Recently the Survival With ORal D-sotalol (SWORD) study showed increased mortality in patients treated with the pure class III agent D-sotalol (1), leading to the discontinuation of its clinical investigation. In the SWORD study, D-sotalol was tested against placebo in patients with ventricular ectopy after myocardial infarction. These results helped to weaken clinical interest previously placed on class III drugs in treating ventricular arrhythmias after myocardial infarction. Indeed, control of cardiac arrhythmias by selective lengthening of action potential duration (APD) has been a growing concept in cardiac electropharmacology (2). Among class III agents, DL- and D-sotalol have shown antiarrhythmic effects in experimental and clinical investigations. In vivo, studies showed preventive effects of sotalol on atrial and ventricular arrhythmias in anesthetized pigs (3) and dogs (4,5). Clinical reports also showed DL- and D-sotalol antiarrhythmic efficacy against ventricular arrhythmias (6-8) and supraventricular tachycardias (9). On the other hand, sotalol was demonstrated to lengthen APD and refractoriness in both Purkinje fibers (10-12) and ventricular muscle (12,13) in human ventricular myocardium in vitro (14) and in clinical trials (8,15), with no significant effect on the maximal rate of increase of the AP upstroke or intracardiac conduction times (16). Concomitantly, proarrhythmic properties were observed with the pure class III agent D-sotalol in an in vitro model of global ischemia on isolated guinea-pig ventricular free walls, whereas only DL-sotalol had antiarrhythmic action (17). Clinical studies showed risks caused by the recognized proarrhythmic effects of class III drugs (18,19) such as emergence of new inducible ventricular tachycardias (20) or early after depolarization-dependent triggered activity, leading to torsades de pointes development (21,22). However, the mechanisms responsible for anti- and proarrhythmic actions of DL- and D-sotalol are still elusive (23,24), and it has not been ascertained whether β-blocking properties of DL-sotalol may contribute clinically.
During myocardial ischemia, the "border zone," described as the intermediate zone separating normal and hypoxic/ischemic regions (25), has been suggested as a major site of arrhythmias (26-27), although the existence of such a border zone remains controversial (28). More recently, an in vitro study demonstrated the importance of APD dispersion between the normal and the adjacent ischemic myocardial region in the occurrence of spontaneous arrhythmias (29). The in vitro model of ischemia used for the latter study consisted of an experimental bath divided into two compartments by a high-quality rubber membrane (30-32), allowing the simultaneous study of myocardial strips from guinea-pig right ventricle in both normal and ischemia-simulated conditions, thus allowing an in vitro investigation of the border zone. Ischemia is known to modify the class III action of DL-sotalol and of its dextrorotatory isomer (17), but electrophysiologic effects of these agents on ventricular myocardium around the border zone were not previously investigated.
With the aim of clarifying mechanisms possibly responsible for results of the SWORD study, we tested both DL- and D-sotalol, at 5 and 10 μM, in the previously mentioned in vitro model of border zone arrhythmias (30-32). Both electrophysiologic and anti- and proarrhythmic effects of DL-sotalol and D-sotalol were concomitantly measured on AP parameters simultaneously under normoxic and ischemic conditions to evaluate (a) the ischemia-drug interactions, and (b) the incidence of spontaneous arrhythmias around the border zone on ischemia and reperfusion.
Preparation and mounting
Guinea pigs of either sex weighing 300-400 g were killed and exsanguinated. The hearts were quickly removed and placed in Tyrode's solution, gassed with a 95% oxygen/5% carbon dioxide mixture, at room temperature. A standard strip of right ventricular myocardium (12 mm long, 4 mm wide, 1.5 mm thick) was dissected from the free wall and placed in an experimental bath (volume of 5 ml; 30-32) separated into two compartments by a thin latex membrane. A small incision was made at the bottom of the latex partition, and the strip was carefully passed through. The preparation was then pinned, the endocardial surface upward, on the Sylgard base of the bath. This double compartment allowed the two segments of the ventricular strip to be independently superfused at a rate of 2 ml/min. Continuity of the partition was tested at the end of each experiment by means of dye injection (methylene blue) into one of the compartments. Temperature at the level of the double chamber, including that of incoming fluids, was controlled by a circulating thermostat-controlled bath.
Studies were performed in Tyrode's solution oxygenated with a mixture of 95% oxygen and 5% carbon dioxide. The composition of the Tyrode's solution was (in mM): Na+, 135; K+, 4; Ca2+, 1.8, Mg2+, 1.0; H2PO4−, 1.8; HCO3−, 25; Cl−, 117.8; and glucose, 5.5. The pH was 7.35 ± 0.05 (fitted with diluted HCl), and the temperature was maintained at 36.5 ± 0.5°C. Frequent analyses (BG Electrolytes, Instrumentation Laboratory SpA, Milano, Italy) allowed maintenance of pO2 and pCO2 at 510 ± 20 and 34 ± 2 mm Hg, respectively. A modified Tyrode's solution also was used, which differed from the normal one by an elevated K+ concentration (12 instead of 4 mM), decreased HCO3− concentration (9 mM), leading to a decrease in pH (6.90 ± 0.05), decrease in pO2 (510 became 80 mm Hg) by means of gassing the preparation with 95% N2 and 5% CO2, and complete withdrawal of glucose. As previously described (29-31), our modifications of Tyrode's solution, combining hypoxia, hyperkalemia, acidosis, and lack of substrates, are quite similar to those reported by Morena et al. (33) as able to reproduce in vitro the electrophysiologic abnormalities induced in vivo by ischemia. DL- And D-sotalol (Bristol-Myers Squibb, Paris, France) were dissolved in double distilled water.
Stimulation and recordings
The preparations were stimulated via bipolar Teflon-coated steel-wire electrodes positioned near the two extremities of the ventricular strip in each of the two bath chambers. A simple commutator allowed application of stimulations either through one or the other extracellular stimulation electrode. Stimuli were rectangular pulses of depolarizing voltage, 2 ms in duration, and twice the diastolic threshold intensity (∼2-2.5 V) delivered by a programmable stimulator (Biologic SMP-310, Claix, France) at a basic cycle length of 2.22, 1.25, 1, or 0.56 Hz. Preparations needing stronger pulses (>5 V) to elicit APs in the two compartments were eliminated because of a conduction block, possibly produced at the level of the latex separating membrane. During the experiment, stimulation was stopped whenever spontaneous repetitive responses appeared.
In the two compartments, transmembrane potentials were recorded by impaling myocytes with conventional intracellular glass microelectrodes inserted in each side of the preparation near the latex membrane. The micropipettes were pulled from filamented capillary tubes on a single-barreled microelectrode puller, filled with 3 M KCl, and coupled to silver-silver chloride microelectrode holders leading to the double-input stage of a high-impedance capacitance-neutralizing amplifier. The two ball-shaped reference silver-silver chloride electrodes were positioned in the superfusate of each chamber, close to the preparation but at a distance from the recording glass microelectrodes. APs were monitored on a digital memory oscilloscope. The AP data were also digitized and stored on hard disk at a sampling frequency of 8 kHz (Datapac 13.2; University of Normandy, Caen, France), on an IPC computer (386; 33 MHz).
The following AP characteristics were recorded, measured, and stored: resting membrane potential (RMP), AP amplitude (APA), APD at 90% of repolarization (APD90), and maximal upstroke velocity (Vmax) of AP. Measurements were obtained simultaneously from both anterior (superfused with the modified Tyrode's solution: "ischemic" chamber) and posterior chamber (superfused with the normal Tyrode's solution: "normal" chamber). The Datapac system allowed measurements of myocardial conduction times (MCTs) between the two intracellular microelectrodes, by measuring the interval separating the Vmax peaks of the two APs.
During ischemia and on reperfusion, two types of electrical disturbances were recorded:
- Myocardial conduction blocks. These blocks were (a) unidirectional, in the direction from ischemic zone (IZ) toward normal zone (NZ) or conversely, and (b) bidirectional, between these two regions or when the IZ became inexcitable, whatever the compartment receiving the stimulation. Blocks were coded present = 1, absent = 0; and
- Spontaneous repetitive responses such as singles, couplets, or triplets of APs, salvos (four to nine spontaneous APs), and sustained arrhythmias (salvos of >10 spontaneous APs at a frequency ≥3 Hz; coded present = 1, absent = 0).
Effects of DL- and D-sotalol on AP parameters and MCTs. During a 120-min equilibration period, the two compartments were superfused with normal Tyrode's solution, and the right ventricular muscle was stimulated at a frequency of 1 Hz. Then one chamber (compartment called IZ) was superfused for 30 min with modified Tyrode's solution (referred to as ischemia phase) and then superfused again with normal Tyrode's solution for 30 min (referred to as reperfusion phase), while the second compartment was still under normoxic conditions [normal zone (NZ)]. During both ischemia and reperfusion phases, DL- or D-sotalol (5 and 10 μM) were added to the superfusion fluids, allowing the simultaneous recordings of their electrophysiologic effects in normoxic conditions and during ischemia followed by reperfusion. The results obtained with these four groups [DL-sotalol, 5 μM (n = 6); DL-sotalol, 10 μM (n = 7); D-sotalol, 5 μM (n = 6); and D-sotalol, 10 μM (n = 7)] were compared with those recorded during the 60 min of ischemia/reperfusion in the absence of drugs (control group, n = 8). Two APs were recorded simultaneously in each compartment before the ischemic period, serving also as control. Then recordings were performed every 5 min during both ischemia and reperfusion phases. The stimulation side was changed during signal acquisition, thus allowing MCT measurements in the two conduction directions (NZ to IZ and conversely).
Effects of DL- and D-sotalol on ischemia- and reperfusion-induced arrhythmias. Simultaneous with AP parameters, myocardial conduction blocks and spontaneous arrhythmias were recorded and stored during both ischemia and reperfusion phases, on the 35 preparations related previously. In addition, another 42 reference preparations investigated in our laboratory (29), with varying degrees of ischemia-simulated duration and stimulation rate, were summed. The aim was to include a larger number of control experiments (eight plus 42) to analyze and take into account random variability of several parameters, later entered as covariates in a robust multivariate model of event prediction. Ischemia-phase durations (either 15 or 30 min) and stimulation rate (2.22, 1.25, 1, or 0.56 Hz) in these latter preparations were chosen in a randomized manner (29). Statistical analysis of spontaneous arrhythmia incidence during ischemia and reperfusion was then performed in 77 experiments overall.
Student's t test was used to assess the significance of changes observed in AP parameters. To define the correlation between the measured parameters and the incidence of arrhythmias, aimed at taking into account random variability of measured parameters, multiple linear regression (34) was used to study the relation between the incidence of a given event (here the occurrence of either ischemia or ischemia-reperfusion spontaneous repetitive responses, coded 0 when absent or 1 when present) and a set of explanatory variables (covariates).
Suppose that t is survival time (the time of event occurrence) and X1, X2,. . . .Xp are the dependent or explanatory variables. Let Y = Ln(t) be the dependent variable where Ln = natural logarithm. Then the model assumes a linear relation between Ln(t) and the Xs. The model equation is Equation (1) where e is an error term. This model is known as a log-linear regression model because the log (as natural logarithm) of survival time is a linear function of the Xs. Because Ln(t) is usually not normally distributed, the method of maximal likelihood is used to obtain estimates of βis and their standard errors. When the Weibull's distribution is assumed, the log-linear is sometimes known as the accelerated-life or accelerated-failuretime model. The Weibull's distribution shows that the density function of event occurrence very rapidly decreases with time [as is the case with the occurrence of both ischemia and reperfusion spontaneous repetitive responses in our model (29)], this decrease being most rapid during the initial part of the follow-up (34).
By using the Weibull's model, the probability of an event might be easily calculated by the formula: Equation (2) where p is the probability or risk of a given event, t = risk prediction interval of interest, m = the Weibull's equation evaluated at the appropriate value for each risk factor, and s = the extreme value scale parameter.
To obtain the Weibull's equation m, the following applies Equation (3) where α = model constant, βis are the coefficients, and Xs are arbitrary (generally mean) values of explanatory variables.
Although for Eq. 3, all βis contribute, significant (p < 0.05) contribution of explanatory variables might be concluded by β coefficient/SE > | 1.96 |. Finally, the algebraic sign of the β coefficient is important to conclude for risk (when negative) or protective (when positive) capability of a given factor X.
Setting adequate specifications (32), computer software may actually be used to obtain curves where the presence of given contributory or explanatory factors or both might be illustrated. In the case described here, in comparison with controls, we have illustrated the role of D- and DL-sotalol, at both concentrations, on the occurrence of spontaneous arrhythmic events during our experiments. For this purpose, patterns were constructed in which average values were considered for all 10 Xs included in (3), the only difference resulting in variables defining presence or absence of each drug (at two concentrations). The method was such that experiments were largely representative of the electrophysiologic responses of guinea-pig ventricular myocardium because all drug-induced changes were tested here against confident average values as assessed in 50 controls. On the other hand, investigated patterns (on ischemia and ischemia-reperfusion spontaneous repetitive responses) evaluated the respective (relative to controls) contribution of treatment types on the event occurrence, all remaining covariates being constant in the model, which is a commonly used method to investigate the role of risk factors in epidemiologic studies (see 29 for discussion). Patterns obtained in this study illustrated the estimated survival of cardiomyocytes in our in vitro model, namely the arrhythmic event-free rate of myocardial preparations.
Effects of DL- and D-sotalol on AP parameters and MCTs
After 30 min of ischemia, electrophysiologic parameters were compared univariately among controls (no drug, n = 8); DL-sotalol, 5 and 10 μM (n = 6 and n = 7, respectively); and D-sotalol, 5 and 10 μM (n = 6 and n = 7, respectively).
AP duration. APD90 was measured simultaneously in the two compartments during ischemia and reperfusion phases. Figure 1 illustrates APD90 (mean ± SEM) measured in normal and ischemic zones (upper and lower graphs, respectively), in initial conditions (before ischemia phase, corresponding to time 0), and at 10 and 30 min of simulated ischemia, and finally at 30 min of reperfusion, for all groups.
In the NZ (Fig. 1A), APD90 did not vary significantly with normal Tyrode's solution superfusion (from 151 ± 6 to 157 ± 10 ms after 60 min; p > 0.05), whereas a significant increase of APD90 was observed with DL- and D-sotalol (p < 0.05 after 30 or 60 min of drug perfusion). The AP lengthening induced by D-sotalol after 60 min was significantly (p < 0.05) more marked with 10 μM compared with 5 μM (APD90 increased from 150 ± 7 to 176 ± 7 ms, +18 ± 3%, at 5 μM; p < 0.005; and from 156 ± 9 to 208 ± 8 ms, +38 ± 8%, at 10 μM; p < 0.005), whereas the concentration dependence of class III effects was not significant after 10 or 30 min, or between both 5 and 10 μM concentrations with DL-sotalol. For the same concentration of 10 μM, after 60 min, D-sotalol produced class III effects significantly more pronounced than did the racemic compound (from 147 ± 5 to 183 ± 11 ms, +24 ± 5%, with DL-sotalol vs. an increase from 156 ± 9 to 208 ± 8 ms, +38 ± 8%, with D-sotalol; p < 0.05), whereas AP lengthening obtained with 5 μM of drug did not differ significantly between both agents.
In the IZ (Fig. 1B), simulated ischemia significantly decreased APD90 (from 151 ± 4 to 60 ± 7 ms after 30 min of ischemia; p < 0.005); then reperfusion allowed return to initial values of APD90 (155 ± 4 ms after 30 min of reperfusion). In presence of DL- and D-sotalol, at both 5 and 10 μM, shortening of APD90 was similar (p > 0.05) to that observed in control group (at 30 min of ischemia phase, APD90 was 67 ± 7 and 66 ± 8 ms at 5 μM of both DL- and D-sotalol, respectively, whereas APD90 was 58 ± 7 and 74 ± 4 ms at 10 μM DL- and D-sotalol, respectively, vs. 60 ± 7 ms in control group). During reperfusion, APD90 increased to reach values measured in the NZ. At 30 min of reperfusion, APD90 was significantly lengthened compared with initial values, in presence of both agents, with a concentration-dependent class III effect significant (p < 0.05) only with D-sotalol (APD90 increased from 151 ± 7 to 169 ± 7 ms, +12 ± 2%, at 5 μM; p < 0.005; and from 154 ± 8 to 195 ± 10 ms, +27 ± 2%, at 10 μM; p < 0.005).
Maximal upstroke velocity of AP (Vmax) and MCTs. During ischemia and reperfusion phases, Vmax and MCT were recorded simultaneously in the NZ and IZ. Variations of Vmax during both ischemia and reperfusion phases are illustrated in Fig. 2 for all groups.
In the NZ (Fig. 2A), Vmax did not vary significantly with normal Tyrode's solution superfusion (from 242 ± 30 to 221 ± 32 V/s after 60 min; p > 0.05). Simulated ischemia induced a significant Vmax decrease (from 248 ± 17 to 52 ± 20 V/s at 30 min of ischemia period; p < 0.01), which was completely reversed by reperfusion (248 ± 27 V/s at 30 min; Fig. 2B). Both DL- and D-sotalol, at both 5 and 10 μM, had significant effect on neither Vmax (NZ, Fig. 1A) nor on Vmax changes induced by ischemia and reperfusion (IZ, Fig. 1B) compared with control experiments.
MCT (not illustrated) was significantly increased during the ischemia phase in both conduction directions (NZ to IZ and IZ to NZ increased 201 ± 19% and 134 ± 19%, respectively, after 30 min of ischemia). Class III agents D- and DL-sotalol showed no statistically significant action on MCT.
Other AP parameters. Simulated ischemia induced a significant AP-amplitude decrease and resting membrane depolarization (from 120 ± 2 to 78 ± 6 mV; p < 0.01; and from −88 ± 1 to −60 ± 1 mV; p < 0.005; respectively, at 30 min of ischemia phase). APA and RMP returned to initial values after 10 min of reperfusion. No significant effect of DL- and D-sotalol was observed on both APA and RMP (not illustrated).
Representative AP recordings, performed simultaneously in IZ and NZ, in the absence (control) or in the presence of D-sotalol (10 μM), are illustrated in Fig. 3. The APs were recorded in initial conditions (normal Tyrode's solution in the two compartments) and at the end of both ischemia and reperfusion phases, in the same cell in each myocardial zone. The APD90 values, measured in IZ and NZ, are given under tracings of Fig. 3, illustrating the class III effect of D-sotalol, 10 μM, observed only in NZ during ischemia phase. Other AP parameters are not shown because of the absence of a significant effect of D-sotalol, 10 μM. Similar changes were observed with D-sotalol, 5 μM, and DL-sotalol, 5 and 10 μM (not illustrated).
Effects of DL- and D-sotalol on ischemia- and reperfusion-induced arrhythmias and their multivariate prediction
For these aims of the study, 77 experiments were considered (27 treated preparations with DL- or D-sotalol and 50 control experiments).
The baseline descriptive statistics of variables considered in all 77 experiments included in this investigation are summarized in Table 1, which provides the means to assess the variable distribution and the 95% confidence intervals. All electrophysiologic variables measured in NZ were normally distributed, except for Vmax, which appears rightward skewed. Conversely, all variables measured in IZ were not normally distributed (leftward skewed for APA and RMP and rightward skewed for APD90 and Vmax). The same holds true among variables used to randomize the 42 reference experiments summed up with this part of the study, for stimulation rate (right-ward skewed), whereas ischemic time was normally distributed. In these latter 42 reference experiments, identical distributions were observed (29). Because distributions observed in the 42 reference experiments and in the other 35 (result not shown) were superposable (Table 2), AP parameters as observed in the former group may be mixed with those seen in the latter D- and DL-sotalol groups and the remaining eight controls, and the differential results later obtained after drugs may be compared with 50 controls.
The distribution of the parameters used in the Weibull's analysis among the five groups is given in Table 2. All treated preparations underwent an ischemia period of 1,800 s and were stimulated at 1 Hz, whereas for the control group, ischemic time and stimulation rate were slightly different because of the randomization of these variables in 42 of the 50 experiments. However, for these latter variables, there was no significant difference between control and treated groups. There was no significant difference in electrophysiologic variables, measured either in NZ or in IZ, between control and treated groups (p > 0.05), except for APA in NZ, which was lower in the 10 μM D-sotalol group (p < 0.05) and Vmax in IZ, which was higher in the 10 μM DL-sotalol group (p < 0.01).
The incidence of spontaneous repetitive responses and myocardial conduction blocks during either ischemia or reperfusion phases among groups is given in Table 3. During ischemia phase, distributions of spontaneous repetitive responses and myocardial conduction blocks between groups were different. Blocks occurred in 28% of control preparations, whereas they were enhanced by DL-sotalol, 10 and 5 μM, and D-sotalol, 10 and 5 μM (71, 67, 50, and 50% of treated preparations, respectively). The occurrence of total spontaneous repetitive responses was enhanced by these agents in 86, 83, 62, and 83% of treated preparations versus 32% of the control group, respectively. Among total ischemia arrhythmic events, severity of rhythm disturbances was enhanced by DL-sotalol, 10 and 5 μM, and D-sotalol, 10 and 5 μM: 83, 100, 80, and 60%, respectively, of total spontaneous arrhythmias were of sustained type (salvos >10 AP) versus 19% of such arrhythmic events in the control group. During reperfusion phase, no myocardial conduction block occurred, and all ischemia-induced blocks were abolished in the first 5 min of reperfusion. Spontaneous repetitive responses occurred in 88% of control experiments. Similar occurrence was obtained in the presence of DL-sotalol, 5 μM, and D-sotalol, 10 and 5 μM (87, 83, and 100%, respectively), whereas only 57% of preparations treated with DL-sotalol, 10 μM, had spontaneous repetitive responses. However, all treated preparations with spontaneous arrhythmic events showed a more marked degree of arrhythmic disturbances than did the control preparations: 75, 80, 100, and 83% of reperfusion repetitive responses observed in presence of DL-sotalol, 10 and 5 μM, and D-sotalol, 10 and 5 μM, respectively, were sustained activities versus 36% of total arrhythmic events observed in the control group.
Representative electrical disturbances are shown in Fig. 4, an example of myocardial conduction block (upper traces) and two examples of spontaneous repetitive responses such as salvos (middle traces) and sustained activities (lower traces). The conduction block illustrated here (top of Fig. 4) is unidirectional in the IZ to NZ direction: stimulus applied in NZ induced responses in NZ and then in IZ, whereas stimulus applied in IZ induced response only in IZ; thus the propagation of the signal did not reach the NZ. As described in Methods, stimulation was stopped during the occurrence of spontaneous sustained responses; the example shown in Fig. 4 (bottom) illustrates persistence of these spontaneous sustained activities after stimulation was stopped.
The results of the Weibull's model in predicting either ischemia or reperfusion spontaneous repetitive responses in the overall study (N = 77) are reported in Table 4. Covariates were included to assess the predictive contribution of the variables used to randomize the 42 control experiments, along with covariates in both NZs and IZs and presence of DL- and D-sotalol at 10 and 5 μM. In the reperfusion total spontaneous repetitive responses, spontaneous repetitive responses and myocardial conduction blocks on ischemia were included. In prediction of ischemia spontaneous arrhythmic events (χ2 = 24.79; p = 0.0367), neither variables used for the randomization nor covariates measured in NZ and IZ were significantly related to the occurrence of arrhythmias, whereas the presence of DL-sotalol, 10 and 5 μM, and D-sotalol, 5 μM, were significant predictors (β = −2.26, t = −2.80; β = −1.85, t = −2.35; and β = −2.04, t = −2.51, respectively). Although the predictive contribution of the presence of D-sotalol, 10 μM, was not significant, the t-value sign remained negative (β = −1.40; t = −1.69; 0.1 > p > 0.05), suggesting proarrhythmic activity. Marked proarrhythmic properties of class III agents at both concentrations, during simulated ischemia, are illustrated in Fig. 5, which represents the estimated survival curves for ischemia total spontaneous repetitive responses, based on the accelerated-failure-time Weibull's model. The estimated survival (arrhythmic event-free rate) of control experiments (n = 50), for an ischemia time of 1,800 s, is 0.68, whereas in presence of DL-sotalol, 5 and 10 μM, and D-sotalol, 5 and 10 μM, the estimated survival of preparations decreased to 0.17, 0.08, 0.12, and 0.30, respectively.
In prediction of reperfusion total spontaneous repetitive responses (Table 4; χ2 = 46.74; p = 0.0001), the contribution of stimulation rate, APA in NZ, and Vmax in IZ were statistically significant (β = −0.0009, t = −2.05; β = −0.08, t = −3.32; and β = −0.01, t = −3.39, respectively), as was the presence of ischemia spontaneous arrhythmias, which enhanced significantly the risk of reperfusion total spontaneous arrhythmic events (β = −1.12, t = −3.56). Conversely, a longer ischemic time and the presence of DL-sotalol, 10 μM, showed significant protective properties (β = 0.0021, t = 4.84; and β = 2.57, t = 3.52, respectively). Pro- or antiarrhythmic activities of DL- and D-sotalol at 5 and 10 μM during reperfusion are shown in Fig. 6, which gives, for each experimental group, the estimated survival of preparations (arrhythmic event-free rate) for ischemia-reperfusion total spontaneous repetitive responses. The accelerated-failure-time Weibull's model, based on ischemia-reperfusion time of 3,600 s, shows the marked antiarrhythmic contribution of DL-sotalol, 10 μM (estimated survival: 0.64 vs. 0.0 in the control group). Conversely, DL-sotalol, 5 μM, and D-sotalol, 5 and 10 μM, groups presented estimated survival curves with accelerated kinetics of decreasing phase, suggesting tendencies to proarrhythmic effects, according to negative t values given in Table 4 for the prediction of reperfusion total arrhythmias (t = −0.94, t = −1.48, and t = −0.45, respectively). For understanding of results, the antiarrhythmic effect of DL-sotalol, 10 μM, illustrated in Fig. 6, demands considering the accelerated-failure-time model, based here on an ischemia-reperfusion time of 3,600 ms, which also takes into account the occurrence of total spontaneous repetitive responses during the ischemia phase. Thus results obtained in Fig. 6 take into consideration pro- or antiarrhythmic effects of agents investigated on ischemia spontaneous total arrhythmic events, which are of importance regarding the high proarrhythmic action of DL-sotalol, 10 μM, during the ischemia phase. Finally, considering the contributory effect of a reduced Vmax and the protective role of DL-sotalol, 10 μM, on the occurrence of reperfusion spontaneous arrhythmias, it is useful to note in Table 2 a significantly more rapid Vmax in the DL-sotalol 10 μM group than in the control group (p < 0.01) in the IZ, Vmax being measured just before the occurrence of the first ischemia spontaneous repetitive response. However, there was no significant action of DL-sotalol, 10 μM, on Vmax in the IZ, as shown in Fig. 2B, suggesting therefore that (a) ischemia spontaneous arrhythmias occurred in the DL-sotalol 10 μM group at a moment when Vmax was not yet dramatically reduced by simulated ischemia, and (b) the antiarrhythmic effect of DL-sotalol, 10 μM, during the reperfusion phase cannot be related to an action of the latter class III agent on Vmax.
To determine whether the antiarrhythmic action of DL-sotalol, 10 μM, during reperfusion might be mediated by its β-blocking properties, we tested the pure β-blocking agent propranolol on border zone arrhythmias occurring in our in vitro model of simulated ischemia-reperfusion. Results obtained on seven preparations (not illustrated), revealed that, in the presence of propranolol, 10 μM, the incidence of spontaneous repetitive responses during the ischemia phase was not significantly changed (in 57% of treated preparations vs. 32% in controls, NS), whereas propranolol, 10 μM, exhibited antiarrhythmic efficacy on reperfusion-related spontaneous repetitive responses (in 14% of treated preparations vs. 88% in controls; p < 0.05, Exact test).
In brief, our results show that (a) DL- and D-sotalol, at 5 and 10 μM, lengthened significantly the APD under normoxic conditions, whereas these class III effects were lost in ischemic tissue. Both agents were unable to prevent the ischemia-induced APD shortening; (b) DL- and D-sotalol, at 5 and 10 μM, showed proarrhythmic effects during simulated ischemia, whereas DL-sotalol, 10 μM, prevented spontaneous responses during reperfusion, similar to what was observed with the pure β-blocking agent propranolol, 10 μM; (c) DL- and D-sotalol, at 5 and 10 μM, enhanced the severity of spontaneous arrhythmic events during the occurrence of both ischemia- and reperfusion-induced arrhythmias; and (d) except DL-sotalol 10 μM, no other group, including controls, was free of spontaneous repetitive responses during 3,600 s of ischemia-reperfusion in our model, which may indicate a significant antiarrhythmic capability of DL-sotalol at 10 μM compared with the lower concentration of 5 μM and with both 5 and 10 μM D-sotalol.
Control of cardiac arrhythmias by the selective lengthening of the APD has been a promising challenge in cardiac electrophysiology, regarding the antiarrhythmic properties of DL- and D-sotalol elucidated from several in vivo and clinical investigations (3-9). From the Student's t test analysis performed in our experiments, it appears that DL- and D-sotalol (5 and 10 μM) exhibited different actions under normoxic and ischemic conditions, by lengthening APD90 in the normal tissue without preventing ischemia-induced APD90 shortening in the adjacent myocardial region. On the other hand, the multivariate statistical analysis based on the Weibull's model illustrated proarrhythmic activities of DL- and D-sotalol (5 and 10 μM) during simulated ischemia and antiarrhythmic effects of DL-sotalol, 10 μM, on reperfusion-induced spontaneous repetitive responses. This latter analysis did not show statistical significance as concerns the relation between ischemia-induced spontaneous arrhythmias and measured AP parameters, whereas AP amplitude and Vmax predicted the occurrence of reperfusion-induced spontaneous arrhythmias. Finally, antiarrhythmic effects were obtained during reperfusion only with DL-sotalol, 10 μM (and with propranolol, 10 μM), which, unlike D-sotalol, has β-blocking properties.
Our results show that both DL- and D-sotalol (5 and 10 μM) are unable to prevent ischemia-induced APD shortening on guinea-pig isolated right ventricular myocardium, as previously described by Pasnani and Ferrier (17). However, these authors (a) used a concentration of 100 μM DL- and D-sotalol, which is higher than those used in our study, and more distant from the therapeutic plasma concentration range (10-30 μM), as reported in patients (35); and (b) performed their simulated ischemia globally by perfusion of the total isolated right ventricular free walls with a modified Tyrode's solution. Compared with their experiments, our in vitro ischemia-reperfusion model allowed simultaneous measurements of drug effects on both ischemic and normal adjacent myocardial zones of the same preparation. Moreover, we were able to observe the existence of class III actions of DL- and D-sotalol in the normal region, adjacent to the IZ, in which these effects were lost. These differential class III efficacies of DL- and D-sotalol on normal and ischemic tissues were expected, considering ionic conductances involved in electrophysiologic effects of both sotalol and ischemia. The AP shortening occurring under ischemic conditions has been attributed mainly to the activation of a potassium conductance dependent on intracellular adenosine triphosphate (ATP; IK-ATP; 36), whereas sotalol was mostly effective on the delayed outward-rectifying potassium current (IKr; 10,18). The depressing activity of sotalol on IKr was likely not sufficient for counteracting the IK-ATP-induced AP shortening during simulated ischemia. The different effects of class III agents in normal and ischemic myocardium were significant, considering the role of APD dispersion in the genesis of ventricular arrhythmias (29). A previous study (29) showed the promoting role of APD dispersion in the occurrence of spontaneous repetitive responses in our in vitro model of border zone arrhythmias. More recently, a clinical study (37) using simultaneous monophasic AP recordings from two sites of the right ventricle clearly proposed a link between the dispersion of repolarization and the inducibility of monomorphic ventricular tachycardia. Our data might confirm the importance of the differential class III effects of DL- and D-sotalol under different physiological conditions, around the border zone delimiting normal and ischemic myocardial tissues.
In our investigation, both DL- and D-sotalol, at concen-trations of 5 and 10 μM, showed proarrhythmic properties on spontaneous repetitive responses during simulated ischemia. In their in vitro model of global ischemia, Pasnani and Ferrier (17) found opposite effects between DL-sotalol and the dextrorotatory isomer D-sotalol on ischemia-induced arrhythmias: abolishment and enhancement, respectively, of arrhythmias at 100 μM. The latter data are consistent with our findings, considering the two different in vitro models used. Pasnani et al. (17,38) obtained arrhythmias showing the characteristics of a transmural reentry, whereas spontaneous arrhythmic events occurring in our model might be explained not only by transmural reentry but also by reentry occurring around the border zone between normal and ischemic regions. It has often been stated that the occurrence of reentry arrhythmias requires an unidirectional myocardial block. Actually, our results show an increase in the incidence of conduction blocks occurring between the two adjacent myocardial zones in the ischemia phase with DL-sotalol, 5 and 10 μM, in our experimental conditions, whereas DL-sotalol, 100 μM, prevented the occurrence of transmural blocks in the in vitro model of global ischemia (17). These differences arising from the experimental results, and probably related to the conduction pathway inducing conduction blocks, might explain, at least partly, the proarrhythmicity of DL-sotalol on border zone arrhythmias versus antiarrhythmicity found on arrhythmias occurring during a global simulated ischemia.
In vivo, restoration of coronary flow toward previously ischemic myocardium may elicit severe arrhythmias. These pathophysiologic alterations may be attributed to several factors, such as depletion of high-energy phosphates, sodium or calcium overload or both, or implication of reactive oxygen species such as hydrogen peroxide, superoxide radical, or hydroxyl radical. Pharmacologic agents such as free radical scavengers (39), calcium antagonists (40), and antiarrhythmic agents (41) may help in preventing ischemia-reperfusion injury. In our model of in vitro ischemia, only DL-sotalol, 10 μM, caused antiarrhythmic effects during reperfusion. The other compounds were ineffective against the recorded reperfusion arrhythmias. In their in vitro model of global ischemia, Pasnani and Ferrier (17) observed antiarrhythmic effects with DL-sotalol during reperfusion but not with the predominantly class III isomer, D-sotalol. However, the comparison between our results and the latter study requires care, considering the different concentrations of drugs and the experimental models of ischemia. Moreover, in our experimental conditions, reperfusion induced more frequent spontaneous arrhythmias than reperfusion after a global ischemia (in 86% of preparations in our model vs. 73% in the model of Pasnani and Ferrier); this might be the result of the presence of the border zone, which may worsen reperfusion arrhythmias. The antiarrhythmic effects observed with DL-sotalol, 10 μM, during reperfusion cannot be related to variations of the AP amplitude and the Vmax, as suggested by the results of the multivariate statistical analysis based on the Weibull's model, because these latter AP parameters were not affected by this class III agent. This antiarrhythmic action might be explained by an increase of refractoriness in the nonischemic part, allowing termination of arrhythmias from reentrant circuits and also by the β-blocking effect of DL-sotalol, 10 μM. The antiarrhythmic efficacy obtained in our model with the pure β-blocking agent propranolol, 10 μM, indicates that arrhythmias observed during the reperfusion phase might be partly caused by adrenergic-system stimulation. On the other hand, it could explain the inability of D-sotalol, devoid of β-blocking properties, to prevent these reperfusion-induced spontaneous repetitive responses. It was recently shown that, during sympathetic overstimulation in exercising dogs with acute myocardial ischemia, both DL-sotalol and propranolol were protective against the onset of ventricular fibrillation, whereas D-sotalol was not (23). In this latter in vivo study, however, the antiarrhythmic efficacy of β-blocking agents was related in part to their bradycardic action. In isolated guinea-pig ventricular myocytes, adrenergic stimulation by isoproterenol reduced the prolongation of AP induced by D-sotalol (24), suggesting that, in the presence of enhanced adrenergic activity, as might likely occur in our in vitro model of reperfusion, the electrophysiologic effects of D-sotalol are impaired. On the other hand, results obtained in our laboratory with the pure class III agent dofetilide, 5, 10, and 50 nM, indicate that, as seen with D-sotalol, dofetilide does not contribute to the prevention of ischemia-reperfusion repetitive responses in our model (unpublished data).
Although the existence of a border zone between normal and hypoxic/ischemic regions is still controversial (28), such border zone, associated with inhomogeneous distribution of electrophysiologic properties, anatomic changes, and biochemical changes, seems well established as a major site of arrhythmia origin (26,27). On the other hand, injury currents, thought to be a possible mechanism leading to arrhythmias, such as automatic activities, focal reexcitations, or reentry arrhythmias (42), have a recognized origin in the border zone, as indicated by investigations in isolated porcine and canine hearts (43). More recently in sheep Purkinje fibers, Kupersmith et al. (44) clearly demonstrated that APDs and membrane potential inhomogeneities led to electrotonic transmission of an injury current to border zones adjacent to zones of abnormal APD prolongation and that this injury current led to triggered activities. All these findings support interest in the study of the electrophysiologic effects of drugs modifying APD, such as class III agents, on the border zone separating normal and ischemic myocardial regions, as well as in effects on the border zone arrhythmias. Taking into account the recent results of the SWORD study (1), our results might clarify the electrophysiologic effects of these agents and help the under-standing of the proarrhythmic activities of DL- and D-sotalol at pharmacologic concentrations of 5 and 10 μM.
The validity and the relevance of our in vitro model with regard to the production of arrhythmias during ischemia and reperfusion was previously discussed (30), recognizing this model as able to produce spontaneous arrhythmias similar to those observed in vivo during coronary occlusion and after reperfusion (45) or in humans during coronary transluminal angioplasty or enzymatic recanalization (46). However, the model of simulated ischemia-reperfusion used in our experiments has some limitations when considering the anatomic concept of the border zone. Indeed, the partition between ischemic and normal myocardial regions in our experiments was narrow and regular. In diseased states, this zone may be larger, as in chronic infarction and fibrosis, and may be the source of several arrhythmic states; this is less clear for acute ischemia. Moreover, in vivo, the demarcations between diseased and normal tissue are probably irregular. However, this does not modify the implications of our findings for the understanding of simultaneous electrophysiologic and proarrhythmic effects of DL- and D-sotalol around the myocardial border zone.
In conclusion, our work relates, in an in vitro investigation of a single cardiac preparation, the electrophysiologic effects of DL- and D-sotalol on both normal and ischemic tissues separated by a border zone. Our results provide evidence for proarrhythmic activities of DL- and D-sotalol on the border zone arrhythmias occurring during simulated ischemia and reperfusion (except with DL-sotalol, 10 μM, on reperfusion), and different class III efficacy of DL- and D-sotalol on both normal and ischemic adjacent myocardial regions. These and other data (17,23,24) might be useful to help explain the results of the SWORD study (1). Our findings confirm caution in the clinical use of presently available class III agents in the presence of myocardial ischemia, and the clinical interest that could arise from in vitro electrophysiologic studies of cardioactive drug effects simultaneously explored in normal and diseased adjacent myocardial regions.
Acknowledgment: This study was supported by a grant from Bristol-Myers Squibb. We thank Mrs. Veronique Sabalos, engineer, and Mr. Michel Morel, engineer, for their help in the computer treatment of signals acquired by DATAPAC.
1. Cobbe SM. Class III antiarrhythmics: put to the SWORD? Heart
2. Singh BN, Nademanee K. Control of cardiac arrhythmias by selective lengthening of repolarization: theoretic considerations and clinical observations. Am Heart J
3. Finance O, Manning A, Chatelain P. Effects of a new amiodarone-like agent, SR 33589, in comparison to amiodarone, D,L-sotalol and lignocaine, on ischemia-induced ventricular arrhythmias in anesthetized pigs. J Cardiovasc Pharmacol
4. Deedwania P. Suppressant effects of conventional beta-blockers and sotalol on complex and repetitive ventricular premature complexes. Am J Cardiol
5. Lynch J, Wilber D, Montgomery D, Hsieh T, Patterson E, Lucchesi B. Antiarrhythmic and antifibrillatory actions of the levo and dextrorotatory isomers of sotalol. J Cardiovasc Pharmacol
6. Roden DM. Usefulness of sotalol for life-threatening ventricular arrhythmias. Am J Cardiol
7. Hohnloser SH, Zabel M, Krause T, Just H. Short- and long-term antiarrhythmic and hemodynamic effects of D,L-sotalol in patients with symptomatic ventricular arrhythmias. Am Heart J
8. Hohnloser S, Meinertz T, Stubbs P, et al. Efficacy and safety of D-sotalol, a pure class III antiarrhythmic compound, in patients with symptomatic complex ventricular ectopy: results of a multicenter, randomized, double-blind, placebo-controlled dose-finding study. Circulation
9. Camm AJ, Paul V. Sotalol for paroxysmal supraventricular tachycardias. Am J Cardiol
10. Carmeliet E. Electrophysiologic and voltage clamp analysis of the effects of sotalol on isolated cardiac muscle and Purkinje fibers. J Pharmacol Exp Ther
11. Varró A, Nakaya Y, Elharrar V, Surawicz B. Effect of antiarrhythmic drugs on the cycle length-dependent action potential duration in dog Purkinje and ventricular muscle fibers. J Cardiovasc Pharmacol
12. Ishii M, Kamiya J, Hashimoto K. Cellular electrophysiological effects of MS-551, a new class III antiarrhyhthmic agent. Drug Dev Res
13. Lathrop DA. Electro-mechanical characterization of racemic sotalol and its optical isomers on isolated canine ventricular trabecular muscles and Purkinje strands. Can J Physiol Pharmacol
14. Montero M, Schmitt C. Differential effects of D-sotalol on endocardial and epicardial action potentials of human ventricular myocardium in dilated cardiomyopathy. J Cardiovasc Pharmacol
15. McComb JM, McGovern B, McGowan JB, Ruskin JN, Garan H. Electrophysiologic effects of D-sotalol in humans. J Am Coll Cardiol
16. Antonaccio MJ, Gomoll AW. Sotalol-pharmacological and antiarrhythmic effects. Cardiovasc Drug Rev
17. Pasnani JS, Ferrier GR. Pro- and anti-arrhythmic effects of D-sotalol and DL-sotalol in an isolated tissue model of ischemia and reperfusion. J Pharmacol Exp Ther
18. Lipka L, Dizon J, Reiffel J. Desired mechanisms of drugs for ventricular arrhythmia: class III antiarrhythmic agents. Am Heart J
19. Hohnloser SH, Singh BN. Proarrhythmia with class III antiarrhythmic drugs: definition, electrophysiologic mechanisms, incidence, predisposing factors, and clinical implications. J Cardiovasc Electrophysiol
20. Reisinger J, Shenasa M, Lubinski A, et al. Clinical implications of pleomorphic ventricular tachycardias on oral sotalol therapy. Eur Heart
21. Hondeghem L, Snyders D. Class III antiarrhythmic agents have a lot of potential but a long way to go: reduced effectiveness and dangers of reverse use dependence. Circulation
22. Claudel JP, Touboul P. Sotalol: From "just another beta blocker" to "the prototype of class III antidysrhythmic compound." Pacing Clin Electrophysiol
23. Vanoli E, Hull SS Jr, Adamson PB, Foreman RD, Schwartz PJ. K+
channel blockade in the prevention of ventricular fibrillation in dogs with acute ischemia and enhanced sympathetic activity. J Cardiovasc Pharmacol
24. Vanoli E, Priori SG, Nakagawa H, et al. Sympathetic activation, ventricular repolarization and Ikr
blockade: implications for the antifibrillatory efficacy of potassium channel blocking agents. J Am Coll Cardiol
25. Hofman H. Interaction between a normoxic and a hypoxic region of guinea pig and ferret papillary muscle. Circ Res
26. Janse MJ, Cinca J, Morena H, et al. The "border zone" in myocardial ischemia: an electrophysiological, metabolic and histochemical correlation in the pig heart. Circ Res
27. El-Sherif N, Mehra R, Gough WB, Zeiler RH. Ventricular activation patterns of spontaneous and induced ventricular rhythms in canine one-day-old myocardial infarctions: evidence for local and reentrant mechanisms. Circ Res
28. Hearse DJ, Yellon DM. The "border zone" in evolving myocardial infarction: controversy or confusion? Am J Cardiol
29. Schiariti M, Puddu PE, Rouet R. Multivariate prediction of spontaneous repetitive responses in ventricular myocardium exposed in vitro to simulated ischemic conditions. Int J Cardiol
30. Rouet R, Adamantidis M, Honore E, Dupuis B. In vitro abnormal repetitive responses in guinea pig ventricular myocardium exposed to combined hypoxia, hyperkalemia and acidosis. J Appl Cardiol
31. Puddu PE, Rouet R. Are different pharmacokinetic properties of lidocaine versus propafenone in normal as compared to ischemic myocardium responsible for different electrophysiologic effects which help preventing arrhythmias? [Editorial comment]. Cardiologia
32. Belichard P, Pruneau D, Rouet R, Salzmann JL. Electrophysiological responses of hypertrophied rat myocardium to combined hypoxia, hyperkalemia and acidosis. J Cardiovasc Pharmacol
33. Morena H, Janse MJ, Fiolet JWT, Krieger WJG, Crijns H, Durrer D. Comparison of the effects of regional ischemia, hypoxia, hyperkalemia and acidosis on intracellular and extracellular potentials and metabolism in isolated porcine heart. Circ Res
34. Afifi AA, Clark V. Computed-aided multivariate analysis.
2nd ed. New York: Van Nostrand Reinhold, 1990:353-8.
35. Neuvonen PJ, Elonen E, Tanskanen A, Tuomilehto J. Sotalol prolongation of the QT-c interval in hypertensive patients. Clin Pharmacol Ther
36. Coetzee WA. ATP-sensitive potassium channels and myocardial ischemia: why do they open? Cardiovasc Drug Ther
37. Yuan S, Wohlfart B, Olsson SB, Blomström-Lundqvist C. The dispersion of repolarization in patients with ventricular tachycardia. Eur Heart J
38. Ferrier GR, Guyette CM. Ventricular tachycardia in an isolated guinea pig ventricular free wall model of ischemia and reperfusion. J Cardiovasc Pharmacol
39. Myers ML, Bolli R, Lekich RF, Hartley CJ, Roberts R. Enhancement of recovery of myocardial function by oxygen free-radical scavengers after reversible regional ischemia. Circulation
40. Tilton RG, Watts JA, Land MP, Larson KB, Sutera SP, Williamson JR. Discordant effects of nisoldipine on coronary vascular resistance and permeability changes during reflow after ischemia in isolated rabbit hearts. J Mol Cell Cardiol
41. Manning AS, Come R, Isted K, Coltart DJ, Hearse DJ. Reperfusion-induced ventricular fibrillation: modification by pharmacological agents. Adv Myocardiol
42. Duo GS, Hoff P, Kupersmith J. Repolarization interactions between cardiac segments of varying action potential duration. Am Heart J
43. Janse MJ, VanCapelle JL, Morsink H, et al. Flow of "injury" current and patterns of excitation during early ventricular arrhythmias in acute regional myocardial ischemia in isolated porcine and canine hearts: evidence for two different arrhythmogenic mechanisms. Circ Res
44. Kupersmith J, Li ZY, Maldonado C. Marked action potential prolongation as a source of injury current leading to border zone arrhythmogenesis. Am Heart J
45. Corr PB, Witowski FX. Potential electrophysiologic mechanisms responsible for dysrhythmias associated with reperfusion of ischemic myocardium. Circulation
46. Tzivoni D, Keren A, Granot H, Gottlieb S, Benhorin J, Stern S. Ventricular fibrillation caused by myocardial reperfusion in Prinzmetal angina. Am Heart J