Meta-analyses showed that exercise training decreased mortality in patients after myocardial infarction, but no single adequately powered trial has shown this benefit. The purpose of this study was to evaluate the effect of cardiac rehabilitative exercise training on prognostic variables derived from the exercise tolerance test and whether these changes had an effect on predicted cardiovascular and all-cause mortality using externally validated risk scores. Two hundred ten consecutive patients who completed a 12-week cardiac rehabilitation program were analyzed. Peak MET level, heart rate recovery, chronotropic index, Duke prognostic score, and Cleveland Clinic Foundation risk score obtained at program entry and exit were compared. All prognostic variables and risk score–derived predicted 5-year mortality rates improved significantly after cardiac rehabilitation. Exercise capacity increased by 32% (6.6 ± 2.7 to 8.7 ± 2.9 METS; p <0.0001), heart rate recovery increased by 20% (15 ± 9 to 18 ± 10 beats/min; p <0.0001), and chronotropic index increased by 11% (56 ± 22% to 62 ± 22%; p <0.0001). Duke prognostic score 5-year predicted cardiovascular mortality rate decreased by 33% (6 ± 4% to 4 ± 2%; p <0.0001), and Cleveland Clinic Foundation risk score 5-year predicted total mortality rate decreased by 40% (5 ± 7% to 3 ± 5%; p <0.0001). These improvements remained significant in high-risk subgroups. When controlled for age, gender, diabetes, MET level achieved, ejection fraction, and β-blocker use, changes remained significant (p <0.0001) for each outcome measure. In conclusion, exercise training in supervised cardiac rehabilitation programs significantly improved exercise capacity, measures of chronotropic dynamics, and prognostic risk scores. These differences translate into a decrease in predicted mortality.