Effect of Vitamin D Supplementation on Glucose Metabolism Biomarkers—A Meta-analysis of Randomized Controlled Trials

Observational studies have linked higher serum concentrations of 25(OH)D with a reduced risk of diabetes. In order to assess the effect of vitaminD supplementation (VDS) on glucose metabolism biomarkers, a systematic review with meta-analysis of randomized-controlled trials (RCTs) was undertaken. RCTs published until February 2017 conducted in adults were selected if they examined the effect of VDS on glycosylated haemoglobin (HbA1c), fasting glucose (FG), fasting insulin (FI), or HOMA-IR. For each biomarker, summary mean differences (MDs) between intervention and placebo groups were computed using a random-effect model. A total of 48 (n=5,100 subjects), 80 (n=9,565), 58 (n=7,841), and 57 (n=7,763) RCTs were selected for the analyses on HbA1c, FG, FI and HOMA-IR, respectively, with a majority of studies conducted in Europe and the Middle-East. The summary MDs (95% confidence intervals) were: -0.06% (−0.12, 0.00) for HbA1c, -0.12 mmol/L (-0.20, -0.05) for FG, -8.53 pmol/L (-13.17, -3.90) for FI, and −0.35 (-0.54, -0.15) for HOMA-IR. High heterogeneity between studies was observed for all biomarkers (I² > 70%) and no evidence for publication bias was noted. Important differences appeared between countries. In the Middle-East, the results were -0.25 (−0.42, -0.08) for FG, -17.74 (-26.36, -9.11) for FI and -0.71 (-1.09, -0.32) for HOMA-IR, with high heterogeneity (I²=76%, 79%, 79% respectively). In Europe, the results were 0.01 (-0.06, 0.07) for FG, −0.59 (-6.86, 5.68) for FI and -0.04 (-0.19, 0.10) for HOMA-IR, with lower heterogeneity (I²=21%, 48%, 0% respectively). In studies with high dose of vitamin D (≥ 100 µg/d) and subjects with baseline 25(OH)D ≤ 20 ng/mL, the summary MDs for HB1Ac and FG were -0.09% (-0.24, 0.07) and -0.mmol/l (-0.16, 0.06). The influence of VDS on glucose metabolism biomarkers seems to be marginal. The reasons underlying the heterogeneity of results across countries is unknown but differences in the conduct of RCTs cannot be excluded. Disclosure C. Pizot: None. P. Mullie: None. A. Macacu: None. M. Dragomir: None. P. Boyle: Other Relationship; Self; Sanofi. P. Autier: Other Relationship; Self; Sanofi.