Pancreatoduodenectomy: Role of drain fluid analysis in the management of pancreatic fistula

Background: Pancreatic fistula remains an important cause of death following pancreatoduodenectomy. There is still uncertainty regarding the use of drains following pancreatoduodenectomy with recent reports suggesting that it might be harmful with increased complications. We evaluated the use of drain fluid analysis in the management of patients following pancreatoduodenectomy. Methods: A prospective study was conducted on all patients undergoing pancreatoduodenectomy at two hospitals between April 2004 and August 2006. Drain fluid analysis was carried out from day 3 to day 5. These data were collected with the clinical pictures of the patients and with subsequent radiological assessment. Results: Fifty consecutive patients underwent modified extended pancreatoduodenectomy for a periampullary tumour. In patients with no clinical evidence of a fistula, the mean postoperative drain fluid amylase levels were as follows: on postoperative day 3 it was 262 U/mL (standard error of mean 69), on postoperative day 4 it was 112 U/mL (standard error of mean 47) and on postoperative day 5 it was 125 U/mL (standard error of mean 64). Only three (6/6, 50%) of these patients had clinical features suggestive of a leak and were found to have a pancreatic fistula on subsequent imaging. There was no correlation between the total or mean volumes of drainage and development of a pancreatic fistula. Conclusion: The drain fluid analysis did not provide additional information that was not already evident from the clinical picture of the patient. Drain fluid analysis had no effect on patients with a biochemical leak only. Patients who had a significant disruption of their pancreatic anastamosis did not need biochemical analysis as the character, that is, turbidity of the drain fluid was an equally reliable indicator of the underlying pathology.