Background Hand and foot surgery in Rheumatoid Arthritis (RA) is accepted as a surrogate marker for joint destruction and the result of failed medical treatment. The ability to predict this outcome is currently limited but has the potential to guide clinicians in early management decisions.
Objectives This group has already reported on the value of standard clinical and laboratory measures for predicting joint surgery in RA [EULAR2012]. This study examines a subset of this cohort with damage scores of hands and feet (Larsen).
Methods Patients were recruited within 2yrs of disease onset and prior to DMARD therapy to the Early RA Study (ERAS, n=1465, 1986-1998, median follow up 10yrs). Standard clinical & laboratory measures, and x-rays of hands and feet, were performed at baseline, and then yearly. Treatment of patients included disease modifying, steroid and biologic therapies according to standard UK practices for management of hospital based RA patients. Radiographs of hands and feet were scored according to Larsen in a subset (n=1186) to include wrists, mcp, pip & mtp joints. Source data of orthopaedic interventions included patient report and medical records, Hospital Episode Statistics (HES) and the National Joint Registry. Length of follow up was based on the National Death Registry. Joint surgery of hands and feet was mainly synovectomy, arthroplasty and fusion.
Results 1146 patients had Larsen scores of hands & feet at 0,1,2,3yrs, and included a total score, and 3 domain scores of wrists, MCP & PIP, & MTP joints. Joint surgery was performed in a total of 553 patients (38%), of whom 159 had at least one orthopaedic procedure for RA of a wrist, hand or forefoot joint, at a median of 10, 7 & 8.8yrs respectively. ROC analysis was performed using the first 3yrs of Larsen scores to identify suitable cut-off points of total and domain scores to predict surgery of the hands and feet. A Cox regression model with competing risk and controlling for age at disease onset, sex and baseline disease activity indicated that having a Larsen score >=10 within the first 3yrs increased the risk of hand or foot surgery by more than two-fold (SHR=2.58; P<0.001, 95% CI 1.63-4.08). Results for specific joints were as follows: wrist (SHR 2.67, P<0.001,95%CI 1.46-4.83), MCP/PIP (SHR=2.77, P<0.003, CI 1.43-5.38), MTP (SHR=2.98, P<0.001, CI 1.58-5.62). Graphic displays will demonstrate the cumulative hazard ratios over time of both the total and individual domain scores for eventual need for hand/foot surgery.
Conclusions Orthopaedic surgery is an uncommonly reported outcome in RA and difficult to predict. It is an important and frequent occurrence in long term RA, reflecting structural joint damage. Larsen scores in first 3yrs of RA add predictive value for eventual need for hand & foot surgery.
Acknowledgements With acknowledgement and thanks to the ERAS clinicians and nurses.
Disclosure of Interest None Declared