Background Early disease markers can predict outcome in RA; however, patient-reported factors are also important yet their impact on disease outcomes, especially in the long-term is less-well established. We and others have recently demonstrated that patients with a high contribution of patient reported factors to disease activity (DAS) scores may display more favourable radiographic outcomes, although their potential impact on surgical outcomes remains unknown. In this study we examine associations between baseline standard clinical, laboratory and radiographic markers along with the effect of the proportion of DAS attributable to patient-reported components (joint tenderness and visual analogue score) termed DAS-P1 on subsequent orthopaedic surgery rates.
Objectives To examine the prediction of early disease markers and patient-reported factors on the timing and type of orthopaedic surgery.
Methods Data from the Early RA Study, ERAS, (1986-1999 n=1465) and the Early RA Network, ERAN, (2002-2012 n=1236) were linked to national datasets (Hospital Episode Statistics, National Joint Registry and the Medical Research Information Service). Standardized laboratory, radiographic and clinical variables including DAS and DAS-P were recorded at baseline in both cohorts, treatments based on conventional practice/guidelines of the time. Correlation analysis and multivariate competing-risks survival analysis were used to estimate sub-hazard ratios for intermediate (e.g. hand/foot) and major (large joint replacements) orthopaedic surgery, using baseline DAS and DAS-P (r<0.40). DAS-P was calculated only for patients with active disease (DAS≥3.2) in order to minimise the variation derived from calculations for those with small denominators.
Results In multivariate regression models earlier recruitment year, lower BMI and haemoglobin predicted higher risk for both intermediate and major surgery (p<0.05), and female gender only for intermediate (p<0.001). Standard DAS had no predictive value for either type of surgery. In the same models, DAS-P at baseline significantly predicted intermediate surgery (35% increase in risk, p<0.001); the effect on major surgery was similar, but did not reach statistical significance. After adjusting for gender and age at disease-onset DAS-P was correlated with worse HAQ (r=0.31, p<0.001) at baseline.
Conclusions The variation in power of different baseline variables to predict different types of intervention suggests contributions from different pathological and/or patient-level factors. The predictive value of the DAS-P index is in the opposite direction to that expected from previous findings that high contributions of patient reported factors to DAS are associated with lower radiographic scores of joint damage. Our findings suggest that non-inflammatory/non-structural factors might influence the decision to undertake joint surgery. The findings support that adequately addressing patient-level factors including pain, along with standard disease markers is important for improving long-term disease outcomes.
McWilliams DF, Zhang W, Mansell JS et al. Predictors of change in bodily pain in early rheumatoid arthritis: an inception cohort study. Arthritis Care Res 2012;64:1505-13.
Disclosure of Interest None declared