Patients selection
A retrospective study involving 226 patients with HCCA who underwent EBD or PTBD at the Department of Hepatobiliary and Echinococcosis Surgery, Digestive and Vascular Surgery Center, the First Affiliated Hospital of Xinjiang Medical University from January 2016 to December 2020 was conducted. Study patients were selected according to the following eligibility indications: (i) Age greater than 18 years old without gender limitations; (ii) Preoperative diagnosis of HCCA through imaging; (iii) Biochemical examination, especially serum bilirubin and direct bilirubin index, suggested obstructive jaundice; (iv) Received primary EBD or PTBD due to obstructive jaundice caused by HCCA. However, patients were excluded from the study according to the following criteria: (i) Exploratory laparotomy was performed before operation; (ii) Received preoperative chemotherapy, radiotherapy and other anti-tumor treatment; (iii) Contraindications or poor tolerance to EBD or PTBD treatment; (iv) Biochemical examination suggested non-obstructive jaundice; (v) Patients with congenital malformations of the biliary system; (vi) Severe impaired liver functions and grade C Child-Pugh classification; (vii) Patients with biliary stones or other benign biliary strictures; (viii) Patients with previous surgical history of common biliary duct; (ix) Patients accompanied with biliary tract infection or pancreatitis at admission; (x) Woman patients with pregnancy. All patients routinely underwent preoperative abdominal US, abdominal CT, and magnetic resonance cholangiopancreatography (MRCP).
According to different modes of biliary drainage, the patients were divided into EBD group (n=68) ans PTBD group (n=158). Among these patients, 26 cases were excluded due to previous surgical history of common biliary duct including ten cases in EBD group and 16 cases in PTBD group. Moreover, seven patients in EBD group were also excluded due to benign hepatobiliary diseases. Due to incomplete postoperative clinical data, 29 patients, including five cases in EBD group and 24 cases in PTBD group, were also excluded. In addition, 31 patients with non-obstructive jaundice were excluded, who were included five cases in EBD group and 26 cases in PTBD group. Ten patients, including three cases in EBD group and seven cases in PTBD group, were excluded due to poor liver functions, which were classified into Child-Pugh grade C. One patient was excluded due to combined PTBD and EBD therapy, and eight patients in PTBD group were excluded due to open surgery before PTBD. Then, 114 patients were included in this study, including 37 cases in EBD group and 77 cases in PTBD group, among whom 11 patients in EBD group and two patients in PTBD group failed operation. Two group patients were 1:1 matched using propensity score matching (PSM) analysis [17], and 19 pairs of HCCA patients were used for further analysis. Detailed study diagram was shown in Fig.1.
Operation techniques
In EBD group, duodenoscopy was used to cut the Oddi sphincter at the duodenal papilla, and the contrast catheter was inserted. Then, contrast agent was injected to develop the biliary duct and pancreatic duct. When the obstruction site was confirmed, the guide wire was inserted along the contrast catheter. Finally, the stent or nasobiliary duct was inserted along the guide wire after the guide wire retrogradely passes through the obstruction site.
In PTBD group, dilated biliary duct was punctured under digital subtraction angiography (DSA) monitoring. After extracting the bile, the guide wire was inserted into the biliary duct along the puncture needle. Then, a stent or drainage tube for internal drainage or internal and external drainage was placed after the guide wire passes through the obstruction site. If the guide wire failed to pass through the obstruction site, external drainage was performed.
Data collection and follow up
General situation of the patients, including gender, age, body mass index (BMI), hypertension, diabetes, coronary heart disease (CAD), hepatitis was collected through reviewing medical records. In addition, collected indicators for liver function were as follows: Child-Pugh grading, Bismuth-Corlette classification, preoperative and postoperative total bilirubin (TBIL), aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ- Glutamyl transpeptidase (γ-GGT) and alkaline phosphatase (ALP). Postoperative complications were included electrolyte disorder, acute pancreatitis, biliary bleeding, biliary tract infection, falling of stent/drainage tube, bile leakage and hypoproteinemia. Clinical data related to patient hospitalization, such as postoperative hospital stay, hospitalization expenses and operation time were also collected.
The PTBD group should wait for 2-4 hours after operation until the anesthetic reaction disappears, and patients without nausea and vomiting and other uncomfortable symptoms can gradually resume a liquid or semi-liquid light diet. In the EBD group, fasting was routine after operation, and the diet should be decided according to the blood biochemical indexes of patients in the early morning of the next day. If the indexes of white blood cell, serum amylase and serum lipase were not high, the liquid or semi-liquid light diet could be resumed gradually.
The liver function related indexes such as TBIL, AST and ALT were measured again at the same time of drainage in both groups to assess the drainage effect. Patients were observed at postoperative follow-up from this biliary drainage until discharge or until the next surgical intervention.
Definition of Events
Technical success is defined as the successful placement of a stent or catheter for drainage at the site of malignant lesion invasion. Definition of acute pancreatitis: (i) The patient suddenly presents with acute persistent mid-upper abdominal pain. (ii) Serum amylase exceeds the upper limit of normal by more than 3 times. (iii) Imaging changes of pancreatic inflammation such as pancreatic enlargement and peripancreatic exudation on ultrasound or CT examination. The diagnosis of acute pancreatitis can be confirmed by meeting two of the above three diagnostic criteria. Bile duct infection was defined as a patient with a temperature > 38.5°C and no other demonstrable cause, along with biochemical evidence of cholestasis and infection (increased C-reactive protein and white blood cells). Procedure-related biliary bleeding after biliary drainage was defined as bleeding that required blood transfusion or additional intervention, or bleeding that resulted in a 2 g/dL drop in hemoglobin levels. Definition of stent or drainage tube dislodgement: (i) stent dislodgement: postoperative patients undergoing abdominal X-ray or endoscopic retrograde cholangiopancreatography (ERCP) clearly have a biliary stent that is not in its original position and has partially dislodged or completely dislodged; (ii) drainage tube dislodgement: postoperative patients with drainage tubes dislodged from the puncture site due to loose or damaged body surface drainage tube fixation devices. Bile leak is defined as a patient presenting with postoperative signs of peritoneal irritation, bile aspirated by laparotomy or bile accumulation in the abdominal cavity found by reoperation.
PSM analysis
In order to avoid confounding factors and differences caused by intrinsical baseline characteristics, matched patients was selected to compare the short-term efficiency between groups through PSM analysis [17]. Logistic regression model including 14 variables was used to estimate propensity score, which were as follows: gender, age, BMI, hypertension, diabetes, CAD, hepatitis, Child-Pugh grading, Bismuth-Corlette type [18], preoperative TBIL, preoperative AST, preoperative ALT, preoperative γ-GGT and preoperative ALP. The nearest neighbor matching method was used to 1:1 match the two group patients within 0.02 standard deviation without replacement.
Statistical analysis
Statistical analysis was performed using SPSS version 26.0 (SPSS Inc, Chicago, IL, USA). Quantitative data confirming to normal distribution were presented as the mean ± standard deviation, and statistical comparison between groups was made by Student’s t test. Relatively, quantitative data without normal distribution were presented with the median (Q1, Q3), and statistical comparison between groups was made by Mann-Whitney U test or Wilcoxon rank sum test. Qualitative data were presented with ratio or percentage (%), and statistical comparison between groups was made by Chi square test or Fisher exact probability test. Statistical significance was set at the 5% level and P< 0.05 was considered statistically significant.