It is unclear whether patients with early radiographic osteoarthritis (OA) but severe hip symptoms benefit from total hip replacement (THR). We aimed to assess which factors were associated with successful THR in this patient group. From a consecutive series of 1,935 patients undergoing THR we identified 70 (3.6%) patients with early OA (Kellgren and Lawrence (KL) grades 0-2). These were compared with 200 patients with advanced OA (KL grades 3–4). Outcomes were Oxford Hip Scores (OHS), EQ5D and EQ-VAS scores; compared pre-operatively with one year post-operatively. We investigated which clinical and radiographic (plain x-ray, CT, MRI) features predicted successful THR in the early OA group. Success was defined as reaching a postoperative OHS≥42. The early OA group were significantly younger (61 vs 66 years; P=0.0035). There were no significant differences in body mass index, ASA grade or gender. After adjusting for confounders, the advanced OA group had a significantly greater percentage of possible change (PoPC) in OHS (75.8% versus 50.4%; P<0.0001) and improvement in EQ5D (0.151 versus 0.002; P<0.0001). There were no significant differences in complication, revision or readmission rates. In the early OA group, we identified 16/70 (22.9%) patients who had a ‘successful’ THR. Of those with early OA, 38 patients had pre-operative CT or MRI scans. Patients who had a ‘successful’ THR were significantly more likely to have subchondral cysts on CT/MRI (91.7% versus 57.7%; P=0.0362). The presence of cysts on CT/MRI was associated with a significantly greater PoPC in OHS (61.6% versus 38.2%; P=0.0353). The combination of cysts and joint space width (JSW) <1mm was associated with a PoPC of 68%. Plain radiographs were found to significantly underestimate the narrowest JSW compared to CT/MRI (2.4mm versus 1.0mm; P<0.0001). We advise caution in performing THRs in patients with early OA (KL grades 0-2) on plain radiographs. We advocate pre-operative cross-sectional imaging (CT or MRI) in these patients. In the absence of cysts on cross-sectional imaging, a THR seems unlikely to provide a satisfactory outcome.
