A meta-analysis of the association between body mass index and risk of vertebral fracture

We conducted a meta-analysis of prospective studies to assess the association between BMI and incident vertebral fracture. We found that as body mass index (BMI) increases, the risk of vertebral fracture decreases in men, but not in women, suggesting possible gender differences in the relationship of BMI with risk of vertebral fracture. Recent evidence suggests that the relationship between BMI and fracture risk may be site-specific. We conducted a systematic review and meta-analysis of prospective studies to investigate the association between BMI and risk of incident vertebral fracture. PubMed and Embase were searched for relevant articles published from inception through February 15, 2017. Extracted relative risks (RR) from the prospective studies were pooled using random-effects meta-analysis. Six studies were included, with a total of 105,129 participants followed for 3 to 19 years. The pooled RR (95% confidence interval [CI]) for vertebral fracture per each standard deviation increase in BMI was 0.94 (95% CI = 0.80-1.10) with significant heterogeneity (I (2) = 88.0%, p < 0.001). In subgroup analysis by gender, we found a significant inverse association between BMI and risk of vertebral fracture in men (RR = 0.85, 95% CI = 0.73-0.98, n = 25,617 participants) but not in women (RR = 0.98, 95% CI = 0.81-1.20, n = 79,512 participants). Across studies of women not adjusting for bone mineral density (BMD), there was no significant association between BMI and risk of vertebral fracture (RR = 0.91, 95% CI = 0.80-1.04, p = 0.18, n = 72,755 participants). However, BMI was associated with an increased risk of vertebral fracture in studies of women that adjusted for BMD (RR = 1.28, 95% CI = 1.17-1.40, p < 0.001, n = 6757 participants). Substantial heterogeneity was found among studies of women (I (2) = 90.1%, p < 0.001), which was partly explained by the adjustment for BMD (adjusted R (2) = 61%). We found no evidence of publication bias (p = 0.40). In conclusion, our findings suggest that there might be gender differences in the relationship of BMI with risk of vertebral fracture. Further research is needed, including the assessment of other measures of adiposity, such as visceral adiposity, on the risk of vertebral fracture.