Persisting ratio of total amylase output in drain fluid can predict postoperative clinical pancreatic fistula

Background/purpose A consistent predictor for pancreatic fistula (PF) development in the early period after pancreatic resection is still lacking. Patients and methods A total of 54 consecutive patients undergoing pancreatic resection were enrolled between June 2007 and April 2010. Pancreaticoduodenectomy (PD) and distal pancreatectomy (DP) were performed in 38 and 16 patients, respectively. For the purpose of finding an early predictor for PF development, we investigated drain amylase levels (d-Amy, IU/mL), drain output volume (d-Vol, mL/day) and drain amylase output (Amy-V, IU/day) on postoperative days (POD) 1 and 3. Amy-V was calculated as the product of d-Amy and d-Vol, and was expressed as the sum of values obtained from all drains. In addition, the ratio of d-Amy or Amy-V on POD3 to that on POD1 was calculated as the persisting ratio in each patient. Results The overall incidence of clinical PF (International Study Group on Pancreatic Fistula Grade B and C) was 16.7%, occurring in 13.1% after PD and 25% after DP. All PF occurred in cases with a remnant pancreas of soft texture. There was no significant difference in d-Amy, d-Vol, or Amy-V on POD1 and POD3 between patients with and without clinical PF. The persisting ratio of Amy-V was significantly lower in patients without clinical PF compared to those with clinical PF (P = 0.029). Furthermore, the persisting ratio of Amy-V was significantly lower in patients with Grade A PF compared to those with Grade B PF (P = 0.03). Conclusion The persisting ratio of drain amylase output is a new significant predictive factor for clinical PF development.