Professor El-Omar has chosen Dr Alex Adams and Professor Jack Satsangi to do the next #GUTBlog. Dr Adams and Professor Satsangi are from the Translational Gastroenterology Unit, University of Oxford, Oxford, UK. The #GUTBlog focusses on the paper “Early management of acute severe UC in the biologics era: development and international validation of a prognostic clinical index to predict steroid response” which was published in paper copy in GUT in March 2023.
Dr Alex Adams and Professor Jack Satsangi write:
“Acute severe colitis (ASC) is a severe exacerbation of ulcerative colitis (UC) involving frequent bloody diarrhoea and signs of systemic illness necessitating hospital admission for intensive management. The criteria used to define ASC (more than six bloody motions over 24 hours together with fever, tachycardia, anaemia, or elevated CRP/ESR) were widely adopted following Truelove and Witts’ landmark study in 1955, since which parenteral corticosteroids have remained first-line therapy.
The current standard of care in the medical management of ASC has developed over the last 20-30 years; and now involves treatment with intravenous corticosteroids for the initial three days of admission, at which point response is assessed, and further medical rescue therapy or surgical intervention is considered if the response is insufficient. In practice, less than 50% of ASC cases demonstrate full response to first-line steroid treatment, meaning there may be an opportunity to improve both patient experience and possibly outcomes if an early determination of response to corticosteroids can be made.
An episode of ASC may be the first presentation for an individual with UC (approximately 30% of the cases in our study), meaning that they have no previous exposure to IBD therapies. However, in the era of biologics, many patients have been exposed to biological therapies prior to admission with ASC. Additionally, the widespread use of biologics as medical rescue therapy or as maintenance therapy in those responding to rescue therapy are all likely to have impacted outcomes both acutely and in the longer term.
Outcomes in the biologics era
We compared ASC outcomes within a single specialist centre (the Translational Gastroenterology Unit, John Radcliffe Hospital, Oxford, UK) between 1992-3, and 2015-19. Over this time the response to the initial management has not changed with 40-41% of patients responding to treatment with three days of IV corticosteroids. Outcomes after day 3 have improved however, with rates of emergency colectomy for ASC reducing from 29% to 15% (8-11% in our contemporaneous Australian and Indian cohorts in this study). This halving in colectomy rate coincides with a doubling of the rates of rescue therapy (from 27% receiving ciclosporin in 1992-3, to 54% equally split between ciclosporin and infliximab in 2015-19). Despite the pleasing symmetry we must remember that other changes in ASC management will have contributed to the change in outcomes such as adoption of evidence-based management protocols.
Outcomes following the acute admission have also improved, with colectomy rates within a year halving in Oxford (from 43% to 21%) and readmission with a further episode of ASC reduced from 35% to 12%. Despite the overall reduction in colectomy within a year, in both periods patients who required medical rescue therapy had an 8-fold increased risk of colectomy within a year compared to patients who responded completely to IV corticosteroids. A response to rescue therapy does not end the increased risk to the patient and there is ample room to improve outcomes for this group.
Early identification of steroid non-responders
To maximise usefulness to clinicians we aimed to create a simple clinical index using points assigned to routinely measured clinical parameters surpassing certain thresholds, rather than a more complex regression model or machine learning model. We restricted our analysis to parameters measured at admission thereby looking to predict rather than assess response to steroids, and using multiple international cohorts allowed rigorous validation of our findings.
Using the Oxford 2015-19 cohort we found the best index to predict steroid response consisted of four points for each of CRP≥100 mg/L, albumin≤25 g/L, UCEIS≥4, and UCEIS≥7. When testing this in cohorts from Australia and India we found that scores of 0 and 4 once again correlated with complete response and non-response to steroids respectively. Our proposed implementation for this index is that a score of 3 or 4 should be used as an indication for early consideration of medical rescue therapy or surgery as it is associated with a steroid failure rate of 84% with an odds ratio of 12, and would prevent one delayed treatment for every 8 patients assessed. We would recommend lower scores be treated routinely, with a score of zero being particularly reassuring as we found that all patients scoring zero responded to steroids.
Many studies dating back to the 1970s have identified features associated with poorer outcomes in ASC, including inflammatory markers in the blood and stool, endoscopic severity, and systemic upset; for example, the 1992-3 Oxford cohort in this study was the source for the commonly used day 3 CRP and stool frequency criteria to judge steroid response. One criticism that could be levelled at these past efforts, is that insufficient effort has been spent on validation of previous results. We hope that the evidence we have gathered across three continents, demonstrating consistent results despite differences in genetic background, age, disease duration, and previous treatment exposure, will allow this index to be translated into clinical practice more fully than previous findings. Whether through changing diet and environment, or increased recognition, IBD is an increasingly global condition and while no study can be expected to sample a truly representative global population, testing findings across distinct populations is an increasingly important step in validation.
Early identification of patients likely to require medical rescue therapy or surgery will be of immediate benefit in terms of patient counselling and resource management, though further work to optimize the management of these high-risk patients will be required. We suggest that trials to assess the use of alternative strategies to corticosteroid monotherapy in patients who are predicted to be high-risk of steroid non-response on admission using this index are now warranted.
Note:
Dr Alex Adams, Dr Vipin Gupta and Dr Waled Mohsen are joint first authors of the paper.