Purpose: We conducted a meta-analysis of individual studies reporting an association between serum vitamin D levels and AMD.
Methods: Relevant studies evaluating the association between serum vitamin D levels and AMD risk were identified by systematically searching four electronic literature databases (Ovid Medline, PubMed, EMBASE, and ISI Web of Science) censored by June 2015. Due to the heterogeneity of studies in categorizing serum vitamin D levels, all individual odds ratios (ORs) were recalculated and transferred for an increase of serum vitamin D levels by 10 ng/ml. Summary ORs and 95% confidence intervals (CIs) of AMD risk per 10-unit increase of serum vitamin D were obtained using standard meta-analysis. Publication bias was evaluated using funnel plots and Kendall's rank correlation tests.
Results: Ten individual studies were included and pooled in this meta-analysis. Meta-analysis of studies on AMD risk led to a pooled OR (95% CI) of 0.91 (0.69-1.22) for an increase of 25-hydroxy vitamin D by 10 ng/mL (P = 0.12). No indication for publication bias was found, but substantial heterogeneity was obtained (I2 = 79.7%, P < 0.01). Estimates from subgroup analyses also did not show statistically significant associations of serum vitamin D levels with different stages (early AMD, late AMD, and advanced AMD) and subtypes of AMD (neovascular AMD and nonneovascular AMD; P > 0.05).
Conclusions: There is no evidence to indicate an inverse association between serum vitamin D levels and any stages and subtypes of AMD risk, but opposite results from the United States and Korea resulted in this nonsignificance. Potential difference across various study designs might exist, based on few studies reporting in heterogeneous manners so far. More studies are needed to further confirm the causality of vitamin D and AMD, especially longitudinal studies.