To asses drains amylase (DA) cut-offs for the risk of clinically relevant postoperative pancreatic fistula (POPF) and define the optimal timing of drains removal based on daily DA assay and abdominal CT scan finding after pancreatoduodenectomy (PD). Different algorithms able to identify patients at higher risk of POPF and to assess the optimal time for drains removal after PD have been proposed. The most accurate DA cut-offs in the assessment of the risk of clinically relevant POPF were retrospectively identified. Data from a prospective trial for optimal timing of drains removal were analyzed. Then, to validate the cut-offs identified in the first phase, they were applied to the patients enrolled in the prospective trial. Patients with POD1 DA ≥ 666 U/L were at higher risk of clinically relevant POPF (p 0.0001). POD3 DA value ≥ 252 U/L predicted 88% of clinical relevant fistulas. POD3 DA level ≥ 207 U/L was able to predict 68% of biliary fistulas. Patients with abdominal collection ≥ 5 cm, showed a significantly higher rate (60% vs. 23%, p < 0.001) of biliary fistula. Timing of drains removal did not influence complications. Drains amylase levels predict clinically relevant POPF. Drains should be maintained up to POD3; in case of POD1 DA levels < 666 U/L and POD3 DA levels < 252 U/L drains could be removed. In case of POD3 DA levels, ≥ 207 the routine use of abdominal CT scan in the same day could be justified to detect collections ≥ 5 cm and maintain drains beyond the POD3.
Keywords: Pancreas anastomosis; Pancreatic fistula; Pancreatic surgery; Pancreatic surgery complications; Surgical drains management.