Abstract Background The EU-CaRE project aims to obtain the evidence base to improve, tailor and optimise cardiac rehabilitation (CR) programmes regarding sustainable effectiveness, cost-effectiveness and participation level in the elderly. Purpose To compare the effect of CR programmes in 7 European countries on the main outcome. Methods The observational EU-CaRE study is a prospective study with eight participating CR sites in seven countries (Denmark, France, Germany, Italy, the Netherlands, Spain and Switzerland). Consecutive patients age 65 or older with coronary heart disease or valve surgery undergoing CR were included. Data were obtained at baseline (T0), at the end of CR (T1) and at 1 year (T2). VO2peak was assessed by cardiopulmonary exercise test. We compared changes in VO2peak across sites in multilevel regression models for repeated measures, adjusting for age, gender, baseline VO2peak, comorbidities, CVD risk factors and educational attainment. Results are reported separately for patients undergoing surgery (CABG/valve replacement) and other indications (MI/stable CAD/PCI). Results 1633 patients were included, and VO2peak was available at all three timepoints for 1243 patients (76%), 492 undergoing surgery and 751 for other indications. Lag time from index event to first CPET and baseline VO2peak varied significantly across centres. Mean age was 72.9 (range 65–90), mean VO2peak at T0 was 16.49 (SD 4.94), at T1, 18.55 (5.33) and at T2 19.03 (5.43) ml/kg/min. For patients not undergoing surgery mean improvement T0-T1 was 1.57 (2.98) and T0-T2 was 1.40 (3.17) ml/kg/min (both p<0.001). In the surgery patients as expected baseline VO2peak was lower but in contrast to non-surgery continued to improve after CR: T0-T1 improvement was 2.77 (2.76) and T0-T2 improvement was 4.29 (3.83) (both p<0.001). At one-year follow-up surgery and non-surgery patients had similar VO2peak (p=0.59). The greatest mean improvement in a site from baseline to one year was 2.04 (3.74) and the smallest 0.59 (2.54) ml/kg/min for non-surgery and, correspondingly 6.13 (4.83) and 1.29 (2.06) ml/kg/min for surgery patients (both multivariable adjusted p for differences between sites p<0.001). In addition to age and gender, baseline and one-year VO2peak was lower in patients with diabetes, hypertension, smokers, patients with other co-morbidity, lower educational attainment and lower mental component score of the SF36. These factors had no systematic impact on the effect of CR. VO2peak at baseline, after CR and 1 year Conclusions The study provides high-quality data on the effectiveness of current CR programmes in Western Europe. Overall, CR was effective in achieving significant improvements in exercise and maintaining these effects. There were significant differences in effect between sites indicating a potential for improvement by adapting CR programs. Acknowledgement/Funding Horizon 2020