Randomized clinical trial of pancreaticogastrostomy versus pancreaticojejunostomy on the rate and severity of pancreatic fistula after pancreaticoduodenectomy

BackgroundAnastomotic leakage of pancreaticojejunostomy (PJ) remains the single most important source of morbidity after pancreaticoduodenectomy (PD). The primary aim of this randomized clinical trial comparing PG with PJ after PD was to test the hypothesis that invaginated PG would result in a lower rate and severity of pancreatic fistula. MethodsPatients undergoing PD were randomized to receive either a duct-to-duct PJ or a double-layer invaginated PG. The primary endpoint was the rate of pancreatic fistula, using the definition of the International Study Group on Pancreatic Fistula. Secondary endpoints were the evaluation of severe abdominal complications (Clavien-Dindo grade IIIa or above), endocrine and exocrine function. ResultsOf 123 patients randomized, 58 underwent PJ and 65 had PG. The incidence of pancreatic fistula was significantly higher following PJ than for PG (20 of 58 versus 10 of 65 respectively; P=0014), as was the severity of pancreatic fistula (grade A: 2 versus 5 per cent; grade B-C: 33 versus 11 per cent; P=0006). The hospital readmission rate for complications was significantly lower after PG (6 versus 24 per cent; P=0005), weight loss was lower (P=0025) and exocrine function better (P=0022). ConclusionThe rate and severity of pancreatic fistula was significantly lower with this PG technique compared with that following PJ. Registration number: ISRCTN58328599 (http://www.controlled-trials.com).