Predictors of prolonged pancreatic fistula in patients undergoing pancreatoduodenectomy (whipple procedure)

Abstract

Presenter: May Tee MD, MPH | Mercy Medical Center Background: Postoperative pancreatic fistula (POPF) contributes to substantial morbidity associated with pancreatoduodenectomy (PD). Risk stratification for development of POPF, based on factors such as pancreatic duct size, pancreatic texture, and intraoperative blood loss, may mitigate severity of POPF by treatment with antibiotics, additional drains, or feeding tubes. We aimed to examine predictors of prolonged POPF (POPF at 30 days or greater) in patients undergoing PD to provide further stratification of patients at risk. Methods: The American College of Surgeons National Surgical Quality Improvement Program pancreatectomy procedure targeted participant use file was queried from 2015-2017, when persistent POPF at postoperative day 30 was captured. All PD procedures were extracted with emergent cases excluded. Univariable and multivariable analyses were conducted to examine the association between persistent POPF and 30-day surgical outcomes. Results: N = 12,014 patients undergoing open and minimally invasive PD were evaluated. On univariate analyses, persistent POPF was associated with the following outcomes: surgical site infection (SSI) superficial (11.0% vs. 7.4%, P < 0.001), deep SSI (3.1% vs. 1.2%, P < 0.001), organ space SSI (44.1% vs. 13.1%, P < 0.001), wound dehiscence (2.3% vs. 1.0%, P < 0.001), pneumonia (6.7% vs. 3.5%, P < 0.001), reintubation (9.2% vs. 3.2%, P < 0.001), pulmonary embolus (2.9% vs. 1.1%, P < 0.001), failure to wean off mechanical ventilation (9.2% vs. 2.4%, P < 0.001), renal insufficiency (1.6% vs. 0.6%, P < 0.001), renal failure (2.4% vs. 0.8%, P < 0.001), myocardial infarction (2.3% vs. 1.1%, P=0.001), deep vein thrombosis (6.8% vs. 2.6%, P < 0.001), sepsis (19.2% vs. 8.5%, P < 0.001), septic shock (9.9% vs. 2.4%, P < 0.001), length of stay greater than 12 days (58.9% vs. 23.7%, P < 0.001), discharge to destination other than home (28.9 vs. 12.8%, P < 0.001), return to operating room (13.5% vs. 4.9%, P < 0.001), readmission to hospital (23.8% vs. 15.8%, P < 0.001), delayed gastric emptying (35.2% vs. 15.4%, P < 0.001), percutaneous drain placement (33.3% vs. 10.9%, P < 0.001), and postoperative C. difficile colitis (3.3% vs. 1.8%, P=0.003). On multivariable analyses, independent predictors of prolonged POPF include: male sex (OR=1.6, P < 0.001), obesity (OR=1.5, P < 0.001), operative time greater than 420 minutes (OR=1.2, P=0.03), preoperative chemotherapy (OR=0.6, P=0.002), postoperative blood transfusion (OR=1.7, P < 0.001), pancreatic duct < 3mm (OR=1.7, P=0.002), soft pancreatic texture (OR=2.4, P < 0.001), and drain to suction (OR=1.4, P=0.015). Conclusion: Prolonged POPF is a major source of morbidity following PD but not mortality. Further risk stratification for POPF may be considered for male patients with increased body mass index undergoing prolonged PD procedures. Significant perioperative blood loss requiring transfusion, small pancreatic duct, and soft pancreas remain important predictors of prolonged POPF. Interestingly, drain placed to suction was associated with increased odds of prolonged POPF but neoadjuvant chemotherapy appeared to be protective against prolonged POPF.