A pilot study of Spring Stopper Stents: Novel partially covered self-expandable metallic stents with anti-migration properties for EUS-guided hepaticogastrostomy

ABSTRACT

Background and Objectives:

EUS-guided hepaticogastrostomy (EUS-HGS) is an effective salvage procedure when conventional endoscopic transpapillary biliary drainage is difficult or fails. However, the risk of stent migration into the abdominal cavity has not been resolved completely. In this study, we evaluated a newly developed partially covered self-expandable metallic stent (PC-SEMS) that has a spring-like anchoring function on the gastric side.

Methods:

This retrospective pilot study took place at four referral centers in Japan between October 2019 and November 2020. We enrolled 37 cases consecutively who underwent EUS-HGS for unresectable malignant biliary obstruction.

Results:

The rates of technical and clinical success were 97.3% and 89.2%, respectively. Technical failures included one case in which the stent was dislocated during the removal of the delivery system, requiring additional EUS-HGS on another branch. Early adverse events (AEs) were observed in four patients (10.8%): two with mild peritonitis (5.4%) and one each (2.7%) with fever and bleeding. No late AEs were observed during the mean follow-up period of 5.1 months. All recurrent biliary obstructions (RBOs) were stent occlusions (29.7%). The median cumulative time to RBO was 7.1 months (95% confidence interval, 4.3 to not available). Although stent migration in which the stopper was in contact with the gastric wall on follow-up computed tomography was observed in six patients (16.2%), no migration was observed.

Conclusions:

The newly developed PC-SEMS is feasible and safe for the EUS-HGS procedure. The spring-like anchoring function on the gastric side is an effective anchor preventing migration.

INTRODUCTION

EUS-guided biliary drainage (EUS-BD), including EUS-hepaticogastrostomy (EUS-HGS) and EUS-choledochoduodenostomy (EUS-CDS), has been increasingly reported as an alternative procedure in cases of failed or difficult endoscopic retrograde cholangiopancreatography for malignant biliary obstruction (MBO).[1-5] Although EUS-BD has high rates of technical and clinical success, the rate of adverse events (AEs), including bile leakage, stent migration, and hemorrhage, is high.[6-9] Among the AEs, intraperitoneal migration of a covered self-expandable metallic stent (C-SEMS) resulting in peritonitis can be fatal.[10-12] A long C-SEMS (10 or 12 cm in length) with sufficient intragastric length has been recommended to prevent migration.[13] However, intraperitoneal migration cannot be completely prevented with a long C-SEMS because the gastric wall returns to its original position after a stent placement.[14] Thus, a C-SEMS with anti-migration properties at both ends is needed to prevent migration. We developed a partially covered SEMS (PC-SEMS) with anchoring properties (Niti-S Spring Stopper Stent [SSS]; Taewoong Medical, Gimpo, Korea). We evaluated the SSS, which has a spring stopper at the gastric side and uncovered potion at the hepatic side, to prevent proximal and distal migration.

PATIENTS/MATERIALS AND METHODS

Study design

This study involved a multicenter retrospective chart review. The aim of this study was to evaluate the safety and effectiveness of a newly designed PC-SEMS with anti-migration properties during the EUS-HGS procedure. The following four referral and academic centers participated: Juntendo University Hospital, Cancer Institute Hospital, Saitama Medical Center of Saitama Medical University, and The University of Tokyo Hospital. This study was reviewed and approved by the institutional review board of each institution, and informed consent was obtained with an opt-out form on the website.

Patients

Patients who underwent EUS-HGS for unresectable MBO using SSS at the four referral and academic centers between October 2019 and November 2020 were consecutively enrolled retrospectively. The inclusion criteria were (1) unresectable MBO and (2) failed or difficult conventional endoscopic transpapillary drainage. The exclusion criteria were (1) the presence of massive ascites and (2) severe coagulopathy. Written informed consent was obtained from all patients for the EUS-HGS procedure.

Spring Stopper Stent

The basic structure is a cross-wire design with a nickel-titan alloy. The stent is covered with silicone, but 15–20 mm on the hepatic side is uncovered [Figure 1] to prevent distal migration out of the liver or obstruction of the biliary branch. The outside of the SSS has exposed wires. The exposed wires contribute to friction with the stomach wall and adherence to the bile duct wall, which may be useful in preventing migration. The length of the uncovered hepatic portion is increased compared to a long PC-SEMS (Niti-S; modified Giobor Stent, Taewoong Medical, Seoul, Korea) to prevent distal migration on the hepatic side.[13] The 1 cm from the end of the gastric side has a spring-like anchoring function (spring stopper) with a double-covered SEMS to prevent proximal migration into the abdominal cavity. The anchoring ability of the spring stopper is high compared to other SEMS, including the HANAROSTENT (MI Tech, Seoul, Korea), HILZO (BCM Gyeonggi-do, Korea), and Wallflex (Boston Scientific, Natick, MA, USA), used in our previous experimental study.[15] The stents used in this study were 8 mm in diameter and 8 cm, 10 cm, or 12 cm in total length. The stent delivery system was 8.5 French Gauge (Fr) with a tapered tip to allow smooth insertion.

Figure 1.

Details of a Spring Stopper Stent (Taewoong Medical, Gimpo, Korea). (a) The proximal end has a 15-mm or 20-mm uncovered portion to prevent the peripheral branch obstruction and distal migration. The 1 cm from end of gastric side provides a spring-like anchoring function (spring stopper) with a double-covered membrane for antimigration property to prevent proximal migration. The stents used in this study were 8mm in diameter, and 8 cm or 10 cm or 12 cm in total length. (b) When the spring stopper is pulled toward the liver, the stopper part deforms and the anchoring ability works

EUS-hepaticogastrostomy procedure

All procedures were performed with a conventional curved linear array echoendoscope (EG 580 UT, Fujifilm, Tokyo, Japan; UCT-260, Olympus Medical Systems, Tokyo, Japan). Briefly, a standard 19-gauge needle was used to puncture the left intrahepatic bile duct under EUS guidance in color Doppler mode to avoid the intervening blood vessels. A 22-gauge needle was used if the targeted bile duct was not sufficiently dilated. The B3 branch bile duct was preferred as the puncture site over the B2 branch bile duct because a B2 puncture can cause a transesophageal puncture, leading to mediastinitis. After the contrast medium was injected into the bile duct, a 0.025-inch guidewire (Visiglide2, Olympus Medical Systems; EndoSelector, Boston Scientific Japan, Tokyo, Japan) was advanced. The fistula was dilated initially with an ultra-tapered mechanical dilator (ES Dilator, Zeon Medical, Tokyo, Japan) and/or a balloon dilator (REN, Kaneka Medix, Tokyo, Japan). A cautery dilator (Cyst-Gastro Set; Endoflex, Niederrhein, Germany, Fine 0.025; Medicos Hirata, Osaka, Japan) was used when mechanical or balloon dilation failed. A 0.018-inch guidewire (Fielder, Asahi Intech., Tokyo, Japan) was used when a 22-gauge needle was used and exchanged for a 0.025-inch guide wire after the puncture tract was dilated. The SSS was deployed using an intra-scope channel stent-release technique from the left intrahepatic bile duct to the stomach, with the only uncovered portion located in the bile duct. In all cases, the intragastric length was approximately 5 cm or greater at the time of stent placement. A follow-up computed tomography (CT) scan was taken after EUS-HGS even if the patient was asymptomatic.[16]

Study outcomes

The primary outcome was the technical success rate. Secondary outcomes were the clinical success rate; stent migration rate; the procedure time; procedure-related AEs, such as abdominal pain, biloma, peritonitis, bleeding, cholangitis, and perforation; recurrent biliary obstruction (RBO); and time to RBO (TRBO).

Definitions

Technical success was defined as the successful deployment of an SSS in an appropriate position. Clinical success was defined as a reduction in serum total bilirubin of 50% or < 2 mg/dL within 2 weeks.[17] The procedure time was defined as the time between the puncturing of the bile duct and the placement of the stent. RBO was defined as a composite endpoint of either occlusion or migration, and TRBO refers to the time from an SSS placement to RBO according to the Tokyo criteria 2014.[17] AEs were divided into early (within 14 days) and late (later than 14 days) AEs and graded according to the severity grading system of the American Society for Gastrointestinal Endoscopy lexicon.[18] Abdominal pain was defined as pain without the fluid collection on CT. Biloma was defined as the presence of fluid surrounding the puncture site without abdominal pain. Peritonitis was defined as the presence of clinical symptoms of peritoneal inflammation and corresponding accumulation of fluid on CT. The stent lengths of the intrahepatic, intraperitoneal, and intragastric portions were measured on follow-up CT after EUS-HGS.

Statistical analysis

The data were analyzed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan).[19] The results were presented as frequencies (%) or medians with interquartile ranges (IQRs). We calculated the median cumulative survival time and TRBO with 95% confidence intervals (CIs) using Kaplan–Meier analyses.

RESULTS

Patients

Thirty-seven patients (22 males and 15 females) underwent EUS-HGS using the SSS between October 2019 and November 2020 and were included in the study. The patients’ baseline characteristics are summarized in Table 1. The most common primary cancer-causing MBO was pancreatic cancer in 54.1% of cases (n = 20). The major indications for EUS-HGS were difficulty accessing the papilla due to duodenal tumor invasion in 54.1% of cases (n = 20). Twenty-two cases (59.5%) underwent EUS-HGS as primary BD, and the remaining 15 cases (40.5%) underwent prior transpapillary BD and switched to EUS-HGS.

Table 1.

Patients’ baseline characteristics (n = 37)

	n (%)
Age (years), median (IQR)	70 (62–76)
Male	22 (59.5)
Cause of biliary obstruction
Pancreatic cancer	20 (54.1)
Biliary tract cancer	6 (16.2)
Gastric or duodenal cancer	4 (10.8)
Hepatocellular carcinoma	1 (2.7)
Metastatic lymph nodes	6 (16.2)
Indication for EUS-HGS	16 (14.0)
Duodenal tumor invasion	20 (54.1)
Difficult approach to the targeting bile duct	7 (18.9)
Refractory to transpapillary stenting	5 (13.5)
Altered anatomy	3 (8.1)
Unsuccessful biliary cannulation	1 (2.7)
History of ERCP related AEs	1 (2.7)
Biliary stricture
Distal	22 (59.5)
Hilar	15 (40.5)
Prior biliary drainage	14 (37.8)
Duodenal stent	10 (27.0)
Total bilirubin, median (IQR) (mg/dL)	4.9 (2.2–8.3)
Alkaline phosphatase, median (IQR) (U/L)	1181 (744–1914)
C-related protein, median (IQR) (mg/dL)	3.0 (1.6–5.4)
Cumulative survival time, median (95% CI) (months)	4.0 (2.0–6.1)

IQR: Interquartile range; CI: Confidential interval; EUS-HGS: EUS-guided hepaticogastrostomy; AEs: adverse events

Details of the procedure are summarized in Table 2. The most frequent puncture site was B3 (78.4%), and the median diameter of the punctured bile duct was 4.6 mm (IQR, 3.0–8.0). The most selected stent was 10 cm in length with a 15-mm uncovered portion in 45.9% of cases (n = 17). The median stent lengths of the intrahepatic, intra peritoneal, and intra gastric portions measured on CT were 47 mm (IQR, 38–53), 18 mm (IQR, 8–23), and 41 mm (IQR, 32–49), respectively.

Table 2.

Procedure details

	n (%)
Procedure time, median (IQR) (min)	18 (15–24)
Puncture site
B2	8 (21.6)
B3	29 (78.4)
Diameter of punctured bile duct, median (IQR) (mm)	4.6 (3.0–8.0)
Puncture needle and guidewire
19G needle and 0.025 inch guidewire	33 (89.2)
22G needle and 0.018 inch guidewire	4 (10.8)
Fistula dilation
Bougie	20 (54.1)
Balloon	27 (73.0)
Cautery	1 (2.7)
Total stent length (cm)/uncovered length (mm)
8/15	1 (2.7)
8/20	2 (5.4)
10/15	17 (45.9)
10/20	12 (32.4)
12/15	3 (8.1)
12/20	2 (5.4)
Intrahepatic stent length, median (IQR) (mm)	47 (38–53)
Intraperitoneal stent length, median (IQR) (mm)	18 (8–23)
Intragastric stent length, median (IQR) (mm)	41 (32–49)

IQR: Interquartile range

Technical and clinical success and adverse events

The clinical outcomes are summarized in Table 3. The rates of technical success and clinical success were 97.3% (n = 36) and 89.2% (n = 33), respectively. Technical failures included one case in which the stent was dislocated during the removal of the delivery system, requiring additional EUS-HGS on another branch. Of the clinical failures, two cases were poor drainage due to bleeding (one with spontaneous bleeding from the primary tumor apart from the SSS, and one with bleeding after cautery dilation requiring transcatheter arterial embolization) and two cases were liver failure due to progression of liver metastasis.

Table 3.

Details of clinical outcomes

	n (%)
Technical success	36 (97.3)
Clinical success	33 (89.2)
Early AEs	4 (10.8)
Fever	1 (2.7)
Peritonitis	2 (5.4)
Bleeding	1 (2.7)
Late AEs	0
RBO	11 (29.7)
Hyperplasia	7 (18.9)
Food impaction	3 (8.1)
Hemobilia	1 (2.7)
Migration	0
The spring stopper portion moved in contact with the gastric wall on CT the next day after HGS	6 (16.2)
Length of used stent, 8/10/12 cm	1/5/0
TRBO in cases with RBO, median (IQR) (months)	4.3 (2.7–6.2)
Cumulative TRBO, median (95% CI) (months)	7.1 (4.3–NA)

RBO: Recurrent biliary obstruction; TRBO: Time to recurrent biliary obstruction; IQR: Interquartile range; CI: Confidential interval; HGS: Hepaticogastrostomy; CT: Computed tomography; NA: Not available; AEs: Adverse events

Early AEs were observed in 10.8% of cases (n = 4), including fever in 2.7% (n = 1), peritonitis in 5.4% (n = 2), and bleeding in 2.7% (n = 1). There were no late AEs. The severity of the early AEs was mild in one case (fever case) and moderate in three cases. One case of peritonitis was caused by bile leak after tract dilation and required percutaneous drainage, and the other case of peritonitis required placing an additional SEMS because the uncovered portion of the SSS was located extrahepatically. One case of bleeding was caused by cautery dilation and required transcatheter arterial embolization. All AEs were procedure related and not stent related.

The spring stopper portion had come into contact with the gastric wall on follow-up CT in 16.2% of cases (n = 6), although the stent length in the stomach was sufficient at the time of SSS placement [Figure 2]. RBO was observed in 29.7% of cases (n = 11), all due to an occluded stent. As a result, no cases of migration were observed during the mean follow-up period of 5.1 months. Among the RBO cases, the most common causes of occlusion were mucosal hyperplasia at an uncovered portion in 63.6% of cases (7/11). The median cumulative TRBO was 7.1 months [95% CI, 4.3 to not available; Figure 3].

Figure 2.

The high-risk case of stent migration using a SSS. (a) Fluoroscopic image of EUS-HGS, SSS was deployed using intra-scope channel stent release technique. (b) Endoscopic images of SSS at the placement, Stent length in the stomach was enough long. (c) Plane CT taken on the next day after EUS-HGS, Spring stopper part moved in contact with the gastric wall. (d) Endoscopic images of SSS on the next day after EUS-HGS, Spring stopper part was closely attached to the stomach wall to prevent migration. EUS-HGS: EUS-guided hepaticogastrostomy; CT: Computed tomography; SSS: Spring Stopper Stent

Figure 3.

Kaplan–Meier curve of time to recurrent biliary obstruction. The median time to recurrent biliary obstruction was 7.1 months (95% CI, 4.3-NA). CI: Confidence interval; NA: Not available; RBO: Recurrent biliary obstruction

Reintervention via the SSS was successfully performed in 90.9% of RBO cases (10/11). Reintervention failures included one case in which the guidewire failed to pass the hyperplasia even with a trimmed stent. An additional stent was placed to cover the stricture in six cases (uncovered metal stent in three and plastic stent in three) of RBO due to hyperplasia. Balloon sweeping was performed in all the three cases of RBO due to food impaction. Balloon sweeping followed by placement of a plastic stent was performed in one case of RBO due to bleeding. Stent trimming using argon plasma coagulation (APC) was not required even with the long intragastric length in the 10 successful cases of reintervention. Reintervention could be performed through the gastric stent end in seven cases and the stent mesh in three cases.

DISCUSSION

In this study, EUS-HGS using the SSS, a newly designed C-SEMS with a spring-like anchoring system, resulted in high rates of technical and clinical success with an acceptable AE rate. Migration was not observed, and there were six cases (16.2%) in which the spring stopper effectively prevented proximal migration into the abdominal cavity.

A stent can migrate during deployment but also after being placed.[10-12] Intra-scope channel stent release was effective to prevent migration during the deployment step and shortens the distance between the liver and the stomach.[20] However, our previous study revealed that the gastric wall returns to its original position but that distance between the liver and the stomach cannot be determined during EUS-HGS.[14] Even when sufficient intragastric stent length was observed with the endoscopic view, it was occasionally shorter on the follow-up CT after EUS-HGS, and additional endoscopic rescue techniques such as “Crisscross” and “ClipFlap” were required.[21,22] Therefore, a long C-SEMS is required for EUS-HGS and favorable for anchoring at both sides of the stent.[13,23,24] Some reports have evaluated stents with anti-migration properties on the proximal and distal sides. Park et al. reported two newly developed SEMSs with anti-migration properties for EUS-BD.[25] They reported a fully covered SEMS (FC-SEMS) with anti-migration flaps on the proximal and distal ends and a PC-SEMS with anti-migration flaps at the distal end and a funnel-shaped uncovered portion at the proximal end. Neither C-SEMS migrated in 32 cases (20 for EUS-HGS, 12 for EUS-CDS). Cho et al. reported the long-term outcomes of a newly developed hybrid metal stent with a long uncovered portion on the proximal side and flaps on the distal and proximal sides of the covered portion in EUS-BD.[26] Endoscopic examinations revealed that some anchoring flaps were embedded in the stomach mucosa. In our study, there were six cases in which the stopper part of the gastric end was in close contact with the gastric wall despite the use of a longer stent. Proximal migration on the gastric side would have occurred in these cases if SEMS without anti-migration properties had been used instead of the SSS. We measured the anchoring force of the stents used for EUS-BD, and SSS showed stronger anchoring force than HANAROSTENT, HILZO, and Wallflex.[15] There is one report of a case in which the hepatic side of the SSS migrated into the abdominal cavity during stent deployment.[27] In that report, the SSS could not be pulled from the stomach wall without additional dissection around the puncture site. These in vitro and in vivo results reveal that the anchoring ability of the spring stopper is extremely high. In addition, the uncovered length on the hepatic side is also important for preventing distal migration out of the liver. The length of the uncovered hepatic portion should be increased to prevent distal migration of the hepatic side because of strong anchoring ability on gastric side. Therefore, the uncovered length of the SSS (15 or 20 mm) is longer than the uncovered length of a modified Giobor stent (10 mm).[13] As a result, there were no cases of migration during the mean follow-up period of 5.1 months. A sufficiently long stent with an effective anchoring system on both sides is necessary to reduce the risk for migration.

The main cause of RBO in PC-SEMS was hyperplasia at the uncovered portion. The SSS has a longer uncovered portion (15 or 20 mm) than a modified Giobor stent (10 mm), and the risk of hyperplasia was concerned to be higher. However, the incidence of hyperplasia of the SSS (18.9%) is comparable to that of the modified Giobor stent (23%).[24] Managing hyperplasia with a plastic stent is occasionally difficult and requires the placement of an additional SEMS.[24] Matsubara et al. reported a case in which bile duct radiofrequency ablation was useful for hyperplasia.[28] Management of hyperplasia should be established in the near future. On the contrary, FC-SEMS is favorable for preventing hyperplasia. The usefulness of FC-SEMS with a thin delivery system has been reported.[29-31] Although FC-SEMS is useful for preventing hyperplasia, dislocation or migration has been reported in 0%–20% of cases. FC-SEMS with an effective anchoring system on both sides is warranted to prevent hyperplasia and migration.

Although a long C-SEMS effectively prevents migration, a long stent can be difficult during the reintervention procedure when the stent is occluded. Trimming the SEMS with APC is useful[32] but cumbersome. The opportunity to perform reintervention through the stent mesh is increasing because of its simplicity.[33] In our study, reintervention was successfully performed through the gastric stent end in seven cases and the stent mesh in three cases. Reintervention could be performed without difficulty even if the intragastric length was long.

This study has several limitations. This was a retrospective pilot study with a relatively small number of patients, a short follow-up period, and a single arm without a control. A further large-scale prospective study with a control arm is required to confirm our outcomes.

CONCLUSIONS

In conclusion, we report here the safety and effectiveness of SSS for EUS-HGS in this retrospective study. No stent migrated after EUS-HGS. The spring-like anchor on the gastric side effectively prevented proximal migration.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.