Experimental models of torsades de pointes.

Torsades de pointes is the most typical ventricular tachycardia involving QT-interval prolongation. It is a rather unusual but potentially lethal ventricular tachycardia with a distinctive morphology favored by bradycardia, antiarrhythmic drugs and hypokalemia and requires specific treatment. Torsades de pointes has been shown to be related to bradycardia-dependent early afterdepolarizations (EAD) and/or increased dispersion of repolarization. However, although EAD can be obtained relatively easily in vitro with quinidine or sotalol, torsades de pointes are very difficult to reproduce in animal models. The models of torsades de pointes which have been proposed can be categorized as morphological, EAD-related or pharmacological models. The purpose of the 'morphological' models was to reproduce the twisting of QRS axis typical of torsades de pointes, with no consideration of other aspects such as long QT or bradycardia. These models were produced by epicardial electrical or chemical (aconitine) stimulation at two distant ventricular sites or by overdosing of quinidine in dogs with acute myocardial infarction. The second type of model focused on the conditions producing EAD in vitro. Ventricular tachycardias were obtained in anesthetized dogs using toxics such as cesium or anthopleurine, both producing EAD in vitro. These ventricular tachycardias were shown to be sensitive to magnesium, heart rate and autonomic tone, but torsades de pointes remained rare, at least after cesium injections. The pharmacological models that could be used to study the QT-dependent proarrhythmic effects of drugs are the anesthetized rabbit with alpha-adrenergic stimulation, and the conscious dog model with chronic AV-block and diuretic-induced hypokalemia. Methoxamine-treated anesthetized rabbits develop ventricular tachycardias during clofilium infusions. These ventricular tachycardias, although appearing at very high heart rates, have typical torsades de pointes aspects and are often associated with giant QT waves. The specificity of the model remains to be tested. In our conscious bradycardic and hypokalemic dogs, quinidine and sotalol but not flecainide, propranolol or lidocaine induced QT-dependent arrhythmogenic effects and torsades de pointes. Efficacy of high rate stimulations and magnesium were repeatedly observed. This demanding model, especially designed for qualitative drug comparisons, is also well suited to studies on the mechanisms of initiation of torsades de pointes. The pertinence of these models for estimating the risk of QT-dependent proarrhythmias associated with non-antiarrhythmic agents remains to be tested.