Objective: To assess the safety of the pancreatic anastomosis after pancrea
tico-duodenectomy (PD).
Design: Non-randomized prospective trial in consecutive patients.
Setting: University hospital.
Subjects: 171 consecutive patients with resectable periampullary cancer (80
%) or intractable pain due to chronic pancreatitis (20%) undergoing PD.
Interventions: Pancreaticojejunostomy (PJ) and pancreaticogastrostomy (PG).
Main outcome measures: Mortality and morbidity rates due to anastomotic lea
k following PJ and PG.
Results: 91 PJ and 80 PG patients were comparable for age, gender, total bi
lirubin, ASA grading, indication for PD, operating time, pancreas texture,
blood loss and replacement. The rate of pancreatic fistula was significantl
y higher in PJ patients (13%) than in PG patients (3.7%) (12 vs. 3, p = 0.0
29). Overall death rate was significantly higher after PJ (12%) than after
PG (3.7%) (11 vs. 3, p = 0.047). Fatal outcome due to pancreatic leak (3 vs
, 1, p = 0.83) and other death rates (8 vs. 2, p = 0.14) were not significa
ntly different in PJ and PG groups, respectively.
Conclusion: PJ was associated with significantly higher pancreatic leak rat
e than PG. However, there was no statistically significant difference in mo
rtality rates directly related to pancreatic leak.