Evaluation of the FRAX and Garvan fracture risk calculators in older women

Abstract

Fracture risk calculators estimate the absolute risk of osteoporotic fractures. We investigated the performance of the FRAX and Garvan Institute fracture risk calculators in healthy, older, New Zealand, postmenopausal women with normal bone mineral density (BMD) for their age. Fractures were ascertained in women initially enrolled in a 5-year trial of calcium supplements and followed on average for 8.8 years. Baseline data (1422 women, mean age 74 years, mean femoral neck BMD T-score -1.3) were used to estimate fracture risk during follow-up using the FRAX and Garvan calculators. The FRAX-New Zealand tool was used both with and without baseline BMD. The discrimination of the calculators was assessed using the area under the curve (AUC) of receiver operating characteristic curves. The calibration was assessed by comparing estimated risk of fracture with fracture incidence across a range of estimated fracture risks and clinical factors. For each fracture subtype, the calculators had comparable moderate predictive discriminative ability (AUC range: hip fracture 0.67-0.70; osteoporotic fracture 0.62-0.64; any fracture 0.60-0.63) that was similar to that of models using only age and BMD. The Garvan calculator was well calibrated for osteoporotic fractures but overestimated hip fractures. FRAX with BMD underestimated osteoporotic and hip fractures. FRAX without BMD underestimated osteoporotic and overestimated hip fractures. In summary, none of the calculators provided better discrimination than models based on age and BMD, and their discriminative ability was only moderate, which may limit their clinical utility. The calibration varied, suggesting that the calculators should be validated in local cohorts before clinical use.