Clavien-Dindo classification and risk factors of complications after pancreaticoduodenectomy

Abstract

Objective: To analyze Clavien-Dindo classification and risk factors of complications after pancreaticoduodenectomy and investigate the relationship between the major risk factors and Clavien-Dindo classification of complications. Methods: The retrospective case-control study was adopted. The clinical data of 200 patients who underwent pancreaticoduodenectomy at the Third Affiliated Hospital of Inner Mongolia Medical University from January 2010 to June 2015 were collected. The patients underwent Whipple procedure or pylorus-preserving pancreaticoduodenectomy according to the tumor site. Observation indicators included: (1) postoperative complications using Clavien-Dindo classification; (2) univariate and multivariate analyses: patients’ basic information, surgery-related factors, pancreas-related factors; (3) relationship between independent risk factors and Clavien-Dindo classification of complications after pancreaticoduodenectomy. The chi-square test was applied to univariate analysis and categorical data. The comparison between groups was done by using independent samples nonparametric test (Kolmogorov-Smirnov Z), and multivariate analysis was done by using Logistic regression model. Results: (1) Postoperative complications: Of 200 patients, 122 underwent Whipple procedure and 78 underwent pyloruspreserving pancreaticoduodenectomy, including 6 cases combined with vascular reconstructions and 1 case with RFA of liver tumors. Ninety-eight patients had postoperative complications, including 41 patients with no less than 2 types of complications. After surgery, pancreatic fistula was detected in 80 patients, including 42 cases with grade A, 28 cases with grade B and 10 cases with grade C; incisional infection in 29 patients; gastric retention in 24 patients; intra-abdominal infection in 16 patients; intraabdominal hemorrhage in 10 patients, including 8 patients receiving interventional treatment; biliary leakage in 7 patients and unplanned reoperation in 2 patients. Three patients were dead during hospitalization. The incidences of complications in grade I, II, III (III a and III b), IV and V of Clavien-Dindo classification were 28.00% (56/200), 13.00% (26/200), 5.00% (10/200), 1.50% (3/200) and 1.50% (3/200). (2) Univariate and multivariate analyses: The results of univariate analysis showed that body mass index (BMI) and pancreas texture were risk factors affecting complications after pancreaticoduodenectomy (χ 2 = 6.483, Z = -3.189, p < .05). The results of multivariate analysis showed that BMI > 23.9 kg/m 2 and soft pancreas were independent risk factors affecting complications after pancreaticoduodenectomy (OR = 2.044, 1.649, 95% confidence interval: 1.212-3.447, 1.194-2.275). (3) The relationship between independent risk factors and Clavien-Dindo classification of complications after pancreaticoduodenectomy was analyzed. There were statistically significant differences between BMI or pancreas texture and Clavien-Dindo classification of complications after pancreaticoduodenectomy (χ 2 = 13.897, 27.077, p < .05). Conclusions: Clavien-Dindo classification of complications after pancreaticoduodenectomy contributes to comprehensive comparison and evaluation, and this type of classification in this study mainly refers to grade I and II. Reducing BMI and good management of pancreatic stump may improve Clavien-Dindo classification of complications after pancreaticoduodenectomy.