Medical treatment of rheumatoid arthritis (RA) aims at controlling synovitis and arresting erosive disease. Orthopaedic surgery deals with joint destruction, reflecting the severity of damage, but also deals with local persistent synovitis and tenosynovitis. Along with altering medical treatment strategies, rheumatologists and surgeons feel they have changed the type and timing of surgery in recent years. We performed a cross sectional study on the use of orthopaedic surgery in RA, to provide data on types of procedures and evolution of timing and type of surgery.

A cross sectional study (January to April 2004) was carried out in two university and three non-university outpatient clinics. Consecutive outpatients with RA¹ (n = 285) were considered, except for one centre that included patients attending the day care centre for treatment with infliximab (n = 20). An examiner questionnaire assessed disease characteristics and medical and surgical treatment specific for RA. Joint-sacrificing procedures refer to surgery where the joint cartilage is either replaced or resected with or without subsequent joint fusion and non-joint-sacrificing surgery refers to all articular or periarticular procedures that do not result in the resection or replacement of the cartilage.

Time to surgery was assessed by Kaplan-Meier analysis. Patients were dichotomised in two cohorts according to the year of diagnosis (before or after 1990) because a clear change was seen from this year onward. Cox proportional hazard analysis explored the influence of belonging to one of the cohorts on each type of surgery multivariate.

The mean (SD) age of the 285 patients (70% university) was 57 (12) years and the mean disease duration 9.7 years. Two hundred and fifty four (89%) patients currently used one or more disease modifying antirheumatic drug, including a biological agent in 37%. Seventy one patients belonged to the early and 214 to the late cohort.

One hundred and six (37%) patients underwent 274 procedures. Sixty seven (24%) reported multiple interventions. Median disease duration at the first operation was 7 years (mean (SD) 8.0 (6.9) years). Table 1 provides an overview of the frequencies and timing of all types of surgical procedures. A more detailed table is available online in Appendix 1, available at http://www.annrheumdis.com/supplemental.

Figures 1A and B show that patients in the recent cohort were 1.79 times (95% confidence interval 1.10 to 2.93) more likely to have earlier joint surgery than those in the later joint surgery cohort, specifically earlier non-joint-sacrificing procedures (hazard ratio 6.35 (95% confidence interval 2.31 to 16.95)).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 Time to (A) first joint-sacrificing procedure and (B) first non-joint-sacrificing procedure for patients diagnosed before, compared with those diagnosed after, 1990.

In a cross sectional group of patients with RA treated by a rheumatologists, 28% had undergone joint-sacrificing and 19% non-joint-sacrificing orthopaedic procedures. Of those having surgery, 63% had multiple procedures. Remarkably, there was a shift towards earlier non-joint-sacrificing surgery in the group diagnosed after 1990.

Compared with reported values, the prevalence of procedures showed similar trends.^2–8 However, in a population of patients with RA sampled in Canada a decline in orthopaedic surgery among the later cohorts was reported.⁶ A survey from Sweden confirmed a decrease in admissions of at least one night for lower limb surgical procedures between 1987 and 2001.⁹

During the past 15 years, increasing expectations from patients and healthcare providers have focused on tighter disease control and optimal functioning. Accordingly, the increase in non-joint-sacrificing surgery (synovectomy) points to the role of surgery in controlling local inflammation. If our results can be confirmed in other regions, it will be important to assess the long term effectiveness of early non-joint-sacrificing surgery as an adjuvant to medical treatment.