The value of serum amylase and drain fluid amylase to predict postoperative pancreatic fistula after pancreatoduodenectomy: a retrospective cohort study

Purpose Serum and drain amylase have been identified as important predictors of postoperative pancreatic fistula (POPF) and might be useful to guide postoperative drain management after pancreatoduodenectomy. We aimed to determine and compare the value of serum amylase and drain fluid amylase to predict postoperative pancreatic fistula after pancreatoduodenectomy. Methods This retrospective cohort study included patients after pancreatoduodenectomy from 2012 to 2019. The primary endpoint of our study was grade B/C POPF. Serum amylase on postoperative day 1 (SA-1) and drain fluid amylase on postoperative day 2 (DFA-2) were analyzed. Results A total of 92 of 437 patients (21.1%) developed a grade B/C POPF. SA-1 was higher in patients who developed a grade B/C POPF (336 U/L vs. 97 U/L, p<0.001). Similarly, DFA-2 was higher in patients who developed a grade B/C POPF (1764 U/L vs. 78 U/L, p<0.001). SA-1 and DFA-2 had similar predictive accuracy (AUC: 0.82 vs. 0.85, respectively, p=0.329). Patients with SA-1<100 U/L (n=178) had a risk of 2.2% of developing grade B/C POPF, compared to 38.2% in patients with SA-1 >100 U/L (n=207). Patients with DFA-2<100 U/L (n=141) had a risk of 0% of developing grade B/C POPF, compared to 36.2% in patients with DFA-2>100 U/L (n=196). SA-1 and DFA-2 were strongly associated at a cut-off of 100 U/L (p<0.001, 89% concordance rate). Conclusion Postoperative serum and drain amylase values below 100 U/L both effectively rule out POPF after pancreatoduodenectomy. The advantage of serum amylase measurement is that it can be used in patients who are managed without surgical drains.