A multicenter prospective study on incidence and risk factors of postoperative pancreatic fistula after radical gastrectomy: a report of 2 089 cases

Objective To investigate the incidence of postoperative pancreatic fistula (POPF) and its risk factors after radical gastrectomy. Methods The prospective study was conducted. The clinicopathological data of 2 089 patients who underwent radical gastrectomy in 22 medical centers between December 2017 and November 2018 were collected, including 380 in the Zhongshan Hospital of Fudan University, 351 in the Renji Hospital of Shanghai Jiaotong University School of Medicine, 130 in the Ruijin Hospital of Shanghai Jiaotong University School of Medicine, 139 in the Peking University Cancer Hospital, 128 in the Fujian Provincial Cancer Hospital, 114 in the First Hospital Affiliated to Army Medical University, 104 in the First Affiliated Hospital of Nanchang University, 104 in the Affiliated Hospital of Qinghai University, 103 in the Weifang People′s Hospital, 102 in the Fujian Medical University Union Hospital, 99 in the First Affiliated Hospital of Air Force Medical University, 97 in the Sir Run Run Shaw Hospital Affiliated to Zhejiang University School of Medicine, 60 in the Hangzhou First People′s Hospital Affiliated to Zhejiang University School of Medicine, 48 in the Fudan University Shanghai Cancer Center, 29 in the First Affiliated Hospital of Xi′an Jiaotong University, 26 in the Lishui Municipal Central Hospital, 26 in the Guangdong Provincial People′s Hospital, 23 in the Jiangsu Province Hospital, 13 in the First Affiliated Hospital of Sun Yat-Sen University, 7 in the Second Hospital of Jilin University, 4 in the First Affiliated Hospital of Xinjiang Medical University, 2 in the Beijing Chao-Yang Hospital of Capital Medical University. Observation indicators: (1) the incidence of POPF after radical gastrectomy; (2) treatment of grade B POPF after radical gastrectomy; (3) analysis of clinicopathological data; (4) analysis of surgical data; (5) risk factors for grade B POPF after radical gastrectomy. Measurement data with normal distribution were represented as Mean±SD, and comparison between groups was analyzed using ANOVA. Count data were described as absolute numbers or percentages, and comparison between groups was analyzed using the chi-square test. Univariate analysis was conducted using the t test or chi-square test based on data excluding missing data of clinico-pathological and surgical data. Multivariate analysis was conducted using the Logistic regression model based on factors with P<0.20 in univariate analysis. Results There were 2 089 patients screened for eligibility, including 1 512 males, 576 females and 1 without sex information, aged (62±11)years. The body mass index (BMI) was (23±3)kg/m2. (1) The incidence of POPF after radical gastrectomy: the total incidence rate of POPF in the 2 089 patients was 20.728%(433/2 089). The incidence rates of biochemical fistula, grade B pancreatic fistula, and grade C pancreatic fistula were 19.627%(410/2 089), 1.101%(23/2 089), 0, respectively. (2) Treatment of grade B POPF after radical gastrectomy: 2 of 23 patients with grade B POPF after radical gastrectomy had drainage tube placed for more than 21 days and received anti-infective therapy. Four of 23 patients with grade B POPF after radical gastrectomy had ascites detected by imaging examination, of which 2 received peritoneal drainage guided by ultrasound, 1 received failed puncture drainage, 1 received no puncture drainage, and they were given anti-infective therapy. Eleven of 23 patients with grade B POPF after radical gastrectomy had no ascites detected by imaging examinations, and they were given anti-infective therapy and inhibitors of pancreas secretion for clinical manifestation as fever or elevated white blood cells. Six patients with no typical clinical manifestations were given somatostatin to inhibite pancreas secretion and prolonged duration of abdominal drainage tube placement (with a median time of 7 days). All the 23 patients recovered well after treatment, without reoperation. (3) Analysis of clinicopathological data: for the 2 089 patients, BMI, cases with or without neoadjuvant therapy were (23±3)kg/m2, 1 487, 160 of patients without pancreatic fistula, (23±3)kg/m2, 386, 22 of patients with biochemical fistula, and (24±3)kg/m2, 22, 1 of patents with grade B pancreatic fistula, showing significant differences between the three groups (F=5.787, χ2=8.269, P<0.05). (4) Analysis of surgical data: for the 2 089 patients, cases with open surgery, laparoscopic assisted surgery, totally laparoscopic surgery (surgical method), cases with D1 lymph lode dissection, D2 lymph lode dissection, and other lymph lode dissection (range of lymph lode dissection), cases with no omentectomy, partial omentectomy, and total omentectomy (range of omentectomy), cases with no usage of energy facility, usage of CUSA, LigaSure, LigaSure+ CUSA as energy facility, cases with or without biological glue, the number of lymph node dissection were 737, 624, 292, 24, 1 580, 51, 418, 834, 381, 63, 1 530, 23, 16, 1 431, 201, 33±14 of patients without pancreatic fistula, 146, 189, 74, 11, 389, 9, 110, 171, 128, 35, 359, 6, 9, 378, 31, 31±14 of patients with biochemical fistula, and 14, 5, 4, 0, 20, 3, 6, 13, 4, 2, 18, 1, 2, 22, 1, 37±16 of patients with grade B pancreatic fistula, showing significant differences between the three groups (χ2=15.578, 9.397, 15.023, 28.245, 8.359, F=4.945, P<0.05). (5) Risk factors for grade B POPF after radical gastrectomy: results of univariate analysis showed that usage of energy facility was a related factor for grade B POPF after radical gastrectomy (χ2=9.914, P<0.05). Results of multivariate analysis showed that laparoscopic assisted surgery, combined evisceration, application of LigaSure + CUSA, the number of lymph lode dissection were independent factors for for grade B POPF after radical gastrectomy (odds ratio=0.168, 3.922, 9.250, 1.030, 95% confidence interval: 0.036-0.789, 1.031-14.919, 1.036-82.602, 1.001-1.059, P<0.05). Conclusions The incidence of grade B POPF after radical gastrectomy is relatively low. Laparoscopic assisted surgery, combined evisceration, application of LigaSure + CUSA, and the number of lymph lode dissection are independent risk factors for grade B POPF. Trial Registration: This study was registrated at ClinicalTrial.gov in United States with the registration number of NCT03391687. Key words: Gastric neoplasms; Radical resection; Postoperative pancreatic fistula; Influencing factor; Prospective study