In "post hoc" subgroup analyses, a simple classification system for patients, based on the presence or absence of findings indicative of electrical and/or mechanical complications early during short-term hospitalization, was applied to the data from the Beta-Blocker Heart Attack Trial (BHAT). In the largest subgroup of BHAT patients who had no reported complications, the 25 month mortality was low and the observed benefit of propranolol therapy small. Patients with electrical complications only had intermediate mortality and a pronounced effect of treatment was observed. Those with mechanical complications had the highest mortality and experienced an intermediate relative benefit of beta-blocker treatment. They also reported the most adverse effects. Post hoc analyses should always be interpreted cautiously. It is important to determine whether these findings are present in other completed beta-blocker trials. On the basis of these analyses alone it is suggested that the present practice of prescribing beta-blockers in postinfarction patients should not be altered.