Pancreatic fistula and relative mortality in malignant disease after pancreaticoduodenectomy. Review and statistical meta-analysis regarding 15 years of literature.

This paper analysed the literature published in the last 15 years regarding the onset of pancreatic fistula after pancreaticoduodenectony carried out for tumours in the periampullary region, in the head of the pancreas and in the distal common bile duct. Out of 8370 pancreatic resections we were able to go by only 2684 cases, which showed the type of treatment used in the remaining stump, the rates of leakage and relative mortality. The data collected were analysed statistically using the Cochran test and or the chi 2, evaluating the possible significant difference relative to the various methods of reconstruction. The onset of pancreatic fistrula was found to be statistically more frequent after ligation of the stump than after pancreatico-jejunal anastomosis (p = 0.001). Comparing the pancreatico-jejunal end-to-side anastomosis, to pancreatico-jejunal end-to-end and wirsung-jejunal end-to-side anastomosis the first one had shown a significantly higher rate only for leakage (respectively p = 0.008 and p = 0.010). The occlusion of the wirsung duct with biological substances showed better results compared to ligation (p = 0.001) only as regards onset of the fistula, while the comparison between the occlusion and the three types of anastomosis did not show any statistically difference, except for the pancreatico-jejunal end-to-side anastomosis, in which it was significant only as regards leakage (p = 0.009). The statistical analysis between pancreatico-gastrostomy and pancreatico-jejunal anastomoses indicated that the first technique had a lower morbidity rate than pancreatico-jejunal end-to-side (p = 0.001), pancreatico-jejunal end-to-end (p = 0.010) and wirsung-jejunal end-to-side (p = 0.011). We analysed and compared the results obtained before and after 1975, in order to discover whether was an improvement in the prevnetion or in the treatment of such a complication and its consequences. Furthermore, we tried to establish whether the transanastomotic drainage, the site of the neoplasm, the texture of pancreatic parenchyma and the patient's age could in any way influence the onset and course of the fistula.