Decreased precision for BMD measurements in the prosthetic knee using a non-knee-specific software.

Non-knee-specific software has been widely used for bone mineral density (BMD) measurement at the knee, which coincide with a higher measurement error. This study was conducted to test whether software developed for BMD measurement at the hip could be used in the knee by using dual-energy X-ray absorptiometry (DXA) and whether knee rotational alignment or differences in soft tissue substitutes would influence the precision of the BMD measurement. Twelve knees operated with total knee arthroplasty were included. The knees were scanned four to six times each in four rotational positions using DXA with either a Plexiglas rod or rice bags as a soft tissue substitute. The BMD was measured in one manually defined region at the proximal tibia by using the software of Orthopedic. The precision (coefficient of variation) at the knee was 5.1-9.0% when using the Plexiglas rod and 7.3% when using rice bags. The poor precision was mainly explained by inconsistency in tissue baseline establishment and/or by low baseline bone density (Adj R2 = 0.78-0.90, p < 0.000). Neither intentional rotation of the knee nor using different soft tissue substitutes significantly altered the precision. The results suggest that the use of a non-knee-specific software in the prosthetic knee is not appropriate.