Relationship between body mass index and fracture risk is mediated by bone mineral density

The relationship between body mass index (BMI) and fracture risk is controversial. We sought to investigate the effect of collinearity between BMI and bone mineral density (BMD) on fracture risk, and to estimate the direct and indirect effect of BMI on fracture with BMD being the mediator. The study involved 2199 women and 1351 men aged 60 years or older. BMI was derived from baseline weight and height. Femoral neck BMD was measured by dual-energy X-ray absorptiometry (DXA; GE-LUNAR, Madison, WI, USA). The incidence of fragility fracture was ascertained by X-ray reports from 1991 through 2012. Causal mediation analysis was used to assess the mediated effect of BMD on the BMI-fracture relationship. Overall, 774 women (35% of total women) and 258 men (19%) had sustained a fracture. Approximately 21% of women and 20% of men were considered obese (BMI ≥ 30). In univariate analysis, greater BMI was associated with reduced fracture risk in women (hazard ratio [HR] 0.92; 95% confidence interval [CI], 0.85 to 0.99) and in men (HR 0.77; 95% CI, 0.67 to 0.88). After adjusting for femoral neck BMD, higher BMI was associated with greater risk of fracture in women (HR 1.21; 95% CI, 1.11 to 1.31) but not in men (HR 0.96; 95% CI, 0.83 to 1.11). Collinearity had minimal impact on the BMD-adjusted results (variance inflation factor [VIF] = 1.2 for men and women). However, in mediation analysis, it was found that the majority of BMI effect on fracture risk was mediated by femoral neck BMD. The overall mediated effect estimates were -0.048 (95% CI, -0.059 to -0.036; p < 0.001) in women and -0.030 (95% CI, -0.042 to -0.018; p < 0.001) in men. These analyses suggest that there is no significant direct effect of BMI on fracture, and that the observed association between BMI and fracture risk is mediated by femoral neck BMD in both men and women.