Bone mineral density of the femoral neck in resurfacing hip arthroplasty Precision of DXA biased by region of interest and rotation of the hip

Background and purpose Resurfacing total hip arthroplasty (RTHA) may preserve the femoral neck bone stock postoperatively. Bone mineral density (BMD) may be affected by the hip position, which might bias longitudinal studies. We investigated the dependency of BMD precision on type of ROI and hip position. Method We DXA-scanned the femoral neck of 15 resurfacing patients twice with the hip in 3 different rotations: 15 degrees internal, neutral, and 15 degrees external. For each position, BMD was analyzed with 3 surface area models. One model measured BMD in the total femoral neck, the second model divided the neck in two, and the third model had 6 divisions. Results When all hip positions were pooled, average coefficients of variation (CVs) of 3.1%, 3.6%, and 4.6% were found in the 1-, 2-, and 6-region models, respectively. The externally rotated hip position was less reproducible. When rotating in increments of 15 degrees or 30 degrees, the average CVs rose to 7.2%, 7.3%, and 12% in the 3 models. Rotation affected the precision most in the model that divided the neck in 6 subregions, predominantly in the lateral and distal regions. For larger-region models, some rotation could be allowed without compromising the precision. Interpretation If hip rotation is strictly controlled, DXA can reliably provide detailed topographical information about the BMD changes around an RTHA. As rotation strongly affects the precision of the BMD measurements in small regions, we suggest that a less detailed model should be used for analysis in studies where the leg position has not been firmly controlled.