How Do We Predict the Clinically Relevant Pancreatic Fistula After Pancreaticoduodenectomy?-An Analysis in 244 Consecutive Patients

Background The most important problem in pancreatic fistula is whether one can distinguish clinical pancreatic fistula, grade B + C fistula by the International Study Group on Pancreatic Fistula (ISGPF), from transient pancreatic fistula (grade A), in the early period after pancreaticoduodenectomy (PD). It remains unclear what predictive risk factors can precisely predict which clinical relevant or transient pancreatic fistula when diagnosed pancreatic fistula on POD3 by ISGPF criteria. Methods We analyzed the predictive factors of clinical pancreatic fistula by logistic regression analysis in 244 consecutive patients who underwent PD. Pancreatic fistula was classified into three categories by ISGPF. Results The rate of pancreatic fistula was 69 of 244 consecutive patients (28%) who underwent PD. Of these, 47 (19%) had grade A by ISGPF criteria, 17 patients (7.0%) had grade B, and five patients (2.0%) had grade C. The independent risk factor of incidence of pancreatic fistula is soft pancreatic parenchyma. However, soft pancreatic parenchyma did not predict underlying clinically relevant pancreatic fistula. The independent predictive factors of clinically relevant pancreatic fistula were serum albumin level <= 3.0 g/dl on postoperative day (POD) 4 and leukocyte counts >9,800 mm(-3) on POD 4. Positive predictive value of the combination of two predictive factors for clinical relevant pancreatic fistula was 88%. Conclusions The combination of two factors on POD4, serum albumin level <= 3.0 g/dl and leukocyte counts >9,800 mm(-3), is predictive of clinical relevant pancreatic fistula when diagnosed pancreatic fistula on POD 3 by ISGPF criteria.