Patients with acute myocardial infarction (MI) with bundle branch block (BBB) have a poor prognosis, but distinction between left (L)- and right (R)-sided BBB is seldom made in epidemiological studies. We studied long-term mortality associated with RBBB and LBBB in the TRAndolapril Cardiac Evaluation (TRACE) study. TRACE screened consecutive patients presenting with an MI and recorded clinical, electro- and echo-cardiographic variables. Subsequently, deaths were recorded during a minimum follow-up of 15 years. In total, 6676 consecutive patients with MI were hospitalized at 27 centres in Denmark. Of these, 533 (8%) had BBB, of whom 260 (4%) had RBBB and 273 (4%) had LBBB. Overall, 5196 (78%) patients died, 256 (94%) with LBBB and 235 (90%) with RBBB compared with 4705 (77%) of those without BBB (P < 0.001). In multivariable analyses, hazard ratios (HRs) of RBBB and LBBB were 1.23 [95% confidence interval (CI), 1.07-1.42] and 1.05 (95% CI, 0.91-1.20), respectively. There was interaction between each type of BBB and left ventricular (LV) systolic function (P = 0.02). Right BBB was associated with a worse prognosis in patients with reduced LV systolic function [HR = 1.31 with wall motion index (WMI) ≤ 1.5 (95% CI, 1.11-1.55] while LBBB had a poor prognosis in patients with preserved LV systolic (HR if WMI > 1.5, 1.70; 95% CI, 1.12-2.57). Right BBB was a predictor of increased mortality in patients with reduced LV systolic function, whereas LBBB was a marker of increased mortality in patients with preserved LV systolic function.