JAK Inhibitors in Rheumatoid Arthritis — Living Meta-Analysis (RapidMeta)

Abstract

Do trials in rheumatoid arthritis involving JAK inhibitors provide a significant clinical effect on the odds of an ACR20 response? Four placebo-controlled phase-3 randomised trials (ORAL Solo [tofacitinib], RA-BEAM [baricitinib], SELECT-NEXT [upadacitinib], and FINCH 1 [filgotinib]; 2,228 analysed participants) were combined in the RapidMeta rheumatology living review in a web-based, fully tracked pipeline. ACR20 responder counts from each trial's ClinicalTrials.gov results record were pooled as odds ratios by random-effects (REML) on the log scale, with a Hartung-Knapp-Sidik-Jonkman-adjusted interval and back-transformation to the original measurement. The pooled odds ratio was 3.40 (95% CI 2.89 to 3.99): ACR20 response was roughly 3.4 times more likely with a JAK inhibitor than placebo. Between-trial heterogeneity was absent (I2 = 0%, Cochran's Q(3) = 0.74, p = 0.86, tau2 = 0), and the 95% prediction interval (2.61 to 4.42) excluded no benefit. Results are registered into a reproducibility capsule containing a machine-readable config, an interactive reader, and a Vancouver reference pack. As a class-level living synthesis of one pivotal trial per agent, the dashboard does not capture dose-ranging, head-to-head or longer-term safety outcomes, and cannot substitute for adjudicated trial-level review of risk of bias, switched endpoints, or missing subgroup data.