P546 Comparative efficacy and safety of upadacitinib versus ustekinumab as induction therapy in patients with moderately to severely active ulcerative colitis: A matching-adjusted indirect comparison

Abstract Background Upadacitinib (UPA), an oral Janus kinase inhibitor, and ustekinumab (UST), a parenteral IL-12/23 are approved therapies for patients (pts) with moderately to severely active Ulcerative Colitis (UC). In the absence of head-to-head data, we conducted placebo (PBO)-anchored matching-adjusted indirect comparison (MAIC) compare efficacy safety UPA vs UST during induction. Methods Pts received 45 mg oral, once daily 8 weeks or PBO in phase 3 studies U-ACHIEVE (NCT02819635) U-ACCOMPLISH (NCT03653026), 6 mg/kg IV at week 0 8-week follow up study UNIFI (NCT02407236). Efficacy outcomes were stratified by previous biologic exposure (ie, no prior [bio-naïve] inadequate response, loss intolerance biologics [bio-failed]). Baseline characteristics age, gender, extent duration disease, total Inflammatory Bowel Disease Questionnaire score, Mayo high-sensitivity CRP, faecal calprotectin levels, weight, UC medication use from trials weighted match trial. Week clinical remission per full score (FMS; FMS ≤2 subscore >1), response (decrease ≥3 points ≥30%, plus decrease rectal bleeding [RBS] ≥1 absolute RBS 1), endoscopic improvement (endoscopic histologic-endoscopic mucosal (HEMI). Safety evaluated overall population adverse events (AEs) serious AEs (SAEs), comprising only assessments consistently collected across trials. Numbers needed treat/harm (NNT/NNH) calculated as inverse difference proportions achieving efficacy/safety between UST. Results The MAIC included 824 625 pts data 839 639 data. A greater proportion receiving bio-naïve bio-failed groups achieved remission, improvement, HEMI after weighting (p≤0.01, Table 1) NNTs <10. pts, differences 0.172, 0.146, 0.312, 0.255, respectively, 0.105, 0.235, 0.135, 0.118 pts. SAEs not statistically different Conclusion During induction, similar was observed based on MAIC. Potential bias due unobserved confounders may exist methodology. Further analyses assess longer-term warranted.