Introduction: Dislocation is the most common complication resulting in re-operation following total hip arthroplasty, accounting for 33.5% of revisions. This study investigates the relationship between bearing surface and the risk of revision due to dislocation.

Materials and Methods: Analysis was based on 110,239 primary total hip arthroplasties with a primary diagnosis of osteoarthritis. Data were collected by the Austra-lian Orthopaedic National Joint Replacement Registry from September 1999 to December 2007. The bearing surfaces were: 20627 (18.7%) ceramic-on-ceramic, 14001 (12.7%) ceramic-on-polyethylene, 12208 (11.1%) metal-on-metal, and 62437 (56.6%) metal-on-polyethylene. In 966 (0.8%) hips the bearing surface was unknown.

Results: There were 862 (0.8%) hips revised due to dislocation, with a rate of 0.3 revisions per 100 component years. Survival analysis with an end point of revision due to dislocation was performed. Revision for dislocation is potentially associated with variables other than bearing surface (including age and femoral component head size). Therefore analyses were stratified by femoral head size (≤28mm and > 28mm), and age (< 65 years and ≤65 years). There is a significantly higher rate of revision for dislocation in ceramic-on-ceramic bearing surfaces compared to metal-on-polyethylene bearing surfaces after adjustment for age, sex and head size in the head size < =28mm/Age < 65 group (hazard ratio = 1.53, 95% C.I. = 1.02 to 2.30, p=0.041) and the head size > 28mm/Age > =65 group (hazard ratio = 1.73, 95% C.I. = 1.10 to 2.74, p=0.016).

Discussion: Ceramic-on-ceramic bearing surfaces have a higher risk of revision due to dislocation in the femoral head sizes and ages discussed above, compared with metal-on-polyethylene. Possible mechanisms for this observed difference may include patient selection, the limits to head and liner offset options when using ceramic bearing surfaces or higher rates of revision after dislocation due to ceramic head or liner damage. However our results are based on a seven year follow-up, and higher rates of late dislocation with polyethylene bearings may be observed in association with higher wear rates compared with ceramic liners.