Performance of FRAX and FRAX-Based Treatment Thresholds in Women Aged 40 and Older: The Manitoba BMD Registry.
J Bone Miner Res. 2019 Mar 28;: Authors: Crandall CJ, Schousboe JT, Morin SN, Lix LM, Leslie W
We examined among women aged ≥40 years the performance of the Fracture Risk Assessment Tool (FRAX) and FRAX-based osteoporosis treatment thresholds under the U.S. National Osteoporosis Foundation (NOF) and U.K. National Osteoporosis Guideline Group (NOGG) guidelines. We used registry data for all women aged ≥40 years in Manitoba, Canada with baseline bone mineral density (BMD) testing (n = 54,459). Incident major osteoporotic fracture (MOF), hip fracture, and clinical fracture were assessed from population-based health services data (mean follow-up 10.5 years). Age-stratified hazard ratios (HR) were estimated from Cox regression models. We assessed the sensitivity, specificity, positive predictive value (PPV), number needed to screen (NNS), and number needed to treat (NNT) to prevent a fracture (assuming 20% relative risk reduction on treatment) for osteoporosis treatment thresholds under the NOF and NOGG guidelines. Femoral neck T-score and FRAX (with and without BMD) predicted all fracture outcomes at all ages. There was good calibration in FRAX-predicted vs. observed 10-year MOF and hip fracture probability. Overall sensitivity (PPV) for incident MOF was: 25.7% (24.0%) for femoral neck T-score ≤ -2.5, 20.3% (26.3%) for FRAX (with BMD)-predicted 10-year MOF risk ≥ 20% (NOF threshold), 27.3% (22.0%) for FRAX-predicted 10-year MOF risk ≥ age-dependent cutoff (NOGG threshold), 59.4% (19.0%) for the NOF treatment algorithm, and 28.5% (18.4%) for the NOGG treatment algorithm. Sensitivity for identifying incident MOF varied by age, ranging from 0.0%- 26.3% in women 40-49 years-old and from 49.0% to 93.3% in women aged 80+. The gradient of risk for fracture prediction from femoral neck T-score and FRAX (with and without BMD) as continuous measures was strong across the age spectrum. The sensitivity and PPV of the strategies based on dichotomous cutoffs are low, especially among women aged 40-49 years (who have lowest incidence rates). Threshold-based approaches should be reassessed, particularly in younger women. This article is protected by copyright. All rights reserved. PMID: 30920022 [PubMed - as supplied by publisher]