Three to twelve (7.1 +/- 3.2) months after myocardial infarction, subjects under 54 (45.0 +/- 4.7) yr were assigned randomly to high-intensity (HIE, n = 37) or low-intensity (LIE, n = 42) exercise programs. Cardiac outputs (Q) during graded bicycle ergometer exercise were measured by a CO2-rebreathing method on entry and after 6 and 12 mo of training. The initial exercise Q was low in relation to work load, due to a low stroke volume (SV). Over the year of training, the predicted maximal O2 intake of the HIE group increased significantly (from 26.0 to 30.3 ml.kg-1.min-1), while that of the LIE group showed no significant alteration. During the first 6 mo, the heart rate of the HIE group was significantly reduced at each work level. There was an associated widening of arteriovenous oxygen difference, but SV was unchanged. These findings were attributed to extracardiac factors, including a redistribution of blood flow, biochemical changes in the trained muscle, and a secondary reduction of sympathetic drive. Over the second 6 mo, the SV of the HIE group increased 10%; this may reflect an increase of intrinsic myocardial contractility that develops if high-intensity training is sustained. The LIE group showed no major changes of cardiovascular function over the year of observation.