Ventricular tachyarrhythmia initiation was compared using unipolar cathodal, anodal and bipolar programmed stimulation at 21 sites in 5 normal adult mongrel dogs and 67 noninfarct sites in 16 dogs 3-5 days after experimental myocardial infarction. For this purpose, the minimum number of extrastimuli required for tachyarrhythmia initiation was determined in each pacing mode using twice cathodal threshold current for the drive beats and all extrastimuli except the last. The current and pacing mode were varied for the last extrastimulus (S2, S3 or S4). In the 5 normal dogs, ventricular fibrillation was reproducibly inducible from only 1/21 sites, and only in the cathodal mode. In 15/16 (94%) of the myocardial infarction dogs, a sustained ventricular tachycardia or ventricular fibrillation could be reproducibly initiated with either one (4 dogs), two (5 dogs) or three extrastimuli (6 dogs). Diastolic excitability thresholds were 0.08 ± 0.03, 0.30 ± 0.17, and 0.09 ± 0.04 mA (median ± SD) for unipolar cathodal, anodal and bipolar pacing, respectively (p < 0.001 for anodal vs. cathodal and bipolar). The median absolute current required for ventricular tachyarrhythmia initiation was also highest with anodal pacing (0.72 ± 0.77 mA), versus both the cathodal and anodal modes (0.18 ± 0.28 and 0.20 ± 0.28 mA, respectively, each p < 0.001) but was comparable in all three modes relative to the threshold current (2.0, 2.4 and 2.6 mA for cathodal, anodal and bipolar pacing, respectively) required for initiation. Overall, ventricular tachyarrhythmia initiation was concordant in all three modes at 58/67 (87%) sites and discordant at only 9/67 (13%) sites (p < 0.001). Moreover, there was no difference in either the pattern of arrhythmia initiated in each of the pacing modes with