Does postoperative drain amylase predict pancreatic fistula after pancreatectomy?

Background: Previous studies suggest that after pancreatectomy, drain fluid amylase obtained on postoperative day 1 (DFA1) >5,000 U/L correlates with the development of postoperative pancreatic fistula (PF).(1,2) We sought to validate whether DFA1 is a clinically useful predictor of PF and to evaluate whether DFA1 correlates with PF severity.

Study design: Using a prospective database, we reviewed records from patients having pancreatectomy between 2010 and 2012. Presence and grade of PF were determined using the consensus guidelines from the International Study Group on Pancreatic Fistula (ISGPF).(1) RESULTS: Sixty-three patients who underwent pancreatectomy had a documented DFA1. There were 27 (43%) who developed PF: 2 (7%) were grade A, 18 grade B (67%), and 7 were grade C (26%). Median DFA1 in patients with PF (4,600 U/L, range 32 to 16,900 U/L) was significantly higher than in those without PF (45 U/L, range 2 to 5,840 U/L; p < 0.001). When DFA1 was analyzed at varying cutoff values, correlation of DFA1 with PF was high. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were assessed at varying levels of DFA1. Highest sensitivity (96%) and NPV (96%) were obtained with a cutoff DFA1 of <100 U/L. On multivariate analysis, DFA1 >100 U/L was the only significant predictor of PF when controlling for gland texture, duct size, pathology, and neoadjuvant radiation. There was no statistically significant relationship between DFA1 and PF grade.

Conclusions: In patients undergoing pancreatic resection, a cutoff DFA1 of 100 U/L resulted in high sensitivity and NPV. Early drain removal may be safe in these patients. Further studies are recommended to validate the role of DFA1 in excluding PF and assisting in management of surgical drains.