RESUMEN
BACKGROUND: There are no established predictors for deciding between upfront surgery and PBD in pancreatic head malignancy. Once PBD is present, the ideal drainage-time remains elusive. The aim was, to identify predictors in jaundiced patients and ideal PBD-duration. METHODS: Analysis of 304 patients with pancreatic head malignancy (56% with PBD, n = 170) undergoing pancreaticoduodenectomy was performed. Postoperative morbidity and survival were analyzed. RESULTS: Postoperative complications increased after PBD (98.2% vs. 88.8%; p < 0.001). Patients with PBD received more postoperative antibiotics (42.4% vs. 21.6%; p < 0.001) and wound infections were increased (21.4% vs. 9.4%; p = 0.006). INR predicted postoperative morbidity (p = 0.026), whereas serum-bilirubin (p = 0.708), leucocytes (p = 0.158) and MELD-score (p = 0.444) had no impact. Complications were not different between long (>4 weeks) and short (<4 weeks) PBD-duration (p = 0.608). No life-threatening complications (CDIV + V) occurred after long drainage (0.0% vs. 8.9%; p = 0.028) and long-term survival was not compromised. CONCLUSIONS: INR is a suitable predictor for postoperative outcome, while serum-bilirubin levels had no predictive value. The INR can help deciding between PBD and upfront surgery. If PBD is inevitable, drainage duration of >4 weeks reduced major complications. CATEGORY: Clinical study.