Long-term Outcome of Patients With Acute Severe Ulcerative Colitis Responding to Intravenous Steroids

Robert Salameh; Julien Kirchgesner; Matthieu Allez; Franck Carbonnel; Antoine Meyer; Jean-Marc Gornet; Laurent Beaugerie; Aurelien Amiot


Aliment Pharmacol Ther. 2020;51(11):1096-1104. 

In This Article


Herein, we studied the largest cohort of patients with ASUC defined according to the modified Truelove and Witts criteria who responded to IVS and who had long-term outcome data. After 5 years of follow-up, the probability of colectomy was less than 10%, while 60% of patients experienced clinical relapse due to failure of the chosen maintenance strategy. A lower rate of relapse was observed in patients with early response to IVS and in patients treated with anti-TNF rather than 5-ASA or immunomodulators.

ASUC is a severe and life-threatening condition.[18] Patients experiencing ASUC are at higher risk of colectomy, up to 40%.[2,19] In our study, the probabilities of colectomy were 9% and 12% at 5 and 8 years, respectively. These results are close to those observed in the overall population of patients with UC, even in the anti-TNF era.[20,21] Combining clinical and biological markers of severity has been used to predict the short-term risk of colectomy in patients with ASUC.[13,22–24] In our cohort, the risk of colectomy was only predicted by a Lichtiger index >12 on admission (HR = 10.12, 95% CI [1.30–78.55], P = 0.03). These data suggest that the risk of colectomy in patients with ASUC is highly conditioned by the responsiveness to steroids as well as the severity of the flare itself.

Unlike CD, where top-down or accelerated step-up strategies are recommended for preventing bowel damage, choosing the best maintenance therapy for UC may be controversial. ECCO recently considered that ASUC is an indicator of disabling disease and requires accelerated step-up strategies with thiopurine for patients naïve to immunomodulators and biological agents in thiopurine-refractory patients.[5] In our study, patients treated with 5-ASA as maintenance therapy had a higher rate of relapse over time compared with patients treated with immunomodulators and anti-TNF agents (Figure 1). A recent study has recently evaluated the benefit of immunomodulators in hospitalised patients responding to IVS irrespective of ASUC criteria and found no benefit of early immunosuppression on the risk of colectomy.[25] The benefits of anti-TNF over immunomodulators are questionable in this setting, but patients treated with biological agents were excluded from the latter study. To date, there is only one head-to-head study that has assessed the efficacy of the latter drugs to maintain remission after clinical remission was induced with steroids; it showed the superiority of the combination of infliximab plus thiopurine compared to either thiopurine or infliximab alone.[26] No data are currently available in the setting of patients with ASUC responding to IVS. In the present study, clinical relapses were observed in 42% and 60% at 1 and 5 years respectively. We showed that the use of anti-TNF as maintenance therapy after ASUC responded to IVS, irrespective of previous exposure to immunomodulators and/or biological agents, was associated with a lower risk of relapse over time. Further studies should assess the benefit of pre-emptive use of anti-TNF in patients experiencing ASUC. The benefit of infliximab in patients with ASUC responding to IVS is currently being evaluated in a randomised controlled trial in comparison with thiopurine (the ACTIVE trial, EUdraCT 2014-005212-42).

While two-thirds of patients responded to IVS for ASUC, most of the efforts have been made to identify patients with steroid resistance and colectomy during the index hospitalisation. In this setting, the Oxford index combines clinical (number of daily bowel movement) and biological markers (CRP), with a prediction accuracy of 85%.[13] Other indices have been developed during the past decades, with similar accuracy, but are less simple to use in daily clinical practice.[13,22–24] There is evidence that achieving complete clinical remission during the index hospital admission improves long-term outcome and delays the need for colectomy.[13,27] In a prospective series at the Oxford University Hospital, the risk of further relapse and further colectomy after ASUC responded to IVS was lower in the 42% of patients with complete response (≤3 stools per day without visible bleeding on day 7) than in the 29% of patients with partial response. Overall, the risk of colectomy was 40% at 6 months and 85% over the next decade in patients with partial response, compared with 15% in the complete responders.[27] In our study, the presence of an early response with fewer than six stools per day by day 3 and a partial Mayo Clinic score less than 3 by day 5 was associated with a lower risk of clinical relapse over time. It is conceivable that patients with incomplete early response to IVS may benefit from a top-down strategy that takes into account the subsequent maintenance therapy.

We have to acknowledge several limitations of the present study. First, this is a retrospective observational study in which the choice of maintenance therapy was at the investigators' discretion. However, we adjusted the multivariate models according to the previous exposure to immunomodulators and/or biological agents, and no difference was found when comparing the first and last halves of the 12-year inclusion period. Second, patients were recruited from four tertiary care IBD centres, which could have introduced biases in terms of patients' recruitment and choice of maintenance therapy. Indeed, only 142 of 347 (40.9%) patients with ASUC responded to steroids allowing hospital discharge and switch to oral steroids a lower proportion than expected.[4] However, we selected patients with ASUC according to the modified Truelove and Witts criteria rather than based on the Lichtiger index or the fact that patients received IVS, which allowed us to study patients readily identified as having ASUC.[18] Furthermore, we chose to exclude patients even in case of clinical response if they have been treated with second-line medical therapy and the absence of discontinuation of IVS to oral steroids due to partial clinical response. Third, there was no standardised protocol of follow-up or treatment escalation or de-escalation, but patients were systematically followed up with three to six visits by physicians. Our study also has strengths. Patients were recruited from a standardised hospital in-patient diagnostic dataset, allowing the identification of all consecutive patients with UC with hospitalisation of at least 48 hours and promoting generalisability of results to real-world patient populations.