Impact of BMI on response to Janus kinase inhibitors in rheumatoid arthritis: an individual patient data meta-analysis of randomised controlled trials.

Obesity is common in patients with rheumatoid arthritis and is associated with poorer disease outcomes. We aimed to evaluate the association between BMI and clinical response to Janus kinase (JAK) inhibitors in patients with rheumatoid arthritis using individual patient data from the JAK inhibitor randomised controlled trial programmes. In this individual patient data meta-analysis, ClinicalTrials.gov was searched from database inception to Jan 13, 2025, for phase 3, randomised controlled trials of tofacitinib, baricitinib, upadacitinib, and filgotinib in adult patients with rheumatoid arthritis, with a placebo or active comparator. Individual patient data were obtained via the Vivli data-sharing platform. The primary outcomes were a 20% improvement based on American College of Rheumatology (ACR) core set variables (ACR20) and Disease Activity Score based on the evaluation of 28 joints and C-reactive protein (DAS28-CRP), at the trial-defined primary efficacy timepoint. BMI was analysed as a continuous and categorical variable using one-stage mixed-effects models, and validated using a two-stage individual patient data meta-analysis. Risk of bias was assessed using the Cochrane Risk of Bias 2 tool. There was no involvement of people with lived experience in the study design or conduct. This study is registered with PROSPERO (CRD420251180005). 16 trials contributed individual patient data from 11 883 participants (9588 [80·7%] women and 2293 [19·3%] men, and two with missing sex data). Mean age was 52·9 years (SD 12·3). 8493 (71·5%) of 11 883 participants were White, 2142 (18·0%) were Asian, 439 (3·7%) were Black, and 809 (6·8%) were of other ethnicity. 7765 received tofacitinib, upadacitinib, or baricitinib. Filgotinib trials were excluded due to insufficient variables. Median BMI was 26·8 kg/m2 (IQR 23·2-31·3) and 3676 (30·9%) patients had class 1-3 obesity. Higher BMI was associated with reduced JAK inhibitor efficacy across all endpoints. Compared with patients with healthy weight, adjusted relative risk for ACR20 response was 0·94 (95% CI 0·91-0·98) for patients with overweight (25 to <30 kg/m2), 0·92 (0·89-0·96) for class 1 obesity (30 to <35 kg/m2), 0·88 (0·81-0·96) for class 2 obesity (35 to <40 kg/m2), and 0·78 (0·71-0·85) for class 3 obesity (≥40 kg/m2). For DAS28-CRP, adjusted mean differences compared with healthy weight were 0·14 (95% CI 0·06-0·22) for overweight, 0·21 (0·12-0·31) for class 1 obesity, 0·30 (0·16-0·44) for class 2 obesity, and 0·54 (0·37-0·70) for class 3 obesity. No corresponding BMI gradient was observed in the placebo group. Between-study heterogeneity was low to moderate (I2=0-40%). Overall risk of bias in individual studies was low. Higher BMI was associated with poorer response and a reduced treatment benefit with JAK inhibitors, with obesity acting as an effect modifier. These findings support integration of weight management into routine care for rheumatoid arthritis and improved BMI representation in future trials. None.