The value of drain fluid amylase as a predictor of postoperative pancreatic fistula after pancreaticogastrostomy.

BACKGROUND Drain fluid amylase is commonly used as a predictor of pancreatic fistula after pancreaticoduodenectomy (PD). This study aimed to determine the ideal cut-off value of drain fluid amylase on postoperative day 1 (DFA1) for predicting pancreatic fistula after pancreaticogastrostomy (PG). METHODS Prospective data of 272 consecutive patients undergoing PG between 2010 and 2020 was collected and analysed to determine the postoperative pancreatic fistula (POPF) risk factors. RESULTS The incidence of POPF was 143 cases (52.6%). The median DFA1 in patients with POPF was significantly higher than that of patients with NO-POPF (5483 versus 311, P < 0.001). DFA1 correlated with POPF in the area under the curve (AUC) of 0.84 (P < 0.001). When DFA1 was 2300 U/L, Youden index was the highest, with a sensitivity of 72.7% and a specificity of 82.9%. Logistic regression analysis showed that DFA1 ≥ 2300 U/L was an independent predictor of POPF (P < 0.001; OR: 12.855; 95% CI: 7.019-23.544). The AUC of DFA1 and clinically relevant postoperative pancreatic fistula (CR-POPF) was 0.674 (P < 0.001). CONCLUSION DFA1 ≥ 2300 U/L can be used as an independent predictor of POPF after PG. DFA1 ≥ 3000 U/L can predict the occurrence of CR-POPF, when DFA1 ≥ 3000 U/L, the patients should be observed closely active for complications.