Multiple studies have explored the association between serum or plasma vitamin B12 status and obesity, in part because of the relationship between elevated homocysteine concentrations and atherosclerosis. This review will address the inconsistent finding of these studies with the objective of determining whether vitamin B12 concentrations are lower in people with higher body mass indices. MEDLINE and EMBASE were searched to February 2017. Observational studies in general and clinical populations comparing serum/plasma B12 concentrations across groups of different body mass indices were selected. We did network and pairwise meta-analyses of serum/plasma B12, folate and homocysteine using frequentist techniques. Evidence-based items potentially indicating risk of bias were assessed. Of 844 citations, we identified 19 eligible observational studies with 7,055 participants. The overall network, while showing no significant inconsistency between indirect and direct comparisons (P=0.34), was qualitatively inconsistent. Based on the results of the meta-regression, in an exploratory sub-network meta-analysis where obesity groups were combined, we excluded disease-specific populations and studies with inadequate description of populations. The direction of the indirect and direct evidence was consistent. The pairwise results from this sub-network showed lower levels of B12 in people with higher body mass indices: obesity versus control difference in means (MD) -56pmolL-1 (95% CI -90, -23), obesity versus overweight MD -21pmolL-1 (95% CI -37, -5) and overweight versus control MD -51pmolL-1 (95% CI -51, -24). Heterogeneity remained very large for most comparisons, and all the studies carried a high risk for bias. This review did not establish an inverse association (or J-curve) between serum or plasma B12 concentrations and body mass index, but the direct pairwise evidence is consistent with an inverse association and supports further investigation.
Read full abstract