Using a community-based population of patients with acute myocardial infarction (AMI), we sought to: (1) determine the prevalence of bundle branch block (BBB) on the presenting electrocardiogram (ECG), (2) compare the clinical characteristics and the treatment administered to patients with and without BBB, and (3) determine the association of BBB with mortality. We analyzed the admission ECGs of 894 consecutive patients with AMI from Olmsted County, Minnesota, seen at our institution from January 1988 to March 1998. Of these, 53 had left BBB (LBBB) (5.9%) and 60 had right BBB (RBBB) (6.7%). Patients with BBB were more likely to be older, have a history of AMI or hypertension, and to be in Killip class >I at presentation. They were less likely to receive primary reperfusion therapy, beta blockers, or heparin, but more likely to receive angiotensin-converting enzyme inhibitors. They had lower mean predischarge ejection fractions (38 +/- 16% vs 50 +/- 15%, p <0.0001). In-hospital mortality was 13.3%, 17.0%, and 9.1% for patients with RBBB, LBBB, and no BBB, respectively (p = 0.11). Respective postdischarge survival at 1, 3, and 5 years was 80%, 60%, and 50% in the RBBB group, 78%, 56%, and 51% in the LBBB group, and 92%, 85%, and 76% in the group without BBB (p <0.0001). Although BBB was not an independent predictor of mortality on multivariate analysis, the presence of transient or persistent BBB with AMI is an easily recognized clinical marker of increased mortality. Our conclusion from this study is that in a community-based population, patients who had LBBB or RBBB at the time of AMI had lower predischarge ejection fractions and higher in-hospital and long-term unadjusted mortality.