Abstract
Background
Temporary placement of removable, fully covered, self-expandable metal stents (fcSEMS) for treatment of benign biliary strictures (BBS) has been reported to be effective. However, the optimal extraction time point remains unclear and stent migration has been a major concern.
Objective
The objective of this study was to evaluate the efficacy and safety of this treatment modality using an fcSEMS with a special antimigration design and prolonged stent indwell time.
Methods
We performed a prospective, single-arm study at six tertiary care centers in Europe. Patients with BBS underwent endoscopic or percutaneous implantation of an fcSEMS (GORE® VIABIL® Biliary Endoprosthesis, W.L. Gore & Associates, Flagstaff, AZ, USA). The devices were scheduled to be removed nine months later, and patients were to return for follow-up for an additional 15 months.
Results
Forty-three patients were enrolled in the study. Stricture etiology was chronic pancreatitis in the majority of patients (57.5%). All fcSEMS were placed successfully, either endoscopically (76.7%) or percutaneously (23.3%). Stent migration was observed in two patients (5.2%). Primary patency of the SEMS prior to removal was 73.0%. All attempted stent removals were successful. At removal, stricture was resolved or significantly improved without need for further therapy in 78.9% of patients. Stricture recurrence during a follow-up of two years post-implant was observed in two patients.
Conclusions
Temporary placement of the fcSEMS is a feasible, safe and effective treatment for BBS. The design of the device used in this study accounts for very low migration rates and facilitates easy stent retrieval, even after it has been in place for up to 11 months.
Introduction and background
Endoscopic stent therapy is considered as first-line therapy for benign biliary strictures (BBS). 1 Repeated endoscopic balloon dilation followed by insertion of multiple plastic stents (PS) has been shown to be highly effective for treatment of those strictures.2–4 However, because of the limited patency of PS, this strategy usually requires four or five endoscopic interventions over a time period of approximately one year. Moreover, progressive bile duct dilation by inserting an increasing number of stents is technically demanding, as up to six stents side by side may be required. 2 Those drawbacks might not only put patients at risk of complications but also might have a negative impact on adherence to therapy.
Multiple studies have shown superior patency of self-expandable metal stents (SEMS) compared to PS in patients with unresectable malignant bile duct strictures. 5 With the introduction of fully covered SEMS (fcSEMS), which exhibit good removability, those devices have increasingly been used for treatment of benign biliary conditions such as strictures, leaks and bleeding. In the 2012 Guideline from the European Society of Gastrointestinal Endoscopy (ESGE), temporary insertion of fcSEMS was considered as an “investigational option” for treatment of BBS. 6 In recent years, there was increasing evidence supporting safe removability and treatment efficacy of fcSEMS in this indication.1,4,7–9 However, the covering of the SEMS prevents tissue ingrowth and therefore not only facilitates removability but also increases risk of device migration. Migration rates vary greatly depending on stricture etiology and type of SEMS but have been roughly reported to reach about 30%–40%.1,10–13 Moreover, the optimal time point of stent retrieval remains unclear. Too short stent indwell times may result in insufficient long-term stricture resolution, whereas the risk of stent-associated complications is likely to increase over time. Moreover, longer indwell time may impair removability. Different treatment protocols have been used in available studies, mainly reporting about indwell times of three to six months.8–10,12–21
We report the results of an international prospective, multicenter study designed to assess safety and efficacy of temporary placement of a fully expanded polytetraflouroethylene, fluorinated ethylene propylene (ePTFE/FEP)-covered SEMS for treatment of BBS using a protocol with a prolonged stent indwell period.
Methods
Design
We conducted a multicenter, prospective, single-arm study (NCT01343160) in three European countries at six tertiary referral centers. The study was approved by the independent ethics committee at each participating center. The objective of the study was to evaluate the safety and efficacy of temporary fcSEMS treatment in patients with BBS. Endpoints of the study were:
technical success: safe and successful endoscopic or percutaneous deployment of the fcSEMS; safe and effective stent removal; clinical success: successful treatment of the stricture upon implant; stent patency; viability of treatment (defined as stricture resolution or improvement with no requirement of re-stenting after fcSEMS removal); and stricture patency/recurrence on follow-up visits.
Patients were also assessed for stent-associated complications.
Patients
After obtaining written informed consent, patients ≥ 18 years with BBS and necessity of stent treatment were included in the study. Treatment-naïve as well as pretreated patients were eligible. Patients with stricture anatomy ruling out fcSEMS use (defined as distance of stricture < 1 cm from the hilus), patients with a previous bare metal stent treatment, patients with malignant biliary disease, pregnant patients and patients participating in other clinical studies were not considered to be eligible for study participation.
Endoscopic/percutaneous procedures and stents
For stent treatment, we exclusively used the GORE® VIABIL® Biliary Endoprosthesis (W.L. Gore & Associates, Flagstaff, AZ, USA). It is a self-expanding nitinol stent-graft that is fully covered end to end with a nonporous, nonbiodegradable membrane made of expanded ePTFE and FEP. The device is equipped with multiple ePTFE/FEP-covered wire loops (“fins”) that serve as anchors to prevent device migration (see Figure 1). Radiopaque gold rings are incorporated into each end of the prosthesis to facilitate stent visibility during interventions. The stent is available in different lengths (4, 6, 8, 10 cm) and diameters (8 and 10 mm). The device is also available with proximal transmural drainage holes. Those stents were not used in this study because they are not intended to be removed since ductile tissue ingrowth through the side holes over time may impede device removal.
FCSEMS used in this study (GORE® VIABIL® Biliary Endoprosthesis, W.L. Gore & Assoc., Flagstaff, AZ, USA). The Nitinol SEMS is fully covered end-to-end with an ePTFE/FEP membrane. The stent is equipped with fully covered fins (arrows) to prevent migration.
The fcSEMS were typically placed either endoscopically or percutaneously over a 0.035-inch guidewire under fluoroscopic visualization. For endoscopic placement, an 8.5 F delivery catheter was used. For percutaneous delivery, the stent is available with a 40 cm delivery catheter and a 10 F outer sheath. All stents were placed either intraductally or transpapillary. Pre- and post-deployment balloon dilation was not mandatory but rather left to the decision of the physician. Stent removal was scheduled nine months after stent placement. Endoscopic stent removal was achieved by grasping the distal end with rat-tooth forceps or with a polypectomy snare. Extraction was performed by gentle retraction of the endoscope under fluoroscopic vision. Percutaneous retrieval was performed using a 12 F Introducer by grasping the proximal end of the stent with a forceps and pulling the stent through the introducer sheath.
A cholangiogram after device removal was performed and assessed for stricture resolution. Insufficient stricture resolution was defined as improvement in degree of stricture of <50% in combination with insufficient/delayed clearance of contrast medium from the common bile duct. In case of insufficient stricture resolution, decision for further therapy was left to the treating physician.
Assessment and follow-up
Prior to treatment, all patients were assessed for clinical symptoms of obstruction and medical and surgical history was obtained. Lab tests prior to treatment included aspartate aminotransferase (AST), alanine aminotransferase (ALT), bilirubin, C-reactive protein (CRP), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT) and white blood count. Furthermore, patients received abdominal ultrasound and/or computed tomography (CT) scan prior to treatment. Before stent implantation, a cholangiogram was performed to assess location, length and degree of stricture. Telephone interviews assessing symptoms of biliary obstruction using a standardized protocol were conducted at one and six months post-deployment. Additionally lab tests as described above were completed one month post-deployment. At nine months post-fcSEMS removal, imaging (abdominal ultrasound or CT), lab tests and clinical symptoms were completed. Cholangiogram was performed prior to and after device removal. Lab tests were again completed one month post-removal. Telephone interviews to assess recurrent biliary obstruction were completed at one, three and 15 months post-removal.
In case of recurrent biliary obstruction during stent treatment or after stent removal, endoscopic retrograde cholangiopancreatography (ERCP) was again performed to assess for stent obstruction, stent migration or stricture recurrence (see Figure 2). Decision for further therapy was left to the treating physician.
FCSEMS treatment in a patient with chronic pancreatitis. (a) ERCP showing benign stricture of the distal common bile duct (and proximal stricture of the pancreatic duct). (b) A FCSEMS (GORE® VIABIL® Biliary Endoprosthesis, W.L. Gore & Assoc, Flagstaff, AZ, USA) was placed. The stricture resolved completely after FCSEMS insertion. Patient also received placement of a 7F plastic stent in the pancreatic duct. (c) Endoscopic view after FCSEMS placement. (d) Endoscopic FCSEMS removal after 9 months using a polypectomy snare. (e) Cholangiogram after FCSEMS removal showing only minimal residual stricture in the common bile duct, not requiring further therapy.
Data analysis
The study data was centrally managed by the coordinating center in Ludwigsburg. The statistical analysis was performed using SAS 9.3 by the coordinating center in assistance with W.L. Gore and Associates. The descriptive analysis consisted of mean, median, standard deviation for the numeric variables and percentages for the qualitative variables. Patency analysis was performed using the Kaplan Meier method. For analysis of the patient characteristics and stent placement data, all patients who were enrolled and treated with the fcSEMS were included. After dropout of four patients, all further endpoints were analyzed in a per-protocol fashion.
Results
Patients
Between January 2011 and April 2013, 47 patients were screened for eligibility. Forty-three patients were enrolled in the study (see flowchart in Figure 3). Stricture etiology in those patients was: chronic pancreatitis (CP) (24), anastomotic stricture at hepaticojejunostomy after pancreatic head resection (seven), surgical bile duct injury after cholecystectomy (four), recurrent common bile duct stones (four), post-sphincterotomy stenosis (one), autoimmune cholangitis (one) and cryptogenic (two) (see Table 1). Thirty-three patients (76.7%) had received prior treatment with balloon dilation and/or PS insertion whereas 10 (23.8%) patients were treatment naïve (see Table 1).
Flow chart showing patient recruitment and further treatment. Numbers indicate absolute number of patients. Patient characteristics GGT: gamma-glutamyl transferase; CBD: common bile duct.
Stent placement
In 33 patients (76.7%) the fcSEMS were placed endoscopically, whereas 10 patients (23.3%) received percutaneous stent insertion. The reason for percutaneous stent insertion was altered post-surgical anatomy in six cases. The other four patients had normal biliary anatomy but endoscopic cannulation of the bile duct had failed. Of the 10 patients, four had a history of prior treatment with balloon dilation.
Stent placement data
Stent indwelling and removal
Percentages were calculated by per-protocol analysis (PP-A).
Of the 10 patients who primarily underwent percutaneous stent, one died, one was excluded from the study because of malignant disease and in one case endoscopic removal was not possible because of duodenal stenosis (patient underwent Whipple). Of the remaining seven patients, four underwent percutaneous and three endoscopic removal.
Defined as stricture resolution without necessity of further treatment.
Stent indwelling, patency, stent-related complications
Median duration of stent therapy was 272 days (standard deviation 58.7; range: 7–336). Estimated primary patency at nine months was 73.0% (see Figure 4 for Kaplan Meier plot). Device occlusion due to intraluminal sludge formation occurred in four patients. One patient developed recurrent cholangitis without evidence of device occlusion. Another patient developed biliary obstruction due to bile duct stones proximal to the stent. One case of acute pancreatitis with dilation of the pancreatic duct was observed (see Table 4).
Stent patency. Kaplan Meier plot showing percentage of patients with patent stents. Censored patients are indicated with vertical lines, 95% Confidence intervals are indicated in light blue. Follow up Percentages calculated by per-protocol analysis (PP-A)
Stent removal
Two patients died before scheduled stent removal. Causes of death were non-device-related cardiovascular events. One patient was excluded from the study after stent placement because the biliary stricture was revealed to be malignant. Another patient refused further participation in the study. Hence, 39 patients remained with indication for stent removal. In four patients, the SEMS was removed percutaneously, and in the others it was removed endoscopically. In 29/39 patients (74.4%) the endoprosthesis was removed per protocol after completed therapy. In 10/39 patients (25.6%), the device had to be removed prematurely because of the following reasons: device occlusion (five), acute cholecystitis due to overstenting of the cystic duct (one) recurrent cholangitis without evidence of occlusion (one), bile duct stones proximal to the SEMS (n = 1), acute pancreatitis (one), and necessity of surgery due to duodenal stenosis in a patient with CP (see flowchart in Figure 3 and Table 3).
In one patient with symptoms of recurrent biliary obstruction/device occlusion, the fcSEMS could not be removed because the patient had developed severe duodenal stenosis due to CP and the papilla could not be reached endoscopically. This patient underwent surgical pancreatic head resection. All other attempted fcSEMS removals were successful. Upon stent retrieval, one minor papillary bleeding occurred, which stopped spontaneously. Another patient developed abdominal pain and CRP elevation due to local hematoma and infection of the chest wall after percutaneous stent retrieval. There was no evidence of perforation, abscess or local bleeding. The patient was treated with analgesics and with intravenous (i.v.) antibiotics for one week and the symptoms resolved completely.
Stent migration
During stent indwelling, one distal stent migration was observed one month after placement; the stent was extracted and a new fcSEMS was inserted (this patient was not counted as a treatment success in the further per protocol analysis (PP-A)). Upon cholangiogram prior to scheduled device removal after nine months, one stent in another patient with anastomotic stricture at a heapticojejunostomy was migrated distally. The stent was retrieved percutaneously. The stricture had resolved completely and no further therapy was necessary. Hence, total migration rate was 5.2% (2/38 patients, PP-A) (see Table 3).
Stricture resolution
Stricture resolution upon stent retrieval could be assessed in 38 out of 43 patients (see flowchart). Viability of treatment, defined as stricture resolution not requiring further therapy, was observed in 30 out of 38 patients (78.9%). When calculated by intention-to-treat analysis (including patients who were excluded or died after SEMS insertion), viability of treatment was 69.7%. Eight patients required further endoscopic therapy due to insufficient or no improvement of the stricture. Of those, four patients (50%) had undergone nine months of fcSEMS indwell per protocol; the other four had undergone premature stent removal because of stent-related complications.
Further therapy of the eight patients with insufficient stricture resolution included: re-placement of an fcSEMS (three patients), placement of (multiple) PS (four patients) and balloon dilation without stent insertion (one patient) (see also flowchart in Figure 3).
In the subgroup of patients with CP, stricture resolution could be assessed in 21 out of 24 patients. In this subgroup, stricture resolution without necessity of further treatment was observed in 16/21 patients (76.2%). In the patients with strictures of other etiology, stricture resolution was observed in 14 out of 17 patients (82.3%).
Stricture recurrence/follow-up
Of the 30 patients with stricture resolution after fcSEMS removal, complete follow-up data could be obtained in 25 patients. Two patients died before completion of follow-up five and seven months post-stent removal. Causes of death were not related to biliary obstruction. Another three patients were lost to follow-up after fcSEMS removal (see flowchart). During the follow-up period, stricture recurrence was observed in two of 25 patients (8.0%) (see Table 4). Both patients underwent further endoscopic therapy. Etiology of stricture in these two patients was one each CP and recurrent bile duct stones.
Discussion
Recent studies have shown that temporary placement of an fcSEMS is feasible and effective for BBS.7–10,22 However, the optimal duration of stent indwell is still unclear. Most available studies have investigated fcSEMS treatment with relatively short indwell times, often ranging from three to six months.8–10,12–21 In our study, we chose an extended indwell time of nine months. The rationale for this was that studies on PS treatment of BBS have shown that a stent treatment time of about one year is usually necessary to achieve durable stricture resolution.2,3 On the other hand, clinical experience as well as several studies have shown that risk of stent occlusion increases after six months. In our study, stent patency indeed decreased after six months but was still acceptable at nine months before scheduled retrieval. Moreover, patients requiring premature stent removal showed a high percentage of insufficient stricture resolution (50%). Extended indwell time is also supported by a recent prospective, multinational study by Devière and colleagues that even reported on nine to 11 months of stent treatment and showed favorable results especially of the subgroup of patients with CP. 7 Despite extended indwell, we observed good device removability. The papilla could not be reached in one patient with CP who had developed severe duodenal stenosis and required a Whipple operation. In all other patients, the fcSEMS could be removed successfully. These results show that the complete end-to-end ePTFE/FEP covering effectively prevents tissue ingrowth, ensuring safe and effective device removal.
The vast majority of other studies on treatment of BBS exclusively report on endoscopic fcSECMS implant and removal. In contrast, 23% of patients in our cohort underwent percutaneous insertion because of altered postoperative anatomy or failed endoscopic biliary cannulation. In four of these patients, the devices were successfully removed percutaneously using a rat tooth forceps. The nitinol exoskeleton and the thin material of the cover are easily recompressible to facilitate removal through a 12 F introducer sheath. Although all extractions were successful, there was one case of chest wall hematoma and local infection after extraction. However, this event was most likely not due to biliary trauma caused by the fcSEMS extraction itself because the device had been safely pulled through the 12 F introducer. More likely, the hematoma resulted from the percutaneous puncture of the non-dilated intrahepatic bile ducts (which was difficult and required multiple attempts in this case) followed by introduction of the large-diameter sheath. Although technically feasible, percutaneous SEMS removal is generally more difficult than endoscopic removal. For this reason, we attempted endoscopic removal prior to the percutaneous approach whenever possible.
Prevention of tissue ingrowth by the covering membrane of the SEMS results in good removability properties but may result in an increased risk of device migration. This puts the patient at risk for recurrent biliary obstruction and usually leads to additional endoscopic interventions, which themselves can lead to further complications and will cause additional costs. Reported migration rates vary greatly in published studies depending on device as well as stricture etiology. An early multicenter study by Tarantino et al.12,18 reported stent migration in 24.2% of patients. Modifications in stent design such as flared ends, anchoring fins, convex margins or anchoring with double pigtail plastic stent have been proposed to reduce risk of migration.1,7–9,13,19,23,24 In a retrospective multicenter study by Kahaleh and colleagues, an fcSEMS with flared ends was used to treat patients with mixed stricture etiologies. Stent migration occurred in 10.5% of patients. 9 Similar results were reported in a United States (US) retrospective multicenter study by Saxena and colleagues. 8 Walter and colleagues investigated a nitinol stent with irregular cell sizes, flared ends and a mixed silicone/PTFE cover. This design could not effectively prevent migration, which was observed in 31% of patients. 13 The recent multicenter study by Devière and colleagues using the WallFlex RX Biliary Stent (Boston Scientific Corp, Natick, MA, USA) reported device migration rates below 5% at six months, but the migration increased over time to 17%–18% at 12 months in the subgroup of CP patients. In the subgroup of liver transplant and post-cholecystectomy patients, stent migration was particularly high, with migration rates of 74.7% and 66.7%, respectively. 7 It is noteworthy that in this study only device migrations that required reintervention and re-stenting have been reported. In our study, we observed only two device migrations, one of them requiring reintervention. Total migration rate was 5.2%. One reason for this very low stent migration rate compared to the multicenter trial by Devière and colleagues and other trials may be that we did not include post-liver transplant strictures. As mentioned, risk of device migration is particularly high in this subgroup compared to post-surgical or CP-related strictures. However, the stent design may be the main factor for the particularly low migration rate observed in our cohort. First, the fcSEMS used is equipped with multiple elevated wire loops that serve as anchors and are designed to prevent migration. Park and colleagues compared fcSEMS with anchoring flaps versus fcSEMS with flared ends and found that the stents with anchoring flaps showed significantly lower migration rates compared to the stents with the flared ends (0% versus 33%, respectively), indicating the efficacy of this antimigration stent design feature. 19 An early US monocentric pilot study reported a very low incidence of stent migration in a cohort of 44 patients using the same fcSEMS as in our study. 14 The anchoring “fins” of the FCSEMS used in our study seem not to impair removability because they are fully ePTFE/FEP covered and are designed to easily fold backward into the device upon application of longitudinal removal force. The study by Mahajan et al. observed ulceration and slight bleeding upon cholangioscopy after device removal. 14 This is the only report of “suspected fin ulcerations”; however, the authors did not report any relevant complications or sequelae with this observation. Although we did not perform post-removal cholangioscopy, we did not observe severe removal complications related to the anchoring fins in our cohort. The chest wall hematoma that occurred in one patient was probably not attributed to the fins but rather to the difficult percutaneous puncture followed by insertion of a 12 F introducer sheath. An experimental study by Isayama and colleagues demonstrated that the fcSEMS used in our study provides very low axial force to prevent kinking and other adverse effects, yet has a medium level of radial force that is strong enough to keep the stricture open but not too high as to cause damage to the biliary duct wall. 25
Viability of stent treatment, defined as stricture resolution upon device removal without necessity of further treatment, was observed in 78.9% of patients in our study. In the subgroup of patients with CP, which represented the majority of patients in our cohort, stricture resolution was 76.2% compared to 82.3% in the patients with strictures of other etiology. Although reported success rates for CP-related strictures have been low in older studies,26,27 more recent trials report more favorable results that are comparable to those of our study. In the multicenter trial by Deviere et al., stricture resolution in CP patients and post-cholecystectomy patients was 79.7% and 72.2%, respectively. 7 Walter and colleagues reported a total immediate success rate of 80%; success rates did not differ significantly between CP patients and patients with strictures of other etiology. 13 Two other recent multicenter studies, which included a patient cohort similar to ours, report comparable short-term success rates.8,9
The patients in our trial were followed for a total of two years (15 months post-stent removal). Of the 25 patients with stricture resolution who completed the follow-up, we observed two stricture recurrences, accounting for a total recurrence rate of 8%. Stricture recurrence in these two patients was suspected based on clinical symptoms of biliary re-obstruction and was then confirmed via ERCP. There might be a discrepancy between clinical symptoms and imaging or laboratory criteria of biliary obstruction. However, our follow-up protocol does reflect daily clinical practice, as patients with BBS in the long term are usually followed clinically based on symptoms such as occurrence of jaundice or cholangitis. Reported long-term success rates differ significantly in available studies. Walter and colleagues reported an initial success rate of 80% whereas 21% of patients developed stricture recurrence (long-term success rate 63%), respectively. 13 The reason for the difference in long-term stricture patency compared to our results may be the much shorter stent indwell time, which was only three months. Poley and colleagues reported a total success rate of 61% within 15 months of follow-up; duration of stent treatment ranged from two to eight months. 27 The study by Saxena and colleagues report very few (five out of 123 patients) stricture recurrences when using a stent indwell time of six months; 8 Wagh and colleagues report similar results. 10 In the study by Devière and colleagues, who used even longer indwell periods, favorable recurrence rates (14.8%) were reported. 7 Hence, stent indwell times of at least six months may be favorable regarding long-term treatment success at least in patients with BBS not related to liver transplantation.
The concept of insertion of an increasing number of PS was shown to be highly effective for treatment of BBS and is currently the only endoscopic treatment modality for which long-term data (>2 years) are available. Costamagna and colleagues reported low stricture recurrence rates in a 10-year follow-up study in patients with different stricture etiologies, suggesting that the initial treatment effect is durable. 3 However, multiple endoscopic interventions are necessary to achieve stricture resolution. There is a recent randomized controlled study investigating treatment with multiple PS versus fcSEMS in patients with CP. Long-term success rate was impressively high, reaching 90% and 92% in the PS versus the fcSEMS group, respectively. 28 Patients were treated for six months but underwent scheduled re-ERCP after three months even in the fcSEMS group, which may not be necessary (and may even obviate the advantage of fcSEMS treatment versus PS) given the favorable three-month patency for fcSEMS reported in most studies. Taken together these data suggest that the previous preference for PS for benign and (covered) metal stents for malignant lesions29,30 no longer holds and fully covered metal stents are increasingly and effectively making inroads into the realm of benign pancreaticobiliary strictures.
The present study does have several limitations. Although conducted prospectively in an international multicenter setting, it is an observational study without a control group. Hence, direct comparison to plastic stenting or surgery was not possible. The number of patients was limited and even decreased in the course of the study because five patients died and three patients were lost to follow-up. However, the results regarding treatment success were comparable to other studies with larger number of patients. Another limitation is that stricture resolution was assessed by the endoscopist/interventional radiologist and not by central independent reviewers, so overestimation of stricture resolution cannot be excluded. Further endoscopic therapy in case of insufficient stricture resolution was also not standardized but rather left to the individual decision of the endoscopist/interventional radiologist. However, the low rate of stricture recurrence points to correct assessment of initial treatment success.
In conclusion, the present study shows that endoscopic as well as percutaneous fcSEMS treatment with prolonged indwell time is feasible and effective for BBS. Removability of the ePTFE/FEP-covered SEMS used in this study was excellent and migration rate was particularly low.
Footnotes
Declaration of conflicting interests
Prof. Caca has received lecture fees from W.L. Gore & Associates. The other authors have nothing to declare.
Funding
This work was supported by W.L. Gore & Associates, which provided financial support to the coordinating center (Ludwigsburg) and all participating centers.