Serum 25-hydoxyvitamin D concentrations in relation to Hashimoto's thyroiditis: a systematic review, meta-analysis and meta-regression of observational studies

Purpose Available evidence on the relation between vitamin D status and Hashimoto's thyroiditis (HT) remains inconsistent. We conducted a meta-analysis of serum 25-hydoxyvitamin [25(OH)D] concentrations in HT, and examined how the strength of this relationship varies as a function of several moderating factors. Methods Twenty-six observational, case-control studies, published before Feb 20, 2018, were located using Google Scholar, PubMed, Web of Science, SCOPUS, LILACS and SCIELO. Study quality was assessed and random-effects models were used, along with univariate mixed-effect meta-regression, for all analyses. Results The 25 studies (2695 cases, 2263 controls) confirmed lower serum 25(OH)D concentrations in HT compared to healthy controls, with Cohen's d - 0.62 (95% CI - 0.89, - 0.34; P = 1.5 x 10(-5)) and substantial heterogeneity between studies. HT showed an odds ratio (OR) of 3.21 (1.94-5.3; P = 5.7 x 10(-6)) for 25(OH)D deficiency (cut-off 20 ng/mL) against healthy controls. A corrected Cohen's d of - 0.43 [(- 0.76, - 0.09), P = 0.013] was obtained by trim-and-fill adjustment for publication bias. The association was consistent across Asian and European studies, pediatric and adult population, high- and moderate-quality studies. Near-equatorial latitudes (< 35 degrees N/S, P = 3.4 x 10(-4)) and moderate-income economy (gross national income (GNI) 1000 < US$ < 12,000, P = 0.012) were associated with more discrepant 25(OH)D concentrations between the groups. Higher latitude (P = 0.0047), and higher mean body mass index (P = 0.006, 10 studies) were associated with smaller Cohen's d by univariate meta-regression, with evidence of nonlinear moderation by GNI (P = 3.5 x 10(-6)), and mean serum thyrotropin in affected individuals (P = 0.017, 21 studies). Conclusion The present work shows a significant association between circulating 25(OH)D and HT, partly resolves mixed findings by identifying the empirical moderators contributing to overall heterogeneity, and highlights HT patient groups and the conditions under which the association is strongest.