Eighteen comparative studies (3,139 patients, 3,404 hips) were included in the review. Fifteen studies were randomised controlled trials, two were non-randomised trials and one study included both a randomised and non-randomised arm. Study follow-up ranged from three months to 8.1 years. Four randomised trials were classified as being moderate to high quality, five as moderate quality and six as low quality. Where reported, between 65% and 100% of patients were followed up.
Functional outcomes: Comparison of metal on metal bearings versus metal on polyethylene at two years showed improvements in Harris hip scores associated with metal on polyethylene (-2.4, 95% CI -4.5 to -0.3, I2=0%; four studies). No other comparisons using Harris hip scores were significantly different. Sensitivity analyses did not significantly alter the overall findings. One of six studies that used alternative functional measurement scores showed substantially better functioning with metal on polyethylene compared to metal on metal bearing surfaces.
Surgical revision and dislocation: There were no statistically significant differences in occurrence of surgical revisions or dislocation between metal on metal versus metal on polyethylene bearings (two studies). One of two studies that compared metal on metal with ceramic on polyethylene reported substantially higher occurrence of dislocation in the ceramic on polyethylene group. One of two studies that compared ceramic on ceramic with metal on polyethylene bearings showed substantially lower occurrence of revision in the ceramic on ceramic arms. There were no other significant differences in surgical revision for any other comparison, including ceramic on ceramic versus ceramic on polyethylene (five studies).
Findings from six national registries were reported. These were mixed or conflicted with the comparative study findings. Further details were reported in the review.