Remodeling of infarcted and noninfarcted ventricular regions, infarct expansion, shape distortion, and global left ventricular (LV) dilation influence LV performance and survival. The effect of chronic exercise, initiated early or later after infarction, on remodeling, hemodynamics, and survival has not been studied. A total of 156 rats were randomized after coronary artery occlusion or sham operation to remain sedentary or to start with swim training 4 days or 21 days after coronary occlusion, which was continued over 8 weeks (6 days per week, 90 minutes per day). These intervals after coronary artery ligation were chosen because final size of infarction is well reached after 4 days; histological evolution of scar healing is still in progress, and after 21 days, histological scar healing is completed. At 8 weeks, hemodynamics were measured and LV dilation quantitated by passive pressure-volume curves. In groups with small (< or = 35%) and large (> 35%) infarcts, the area enclosed by endocardial circumference, infarct size, LV diameter, scar thickness, and septal thickness were measured in stained transverse serial LV sections to assess aneurysmal shape distortion and the response of infarcted and noninfarcted myocardia. Survival was not influenced by infarction or exercise alone. In rats with small infarcts, LV volume and shape and long-term survival were not altered by chronic exercise initiated early or late after coronary artery ligation. Mortality rose in animals with large infarction as a result of exercise (P < .0001) and was 47.6% with early exercise and 26.7% with late exercise (P < .05, early versus late). Infarct size in rats with early exercise (48 +/- 3%) was similar compared with infarct size of rats with late exercise (46 +/- 2%, P = NS compared with early exercise). Exercise did not affect LV dysfunction (assessed by systolic and end-diastolic pressures and dP/dtmax) in survivors of small and large infarctions. LV volumes increased (P < .05) in sedentary rats by large infarction (n = 13, 0.48 +/- 0.04 mL) compared with volumes after sham operation (n = 33, 0.33 +/- 0.03 mL) and with exercise (early, n = 11, 0.56 +/- 0.04 mL; late, n = 11, 0.65 +/- 0.04; P < .05 versus sedentary). In nonsurvivors from early exercise, the area enclosed by LV endocardial circumference (which corresponds to LV volume) was increased by 195%, LV diameter was increased by 60%, and scar thickness was reduced by 37% (P < .05 versus respective control). Septal thickness increased in survivors by exercise (+25%) but decreased (-28.6%) in nonsurvivors (P < .0001 versus respective control). Endurance training in rats after small infarction, whether started early or late after left coronary artery ligation, was well tolerated without changes in LV volume, shape, hemodynamics, and long-term survival. Endurance training in rats with large infarction decreased overall survival (P < .0001). In survivors from late exercise, training caused aggravation of global LV dilation without additional shape changes. Endurance training after large infarction caused aggravation of remodeling to a degree that was not compatible with life in 27% of the rats with late exercise and in 48% with early exercise after coronary artery ligation, despite similar exercise. This was explained by extensive remodeling that was most pronounced in nonsurvivors from early exercise. In these rats, severe global LV dilation, distortion of LV shape, scar thinning, and a paradoxic reduction of septal thickness, ie, mismatch of infarcted and noninfarcted myocardia, were observed.