The Winning Line Up: Positioning Advanced Therapies for Ulcerative Colitis

In a span of less than 12 months, 3 agents were approved for moderate to severe ulcerative colitis (UC) worldwide—first, ozanimod, an S1P receptor modulator in May 2021; then filgotinib, a JAK1 inhibitor in November 2021 in Europe, and finally upadacitinib (UPA) in March 2022.¹ These approvals came after a several-year dearth of newly approved medications in the field, and after the same class of medications, anti-tumor necrosis factor agents, have dominated prescribing practices based on clinical response and remission rates for the last 20 years. As such, some of the most well-attended topics at leading inflammatory bowel disease (IBD) conferences explore positioning of advanced therapies for UC. These conversations also consume board rooms of pharmaceutical companies, insurers, and policy makers across the nation.

When it comes to positioning therapies in IBD, there are less than a handful of head-to-head trials directly comparing agents to one another with positive results in order to best answer this question: VARSITY² and SEAVUE³ [GARDENIA did not meet primary endpoint when comparing etrolizumab with infliximab monotherapy for moderate to severe UC].⁴ Both trials had somewhat surprising and paradigm-questioning results: VARSITY found vedolizumab superior to adalimumab for clinical remission and endoscopic improvement in moderate to severe UC, and SEAVUE found both agents (UST and ADA) to be highly effective across outcomes in bio-naive patients with moderate to severe Crohn’s disease. Finally, because the FDA requires demonstration of superiority of a new agent over placebo in order to be approved, as opposed to an existing therapeutic comparator, incentives are misaligned for a slew of additional direct-comparison trials to become available in the near future—companies may shirk the opportunity to go toe to toe with others.

In order to best digest the study published by Panaccione et al in the landscape of IBD therapies, we will briefly review the concepts of evidence-based medicine (EBM) and comparative effectiveness research (CER).⁵ EBM started in 1981 under the tutelage of David Sackett, and later Gordon Guyatt, in the 1990s, both of Canada’s McMaster University as a framework through which medicine could be practiced most efficiently in a nation with universal healthcare.⁶ CER, as per the Institute of Medicine, is used to “assist consumers, clinicians, purchasers, and policy makers to make informed decisions that will improve health care at both the individual and population levels.”⁷ A technique used in CER, network meta-analysis (NMA), allows for indirect comparisons to infer the overall efficacy of 1 drug compared with another. Using data from randomized controlled trials, researchers are able to estimate relative effectiveness between agents using networks—that is Drug A versus Drug C and Drug B versus Drug C, then connecting the dots between all 3.⁸ Relative ranks of different therapies are then described using probabilities known as “surface under cumulative rankings” or SUCRA. In certain scenarios, a higher value (closer to 1) is a favorable statistic for example clinical efficacy; in other, a lower value (closer to 0) is preferred for example safety.

In this study by Panaccione et al,⁵ authors set out to evaluate the efficacy and safety of advanced therapies for the treatment of moderate to severe UC. In addition to traditional division by induction and maintenance results, their analysis is unique in that it combined those time sequences to simulate absolute treatment efficacy—what patients may experience across time. Their data contribute meaningfully to the pool of NMAs in a space where inadequate data preclude (many) head-to-head comparisons. Two of the authors were tasked with evaluating the quality and appropriateness of inclusion of candidate trials using GRADE methodology and the Cochrane Risk of Bias tool. This technique ensures transitivity or comparability of a trial data to enhance the strength of conclusions suggested. As a mentor once said: “Garbage in, garbage out,” describing the reciprocal importance of ensuring quality data are used for these mathematical extrapolations.

Slightly less than half of potential trials made the cut (23/48)—notable exclusions included phase II data, the VARSITY trial and non-Mayo score-based RCTs. Study populations were similar, with minor heterogeneity in patient weight, disease duration, extent of disease, and concomitant medications. This trial placed limits on the induction phase to 6–10 weeks post-baseline and thus any patients who had responded to drug after the induction phase in the original RCT would be excluded from analysis. Endoscopic scoring was not uniformly performed by a central reader and there were slight differences in time frame for evaluation of induction response.

Panaccione et al’s⁵ network encompassed more than 8800 patients, split between biologic-naive (5080) and biologic-exposed (2648). This distinction is important as the discussion around medication choice for the newly diagnosed ­patient differs significantly from those who have tried numerous prior therapies. Singh et al had 3700 bio-naive and nearly 1600 bio-exposed patients.⁸ Additional patient networks published by Lasa and Burr numbered more than 10 000 and approximately 12 000 patients, respectively.^9,10

In this analysis, UPA, with a SUCRA of 99% and 97% for bio-naive patients for clinical response and clinical remission, respectively, was suggested as the top-ranking agent in both bio-naive and bio-exposed patients with UC in induction efficacy. In the maintenance phase in terms of clinical remission, tofacitinib—10 and 5 mg—and filgotinib had highest ranking probabilities with SUCRAs of 87%, 84%, and 79%, respectively; UPA came in at 72%, suggesting a strong class effect from the JAK family of small molecules. For bio-exposed patients, UPA had nearly double-digit odds ratios and SUCRAs >92% for each outcome tested: clinical response and remission and endoscopic improvement in both induction (45 mg dose) and maintenance phases (30 mg dose).

Despite the early concerns about risk of major cardiovascular events and venous thromboembolism that led to a black box warning for tofacitinib in 2021, as subsequently applied to JAK inhibitor class, these authors elaborated a 7380 patient-induction and 4841 patient-maintenance network , there were no worrisome comparative risk for this class—all of the therapies considered were similarly “low-risk”; and UPA had a relatively safer OR (0.2) versus placebo when considering discontinuation due to AEs in the induction phase with a statistically significant SUCRA of 90.1% and a similar finding for ustekinumab and vedolizumab during maintenance (0.2 and 0.4, respectively).

Critics of NMA point out the stringent inclusion criteria for inclusion in industry-funded trials can make generalizability limited. Furthermore, some suggest clinical trials have somewhat low external validity, by excluding those with mild disease, young patients (<18 yo), and those who have failed multiple biologics. Others point out having one-third of authors with overt ties to the manufacturer of the highest performing agent can pose a threat to bias in the study conclusions. However, as a field, IBD has long held a close relationship with industry, and as prescribers, we are tasked with evaluating evidence using our own critical lens and EBM training. One additional consideration inherent in the technique of NMA involves the assumption that only induction phase-responders are included in subsequent analyses; eliminating those who might become late responders, a pleasant surprise in clinical practice.

This article offers providers additional data points for their day-to-day practice of EBM for these with UC, making it easier to find a winning lineup for your roster. Using shared-decision making practices, patients and providers can choose the best drug for THAT patient using NMA such as this one. By combining this and others with patient preferences (eg, route of administration), clinical judgment and experience (eg, immunogenicity), as well as guidelines published by AGA, ACG, and ECCO, prescribers can have supporting evidence to gain insurance approval of the desired therapy without (too much) delay, an unfortunate circumstance in the American healthcare system that costs patients’ valuable time.¹¹

Conflicts of Interest

E.M.F. holds the position of Associate Editor for Crohn’s and Colitis 360 and has been recused from reviewing or making decisions for the manuscript. Abbvie—research support, speakers bureau; BMS—research support, speakers bureau, Janssen—research support, speakers bureau; Takeda—research support.