In a prospective study, the influence of the length of the time interval on spontaneous variability was investigated in 100 patients with CAD or IDC and untreated ventricular arrhythmia of Lown grade IV. Patient follow-up was carried out over 260 +/- 387 days. In each of the 498 ambulatory Holter tapes, the mean hourly arrhythmia count (AC) of couplets and salvos was verified. The variability of ACs between two Holter ECGs was defined as the logarithm of the quotient AC day 2(n + 0.01)/AC day 1(n + 0.01). The spontaneous distribution of variability quotients (means +/- 2 SD) was defined separately for couplets and salvos and for each of four ranges of control intervals (0-6 days, 7-89 days, 90-364 days, greater than or equal to 365 days). The percentage change in arrhythmia count necessary to establish drug efficacy (R), was calculated according to the formula R(%) = (10(0) - 10(-2SD].100, whereas the percentage change necessary to prove aggravation of arrhythmia (A) was assessed by the formula A(%) = (10(0) + 10(+2SD].100. For couplets, R extended from 90%, 94%, 98% to 99%; A increased from 1114%, 1895%, 6153% to 14032%, respectively. For salvos, R remained almost unchanged at a high level with 95%, 98%, 98%, 99%. The figures of A were 2189%, 4650%, 5698% and 9650%, respectively. It is concluded that the spontaneous variability of complex ventricular arrhythmias is remarkably high with short control intervals and increases further with longer ones.(ABSTRACT TRUNCATED AT 250 WORDS)