Due to reduced mortality the number of pancreatic resections has increased during the last two decades. However, there is no consensus for many factors like indication/contraindication for surgery, type of resection, use of octreotide, indication for portal vein resection etc., because of the lack of evidence based data. Furthermore, current controversial data of ‘hospital volume’ contribute to the discussion of how and where pancreatic surgery should be performed. National comparative data regarding pancreatic surgery do not exist in Germany. To demonstrate current trends and/or differences in pancreatic surgery in Germany data were gained from six large surgical departments 1999 – 2003.

METHODS:
A survey with a standardized questionnaire was performed at six large surgical departments in Germany. All pancreatic resections from 1999 to 2003 were evaluated with annual numbers. Operative indications and techniques were assessed as well as all complications including their definitions. Overall 1354 resections were performed at the six hospitals from 1999-2003. For this study 1083 pancreatic head resections were further analysed.

RESULTS:
Annual hospital volume for pancreatic head resection was between 12 and 61. In cancer two centers performed classical Whipple procedures whereas four performed pylorus-preserving resections. Arterial infiltration is judged as contraindication to curative resection by all groups. Portal vein resection, however, varied large and was performed in between three and 27 % in malignant tumors. Duct-to-mucosa anastomoses are perfomed in only two of six centers. The frequency of pancreatic leakage did not vary much (8 – 15%), despite slight differences in definition. Only two of six centers still use octreotide routinely with no statistical differences regarding pancreatic leakage to the centers not using octretide. Peroperative morbidity was in the range between 35 and 48%. Average mortality was 2.8% (range 1.3-6%) and showed no correlation to hospital volume, surgical technique or frequency of portal vein resection.

CONCLUSIONS:
Our large survey of pancreatic resections performed in experienced centers in Germany demonstrates that pancreatic head resections can be performed with low mortality. Rather equal frequencies of mortality, morbidity and pancreatic leakage are the result of center experience and are probably independent of technical details like type of resection, anastomotic technique or use of octreotide.