The effect of vitamin D supplementation on glucose metabolism in type 2 diabetes mellitus: A systematic review and meta-analysis of intervention studies.

Design: randomized controlled trials (open-label and blinded); some cross-over designs with initial-arm data used for meta-analysis; participants: adults with type 2 diabetes; 29 trials totaling 3,324 participants; conducted across multiple regions (Iran, Europe, Asia, Australia, UAE, Africa, South America, and multinational); baseline vitamin D status varied, with several trials enrolling Vitamin D–deficient individuals; co-interventions allowed in some studies.

Oral vitamin D supplementation (mostly vitamin D3, cholecalciferol); regimens varied across trials: daily, weekly, or single bolus dosing; doses ranged from 400 IU daily to a 450,000 IU single dose; treatment duration ranged from 8 weeks to 5.5 years (median ~12 weeks); some regimens included concomitant calcium supplementation or calcium/vitamin D–fortified foods; one trial used vitamin D2 (ergocalciferol).

Vitamin D supplementation produced a modest reduction in HbA1c by 0.32 percentage points versus placebo (95% CI −0.53 to −0.10) with substantial heterogeneity (I2 = 91.9%); no overall effect on fasting glucose (−2.33 mg/dL, 95% CI −6.62 to 1.95; I2 = 59.2%). In studies where baseline vitamin D deficiency was corrected (n=7), HbA1c reductions were larger (−0.45%, 95% CI −1.09 to 0.20) and fasting glucose reductions tended to be greater (−7.64 mg/dL, 95% CI −16.25 to 0.97) but not statistically significant. Some regimens with lower vitamin D doses or shorter treatment lengths showed more favorable HbA1c effects; overall, evidence is heterogeneous. Conclusion: Vitamin D may offer a modest HbA1c improvement in adults with type 2 diabetes, but does not reliably improve fasting glucose; benefits may be more evident when vitamin D deficiency is corrected, though larger, high-quality trials are needed, especially in deficient populations.

Substantial heterogeneity across studies (HbA1c I2 ≈ 92%); many trials of poor quality; short and variable treatment durations; wide range of vitamin D doses and regimens; potential confounding from diabetes medication changes; limited generalizability to individuals with normal glucose metabolism; inconsistencies in vitamin D measurement methods; possible publication bias; co-supplementation with calcium in some trials may confound results.