P543 Patients in trials of moderate-to-severe UC who are exposed to placebo are more likely to suffer harm than those receiving active treatment: The lack of clinical equipoise in traditional study designs

Abstract

Abstract Background Ulcerative Colitis (UC) is a relapsing-remitting inflammatory disease of the colon and rectum. There are many challenges to the recruitment of eligible individuals into clinical trials for IBD, including the availability of multiple standard-of-care treatment options and patient refusal to participate in randomized placebo (PBO)-controlled trials due to fear of receiving PBO. Patient harm across clinical trials for moderate-to-severe UC has not been adequately examined. We sought to quantify the potential harm to patients recruited to PBO. Methods We reviewed phase 3 clinical trials of therapies for moderate-to-severe UC including anti-TNF-alpha (infliximab, adalimumab, golimumab), anti-integrin (vedolizumab), anti-interleukin (IL) 12/23 (ustekinumab), anti-IL23 (mirikizumab, guselkumab), Janus kinase inhibitors (tofacitinib, upadacitinib), and sphingosine-1 phosphate receptor modulators (ozanimod, etrazimod). We analyzed the proportions (# of pts. with events/cohort size) of adverse events (AEs), serious adverse events (SAEs), disease exacerbation AEs (DEAEs), and disease exacerbation SAEs (DESAEs). We calculated Number Needed to Harm (NNH = 1/Incidence rate in cohort) for PBO (NNHPBO) and experimental therapy (NNHRx) in each trial. Maintenance phase NNHs were categorized as treat-through (TT) or re-randomized (RR). Results We identified an increased probability of AEs in the PBO arm of 7/11 induction (NNHPBO= 2.18-11.06; NNHRx=2.47-16.52), 1/8 TT maintenance, and 5/11 RR maintenance (NNHTTPBO = 1.2-3.1; NNHTTRx =1.1-16.4; NNHRRPBO = 1.3-10.8). We identified an increased probability of SAEs for 8/11 induction (NNHPBO=14.5-58.0; NNHRx=22.13-101.0), 4/8 TT, and 5/11 RR (NNHTTPBO = 3.9-23.3; NNHTTRx =4.2-29.9; NNHRRPBO = 8.8-35.0). Patients were at increased probability of DEAEs in the PBO arm of 8/11 induction (NNHPBO=7.38-43.20; NNHRx=12.50-106.33), 7/8 TT, and 7/10 RR (NNHTTPBO = 3.0-11.1; NNHTTRx =4.3-15.0; NNHRRPBO = 2.7-25.2). DESAEs were at increased probability of harm in the PBO arm of 8/9 induction (NNHPBO=29.00-116.00; NNHRx=47.43-119.75), 3/5 TT, and 7/10 RR (NNHTTPBO = 14.4-52.0; NNHTTRx =13.2-230.0; NNHRRPBO = 11.2-25.2). Conclusion In PBO-controlled clinical trials of therapies for moderate-to-severe UC, there are increased AEs and SAEs for patients who receive PBO, especially for DEAEs and DESAEs. Induction treatment responders re-randomized for maintenance to PBO experienced greater harms (i.e. lower NNH) than TT PBO patients. These findings raise concern about clinical equipoise in moderate-to-severe UC clinical trials and support ongoing efforts to design trials with active comparator arms and adaptive or platform methodologies.