0.05) and LV APD, and 2) the presence of multiple fered with the idioventricular rhythm, leading to APD adapta- EBs (Table 2). No significant difference in the number of single tions. The fact that this APD adaptation was most pronounced EBs was observed between the two groups (19 Ϯ 10 vs. 30 Ϯ 50 in the left ventricle led to further increases in the ⌬APD and to EBs). the occurrence of spontaneous TdP. Regression analysis between QT duration and other repo- Induction of TdP after programmed electrical stimulation. larization variables. The measured repolarization with the After d-sotalol, programmed electrical stimulation resulted in MAP catheters showed good correlation with the QT duration the induction of TdP in five dogs. Programmed electrical (Table 3) (LV APD: y ϭ 0.83x ϩ 51; RV APD: y ϭ 0.55x ϩ stimulation could not be performed in four of nine dogs with 109). There was also a positive although weak correlation spontaneously inducible TdP almokalant because ectopic ac- between ⌬APD and QT duration (y ϭ 0.27x Ϫ 58, r2 ϭ 0.32). tivity and periods of TdP prohibited this (Fig. 5, Table 2). JACC Vol. 30, No. 6 VERDUYN ET AL. 1579 November 15, 1997:1575–84 RELEVANT FACTORS FOR THE INITIATION OF TdP

Figure 1. Time-dependent behav- ior of LV and RV APD during and after administration of d-sotalol. LV and RV APD is depicted (y- axis) just before and up to 20 min after the start of d-sotalol (x-axis). d-Sotalol increases the LV and RV APD with different sensitivity: The increase is more pronounced for the LV APD, leading to an increase in ⌬APD from 75 ms (0 min) to 150 ms at 8 min. During the observation period of 10 min, EADs were observed in the left ventricle; however no EBs developed, nor did a spontaneous TdP arrhythmia arise. Therefore, no dynamic changes in APD length occur, and the curves of both APDs are smooth. Programmed electrical stimulation is performed starting at 20 min after the start of d-sotalol. The APD of both ventricles is de- creased, leading to a ⌬APD of ϳ60 to 90 ms. Pacing does not lead to the induction of TdP (see Fig. 6).

Pacing induced TdP in four of the other ﬁve dogs (Table 2). Of before pacing (Fig. 1 and 6). The APDs are smooth, and no the ﬁve dogs without spontaneous arrhythmias, pacing resulted ectopic activity or TdP can be induced. After almokalant, this in TdP in three. Therefore, a total of 12 of 14 dogs (9 sponta- dog demonstrated an ⌬APD of 115 ms, with EADs in both neous, 3 after pacing) developed TdP after almokalant. ventricles. Pacing now resulted in the induction of a spontane- A representative illustration of the difference in response to ously terminating TdP. Dogs with inducible TdP by pacing had d-sotalol and almokalant in the same dog is shown in Figure 6 a higher ⌬APD (120 Ϯ 55 vs. 75 Ϯ 30 ms, p Յ 0.05) than dogs (the same dog as in Fig. 1). After d-sotalol, ⌬APD is 90 ms with noninducible TdP.

Figure 2. Time-dependent (dynamic) behav- ior of LV and RV APD during and after administration of almokalant. Format as in Figure 1. Under control condition 1 (see also Panel 1, Fig. 3), the LV APD is longer than the RV APD, leading to a ⌬APD of 35 to 45 ms. Directly after the start of almokalant, the LV and RV APD start to increase. This increase is not synchronous: The LV APD increase is more pronounced, leading to a more pronounced ⌬APD. By ϳ3 min (condition 2 [see also right panel, Fig. 3]), single EBs develop that are preceded by the occurrence of EADs in the left ventricle (see also right panel, Fig. 3). These EBs lead to variations in the LV APD and to a more pronounced ⌬APD. By ϳ4.5 min after the start of almokalant, multiple EBs (mEBs) develop, resulting in a ⌬APD of 160 ms. Just before the spontaneous TdP arrhythmia (condition 3 [see also Fig. 4]), the ⌬APD has reached a value of 200 ms because of the dynamic reaction of the LV APD to the multiple EBs. Each acceleration in rate by the EBs is followed by a transient decrease, especially in the LV APD, leading to a shorter ⌬APD and a temporary inability of TdP to occur. 1580 VERDUYN ET AL. JACC Vol. 30, No. 6 RELEVANT FACTORS FOR THE INITIATION OF TdP November 15, 1997:1575–84

Figure 3. Occurrence of EADs and single EBs after almokalant (0.12 mg/ kg for 10 min). Three electrocardiographic leads (I, II and aVR) and two MAPs (left and right ventricles) are illustrated at a paper speed of 25 mm/s. These tracings are from data from the same dog as in Figure 2. Baseline conditions are shown in Panel 1. The action potentials are smooth, and ⌬APD is 35 ms. Three minutes after the start of almokalant (Panel 2), the QT interval and APDs are increased. An EAD is visible in the LV MAP (arrow), and the ⌬APD has increased to 95 ms. EBs arise, leading to variations in the LV APD, from 435 to 505 ms. Because the RV APD in this dog does not show similar changes, ⌬APD varies accordingly from 85 to 140 ms.

Sequence of drug administration. When d-sotalol was accepted variables, bradycardia and (class III induced) pro- compared with almokalant, all dogs with inducible TdP during longed QT time, EADs and dispersion of repolarization have d-sotalol also had inducible TdP with almokalant (Table 2). been indicated. By using MAP catheters, the in vivo appear- These observations were independent of the sequence of ance of EADs has been associated with EAD-triggered EBs experiments. In two dogs, d-sotalol was administered in a third and a changed or more marked configuration of the T or TU experiment. The response was similar to that in the first waves (21). Similarly, an increased intraventricular dispersion experiment, thereby confirming reproducibility. has been found in patients with congenital or acquired TdP (12). However, only one case report has shown a causal Discussion relation between the occurrence of EAD-triggered EBs and ⌬APD of repolarization to initiate spontaneous acquired TdP The exact contribution of the different factors leading to the (15). In the present study we tried to elucidate the relevance of initiation of TdP is still a matter of discussion. Besides the both variables and acquired spontaneous and pacing-induced

Figure 4. Short-lasting spontaneous TdP arrhythmia after almokalant. Same dog as in Figure 3. Six minutes after almokalant, more EADs have developed (arrow) that now clearly trigger multiple EBs (aster- isk). The ⌬APD has increased to 200 ms. This sequence of events leads to a spontaneous, short-lasting TdP arrhythmia. JACC Vol. 30, No. 6 VERDUYN ET AL. 1581 November 15, 1997:1575–84 RELEVANT FACTORS FOR THE INITIATION OF TdP

Figure 5. Repetitive ectopic activity and TdP after almokalant. Recordings are shown at a paper speed of 10 mm/s. Before administration of almokalant (CONTROL), a stable cycle length (1,510 ms) with a QT duration of 430 ms is present. Administration of almokalant (0.12 mg/kg for 10 min) led to the occurrence of EBs at 3 min. From that time to 30 min after the start of the infusion, episodes of multiple EBs and TdP alternated with the normal idioventricular rhythm. In this ﬁgure a continuous recording of 225 s starting 5 min after start of almokalant is shown. Almokalant has increased the cycle length to 1,745 ms and the QT duration to 680 ms. Furthermore, the T wave has become more negative. At the end of this T wave, an EB starts a period of TdP. After cessation, another period of TdP evolves. After a change in the site of origin of the idioventricular rhythm (different QRS conﬁgurations), a new episode of TdP now starts after a short–long–short sequence.

TdP by comparing the proarrhythmic effects of almokalant ing on the specific circumstances (baseline or class III), we put with d-sotalol. less emphasis on bradycardia and demonstrated the lack of Factors contributing to interventricular dispersion. In a correlation between ⌬APD and QT duration, indicating that recent paper (17), we demonstrated that ⌬APD is bradycardia ⌬APD is an independent factor. dependent. Both d-sotalol and almokalant prolong the cycle Similar findings were observed with MgSO4 in a previous length of the idioventricular rhythm equally. Almokalant in- report (17), where MgSO4 administration after d-sotalol re- creased the different repolarization variables to a greater turned the ⌬APD to baseline, but the other repolarization extent than d-sotalol. In addition, EADs occurred more fre- variables (QT duration and both APDs) where still signifi- quently after almokalant, contributing to ⌬APD. Furthermore, cantly prolonged (17). we indicated that the appearance of especially multiple EBs Factors associated with spontaneous TdP. In order of leads to further prolongation of the ⌬APD. This abnormal rate appearance, spontaneous occurrence of TdP after almokalant adaptation is especially observed in the LV APD and is present (Fig. 2 to 4) was preceded by 1) an inhomogeneous increase in in the post-extrasystolic beat after temporary increases in rate. left and right ventricular APD, 2) the occurrence of (sub- Longer periods of a increased rate show a normal rate threshold) EADs, 3) the occurrence of single followed by adaptation (i.e., shortening of the APD). multiple EBs, and 4) a further increase in ⌬APD resulting QT duration can be best predicted on the basis of the LV from the dynamic response of the left ventricular APD because APD. Because the LV APD determines ⌬APD, the latter of frequency changes. Spontaneous TdP was associated with variable is also correlated with QT duration, although the the appearance of multiple EBs and with a large ⌬APD. The regression coefficient is much lower (Table 3). By concentrat- EBs seemed to be generated by the EADs, as clearly shown in Figure 4 and indicated by the fact that they always appeared Table 3. Regression Analysis of QT Duration and Other after the occurrence of EADs. Repolarization Variables It should be mentioned that in the experiments with almokalant, electrophysiologic measurements were often ham- All Values Baseline Class III Agent (n ϭ 42) (n ϭ 21) (n ϭ 21) pered by the recurrent presence of EBs and episodes of TdP. This recurrence will lead to an underestimation of the electro- 2 2 2 r p Value r p Value r p Value physiologic effects of almokalant. LV APD 0.7 Ͻ 0.01 0.64 Ͻ 0.01 0.47 Ͻ 0.01 Factors associated with pacing-induced TdP. When spon- RV APD 0.62 Ͻ 0.01 0.58 Ͻ 0.01 0.34 Ͻ 0.01 taneous episodes of TdP occur, pacing will normally also lead ⌬APD 0.33 Ͻ 0.01 0.15 NS 0.12 NS to the reproducible induction of these arrhythmias (Table 2) Abbreviations as in Table 1. (16). As shown in a previous study (17) and confirmed in the 1582 VERDUYN ET AL. JACC Vol. 30, No. 6 RELEVANT FACTORS FOR THE INITIATION OF TdP November 15, 1997:1575–84

Figure 6. Effect of programmed electrical stimulation on inducibility In vitro studies (21) using isolated canine myocytes showed of TdP after d-sotalol and almokalant in the same dog. Three that under baseline conditions, the APD of the left endocardial simultaneously recorded surface electrocardiographic leads and two endocardial MAPs are presented. Paper speed is 10 mm/s. This figure (and epicardial) cells is longer than that of the right ventricle, shows the effect of programmed electrical stimulation after d-sotalol thereby creating ⌬APD. This difference has been described (Panel 1) and after almokalant (Panel 2) in the same dog as in Figure (21) as even more pronounced for the M cell, located within 1. After d-sotalol with a ⌬APD of 90 ms, programmed electrical the myocardium. stimulation results in TdP. In contrast, almokalant induction of TdP is Like the Purkinje fiber, the M cell has greater sensitivity for observed after programmed electrical stimulation with a prepacing ⌬APD of 115 ms. class III agents in its ability to prolong the APD and to develop EADs (21). The increase in ⌬APD in our experiments after class III drugs is primarily due to a larger increase in the APD present study, the main difference between dogs with pacing- of the left ventricle. The effect of almokalant (or d-sotalol) inducible and noninducible TdP is ⌬APD. Prevention or could be different for the various tissues involved in the heart, suppression, or both, of pacing-induced TdP resulted from the thus explaining the ⌬APD. disappearance of EADs (13,16) and diminishment of ⌬APD The M cells are not only present in dogs (21); recent studies (22) also report the existence of these M cells in normal human (17) by MgSO4. The finding that the presence of single EBs does not differ myocardium. It is not known to what extent the M cells between dogs with pacing inducible and noninducible TdP contribute to the APD, as recorded by the endocardially placed indicates that these beats occur during insufficient ⌬APD or MAP catheter. that the frequency changes induced by these beats do not lead Regional appearances of EADs can also explain the in- to a marked variation in APD (⌬APD). That pacing is capable crease in ⌬APD. d-Sotalol predominantly induced EADs in of inducing TdP is possibly related to the coupling interval of the left ventricle, whereas almokalant produced EADs in both the paced beats or the generation of multiple beats. Until now ventricles. Because administration of almokalant led to a more we have not been able to determine the dynamic changes pronounced increase in ⌬APD than administration of d- within the pacing train or the relevance of the pacing site for sotalol, the relative contribution of the amplitude of the EADs the induction of TdP. might also be important. Possible causes of interventricular dispersion. In dogs Almokalant and d-sotalol: proarrhythmic and electrophysi- with chronic complete AV block, ⌬APD is present under ologic effects. The proarrhythmic potential of class III agents baseline conditions and shows bradycardia dependence: De- is well known (1–9). Prolongation of repolarization is the creasing heart rate will increase ⌬APD (17). By their effect on mechanism of the antiarrhythmic effect of class III drugs, but heart rate, class III agents will thus increase the interventric- prolongation is also associated with the risk of development of ular differences in APD. TdP. Measurement of the absolute QT interval does not JACC Vol. 30, No. 6 VERDUYN ET AL. 1583 November 15, 1997:1575–84 RELEVANT FACTORS FOR THE INITIATION OF TdP predict proarrhythmic potential in clinical conditions (5), as nism of TdP (21,38). In addition to their relevance for the shown in our animal model (23). Currently, QT dispersion (as occurrence of EADs, differences in APD have been shown (39) a variable of nonhomogenous repolarization) is frequently to create the substrate for a reentrant tachycardia. Whether mentioned (24) as a possible tool for predicting proarrhythmic reentry plays a role in the initiation and continuation of TdP is risk as well as antiarrhythmic efficacy. The QT dispersion still unclear. It is conceivable that reentry succeeds or alter- measured in multiple ECG leads correlates with the dispersion nates with triggered activity (38,40), or both, and is certainly in regional repolarization measured by epicardial MAPs (25). the case when TdP progresses to ventricular fibrillation. Because QT dispersion shows large variation due to different ⌬APD is not likely to play a role in reentrant arrhythmias methodologies and different patient characteristics, the ques- because spatial dispersion is located so far apart. Still the tion of its use as a general arrhythmic marker has yet to be polymorphic appearance of TdP can be explained by ⌬APD. resolved (26). The appearance of EBs from a single foci can lead to TdP It is thus important to screen new class III agents for their when this “monomorphic tachycardia” encounters continu- proarrhythmic potential in an animal model. Most animal TdP ously shifting areas of repolarization. The situation becomes models have used nonclinically relevant drugs such as cesium more complicated when triggered EBs arise from multiple sites (10,13,27) or Bay K8644 (28). The only exceptions are an in the ventricles. awake AV block dog model with hypokalemia and beta- Conclusions. Spontaneous TdP is associated with prolon- blockade followed by antiarrhythmic drugs (29) and a rabbit gation of repolarization, the occurrence of EADs, the appear- model (30). In the awake dog it is not yet possible to measure ance of (multiple) EBs and ⌬APD. The first three factors all MAPs and thereby correlate proarrhythmic findings with contribute to a further increase in ⌬APD. These findings also EADs and APD. In the rabbit, TdP develops at relatively fast show that the present canine model of TdP can be used to heart rates after administration of class III agents and concom- screen new class III agents for their proarrhythmic potential. itant alpha-adrenergic stimulation. To our knowledge, comparison between different agents in the same animal concomitant with MAP recordings have not been described until the References present study (29–31). Because the response of the dog is maintained over weeks, 1. Colatsky TJ, Follmer CH, Starmer CF. Channel specificity in antiarrhythmic drug action: mechanism of potassium channel block and its role in suppress- serial comparison for screening of proarrhythmic effects of ing and aggravating cardiac arrhythmias. Circulation 1990;82:2235–42. (antiarrhythmic) drugs is feasible. Although this information 2. Hondeghem LM, Snyders DJ. Class III antiarrhythmic agents have a lot of cannot be translated into the exact incidence of TdP in the potential but a long way to go: reduced effectiveness and dangers of reverse use dependence. Circulation 1990;81:687–90. patient population, it allows an estimation of the relative risk 3. Singh BN, Wellens HJJ, Hiraoka M, editors. Electropharmocological Con- of drugs and their risk/benefit ratio. trol of Cardiac Arrhythmias: To Delay Conduction or to Prolong Refracto- Almokalant and d-sotalol both prolong the APD by block- riness. Armonk (NY): Futura, 1994. ade of Kϩ currents (32–34). Although almokalant is a selective 4. McNeil DJ, Davies RO, Deitchman D. Clinical safety profile in the treatment of arrhythmias. Am J Cardiol 1993;72:44A–50A. blocker of the rapid component of the delayed rectifier current 5. Nguyen TP, Scheinman MM, Seger J. Polymorphous ventricular tachycardia: (IKr) (32,34), the published data are not consistent for d- clinical characterization, therapy and QT interval. Circulation 1986;74: sotalol, which might also affect other Kϩ channels, such as the 340–9. inward rectifier current (I ), the slow component of the 6. Hohnloser SH, Woosley RL. Sotalol. N Engl J Med 1994;331:31–8. K1 7. Hohnloser SH, Klingenleben T, Singh BN. Amiodarone-associated proar- delayed rectifier current (IKs) and transient outward current rhythmic effects: a review with special reference to torsade de pointes (Ito) (34–36). In addition, d-sotalol has some beta-blocking tachycardia. Ann Intern Med 1994;121:529–35. effects (Ͻ5% of dl-sotalol [6]), whereas almokalant is devoid of 8. Wiesfeld ACP, Crijns HJGM, Bergstrand RH, Almgren O, Hillege HL, Lie KI. Torsade de pointes with almokalant, a new class III antiarrhythmic drug. beta-blocking activity (33). Am Heart J 1993;126:1008–11. Limitations and implications. In the present study we 9. Jackman WM, Friday KJ, Anderson JL, Aliot EM, Clark M, Lazzara R. The randomly placed two MAP catheters, one in the left and the long QT syndromes: a critical review, new clinical observations and a other in the right ventricle, to provide local information about unifying hypothesis. Prog Cardiovasc Dis 1988;31:115–72. 10. Levine JH, Spear JF, Guarnieri T, et al. Cesium chloride-induced long QT the differences in repolarization between the two ventricles syndrome: demonstration of afterdepolarizations and triggered activity in (⌬APD). Possible transmual or intraventricular differences in vivo. Circulation 1985;72:1092–103. APD were not assessed in this study. Placement of multiple 11. Carlsson L, Abrahamsson C, Andersson B, Du¨ker G, Schiller-Linhardt G. Proarrhythmic effects of the class III agent almokalant: importance of catheters with d-sotalol revealed that the intraventricular infusion rate, QT dispersion, and early afterdepolarizations. Cardiovasc Res difference amounted only to 20% to 30% of the ⌬APD (17). 1993;27:2186–93. We do not know the type and number of cells contributing to 12. Bonatti V, Rolli A, Botti G. Monophasic action potential studies in human subjects with prolonged ventricular repolarization and long QT syndromes. the MAP. Similarly, whether the presence of EADs on the Eur Heart J 1985;6 Suppl:131–43. MAP is real or based on differences in transmural repolariza- 13. Bailie DS, Inoue H, Kaseda S, Ben David J, Zipes DP. Magnesium tion is not known (37). However, interventions specifically suppression of early afterdepolarizations and ventricular tachyarrhythmias directed to suppression of EAD formation, such as pacing and induced by cesium in dogs. Circulation 1988;77:1395–402. 14. Surawicz B. Electrophysiological substrate of torsade de pointes: dispersion drugs, leads to the disappearance of these EADs (16,17). of repolarization or early afterdepolarizations. J Am Coll Cardiol 1989;14: Transmural differences have been implicated in the mecha- 172–82. 1584 VERDUYN ET AL. JACC Vol. 30, No. 6 RELEVANT FACTORS FOR THE INITIATION OF TdP November 15, 1997:1575–84

15. Habbab MA, El-Sherif N. Drug induced torsade de pointes: role of early hypertrophy predisposes to ventricular tachycardia induction by phase 2 afterdepolarizations and dispersion of repolarization. Am J Med 1990;89: early afterdepolarizations after administration of Bay K 8644. J Am Coll 241–6. Cardiol 1992;20:1576–84. 16. Vos MA, Verduyn SC, Gorgels APM, Lipscei GC, Wellens HJJ. Reproduc- 29. Weissenburger J, Davy JM, Chezalviel F, et al. Arrhythmogenic activities of ible induction of early afterdepolarizations and torsade de pointes arrhyth- anti arrhythmic drugs in conscious hypokalemic dogs with atrioventricular mias by d-sotalol and pacing in the dog with chronic atrioventricular block. block: comparison between quinidine, lidocaine, flecainide, propranolol and Circulation 1995;91:864–72. sotalol. J Pharmacol Exp Ther 1991;259:871–83. 17. Verduyn SC, Vos MA, van der Zande J, van der Hulst FF, Wellens HJ. Role 30. Carlsson L, Almgren O, Duker G. QTU prolongation and torsades de of interventricular dispersion of repolarization in acquired torsade de pointes induced by putative class III antiarrhythmic agents in the rabbit: pointes arrhythmias: reversal by magnesium. Cardiovasc Res 1997;34:453– etiology and interventions. J Cardiovasc Pharmacol 1990;16:276–85. 63. 31. Buchanan LV, Kabell G, Brunden MN, Gibson JK. Comparative assessment 18. Vos MA, Gorgels APM, Lipscei GC, de Groot SHM, Leunissen JDM, of ibutilide, d-sotalol, clofilium, E4031, and UK-68,798 in a rabbit model of Wellens HJJ. Mechanism specific anti-arrhythmic effects of the potassium proarrhythmia. J Cardiovasc Pharmacol 1993;22:540–9. channel activator levcromakalim against repolarization dependent tachycar- 32. Sanguinetti CM, Jurkiewicz NK. Two components of cardiac delayed dias. J Cardiovasc Electrophysiol 1994;5:731–42. rectifier Kϩ current: differential sensitivity to block by class III antiarrhyth- 19. Duker G, Almgren O, Carlsson L. Electrophysiologic and hemodynamic mic agents. J Gen Physiol 1990;96:195–215. effects of H 234/09 (almokalant), quinidine, and (ϩ)sotalol in the anesthe- 33. Almgren O. Electropharmacological profile and antiarrhythmic potential of tized dog. J Cardiovasc Pharmacol 1992;20:458–65. almokalant (H234/09). In: Singh BN, Wellens HJJ, Hiraoka M, editors. 20. Damiano BP, Rosen MR. Effects of pacing on triggered activity induced by Electropharmocological Control of Cardiac Arrhythmias: To Delay Conduc- early afterdepolarizations. Circulation 1984;69:1013–25. tion or to Prolong Refractoriness. Armonk (NY), Futura, 1994:569–75. 21. Antzelevitch C, Sicouri S. Clinical relevance of cardiac arrhythmias gener- 34. Wettwer E, Grundke M, Ravens U. Differential effects of the new class III ated by afterdepolarizations: role of M cells in the generation of U waves, triggered activity and torsade de pointes. J Am Coll Cardiol 1994;23:259–77. agents almokalant, E-4031 and d-sotalol and of quinidine, on delayed rectifier currents in guinea pig ventricular myocytes. Cardiovasc Res 1992; 22. Drouin E, Charpentier F, Gauthier C, Laurent K, Le Marec H. Electro- 26:1145–52. physiologic characteristics of cells spanning the left ventricular wall of human heart evidence for the presence of M cells. J Am Coll Cardiol 1995;26:185– 35. Carmeliet E. Electrophysiologic and voltage clamp analysis of the effects of 92. sotalol on isolated cardiac muscle and Purkinje fibers. J Pharmacol Exp Ther 23. Verduyn SC, Vos MA, Gorgels APM, Leunissen JDM, van der Zande J, 1985;232:817–25. Wellens HJJ. QT time: predictor of the occurrence of acquired torsade de 36. Varro A, Nanasi PP, Lathrop DA. Effect of sotalol on transmembrane ionic pointes arrhythmias [abstract]. PACE 1994;17:826. currents responsible for repolarization in cardiac ventricular myocytes from 24. Hii JTY, Wyse DG, Gillis AM, Duff HJ, Solylo MA, Mitchell LB. Precordial rabbit and guinea pig. Life Sci 1991;49:PL7–9. QT time interval dispersion as a marker of torsade de pointes: disparate 37. Nesterenko VV, Antzelevitch C. Factors responsible for “apparent” early effects of class 1A antiarrhythmic drugs and amiodarone. Circulation 1992; afterdepolarizations (EADs) in monophasic actionpotential (MAP) record- 86:1376–82. ings: a model study [abstract]. PACE 1995;18:830. 25. Zabel M, Portnoy S, Franz MR. Electrocardiographic indexes of ventricular 38. El-Sherif N, Restivo M, Caref EB, Piracha M. Electrophysiologic mecha- repolarization: an isolated heart validation study. J Am Coll Cardiol nism(s) of torsade de pointes ventricular tachyrrhythmias (TdP): tridimen- 1995;25:746–52. sional mapping of ventricular activation [abstract]. Circ Res 1995;92 Suppl 26. Statters DJ, Malik M, Ward DE, Camm AJ. QT dispersion: problems of I:I–641. methodology and clinical significance. J Cardiovasc Electrophysiol 1994;5: 39. Kuo CS, Munakata K, Reddy P, Surawicz B. Characteristics and possible 672–85. mechanism of ventricular arrhythmia dependent on the dispersion of action 27. Nayebpour M, S Nattel. Pharmacologic response of cesium induced ventric- potentials. Circulation 1983;67:1356–67. ular tachyarrhythmias in anesthetized dogs. J Cardiovasc Pharmacol 1990; 40. Abildskov JA, Lux RL. The mechanism of simulated torsade de pointes in a 15:552–61. computer model of propagated excitation. J Cardiovasc Electrophysiol 28. Ben-David J, Zipes DP, Ayers GM, Pride HP. Canine left ventricular 1991;2:224–37.