A risk-scoring model for predicting late postoperative hemorrhage following pancreatoduodenectomy: development and...

Post-pancreatectomy hemorrhage (PPH) represents a life-threatening complication following pancreatoduodenectomy (PD). This study aimed to develop and externally validate a clinically applicable risk-scoring model to predict its occurrence.

Patients who underwent curative-intent PD were included in the study. The risk-scoring model for predicting late PPH was developed in the training cohort, and the performance of the model was subsequently validated in an external validation cohort.

Of 405 eligible patients, 300 formed the training cohort and 105 the external validation cohort. Late PPH occurred in 7.7% and 7.6% of patients, respectively. Late PPH significantly impacted the long-term prognosis, with a median survival of 20.7 months compared to 35.2 months (P = 0.009). The most frequent site of hemorrhage was the common hepatic artery, accounting for 22.6% of cases. Multivariate logistic analysis identified body mass index (BMI), preoperative total bilirubin (TBIL), preoperative prothrombin time (PT), and clinically relevant postoperative pancreatic fistula (CR-POPF) as independent risk factors associated with late PPH. By integrating these four factors, the predictive model demonstrated concordance indices of 0.863 and 0.825 in the training and validation cohorts, respectively. The model's discriminative capacity was further assessed by categorizing the predicted probabilities of late PPH into two risk groups: low-risk (score ≤2) and high-risk (score >2).

We developed and externally validated a simple risk-scoring model for late PPH after PD, using routinely available clinical variables to predict risk and stratify high-risk patients for targeted monitoring and prevention.