Abstract
OBJECTIVE: To evaluate the efficacy of intra--articular (IA) glucocorticoid for knee or hip osteoarthritis (OA) in specific subgroups of patients according to the baseline severity of pain and inflammatory signs using individual patient data (IPD) from existing trials. Furthermore, this study aims to assess if a baseline pain cut-off was associated with clinically important effectiveness of IA glucocorticoid. This is an update of an IA glucocorticoid IPD meta-analysis by the OA Trial Bank. METHOD: Randomized trials evaluating one or more IA glucocorticoid preparations in hip and knee OA, published to May 2018 were selected. IPD of patient and disease characteristics and outcome measures were acquired. The primary outcome was pain severity at short-term follow-up (up to 4 weeks). Potential interaction effect of severe pain (≥70 points, 0-100 scale) and signs of inflammation at baseline were studied using a two-stage approach with general liner model followed by random effects model. Analysis of trend was conducted, assessing if a baseline pain cut-off was associated with the threshold for clinically important treatment effect of IA glucocorticoid compared to placebo. RESULTS: Four out of 16 eligible randomized clinical trials (n = 641) were combined with the existing OA Trial Bank studies (n = 620), yielding 1261 participants from eleven studies. Participants with severe baseline pain compared to those with less severe pain had greater pain reduction at mid-term (around 12 weeks) (mean reduction: -6.90 (95%CI -10.91; -2.90)), but not at short- and long-term. No interaction effects were found between inflammatory signs and IA glucocorticoid injections compared to placebo at all follow-up time-points. Analysis of trend demonstrated treatment response to IA glucocorticoid from baseline pain levels >50 (0-100 scale) and above. CONCLUSION: This updated IPD meta-analysis demonstrated that participants with severe pain compared to those with less severe pain at baseline experienced significantly more pain relief with IA glucocorticoid compared with placebo at mid-term.