Sixty consecutive patients (31 men and 29 women; mean age, 64 years; range, 36-81 years) with a malignant hilar biliary obstruction and contralateral portal vein steno-occlusion, i.e., severe stenosis (> 90% luminal narrowing on computed tomography [CT]) or complete occlusion (encasement of the portal vein with no demonstrable flow or an intraluminal filling defect), caused by tumor invasion or pre-operative embolization for resection from October 2010 to October 2013, were treated by percutaneous placement of biliary metallic stents. Thirty-nine of these patients had a portal vein occlusion. Twelve had severe portal vein stenosis caused by underlying malignant tumors, whereas nine had a portal vein occlusion caused by preoperative portal vein embolization using vascular plugs, coils, or Gelfoam particles. A curative surgical treatment could not be performed in these nine patients due to major vessel invasion or peritoneal seeding. Demographic, clinical, and laboratory data were collected from the medical records of all patients or from the electronic patient information database. Accurate and definitive follow-up data were obtained for all 60 patients. This retrospective study was approved by our Institutional Review Board, and the requirement for written, informed consent from each patient was waived.

The patient characteristics are presented in Table 1. The diagnosis was confirmed histopathologically in 46 patients (77%) through a percutaneous transbiliary forceps biopsy in 25 patients and by endoscopic brush cytology and forceps biopsy in 21 patients. The diagnoses were presumed in the other 14 patients (23%) based on CT and magnetic resonance cholangiopancreatography results. Patients were included if they had a malignant biliary hilar obstruction and a unilobar portal vein steno-occlusion or embolization that could not be treated surgically due to unresectability, late tumor stage, advanced age, co-morbidities, or if endoscopic attempts to drain obstructed bile ducts were unsuccessful. Exclusion criteria included abnormal coagulation status (international normalized ratio value ≥ 1.5), platelet count ≤ 50000, or a generally poor health status (Eastern Cooperative Oncology Group performance status grade 3-4).

Computed tomography or magnetic resonance cholangiography was performed to access patient anatomy and determine the most appropriate interventional approach before percutaneous transhepatic biliary drainage (PTBD). All PTBD procedures were performed in the peripheral bile duct, and the patients were under conscious sedation with intravenous pethidine hydrochloride (Demerol; Keukdong, Seoul, Korea) and with local anesthesia using intramuscularly injected lidocaine (Jeil, Daegu, Korea) at the PTBD site. The intrahepatic bile duct was punctured using a 21-gauge Chiba needle (Cook Medical, Bloomington, IN, USA) under fluoroscopic guidance or ultrasonography, and a 8.5-Fr drainage catheter was inserted to relieve jaundice or cholangitis before inserting the stent. Stents were placed for less than 2 weeks after PTBD; thus, allowing any cholangitis to be treated. Antibiotics were administered intravenously 2 hours before the procedures and for at least 48 hours after. Intravenous antibiotic therapy was maintained in patients with cholangitis until the treatment was successful.

The stent insertion procedure was as follows. A 5-Fr Kumpe catheter (Cook Medical) and a 0.035-inch guidewire (Radifocus Guide Wire M; Terumo, Tokyo, Japan) were advanced across the hilar obstruction to the common bile duct or duodenum. Balloon dilation was performed in all patients before placing the stent using a 6-8-mm-diameter angioplasty balloon catheter (Synergy; Boston Scientific, Galway, Ireland).

All stents deployed were uncovered (Zilver [Cook Medical], Sentinol stents [Boston Scientific]), partially expanded polytetrafluoroethylene (ePTFE)-covered (Hercules [S&G Biotech, Seongnam, Korea], Comvi [TaeWoong Medical, Gimpo, Korea], or GD stents [TaeWoong Medical]). The Zilver and Sentinol stents were made of nitinol with laser cutting and have high flexibility and minimal foreshortening during deployment. All covered stents were partially ePTFE-covered nitinol stents. The partially ePTFE-covered stents had bare extensions of 2 or 3 cm at the proximal end to prevent tumor overgrowth, migration, and intrahepatic bile duct occlusion. All stents were available in 8 and 10 mm diameters and in 6 and 8 cm lengths. The type, diameter, and length of each stent were decided by physician's preference.

Our stent deployment system was introduced over a 0.035-inch, 150-cm long stiff hydrophilic guidewire (Terumo) or over an extra-stiff Amplatz guidewire (Cook Medical) and was deployed across the hilar obstruction to the common bile duct to cover the bile duct approximately 1-2 cm proximal and 2-3 cm distal to the obstruction to prevent tumor overgrowth. Post-stenting balloon dilation was not performed in any patient. After the procedure, an 8.5-Fr or 10-Fr temporary drainage catheter (Cook Medical) was placed just proximal to the stent. The external drainage catheter was irrigated frequently for 1-2 days after placement to remove any possible blood clots or sludge. The temporary drainage catheter was removed after 2-3 days of catheter clamping, and without further intervention if free contrast flowed through the stent into the common bile duct and the duodenum. Information regarding the current health status or death of patients was obtained by telephone or from their medical records.

The study endpoints were the rates of technical success, complications, successful internal drainage, stent patency, and patient survival time. Technical success was defined as stent deployment in an appropriate position across the stricture with good contrast passage through the stent. Complications were classified as major and minor according to the guidelines of the Society of Interventional Radiology Standards of Practice Committee (15). Successful internal drainage was defined as successful removal of the temporary drainage catheter and a decrease in serum bilirubin level to < 75% of the pretreatment value within the first month following placement of the stent. Patient survival time was defined as the time interval between initial stent placement and death. Stent patency was defined as the period between initial stent placement and recurrence of the obstruction or death with a patent stent. A stent was assumed to be patent at the time of death if serial serum bilirubin levels were normal or only mildly elevated (< 3 mg/dL). Stent occlusion was defined as a radiologically confirmed biliary obstruction with a serum bilirubin level > 3 mg/dL or as any condition requiring a repeat intervention to improve biliary drainage. Sludge formation was defined as sludge occluding the stent lumen and that could be cleared by a biliary balloon sweep. Tumor ingrowth was defined as tissue occluding the stent lumen and that could not be cleared by a biliary balloon sweep. Tumor overgrowth was defined as a new stricture occurring at the proximal or distal end of a stent.

All statistical analyses were conducted using SPSS ver. 14.0 software (SPSS Inc., Chicago, IL, USA) A p value < 0.05 was considered statistically significant. The paired-sample t test was used to compare serum bilirubin levels before and after stent placement. Patient survival time and stent patency were estimated using a life-table analysis, according to the Kaplan-Meier method. The comparison of stent patency between covered and uncovered metallic stents was conducted using the log-rank test.

Stents were successfully placed in all 60 patients. Control cholangiography performed immediately following stent placement confirmed correct positioning of all stents. None of the covered stents migrated after deployment. The 60 patients received a total of 70 stents. A single stent was sufficient to relieve the malignant hilar obstruction in 51 patients (Fig. 1), whereas two stents were placed in the ipsilateral bile ducts of the remaining nine patients with multiple separation of the ipsilateral, intrahepatic bile duct (Fig. 2). Twenty-seven patients received 33 covered stents, and 33 patients received 37 uncovered stents of different sizes.

Procedural-related minor complications occurred in six (10%) patients. Two patients experienced self-limiting hemobilia that resolved completely without a transfusion 1-3 days following stent placement. Cholangitis occurred in four patients, two of whom underwent placement of a covered stent and two who underwent placement of an uncovered stent. However, all four patients with cholangitis showed favorable clinical courses within 4 days following antibiotic treatment and without the need for additional intervention. Two patients developed acute cholecystitis 11 and 350 days, retrospectively, following placement of a biliary stent and underwent percutaneous gallbladder drainage.

The follow-up cholangiography performed 3-6 days after placing the stents showed adequate decompression of the biliary ducts in 54 patients. The temporary drainage catheters were removed in these 54 patients after confirming good passage of contrast medium through the stent. The temporary drainage catheters could not be removed in six patients who underwent placement of uncovered (n = 3) and covered (n = 3) stents due to persistently high serum bilirubin levels (n = 5) or stent dysfunction (n = 1). One patient had an uncovered stent dysfunction caused by early tumor ingrowth; however, the temporary drainage catheter could not be removed due to the patient's refusal to undergo additional stent placement. Mean serum bilirubin level, which was 10.35 ± 8.0 mg/dL before drainage, decreased significantly to 1.96 ± 2.3 mg/dL 1 month after placing the stent (p < 0.001). Successful internal drainage was achieved in 54 (90%) of the 60 patients.

All patients were clinically followed until their death or until the end of the study, and the cutoff date for data analysis was December 31, 2013. During the mean follow-up period of 268 days (range, 27-862 days), 55 patients died, and five patients survived. One patient died within 30 days after placement of a stent due to rapid progression of underlying malignancies; death was not directly related to the procedure. According to the Kaplan-Meier analysis, the median survival time was 210 days (95% confidence interval [CI], 135-284 days) (Fig. 3).

According to the Kaplan-Meier analysis, median stent patency time was 133 days (95% CI, 94-171 days), and the cumulative stent patency rates at 1, 3, 6, 9, and 12 months were 100%, 97%, 46%, 28%, and 21%, respectively (Fig. 4). No significant difference in stent patency was observed between covered and uncovered stents (p = 0.646) (Fig. 5). Stent dysfunction occurred in 16 (29.6%) of 54 patients after a mean of 159 days (range, 65-321 days). Stent dysfunction caused by tumor ingrowth occurred in all 10 patients who received an uncovered stent, whereas stent dysfunction caused by sludge incrustation (n = 5) or tumor overgrowth (n = 1) occurred in only six patients who received a covered stent (Table 2). Stent dysfunctions were managed using the PTBD procedure in 16 patients. Four patients with a stent dysfunction caused by tumor ingrowth (n = 3) or sludge incrustation (n = 1) received another covered stent, resulting in good internal drainage before these patients died. An additional metallic stent was not placed due to disease progression and poor general health in the remaining 12 patients.