Directly measured peak aerobic capacity or oxygen uptake is a powerful predictor of prognosis in individuals with cardiovascular disease. Women enter phase 2 cardiac rehabilitation (CR) with lower and their response to training, compared with men, is equivocal. We analyzed at entry and exit in patients participating in CR and improvements by diagnosis to assess training response. We also identified sex differences that may influence change in . The cohort included consecutive patients enrolled in CR between January 1996 and December 2015 who performed entry exercise tolerance tests. Data collected included demographics, index diagnosis, , and exercise training response. The cohort consisted of 3925 patients (24% female). There was a significant interaction between baseline and diagnosis (P < .001), with percutaneous coronary intervention and myocardial infarction greater than other diagnoses. Surgical patients demonstrated greater improvement in than nonsurgical diagnoses (n = 1789; P < .001). Women had lower than men for all diagnoses (P < .02) and demonstrated less improvement (13 vs 17%, P < .001). Percent improvement using estimated metabolic equivalents of task (METs) were similar for women and men (33 vs 31%, P = NS). Despite overall increases in , 18% of patients (24% women, 16% men) failed to demonstrate any improvement (exit ≤ entry ). While there were no differences in training effect estimated by METs, directly measured showed a significantly lower training response for women despite adjusting for covariates. In addition, 18% of patients did not see any improvement in . Alternatives to traditional CR exercise programming need to be considered.