BACKGROUND Although programmed electrical stimulation is widely used for provoking sustained ventricular tachycardia (VT), the mechanism by which repetitive extrastimulation evokes VT is still little understood. Specifically, it is not clear why several closely coupled extrastimuli are frequently required to induce VT. Although regularly paced human ventricular myocardium exhibits a near constant relation between myocardial repolarization and refractoriness, the effect of repetitive extrastimulation on the relation between repolarization and excitability in the human heart and its relevance for arrhythmia induction by programmed stimulation are unknown. We hypothesized that the induction of VT by repetitive extrastimulation is facilitated by an altered relation between repolarization and refractoriness, and this leads to disturbances in ventricular impulse propagation, which trigger the onset of VT. RESULTS Twenty-one patients undergoing routine electrophysiological study were paced from the right ventricular apex and outflow tract endocardium with monophasic action potential-pacing catheters placed at both sites simultaneously Monophasic action potential durations (APDs) and effective refractory periods (ERPs) were measured simultaneously at each site, during regular stimulation (S1-S1) at 400-ms cycle length and during three consecutive extrastimuli (S2 through S4) at the closest coupling intervals at which all three extrastimuli still resulted in capture. Measurements further included the repolarization level at which the earliest capture occurred, the ratio between ERP and APD, and the propagation time between the pacing and distant recording site. APD and ERP both shortened progressively with each extrastimulus. APD at 90% repolarization decreased from a baseline (S1) of 238.1 +/- 19.7 ms by 14.9% at S2, 18.9% at S3, and 22.9% at S4 (P < .0001, S1 versus S4). ERP decreased from 233.1 +/- 19.7 ms (S1) to 180.0 +/- 41.9 ms (S3) (P < .0001, S1 versus S3). While ERP shortening occurred mainly on the basis of APD shortening, there was an additional factor that contributed to ERP shortening independent of APD shortening. Each consecutive extrastimulus was able to elicit a propagated response at earlier repolarization levels than the previous one: the earliest capture for S2 occurred at 85.5 +/- 10.2% of complete repolarization, for S3 at 83.9 +/- 10.5%, and for S4 at 78.4 +/- 11.2% (P < .05 for S2 versus S3; P < .05 for S3 versus S4; P < .01 for S2 versus S4). This progressive "encroachment" of the earliest capture stimulus onto the preceding repolarization phase (at progressively less repolarized levels) correlated with a progressive delay of impulse propagation between the pacing site and the second recording site: propagation time increased from baseline (S1) by 10.5 +/- 1.3% with S2 to 19.0 +/- 1.6% with S3 and to 22.5 +/- 2.8% with S4 (P < .05, S4 versus S1). VT was induced in 11 of 21 patients. Nine of these had VT induced only when significant encroachment of extrastimuli on the preceding repolarization phase (< 81.3 +/- 7.0%) and associated conduction slowing (> 16.6 +/- 1.8%) were present. CONCLUSIONS Repetitive extrastimulation not only shortens APD and subsequently ERP but also alters the ERP/APD relation by allowing capture to occur at progressively less complete repolarization levels. This progressive encroachment onto the preceding repolarization phase is associated with impaired impulse propagation and a high incidence of VT induction. This may help explain how repetitive, closely coupled extrastimulation induces ventricular tachycardia in the human heart.