Background/Purpose: Axial spondyloarthritis (axSpA) is associated with osteoporosis and fragility fracture risk that may not be reflected by areal bone mineral density (aBMD) measurements alone. DXA-derived three-dimensional (3D) analysis has emerged as a promising method for assessing structural bone properties of the femur, showing good correlation with quantitative computed tomography (1).
We aimed to investigate the discriminative performance of DXA-derived 3D parameters of the femur in comparison to aBMD and trabecular bone score (TBS) for pre-existent fragility fractures (FFx).

Methods: This cross-sectional analysis is part of the Rh-GIOP prospective observational cohort study focusing on bone health in inflammatory rheumatic diseases. We analyzed baseline visits of axSpA patients fulfilling the ASAS 2009 classification criteria. Using 3D-Shaper software (v2.12, 3D-Shaper Medical, Barcelona, Spain), we derived several parameters (integral, trabecular and cortical volumetric BMD, cortical thickness and cortical surface BMD) from hip DXA scans. Logistic regression analysis was used to develop discrimination models incorporating DXA-based T-scores, TBS, and 3D-Shaper parameters to detect prevalent FFx, vertebral fractures (VFx), and non-vertebral fractures (NVFx). Area under the curve (AUC) with 95% confidence intervals were calculated.

Results: Eighty patients with axSpA (64% male) were analyzed. The prevalence of FFx was 34%, and 15% had a history of VFx. DXA T-scores showed AUC values of 0.62 (0.48; 0.75), 0.54 (0.39; 0.70), and 0.64 (0.50; 0.78) for any FFx, VFx, and NVFx, respectively. DXA-derived 3D parameters showed better discrimination for fragility fractures, with AUC values of 0.75 (0.64; 0.86), 0.81 (0.69; 0.94), and 0.72 (0.59; 0.85). Incorporating these parameters in a logistic regression model with DXA and TBS further improved the discriminative ability (any FFx: AUC 0.78 (0.67; 0.89), VFx: AUC 0.86 (0.74; 0.98), NVFx: AUC 0.75 (0.63; 0.88)). TBS alone or in combination with DXA did not improve the discrimination of the model.

Conclusion: Our results suggest that DXA-derived 3D hip parameters offer superior discrimination of pre-existent FFx compared to DXA or TBS alone in patients with axSpA. Future longitudinal analysis will determine whether DXA-derived 3D femoral parameters can improve fracture prediction in clinical practice.

« Back to ACR Convergence 2024

ACR Meeting Abstracts - https://acrabstracts.org/abstract/dxa-derived-3d-parameters-of-the-hip-discriminate-prevalent-fragility-fractures-in-individuals-with-axial-spondyloarthritis/