Using a new method for long-term recording of monophasic action potentials from the human heart, we studied in 17 patients the effects on ventricular action potential duration (APD) of three clinically pertinent cycle length perturbations: (1) single extrastimuli, (2) abrupt sustained rate acceleration and deceleration, and (3) different steady-state cycle lengths. Results were: (a) APD after single extrastimuli at progressively longer cycle lengths were related to the extrastimulus cycle length with a biphasic electrical restitution curve which after an initial steep rise and a subsequent transient descent rose again more gradually to a plateau at cycle lengths above 800-1,000 ms. (b) After a sustained step decrease in cycle length, the first APD shortened abruptly while final steady-state adaptation required up to several minutes. The transition between the rapid and slow phase of APD change was characterized by a variable alternans of APD which correlated inversely with the preceding diastolic interval. (c) In the steady state, APD correlated linearly with cycle length, increasing an average of 23 ms per 100 ms cycle length increase (r = 0.995). The divergence between steady-state and non-steady-state APD, and the slowness of steady-state adaptation, are important factors to be considered in clinical electrophysiologic studies and in rate correction algorithms of APD or QT intervals, respectively.
