In cardiac ventricular myocytes, an imbalance between inward and outward currents may prolong action potential duration. This makes the heart vulnerable to the occurrence of so-called torsade de pointes (TdP) arrhythmias, which are life-threatening ventricular tachycardias that create rapid fluctuations of QRS complexes around the isoelectric line on the human ECG. In patients, there are several independent risk factors or ‘challenges’ for TdP arrhythmias, including hypokalaemia, bradycardia, genetic and drug-induced long QT syndromes and chronic congestive heart failure (CHF) (Roden, 2004). A single ‘challenge’ to cardiac ventricular repolarization, for example, the reduction of a single membrane ion current, usually does not result in repolarization-dependent arrhythmias. Apparently, the heart has a reserve, commonly referred to as ‘repolarization reserve’ (Varro and Baczko, 2011), and multiple challenges are therefore usually required in order to provoke arrhythmia. Often, QT-prolonging drugs associated with TdP arrhythmias are the final challenge that exceeds the reserve, resulting in proarrhythmia. Quantification of the repolarization reserve, however, remains difficult. Although a number of surrogate parameters have been suggested (Thomsen et al., 2006), such as temporal or spatial dispersion of action potential duration, optimal quantification of repolarization reserve still requires testing of susceptibility to arrhythmias, where the cumulative severity of the challenges required to exceed the reserve then provides an estimation of the reserve. Interestingly, some drugs, including those that block the inward L-type calcium current (ICa-L), have been shown to be effective against drug-induced arrhythmias, by counteracting one or more of the predisposing challenges (Oros et al., 2010). Only a few experimental large animal models mimicking CHF have been developed. Currently, the efficacy in which ICa-L inhibition prevents or suppresses early after depolarizations (EAD) and polymorphic ventricular tachycardia in CHF is not clear and difficult to predict since calcium-handling disturbances are apparent in this disease (Janse, 2004). Moreover, in a setting of CHF, this apparent simple antiarrhythmic approach has to deal with conflicting imperatives such as antiarrhythmic action versus haemodynamic tolerance. In this issue of the BJP, Milberg et al. (2012a) report the outcome of ICa-L block by verapamil, a well established antiarrhythmic compound, on arrhythmic end points in a rabbit model of non-ischemic CHF with long-QT characteristics (Milberg et al., 2012a). CHF was generated by continuous right ventricular rapid pacing and, subsequently, Langendorff-perfused sham and CHF hearts were subjected to a number of additional challenges in order to provoke arrhythmias: bradycardia, ectopic ventricular activation, severe hypokalaemia and erythromycin-mediated IKr block. Repolarization was prolonged to some extent in CHF but spatial dispersion was not affected at baseline. Only after IKr block, especially transmural dispersion was increased to a larger extent in CHF. Arrhythmias were observed, but their number in hearts from sham animals (four of 11 hearts; 36%) was not significantly different from rabbit hearts with CHF (eight of 11; 73%; P = NS). Unfortunately, surrogate parameters were only reported for normokalemic circumstances, when the repolarization reserve was challenged less severely and thus could not directly be associated with the arrhythmic end point. Remarkably, the findings of the same group published recently (Frommeyer et al., 2011), in which the rabbit hearts were used to analyse the proarrhythmic effect of the IKr blocker sotalol were in favour of the CHF model used here. In this CHF group, sotalol induced EADs (as estimated from monophasic action potential morphology) and TdP in 16 of 18 (89%) hearts compared with seven of 14 (50%) hearts in the sham group. When we solely compare arrhythmia incidence based on these numbers, a P-value of 0.023 is obtained (two-tailed Fisher exact test). However, in both studies, and yet another [seven of 14 (50%) Milberg et al., 2012b], the pronounced incidence of arrhythmias in the sham hearts represents a potential limitation. Nevertheless, verapamil was demonstrated to be an efficient antiarrhythmic drug in this setting, and importantly, we may thus conclude that effectiveness of ICa-L block as antiarrhythmic treatment persists in an isolated rabbit heart model where CHF is added. The next hurdle will be to reach similar conclusions in an in vivo model where CHF is a more prominent proarrhythmic factor. Mechanisms of the antiarrhythmic potential of verapamil against repolarization-dependent arrhythmias have been ascribed to shortening of the QT interval and decreases in beat-to-beat variability of action potential duration (Oros et al., 2010; Bourgonje et al., 2011), and now Milberg et al. (2012a) show that it counteracts spatial dispersion in a CHF heart too. Promising as it seems, verapamil is contraindicated in CHF, especially in cases with severe systolic dysfunction and reduced fractional shortening (Chew et al., 1981). As verapamil inhibits the systolic calcium flux and, consequently, contractility, it is negatively inotropic, and this makes verapamil probably a poor choice in the clinic; certainly, when considering that the concentration used by Milberg et al. (0.75 µM) was unable to suppress arrhythmias completely. In the in vivo complete atrial-ventricular block dog model, verapamil plasma levels of around 0.5 µM clearly were antiarrhythmic but also lowered left ventricular pressure (Oros et al., 2010). Upon titrating verapamil, antiarrhythmic activity could not be observed without a drop in left ventricular pressure (Bourgonje et al., 2011). Other calcium channel antagonists might be a better option, however, and the authors themselves advocate second-generation ICa-L blockers. Take, for instance, nifedipine that more strongly affects smooth than striated muscle (Millard et al., 1983), where lowering peripheral resistance would compensate for negative inotropy. This may still have a major drawback because, in order to preserve blood pressure where contractility is reduced and vessels are dilated, heart rate must increase, which is also unfavourable for an already weakened heart. Obviously, it would be hard to predict the individual effects on vasodilatation and cardiac contractility, and where they would counterbalance each other in a haemodynamically challenged heart under neurohumoral influence. This should be approached experimentally. Furthermore, while inhibiting systolic calcium may be worrisome, in the case of diastolic dysfunction, calcium channel antagonism might be beneficial by improving coronary flow and muscle relaxation. As answering these questions is beyond the opportunities offered by the model of Milberg et al. (2012a), other models should be employed to address these intriguing possibilities. In conclusion, the study of Milberg et al. (2012a) demonstrates the efficacy of verapamil as an antiarrhythmic agent in the setting of CHF and provides basic science insights into its mechanism of action of reducing spatial dispersion. Further studies are required to pinpoint the contribution of CHF to arrhythmogenesis in this model, to recapitulate the findings in models where CHF is a more pronounced proarrhythmic challenge and to validate the antiarrhythmic efficacy and demonstrate clinical feasibility, of ICa-L block in in vivo models of CHF.