Prospective validation of a preoperative risk score model based on pancreatic texture to predict postoperative pancreatic fistula after pancreaticoduodenectomy.

BACKGROUND: In 2015, basing on objective preoperative factors related to pancreas remnant texture (body mass index, Wirsung duct size and preoperative diagnosis), we proposed a score model to predict the risk of postoperative pancreatic fistula after partial pancreatectomies. The aim of the present study was to prospectively validate this preoperative predictive risk score for postoperative pancreatic fistula after pancreaticoduodenectomy. METHODS: Prospective study of consecutive patients who underwent pancreaticoduodenectomy in which a preoperative risk score, based on factors related to the pancreatic texture, was calculated. The risk score model was tested by comparison with subjective intraoperative assessment of the pancreas remnant texture and drain amylase value on postoperative day 1. Sensitivity, specificity, positive and negative likelihood ratio and area under the curve were calculated. RESULTS: Eighty-four patients who underwent pancreaticoduodnectomy were analyzed. Clinically relevant pancreatic fistulas rate was 40.6%. The risk score model with a cut-off of 6 increased the odds of pancreatic fistula approximately 3 fold but it was not independently related to it. On the contrary, considering a cut-off of 5, the risk score model increased the odds of pancreatic fistula 11-16 fold and it was independently related to it. The new risk score model and pancreatic texture had high sensitivity (97% and 88%, respectively) and low specificity (34% and 60%, respectively) while the amylase drain value had low sensitivity (44%) and high specificity (92%). CONCLUSIONS: The preoperative risk score model with a cut-off of 5 was a useful predictor of clinically relevant pancreatic fistula after pancreaticoduodenectomy. The drain amylase value represents a complementary factor to the risk score in predicting a pancreatic fistula.