Comparative Effectiveness of Bimekizumab in Psoriatic Arthritis: A Model-Based Meta-Analysis of American College of Rheumatology Response Criteria

Abstract

A model-based meta-analysis (MBMA) was conducted to compare the efficacy of bimekizumab with other psoriatic arthritis (PsA) treatment regimens using ≥ 20%/50%/70% improvements in American College of Rheumatology (ACR) criteria (ACR20/50/70) for patients with PsA. Forty-nine trials of 16 drugs were identified in the literature, comprising 21,340 patients. Trial-level covariates, including prior biologic use, concomitant methotrexate use, time since diagnosis, trial completion year, and active comparator were considered for exploratory models. The final model was selected using leave-one-out cross-validation (LOO CV) to assess predictive performance based on prespecified criteria. LOO CV was conducted for 15 trials; the final model demonstrated that 91.5% (952/1,040) of the observed treatment differences, and 96.1% of the observed ACR20/50/70 response rates were within the 95% prediction interval (PI). Median ACR50 response rates (95% PI) at week 16 in biologic-naïve patients were predicted to be 44% (40–49%) for bimekizumab 160 mg, among the highest of all treatments analyzed. Response rates for secukinumab 150 mg and risankizumab 150 mg were 28% (25–32%) and 27% (24–31%), respectively. The MBMA was also used to predict the probability of success (PoS) of potential head-to-head trials using ACR50 response as the end point with varying sample sizes: vs. secukinumab 150 mg, the PoS for bimekizumab 160 mg was 62% (N = 200) and 90% (N = 400). Versus risankizumab 150 mg, the PoS for bimekizumab 160 mg was 68% (N = 200) and 94% (N = 400). In summary, a predictive MBMA described ACR20/50/70 outcomes in PsA, allowing accurate and precise treatment comparisons and robust PoS calculations.