Exercise Interventions and Glycemic Control in Patients With Diabetes

Abstract

To the Editor: The meta-analysis of exercise interventions in type 2 diabetes by Mr Umpierre and colleagues concluded that “aerobic, resistance, and combined training are each associated with HbA1c decreases.” However, we are not convinced that “the magnitude of this reduction is similar across the 3 exercise modalities.” Key factors likely to influence the magnitude of hemoglobin A1c (HbA1c) reduction achieved by a glucoselowering intervention include baseline HbA1c and use of an intention-to-treat analysis. When baseline HbA1c is higher, greater reductions are likely to be achieved. Six of the 20 aerobic exercise trials had baseline HbA1c greater than 8.5% vs none of the trials evaluating combined aerobic and resistance training. Six of the 20 aerobic exercise trials included more than 3 weekly exercise sessions vs none of the resistance exercise studies and only 1 combined exercise study. When intention-to-treat analyses are performed, smaller HbA1c reductions are likely to be observed than in analyses that exclude subjects who drop out. Among the aerobic exercise studies, only 2 of 20 trials used intention-to-treat analyses; these 2 trials included 132 of 490 subjects (27%) performing only aerobic exercise. Among trials evaluating combined exercise, 2 of 7 trials used intention-to-treat analyses; these included 140 of 261 participants (54%) performing combined exercise. Therefore, bias introduced by aerobic exercise–only studies having higher baseline HbA1c levels, greater exercise frequency, and proportionally fewer intention-to-treat analyses may have resulted in relative overestimation of the effect of aerobic exercise alone. The trials by Church et al and Sigal et al were the largest in the meta-analysis and were among the highest in methodological quality. Both used intention-to-treat analyses. Unlike most of the other studies, these trials included direct comparisons of combined aerobic and resistance training vs aerobic training alone. In both studies, combined training produced the strongest results. The Church et al study found that combined training, but not aerobic training, reduced HbA1c compared with controls, and combined training also resulted in greater decreases in body fat and increases in strength. The Sigal et al trial found that combined exercise resulted in a significantly greater decrease in HbA1c than aerobic training or resistance training alone (−0.97%, −0.51%, and −0.38% vs control, respectively). We think that the best evidence supports combined aerobic and resistance training as the most effective exercise regimen for improving glycemic control in people with type 2 diabetes.