Previously, radiostereometric analysis following hip revision performed using impacted morsellised allograft bone and a cemented Exeter stem has shown continuous subsidence of the stem for up to five years. It is not known whether the subsidence continues thereafter. In our study, 17 of 25 consecutive osteo-arthritic patients with aseptically loose stems who underwent first-time revision using impacted morsellised allograft bone and a cemented Exeter stem were followed by yearly radiostereometric examinations for nine years. The mean subsidence at six weeks was 1.1 mm (0.1 to 2.3), from six weeks to one year 1.3 mm (0 to 2.6), from one to five years 0.7 mm (0 to 2.0), and from five to nine years 0.7 mm (0.1 to 3.1). That from six weeks to nine years was 2.7 mm (0 to 6.4) (95% confidence interval 2.0 to 3.5). The Charnley pain score significantly improved after revision, and was maintained at nine years, but walking ability deteriorated slightly as follow-up extended. Of the eight patients who were not followed for nine years, two had early subsidence exceeding 11 mm.

Our findings show that in osteo-arthritic patients who undergo revision for aseptic loosening of the stem using impacted morsellised allograft bone and a cemented Exeter stem, migration of the stem continues over nine years at a slower rate after the first year, but without clinical deterioration or radiological loosening.

Several methods of measuring the migration of the femoral component after total hip replacement have been described, but they use different reference lines, and have differing accuracies, some unproven. Statistical comparison of different studies is rarely possible.

We report a study of the EBRA-FCA method (femoral component analysis using Einzel-Bild-Röntgen-Analyse) to determine its accuracy using three independent assessments, including a direct comparison with the results of roentgen stereophotogrammetric analysis (RSA).

The accuracy of EBRA-FCA was better than ±1.5 mm (95% percentile) with a Cronbach’s coefficient alpha for interobserver reliability of 0.84; a very good result.

The method had a specificity of 100% and a sensitivity of 78% compared with RSA for the detection of migration of over 1 mm. This is accurate enough to assess the stability of a prosthesis within a relatively limited period. The best reference line for downward migration is between the greater trochanter and the shoulder of the stem, as confirmed by two experimental analyses and a computer-assisted design.