Bone microarchitecture, volumetric or areal bone mineral density for discrimination of vertebral deformity in adults: a cross-sectional study

Introduction/background: Both areal bone mineral density (aBMD) and bone microarchitecture have been associated with vertebral deformity (VD), but there are limited data on the utility of bone microarchitecture measures in combination with aBMD in discriminating VD. This study aimed to describe whether radial bone microarchitecture measures alone or in combinations with radial volumetric bone mineral density (vBMD) or spine aBMD can improve discrimination of VD in adults.

Methods: Data on 196 subjects (mean age (standard deviation, SD) = 72 (7) years, female 46%) were utilized. VD of T4-L4 and spine aBMD were measured using dual-energy X-ray absorptiometry. VD was defined if anterior to posterior height ratio was more than 3-SD, 4-SD below, or >25% decrease compared with the sex-matched normal means. Bone microarchitecture parameters at distal radius were collected using high-resolution peripheral quantitative computed tomography and analyzed using StrAx.

Results: The strongest associations were seen for the cortical thickness (odds ratios (ORs): 2.63/SD decrease for 25% and 2.38/SD decrease for 3-SD criterion) and compact cortical area (OR: 3.33/SD decrease for 4-SD criterion). The area under the receiver operating characteristic curve (AUC) for spine aBMD for VD was 0.594, 0.597, and 0.634 for 25%, 3-SD and 4-SD criteria, respectively (all p < 0.05). Compact cortical area, cortical thickness and compact cortical thickness alone had the largest AUCs for VD (0.680-0.685 for 25% criterion, 0.659-0.674 for 3-SD criterion, and 0.699-0.707 for 4-SD criterion). Adding spine aBMD or radial vBMD to each cortical measure did not improve VD discrimination (∆ AUC 0.8%-2.1%).

Conclusions: Cortical measures had the best utility for discriminating VD when used alone. Adding either spine aBMD or radial vBMD did not improve the utility of cortical measures.