Humeral BMD can be Measured With DXA and Is Lower in the Surgical Arm After Total Shoulder Arthroplasty

As arthroplasty leads to bone loss, we hypothesized that humeral bone mineral density (BMD) is lower after total shoulder arthroplasty (TSA) in the operative versus non-operative arm. However, there is no clinical approach to measure humeral BMD with dual-energy x-ray absorptiometry (DXA). The purposes of this pilot study were to develop DXA methodology to measure humerus BMD, propose humerus regions of interest (ROIs), compare TSA BMD to the non-operative arm, correlate humeral BMD with standard sites, and evaluate measurement reproducibility.Thirty-eight adults 1-5 years post-TSA had standard clinical DXA scans plus full humerus scans using the atypical femur fracture feature; precision was assessed in a subset (n = 32). Six custom ROIs were used to measure BMD throughout the humerus. Radius and humeral BMD were compared between arms by paired t-test and correlated ipsilaterally using Pearson's Correlation. Custom ROI BMD precision was assessed using the International Society for Clinical Densitometry (ISCD) advanced precision calculator.Study included 38 subjects (24M/14F), with mean (SD) age and time post-surgery of 69.6 (7.7) years and 2.5 (1.3) years respectively. BMD was lower (p < 0.01) at all custom humerus sites (3.8% to 8.2%) on the surgical side but not different at radius sites. Humeral BMD correlated positively with ipsilateral ultra-distal and 1/3 radius (r = 0.54 to 0.86; p < 0.05). Custom BMD precision (%CV) ranged from 6.0-16.0%.Humerus BMD can be measured using DXA and is lower in the TSA arm. Radius BMD correlated with humeral BMD but was not lower in the surgical arm. BMD precision was worse than usual clinical sites; use of software optimized for the femur is a notable limitation and likely contributes to suboptimal precision. Further study to assess the clinical utility of humeral BMD is needed. Automation and optimization of these measurements should improve precision.