Background:​ Obese older adults have higher bone mineral density (BMD), but other characteristics of obesity may predispose these individuals to fracture. We aimed to compare clinical risk factors for incident fracture in obese and non-obese older men.

Methods:​ Body mass index (BMI) and body fat percentage were assessed at baseline by anthropometry and dual-energy X-ray absorptiometry, respectively, in 1625 community-dwelling men aged ≥70 years. Categories for non-obese and obese were <30 kg/m2and ≥30 kg/m2 for BMI, and <30% and ≥30% for body fat percentage. Sociodemographic, medical, physical function and blood biochemistry parameters were collected at baseline. Self-reported incident fractures of any type and cause (excluding pathological fractures and fractures of hands, fingers, feet, toes and the skull) were confirmed by radiographic reports and recorded up to 9 years. Hip fractures were followed for 14 years using data linkage.

Results:​ Prevalence of obesity was 27% according to BMI and 44% according to body fat percentage. There were no differences in incidence of any fracture between non-obese and obese men by BMI (10.7 vs 9.3%, respectively;​ P > 0.05) or body fat percentage (10.2 vs 10.6%, respectively;​ P > 0.05). Significant interactions were observed demonstrating that dementia increased hazard for incident any and hip fracture in non-obese men (adjusted hazard ratio 7.08;​ 95% CI 3.27–15.36 and 8.36;​ 3.13–22.31, respectively) but not obese men. Past-year falls increased hazard for any fracture in obese men (2.86;​ 95% CI 1.60–5.10) but not non-obese men while higher luteinizing hormone concentrations reduced hazard for hip fracture in obese men (0.91;​ 0.85–0.97 per IU/L) but not non-obese men.

Conclusions:​ In community-dwelling older men, obesity does not protect against incident fracture. Assessments of falls history and gonadotrophin levels, in addition to established clinical risk factors for fracture, may contribute to improvements in fracture prediction in obese older men.