Search

Background: Intra-articular corticosteroid injections (IACI) are a commonly used treatment for painful knee osteoarthritis (OA). Response to treatment varies the reason for which is unclear. Further there are no data concerning the impact of accuracy of injection and psychological factors including illness perceptions, pain catastrophizing and depression on outcome following IACI.Objectives: i) to undertake a systematic review looking at predictors of response to IACI in patients with symptomatic knee OA and, ii) to determine the role of psychological factors and accuracy of injection in predicting response to IACI.Methods: A systematic review was conducted using electronic databases for randomised trials and observational studies looking at predictors of response to IACI in knee and hip OA. An observational study of 141 consenting patients (105 primary OA and 36 secondary OA in the context of well controlled rheumatoid arthritis) receiving routine IACI as part of clinical care for knee OA was conducted including baseline assessment and outcome assessments at 3 and 9 weeks. Response was defined as at least 40% reduction of pain from baseline, using the Western Ontario and McMaster Universities Arthritis Index (WOMAC). Assessment included ultrasound (US) for features of synovial inflammation), radiographs, and assessment of psychological factors including the revised illness perception questionnaire (IPQR). Accuracy of injection was assessed using US. Characteristics of responders and non-responders to IACI at 3 and 9 weeks were determined using univariate statistics and significant factors entered into logistic regression models. Results: The systematic review found no consistent evidence for any disease or non-disease related predictor of response and no systematic exploration of the effects of psychological factors or accuracy of injection on treatment response. In the observational study, 83 (53%) of 141 subjects were responders to IACI at 3 weeks and 56 (44%) at 9 weeks. In univariate analysis, responders to treatment had higher scores for the IPQR domain treatment control and lower scores for IPQR consequences, depression and pain catastrophizing at both 3 and 9 weeks. Physical and patient related factors, including accuracy of injection and US features, were not associated with outcome, with the exceptions of higher baseline pain and previous experience of injection being associated with non-response at 9 weeks. In multiple regression, treatment control was the only independent predictor of response at 3 weeks. At 9 weeks, treatment control, consequences and depression were independent predictors of treatment outcome. Conclusion: In this observational study illness perceptions and depression predicted the outcome of IACI at 3 and 9 weeks. By contrast, physical factors including accuracy of injection did not influence outcome. Further work is needed to replicate these findings and elucidate mechanisms for these effects.