[Prevention of pancreatic fistula after pancreatoduodenectomy]

Prospective randomized investigation of the efficiency of somatostatin analogues and glucocorticoids in pancreatic fistula prevention after pancreatoduodenectomy by using.

In period from December 2018 till March 2020 78 patients underwent pancreatoduodenectomy for pancreatobilliary tumors in department of abdominal surgery of National Medical Research Center of Surgery named after A.V. Vishnevsky. Intraoperative frozen section investigation of pancreatic functioning acinar structures (FAS) was held for all patients. 38 patients had more than 40% of FAC and were related with high risk of pancreatic fistula (PF), while 40 patients with less than 40% FAC were included in low risk of PF group. In both groups patients were randomized to main and control subgroups. In main subgroup of high risk group patients combination of somatostatin analogues and glucocorticoids was used, while in control subgroup patients received only somatostatin analogue. In low risk of PF group patients of main subgroup preventively got somatostatin analogue, while control group patients had no specific prophylaxis of PF. To assess the effect of drug prophylaxis on the development of pancreatic fistula we used logistic regression models with the inclusion of the drug use factor as an independent variable.

25 patients were included in main subgroup of high risk group. Clinically relevant pancreatic fistula (CRPF) developed in 14 (56%) cases. From 13 patients of control subgroup CRPF developed in 5 (38%) cases. In main subgroup of low risk group 18 patients were included and 3 (16%) of them had CRPF. In control subgroup were 22 patients and there were no cases of CRPF.

In our series combination of somatostatin analogue and glucocorticoid didn't show efficiency in prevention of CRPF in high risk patients, although difference between subgroups wasn't statistically significant (p=0.34). In low risk group patients prophylactic use of somatostatin analogue also didn't show decline of CRPF incidence and the difference between subgroups also wasn't statistically significant (p=0.46).