Effect of monotherapy with conventional synthetic disease-modifying anti-rheumatic drugs or glucocorticoids on radiographic progression in rheumatoid arthritis: a network meta-analysis of 64 treatment arms from 31 randomized controlled trials

Objective
In a previous network meta-analysis (NMA) of randomized controlled trials (RCTs), we analysed the effects of 27 potential conventional synthetic disease-modifying anti-rheumatic drugs (csDMARDs), glucocorticoid (GC), and placebo in patients with rheumatoid arthritis (RA), using tender joint count as the primary outcome. The purpose of the present NMA was to investigate the relative effects of csDMARDs, GC, and placebo on radiographic joint destruction.

Method
We identified 31 RCTs investigating 13 csDMARDs, GC, and placebo used in monotherapy, and used WinBUGS software to conduct an NMA, metaregressions, and subgroup analyses for possible confounders. The percentage annual radiographic progression rate (PARPR) was the primary outcome, while the standardized mean difference was used in a sensitivity analysis.

Results
Leflunomide, sulfasalazine, and injected gold were more favourable than placebo and neither more nor less favourable than methotrexate. Although the effect size was equivalent with methotrexate, GC was not statistically better than placebo with the PARPR method. Azathioprine was less favourable than methotrexate and the remaining drugs were either not different from placebo or insufficiently investigated for robust conclusions.

Conclusion
Our study confirms that the present routine csDMARDs, methotrexate, leflunomide, and sulfasalazine, have inhibitory effects more favourable than placebo on joint destruction in RA.