Serum 25-hydroxy vitamin D levels in relation to body mass index: a systematic review and meta-analysis

Background Although several cross-sectional studies have investigated serum vitamin D levels in relation to body mass index (BMI), findings are inconsistent. Objective This systematic review and meta-analysis of published cross-sectional data was conducted to summarize the evidence on the link between serum vitamin D levels and BMI in adults. Methods PubMed, ISI Web of Science, Scopus and Google scholar database were searched to May 2012 for all relevant published papers. We found 34 articles that reported the correlation coefficients between serum 25-hydroxy vitamin D (25(OH)D) levels and BMI in apparently healthy adults (>18 years). The primary analysis was done on these 34 papers that reported 37 correlation coefficients. To find the source of between-study heterogeneity, our secondary analysis was confined to eight studies that had used random sampling method and reported the correlations for the whole population. Results Our meta-analysis on 34 relevant papers revealed an overall significant inverse, but weak, association between serum 25(OH)D levels and BMI (Fisher's Z=0.15, 95% CI: 0.19, 0.11) with a significant heterogeneity between studies. In the subgroup analysis based on gender and study location (East vs. West), the inverse associations were significant in both genders (male: Fisher's Z=0.11, 95% CI: 0.14, 0.08 and female: 0.14, 95% CI: 0.21, 0.08) and both study locations (East: 0.09, 95% CI: 0.14, 0.04 and West: 0.23, 95% CI: 0.31, 0.17). In the subgroup analysis based on developmental status of countries, the weak inverse association remained significant in developed countries (0.17, 95% CI: 0.21, 0.14), but not in developing nations (0.10, 95% CI: 0.20, 0.01). Using meta-regression, we found that latitude (P=0.91) or longitude (P=0.2) of cities did not significantly contribute to the computed effect sizes. When we restricted our analysis to eight selected studies that used random sampling method, we reached the same findings. In this analysis, gender and developmental status of countries explained the between-study heterogeneity. Conclusion There is a significant inverse weak correlation between serum 25(OH)D levels and BMI in adult population, except for women living in developing countries. Further research particularly in developing countries and populations living near the equator is needed.