Fracture risk prediction based on multivariable diagnostic models incorporating dual-energy X-ray absorptiometry based bone mineral density, trabecular bone score, and clinical risk factors in women aged ≥ 50 years: a prospective study

Abstract

According to the available reports, there is a major treatment gap in the management of osteoporotic fractures in Poland. It is crucial to identify those at high risk of low-energy fractures who ought to be treated. The study aimed to assess the diagnostic value of dual-energy X-ray absor­ptiometry (DXA) and the DXA-derived trabecular bone score (TBS) in fracture risk prediction in Polish women ≥ 50 years of age at risk of osteoporosis. Additional measures included the analysis of how incorporating other risk factors may improve the fracture prediction model. From a database of 5,000 randomly selected women, a total of 411 women aged ≥ 50 years were finally included – those who had a DXA examination followed by a complete medical questionnaire performed. After 5–9 years of observation, participants completed a follow-up questionnaire similar to the baseline questionnaire. In addition, TBS analysis was performed retrospectively using lumbar spine DXA scans. Then their fracture risk was calculated using the FRAX algorithm. Most patients with osteoporotic fractures were in a group with a degraded bone microarchitecture and a diagnosis of osteoporosis. Fracture patients without osteoporosis (T-score > –2.5) were mostly (73%) characterised by a partially or significantly degraded bone microarchitecture. The highest predictive value (AUC 0.6) was observed in relation to the models FRAX BMD and FRAX BMD-TBS (p < 0.05), which included both skeletal and other clinical risk factors. Also, the outcomes for all other analysed models (FRAX BMI + TBS, FRAX BMD + TBS, FRAX BMI + BMD spine + TBS) were statistically significant. Indirect bone microarchitecture analysis using DXA-derived TBS predicts fragility fracture risk independently of spine BMD and with comparable predictive performance in postmenopausal women in Poland. Further studies on fracture prediction models in patients with osteopenia are required.