To assess the effects of thrombolysis and reperfusion on late potentials after myocardial infarction, 101 patients (79 men, age 63.2 +/- 10.5 years) underwent signal-averaged ECG studies at 10.7 +/- 9.2 days, with the use of a 40 to 250 Hz band-pass filter. Patients were divided into four groups: (1) 54 patients treated with thrombolytic agents at 2.8 +/- 1.1 hours, with 81% "early" patency/reperfusion (TIMI grades 2 and 3); (2) 47 patients treated conventionally with 45% "late" patency/reperfusion; (3) 56 patients with patency (TIMI grades 2 and 3); and (4) 26 patients without patency (TIMI grades 0 and 1). A late potential was present when greater than or equal to 2 of 3 defined criteria were present. There was a significant difference in the incidence of late potentials between groups 1 and 2 (22% vs 43%, respectively; p = 0.048) and between groups 3 and 4 (18% vs 50%, respectively; p = 0.006). Late potentials also tended to occur less often after "early" than after "late" patency/reperfusion (12.5% vs 25%). The odds ratio for developing a late potential was 0.39 for thrombolysis versus no thrombolysis (p less than 0.05) and 0.22 for patency/reperfusion (TIMI grades 2 and 3) versus no patency/reperfusion (TIMI grades 0 and 1) (p less than 0.05). By analysis of covariance the effects of thrombolysis on late potentials were entirely explained by reperfusion. Thus the risk of late potentials after myocardial infarction is high but is reduced by thrombolysis and reperfusion. In addition, the effectiveness of "early" reperfusion appears to be greater than that of "late" but requires further clarification.