Acute Severe Ulcerative Colitis: An International Delphi Consensus on Clinical Trial Design and Endpoints

Abstract

Background & Aims

Interventional clinical trials in ASUC are characterised by substantial heterogeneity due to a lack of consensus in several key areas of trial design – this impedes clinical research efforts to identify novel therapies. The objective of this initiative was to achieve the first consensus and provide clear position statements on ASUC trial design.

Methods

A modified Delphi consensus approach was employed with a panel of twenty clinicians with international representation and expertise in ASUC trial design and delivery. Agreement was defined as at least 75% of participants voting as ‘agree’ with each statement.

Results

In total, thirty statements achieved consensus and were approved. Statements centred on proposing suitable eligibility criteria (disease extent, disease severity, prior therapy exposure), optimising trial design (randomisation, stratification, corticosteroid handling, timing of assessments), and recommending primary and secondary endpoints alongside defining key efficacy outcomes (clinical and endoscopic reponse and remission, treatment failure, quality of life).

Conclusions

The expansion of drugs to treat moderate-severe UC over the past decade, particularly the rapidly acting Janus kinase inhibitors, is promising and has reignited the interest in identifying suitable therapeutic candidates for ASUC. Clinical trials in this high-risk population are challenging to conduct and this consensus provides a framework for future trials to advance drug development.

Introduction

Acute severe ulcerative colitis (ASUC) is a potentially life-threatening medical emergency affecting up to 25% of patients with ulcerative colitis (UC) at some point during the disease course. It remains associated with a 10–15% risk of colectomy and a 1% risk of mortality.^1–4 Effective and timely medical management is necessary to reduce the inflammatory burden rapidly and minimise these risks.

In contrast to the marked expansion of advanced therapies in the outpatient setting for moderate-severely active UC, there is a paucity of randomised trials in ASUC to guide management for this high-risk subgroup. For example, the initial treatment paradigm for the induction of remission with intravenous corticosteroids remains unchanged since the landmark randomised placebo-controlled trial (RCT) of cortisone almost 70 years ago.⁵ Options for medical rescue or salvage for steroid non-responders have remained limited to ciclosporin and infliximab for the past two decades – both of which have limitations in clinical practice. For those who do respond to inpatient steroid treatment, the first RCT to investigate treatment failure during maintenance treatment was only published in 2024.⁶ These realities underscore the inherent difficulties in conducting clinical trials in patients with ASUC.

To understand the clinical trial landscape for hospitalised UC flares, we recently conducted a systematic review that identified 33 RCTs investigating the safety and efficacy of 23 distinct interventions.⁷ This review identified substantial heterogeneity in definitions, intervention designs, end-points, timing of assessments, and disease activity indices used.⁷ This heterogeneity across prior trials is due to the lack of standardisation for definitions of disease severity for trial enrolment and for definitions of treatment response, remission, and treatment failure. The objective of this initiative was to develop a consensus and provide clear position statements on these key areas to optimise the conduct of future ASUC trials. This includes early intervention ASUC trials (prior to patients being declared steroid non-responders) and trials of rescue therapy (in the event of steroid failure).

Methods

Setting and Participants

An international consensus meeting was held virtually on July 3^rd, 2024. A global representation of twenty clinicians across fourteen countries and five continents agreed to participate in this initiative and were selected through purposive sampling based on their expertise in ASUC and clinical trials.

Statements and Expert Consensus

Results from the systematic literature review were recently published and the findings were shared and discussed among the meeting participants prior to the consensus.⁷ Members of the steering committee (SH, VJ, BES, PSD, SS, SD, and LPB) formulated and refined a total of 36 statements in advance of the meeting. We employed a modified Delphi consensus approach with two rounds of voting through a virtual platform.⁸ New statements could also be suggested and voted for during the meeting.

In the first round, the panel members were presented with the statement and asked to rate degree of agreement based on a five-point Likert scale: fully agree, mostly agree, neutral, slightly disagree, or mostly disagree. Statements were considered approved if at least 75% of participants fully agreed or mostly agreed. In the second round, statements not unanimously accepted were rediscussed, rephrased, and re-voted on. If no consensus was reached in the second round of voting, the statement was excluded. At each round, panellists were independent and blinded to the votes of others to minimise the potential for individuals to influence others. All participants were invited to vote, although it was not mandatory. Two gastroenterologists who chaired the meeting (SH and LPB) did not vote. After the second round, the statements that met the pre-defined consensus threshold were used to form the basis of this manuscript. All participants were involved in drafting the manuscript and approved its final version.

Results

In total, thirty statements (out of 36 initially proposed) achieved consensus and were approved following this meeting. Of these, fourteen statements were discussed and reformulated after the first round of voting. Two statements were merged with other statements as they addressed similar themes. Four statements did not achieve consensus after two rounds of voting and were excluded (Supplementary Table 1). No new statements were proposed. Figure 1 displays a schematic of the modified Delphi process.

Figures.

Flowchart of modified Delphi process

Eligibility Criteria

Inclusion and exclusion criteria are important components of ASUC clinical trials to define the patient population under investigation. Eligibility criteria should ensure safety, and balance the need to minimise heterogeneity that can mask the findings of the investigational medicinal product (IMP), and the need to produce data that are generalisable to a broader patient population. Consensus was achieved for including patients with colonic inflammation of at least 15 cm from the anal verge. There was also agreement that to be enrolled into an ASUC trial, patients should be unwell enough to require hospitalisation, require treatment with intravenous steroids (according to the treating clinician), and have severe disease defined by the Truelove and Witts criteria. The role of biomarkers in defining disease severity was discussed and recommended an elevated CRP (above the local reference range) within 48 hours of screening; a similar proposed statement for faecal calprotectin did not achieve consensus. Additionally, the importance of endoscopic assessment to include patients with a Mayo endoscopic subscore of 2 or more prior to trial enrolment also received consensus. Table 1 summarises the scoring systems included in the approved consensus statements. Table 2 highlights the list of statements that achieved consensus for eligibility criteria. A detailed rationale for the inclusion of these statements is provided in the Supplementary Material and covers areas such as ASUC diagnosis, disease extent, disease activity, disease severity, and endoscopic assessment for eligibility.

Table 1:

Description of scoring systems and definitions used in approved statements

Scoring system	Year Developed	Description	Commonly used definitions in prior ASUC clinical trials
Truelove and Witts5	1955	Stratifies disease severity based on stool frequency and presence of ≥1 features of systemic disturbance (fever ≥37.8, pulse >90 bpm, haemoglobin ≤10.5 g/dL, erythrocyte sedimentation rate ≥30 mm/hour)	ASUC defined as stool frequency ≥6 with blood together with ≥1 positive measures of systemic involvement
Lichtiger index33	1990	Stratifies disease severity based on eight clinical parameters with a maximum score of 21. Measured components include stool frequency, nocturnal diarrhoea, rectal bleeding, faecal incontinence, abdominal pain or cramping, abdominal tenderness, general well-being, and need for anti-diarrhoeal drugs. Also known as the modified Truelove and Witts severity index.	ASUC defined as a Lichtiger score of >10. Treatment response defined as a score a score <10 points on two consecutive days, or a reduction of ≥3 points compared to baseline.
Mayo score28	1987	Standardises disease severity and assesses response to treatment over time. Also known as disease activity index. Comprised of 4 parts: stool frequency (SF), rectal bleeding (RB), endoscopic severity (ES) and physician’s global assessment (PGA), each scored from 0–3. There have been several iterations of this score. A patient-reported outcome-2 combines the assessment of SF and RB. A modified Mayo score combines the assessment of SF, RB, and ES. A partial Mayo score combines the assessment of SF, RB, and PGA.	PRO2: remission defined as stool frequency of ≤2 with an RB sub score of 0. Modified Mayo scores and partial Mayo scores were scarcely used in ASUC trials.7
Ulcerative colitis endoscopic index of severity	2012	Accurately predicts overall assessment of endoscopic severity. Assesses degree of three endoscopic features that include vascular pattern, mucosal bleeding, and erosions and ulcers.	Not previously used in an ASUC trial
Oxford criteria13	1996	Predicts the need for colectomy on day 3 after intensive treatment with intravenous corticosteroids. Of patients with ≥8 stools on that day, or a stool frequency between 3–8 with a CRP > 45 mg/l, 85% require colectomy.	Oxford criteria used to enrol patients into trials assessing safety and efficacy of salvage therapy.

Abbreviations: ASUC, acute severe ulcerative colitis; bpm, beats per minute; CRP, C-reactive protein; dL, decilitre; ES, endoscopic score; g, gram; IV, intravenous; mcg, microgram; PGA, physician’s global assessment; PRO, patient reported outcome; RB, rectal bleeding; SF, stool frequency.

Table 2:

List of approved statements achieving consensus on eligibility criteria

Number	Proposed Statement	Consensus (Voters n/N)
Eligibility Criteria
1	Eligible patients should have an established diagnosis of UC, or a high index of suspicion of UC at first presentation	93% (14/15)
2	Patients with a clinical suspicion of infection may be enrolled into an ASUC clinical trial where the physician suspects UC is the most likely diagnosis. Patients subsequently found to have an infection should be excluded.	80% (12/15)
3	Eligible patients should have colonic inflammation extending for ≥15cm from the anal verge	87% (13/15)
4	Eligible patients should have severe disease as defined by the Truelove and Witts criteria i.e., ≥6 stools with blood and one or more of following: haemoglobin <10.5 g/dL, ESR>30 mm/h (or CRP >30), fever >37.8°C, or tachycardia >90/min.	80% (12/15)
5	Eligible patients should have a requirement for intravenous corticosteroids in the judgement of the treating clinician	80% (12/15)
6	Eligible patients should be hospitalised with clinically confirmed or suspected ASUC and a stool frequency ≥6/day and an elevated CRP within 48h of screening but the optimal CRP cut off level remains to be determined	100% (15/15)
7	For ASUC trials of early intervention, prior to corticosteroid failure, there should be confirmation of Mayo endoscopic sub-score ≥2 level of disease activity beyond rectum. Endoscopic assessment within 28 days of admission is acceptable if no major treatment adjustment has been made. The index scope should report both Mayo score and the UCEIS.	87% (13/15)
8	For ASUC trials of medical salvage, following intravenous corticosteroid failure, there should be confirmation of Mayo endoscopic sub-score ≥2 level of disease activity beyond rectum with an in-hospital endoscopy prior to trial enrolment. This index scope should report both Mayo score and the UCEIS.	80% (12/15)
9	Populations exposed to multiple prior advanced therapies should be included when considering eligibility for ASUC trials	93% (14/15)

Abbreviations: ASUC, acute severe ulcerative colitis; cm, centimetre; CRP, C-reactive protein; dL, decilitre; ESR, erythrocyte sedimentation rate; g, gram; UC, ulcerative colitis; UCEIS, ulcerative colitis endoscopic index of severity.

Trial Design

Randomisation and masking

Table 3 shows the list of statements reaching consensus for trial design. Stratified randomisation first groups patients into strata according to clinical features that may influence outcome risk, and then assigns patients to treatment according to separate randomisation schedules.⁹ Prior or current use of advanced therapies, such as monoclonal antibodies and/or small molecule inhibitors, is an important factor that should be considered when stratifying randomization (93% agreement). As part of this statement, there was also consensus for not limiting multiple advanced therapy-exposed populations for ASUC trial inclusion. Stratification by pre-hospitalisation systemic corticosteroid use was considered less straightforward due to varying regimens and the high prevalence of outpatient steroid use. Nonetheless, separating regimens into high dose versus low dose (e.g., doses above and below 20 mg prednisolone or its equivalent) over a specified period may be one way to circumvent this if trialists deem this an important stratum to incorporate. Although the optimal number of strata for an ASUC trial is informed by study size, the panel highlighted the importance of keeping the number of variables to a minimum to reduce the challenges of balanced recruitment across treatment arms.

Table 3:

List of approved statements achieving consensus on study design

Number	Proposed Statement	Consensus (Voters n/N)
Study Design
10	Randomisation should be stratified by prior biologic and small molecule use	93% (14/15)
11	Participants and investigators should ideally be masked to treatments assigned for the duration of the trial	100% (15/15)
12	Dose and duration of intravenous steroid therapy should be clearly documented, and sensitivity analysis performed to account for variability in the corticosteroid regimens	80% (12/15)
13	Participants who respond to the treatment should be switched to oral corticosteroids and fixed tapering regimens for the maintenance period should enable steroid withdrawal within eight weeks	93% (14/15)
14	Non-response to intravenous corticosteroids should be assessed 48 hours after steroid initiation (on day 3) and should be defined by a stool frequency of >8 times/day or a stool frequency of >3 times/day and a CRP >45 mg/L (Oxford criteria).	87% (13/15)
15	For trials investigating the efficacy of salvage therapy following sub-optimal intravenous corticosteroid response, randomisation to an active comparator should be infliximab or ciclosporin	100% (15/15)
16	Unless there are absolute contraindications, all patients should receive thromboprophylaxis throughout hospitalisation	100% (15/15)
17	There should be representation from patients with lived experience of ASUC on trial committees to aid trial design and delivery and to ensure patient acceptability	87% (13/15)

Abbreviations: ASUC, acute severe ulcerative colitis; CRP, C-reactive protein; L, litre; mg, milligram.

Only half of prior ASUC trials were conducted where both subjects and investigators were blinded to the group allocation. Though challenging to do as the potential outcomes of ASUC are serious, participants, clinicians, outcome assessors, and data analysts should ideally be masked to treatments assigned for the duration of the trial to minimise bias and maximise the validity of the results (100% agreement). However, if an open-label design is required, best practice strategies to reduce bias should be followed.¹⁰

Corticosteroid handling and response

There was variability with inpatient corticosteroid regimens in previous ASUC trials, which may potentially influence safety and efficacy outcomes.⁷ Hence, the dose, type, and duration of intravenous steroid therapy should be clearly documented, and sensitivity analysis performed to account for any variability (80% agreement). Similarly, outpatient steroid tapering in maintenance trials also varied widely from no comment on how steroids were managed to adaptive tapering at physician discretion and fixed tapered regimens. Adaptive steroid tapering has been shown to impact trial interpretability and patient outcomes, with lower steroid-free remission rates compared to those with fixed regimens.¹¹ Subjects who respond to the treatment should, therefore, be switched to oral corticosteroids with fixed tapering regimens that enable steroid withdrawal within eight weeks during the maintenance period (93% agreement). Recent evidence suggests comparable clinical outcomes with a rapid steroid taper.^6,12 This approach also reduces the risks associated with long-term steroid exposure and offers a more pragmatic approach consistent with routine practice if patients are starting advanced therapies post-discharge.

The gold standard approach of treating ASUC with intravenous steroids reduces mortality and substantial morbidity.⁵ Thus, in early intervention ASUC trials, it should be used as the baseline therapy with the intervention under study used as an adjunct. Establishing intravenous corticosteroid non-response for ASUC trials of rescue therapy should be assessed 48 hours after steroid initiation (on day 3). Non-response should be defined by a stool frequency of >8 times/day or a stool frequency of >3 times/day and a CRP >45 mg/L (Oxford criteria) (87% agreement).¹³ The limitations of the Oxford criteria were acknowledged – these criteria were developed in the pre-biologic era and may not be relevant to the modern day patient population. In the post-biologic era, the Oxford criteria had a sensitivity, specificity and positive predictive value of 62%, 64% and 56%, respectively, in predicting steroid failure.¹⁴ Further, the criteria are based on a CRP measurement and is less reliable in patients that are CRP non-producers. Nonetheless, it is the most established tool available and the day 3 timepoint is important to determine eligibility and randomisation for clinical trials evaluating the efficacy of medical salvage therapy following sub-optimal steroid treatment. For rescue trials, randomisation to an active comparator should be with a drug with proven efficacy in this setting, which presently is limited to infliximab or ciclosporin (100% agreement). More recent scoring systems predicting steroid failure at admission (such as the ADMIT-ASC index or the ACE index) were based on retrospective cohorts and may have a place in future ASUC trial designs, but their utility in the trial setting requires further confirmation before they can be recommended for this purpose.^15,16

The use of systemic corticosteroids further compounds the increased risk of venous thromboembolism driven by hospitalisation and the high inflammatory burden from uncontrolled IBD.^17,18 Unless absolutely contraindicated, there was universal agreement that all patients should receive thromboprophylaxis throughout their inpatient stay (100% agreement).¹⁸ Although some studies show that the risk of venous thromboembolism persists after discharge, the panel acknowledged that there was a lack of data in this area and a wide variability in clinical practice with respect to thromboprophylaxis to make a position statement about this for ASUC trials.^18,19

Patient and public involvement

Patient and public involvement should be actively sought during the design, development, and delivery of clinical trials.^20,21 This method entails collaborating with patients who are directly impacted by the research focus, to design, develop, and deliver the research, ultimately aiming to enhance outcomes. There are two main reasons for this: first, it improves the quality and relevance of research, and second, it upholds the ethical principle of involving patients in decisions that affect them.^20,21 To enable this, there should be representation from patients and carers with lived experience of ASUC on trial committees to gain valuable insights into the acceptability of various aspects of the study (87% agreement). The degree of patient involvement within a trial is dependent on prior experience and research training available.

Therapeutic Endpoints

The final section of the meeting focussed on obtaining consensus for suitable efficacy endpoints for future ASUC clinical trials evaluating novel therapies (table 4). Although endpoints may vary based on trial objectives (e.g. early intervention vs medical salvage) and trial design (induction vs. maintenance), they should be underpinned by similar efficacy assessment tools. For regulators, the ideal primary efficacy assessment tools for UC clinical trials to support marketing approval require both patient-reported and clinician-reported outcome measures.²² The duration of the induction and maintenance phases in an ASUC trial are variable and are contingent on the properties of the intervention utilised and trial objectives. However, given the risk of adverse postoperative outcomes, including the increased risk of mortality and morbidity with delayed surgery, an efficacy endpoint should occur within 14 days (ideally less than 10 days) to justify continued medical over surgical therapy.^23–25 The panel first focussed on appropriate definitions of clinical response and remission as secondary outcome measures and then the merits of combining these as a composite for suitable primary endpoints.

Table 4:

List of approved statements achieving consensus on trial endpoints

Number	Proposed Statement	Consensus (Voters n/N)
Therapeutic Endpoints
18	Clinical response, defined by resolution of rectal bleeding (Mayo rectal bleeding sub-score of 0) and improvement in stool frequency (at least a 1-point reduction in Mayo stool frequency sub-score) (assessed within 14 days of IMP administration) could be a suitable secondary endpoint for an ASUC trial	93% (14/15)
19	Clinical response, defined by a decrease >3 points in the Lichtiger score from baseline and an absolute score <10 for two consecutive days, (assessed within 14 days of IMP administration) could be a suitable secondary endpoint	93% (14/15)
20	Burden of disease activity, measured by daily assessment of rectal bleeding and stool frequency (PRO2) after IMP administration until endpoint assessment, could be a secondary outcome measure.	100% (15/15)
21	Burden of disease activity, measured by daily Lichtiger scores after IMP administration until endpoint assessment, could be a secondary outcome measure.	87% (13/15)
22	Validated measures of generic health-related quality of life (HRQoL) and IBD/UC- specific HRQoL, should be assessed at baseline, and at weeks 12 and 52 and included as secondary outcomes.	84% (13/15)
23	Steroid-free clinical remission, defined as a Mayo disease activity index score ≤2 with an endoscopic subscore ≤1, and a rectal bleeding subscore of 0 should be assessed at weeks 12 (day 90) and week 52 and included as a secondary outcome measure.	80% (12/15)
24	Endoscopic improvement, defined as a Mayo endoscopic subscore of ≤1 without friability, should be assessed at week 12 (day 90) and week 52 and included as a secondary outcome measure.	93% (14/15)
25	Colectomy-free survival during the index ASUC admission or within 30 days of discharge from index admission should be a secondary outcome measure for new ASUC therapies.	100% (15/15)
26	Colectomy-free survival should be assessed at weeks 12 (day 90) and 52, depending on the length of the trial, and included as a secondary outcome measure.	100% (15/15)
27	Acute severe UC trials should record post operative morbidity and mortality in those needing operative management	100% (15/15)
28	Adverse events and serious adverse events (SAE) should be assessed according to common terminology criteria for adverse events (CTCEA) for duration of IMP exposure in the trial.	100% (15/15)
29	Treatment failure at any time defined by the presence of at least one of: the absence of clinical response within 14 days of IMP administration, a relapse between day 7 and 90 (week 12) (Lichtiger score increase of ≥3 from the previous value lasting for ≥3 consecutive days and leading to treatment modification), absence of steroid-free remission at week 12 (day 90) and week 52 (total Mayo ≤2 with an endoscopic subscore ≤1), a severe adverse event leading to treatment interruption, colectomy, death.	80% (12/15)
30	Resolution of rectal bleeding (Mayo rectal bleeding sub-score of 0), improvement in stool frequency (at least a 1-point reduction in Mayo stool frequency sub-score) and avoidance of salvage medical/surgical therapy (assessed within 14 days of IMP administration) is a suitable primary endpoint for an ASUC trial.	86% (12/14)

Abbreviations: ASUC, acute severe ulcerative colitis; CTCEA, common terminology criteria for adverse events; HRQoL, health-related quality of life; IMP, investigational medicinal product; PRO, patient reported outcome; SAE, serious adverse events.

Clinical response and patient-reported outcomes

Measures of clinical response were based on improvement in stool frequency and the resolution of rectal bleeding, which have been key measures of ASUC treatment efficacy in clinical trials since the seminal RCT of cortisone in 1955.⁵ There are no validated patient-reported outcome measures in this setting, but these self-reported symptoms are included in the 6-point version (PRO2) of the Mayo score disease activity index, which correlates well with other disease severity measures.^26,27 The Mayo score is approved by regulators and is the most commonly used tool to support registration for outpatient UC clinical trials (Table 1).²⁸ Both symptoms are highly responsive, and the resolution of rectal bleeding by day seven has been shown to have high prognostic utility for post-discharge corticosteroid-free clinical remission and endoscopic improvement.²⁹ Therefore, the panel agreed that for clinical response, the resolution of rectal bleeding (Mayo rectal bleeding sub-score of 0) and improvement in stool frequency (at least a 1-point reduction in Mayo stool frequency sub-score) (assessed within 14 days of IMP administration) could be a suitable secondary endpoint for an ASUC trial (93% agreement). The Lichtiger score, which incorporates stool frequency and rectal bleeding alongside six other clinical parameters, has also been used as an early efficacy measure in prior ASUC trials. Clinical response, defined by a decrease of >3 points in the Lichtiger score from baseline and an absolute score of <10 for two consecutive days, could similarly be used as a secondary outcome (assessed within 14 days of IMP administration) (93% agreement). Due to their responsiveness, both the PRO2 and Lichtiger scores were deemed suitable for assessing the daily burden of disease activity until endpoint assessment, with an agreement of 100% and 87%, respectively.

Health-related quality of life (HRQoL) is negatively affected in patients with IBD, and its normalisation is a long-term treatment target.³⁰ As such, validated measures of generic HRQoL and IBD/UC-specific HRQoL, should be included as secondary outcomes and assessed at baseline, and at weeks 12, and 52, depending on the length of the trial (84% agreement).

Clinical remission and clinician-reported outcomes

The recommended definition for clinical remission was defined as a total Mayo disease activity index score ≤2 with an endoscopic subscore ≤1 without friability, and a rectal bleeding subscore of 0 should be assessed at weeks 12 (day 90) and 52 and included as a secondary outcome measure (80% agreement). This included a measure of corticosteroid-free remission (clinical remission without concomitant corticosteroid use in patients who were using corticosteroids at baseline) at both time points.

Members of the panel acknowledged that striving for endoscopic remission, typically defined as a Mayo endoscopic subscore of 0, may be unrealistic in ASUC trials given the presence of residual scarring seen with healing of severe inflammation and deep ulceration (though this could be considered an exploratory endpoint along with histological endpoints).³⁰ Hence, endoscopic improvement (Mayo endoscopic subscore ≤1 without friability) was thought to be a more appropriate objective secondary measure (93% agreement) at weeks 12 (day 90) and 52. In-hospital endoscopic assessment for response during the index admission was not deemed appropriate as it was very unlikely interventions would be responsive for endoscopic outcomes. For central reading, which was considered challenging to implement for screening and enrolment at the present time, blinded central reviews of endoscopy recordings at later timepoints should align with the currently accepted standard of registrational trials.

Colectomy-centred endpoints

Approximately 10–15% of patients will require surgery following an admission with ASUC.^1,4 Accurately capturing colectomy rates is an essential outcome in ASUC trials to determine the safety and efficacy of the intervention. There was common agreement that colectomy-free survival during the index ASUC admission or within 30 days of discharge from index admission should be a secondary outcome measure for new ASUC therapies (100% agreement). For medium to longer-term outcomes, colectomy-free survival should be assessed at weeks 12 (day 90) and 52, depending on the length of the trial, and included as a secondary outcome measure (100% agreement). Lastly, and alongside recording adverse events and serious adverse events, ASUC trials should record postoperative morbidity and mortality in those needing operative management, which was poorly recorded in prior trials (100% agreement). These should be collected using validated and structured indices, such as the Clavien-Dindo classification.³¹

Primary endpoints and composite scores

The use of composite outcomes, which combine several components into a single measure, was discussed for its significance as a primary endpoint for ASUC clinical trials. Trials that centre on the short-term efficacy outcomes of an intervention may consider the resolution of rectal bleeding (Mayo rectal bleeding sub-score of 0), improvement in stool frequency (at least a 1-point reduction in Mayo stool frequency sub-score) and avoidance of salvage medical/surgical therapy (assessed within 14 days of IMP administration) as a suitable primary endpoint for an ASUC trial. Using these components in isolation was considered problematic; firstly, treatment escalation thresholds are subject to inter-institutional and inter-individual variation and too subjective, and secondly, using colectomy rates would reduce statistical efficiency and increase the need for a sample size that would be unfeasible due to the low event rate in modern data series. ASUC trials seeking to evaluate medium to longer-term safety and efficacy outcomes beyond hospitalisation require a more comprehensive composite endpoint that includes subsequent relapse of disease, corticosteroid-free remission, and endoscopic improvement (80% agreement).

Conclusion

Hospitalised patients with ASUC represent a high-risk group where therapies have not evolved and kept pace with UC managed in the community. Promising drugs on the horizon, such as the rapidly acting Janus kinase inhibitors, need to be tested in well-designed clinical trials. The systematic review of ASUC trials that informed this consensus identified numerous definitions for disease severity, treatment response, and remission, alongside variations in trial design that impact interpretability, for example, corticosteroid handling. Such discrepancies impede clinical research efforts for drug development in this orphan population.³² While the lack of grading for the level of evidence for each consensus statement was one of the main limitations, this consensus process had two main strengths. First, this is the first consensus of ASUC trials of which the ultimate goal was to provide a more unified and evidence-based approach for designing and conducting interventional trials for ASUC and selecting meaningful therapeutic endpoints. Second, the process assembled an internationally represented panel of experts in trial design and delivery to provide a framework for future ASUC clinical trials.

Data Availability Statement

The data underlying this article are available in the article and in the online supplementary materials.