Effect of folate supplementation on insulin sensitivity and type 2 diabetes: a meta-analysis of randomized controlled trials

Background: Various mechanisms link higher total homocysteine to higher insulin resistance (IR) and risk of type 2 diabetes (T2D). Folate supplementation is recognized as a way to lower homocysteine. However, randomized controlled trials (RCTs) show inconsistent results on IR and T2D outcomes. Objective: The aim of this study was to examine the effect of folate supplementation on IR and T2D outcomes. Design: We conducted a systematic literature search in PubMed, Web of Science, and EMBASE and prior systematic reviews and meta-analyses and identified 29 RCTs (22,250 participants) that assessed the effect of placebo-controlled folate supplementation alone or in combination with other B vitamins on fasting glucose, insulin, homeostasis model assessment for insulin resistance (HOMA-IR), glycated hemoglobin (HbA1c), or risk of T2D. The meta-analysis was conducted using both random-and fixed-effects models to calculate weighted mean differences (WMDs) or risk ratios with 95% CIs. Subgroup analyses were conducted based on intervention type (folate alone or in combination with other B vitamins), as well as analysis based on population characteristics, duration, dose, and change in homocysteine. Results: When compared with placebo, folate supplementation lowered fasting insulin (WMD: -13.47 pmol/L; 95% CI: -21.41, -5.53 pmol/L; P < 0.001) and HOMA-IR (WMD: -0.57 units; 95% CI: -0.76, -0.37 units; P < 0.0001), but no overall effects were observed for fasting glucose or HbA1c. Heterogeneity was low in all meta-analyses, and subgroup analysis showed no signs of effect modification except for change in homocysteine, with the most pronounced effects in trials with a change of >2.5 mu mol/L. Changes in homocysteine after folate supplementation correlated with changes in fasting glucose (beta = 0.07; 95% CI: 0.01, 0.14; P = 0.025) and HbA1c (beta = 0.46; 95% CI: 0.06, 0.85; P = 0.02). Only 2 studies examined folate supplementation on risk of T2D, and they found no change in RR (pooled RR: 0.91; 95% CI: 0.80, 1.04; P = 0.16). Conclusion: Folate supplementation might be beneficial for glucose homeostasis and lowering IR, but at present there are insufficient data to conclusively determine the effect on development of T2D. This trial was registered on the Prospero database as CRD42016048254.