Surgical palliation for pancreatic malignancy: practice patterns and predictors of morbidity and mortality.

Introduction: Most patients with pancreatic cancer present with, or develop, biliary or duodenal obstruction. We sought to characterize palliative surgery utilization in a contemporary cohort and identify patients at high risk of morbidity and mortality.

Methods: The ACS NSQIP database (2005-2011) was queried for patients with a pancreatic malignancy undergoing gastrojejunostomy, biliary bypass, or laparotomy without resection. Univariate analysis and multivariate logistic regression identified factors associated with increased risk of 30-day morbidity or mortality.

Results: Operations for the 1,126 patients undergoing palliative bypass were gastrojejunostomy alone (33%), bile duct bypass alone (27%), both (31%), or cholecystojejunostomy (9%). A major complication occurred in 20% and mortality in 6.5% at 30 days. Risk factors for morbidity and mortality were defined in multivariate models. The number of identified risk factors stratified morbidity from 14.8-50% and mortality from 1.6-50% (p < 0.0001 for each). Laparotomy alone (n = 622) was associated with lower morbidity than bypass (12 vs. 20%, p < 0.0001), but equivalent mortality (5 vs. 6.5%, p = 0.21).

Conclusion: Palliative bypass for pancreatic cancer is associated with a high rate of morbidity and mortality. In select patients, this risk may be prohibitive. Patient selection reflecting predictors of morbidity and mortality may allow for improved outcomes.