In an observational study we wanted to investigate whether ongoing use of aspirin in a cohort of 753 patients with acute myocardial infarction was able to (1) reduce infarct size as assessed by peak creatine kinase and lactate dehydrogenase, (2) increase the number of non-Q-wave myocardial infarctions, and (3) to what extent thrombolytic treatment at admission could modify these outcomes. We used an exposed (aspirin+)/nonexposed (aspirin-) cohort design, adjusting for the effects of confounders (age, previous coronary heart disease, current smoking, and the prior use of beta-blockers and long-acting nitrates) as well as for the modifying effect of thrombolytic treatment. Crude and adjusted effects showed that aspirin reduced infarct size only in patients not receiving thrombolytic treatment at admission to hospital (n = 411 patients). In analyzing the occurrence of non-Q-wave versus Q-wave myocardial infarctions, the outcome was dichotomized. Crude odds ratio (OR) for developing a non-Q-wave myocardial infarction in aspirin users was 2. 63 (2p < 0.001), in the restricted cohort of patients receiving thrombolytic treatment, OR was 3.46 (2p = 0.002), whereas in those not receiving such treatment, OR was 1.81 (2p = 0.007). Adjusting for the effects of confounders, retained aspirin was an independent predictor of non-Q-wave myocardial infarctions, an effect that was probably increased (from 51 to 128%) in those who received thrombolytic treatment. Thus, aspirin seems to produce a shift to less severe manifestations of myocardial infarction, an effect that was increased in patients given thrombolytic treatment at admission to hospital.