Back to 2011 Program

Background: Recent reports suggested that the drain amylase value in drain on the postoperative day one is the useful predictive factor for postoperative pancreatic fistula (POPF) after pancreatic resection. However, the predictive clinical data for clinically relevant POPF after pancreaticoduodenectomy have not been clearly established. Methods: From Dec. 2003 to Oct. 2010, prospectively collected data from 175 consecutive patients who underwent pancreaticoduodenectomy with two-layered duct to mucosa pancreaticogastrostomy were evaluated. The predictive clinical data (WBC, serum amylase, serum albumin, C-reactive protein, drain amylase, drain fluid volume, etc.) for clinically relevant POPF (ISGPF Grade B and C) were analyzed by logistic regression analysis. Results: Of 175 patients, 31 (18%) developed pancreatic fistula by ISGPF criteria; Grade A in 21 patients (12%), Grade B in 8 (5%), and Grade C in 3 (2%). By univariate analysis, drain amylase on postoperative day (POD) 2, 3, 4, 5, C-reactive protein POD 3 and 4 were found to be significantly associated with clinically relevant POPF (p<0.05). By multivariate analysis, the independent predictive factor for clinically relevant POPF was C-reactive protein on POD 4 (p<0.018). Based on the receiver operating characteristic curve analysis, C-reactive protein >18mg/dl on POD 4 displayed the optimal sensitivity (64%) and specificity (93%). Conclusions: C-reactive protein >18mg/dl on POD 4 is the predictive factor for clinically relevant POPF after pancreaticoduodenectomy when diagnose POPF by ISGPF criteria.