Abstract
BACKGROUND: Postoperative day 1-drains amylase (POD1-DA) values are commonly used to predict the risk of pancreatic fistula (PF) after pancreaticoduodenectomy (PD). Perioperative inflammatory biomarkers have been associated to higher risk of complications in different oncological surgeries. Aim of this study was to investigate the utility of the combination of preoperative inflammatory biomarkers (PIBs) with POD1-DA levels in predicting grade C PF. MATERIALS AND METHODS: From a prospective collected database of 317 consecutive pancreaticoduodenectomies, data regarding POD1-DA levels and PIBs as neutrophil-to-lymphocyte ratio (NRL), derived neutrophil-to-lymphocyte ratio (dNRL), platelet-to-lymphocyte ratio (PLR) were analyzed in 227 cases. P-values <0.05 were considered statistically significant. Receiver operating characteristic (ROC) curves defined the optimal thresholds for biomarkers and drains amylase values and their accuracy to predict PF. Furthermore, the Positive Predictive Value (PPV) was computed to evaluate the probability to develop PF combining PIBs and drains amylase values. Combination of drains amylase and different PIBs cut-offs were used to evaluate the risk of grade C PF. RESULTS: A POD1-DA level of 351 U/L significantly predicted PF (sensitivity 82.7%, specificity 76%, AUC 0.836; p < 0.001) with a PPV of 76.5% and a NPV of 82.6%.POD1-DA levels ≥807 U/L significantly predicted grade C PF (sensitivity 72.7%, specificity 64.4%, AUC 0.676; p = 0.004) with a PPV of 17.8% and a NPV of 95.6%.Notably, this last PPV increased from 17.8% to 89% when PIBs, at different cut-offs, were combined with POD1-DA at the value ≥ 807 U/L. CONCLUSION: PIBs significantly improve POD1-DA ability in predicting grade C PF after PD.