Osteoporosis treatment considerations based upon fracture history, fracture risk assessment, vertebral fracture assessment, and bone density in Canada.

Abstract

To inform criteria for pharmacologic treatment in women age 65years and older, we examined subgroups defined from fracture history, MOF calculated with BMD (MOF-BMD), VFA, and BMD T-score using the population-based Manitoba BMD Program registry. The study population consisted of women age>65years was divided into mutually exclusive subgroups based upon fracture history, MOF-BMD ≥ 20%, vertebral fracture on VFA, and osteoporotic BMD T-score. Healthcare records were assessed for the presence of fracture diagnosis codes occurring after DXA assessment. For each subgroup, we estimated the proportion of individuals with BMD T-score in the osteoporotic range, predicted versus observed 10-year MOF probability, hazard ratio (HR) for MOF, and number needed to treat (NNT) for 3years to prevent a fracture event. The study population consisted of 39,475 women (median age 72years). The majority of women (76.8%) selected as being at high risk based on fracture history, MOF-BMD>20%, or vertebral fracture on VFA had a BMD T-score in the osteoporotic range. During a median follow-up of 8years, 5169 (13.1%) sustained one or more incident MOF. Fracture rates and HRs generally paralleled the FRAX prediction, except in women with a positive VFA where predicted risk based upon clinical risk factors prior to VFA underestimated the observed risk. NNT differed by the risk subgroup, and showed a gradient of decreasing NNT (consistent with greater benefit) in individuals with the highest fracture risk. Fracture history, fracture probability from FRAX, targeted vertebral fracture assessment (VFA), and BMD T-score can stratify older women into different levels of risk and treatment benefit. These results are expected to inform clinical practice guidelines in Canada.