Geographic variation of bone mineral density and selected risk factors for prediction of incident fracture among Canadians 50 and older

Abstract

Background Striking geographic variation in the incidence of osteoporotic fracture has been shown in national and international studies. The contributing risk factors for this variation are not fully understood. Objective To determine the geographic variation of bone mineral density (BMD) values, prevalent low-trauma fracture, prior falls, and vertebral deformity and to determine how this variation is related to the geographic variation of incident low-trauma fracture. Methods We studied incident fracture among 2484 men and 6093 women ages 50 and older from CaMos, a randomly-selected population-based longitudinal cohort recruited from within 50 kilometers of nine cities across Canada. Analyses included up to an eight-year follow-up. Results Estimates of fracture incidence are all age-standardized and given per 1000 person-years and CI denotes confidence interval. Among men, the lowest incidence of low-trauma fracture was 3.2 (95% CI: 1.1–7.5) in Quebec and the highest was 11.9 (95% CI: 7.1–18.6) in Calgary, compared with an overall incidence of 7.2 (95% CI: 5.8–8.7). Among women, the lowest incidence of low-trauma fracture was 11.5 (95% CI: 8.5–15.1) in Halifax and the highest was 18.5 (95% CI: 14.6–23.3) in Calgary, compared with an overall incidence of 15.3 (95% CI: 14.1–16.7). The regional variation in low-trauma fractures was similar to variation in hip fracture incidence among women (Pearson correlation, r = 0.46 to 0.76) but not men ( r = − 0.06 to 0.05). We noted significant geographic variation in the prevalence of low BMD, as defined by BMD T-score ≤ − 2.5, however this variation was not directly related to low-trauma fractures or other risk factors. Furthermore, a model including age, BMD, falls, vertebral deformity, and prior clinical fracture was a good predictor of geographic variation of low-trauma fracture incidence in both men ( r = 0.66) and women ( r = 0.84). Conclusions For both men and women, the burden of low-trauma fracture is not related to the prevalence of osteoporosis as defined by BMD, but is related to a more comprehensive assessment of fracture risk including the following: age, BMD, falls, prior fracture, and vertebral deformity.