A Scoring System to Determine Patients' Risk of Colectomy Within 1 Year After Hospital Admission for Acute Severe Ulcerative Colitis

Background & Aims There is consensus on the criteria used to define acute severe ulcerative colitis (ASUC) and on patient management, but it has been a challenge to identify patients at risk for colectomy based on data collected at hospital admission. We aimed to develop a system to determine patients’ risk of colectomy within 1 y of hospital admission for ASUC based on clinical, biomarker, and endoscopy data. Methods We performed a retrospective analysis of consecutive patients with ASUC treated with corticosteroids, ciclosporin, or tumor necrosis factor (TNF) antagonists and admitted to 2 hospitals in France from 2002 through 2017. Patients were followed until colectomy or loss of follow up. A total of 270 patients with ASUC were included in the final analysis, with a median follow-up time of 30 months (derivation cohort). Independent risk factors identified by Cox multivariate analysis were used to develop a system to identify patients at risk for colectomy 1 y after ASUC. We developed a scoring system based on these 4 factors (1 point for each item) to identify high-risk (score 3 or 4) vs low-risk (score 0) patients. We validated this system using data from an independent cohort of 185 patients with ASUC treated from 2006 through 2017 at 2 centers in France. Results In the derivation cohort, the cumulative risk of colectomy was 12.3% (95% CI, 8.6–16.8). Based on multivariate analysis, previous treatment with TNF antagonists or thiopurines (hazard ratio [HR], 3.86; 95% CI, 1.82–8.18), Clostridioides difficile infection (HR, 3.73; 95% CI, 1.11–12.55), serum level of C-reactive protein above 30 mg/L (HR, 3.06; 95% CI, 1.11–8.43), and serum level of albumin below 30 g/L (HR, 2.67; 95% CI, 1.20–5.92) were associated with increased risk of colectomy. In the derivation cohort, the cumulative risks of colectomy within 1 y in patients with scores of 0, 1, 2, 3, or 4 were 0.0%, 9.4% (95% CI, 4.3%–16.7%), 10.6% (95% CI, 5.6%–17.4%), 51.2% (95% CI, 26.6%–71.3%), and 100%. Negative predictive values ranged from 87% (95% CI, 82%–91%) to 92% (95% CI, 88%–95.0%). Findings from the validation cohort were consistent with findings from the derivation cohort. Conclusions We developed a scoring system to identify patients at low-risk vs high-risk for colectomy within 1 y of hospitalization for ASUC, based on previous treatment with TNF antagonists or thiopurines, C difficile infection, and serum levels of CRP and albumin. The system was validated in an external cohort.