Early Drain Amylase Value Predicts the Occurrence of Pancreatic Fistula After Pancreaticoduodenectomy
We read with special interest the article by Ven Fong et al1 published on August 2015. We believe that this study is a cornerstone in pancreatic surgery, bringing a new concept that can greatly contribute to postoperative care in patients undergoing pancreaticoduodenectomy (PD).
Surgical resection remains the treatment of choice for patients with periampullary tumors, representing the only chance to obtain long-term survival.2 Nowadays, with improvements in surgical expertise, anesthesia and postoperative care, this procedure can be accomplished with a mortality rate of less than 5%.3
Pancreatic fistula (PF), especially that one of clinical relevance, is a major complication after PD, affecting up to one-third of patients.4 Early prediction of its occurrence is of considerable interest5; its exclusion in the early postoperative days enables the treatment of patients under fast-track protocol, which includes mobilization and removal of abdominal drains, rapid evolution of oral feeding, shorter length of hospital stay, and cost savings.6 This idea is important because the occurrence of PF changes the surveillance of patients, carrying a high percentage of major complications (bleeding, abdominal abscess, need for invasive procedures, or reoperation), often requiring interruption of enteral nutrition, maintenance of drains, use of antibiotics, and close monitoring of patients.7
The incidence of PF ranges from 2% to 51%, depending on the criteria adopted in several studies.1,2,7 In 2005, the International Study Group of Pancreatic Fistula (ISGPF) defined its manifestation and classified it into 3 levels according to its severity, to make an international standardization that allows comparison between different institutions.8
Using a cutoff point of 600 U/L on the first postoperative day (POD1), Ven Fong et al have established a value related to low likelihood of occurrence of PF. What draws attention in their work is the average length of hospital stay among patients with higher levels of amylase on POD1. In this group, the reported pancreatic fistula rate was 31.4% (44 patients). It is possible, however, that these rates have been underestimated, because many patients have been discharged on POD7. We are carrying out a similar work in our institution, with daily measurements of drains amylase levels and have observed cases of late pancreatic fistulas, on the 9th and 11th POD. These patients showed high amylase levels on POD1, had normalized amylase values (along with a clear appearance of the drain fluid) in the following days, and have developed PF in the postoperative follow-up, including those of clinical significance.
Thereby, we raise some considerations about the series of Ven Fong et al. Perhaps some patients developed PF out of the hospital, underestimating the reported rate. Maybe it is not safe to discharge patients who have a greater likelihood of developing PF, as many serious complications (such as bleeding and need for reoperation) may occur outside the hospital. Finally, possibly the 25.7% readmission rate in this group could be reduced simply keeping these patients under close observation.
We would like to congratulate the authors for their brilliant work and for raising important issues about the treatment of patients undergoing PD, which remains a challenging operation even for the most skilled Hepato-Pancreato-Biliary surgeons.