Relationship between body mass index and fracture risk at different skeletal sites: a nationwide cohort study.

Abstract

The association between body mass index (BMI) and the risk of clinical fractures at different sites is unclear. This study aimed to examine associations between BMI and fractures at different sites in Korean men and women. This study analyzed 285,643 Korean adults (aged 50-80years) who participated in health examinations from 2002 to 2003 and were followed up until 2015. The incidences of osteoporotic fractures were assessed using the International Classification of Diseases (10th revision; ICD-10) and procedure or radiographic codes. After adjusting for confounders, hazard ratios (HRs) were calculated using Cox proportional hazard models for fracture risk. Site-specific associations between BMI and fractures were found without gender difference. Specifically, an L-shaped association was found for clinical vertebral fractures, wherein the adjusted HRs per 5kg/m2 increase were 0.80 (95% confidence interval [CI] = 0.76-0.83) in BMI groups < 25kg/m2 and 0.97 (95% CI = 0.92-1.03) in BMI ≥ 25kg/m2. A linear inverse correlation for wrist fractures was observed, wherein the HRs were 0.83 (95% CI = 0.81-0.86) per 5kg/m2 increase. For proximal humerus fractures, a non-linear U-shape association was found, wherein the adjusted HRs per 5kg/m2 increase were 0.66 (95% CI = 0.50-0.88) in BMI groups < 23kg/m2 and 1.25 (95% CI = 1.08-1.45) in BMI ≥ 23kg/m2. Low BMI was a risk factor for all tested fractures. Obesity was a risk factor for proximal humerus fracture, but it is a protective factor for wrist fracture.