Abstract
BACKGROUND: Optimal management of patients with severe hyperbilirubinemia (> 14.6 mg/dL) undergoing pancreaticoduodenectomy remains controversial. METHODS: Single-center retrospective study of 665 pancreaticoduodenectomies (2007-2024). Patients were stratified: no preoperative biliary drainage (PBD) with high (> 14.6mg/dL, n = 83) or low (< 14.6mg/dL, n = 312) bilirubin; PBD for low bilirubin (n = 140); PBD for high bilirubin (n = 113). PRIMARY OUTCOMES: 90-day mortality, morbidity, and specific complications. RESULTS: Ninety-day mortality was significantly higher in non-PBD patients with high bilirubin (13.3%) compared to other groups (2.9-5.7%, p = 0.001). Overall morbidity was also higher in this group (29.0% vs. 12.8-20.7%, p < 0.001). While PBD groups showed higher surgical site infections (21.2-26.4% vs. 9.6%, p = 0.001), non-PBD high-bilirubin patients demonstrated increased ARDS (6.0% vs. 0-1.4%, p = 0.016) and reoperation rates (18.1% vs. 9.3-11.5%, p = 0.029). Multivariate analysis identified preoperative bilirubin > 14.6mg/dL as an independent predictor of 90-day mortality. CONCLUSION: Preoperative bilirubin ≥ 14.6mg/dL independently predicts perioperative morbidity and mortality following pancreaticoduodenectomy. These findings support preoperative biliary drainage in patients with severe hyperbilirubinemia to optimize surgical outcomes, despite PBD-associated infectious complications.