Vedolizumab Efficacy Exposure-Response Relationship for Ulcerative Colitis Patients (GEMINI I) Based on Causal Inference Analysis

Abstract

Introduction: Vedolizumab (VDZ) is efficacious for the treatment of moderate-severe active ulcerative colitis (UC). It is hypothesized that individuals with high drug clearance may be less likely to achieve positive efficacy outcomes. The aim of this analysis was to evaluate if higher induction exposures of VDZ are associated with improved efficacy outcomes in UC. Methods: Data from a Phase 3 VDZ clinical trial (GEMINI I) included week 6 outcomes, Partial Mayo Score (PMS), Clinical Response (CR), and Clinical Remission (RM), individual-specific covariates, and individual-predicted plasma VDZ concentrations and clearances based on a prior population pharmacokinetic (PK) analysis. The limited dose-ranging information in this study yielded the potential for confounded causal inference about the exposure-response (E-R) relationship, and therefore, a propensity score based case-matching analysis was conducted. A logistic propensity score model was fit to the collection of treated subjects and all controls using all measured covariates as predictors. Matches between active treatment subjects and controls were based on a propensity score caliper of 0.2 times the standard deviation. Quartiles of clearances and predicted Week 6 and steady-state trough VDZ concentrations were compared with outcomes for both unmatched and case-matched data. Results: E-R relationships for all exposure metrics and each outcome were evident in the quartile analysis of the raw data, although relationships were more robust for Week 6 PMS and CR endpoints than for RM. After case-matching adjustment for potential confounding, a clear E-R relationship was still evident for the PMS and CR endpoints. VDZ clearance >0.14 L/day was associated with diminished efficacy outcomes (PMS decrease 37.1 mcg/mL, Week 14 trough, >18.4 mcg/mL, steady state trough >12.7 mcg/mL. Conclusion: The estimated E-R relationships will inform the design of future studies evaluating the impact of dose individualization on clinical outcomes in UC patients with high clearance.