The OA Trial Bank: meta-analysis of individual patient data from knee and hip osteoarthritis trials show that patients with severe pain exhibit greater benefit from intra-articular glucocorticoids.
van Middelkoop M., Arden NK., Atchia I., Birrell F., Chao J., Rezende MU., Lambert RGW., Ravaud P., Bijlsma JW., Doherty M., Dziedzic KS., Lohmander LS., McAlindon TE., Zhang W., Bierma-Zeinstra SMA.
To evaluate the efficacy of intra-articular (IA) glucocorticoids for knee or hip osteoarthritis (OA) in specific subgroups of patients with severe pain and inflammatory signs using individual patient data (IPD) from existing trials.Randomized trials evaluating one or more IA glucocorticoid preparation in patients with knee or hip OA, published from 1995 up to June 2012 were selected from the literature. IPD obtained from original trials included patient and disease characteristics and outcomes measured. The primary outcome was pain severity at short-term follow-up (up to 4 weeks). The subgroup factors assessed included severe pain (≥70 points, 0-100 scale) and signs of inflammation (dichotomized in present or not) at baseline. Multilevel regression analyses were applied to estimate the magnitude of the effects in the subgroups with the individuals nested within each study.Seven out of 43 published randomized clinical trials (n = 620) were included. Patients with severe baseline pain had a significantly larger reduction in short-term pain, but not in mid- and long-term pain, compared to those with less severe pain at baseline (Mean Difference 13.91; 95% Confidence Interval 1.50-26.31) when receiving IA glucocorticoid injection compared to placebo. No statistical significant interaction effects were found between inflammatory signs and IA glucocorticoid injections compared to placebo and to tidal irrigation at all follow-up points.This IPD meta-analysis demonstrates that patients with severe knee pain at baseline derive more benefit from IA glucocorticoid injection at short-term follow-up than those with less severe pain at baseline.