Introduction The aim of this study was to describe the migration pattern of the Exeter stem after revision with morsellised allograft bone and cement, to evaluate if restricted weight bearing had any influence on migration, and to measure, before and after revision, the quality of life comparing it to primary cemented hip arthroplasties.

Methods Forty-one consecutive stem revisions were followed by radiostereometry (RSA, 1. Selvik 1989). The accuracy of the RSA set-up was between 0.3 mm and 0.7 mm. The surgical procedure described by the Exeter group (2. Gie 1993) was used. All were first time revisions for aseptic loosening and all patients had had their primary arthroplasty for osteoarthritis. Bone stock deficiency was classified according to Gustilo and Pasternak. Sixteen were type I, 20 type II, five type III but none was classified as type IV. The Nottingham Health Profile was used to measure quality of life before and after revision.

Results All stems migrated distally and most of them also migrated medially or laterally and posteriorly. Migration was still observed in one third of stems between 1.5 and two year follow-ups. At two years stem subsidence averaged 2.5 mm, medial or lateral migration averaged 1.2 mm and posterior migration averaged 2.9 mm. No correlation to the pre-operative bone stock deficiency was observed. Between two and five years only marginal migration occurred in 12 of the 15 stems followed for five years. No differences in the migration pattern were detected when free weight bearing was allowed immediately after revision in hips without intra-operative skeletal complications compared to when restricted weight bearing was practiced. Most migration occurs within the first two weeks after surgery. NHP scales for pain, physical mobility, sleep and energy scales improved significantly. NHP scores were in all six scales comparable to those of primary arthroplasties. No rerevision was performed and no stem had radiographic sings of loosening.

Conclusions Most migration occurred early after revision and decreased gradually. Marginal migration after two years does not deteriorate the results during the first five years after surgery. Quality of life (patient outcome) after revision with impacted morsellised allograft bone and cement was comparable to that of primary arthroplasties.

In relation to the conduct of this study, one or more of the authors is in receipt of a research grant from a non-commercial source.

Aim of many surgical operations on the spine is very often to achieve a solid fusion between two or more vertebrae (arthrodesis). Describing lumbar spine mobility radiographically has been determined to be very imprecise with measurement errors of three to six mm in the sagittal plane. Using roentgen stereophotogrammetric analysis (RSA) it is possible to perform clinical kinematic lumbar studies with high accuracy. Many experimental studies have presented basic data on the stabilising implant effect in human cadaver lumbar spines, but no study compared the in vivo stabilising effect in different types of implant.

The RSA was performed in a radiographic set-up with two 40° angulated roentgen tubes with simultaneous exposures. A combined reference plate and calibration device with 0.8 mm tantalum balls at known positions was situated between the patient and the uniplanar film cassette, enabling simultaneous calibration and patient examination. At each RSA the patients were examined in supine and erect positions without corset. The translatory movements, induced by the change in position, of the 0.8 mm tantalum balls implanted into the fused vertebrae were calculated by computed data processing. These translations visualised the movements of the most proximal vertebra of the fusion in relation to the most distal.

The present study has demonstrated the in vivo stability of lumbar fusion augmented with transpedicular screws to be adequate. Sagittal translation seems to be easier to elicit than movements along the other three-dimensional axes. A widely decompressed and destabilised vertebra without screw fixation increases the risk for persisting intervertebral translations. The RSA technique described seems to be a good way for comparing the in vivo behaviour of different implant systems