Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) with placement of biliary stents has been used for more than 2 decades [1, 2]. Because of their lower cost, wide distribution, and ease of placement, plastic stents have been conventionally used for benign biliary diseases and malignant biliary obstruction with short-term survival [2]. Those patients with malignant distal biliary obstruction and survival time beyond 6 months have been more frequently treated with self-expandable metal stents (SEMS), given their greater complication-free survival and cost-effectiveness [3–6]. In an attempt to improve patency, however, partially covered metal stents (PCMS) were developed to resist tissue in-growth through the stent interstices [7, 8], a well-demonstrated complication of uncovered metal stents [9]. Indeed, the demonstration of PCMS removability [10, 11] has prompted expanded indications for their use to include placement for distal pancreatico-biliary malignancy regardless of resectability [12, 13].

With the increasing use of PCMS, though, several complications have been reported, including migration [12, 14–16], benign strictures [7, 17] and cholecystitis [18, 19]. Given the limited prospective data regarding PCMS and their wider use in malignant and benign biliary diseases, we sought to evaluate the complications associated with their use as well as their respective management.

Patients and Methods

Patients

A total of 423 patients underwent insertion of metal stents in our institution over a 5-year period and were followed prospectively; 27 patients received uncovered SEMS and 396 received PCMS. All patients in the uncovered SEMS group had malignant hilar lesions. In the PCMS group, 247 patients presented with malignant disease and 149 patients with benign disease. After stent insertion, patients were followed in clinic with liver function tests (LFTs) until stent removal (in cases of benign diseases receiving PCMS), stent dysfunction, or death. The study was approved by our institutional review board; all patients provided written consent for their procedures.

Methods

In the uncovered SEMS group, uncovered Wallstents (40, 60, or 80 mm length) of 10-mm diameter (Boston Scientific, Natick, MA) were placed. In the PCMS group, Wallstents (40, 60, or 80 mm length) of 10-mm diameter, partially covered with Permalume® (Boston Scientific, Natick, MA), were placed. Following biliary sphincterotomy, the origin of the cystic duct insertion was noted in patients having a gallbladder. The PCMS delivery system was advanced proximal to the lesion over a guidewire where the PCMS was positioned and deployed. When anatomically feasible, PCMS were deployed to avoid occlusion of the cystic duct insertion when the gallbladder was present. Removal of PCMS was effected using the snare or rat tooth technique as previously published [10]. Proximal migration was managed with a combination of CRE balloon dilation (Boston Scientific) and rat tooth extraction (Figs. 1, 2, 3, 4, and 5).

Fig. 1
figure 1

Fluoroscopy showing CRE balloon dilation of a proximally migrated PCMS

Full size image
Fig. 2
figure 2

Fluoroscopy showing the PCMS grasped from the inside using a rat tooth forceps

Full size image
Fig. 3
figure 3

Fluoroscopy showing successful extraction of the PCMS from the bile duct

Full size image
Fig. 4
figure 4

Fluoroscopy showing removal of the PCMS through the working channel of the endoscope

Full size image
Fig. 5
figure 5

Cholangiogram post PCMS removal showing enlargement of the bile duct post PCMS removal

Full size image

Definition of Events

Successful PCMS placement was defined as deployment of the PCMS across the lesion effecting biliary decompression and relief of symptoms. Stent patency was defined by the period of time between stent insertion and stent dysfunction. Stent occlusion, in contrast to tumor overgrowth or stricture in the uncovered portions of the PCMS (hyperplasia), was defined as stent dysfunction secondary to biliary debris or sludge. Complications were stratified as early (occurring ≤30 days of PCMS placement) or late (occurring >30 days following PCMS placement). Cholangitis was defined as fever persisting for 24 h in the setting of biochemical cholestasis without another discernible cause. Cholecystitis was defined by appropriate symptoms in the setting of supportive imaging (ultrasound, computed tomography). We also distinguish duodenal from proximal migration, hereto uncommonly cited in the literature.

Repeat biliary intervention was defined as the need for any type of endoscopic, percutaneous, or surgical procedure to improve biliary drainage after insertion of the initial PCMS.

Results

A total of 70 patients (41 males, 29 females) with a mean age of 59 developed complications (18%) in the PCMS group. Complications were observed in 48 patients with underlying malignancy and 22 patients with benign biliary diseases (Table 1). The global mean time to complication was 159 days with a range between 1 and 1,027 days.

Table 1 Characteristics of patients at the time of PCMS insertion
Full size table

Early Complications (≤30 Days) of PCMS and Management

Fourteen patients (3.5%) had early complications. Five patients had benign disease, and nine patients had malignant indications (Table 2). Complications included duodenal migration (n = 3), proximal migration (n = 2), cholecystitis (n = 4), pancreatitis (n = 2), stent occlusion (n = 1), tumor overgrowth (n = 1), and abdominal pain (n = 1).

Table 2 Early complications of PCMS insertion
Full size table

All the patients who suffered from early cholecystitis had underlying malignant disease. Two patients were noted to have had the PCMS covering the cystic duct insertion, one had tumor involvement of the cystic duct, and only one had PCMS below the cystic duct. Three patients were treated with percutaneous cholecystostomies, while one patient had a gallbladder stent placed endoscopically.

Twelve patients underwent PCMS revision. Two patients who experienced stent migration were managed by removal and replacement with plastic stents, while the other three underwent replacement with a PCMS. The patient with early stent occlusion was found to have a large amount of debris within the PCMS and was treated with removal and replacement with new PCMS. The one patient with early tumor overgrowth was also managed by replacement with PCMS. The two patients with pancreatitis and the one experiencing abdominal pain as a result of PCMS placement had their PCMS removed.

Late Complications (>30 Days) of PCMS and Management

Table 3 delineates the 56 patients with long-term complications, of which 19 had benign indications and 37 had malignant ones. Migration was observed in 31 patients (7.8%). Duodenal migration occurred in 24 cases (6.1%), whereas proximal migration occurred in 7 cases (1.7%). The management of migration included replacement with PCMS (n = 14), replacement with plastic stents (n = 5), removal (n = 7), stent within a stent (n = 3), or no intervention (n = 2) secondary to spontaneous migration with resolution.

Table 3 Long-term complications of PCMS insertion
Full size table

The second most common long-term complication of PCMS placement was cholecystitis (n = 9). In six patients, the PCMS covered the cystic duct insertion. There was cancer involvement of the cystic duct in one patient, while two patients who had PCMS placed below the cystic duct developed cholecystitis. Cholecystitis was managed by cholecystectomy in six cases, percutaneous drainage in two cases, and placement of a gallbladder stent in one case.

Three patients developed late cholangitis. They had duodenal cancer, ampullary adenoma, and benign stricture and developed cholangitis at 873, 874, and 1,027 days, respectively. All three patients had their PCMS removed, while the patient with duodenal cancer underwent PCMS replacement.

Tumor overgrowth (n = 3) was managed by replacement with plastic stents in two cases and replacement with PCMS in the other. We also report PCMS stricture in the proximal uncovered portion of the stent in three patients (Fig. 6), all of which were in benign disease; two were managed with replacement with PCMS, while one patient received plastic stents.

Fig. 6
figure 6

Mucosal hyperplasia at the hilum induced by the uncovered portion of a previously placed PCMS

Full size image

Pancreatitis occurring after PCMS placement was managed with PCMS removal. One patient developed a liver abscess 364 days after PCMS placement and was treated with removal of PCMS and replacement with plastic stents. Another patient developed an infected biloma, 106 days after PCMS placement, which was related to a bile leak after cholecystectomy. This patient had removal and replacement with PCMS. Lastly, one patient developed abdominal pain 45 days after PCMS placement. This was managed with PCMS removal only, since the benign biliary stricture had resolved.

Discussion

Expandable biliary metal stents have been developed for the treatment of malignant biliary strictures [3–6]. In the past decade, partially covered self expandable metal stents (PCMS) were developed to prevent tumor growth through the mesh of the stent [20]. Experience with PCMS has been reported by several groups [11, 17, 21–25].

PCMSs provide better biliary decompression than plastic stents because of their larger diameter. In addition, PCMSs have a metal composite covered by a synthetic material, such as silicone, that resists the effects of bile, gastric, or pancreatic secretions [26]. The covering is designed to prevent tumor ingrowth [7, 8], while the 0.5-mm uncovered portions at each end of the stent are meant to prevent migration. Our group has a preference for PCMS for distal biliary lesions, leaving the use of uncovered SEMS for malignant hilar lesions of the biliary tree deemed surgically unresectable. This explains the small number of uncovered SEMS placed during the study period.

In the PCMS group, we encountered 36/396 cases (9.1%) of migration in our study. Duodenal migration accounted for 27 of these cases (6.8%). This is in the range of what has been reported previously [15, 23, 27, 28]. We also report, however, nine cases (2.3%) of proximal migration of the PCMS. This is likely related to the foreshortening of the PCMS with mucosal overgrowth through the uncovered portion of the SEMS, as previously investigated by our group [17]. Although this is a less common mode of migration and less frequently reported by other investigators, its endoscopic management is more challenging and requires the combination of balloon dilation and rat tooth extraction (Figs. 1, 2, 3, 4, and 5). The mean time to revision was 164 and 90 days for duodenal and proximal migrations, respectively. All migration cases were managed with endoscopic removal of the PCMS except in three cases with malignancy undergoing placement of a PCMS within the previously placed PCMS. Inability to remove the PCMS is unusual in expert hands but can be encountered when the PCMS has been in place for more than 6 months [17] or when access to the malfunctioning PCMS is precluded by malignancy, as experienced in the three above patients.

Occlusion rates of PCMS have been reported between 5% and 14% at 6 to 12 months in distal malignant biliary obstruction [24–29]. In our study population, we found that occlusion occurred in only 3/396 (0.8%) cases throughout our follow-up period of 70 months. In malignancy, we describe occlusion as a separate event from tumor overgrowth: that is, stent clogging may occur separately from tumor overgrowth through the stent interstices of the uncovered portions. It has been hypothesized that tumor burden may promote the formation of biliary sludge and debris, thereby leading to occlusion via the accumulation of these particles rather than tumor spread [26]. In cases of benign biliary disease, in contrast, adherence of bacteria to the coating membrane has been thought to result in stent clogging [8, 30, 31]. Interestingly, all our patients with stent occlusion resulting from debris had a malignant indication.

We did not find tumor in-growth or tumor extension through the stent to be a complication. Tumor in-growth through PCMS is regarded as a failure in the covering membrane and has been described elsewhere [32]. Overall, the reduced rates of tumor overgrowth in our study compared to uncovered stents for the treatment of distal malignant biliary obstruction is consistent with other established studies [16].

Stricture in the uncovered portion of the PCMS in benign disease occurred in three cases in our study. This is less frequent than the study published by Cantu et al. that cited a higher, late complication rate [25]. That group, however, included only patients with common bile duct stricture secondary to chronic pancreatitis, known to be a more difficult group to treat, and only removed PCMS once malfunction was observed.

Cholecystitis was the second leading complication in our study. We report an overall rate of 13/396 (3.3%). Prior studies have suggested that PCMS may be more likely to cause cholecystitis than uncovered SEMS, specifically when the covering material overlaps the orifice of the cystic duct. This is felt to be due to the decreased permeability of PCMS and obstruction to bile flow, when directly opposed to the cystic duct [17, 26, 33]. The reported rates of cholecystitis after PCMS placement are between 2.9% and 12% [26, 29, 34]. We hypothesize that our relatively low rate of cholecystitis may be related to our systematic efforts to avoid covering the cystic duct insertion in benign diseases and to our attempt to irrigate the gallbladder during ERCP when it was distended before PCMS placement in malignant diseases [12, 17]. Isayama et al. reported that cholecystitis occurred mainly in patients with tumor involving the orifice of the cystic duct in the setting of unresectable distal biliary malignancy [19]. Occlusion of the cystic duct by tumor was also found to be the major risk factor for cholecystitis by Suk et al. [18]. Their study found that 9/15, or 67%, of patients with cholecystitis had cystic duct involvement by tumor. Another study used intraductal ultrasonography to diagnose cystic duct invasion by tumor and found this to be predictive of cholecystitis [23].

In our study, cholecystitis was found in the setting of tumor involvement in only 2/13 (15%) cases, despite the finding that 11/13 (85%) patients who developed cholecystitis carried primary diagnoses of malignant biliary disease. This is quite different from rates cited by other studies [35]. Indeed, in 8/13 (62%) cases of cholecystitis we report, the cystic duct was covered by our PCMS, arguing for the positioning of the PCMS as a more important risk factor.

On the other hand, since 3/13 (23%) cases of cholecystitis (at 20, 47, and 55 days after insertion) occurred while the stent was placed below the cystic duct, other factors must play a role, such as bacterial colonization after biliary sphincterotomy, gallstone disease, or perhaps contrast injection, as previously suggested in the literature [35–38].

Four patients (1%) developed acute pancreatitis or worsening pancreatitis after PCMS placement. There were two patients with chronic pancreatitis, one with ampullary adenoma, and one with benign biliary stricture. One patient with chronic pancreatitis developed an infected pseudocyst 4 days after PCMS placement for biliary stricture. The infected pseudocyst occurred in the setting of the PCMS covering the PD orifice. This patient required removal of the PCMS and transpapillary drainage. In the patient with benign biliary stricture, ERCP revealed poor pancreatic duct drainage likely contributing to the complication of pancreatitis. The patient with ampullary adenoma underwent ampullary resection with biliary and pancreatic sphincterotomy and likely developed post-procedure pancreatitis.

We observed two cases of abdominal pain complicating PCMS placement, both placed in appropriate position without obstruction. Our study also revealed one case of liver abscess and one infected biloma after cholecystectomy in a patient with pancreatic cancer; both required drainage. We believe neither of these complications was directly related to the PCMS placement.

The present series of patients with PCMS complications and their management is the largest reported from a single center to our knowledge. The efficacy of PCMS must be weighed against their potential complications and their specific management. Further improvements in the structure and composition of PCMS are required to prevent these complications. Indeed, fully covered stents might be the best option to decrease tumor overgrowth, but also to prevent mucosal hyperplasia at the proximal portion of the SEMS. Fenestration of the covered stent at the level of the cystic duct insertion might prevent cholecystitis. With newer covered SEMS offering those options arriving on the market, further multicenter studies should be performed to determine if better efficacy and lower morbidity can be achieved in both benign and malignant biliary diseases.