Bone Abstracts

Background: Type 2 Diabetes Mellitus (T2DM) presents a "bone paradox" where high bone mineral density (BMD) masks skeletal fragility. Trabecular bone score (TBS) assesses microarchitecture, but legacy algorithms (v3) are confounded by abdominal adiposity. This study evaluates whether TBS v4, with enhanced soft-tissue compensation, improves fracture risk assessment across T2DM status.

Methods: Logistic regression predicted major osteoporotic fractures (MOF) in patients over 50. We compared TBS v3 and v4, adjusting for age, sex, and waist-to-height ratio (WHtR). Full models integrated lumbar spine (LS) BMD and T2DM status. Predictive performance was quantified using area under the ROC curve (AUC).

Results: Despite higher BMD values at the lumbar spine and hip (P 0.001, Table 1), T2DM patients had double the hip fracture prevalence (17.1% vs 8.6%, P 0.001). In models using TBSv3, WHtR was significantly associated with fracture (P 0.029), indicating residual confounding by abdominal adiposity. This association disappeared when TBSv4 was used (P 0.239), suggesting improved soft-tissue correction. In the full model including BMD and T2DM, TBSv4 remained an independent predictor of fracture (Or 0.742, P 0.002). Adding TBSv4 to BMD modestly increased AUC from 0.646 to 0.655, although this did not reach statistical significance (P 0.192).

Conclusions: TBS v4 eliminates the abdominal adiposity bias (WHtR) seen in v3. In T2DM, where BMD is often falsely elevated, TBS v4 provides a robust, independent predictor of fracture. While overall AUC improvement was not statistically significant, the removal of soft-tissue interference makes TBS v4 more clinically reliable for metabolic bone assessment.