Comparative effectiveness of vedolizumab and TNF-antagonist therapy in ulcerative colitis: A multicentre consortium propensity scorematched analysis

Background: We aimed to compare the effectiveness of vedolizumab (VDZ) to tumour necrosis factor (TNF)-antagonist therapy for ulcerative colitis (UC). Methods: Using a multicentre, US-based consortium of UC patients treated with VDZ or TNF-antagonist therapy, we performed propensity score matching (1:1) accounting for age, sex, prior UC-related hospitalisation within the previous year, disease extent, disease severity, steroid refractoriness or dependence, and prior TNF-antagonist failure. Treatment response was categorised using the physician global assessment. Using Cox proportional hazard models, we compared cumulative rates of clinical remission (complete resolution of UC-related symptoms), steroid-free remission (on steroids at baseline, tapered off, no repeat steroid prescription for 4 weeks), and endoscopic healing (Mayo endoscopic subscore of 0 or 1). Hazard ratios (HRs) and 95% confidence intervals (CIs) are reported for VDZ compared with TNF-antagonist therapy. Results: The propensity score model accurately predicted treatment status (area under curve 0.73). Of 646 UC patients, 334 were included after matching (n = 167 VDZ,; 49% male; median age 36 years). After adjusting for concomitant steroid use, concomitant immunomodulator (azathioprine, 6-mercaptopurine, methotrexate) use, and number of prior TNF-antagonists used, VDZ-treated patients had statistically significant higher 12-month cumulative rates of clinical remission (54% vs. 37%; HR 1.54, 95% CI 1.08-2.18) and endoscopic healing (50% vs. 42%, HR 1.73, 95% CI 1.10-2.73). Cumulative 12-month rates for steroid-free remission were numerically higher for VDZ-treated patients, but not statistically significant (49% vs. 38%; HR 1.43, 95% CI 0.79-2.60), These findings were consistent when stratified by disease extent and prior TNF-antagonist exposure. Conclusions: After accounting for measurable disease and patientspecific characteristics that may impact biologic effectiveness, we observed that VDZ-treated UC patients had significantly higher 12-month cumulative rates of clinical remission and endoscopic healing, and numerically higher steroid-free remission rates, when compared with TNF-antagonist-treated patients. Randomised controlled trial data are needed to confirm these findings.