Key predictors of surgical intervention in patients with severe UC, PMC, and IC include rapidly progressive disease course, mucosal contact bleeding, ulcerative defects, colonic bleeding, and colonic dilation.
Key Findings
Results
Surgically treated patients were significantly younger than those who received conservative therapy.
The study included 308 patients with severe and fulminant ulcerative colitis (UC), pseudomembranous colitis (PMC), and ischemic colitis (IC).
108 patients received surgical treatment and 200 received conservative therapy.
Age difference between surgical and conservative groups was statistically significant (p=0.025).
Results
Rapidly progressive disease course was significantly associated with surgical intervention.
Surgically treated patients more frequently had a rapidly progressive disease course.
OR 2.85 (95% CI 1.65–4.92; p<0.001).
This was identified as a key clinical predictor of surgical intervention across all three colitis types studied.
Results
Mucosal contact bleeding on endoscopy was a significant predictor of surgical intervention.
Mucosal contact bleeding was identified as an endoscopic predictor of surgery.
OR 1.99 (p=0.020).
Endoscopic findings were among the clinical and demographic data analyzed in the retrospective cohort study conducted between 2015 and 2024.
Results
Ulcerative defects on endoscopy were a significant predictor of surgical intervention.
Ulcerative defects were identified as an endoscopic predictor of surgery.
OR 1.97 (p=0.012).
This finding was identified alongside mucosal contact bleeding as an endoscopic predictor.
Results
Colonic bleeding as a complication was the strongest predictor of surgical intervention among complications analyzed.
Colonic bleeding was significantly associated with surgical intervention.
OR 6.33 (p<0.001).
This was the highest odds ratio among all identified predictors in the study.
Results
Colonic dilation as a complication was significantly associated with surgical intervention.
Colonic dilation was significantly associated with surgical intervention.
OR 3.67 (p=0.013).
Both colonic bleeding and colonic dilation were analyzed among complications alongside clinical, endoscopic, and imaging findings.
Results
The binary logistic regression predictive model demonstrated satisfactory discrimination with high sensitivity but moderate specificity.
Model discrimination was assessed using ROC analysis.
AUC was 0.700 (p<0.001), indicating satisfactory discrimination.
The model achieved high sensitivity of 92.9% and moderate specificity of 33.7%.
Predictors were identified using binary logistic regression with calculation of odds ratios and 95% confidence intervals.
What This Means
This research examined 308 patients hospitalized with severe forms of three types of colon inflammation: ulcerative colitis, pseudomembranous colitis (often caused by C. difficile infection), and ischemic colitis. The study compared patients who ultimately needed surgery (108 patients) with those successfully treated with medications alone (200 patients), looking for warning signs that could predict who would need an operation. The study was conducted over a nine-year period (2015–2024) at a single institution.
This research suggests that several factors are strongly linked to a patient needing surgery. Patients who needed surgery tended to be younger and had disease that worsened rapidly. During colonoscopy (a camera examination of the colon), finding that the lining bleeds when touched or has open ulcers also raised the likelihood of surgery. Among serious complications, active bleeding from the colon (about 6 times more likely to lead to surgery) and abnormal widening of the colon (about 3.7 times more likely) were the strongest warning signs. The researchers combined these factors into a predictive model that was very good at identifying patients who would need surgery (catching about 93% of surgical cases), though it also flagged many patients who ultimately did not need surgery.
This research suggests that clinicians caring for patients with severe colitis can use these specific clinical, endoscopic, and complication-related warning signs to help identify earlier which patients are at high risk of requiring surgical intervention. Earlier identification of high-risk patients could potentially allow for faster decision-making and preparation for surgery when needed. The moderate overall accuracy of the model (AUC 0.700) indicates that while useful as a screening tool, it should be used alongside clinical judgment rather than as a standalone decision-making instrument.
Krotov G, Danilov M, Knyazev O, Tsvirkun V. (2026). [Predictors of surgical intervention in patients with ulcerative, pseudomembranous, and ischemic colitis].. Terapevticheskii arkhiv. https://doi.org/10.26442/00403660.2026.08.203723