Decreasing the infusion rate reduces the proarrhythmic risk of NS-7: confirming the relevance of short-term variability of repolarisation in predicting drug-induced torsades de pointes

Abstract

1. The rate of infusion has been suggested to be important for drug-induced torsades de pointes (TdP) arrhythmias. We investigated the repolarisation-prolonging effects and proarrhythmic properties of NS-7, a neuroprotective drug in development, using two different infusion rates.

2. A fast (5 min intravenously (i.v.)) escalating dosing regimen (0.3 and 3.0 mg kg⁻¹, n=4) of NS-7 was investigated in anaesthetised control dogs in sinus rhythm (SR). This was compared to a slow infusion (60 min i.v.) of one dose (3.0 mg kg⁻¹, n=4) NS-7. The similar dosing regimens were investigated in anaesthetised dogs with chronic, complete AV block (CAVB), an animal model of TdP (n=6).

3. No electrophysiological effects were seen after 0.3 mg kg⁻¹ NS-7. Fast infusion of 3.0 mg kg⁻¹ caused prolongation of repolarisation, for example, heart rate corrected QT interval (QT_c): in SR: 6±1%; in CAVB: 10±7%, which was accompanied by TdP in three of six CAVB dogs. No TdP were seen in SR dogs.

4. Slow infusion did not cause TdP in the same CAVB dogs, although NS-7 caused repolarisation to prolong with a similar magnitude (QT_c: 12±7%) as in the fast-infusion experiment.

5. Short-term variability (STV) is a novel parameter for the prediction of drug-induced TdP analysing the beat-to-beat variability of repolarisation. STV was only increased after the fast infusion in CAVB dogs (2.6±0.3 versus 6.0±1.4 ms, P<0.05), while there was no increase (2.1±0.2 versus 2.5±1.0 ms) after the slow infusion of NS-7.

6. Peak plasma concentrations attained were lower in slow (0.5±0.1 μg ml⁻¹ after 50 min) than in fast-infusion regimen (2.1±0.4 μg ml⁻¹ after 5 min; P<0.05).

7. The results support the conclusion that limiting peak plasma concentration by decreasing the rate of infusion of NS-7 reduces the proarrhythmic risk despite comparable prolongation in repolarisation parameters. The relevance of STV in predicting drug-induced TdP was confirmed.

Introduction

A large number of cardiovascular and noncardiovascular drugs have been withdrawn from the market in the past decade due to postmarketing reports of clinical drug-induced lethal proarrhythmia (Haverkamp et al., 2000; Belardinelli et al., 2003; Redfern et al., 2003). Most of these drugs have in common that they delay the ventricular repolarisation seen as QT prolongation on the ECG (Haverkamp et al., 2000; Belardinelli et al., 2003). However, numerous studies have shown that there is no straightforward relation between prolongation of repolarisation and the potentially lethal arrhythmia known as torsades de pointes (TdP) (Carlsson et al., 1993; Hondeghem et al., 2001a, 2001b; van Opstal et al., 2001b; Antzelevitch & Shimizu, 2002; Milberg et al., 2002; Thomsen et al., 2003, 2004). Still, the regulatory authorities require that any new potential drug is investigated to determine both its repolarisation delaying and proarrhythmic properties.

NS-7 (enecadine) is a multichannel blocker (I_Na, I_Ca and I_K1) reported to show promising efficacy in animal models of ischaemic stroke (Suma et al., 1997; Oka et al., 2000, 2001; Sopala et al., 2002; Tanaka et al., 2002; Katsumata et al., 2003). In preclinical electrophysiological studies, 100 μM NS-7 prolonged the ventricular action potential in guinea-pigs (Satoh, 2003). In the present study, the electrophysiological and proarrhythmic characteristics of the drug were assessed in anaesthetised dogs with normal and electrically remodelled hearts. For the latter purpose, we used the animal model with chronic, complete atrioventricular block (CAVB) that shows a high susceptibility to drug-induced TdP (Weissenburger et al., 1991; Verduyn et al., 1997, 1999; Vos et al., 1998; Thomsen et al., 2003; Chiba et al., 2004).

Previously, the rate of drug administration has been shown to be critical for TdP induction at similar prolongation of the QT intervals in rabbits (Carlsson et al., 1993). Infusion rate has later been included as a risk factor for TdP observed in the clinical setting (Roden, 2004). Therefore, we employed two different infusion rates of the same total dose in this investigation, along with the analysis of beat-to-beat variability of repolarisation (BVR). This methodology was recently demonstrated to be a superior predictor for drug-induced TdP in the preclinical setting (Hondeghem et al., 2001a; Thomsen et al., 2004).

Methods

Animal handling was in accordance with the European directive for the protection of vertebrate animals used for experimental and other scientific purposes (86/609/EU). The committee for experiments on animals of Maastricht University approved all experiments.

General

A total of 28 experiments were performed on 14 anaesthetised mongrel dogs (body weight 22–33 kg) under aseptic conditions. After overnight fasting, anaesthesia was induced by 20 mg kg⁻¹ sodium pentobarbital intravenously (i.v.) and maintained by 0.5% halothane in a mixture of O₂ and N₂O (1 : 2). During artificial ventilation at a frequency of 10–12 min⁻¹, tidal volume (10–15 ml kg⁻¹) was adjusted to maintain end-expired CO₂ concentration between 3.5 and 4.0%. A thermal mattress was used to maintain body temperature. To prevent volume depletion, the dog received 0.5–1.0 l 0.9% saline i.v. Postoperative care included antibiotics (1 g ampicillin intramuscularly (i.m.)) and analgesics (15 μg kg⁻¹ buprenorphine i.m.).

In all, 10 ECG leads and two monophasic action potentials (MAPs) were recorded continuously throughout the experiments. MAP catheters (EP technologies, CA, U.S.A.) were placed under fluoroscopic guidance on the endocardium of the left (LV) and right ventricle (RV). MAP signals were amplified with a customised isolated DC-coupled differential amplifier at a frequency range of 0–500 Hz with a 20-mV calibration pulse. The offset of the amplifier was variable and could be adjusted to the recorded signal. Besides minimal amplitude of 15 mV, the MAP was required to have a smooth repolarisation and a stable configuration in time.

Experimental design

Eight dogs were studied during sinus rhythm (SR) to determine (1) the dose-dependent electrophysiological effects of NS-7 in physiologically normal hearts and (2) the associated plasma concentrations of NS-7. NS-7 was administered at a fast-infusion (5 min) scheme of escalating doses (0.3, 3.0 and 10 mg kg⁻¹ i.v., n=4) separated by 30 min. Furthermore, a slow-infusion (60 min) scheme of 3.0 mg kg⁻¹ (i.v., n=4) was performed.

In nine dogs, complete atrioventricular block was created by radiofrequency ablation. After 4±1 weeks of atrioventricular block, the dogs underwent an anaesthetised dofetilide-inducibility test (25 μg kg⁻¹ i.v.) to determine their sensitivity to drug-induced TdP (van Opstal et al., 2001a). As NS-7 could have a low proarrhythmic potential that could escape detection in this model, we opted to exclude dogs that were not susceptible to this positive control (Thomsen et al., 2003). Of nine dogs with CAVB, two failed the dofetilide-inducibility test and one was lost due to drug-induced ventricular fibrillation. Thus, six dogs were used for evaluating the electrophysiological and proarrhythmic properties of NS-7 in remodelled hearts.

NS-7 was administered in a fast-infusion (5 min) scheme of escalating doses (0.3 and 3.0 mg kg⁻¹ i.v.) separated by 30 min or in a slow-infusion (60 min) scheme of 3.0 mg kg⁻¹ (i.v.). This was performed in a serial, random crossover design with 2 weeks between experiments. One dog was lost during a fast-infusion experiment, before the slow-infusion experiment was performed.

In two dogs, the dofetilide challenge was repeated after NS-7 to ascertain preservation of TdP inducibility. NS-7 and dofetilide were provided by PAION. Both drugs were dissolved in equal volumes of 0.9% saline.

Plasma concentrations

Blood samples were drawn from a dedicated venous access at various time points throughout the experiments. Samples were collected in citrate tubes, centrifuged for 10 min at 4000 r.p.m. at 4°C and stored at −20°C. NS-7 concentrations were determined by Scope International (Hamburg, Germany).

Data analysis

Applying a custom-made computer programme (ECGview), we measured the following parameters offline at a resolution of 4 ms: RR and QT intervals from a unipolar lead positioned on the sixth intercostal space near the edge of the sternum (Ettinger & Suter, 1970). Also, the LV and RV monophasic action potential duration (MAPD) to 50 and 100% repolarisation was determined (MAPD₅₀ and MAPD₁₀₀, respectively). Heart rate corrected QT intervals (QT_c) were calculated according to van de Water's formula (Van de Water et al., 1989). Interventricular dispersion of repolarisation (ΔMAPD) was defined as LV minus RV MAPD.

BVR was determined according to an earlier publication (Thomsen et al., 2004). Briefly, Poincaré plots were drawn from 30 consecutive LV MAPD and short-term variability (STV, Equation 1), representing the mean orthogonal distance to the line-of-identity, was calculated

where D_n represents LV MAPD of beat number n.

All electrophysiological parameters were measured at maximal QT prolongations, which were 10 and 60 min after the start of the fast and slow infusion, respectively.

The frequency of multiple extrasystoles was quantified in 10-min intervals after the administration of NS-7 in CAVB dogs. Extrasystoles were defined as premature ventricular complexes occurring at a coupling interval of less than 600 ms.

TdP was defined as polymorphic ventricular tachycardia of at least five beats. A dog was defined as inducible when >3 TdP occurred. If a TdP degenerated into ventricular fibrillation, electrical cardioversion was applied.

Statistical analysis

Pooled data are expressed as mean±s.d., unless otherwise stated. Comparisons of electrophysiological data were performed with repeated-measures ANOVA followed by a Bonferroni t-test. Statistical significance was acknowledged at P<0.05.

Results

Fast-infusion scheme in control SR dogs

Infusing NS-7 over 5 min resulted in a dose-dependent increase in RR interval (baseline: 460±50 ms; 0.3 mg kg⁻¹: 475±35 ms; 3 mg kg⁻¹: 530±70 ms (P<0.05) and 10 mg kg⁻¹: 540±25 ms (P<0.05)). Prolongation of electrophysiological repolarisation parameters was seen only after the higher doses of 3.0 and 10 mg kg⁻¹ (Figure 1 for QT_c time and Table 1 for 3.0 mg kg⁻¹). Relatively, 3 mg kg⁻¹ NS-7 caused a 5.8±1% prolongation of the QT_c interval from baseline, whereas 10 mg kg⁻¹ caused a 12.6±2% prolongation. The plasma concentrations over the duration of the experiments are depicted in Figure 1, at peak reaching 0.2±0.02, 2.1±0.4 and 6.4±3.1 μg ml⁻¹ after 0.3, 3.0 and 10 mg kg⁻¹ NS-7, respectively.

Figure 1.

Dose-dependent development in QT_c interval and plasma concentration after 5-min infusions of NS-7 in anaesthetised dogs in SR. Three cumulative doses of NS-7 were administered i.v. as indicated by arrows (n=4 dogs). ^*P<0.05 versus value before last administration. Mean±s.e.m.

Table 1.

Electrophysiological data from anaesthetised dogs with normally conducted SR (n=4)

	Control	NS-7
RR (ms)	460±50	530±70*
QT (ms)	265±15	290±20*
QTc (ms)	310±10	330±15*
LV MAPD (ms)	215±5	230±15*
RV MAPD (ms)	210±5	225±5*
ΔMAPD (ms)	5±5	5±5
STV (ms)	0.8±0.1	1.0±0.4

Measurements were performed 10 min after the start of the 5-min infusion of 3.0 mg kg⁻¹ NS-7 when the QT was maximally prolonged.

^*P<0.05 versus control.

Administration of 10 mg kg⁻¹ NS-7 caused a brief fall in the QT_c interval (Figure 1) attributable to a consistent drug-induced transient shortening of the RR interval from 510±45 to 440±20 ms (P<0.05). No statistically significant drug-induced increase was seen in the interventricular dispersion (ΔMAPD) or BVR (STV, Table 1).

Fast-infusion scheme in proarrhythmic CAVB dogs

NS-7 (0.3 mg kg⁻¹) caused no proarrhythmia, while 3.0 mg kg⁻¹ caused reproducible TdP in three of six CAVB dogs, associated with a QT_c prolongation of 1.8±1 and 10.3±7%, respectively. After the fast infusions, the peak plasma concentrations of NS-7 in CAVB dogs (0.2±0.01 and 1.6±1.9 μg ml⁻¹ after 0.3 and 3.0 mg kg⁻¹ NS-7, respectively) were similar to those seen in SR dogs. A representative example of TdP caused by 3.0 mg kg⁻¹ NS-7 is shown in Figure 2. Representative Poincaré plots of the LV MAPD₁₀₀ at control and under the influence of 3.0 mg kg⁻¹ NS-7, either at fast or slow infusion, are also shown in Figure 2. Table 2 summarises the electrophysiological changes induced by 3.0 mg kg⁻¹ NS-7. During fast infusion of 3.0 mg kg⁻¹ NS-7, significant increases in LV MAPD₁₀₀ and STV were observed. On average 9±20 multiple extrasystoles were encountered within the first 10 min after administration. One dog was lost after a drug-induced ventricular fibrillation, where electrical cardioversion was not feasible.

Figure 2.

(a) Representative ECG traces before and after administration of 3.0 mg kg⁻¹ NS-7 in an anaesthetised dog with CAVB. TdP (right-most panel) only occurred after the fast infusion of NS-7. Two ECG leads (II, lead II; LL, precordial lead placed on the left lateral side of the thorax) and LV MAP recordings are shown in each trace. RR intervals are above and QT time below lead LL. MAPD₁₀₀ is below the LV MAP trace. ECG calibrated to 1 mV cm⁻¹. Vertical scale bar, 20 mV on the MAP signal; horizontal scale bar, 1 s. (b) Poincaré plots of the LV MAPD₁₀₀ from the same dog at the two experiments. Open arrow, control; closed arrow, 3.0 mg kg⁻¹ NS-7. A substantial increase in plot area is appreciable after fast administration, but not after the slow administration. STV in the fast-infusion experiment increases from 3.0 ms at control to 8.8 ms after NS-7, while STV changes from 2.3 to 2.1 ms in the slow-infusion experiment.

Table 2.

Electrophysiological data from serial investigations in anaesthetised dogs with CAVB (n=5)

	5-min infusion		60-min infusion
	Control	NS-7	Control	NS-7
RR (ms)	1240±155	1300±115	1255±215	1370±275
QT (ms)	440±45	480±50	450±30	495±50
QTc (ms)	420±40	450±45	425±20	460±30
LV MAPD100 (ms)	365±50	420±40*	350±10	410±55*
LV MAPD50 (ms)	285±30	325±50	275±15	295±50
RV MAPD100 (ms)	300±35	345±35	295±30	340±55
RV MAPD50 (ms)	245±40	280±35	240±35	280±25
ΔMAPD100 (ms)	65±40	75±45	55±30	70±10
ΔMAPD50 (ms)	40±25	45±45	35±30	15±50
STV LV MAPD100 (ms)	2.6±0.3	6.0±1.4*	2.1±0.2	2.5±1.0†
STV LV MAPD50 (ms)	2.1±0.7	4.9±1.6*	1.9±0.6	2.1±0.5†

NS-7 (3.0 mg kg⁻¹) was administered over 5 min or over 60 min. Measurements were performed at 10 and 60 min, respectively, when the QT was maximally prolonged.

^*P<0.05 versus control;

^†P<0.05 versus 5-min infusion.

Slow-infusion scheme in control SR dogs

Administering 3.0 mg kg⁻¹ NS-7 over 60 min caused an increase in RR (550±55 versus 625±40 ms; P<0.05). The QT_c interval was prolonged to a similar extent as with the fast-infusion scheme (8.1±2 versus 5.8±1%; Figure 3). The maximal values were obtained 60 min after the start of the slow infusion versus 10–15 min in the fast-infusion experiments. The corresponding plasma concentrations of NS-7 (Figure 3) show significant differences in peak levels (0.5±0.1 versus 2.1±0.4 μg ml⁻¹ for slow and fast infusion, respectively; P<0.05).

Figure 3.

Comparison of the 5- and 60-min infusion of 3.0 mg kg⁻¹ NS-7 in anaesthetised dogs in SR. QT_c (upper panel) and plasma concentration (lower panel) are depicted over time for the slow (open circles) and the fast (closed circles) i.v. infusion experiments (n=4 dogs). QT_c is similarly increased after both infusion schemes; however, peak plasma concentrations are higher after the fast infusion. ^*P<0.05 versus 0 min. Mean±s.e.m.

Slow-infusion scheme in proarrhythmic CAVB dogs

As opposed to the fast-infusion scheme, slow infusion of 3.0 mg kg⁻¹ caused no TdP, although the LV MAPD and QT_c time (11.9±9 versus 10.3±7%) were prolonged to similar extents (Table 2). The plasma concentration of NS-7 was similar in CAVB and SR dogs after the slow infusions (0.4±0.1 μg ml⁻¹ after 3.0 mg kg⁻¹ NS-7). STV was not altered by the slow infusion of NS-7. Furthermore, STV was significantly higher after a fast administration of 3.0 mg kg⁻¹ NS-7 than after a slow infusion of the same dose (Table 2). No multiple extrasystoles were observed; however, this was not acknowledged as statistically significantly different from the fast infusion, partly due to the very large variation in the frequency in the latter experiments.

Reproducibility of TdP induction with dofetilide in time was confirmed in two experiments.

Discussion

In this study, we show that decreasing the infusion rate of NS-7 limits the peak plasma concentration and decreases or abolishes the risk of TdP. The only electrophysiological parameter that reflected the absence of TdP was BVR, quantified as STV. This confirms the inability of the prolonged QT interval to predict TdP. Furthermore, this study stresses the importance of assessing cardiac electrophysiological safety of drugs in animal models of reduced repolarisation reserve, because the same dose of NS-7 did not cause proarrhythmia in SR dogs.

Proarrhythmic assessment of NS-7

NS-7 is under development as a neuroprotective drug after ischaemic stroke. The beneficial actions of the drug are thought to be based on block of voltage-activated sodium and calcium channels in neuronal tissue (Oka et al., 2000, 2001; Tanaka et al., 2000). Block of I_Kr (IC₅₀: 0.4 μmol l⁻¹; PAION; data on file) may underlie the prolongation of action potential duration observed in the present study and in isolated guinea-pig ventricular myocytes (Satoh, 2003). Specifically, block of I_Kr has been associated with an increased risk of repolarisation-dependent proarrhythmia (Redfern et al., 2003).

Regulatory authorities like the US Food and Drug Administration or the European Medicines Agency encourage that all potential drugs with suspected QT-prolonging properties in humans should be tested preclinically for their ability to delay repolarisation and for their proarrhythmic characteristics. Drug testing in normal hearts is essential for the analysis of delayed repolarisation in general; however, it is not sufficient for the recognition of potential proarrhythmic effects in the diseased heart of a predisposed, vulnerable individual patient. The canine model with CAVB is known to have acquired QT prolongation and a predisposition to TdP and sudden cardiac death (Verduyn et al., 1997; Vos et al., 1998; van Opstal et al., 2001a, 2001c; Thomsen et al., 2003). The CAVB dogs exposed to NS-7 in this investigation were selected on the basis of a positive dofetilide-inducibility test, increasing the sensitivity of the model (Thomsen et al., 2003). The incidence of dofetilide-induced TdP was seven of nine dogs, comparable to earlier investigations (van Opstal et al., 2001a; Thomsen et al., 2003), and was reproducible in the two dogs, which were retested after the NS-7 experiments had been performed.

To be able to compare NS-7 to other drugs in this animal model, we chose our regular infusion time of 5 min as the fast rate (Verduyn et al., 1997; van Opstal et al., 2001b; Thomsen et al., 2003). In an approach to the clinical setting, the slow infusion went over a period of 60 min, which is still faster than the anticipated rate in humans.

To the best of our knowledge, this is the first study to report prolonged cardiac repolarisation and proarrhythmia by NS-7 in the intact animal. A dose-dependent prolongation of the QT_c interval is appreciable after fast administration of NS-7 in both normal (Figure 1) and CAVB dogs. Peak plasma concentrations clearly exceeded the putative effective therapeutic plasma concentration of 15–29 ng ml⁻¹ (Aoki et al., 2001). TdP was only seen after fast infusion of 3.0 mg kg⁻¹ NS-7 in the remodelled hearts of the CAVB dog. The highest dose (10 mg kg⁻¹) was not administered to CAVB dogs, as TdP were already evident at 3.0 mg kg⁻¹.

Relevance of infusion time

As expected, peak plasma concentrations were markedly lower during slow than during fast infusion. Even so, prolongation of the infusion time did not reduce the effects seen on QT_c intervals, LV MAPD or interventricular dispersion, neither in control dogs (Figure 3 for QT_c) nor in CAVB dogs (Table 2). On the other hand, proarrhythmia was completely absent during the slow infusion. Thus, at two different plasma concentrations of NS-7, repolarisation was similarly prolonged, suggesting that a certain drug concentration is sufficient to produce the observed prolongation of repolarisation, but a higher concentration is necessary to destabilise repolarisation, trigger ectopic beats and sustain TdP. This vulnerability is likely restricted to areas of lowest repolarisation reserve (Roden, 1998). In the CAVB dog, these areas seem to be located in the subendocardial regions of the LV (Thomsen et al., 2004).

Our data confirm the finding of Carlsson et al. (1993), who showed that a fast infusion of the I_Kr blocker almokalant (25 nmol kg⁻¹ min⁻¹) produced TdP in nine of 10 methoxamine-treated anaesthetised rabbits, while a slow infusion (5 nmol kg⁻¹ min⁻¹) gave an inducibility of one of eight. This was associated with a 42% increase in QT_c after the slow infusion, but only 30% prolongation after the fast infusion, once more confirming that QT prolongation is not directly convertible into risk of TdP, a conclusion reached by numerous groups (Hondeghem et al., 2001a; van Opstal et al., 2001b; Antzelevitch & Shimizu, 2002; Milberg et al., 2002; Thomsen et al., 2004). We have compared amiodarone and its noniodinated successor, dronedarone, in chronically dosed CAVB dogs and demonstrated equal QT prolongations, contrary to a different proarrhythmic outcome (van Opstal et al., 2001b). Similar conclusions were drawn from dose-dependent investigations employing the antipsychotic drug, sertindole, or the antiarrhythmic drug, D-sotalol (Thomsen et al., 2003, 2004). In isolated rabbit hearts, Hondeghem et al. (2001a, 2001b) showed that parameters like instability and triangulation of the action potential were proarrhythmic, while action potential prolongation was antiarrhythmic. In a comparison between the macrolide antibiotics erythromycin, clarithromycin and azithromycin, it was shown that they all caused similar increases in repolarisation duration; however, only the latter was devoid of proarrhythmia (Milberg et al., 2002). Other antibiotics like gatifloxacin and moxifloxacin have developed TdP in conscious CAVB dogs despite the absence of drug-induced QT prolongation (Chiba et al., 2004). Moreover, the experimental calmodulin inhibitor W-7 was able to suppress drug-induced TdP without shortening QT intervals (Mazur et al., 1999; Gbadebo et al., 2002).

Beat-to-beat variability of repolarisation

Repolarisation reserve has been introduced as a concept to explain susceptibility to arrhythmia (Roden, 1998; Biliczki et al., 2002). A reduction in repolarisation reserve generates an action potential that is more vulnerable to additional challenges upon repolarisation such as drugs with I_Kr blocking properties. Often I_Kr blockers are the final challenge on repolarisation, which unmasks an unidentified predisposition precipitating a lethal arrhythmia. This predisposition can be a congenital (long QT syndrome) or an acquired (metabolic or electrolyte disturbances, heart disease, etc.) ion channelopathy. Quantification of the repolarisation reserve and identification of the vulnerable patient are therefore important questions for many investigators.

BVR is a concept that provides data to identify unsafe drugs. With the use of Poincaré plots, STV is one way of quantifying BVR. A dose-dependent TdP occurrence after D-sotalol was tightly associated with the elevation of STV, while the absence of TdP after chronic, oral amiodarone was reflected in an unchanged STV (Thomsen et al., 2004). The global nature of the QT interval versus the relatively local recording of an MAP signal possibly explains the low predictive power of beat-to-beat variability of the QT interval. Furthermore, the predominant remodelling of the LV may underlie the absence of drug-induced increases in the BVR of the RV MAPD (Thomsen et al., 2004). With the assumption that it is a population of cells located in the left ventricular subendocardium that has the lowest repolarisation reserve, we would not expect to find physiologically significant drug-induced increases of the STV of the epicardial repolarisation (Sicouri & Antzelevitch, 1991). This is the first study to show that within the STV of the LV MAPD, variability is present at both 50 and 100% repolarisation levels, supporting the hypothesis that BVR arises at the plateau of the action potential rather than at the fast repolarisation.

In the present study, we have applied STV and confirmed our earlier observation that an increase in BVR predicts TdP. The only NS-7 induced increase in STV was observed in the fast-infusion experiments in the remodelled CAVB dogs. In normal hearts as well as with the slow infusion of NS-7 in CAVB dogs, STV was not altered and TdP was absent. Thus, BVR is an attractive addition to assess proarrhythmic actions of drugs. Additionally, STV is present prior to drug-induced early after-depolarisation and extrasystoles, which allows time for preventive antiarrhythmic actions.

Novel parameters in drug screening

In vivo, repolarisation-dependent arrhythmias have been associated with a QT variability index in humans (Bilchick et al., 2004) and with triangulation, spatial dispersion and instability of the MAPD and reverse-use dependency of the drug in isolated rabbit hearts (Hondeghem et al., 2001a). Transmural dispersion of repolarisation in arterially perfused canine left ventricular wedge preparations also has predictive value in proarrhythmic drug testing (Di Diego et al., 2003; Antzelevitch et al., 2004). Short-term variability of repolarisation differs from other measures of temporal lability of repolarisation in its consecutiveness. The former evaluates the direct difference between two consecutive beats, while QT variability index and instability rearranges the order of beats.

Conclusions

Limiting peak plasma concentrations of NS-7 by decreasing the infusion rate reduces the proarrhythmic risk in dogs with remodelled hearts considerably. This occurs despite similar prolongations of the repolarisation parameters, like QT intervals. Conversely, an increase in short-term variability of MAPD predicted the proarrhythmic outcome.

Abbreviations

BVR

beat-to-beat variability of repolarisation

CAVB

chronic atrioventricular block

ΔMAPD

interventricular dispersion

LV

left ventricle

MAP

monophasic action potential

MAPD

monophasic action potential duration

QT_c

heart rate corrected QT interval

RV

right ventricle

SR

sinus rhythm

STV

short-term variability

TdP

torsades de pointes arrhythmias