Controversy exists as to the homogeneity of repolarization throughout the canine ventricular wall in vivo. The type of anesthesia has been shown to affect regional differences in monophasic action potential duration and the inducibility of polymorphic ventricular tachycardias (PVTs) in normal canine hearts. This study was conducted to determine refractory patterns and arrhythmia susceptibility in relation to halothane or pentobarbital anesthesia in dogs with chronic atrioventricular block and biventricular hypertrophy. In 12 dogs with chronic atrioventricular block, 60 needle electrodes (12 mm long, four bipolar electrodes, interelectrode distance of 2 mm) were inserted into the left and right ventricle. Six dogs were anesthetized with pentobarbital and six with halothane. Effective refractory periods (ERPs) were determined along 14 randomly selected needles at baseline and after application of almokalant (0.34 mmol/kg) (basic cycle length 1,000 ms, extrastimulus technique). At baseline and on almokalant, ERPs were uniform, independent of the type of anesthesia. With halothane anesthesia, ERPs were significantly longer under both conditions. Almokalant induced not only a prolongation of ERP in both groups but also a significant increase in transmural dispersion of ERP and in maximum dispersion of ERP. However, local refractory gradients were not specific to any muscle layer and did not seem to be related to the occurrence of PVTs. Almokalant did not induce arrhythmias in any dog in the pentobarbital group, but in four of six animals in the halothane group, apparently due to the more marked prolongation in ERP. Independent of the type of anesthesia, hypertrophied hearts of dogs with chronic atrioventricular block exhibit uniform refractory patterns. Longer ERPs with a comparable degree of dispersion on halothane are associated with a high incidence of drug-induced PVTs, whereas shorter ERPs on pentobarbital seem to prevent arrhythmia induction.