Not all Whipple procedures are equal: Proposal for a classification of pancreatoduodenectomies

Abstract

Background: Pancreatoduodenectomy is the standard treatment for pathologies of the pancreatic head and is performed routinely worldwide. The aim of the study was to analyze this procedure in terms of extent of surgery, technical difficulty, and clinical outcomes and thereby provide a standardized surgical categorization of pancreatoduodenectomies for future reference.

Methods: For this cohort study, all patients who underwent pancreatoduodenectomy at a single center within an 18-year period (October 2001 to December 2019) were identified in a prospectively maintained database. Based on technical difficulty and extent of surgery, 4 pancreatoduodenectomy types were proposed: (1) standard pancreatoduodenectomy; (2) pancreatoduodenectomy with portal vein/superior mesenteric vein resection; (3) pancreatoduodenectomy with multivisceral resection; and (4) pancreatoduodenectomy with arterial resection. Patient characteristics, surgical parameters, and perioperative morbidity and mortality were analyzed. The 4 types were compared with regard to their surgical outcomes.

Results: A total of 3,953 pancreatoduodenectomies were performed in the study period. Standard pancreatoduodenectomy (type 1) was the most frequent procedure (n = 2,931, 74.1%), followed by pancreatoduodenectomy with portal vein/superior mesenteric vein resection (type 2: n = 568, 14.4%), pancreatoduodenectomy with multivisceral resection (type 3: n = 415, 10.5%), and pancreatoduodenectomy with arterial resection (type 4: n = 39, 1.0%). Demographic baseline characteristics were clinically comparable among pancreatoduodenectomy types. Mortality within 90-days correlated with the type of pancreatoduodenectomy (type 1: 2.9%; type 2: 4.2%; type 3: 6.3%; type 4: 10.3%; P = .0007). Overall surgical morbidity was 41.7% (type 1), 40.8% (type 2), 52.5% (type 3), and 59.0% (type 4) (P < .0001), including postoperative pancreatic fistula type B/C (type 1: 11.9%; type 2: 7.7%; type 3: 14.7%; type 4: 15.4; P = .0031) and delayed gastric emptying (type 1: 19.4%; type 2: 22.5%; type 3: 22.0%; type 4: 25.6%; P = .187) as the most frequent complications. Relaparotomies were more frequent in type 4 (20.5%) and type 3 (20.6%) than in type 2 (12.0%) or type 1 (10.4%) pancreatoduodenectomy (P < .0001). Intensive care unit stay ≥2 days was more frequent in type 4 (48.7%) compared with type 3 (25.7%) or type 2 (27.1%) and type 1 (18.6%) (P < .0001).

Conclusion: The results show different clinical outcomes for the 4 types of pancreatoduodenectomy. Morbidity and mortality correlate with pancreatoduodenectomy type. The proposed pancreatoduodenectomy classification is useful for reporting pancreatoduodenectomy procedures, enhances the comparability of future studies, may be used for training purposes, and may guide intra and postoperative decision-making.