OBJECTIVE To determine the prevalence and the prognostic value of exercise-induced ventricular arrhythmia (EIVA) in chronic Chagas' heart disease. STUDY DESIGN AND SETTING An open prospective cohort of 130 clinically stable patients at a University Hospital outpatient unit in Rio de Janeiro, Brazil, was followed up at scheduled clinical visits from 1990 through 2007. The endpoint was total cardiovascular mortality. Survival curves (Kaplan-Meier) and a multivariate Cox proportional hazard model were adjusted to determine the association between EIVA and mortality. RESULTS The median duration of follow-up was 9.9 years (range, 132 days to 17 years). EIVA prevalence was 43.1% (95% CI: 34.5-51.7). Thirty-three cardiovascular deaths (25.4%) occurred. The hazard ratio of EIVA for cardiovascular death, after adjustment for age, was 1.84 (P = 0.09). An interaction was found between EIVA and cardiomegaly on x-ray. In the group with cardiomegaly, the hazard of dying was four times greater in the presence of EIVA (P for interaction = 0.05). CONCLUSION In clinically stable chagasic subjects with cardiomegaly, EIVA is a clinically significant marker of total cardiovascular mortality and may be a useful risk stratification tool in this population.