RELIABLE DEXA MEASUREMENT TECHNIQUE FOR METAL-ON-METAL HIP RESURFACING

Abstract

Dual Energy X-ray absorption (DEXA) has been used to measure bone mineral density (BMD) around total hip prostheses. With the recent increase in the use of metal on metal hip resurfacing, such as the Birmingham Hip Resurfacing (BHR), there has been renewed concern over per prosthetic femoral neck fracture and implant loosening. DEXA quantitatively measures bone mineral density and therefore could predict impending loosening and fracture. To the best of our knowledge, there are no recorded studies assessing BMD around metal-on-metal hip resurfacings such as the BHR. Our intention was to produce a reliable method of measuring bone density around a metal-on-metal hip resurfacing, such as the BHR, prior to a prospective study.

We performed DEXA scans on five patients (7 BHR’s), who had undergone resurfacing with the BHR within the last two years, using the Hologic QDR 45000A scanner. Each BHR was scanned twice on the same day with complete patient repositioning between scans. We analysed the data with the Hologic prosthetic hip (v 8.26a: 3) scan analysis software (operating software 9.80D) by identifying a variable number of same-sized regions of interest (ROI) within the femoral neck. These ROI’s were derived from an inter-trochanteric line and the axis of the BHR stem in the femoral neck. Each of the 14 scans was analysed twice, by three of the authors independently; with at least one week between repeat analysis by the same observer. Statistical analysis was carried out by the local University Department of Statistics.

The variation within the same ROI in a given BHR was 0.00353, whereas the variation between all ROI’s was 1.155. The intraclass-correlation was 0.997 (i.e. the correlation between any two assessments of one ROI) with an overall coefficient of variation of 5%. The variation between the two scans for each BHR and between the three assessors was not significant (p=0.87 and p=0.42 respectively). The mean BMD of the individual ROI’s, between the two assessments of the same scans by the same assessor (i.e. intra-observer variation) was lower on the second assessment by 0.0214gcm⁻² (SD=0.0025) representing 0.5% mean density for all ROI’s. This difference was statistically significant (p< 0.001).

This method demonstrates excellent reproducibility of the method. Inter-scan and inter-observer variation was so negligible that a tiny intra-observer variation of 0.5% (of mean bone density) became statistically significant (p< 0.001), despite it making no difference to the overall intraclass-correlation. Statistical advice suggested that this very small difference in mean density (intra-observer) only reached significance due to the highly sensitive measurements and excellent reproducibility.

We have designed and demonstrated a safe, non-invasive and highly reproducible method for scanning BHR implants in vivo using DEXA. We will now use this method to prospectively study our BHR population to detect impending loosening or fracture.