Risk factor analysis and nomogram for predicting massive postoperative gastrointestinal bleeding in patients with Crohn's disease: a multicenter retrospective study.

Background: Massive postoperative gastrointestinal bleeding is a severe postoperative complication of Crohn's disease (CD) with a high mortality rate, and deteriorating patients' recovery. However, there are few related studies, and it lacks effective prevention measures. Therefore, we conducted a multicenter study to explore the risk factors for massive postoperative gastrointestinal bleeding in CD patients. Methods: This study was a multicenter retrospective case-control study. Patients who were diagnosed with CD and underwent gastrointestinal (GI) surgery were enrolled. The control group was matched 1:4 for gender and age. By comparing perioperative medical information between two groups, risk factors were identified through logistic regression analysis. A nomogram was constructed and internal validation was performed by bootstrap resampling. Results: A total of 170 patients were included. Multivariable logistic regression revealed the independent predictors of massive postoperative gastrointestinal bleeding involving the number of previous abdominal surgeries (OR = 2.56, 95% CI = 1.54-4.24), GI bleeding history (OR = 6.17, 95% CI = 1.59-23.97), serum albumin (ALB) (OR = 0.88, 95% CI = 0.81-0.96), and Nutrition Risk Screening 2002 (OR = 1.57, 95% CI = 1.08-2.29). The nomogram achieved an area under the curve (AUC) value of 0.85 (95% CI: 0.76-0.93). In internal validation, the AUC value was 0.976 (95% CI: 0.955-0.997). Calibration curves showed good alignment. DCA demonstrated that the diagnostic model had good clinical efficiency. Conclusion: The risk of massive postoperative gastrointestinal bleeding in CD patients will be increased with a GI bleeding history, more previous abdominal surgeries, higher nutrition risk, and lower ALB level. Our nomogram model is effective and could be a useful tool for prediction.