vom Scheidt, Annika; Grisolia Seifert, Eric Flavio; Pokrant, Carolin; Püschel, Klaus; Amling, Michael; Busse, Björn; Milovanovic, Petar
Background Currently, bone densitometry fails to identify nearly half of those elderly patients at immediate fracture risk. To improve clinical assessment of vertebral fracture risk, we aimed to determine how the DXA-based 2D parameter Trabecular Bone Score (TBS) relates to subregional variability in 3D trabecular microstructure in young and elderly women compared to aBMD. Methods T12 vertebrae from 29 women (11 young: 32 ± 6 years, 18 aged: 71 ± 5 years) were DXA-scanned ex vivo in anterior-posterior (AP) and lateral projection providing vertebral aBMD and TBS. Additionally, aBMD and TBS were measured for three horizontal (superior, mid-horizontal, inferior) and three vertical subregions (anterior, mid-vertical, posterior) and related to 3D microstructure indices, i.e. bone volume per tissue volume (BV/TV), trabecular number and thickness (Tb.N, Tb.Th), based on HRpQCT. Results Subregional high-resolution tomography showed significant differences in trabecular parameters for both age groups: In horizontal subregions, BV/TV was lowest superiorly, Tb.Th was highest mid-horizontally, and Tb.N was lowest mid-horizontally and highest inferiorly. Correspondingly, aBMD varied between horizontal subregions, with differences depending on projection direction. TBS varied only in lateral projections of the aged group, with lower values for the mid-horizontal subregion. In vertical subregions, BV/TV, Tb.N, and aBMD were highest posteriorly for both groups. TBS did not differ between vertical subregions. Regression analysis showed aBMD as a predictor explained more of the variance in subregional 3D microstructure compared to TBS. Stepwise multi-regression analysis revealed only three combinations of subregion, projection, and group where aBMD and TBS were both significant predictors. Conclusions Subregional aBMD reflects variations in trabecular bone microstructure better than subregional TBS for trisected regions. Specifically, lateral aBMD identifies microstructural heterogeneities independent of age and may improve prediction of vertebral strength and susceptibility to specific fracture types.