To assess the occurrence of and predictive factors for orthopaedic surgery in an inception cohort of rheumatoid arthritis (RA) patients recruited and followed prospectively for 5 yr in nine regions in England.

Standard clinical, laboratory and radiological assessments and all interventions were recorded at baseline and yearly in RA patients (less than 2 yrs symptoms) prior to the use of disease-modifying drugs.

One thousand and sixty-four patients completed 5 yr of follow-up. Two hundred and sixty-four orthopaedic procedures for RA were performed in 181 (17%) patients at a median of 36.5 months from baseline. Seventy-five (7%) had replacements of major joints. Risk factors at baseline for large joint replacement surgery were a low haemoglobin concentration [odds ratio scores (OR) 3.4, 95% confidence interval (CI) 2.1-5.8] and high scores for erythrocyte sedimentation rate (ESR) (OR 3.2, CI 1.8-5.3), disease activity (DAS) (OR 2.1, CI 1.2-3.5) and Larsen X-rays (OR 2.6, CI 1.4-4.8). For hand or foot joint surgery (4%), risk factors included female gender (OR 3.2, CI 1.3-7.6), joint score (OR 2.3, CI 1.2-4.3), erosions (OR 2.3, CI 1.1-4.8), DAS (OR 2.4, 1.3-4.5) and Health Assessment Questionnaire score (OR 1.9, CI 1.0-3.6). No significant associations were seen for tendon, soft tissue or other minor procedures (6%). The HLA-DRB1 RA shared epitope was associated with any type of orthopaedic surgery (OR 1.7, CI 1.1-2.7).

Eleven per cent of RA patients treated with conventional drug therapy for 5 yr underwent large- or small-joint surgery, an outcome which could be compared against that for new disease-modifying drugs. Risk factors varied according to type of surgery, but included standard clinical and laboratory measures. In order to reduce the eventual need for surgery, a therapeutic target in the first year of RA is the suppression of disease activity, as measured by haemoglobin and ESR. These are useful details for clinicians, health professionals and patients.