Background: Pancreatic leak (PL) remains a major cause of postoperative morbidity in patients undergoing pancreatic resection. We sought to evaluate the incidence of and identify risk factors for the development of PL in patients undergoing distal pancreatectomy (DP).
Methods: All patients who underwent primary DP (excluding completion pancreatectomy and debridement) between 1/1/1984 and 7/1/2006 were identified. Data on demographics, clinicopathologic features, operative details, complications, and mortality were analyzed. Chi-squared and multivariate logistic regression analyses were performed to identify risk factors for PL.
Results: In a cohort of 704 patients undergoing primary DP, the median age was 58 years, 45% were male, and 80% were white. The indications for DP were benign pancreatic neoplasm (34%), malignant pancreatic neoplasm (31%), other neoplasm (15%), chronic pancreatitis (14%), pseudocyst (3%), and trauma (3%). Splenectomy was performed in 89%. The pancreatic remnant was sutured alone in 83%, stapled alone in 5%, and both stapled and sutured in 9%. Duct ligation was performed in 22%. Perioperative mortality was <1%, but overall morbidity was 33%. PL requiring a change in clinical management was seen in 12% of cases. Development of PL was associated with an increase in perioperative mortality from 1% to 4% (P=0.04) and an increase in median length of stay from 7 to 10 days (P<0.001). Of those with PL, 35% required additional percutaneous drainage, but only 2% required reoperative intervention. Multivariate analysis revealed that malignant neoplasm (odds ratio (OR) 1.3, P=0.29) and chronic pancreatitis (OR 1.6, P=0.12) as indications for DP did not change PL risk as compared to benign neoplasm. However, increased risk of PL was seen when DP was performed for trauma (OR 6.2, P=0.001) or pseudocyst (OR 3.3, P=0.02). Tobacco use (OR 2.0, P<0.001) was associated with increased PL risk, while preoperative diabetes was associated with decreased risk (OR 0.33, P=0.003). Neither staple vs. suture closure of the pancreatic remnant (OR 1.4, P=0.65) nor ligation of the pancreatic duct (OR 2.0, P=0.05) affected PL risk.
Conclusions: This largest reported series of DP demonstrates that this procedure can be performed with low mortality but still carries a substantial risk of morbidity, particularly PL. DP in the trauma setting significantly increases the risk of PL. In contrast to previous studies, PL risk was not associated with surgical management of the pancreatic remnant. These results emphasize the need for prospective randomized trials to evaluate strategies to reduce PL occurrence.