RESUMO
OBJECTIVES: Development of a simple preoperative risk score to predict morbidity related to pancreatic surgery. BACKGROUND: Pancreatic surgery is standardized with little technical diversity among institutions and unchanging morbidity and mortality rates in recent years. Preoperative identification of high-risk patients is potentially one of the rare avenues for improving the clinical course of patients undergoing pancreatic surgery. METHODS: Using a prospectively collected multicenter database of patients undergoing pancreatic surgery (n=703), surgical complications were classified according to the Clavien-Dindo classification. A new scoring system for preoperative identification of high-risk patients that included only objective preoperatively assessable variables was developed using a multivariate regression model. Subsequently, this scoring system was prospectively validated from 2011 to 2013 (n=429) in a multicenter setting. RESULTS: Eight independent preoperatively assessable variables were identified and included in the scoring system: systolic blood pressure, heart rate, hemoglobin level, albumin level, ASA (American Society of Anesthesiologists) score, surgical procedure, elective surgery or not, and disease of pancreatic origin or not. On the basis of 3 subgroups (low risk, intermediate risk, high risk), the proposed scoring system reached an accuracy of 75% for correctly predicting occurrence or nonoccurrence of major surgical complications in 80% of all analyzed patients within the validation cohort (c-statistic index=0.709, P<0.001, 95% confidence interval=0.657-0.760). CONCLUSIONS: We present an easily applied scoring system with convincing accuracy for identifying low-risk and high-risk patients. In contrast to other systems, the score is exclusively based on objective preoperatively assessable characteristics and can be rapidly and easily calculated.