In a prospective longitudinal study the occurrence of cardiac death was investigated and ongoing cardiac remodeling retrospectively analyzed in regard to adverse outcome. A cohort of 50 subjects with chronic Chagas' disease stratified according to Los Andes groups 1, 2 and 3 were followed-up for (mean+/-SD) 84.2+/-39.0 months. Follow-up was abbreviated by ventricular tachycardia ([VT] incidence: 3.0+/-7.0% year(-1)), first atrial fibrillation episode lasting >24 h ([AF] incidence: 3.3+/-1.0% year(-1)), nonfatal embolic stroke (incidence: 1.3+/-1.0% year(-1)), and cardiac death (mortality rate: 2.3+/-0.8% year(-1)). The relative risk of Los Andes group 3, VT, AF and stroke for cardiac death was, respectively, 25.3 (95%CI [3.5-182.6]), 3.0 (95%CI [1.2-7.3]), 3.6 (95%CI [1.2-10.9]) and 1.1 (95%CI [0.2-7.2]). In a multivariate Cox proportional-hazard model, Los Andes group 3 (hazard ratio=24.5; 95%CI [3.2-189.2]; p<0.01) was independent predictor for cardiac death. LAD and LV mass and not LVEF variation rates differed among group 1 (respectively, 0.03+/-0.1 cm year(-1); 2.4+/-5.7 g year(-1) and -0.1+/-2.4% year(-1)), group 2 (0.04+/-0.1 cm year(-1); 3.7+/-8.8 g year(-1) and -0.8+/-1.4% year(-1)) and group 3 (0.13+/-0.1 cm year(-1), p<0.001; 21.7+/-10.1 g year(-1), p<0.001 and -1.4+/-2.5% year(-1), p=0.26). Variables on admission were linearly related to respectively variation rates (r=0.71; p=0.02) composing a first order linear process with 0.07 months(-1) time constant. In chronic Chagas' disease, initial clinical status is an independent predictor for cardiac death and determines the progression rate of myocardial damage.