Safety and function of isolated Roux loop pancreaticojejunostomy after Whipple's pancreaticoduodenectomy.

A novel method of pancreatic anastomosis after proximal Whipple-type resection: classical pancreaticoduodenectomy (PD) or pylorus-preserving pancreaticoduodenectomy (PPPD), has been evaluated over a 5-year period from 1987 to 1992 in 52 patients. Indications for resection included chronic pancreatitis (n = 9) and neoplasms (n = 43). Reconstruction involved a cephalad end-to-end duodeno-/gastro-jejunal anastomosis with a biliary anastomosis 6-8 cm downstream. A separate isolated defunctioned Roux loop was used to construct a duct-to-mucosa (Wirsung-jejunal) pancreaticojejunostomy. Median postoperative stay was 18.0 days (range 11-32 days); three deaths (operative mortality 5.8%) occurred due to sepsis (subhepatic abscess), profound hypoglycaemia and necrotising pancreatitis respectively. These deaths were not related to pancreatic fistula. There were no pancreatic leaks (defined as greater than 50 ml of amylase-rich fluid for more than 7 days). Postoperative exocrine pancreatic function was good as assessed by re-establishment of preoperative weight (achieved in 35 of 40, ie 88% of surviving PPPD patients), clinical steatorrhoea (present in 10 of 41, ie 24% of surviving patients resected for neoplasm), and the need for pancreatic exocrine supplements (required in only 4 of 41, ie 9.8% of surviving patients resected for neoplasm). Twenty patients considered to have normal pancreatic remnants underwent a p-aminobenzoic acid (PABA) excretion test at 3 to 18 months after operation. Median PABA excretion index was 48% (range 24-100%). Isolated defunctioned duct-to-mucosa pancreaticojejunostomy is a safe procedure offering good functional results after Whipple's PD or PPPD resection.