Effects of exercise, metformin, and their combination on glucose metabolism in individuals with impaired glycemic control: A systematic review and network meta-analysis

ObjectiveTo compare the efficacy of exercise, metformin, and their combination on glucose metabolism in individuals with prediabetes and type 2 diabetes mellitus (T2DM), and rank these treatments by exercise modality and metformin dosage. DesignSystematic review and network meta-analysis. Data sourcesEmbase, Web of Science, PubMed/MEDLINE, and SPORTDiscus were searched until February 2023. Eligibility criteria for selecting studiesRandomized controlled trials (RCTs) of exercise, metformin, or their combined treatments in individuals with prediabetes or T2DM were included. AnalysesWe estimated pooled mean difference (MD) with 95% confidence intervals (CI) for five glucose metabolism variables (i.e., hemoglobin A1c [HbA1c], 2-h glucose during oral glucose tolerance test [OGTT], fasting glucose, fasting insulin, and homeostasis model assessment of insulin resistance [HOMA-IR]) via a random effect model. Subgroup analyses were conducted for prediabetes and T2DM populations. ResultsWe identified 15,872 eligible articles and finally included 375 articles with 378 RCTs, comprising 30,884 participants. When all individuals were pooled, metformin had greater effects than exercise in improving HbA1c (MD -0.65 95% CI [-0.77 to -0.53] vs -0.38 [-0.45 to -0.32] %), 2-h glucose during OGTT (-1.16 [-1.67 to -0.65] vs -0.76 [-1.15 to -0.37]), fasting glucose (-0.99 [-1.14 to -0.83] vs -0.57 [-0.65 to -0.48] mmol/L), and fasting insulin (-2.28 [-3.03 to -1.53] vs -1.47 [-1.85 to -1.09] U/mL), but not in HOMA-IR (-0.36 [-0.77 to 0.04] vs -0.73 [-0.90 to -0.57]). A similar pattern was observed for the subgroup of T2DM patients; however, in prediabetes, exercise showed better efficacy than metformin in HbA1c (-0.17 [-0.23 to -0.11] vs -0.09 [-0.20 to 0.01] %) and 2-h glucose (-0.68 [-1.01 to -0.35] vs -0.04 [-0.51 to 0.43] mmol/L). Exercise + metformin showed a greater efficacy than exercise alone in improving HbA1c (-0.79 [-1.27 to -0.30] %) and fasting glucose (-0.76 [-1.25 to -0.26] mmol/L) when all individuals were pooled. Considering exercise modalities, aerobic interval exercise was most efficient in improving fasting glucose (-0.82 [-1.06 to -0.59] mmol/L), HbA1c (-0.61 [-0.77 to -0.44] %), fasting insulin (-2.22 [-3.34 to -1.10] U/mL), and HOMA-IR (-0.95 [-1.39 to -0.51]). The confidence in evidence was mostly low or very low. ConclusionThe use of exercise, metformin, and their combination are all effective in improving glucose metabolism in individuals with impaired glycemic control, such as prediabetes or T2DM, but the efficacy varies in the five outcome parameters (i.e., HbA1c, OGTT 2-h glucose, fasting glucose, fasting insulin, and HOMA-IR). The efficacy was modified by exercise modality, metformin dosage, and severity of impaired glycemic control. Future clinical trials may further investigate the specific components of the interactive effects of exercise and metformin, such as the timing of exercise and metformin administration, the drug delivery method as well as the effect of additional exercise variables (e.g., exercise frequency and volume). PROSPERO registration number: CRD42023400622.