This randomized study was performed to assess the effects of > 3 hours of physical exercise per week and low-fat diet on collateral formation in nonselected patients with coronary artery disease (intervention group, n = 56). Results were compared with those of patients in a control group (n = 57), who received usual care by their private physicians. Coronary lesions were assessed by quantitative coronary angiography at the beginning and after 1 year of study (n = 92). As previously reported, after 1 year there was a significant retardation of progression of coronary artery disease in the intervention group as compared with the control group. In this study, evaluation of collateral formation revealed no significant difference between both groups, and changes in hemodynamic and metabolic variables or leisure time physical activity were not related to changes in collateral formation. Although progression of the disease was significantly related to an increase in collateral formation, regression was significantly related to a decrease in collateral formation (p < 0.00001). Because patients in the intervention group exercised for > 3 hours/week, and patients with regression of coronary artery disease even dedicated 5 to 6 hours to leisure time physical activity per week, these findings question whether an exercise program within the safety tolerance of patients will be able to induce coronary collateralization in the presence of regression of coronary artery disease.