Interplay of Postpancreatectomy Acute Pancreatitis, Postoperative Pancreatic Fistula, and Mortality after Pancreatoduodenectomy: Insight from a Comprehensive Cohort Study of 1,594 Patients and Development of Predictive Nomograms.

Abstract

Postoperative pancreatic fistula (POPF) and postpancreatectomy acute pancreatitis (PPAP) remain the chief obstacles to optimal outcomes after pancreatoduodenectomy. The mechanistic link between PPAP, clinically relevant POPF (CR-POPF), and mortality resulting from the inability to rescue patients after complications has not been fully elucidated. This study retrospectively analyzed 1,594 consecutive pancreatoduodenectomies performed at a high-volume dedicated pancreatic surgery unit between January 2013 and July 2024. Predictors of PPAP, CR-POPF, and 90-day mortality were determined, and corresponding risk prediction nomograms were developed. A total of 571 patients (35.8%) experienced major morbidity. CR-POPF occurred in 24.3% (grade B 18.4%, grade C 5.9%), and 90-day mortality was 3.4% overall and 9.0% among patients with CR-POPF. PPAP was diagnosed in 10.8% and emerged as the dominant independent predictor of CR-POPF (odds ratio [OR] 11.76, p < 0.001), grade C POPF (OR 5.25, p < 0.001), and 90-day mortality (OR 4.88, p < 0.001). PPAP development was linked to preoperative cholangitis, soft gland texture, and duct diameter less than 3 mm. In the CR-POPF subgroup, failure to -rescue was driven by systemic inflammatory response syndrome (OR 15.05, p = 0.002), fungal sepsis, gram-negative contamination, bile leak, hemorrhage, and pulmonary complications. The dedicated failure-to-rescue nomogram showed an area under the curve of 0.973. PPAP is the pivotal trigger in the CR-POPF-mortality cascade after pancreatoduodenectomy. Refinements to existing fistula risk scores and incorporating PPAP into risk models will enable precise stratification. The developed high-fidelity nomograms can effectively identify patients at imminent risk of mortality, thereby facilitating rescue interventions, after external validation across diverse institutions.