Abstract
Background
The formation of the pancreaticojejunostomy during pancreaticoduodenectomy is the most technically challenging aspect of the procedure, with its failure increasing rates of both morbidity and mortality significantly. Early identification and management of a clinically relevant post-operative pancreatic fistula (CR-POPF) can be critical in reducing the threat of potentially avoidable harm to the patient. The most used indicator for a CR-POPF is the level of drain fluid amylase. There are many different techniques for forming the anastomosis, with considerable analysis but no consensus on superiority. We aimed to look at our centres experience using different techniques and the trends we observed in drain amylases and clinical outcomes.
Methods
A prospective database of all patients in a single UK centre undergoing pancreatic or duodenal resection has been maintained. This includes patient demographics, diagnosis pre and post operatively, operative details and duration, complications, and outcomes. All patients undergoing a pancreaticoduodenectomy between 1st January 2020 and 31st July 2021 were identified and their data retrospectively analysed.
Results
Thirty-three patients underwent a pancreaticoduodenectomy during the study period. The pancreatojejunostomy was formed using a duct-to-mucosa anastomosis in twenty-eight patients and using a dunking technique in five patients.
The mean of the highest drain fluid amylase on post-operative day one for the patients with a dunking anastomosis was 14804.8 (range 3643-43686), on day three 2376.12 (range 167-8008.6) and of the three patients whose drains were in situ at day 5 it was 522.2 (range 31 to 983. An 83.9% reduction in mean drain amylases was observed between Day One and Day Three, followed by a further 78% reduction between day 3 and day. One patient (20%) had a CR-POPF with a grade B fistula, two others had a biochemical leak.
The mean Day One drain amylases for patients with a duct-to-mucosa anastomosis was 71% lower at 4274.5 (range 15.4 to 41755). However this increased by 11.5% by Day Three to 4766.4 (range 5 to 46300) before falling by 64.7% to 1681.9 (range 5 to 13015) on Day Five. Eight patients (28.6%) had a CR-POPF – 3 grade B and 5 grade C fistula - and three patients had a biochemical leak.
Conclusions
In our centre’s experience, the type of anastomosis used to perform the pancreatic reconstruction post pancreaticoduodenostomy significantly impacts the post-operative trend in drain fluid amylase. This is important for clinicians to appreciate in order to avoid premature suspicion of a CR-POPF and prevent potentially unnecessary intervention.