Nomogram for predicting massive abdominal bleeding in duodenal fistula patients with uncontrolled abdominal infection after delayed transfer to tertiary centers: a multicenter prospective study.

Abstract

Background: This study aimed to develop and validate a nomogram for predicting massive abdominal bleeding in patients with duodenal fistulas and uncontrolled abdominal infection following delayed transfer to tertiary centers.

Methods: This multicenter (A, B, C, and D) prospective cohort study included 389 patients from four centers. A nomogram was developed using Least Absolute Shrinkage and Selection Operator regression and multivariate logistic analysis on data from 240 patients (from center A), and validated on 149 patients (from center B, C, and D). The model's performance was assessed using concordance index, calibration plots, and decision curve analysis.

Results: Six predictors were identified: output ≥1000 mL/day, Sequential Organ Failure Assessment score on admission, infection invading the retroperitoneum, laparotomy requirement, endoscopic fistula diameter ≥2 cm, and International Society on Thrombosis and Haemostasis score ≥5. The nomogram showed good predictive accuracy, with concordance indices (C-indices) of 0.87 (95% CI: 0.80-0.91) in the development cohort and 0.82 (95% CI: 0.74-0.88. Calibration plots showed good agreement between predicted and observed bleeding rates.

Conclusion: The developed nomogram offers a tailored approach to risk stratification for massive abdominal bleeding in duodenal fistula patients delayed transfer to tertiary centers. This tool might enhance clinical decision-making and improve patient outcomes.