Bariatric surgery improves glycemic control better than medical therapy; however, the effect of bariatric surgery on HDL function is not well characterized. Serum samples were available at baseline, 1-, and 5-years post procedures, for 90 patients with obesity and type 2 diabetes who were randomized to intensive medical therapy (n = 20), Roux-en-Y gastric bypass (RYGB, n = 37), or sleeve gastrectomy (SG, n = 33) as part of the STAMPEDE clinical trial. We examined serum HDL function by two independent assays, apolipoprotein A-1 exchange rate (AER) and cholesterol efflux capacity (CEC). Compared with baseline, AER was significantly higher at 5 years for participants in all treatment groups, but only increased significantly at 1 year in the RYGB and SG groups. CEC was divided into ABCA1-dependent and independent fractions, and the later was correlated with AER. ABCA1-independent CEC increased significantly only at 5 years in both surgical groups, but did not significantly change in the medical therapy group. There was no significant change in ABCA1-dependent CEC in any group. The increase in AER, but not ABCA1-independent CEC, was correlated with the reduction in body mass index and glycated hemoglobin levels among all subjects at 5 years, indicating that AER as a measure of HDL function would be a better reflection of therapy versus CEC.