Background Orthopaedic surgery is an accepted outcome measure in rheumatoid arthritis (RA) and considered a surrogate marker for joint destruction, but not commonly reported. The need for orthopaedic surgery is expected to be reduced over time with the greater use of more intense therapies for RA.
Objectives To examine orthopaedic surgery rates in RA in two UK multicentre inception cohorts conducted between 1986 & 2011.
Methods The Early RA Study (ERAS) recruited from 1986-1999 (n=1465), the Early RA Network (ERAN) from 2002 (n=1236). Standardised clinical, laboratory and x-ray measures were performed yearly in both cohorts. Symptom onset to baseline was median 6 months in both, and disease modifying, steroid and biologic therapies reflected conventional practice and guidelines of the time frames examined. Source data of all orthopaedic interventions included clinical datasets (patient report and medical records from 1986), which were validated with Hospital Episode Statistics (HES from 1997) and the National Joint Registry (NJR from 2002). Length of follow up was based on the National Death Registry.
Results 558 patients (38%) had 1287 orthopaedic procedures over maximum 25yr follow up in ERAS, 18% had major (e.g. total joint replacements), 16% had intermediate (e.g. wrist synovectomy, forefoot arthroplasty), and 17% minor (e.g. carpal tunnel release). In ERAN (maximum 9yrs follow up) 238 patients (17.5%) had 340 procedures, 6.7% major, 3.5% intermediate, 9.2% minor. The secular declines in orthopaedic surgical rates seen from 1987-2011 (18 to 16 per 1000 patients/year) will be displayed graphically. Declines were mainly in intermediate type surgery of wrist, hands and feet (8 to 6 per 1000 patients/year) and not in total joint replacements (mainly hip & knee). These changes coincided with changes in treatment practice. In the late 1980s and early 1990s, sulphasalazine was the first choice DMARD (70%) and methotrexate second choice, with a gradual reversal of the ratio over time. Median time to first DMARD gradually changed from 3 months to 2 weeks, and from 2000 onwards greater use of combination & triple therapies, and biologics. There was no evidence of major differences in disease characteristics at baseline by year of recruitment, nor in main predictive markers. An unusual finding was low baseline haemoglobin predicted a subgroup who eventually had multiple joint surgery by 5 & 10yrs (OR 2.6, 95%CI 1.5-4.4 & OR 3.0, 95%CI 2.1-4.4 respectively).
Conclusions The need for orthopaedic surgery is a common outcome in RA over time, and reports on this important feature are limited. Hand/foot surgery rates showed a consistent decline from 1986 to 2011, coinciding with secular changes in therapies. Hip/knee replacements were however similar suggesting a different pathological process or variable responses to therapy between large and small joint destructive processes.
Disclosure of Interest None Declared