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

Patients and Methods

Study Population

From December 2017 to January 2019, we conducted a retrospective cohort study in four tertiary care centres in Paris-île-de-France: Henri Mondor, Bicêtre, Saint-Antoine and Saint-Louis University hospitals. All consecutive patients with hospital in-patients with ASUC treated with IVS between January 2006 and December 2017 were evaluated for inclusion. ASUC was defined according to the modified Truelove and Witts criteria: They all had ≥6 episodes of bloody diarrhoea per day and at least one of the following criteria: heart rate ≥90 bpm, temperature ≥37.8°C, haemoglobin ≤10.5 g/dL and/or C-reactive protein (CRP) ≥30 mg/L.[14] Exclusion criteria included failure of IVS leading to colectomy and/or second-line medical therapy with anti-TNF, vedolizumab and/or calcineurin inhibitors; introduction of anti-TNF, vedolizumab and/or calcineurin inhibitors before response to IVS could be assessed; evidence of intestinal superinfection with Cytomegalovirus, Clostridium difficile or other intestinal pathogens; and Crohn's disease (CD) or unclassified inflammatory bowel disease. Response to IVS was defined as (a) a decrease in partial Mayo Clinic score by at least 3 points by day 5 or by at least 30% of the baseline score, with a decrease of at least 1 point on the rectal bleeding subscale or an absolute rectal bleeding score of 0 or 1 from the baseline score at week 0, (b) a Lichtiger index <8 at day 5, (c) the absence of relapse during the first 2 weeks after discharge and (d) discontinuation of IVS to oral steroids with a schedule-tapering protocol.[15] Patients were recruited from the pharmacy database and/or the standardised hospital in-patient diagnostic dataset, including by searching ICD-10 codes. The protocol was approved by the Comité de Protection des Personnes (CPP no. 18–022). No commercial entity had any role in the study. All authors had access to the study data and reviewed and approved the final manuscript.

Patient demographic, clinical and endoscopic characteristics were collected from medical records and included sex, age, smoking habits, history of medical and surgical treatment of UC, family history of UC, date of diagnosis, behaviour of the disease prior to admission, and the dates of admission and discharge. Endoscopic assessment was collected at baseline including either flexible sigmoidoscopy or total colonoscopy. Disease extent was evaluated according to the Montreal classification and collected from medical records.[16] Disease activity was assessed at baseline using (a) total and partial Mayo Clinic scores and Lichtiger index for clinical activity; (b) CRP (mg/L), haemoglobin (g/dL), leucocyte count (/109/L), platelet count (/109/L) and albumin (g/dL) for biological activity; and (c) Mayo Clinic endoscopic subscore and the Ulcerative Colitis Endoscopic Index of Severity (UCEIS) for endoscopic activity.[7,15,17] Mayo Clinic endoscopic subscore and UCEIS were collected from medical records. The calculation of UCEIS has been performed retrospectively based on endoscopic pictures and/or reports in patients included before 2014. Intensive medical therapy consisted of intravenous methylprednisolone at a dose of 0.8 mg per kg per day with fluid and electrolyte resuscitation and thromboprophylaxis with subcutaneous low-molecular-weight heparin. Concomitant use of 5-aminosalycilic acid (5-ASA) and/or steroid enemas was permitted and recorded. The choice of maintenance therapy at discharge was not standardised and was only based on the investigator's choice.


During follow-up, patients were evaluated for disease activity, ongoing UC treatment, body weight and adverse events at days 3 and 5 of IVS and then after discharge at months 3, 6, 12, 18 and 24 and the end of follow-up. Disease activity was assessed during follow-up using (a) partial Mayo Clinic score for clinical activity and Lichtiger index at days 3 and 5; (b) CRP (mg/L), haemoglobin (g/dL), leucocyte count (/109/L), platelet count (/109/L) and albumin (g/dL) for biological activity; and (c) Mayo Clinic endoscopic subscore and the UCEIS for endoscopic activity.[7,15,17]

Outcome Measures

Relapse was defined by (a) active disease according to partial Mayo Clinic score >4 and/or (b) the introduction of oral 5-ASA, steroids, thiopurine, methotrexate, anti-TNF, vedolizumab and/or tofacitinib. Relapses without hospitalisation were previously reported and were not considered adverse events. Clinical remission was defined as a partial Mayo score ≤2 with a combined stool frequency and rectal bleeding subscore of ≤.1[15] Steroid-free clinical remission was defined as clinical remission and the absence of any oral steroid.

Statistical Analysis

All patients were evaluated from the date of discharge to the end of the follow-up period. Patients who discontinued maintenance therapy while in steroid-free clinical remission were censored at the time of discontinuation for the analysis of the risk of relapse. Qualitative data are expressed as number (%), and quantitative data are expressed as mean ± SD or median (interquartile range). Patients' characteristics according to immunomodulator and/or biological agent naïve status were compared using the chi-squared test or the Mann-Whitney test whenever appropriate. Disease activity was compared between days 0, 3 and 5 using the chi-squared, Fisher-exact and Wilcoxon's matched-pair signed-rank test whenever appropriate. The relapse-free and colectomy-free survivals were calculated using the Kaplan-Meier method. The survival distributions were compared using the log-rank test. To identify the independent factors, two Cox proportional hazard models were adjusted according to previous exposure to immunomodulator and/or biological agent with an ascending stepwise procedure. Variables with P < 0.10 in univariate analysis were considered to be potential adjustment variables for the multivariate analysis. The continuous variables were converted to qualitative values by dichotomising around the median value to make two groups of equal size. Variables with P < 0.10 in the univariate analysis were considered potential control variables for the multivariate analysis. All of the analyses were two-tailed, and P -values less than 0.05 were considered significant. All of the statistical evaluations were performed using SPSS statistical software (SPSS Inc., v17). All of the authors had access to the study data and reviewed and approved the final manuscript.