What is a better predictor of clinically relevant postoperative pancreatic fistula (CR-POPF) following pancreaticoduodenectomy (PD): postoperative day one drain amylase (POD1DA) or the fistula risk score (FRS)?

Background Both fistula risk score (FRS) and drain amylase in postoperative day 1 (POD1DA) have been promoted as tools to guide placement and removal of surgical drains following pancreaticoduodenectomy (PD). However, their individual utility has not been compared. Methods A consecutive cohort of PD patients from 2013 to 2015 were identified from a prospectively collected institutional database. Pearson correlation coefficients and receiver operating characteristic (ROC) curves were calculated for FRS (negligible/low vs. moderate/high) and POD1DA of 600 U/L and 5000 U/L as predictors of clinically relevant postoperative pancreatic fistula (CR-POPF). Results The incidence of CR-POPF was 27% in 216 patients. Sensitivity and specificity of FRS, POD1DA >600 U/L, and POD1DA >5000 U/L for predicting CR-POPF were 83% and 55%, 94% and 60%, 33% and 90%. The ROC area under the curve (AUC) for POD1DA >600 U/L (0.764) and FRS (0.749) were not significantly different (p = 0.713). However, POD1DA >5000 U/L (0.615) was significantly worse at predicting CR-POPF (p = 0.015). When FRS and POD1DA >600 U/L were combined; there was no improvement (p = 0.624). Discussion FRS and POD1DA are equally accurate in predicting CR-POPF. Patients with negligible/low FRS or POD1DA <600 U/L should be considered for drain removal.