Postoperative Day 1 Drain Amylase After Pancreatoduodenectomy: Optimal Level to Predict Pancreatic Fistula

Introduction Drain amylase on day 1 (DA-D1) after pancreaticoduodendectomy (PD) to predict occurrence of postoperative pancreatic fistula (POPF) is controversial. In this study, we evaluate the optimal DA-D1 level to predict clinically relevant POPF (CR-POPF). Methods The 2014-2020 NSQIP pancreatectomy-targeted database was queried for patients who underwent elective PD. Perioperative data was extracted to determine development of POPF and CR-POPF per International Study Group of Pancreatic Fistula guidelines. Receiver operative curve (ROC) and Youden's index were used to assess the performance and optimal cutoff for DA-D1 to predict CR-POPF. The DA-D1 value was confirmed with a multivariable logistic regression to determine hazard ratios (HR) for CR-POPF and conditional logistic regression by modified fistula risk score (mFRS) subgroups. Results A total of 6,087 patients with complete perioperative data were included. Mean DA-D1 was 2,897 +/- 8,636 U/L; median drain duration was 5 days. CR-POPF was documented in 544 (8.9%) patients. DA-D1 ROC for CR-POPF had area under the curve of 0.779 (95%CI 0.759-0.798). Youden's index for the CR-POPF ROC coordinates had 77.6% sensitivity and 66.3% specificity, corresponding to DA-D1 values >= 720U/L as an optimal cutoff. CR-POPF was higher for patients with DA-D1 >= 720U/L ( HR 4.6; p = 0.001). Patients DA-D1 < 720U/L with a negligible, low, intermediate, and high mFRS had respectively 1%, 3%, 4%, and 7% rate of CR-POPF. Conclusion DA-D1 < 720U/L after elective PD is a clinically useful predictor of CR-POPF. For patients with negligible to intermediate FRS, surgeons should consider utilizing DA-D1 < 720 U/L for removal of a drain on the first postoperative day.