[Investigations on in-hospital mortality in pancreatic surgery : Results of a multicenter observational study].

backgroundThe rate of hospital mortality (in-hospital mortality) after complex pancreatic resections cannot be used as a decision-making criterion with no further analysis and specification. Such analysis has to provide a risk-adjusted benchmarking including a continuous evaluation taking into account the frequency of a surgical procedure and its competent perioperative management. MATERIAL AND

methodsAs part of the Prospective Evaluation study Elective Pancreatic surgery (PEEP), overall 2003 patients were enrolled over a 3-year time period from 01 January 2006 to 12 December 2008, who underwent elective pancreatic surgery in 27 surgical departments. Included in the study were only hospitals which perform pancreatic resections. In addition to the analysis of the current situation of the operative treatment of pancreatic diseases, the complex aspects of the in-hospital mortality as a main outcome parameter were investigated.

resultsOut of all enrolled patients (n = 2003), 75 patients (3.7%) died during the hospital stay. In the group of 1045 patients with partial pancreaticoduodenectomy (PD), 43 patients did not survive the hospital stay (4.1%). Similarly, such low in-hospital mortality rates were observed after total pancreatoduodenectomy (3.8%) and after left-sided resection of the pancreas (1.9%). With respect to a univariate risk stratification, advanced age and an American Society of Anaesthesiologists (ASA) score of 3 and 4 had a significant impact on in-hospital mortality. Multivariate regression analysis within the PD group revealed an increased need for blood transfusions and a delay in oral feeding as factors closely associated with specific complications with a significant impact on in-hospital mortality. Significant differences in the in-hospital mortality rates were found when comparing hospital volume groups, such as 10-20 vs. >20 cases/year for the 831 Kausch-Whipple procedures for adenocarcinoma and chronic pancreatitis. DISCUSSION: An adequate in-hospital mortality rate in the continuous benchmarking represents an acceptable quality level of structural and therapeutic predictions in pancreatic resections. The participation of surgical departments with complex oncosurgical interventions in clinical multicenter observational studies as a contribution to research on surgical care appears reasonable and recommendable since the results of such studies can provide a contribution to decision-making processes in daily surgical practice.