Background: Gain in VO2 peak after cardiac rehabilitation (CR) following an acute coronary syndrome (ACS), is associated with reduced mortality and morbidity. We have previously shown in CR, that gain in VO2 peak is reduced in Type 2 diabetic patients and that response to CR is impaired by hyperglycemia. Methods: We set up a prospective multicenter study (DARE) whose primary objective was to determine whether good glycemic control during CR may improve the gain in VO2 peak. Sixty four type 2 diabetic patients, referred to CR after a recent ACS, were randomized to insulin intensive therapy or a control group with continuation of the pre-CR antidiabetic treatment. The primary objective was to study the effect of glycemic control during CR on the improvement of peak VO2 by comparing first the 2 treatment groups (insulin intensive vs. control) and second, 2 pre-specified glycemic control groups according to the final fructosamine level (below and above the median). Results: At the end of the CR program, the gain in VO2 peak and the final fructosamine level (assessing glycemic level during CR) were not different between the 2 treatment groups. However, patients who had final fructosamine level below the median value, assessing good glycemic control during CR, showed significantly higher gain in VO2 peak (3.5 +/- 2.4 vs. 1.7 +/- 2.4 ml/kg/min, p = 0.014) and ventilatory threshold (2.7 +/- 2.5 vs. 1.2 +/- 1.9 ml/kg/min, p = 0.04) and a higher proportion of good CR-responders (relative gain in VO2 peak = 16 %): 66 % vs. 36 %, p = 0.011. In multivariate analysis, gain in VO2 peak was associated with final fructosamine level (p = 0.010) but not with age, gender, duration of diabetes, type of ACS, insulin treatment or basal fructosamine. Conclusions: The DARE study shows that, in type 2 diabetes, good glycemic control during CR is an independent factor associated with gain in VO2 peak. This emphasizes the need for good glycemic control in CR for type 2 diabetic patients.< Réduire