Duct-to-Mucosa vs Invagination for Pancreaticojejunostomy after Pancreaticoduodenectomy: A Prospective, Randomized Controlled Trial from a Single Surgeon.

BACKGROUND

Pancreatic fistula (PF) is the most common significant complication after pancreaticoduodenectomy. Invagination and duct-to-mucosa anastomoses are anastomotic techniques that are commonly performed after pancreaticoduodenectomy. There are conflicting data on invagination vs duct-to-mucosa anastomoses about which is superior for minimizing the risk of PF. In addition, all previous studies involved multiple operating surgeons and failed to control for variation in surgeon expertise.

STUDY

DESIGN

This was a randomized controlled study comparing the outcomes of PD between patients who underwent invagination vs those who had duct-to-mucosa anastomoses. All 132 patients were operated on between October 2012 and March 2015 by a single surgeon experienced in both procedures. Pancreatic fistula was the main end point.

RESULTS

Overall and clinically relevant rates of PF rate were 29.5% and 10.6%, respectively. Overall PF rates in the patients treated with invagination vs duct-to-mucosa anastomoses were 30.9% vs 28.5% (p = 0.729), respectively and the corresponding clinically relevant PF rates were 17.6% vs 3.1%, respectively (p = 0.004). Although the overall complication rates were similar in the 2 groups, severe complications were significantly more frequent in the patients treated with invagination (p = 0.013). Duct-to-mucosa anastomosis was also associated with shorter postoperative hospital stay (13 vs 15 days; p = 0.021). There was one perioperative death. Independent variables for the risk of PF were the diameter of the pancreatic duct (greater risk with smaller diameter), the underlying pathology, and male sex.

CONCLUSIONS

Both methods yield similar overall rates for PF, but the rate of clinically relevant PF is lower in patients treated with duct-to-mucosa anastomosis. Additional single-surgeon studies or multi-institution randomized trials controlling for comparable expertise in both procedures should be conducted to confirm these results.