(Pre-) diabetes is frequently undiagnosed and underreported in patients referred for pancreatic surgery

Abstract

Aims: Morbidity and mortality rates after pancreaticoduodenectomy remain still elevated. Recently, different scores to predict morbidity and mortality after pancreaticoduodenectomy have been published. This study aimed to validate the Braga score and the Early Mortality Risk Score (EMRS). Methods: A series of 245 patients who underwent pancreaticoduodenectomy in a single center from 2000 to 2012 was assessed. Complications were graded according to the Dindo-Clavien classification; major complications were defined as grade IIIa to IVb. The Braga score (ranging from 0 to 15) is based on pancreatic texture, main pancreatic duct diameter, blood loss, and ASA score. Score discriminant power was calculated using ROC curves. For the EMRS, only adenocarcinomas were included. Survival time and EMRS (based on age, tumor differentiation, tumor size, and comorbidities) were determined for 120 cancer patients. Relative risks (RR) of postoperative death at 9 and 12 months were calculated. Results: Major complications occurred in 31% of the 245 patients. Pancreatic texture and blood loss were statistically significant for increased morbidity. The areas under curve (C-statistic values) were 0.95 and 0.99 for 4-risk categories and for individual scores, respectively. For the EMRS group, overall median survival was 19 months. EMRS of 4 showed a mortality RR of 5.1 at 9 months (p = 0.048) and of 4.5 at 12 months (p = 0.020). Modified EMRS of 3 also showed a statistically significant increase of the mortality risk at 9 (RR = 3.7, p = 0.025) and 12 months (RR = 3.2, p = 0.017). Conclusions: The Braga score discriminated well between minor and major complications. EMRS of 4 and modified EMRS of 3 are predictors of mortality at 9 and 12 months after oncologic pancreaticoduodenectomy. Our validation studies confirmed that the Braga score could be helpful to detect patients at risk for major complications, while the EMRS may influence preoperative decision-making by predicting the short-term mortality after pancreaticoduodenectomy for adenocarcinoma.