A Clinical Predictive Model for One-year Colectomy in Adults Hospitalized for Severe Ulcerative Colitis

Area of Science:

• Gastroenterology
• Clinical Medicine
• Predictive Analytics

Background:

• Predictive models for colectomy in ulcerative colitis (UC) are crucial for managing high-risk patients.
• Identifying patients requiring colectomy early improves clinical outcomes and resource allocation.

Purpose of the Study:

• To develop a clinical predictive model for one-year colectomy risk in adults hospitalized for severe UC.
• To create a simple scoring system based on admission data for risk stratification.

Main Methods:

• Retrospective analysis of 229 patients hospitalized for severe UC.
• Multivariate regression to identify predictors of one-year colectomy.
• Development of a point scoring system using identified variables and bootstrapping for risk estimation.

Main Results:

• Four variables predicted one-year colectomy: prior biologic exposure (single/multiple classes), inpatient salvage therapy (cyclosporine/TNF-alpha inhibitor), and age <40.
• A scoring system was developed, with colectomy probabilities ranging from 9.4% to 75.0% for scores 1-4.
• A score of zero perfectly predicted no colectomy within one year.

Conclusions:

• Prior biologic exposure, need for inpatient salvage therapy, and younger age are significant risk factors for one-year colectomy in severe UC.
• A simple, validated scoring system can effectively identify and stratify patients at high risk for colectomy during hospitalization.