Objective To investigate the risk factors and treatment of postpancreatico-duodenectomy hemorrhage(PPH).

Methods The retrospective case-control study was conducted. The clinical data of 712 patients who underwent pancreaticoduodenectomy in Peking University First Hospital from January 2012 to November 2021 were collected. There were 392 males and 320 females, aged from 16 to 89 years, with a median age of 62 years. Observation indicators: (1) diagnosis of PPH; (2) analysis of influencing factors for PPH; (3) treatment of PPH. Measurement data with skewed distribution were represented as M(range). Count data were described as absolute numbers or percentages. Univariate analysis was performed using the chi-square test or Fisher exact probability, and multivariate analysis was performed using the Logistic regression model.

Results (1) Diagnosis of PPH. Of the 712 patients, 72 cases had PPH and 7 cases died. The incidence of PPH was 10.11%(72/712), and PPH related mortality was 9.72%(7/72). There were 7 cases of early PPH and 65 cases of delayed PPH. There were 23 cases of mild PPH and 49 cases of severe PPH. (2) Analysis of influencing factors for PPH. Results of univariate analysis showed that preoperative serum total bilirubin (TBil), extended surgery, postoperative pancreatic fistula, postoperative biliary fistula, postoperative abdominal infection were related factors for delayed PPH (χ²=13.17, 3.93, 87.89, 22.77, 36.13, P<0.05). Results of multivariate analysis showed that preoperative serum TBil ≥171 μmol/L, postoperative grade B or C pancreatic fistula, postoperative biliary fistula, postoperative abdominal infection were independent risk factors for delayed PPH (odds ratio=1.91, 8.10, 2.11, 2.42, 95% confidence interval as 1.09-3.33, 4.62-14.20, 1.06-4.23,1.35-4.31, P<0.05). (3) Treatment of PPH. ① Treatment of early PPH. Of the 7 cases with early PPH, 4 cases had mild PPH and 3 cases had severe PPH. The 4 cases with mild PPH were stanched by conservative treatment. The bleeding location of the 3 cases with severe PPH were the posterior wall of pancreatoenteric anastomosis, the pancreatic uncinate stump and the unintentional puncture of the jejunostomy tube of the left upper abdominal wall vessels and the 3 cases were stanched by reoperation. All the 7 cases were discharged without other complications. ② Treatment of delayed PPH. Of the 65 cases with delayed PPH, 19 cases had mild PPH and 46 cases had severe PPH. Of the 19 cases with mild PPH, 18 cases were stanched by conservative treatment including 2 cases died of pancreatic fistula and abdominal infection, 1 case were stanched by endoscope therapy. Of the 46 cases with severe PPH, 18 cases with stable vital signs and slow bleeding were stanched by conservative treatment including 1 case died of infectious toxic shock and the other 28 cases underwent invasive treatment, including 2 cases undergoing gastroscopy, 20 cases undergoing interventional treatment and 6 cases under-going reoperation as the initial treatment. Of the 22 cases taking endoscope or interventional treatment as the initial treatment, 5 cases underwent rebleeding and 2 cases died, with the reblee-ding rate and mortality as 22.7%(5/22) and 9.1%(2/22), respectively. Of the 6 cases taking reopera-tion as the initial treatment, 3 cases underwent rebleeding and 2 cases died, with the rebleeding rate and mortality as 3/6 and 2/6, respectively. There was no significant difference in the rebleeding rate and mortality in patients taking endoscope or interventional treatment as the initial treatment and patients taking reoperation as the initial treatment (P>0.05). Of the 28 cases undergoing invasive treatment, 10 cases underwent secondary surgical treatment, including 6 cases taking reoperation and 4 cases taking interventional treatment as the initial treatment for hemorrhage, and 4 cases died with the mortality as 4/10, and the other 18 cases who did not receive secondary surgical treatment survived. There was a significant difference in the mortality between patients with or without secondary surgical treatment (P<0.05).

Conclusions Preoperative serum TBil ≥171 μmol/L, post-operative grade B or C pancreatic fistula, postoperative biliary fistula, postoperative abdominal infection are independent risk factors for delayed PPH. Surgical treatment should be performed decisively for early severe PPH. For delayed severe PPH patients who undergoing conservative treat-ment without effect, endoscope therapy and interventional treatment should be the first choice, and surgical treatment should be performed if those above procedures not working.