Publication Date



Summary: There was no significant difference between the areas under receiver operating characteristics (AUROCs) and diagnostic indexes (sensitivity, specificity, positive predictive value, negative predictive value) for either major osteoporotic or hip fracture FRAX scores when comparing the unadjusted and trabecular bone score (TBS)-adjusted scores. Introduction: FRAX 10-year probability of fracture can be calculated with adjustment for the TBS. Studies have shown that TBS can improve FRAX assessments in some populations. This study aimed to determine if TBS-adjusted FRAX score is better than the unadjusted score for predicting major osteoporotic fracture (MOF) and hip fracture in Australian men. Methods: This study involved 591 men aged 40–90 years, enrolled in the Geelong Osteoporosis Study. Incident MOF (n = 50) and hip fractures (n = 14) were ascertained using radiological reports. Median follow-up time was 9.5 years (IQR7.5–11.4). Diagnostic indexes were calculated using cut points of ≥20% for MOF and ≥3% for the hip. AUROC curves were also determined for adjusted and unadjusted scores as continuous variables. Results: Sensitivity was higher in the TBS-adjusted scores (MOF 4%, hip 78.6%) than the unadjusted scores (MOF 2%, hip 57.1%), with a decrease in specificity (MOF 98.9 vs 99.3%; hip 79.9 vs 83.9%). When considering TBS-adjusted and unadjusted FRAX as continuous scores, AUROCs were 0.738 and 0.740, respectively, for MOF and 0.849 and 0.848 for the hip. Conclusions: Prediction of fractures by MOF or hip FRAX was not substantially improved by TBS adjustment. There was no difference in AUROCs or diagnostic indexes for cut-off points of ≥20 for MOF and ≥3% for hip FRAX.


Institute for Health and Ageing

Document Type

Journal Article

Access Rights

ERA Access

Access may be restricted.