P688 Comparative Effectiveness of Ustekinumab Versus Vedolizumab in Treating Crohn’s Disease After Failure of Anti-TNF Agents: Real-World Evidence

Abstract

Abstract Background Despite effective anti-TNF agents in treating Crohn’s disease (CD), some recipients experience primary or secondary non-responses, requiring alternative options. Ustekinumab and vedolizumab have not been compared in randomized controlled trials (RCTs) among CD population. Thus, we used real-world data to compare ustekinumab and vedolizumab at 12 weeks after failure of TNF inhibitors. Methods A retrospective study was conducted at a tertiary hospital in Dammam, Saudi Arabia. Patients with CD who had not responded to anti-TNF agents and had never been exposed to vedolizumab and/or ustekinumab were included. Children ≤ 18 years of age, naïve CD patients, CD patients using only anti-TNF agents, and patients who lost follow up were excluded. Primary endpoints were clinical improvements, which were measured by Harvey-Bradshaw Index scores at 12 weeks, and clinical remission, which was measured as an ordinal outcome. Remission clinically, biochemically, and endoscopically; clinical response; cumulative steroid dose; and corticosteroid-free days were secondary endpoints. Using probabilistic Bayesian models and proportional odds models, we analyzed outcomes, and the posterior distribution was used to calculate the probability of treatment effectiveness. A national institutional review board approved the study. Results Five hundred forty-six patients received biological agents for inflammatory bowel disease, of whom 101 received biological agents for CD; 71 patients received ustekinumab, and 30 patients received vedolizumab. The baseline characteristics were similar except for perianal disease, which was frequently reported in ustekinumab arm (P = 0.006). Most of the patients were male (54.5%), with a median age of 32 (IQR: 26.0-38.0). Most patients (51.5%) had stricturing disease in the Ileocolonic site (70.3%). For ustekinumab, the median HBI score was 5 (IQR: 3.0 -8.0) whereas for vedolizumab, it was 5 (IQR: 4.0 -7.0). In table 1, a Bayesian multivariable proportional odds model revealed a 40% reduction in median HBI scores at 12 weeks favoring ustekinumab, with an aOR of 0.60 (95% CI: 0.25 to 1.31) and a probability of effectiveness of 75% for ustekinumab. At 12 weeks, the ordinal outcome scale was 39% lower for ustekinumab arm, with an aOR of 0.61 (95% CI: 0.26 to 1.35) and a probability of effectiveness of 73% for ustekinumab arm. Secondary endpoints showed favorable results for ustekinumab reaching up to a 90% probability of effectiveness. Conclusion Among CD patients who failed anti-TNF therapies, ustekinumab was more effective than vedolizumab. To validate these results and determine these treatments’ relative efficacy in managing CD, further investigations, including prospective studies and RCTs, are essential.