- Email: [email protected]
Effectiveness of a newly designed antireflux valve metal stent to reduce duodenobiliary reflux in patients with unresectable distal malignant biliary obstruction: a randomized, controlled pilot study (with videos)
ORIGINAL ARTICLE
Effectiveness of a newly designed antireflux valve metal stent to reduce duodenobiliary reflux in patients with unresectable distal malignant biliary obstruction: a randomized, controlled pilot study (with videos) Yun Nah Lee, MD,1,* Jong Ho Moon, MD, PhD,1* Hyun Jong Choi, MD, PhD,1 Moon Han Choi, MD,1 Tae Hoon Lee, MD, PhD,1 Sang-Woo Cha, MD, PhD,1 Young Deok Cho, MD, PhD,1 Seo-Youn Choi, MD,2 Hae Kyung Lee, MD, PhD,2 Sang-Heum Park, MD, PhD1 Bucheon, Seoul, South Korea
Background and Aims: In patients with unresectable distal malignant biliary obstruction (MBO), endoscopic biliary drainage by using self-expandable metal stents (SEMSs) is an established palliative treatment. However, the placement of a SEMS across the major duodenal papilla prompts reflux of duodenal contents. In this study, we evaluated stent patency and duodenobiliary reflux caused by a newly developed SEMS with an antireflux valve (ARV) of the windsock type, compared with a conventional covered SEMS (cSEMS) in patients with MBO. Methods: Between January 2013 and September 2014, 77 patients with unresectable distal MBO were assigned randomly to groups treated with an ARV metal stent (ARVMS) group (39 patients) or a conventional cSEMS group (38 patients). In all patients, a barium meal examination was performed to evaluate reflux of barium within the SEMS and intrahepatic bile ducts. The primary outcome was stent patency duration. Secondary outcomes were the rates of technical and clinical success, duodenobiliary reflux on barium meal examination, factors causing stent dysfunction, overall patient survival, and adverse events. Results: Stent placement was technically successful in all patients. The clinical success rates were not statistically significantly different between the ARVMS and cSEMS groups (97.4% vs 97.4%, P Z 1.000). Overall reflux of barium was significantly lower in the ARVMS group than the cSEMS group (7.7% vs 100%, P < .001). The cumulative duration of stent patency was significantly longer in the ARVMS group than in the cSEMS group (median SD, 407 92 vs 220 37 days; P Z .013). On multivariate analysis, complete duodenobiliary reflux (odds ratio, 5.7, P Z .004) and ampullary cancer (odds ratio, 8.98, P Z .012) were identified as independent risk factors for stent dysfunction. There was no significant difference between the 2 groups in overall patient survival or in the incidence of adverse events. Conclusions: The newly developed ARVMS seemed to have a superior duration of stent patency and comparable safety compared with the cSEMS. In addition, the duodenobiliary reflux related to stent dysfunction can be prevented effectively by ARVMS. Further randomized, controlled trials using large numbers of subjects are required to confirm the benefit of SEMSs with antireflux function. (Clinical trial registration number: UMIN000012734.) (Gastrointest Endosc 2015;-:1-9.)
Abbreviations: ARMS, antireflux metal stent; ARV, antireflux valve; ARVMS, antireflux valve metal stent; CI, confidence interval; cSEMS, covered self-expandable metal stent; ePTFE, e-polytetrafluoroethylene; MBO, malignant biliary obstruction; OR, odds ratio; SEMS, self-expandable metal stent. DISCLOSURE: All authors disclosed no financial relationships relevant to this publication. *Authors Lee and Moon contributed equally to this article. Copyright ª 2015 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2015.08.084
www.giejournal.org
Received March 21, 2015. Accepted August 30, 2015. Current affiliations: Digestive Disease Center and Research Institute, Departments of Internal Medicine (1) and Radiology (2), SoonChunHyang University School of Medicine, Bucheon and Seoul, South Korea. Reprint requests: Jong Ho Moon, MD, PhD, SoonChunHyang University School of Medicine, Digestive Disease Center and Research Institute, SoonChunHyang University Bucheon Hospital, 170 Jomaru-ro, Wonmigu, Bucheon 420-767, South Korea. If you would like to chat with an author of this article, you may contact Dr Moon at [email protected].
Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 1
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
In patients with unresectable distal malignant biliary obstruction (MBO), endoscopic biliary drainage by using self-expandable metal stents (SEMSs) is an established palliative treatment for the relief of obstructive jaundice. Many previous studies have indicated that SEMSs are superior to plastic stents for maintaining patency in patients with MBO because of their larger diameter.1-4 However, although effective, a SEMS does not maintain patency for extremely long periods of time; the stent usually becomes obstructed as a result of tumor ingrowth and/or overgrowth, epithelial hyperplasia, biofilm deposition, or sludge. In particular, placement of an SEMS across the major duodenal papilla promotes reflux of duodenal contents and ascending infection.5 In a previous study using a barium examination, free barium reflux through a transpapillary SEMS was observed in all patients, and enteric biliary reflux was suggested as a major cause of cholangitis and as playing an important role in stent occlusion.6 Thus, especially in patients at high risk of duodenobiliary reflux, reduction of duodenobiliary reflux can be an important factor in prolonging stent patency, and several antireflux metal stents (ARMSs) have been developed.7-9 However, the studies were mostly preliminary, and objective results showing the prevention of duodenobiliary reflux by ARMSs have not been reported. In this study, we evaluated stent patency and duodenobiliary reflux of a newly developed SEMS with an antireflux valve (ARV) of the windsock type compared with a “conventional” covered SEMS (cSEMS) in patients with MBO.
METHODS Patients Consecutive patients with histologically confirmed unresectable MBO were enrolled prospectively in this study. The inclusion criteria were (1) obstructive jaundice, (2) MBO revealed by CT and/or magnetic resonance imaging, (3) typical radiological appearance of MBO on ERCP, (4) MBO proximal margin 2 cm or more from the hepatic confluence, (5) not a candidate for curative surgical resection due to tumor stage or operative risks, (6) older than 18 years of age, and (7) ability to provide informed consent. The exclusion criteria were (1) presence of any contraindication to ERCP, (2) previous biliary SEMS placement, surgery, or drainage procedure, and (3) active hepatitis or other hepatic diseases that may cause jaundice. Our institutional review board approved this study. Written informed consent was obtained from all enrolled patients. This study was registered on the UMIN Clinical Trial Registry (UMIN000012734).
Figure 1. A, The newly developed self-expandable metal stent with an antireflux valve (*). B, An 8F modified delivery system having a “bumper” for safe deployment and prevention of creasing of the valve.
designed with an ARV (Fig. 1A). The ARV metal stent (ARVMS) is 10 mm in diameter, and the metal portion is available in 60-, 70-, 80-, 90-, and 100-mm total lengths. The length of the valve portion is 20 mm. The stent is made of nitinol wire and consists of a bare portion with a flared proximal end of 7 mm, a covered body portion, and an antireflux valve portion at the distal end. The antireflux stent consists of an SEMS partially covered by an e-polytetrafluoroethylene (ePTFE) membrane that extends as a pliable tube beyond the distal end of the stent creating a windsock-type ARV. The stent is contained in a standard 8F, pull-back delivery system. The system was slightly modified to ensure safe deployment of the valve and to prevent it from creasing. Attached to the distal core of the delivery system is a novel expandable metal stent– like “bumper” consisting of 20-mm segment of an 8-mm nitinol stent (Fig. 1B). This is situated within the valve and expands and straightens during deployment. A purple marker on the outside of the stent indicates the transition point from the distal stent end to the valve. During deployment of the stent, the purple line is 5 to 10 mm below the papilla in the duodenal lumen, and release of the distal stent end and the valve is performed under endoscopic control. After successful placement of the valve in the duodenal lumen, a delivery system with the metal bumper is retracted and removed from the lumen of the valve, leaving the valve correctly positioned within the duodenal lumen (Video 1, available online at www.giejournal.org).
A newly developed SEMS with a windsocktype ARV
Study design
A new SEMS having an ARV (EGIS Biliary Stent M-Valve; S&G Biotech, Seongnam, South Korea) was modified from a cSEMS (EGIS Biliary Stent; S&G Biotect) and specially
This study was conducted as an open-label, randomized trial at a single center. Once informed consent was obtained and if the patients met the inclusion criteria and
2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2015
www.giejournal.org
Lee et al
Effectiveness of a newly designed antireflux valve metal stent
Figure 2. Endoscopic drainage by using a new self-expandable metal stent (SEMS) with an antireflux valve for malignant biliary obstruction in a patient with pancreatic cancer. A, Cholangiogram showing narrowing of the mid- to distal common bile duct with dilation of the proximal bile duct. B, A new SEMS with an antireflux valve was inserted and C, deployed within the bile duct. D, Fully deployed new SEMS. E, Endoscopic view showing a purple marker indicating the beginning of the valve portion and a bumper on the distal end of the delivery system for protection of the antireflux valve in the introducer. F, The valve portion of fully deployed new SEMS on the duodenum.
demonstrated none of the exclusion criteria, they were assigned randomly to receive an ARVMS (ARVMS group) or a conventional partially cSEMS (cSEMS group) at an allocation ratio of 1:1. Blocked randomization (block size 4) was performed by drawing sealed opaque envelopes. www.giejournal.org
Each envelope contained a card marked with a stent type, and the randomization process was performed by 1 of the participating endoscopists in the ERCP room; after the guidewire was placed correctly and the stricture was passed, the stent was then inserted immediately. Blinding Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 3
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
Figure 3. Barium meal examination for evaluation of duodenobiliary reflux. A, Complete reflux. B, Partial reflux. C, No reflux.
of the endoscopists to stent type was not attempted. The radiologists conducting the interpretation of barium study were blinded to stent assignment.
Endoscopic interventions All procedures were performed by using standardized protocols by 2 experienced investigators using a standard duodenoscope (TJF-260V; Olympus, Tokyo, Japan). An endoscopic sphincterotomy was performed to facilitate stent insertion in all patients. Standard techniques were used to cannulate the biliary tract, and contrast was injected to identify the location and length of the stricture. For patients randomly assigned to the cSEMS group, a conventional partially cSEMS (EGIS biliary stent; S&G Biotect) was used. The new SEMS with an ARV (EGIS Biliary Stent M-Valve; S&G Biotech) was placed in patients assigned to the ARVMS group. The length of the stents was determined according to the length of the bile duct stricture. Positioning of the stent in the center of the stricture was encouraged, and a 5- to 10-mm length of the SEMS should be visible in the duodenum. A stent was inserted 10 to 20 mm proximal to the end of the stricture (Fig. 2).
Evaluation of duodenobiliary reflux Two or 3 days after endoscopic SEMS insertion, a barium meal examination was performed by 2 experienced radiologists in all patients to evaluate the reflux of barium within the SEMS and intrahepatic bile ducts. The duodenobiliary reflux was classified as complete (Fig. 3A), partial (Fig. 3B), or absent (Fig. 3C). Complete reflux was defined as reflux of barium above the proximal end of the SEMS, and partial reflux was defined as a reflux of barium under the proximal end of the SEMS (Video 2, available online at www.giejournal.org).
Follow-up and event definitions To confirm a successful drainage procedure, liver function tests were performed before and 2 to 7 days after stent 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2015
insertion. The patients were scheduled to visit the hospital after 1 month, and liver function tests and a physical examination were performed. Then, clinical and liver function follow-up was conducted at 1-month intervals with a hospital visit. Patients lost to follow-up were analyzed on an intention-to-treat basis and were censored at the time of their last follow-up interview. Technical success was defined as successful stent insertion, based on radiological confirmation on ERCP that the stent was in an appropriate position. Clinical success was defined as at least a 50% decrease in bilirubin levels during the first 7 days after stent insertion. All procedure-related adverse events were recorded according to normal practice.10 Early adverse events were defined as those that occurred during the procedure or within 30 days of stent insertion, and late adverse events were those occurring after this period. Stent dysfunction was considered to occur if the patient had obstructive jaundice and a finding of biliary dilation on imaging studies (CT, magnetic resonance imaging, or US) with or without cholangitis. Endoscopic reintervention was then arranged to confirm and assess the cause of stent dysfunction. If the patients showed recurrent jaundice with high-grade fever, urgent reintervention was also performed, although biliary dilation was not confirmed by imaging studies. Tumor overgrowth was defined as stent occlusion by a tumoral mass over the stent. Tumor ingrowth was defined as stent occlusion by the tumor inside the stent. Migration, proximally or distally, was defined as SEMS movement over the MBO from the original post-placement position, as documented by the final spot radiograph obtained at the initial ERCP.
Outcome measurements The primary outcome was stent patency duration, defined as the time of endoscopic stent deployment until stent dysfunction requiring reintervention, death, loss to follow-up, or patients alive at the end of the study with no occlusion. Secondary outcomes were the rates of www.giejournal.org
Lee et al
Effectiveness of a newly designed antireflux valve metal stent
Assessed for eligibility (n=141) Excluded (n=64) – Surgical treatment: 23 – Multiple stricture: 6 – Close to hilum: 13 – Previous Billroth II operation: 3 – Refused to participate: 19 Randomized (n= 77) Intention-to-treat follow-up
ARVMS (n=39)
cSEMS (n=38)
Lost to follow up (n=3)
Lost to follow up (n=4)
Stent dysfunction (n=7) Died without stent dysfunction (n=26) Alive without stent dysfunction (n=3)
Stent dysfunction (n=14) Died without stent dysfunction (n=19) Alive without stent dysfunction (n=1) Analysis
Figure 4. Flowchart showing the results of inclusion, randomization, and follow-up of the patients. ARVMS, antireflux valve with metal stent; cSEMS, covered self-expandable metal stent.
technical and clinical success, the duodenobiliary reflux on barium meal examination, risk factors causing stent dysfunction, overall patient survival, and adverse events. Patient survival was measured from endoscopic stent deployment to death or the day of last follow-up.
with a multivariate analysis. All statistical analyses were performed using the SPSS software (version 18.0; SPSS Inc, Chicago, Ill). P values <.05 were considered to indicate statistical significance.
RESULTS Statistical analysis This study was intended as a pilot study; therefore, no formal power calculation was performed. The stent patency time and overall patient survival were evaluated by using the Kaplan-Meier method and compared by using the log-rank test. Categorical parameters were expressed as frequencies and proportions and compared by using the c2 test or Fisher exact test. Continuous data are expressed as medians with ranges or SDs and compared by using the Mann-Whitney U test. Univariate and multivariate analyses were performed to identify risk factors for stent dysfunction by using logistic regression analysis. The model included age (<60 vs 60 years), sex, previous plastic stent insertion, stent type (cSEMS vs ARVMS), length of stent, tumor etiology, duodenobiliary reflux on barium meal examination (no or partial vs complete), and adjuvant therapy. For each factor considered to have a potential association with stent dysfunction, the odds ratio (OR) with its 95% confidence interval (CI) was calculated. Factors with P values <.10 on univariate analysis were further analyzed www.giejournal.org
Between January 2013 and September 2014, 141 patients were screened; 64 were excluded for the following reasons: surgical treatment (n Z 23), close to the hilum (n Z 13), multiple strictures of the bile duct (n Z 6), previous Billroth II operation (n Z 3), and refusal to participate (n Z 19). The remaining 77 patients constituted the study cohort; 39 patients were randomly assigned to the ARVMS group and 38 patients to the cSEMS group. Loss to follow-up in the ARVMS and cSEMS groups occurred in 3 patients (7.7%) and 4 patients (10.5%), respectively (P Z .711) (Fig. 4). Baseline characteristics were not significantly different between the ARVMS and cSEMS groups. The causes of MBO were pancreatic cancer (n Z 45), cholangiocarcinoma (n Z 21), ampullary cancer (n Z 9), and metastatic nodes (n Z 2). There was no significant difference between the groups in terms of the underlying diagnoses. In total, 52 patients underwent adjuvant therapy after stent insertion: 29 (74.4%) in the ARVMS group and 23 (60.5%) in the cSEMS group (Table 1). Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 5
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
Stent placement and early clinical outcomes Stent placement was technically successful in all patients. The clinical success rates were not statistically significantly different between the ARVMS and cSEMS groups (97.4% vs 97.4%, P Z 1.00). No stent dysfunction was observed in either group within 30 days after stent placement. Early adverse events were observed in 3 patients (2 cases of mild pancreatitis and 1 case of cholangitis) and 6 patients (4 cases of mild pancreatitis and 2 cases of cholangitis) in the ARVMS group and cSEMS groups, respectively. The early adverse events in the ARVMS group were all considered to be procedure-related events. In 1 patient in the cSEMS group, cholangitis was diagnosed 25 days after stent placement; however, stent dysfunction was not observed by ERCP. There was no significant difference in 30-day mortality between the 2 groups (Table 2).
Stent patency and patient survival The median follow-up durations in the ARVMS and cSEMS groups were 156 days (range 20-507 days) and 189.5 days (range 25-610 days), respectively (P Z .318). The cumulative duration of stent patency in the ARVMS group was significantly longer than that in the cSEMS group (median SD, 407 92 vs 220 37 days; P Z .013, log-rank test) (Fig. 6). During follow-up, the occurrence of stent dysfunction did not differ between the ARVMS and cSEMS groups (P Z .077). Additionally, the detailed causes of stent dysfunction, including sludge occlusion (P Z .224), did not differ between the groups. In the ARVMS group, in total, 3 late adverse events (2 cases of mild pancreatitis and 1 case of hemobilia) were observed, with no significant difference from the cSEMS group (P Z 1.00). The 1 hemobilia case of unknown cause in the ARVMS group occurred 35 days after stent insertion and was treated by reinsertion of a metal stent during ERCP. There was no significant difference in overall patient survival, with the median SD survival being 289 33 days in the ARVMS group and 340 36 days in the cSEMS group (P Z .661, log-rank test) (Table 3).
Risk factors for stent dysfunction Univariate and multivariate analyses of the risk factors for stent dysfunction were performed. Univariate analysis No.
-
Age, y Sex, male/female, no. Total bilirubin, mg/dL Previous plastic stent, no. (%) Duration of plastic stent, days Length of stricture, mm
: 2015
ARVMS (n [ 39)
cSEMS (n [ 38)
P value
74 (46–91)
73 (44–101)
.672
22/17
15/23
.173
4.4 (0.3–20.1) 4.1 (1.4–35.7) 28 (73.7)
1.00
6 (3–150)
7 (3–152)
.317
19 (10–44)
23 (7–700)
.296 .462
6
1 (2.6)
4 (10.5)
7
37 (94.9)
29 (76.3)
1 (2.6)
5 (15.1)
18 (46.2)
16 (42.1)
Tumor etiology, no. (%)
.820 .223
Pancreatic cancer
22 (56.4)
23 (60.5)
Cholangiocarcinoma
10 (25.6)
11 (28.9)
Ampullary cancer
5 (12.8)
4 (10.5)
Metastatic nodes
2 (5.2)
0
29 (74.4)
23 (60.5)
Adjuvant therapy, no. (%)
.790
29 (74.4)
Length of stent, cm, no. (%)
Distant metastasis, no. (%)
A barium meal examination was performed in all 77 patients. In the cSEMS group, the free reflux of barium occurred through the SEMSs in all patients. Overall reflux (complete or partial) of barium was seen in 3 patients of the ARVMS group, which was significantly different from the cSEMS group (7.7% vs 100%, P < .001). Complete reflux in the cSEMS group was also significantly greater than in the ARVMS group (2.6% vs 55.3%, P < .001) (Fig. 5).
-,
Type of stent
8
Duodenobiliary reflux on barium meal examination
6 GASTROINTESTINAL ENDOSCOPY Volume
TABLE 1. Baseline characteristics of the patients in the ARVMS and cSEMS groups
.230
Quantitative data are expressed as median (range). ARVMS, Antireflux valve metal stent; cSEMS, covered self-expandable metal stent.
TABLE 2. Early technical and clinical outcomes of patients in the ARVMS and cSEMS groups Type of stent ARVMS (n [ 39) Technical success, no.
cSEMS (n [ 38)
P value
39
38
d
Clinical success, no. (%)
38 (97.4)
37 (97.4)
1.00
Adverse events, no. (%)
3 (7.7)
6 (15.8)
.310
Mild pancreatitis
2
4
.431
Cholangitis 30-day mortality, no. (%)
1
2
.615
2 (5.1)
1 (2.6)
1.00
ARVMS, Antireflux valve metal stent; cSEMS, covered self-expandable metal stent.
revealed that complete duodenobiliary reflux on barium meal examination (odds ratio [OR], 6.13; P Z .001) and ampullary cancer (OR, 9.13; P Z .006) were risk factors for stent dysfunction. ARVMS (OR, 0.38; P Z.067) and adjuvant therapy (OR, Z 0.53; P Z .236) were not significant factors on univariate analyses. On multivariate analyses, complete duodenobiliary reflux (OR, 5.7; P Z .004) and ampullary cancer (OR, 8.98; P Z .012) were identified as independent risk factors for stent dysfunction. A post hoc power analysis based on a log-rank test yielded a power of approximately 75% (1-b Z 0.7457) to www.giejournal.org
Lee et al
P < .001 100%
P < .001 ARVMS cSEMS
80
P = .013 80
60 55.3% 40 20 0
censored
100
7.7% Reflux
2.6% Complete reflux
Figure 5. Duodenobiliary reflux of patients in the ARVMS and cSEMS groups by using a barium meal examination. ARVMS, antireflux valve with metal stent; cSEMS, covered self-expandable metal stent.
detect estimated differences in median survival time of stent patency, which was 407 days for the 39 patients in the ARVMS group and 220 days for the 38 patients in the cSEMS group under the assumption that the survival function follows the exponential distribution. This power was computed by using STATA/SE 12.0 version with the stpower log-rank command (StataCorp, College Station, Tex).
DISCUSSION In patients with unresectable distal MBO, SEMS placement is regarded as the standard treatment, with a larger luminal diameter and longer patency than those of plastic stents.10,11 However, most SEMSs for distal MBO should be placed at the distal end of the duodenum. As a result, sphincter function is lost, allowing free duodenobiliary reflux of duodenal residue.6 In previous studies, the initial event leading to stent occlusion has been suggested to be the formation of biofilm by the adherence of host proteins and bacteria to the inner wall of the stent.12,13 This biofilm presumably protects the bacteria against antimicrobial agents. Bacterial glucuronidases and phospholipases act on bile components, leading to biliary sludge formation composed of bacterial products, calcium bilirubinate, and calcium fatty acid soaps. In addition, a study recently revealed that duodenobiliary reflux confirmed by using barium meal examination was an independent risk factor associated with CBD stone recurrence.14 Thus, reducing duodenobiliary reflux has been an attractive idea for prolonging the duration of stent patency, and several newly designed stents with antireflux function have been developed. Most of them showed efficacy and safety for palliation of jaundice caused by distal MBO, with a remarkably low risk of cholangitis and with improved stent patency. However, only small numbers of patients were enrolled in studies with nonrandomized www.giejournal.org
Cumulative patency, %
100
Effectiveness of a newly designed antireflux valve metal stent
ARVMS (n=39) 60
40
cSEMS (n=38)
20
0 0
100
200
300
400
500
600
Time, days Figure 6. Kaplan-Meier curve showing the cumulative stent patency of the stent in the antireflux valve metal stent (ARVMS) group and covered self-expandable metal stent (cSEMS) group (P Z .013, log-rank test).
designs.8,15,16 Recently, Hu et al9 reported results of a prospective, randomized, controlled trial in which a new partially covered ARMS and a standard uncovered SEMS were used. In a total of 112 patients with MBO, fewer patients experienced cholangitis in the partially covered ARMS group than in the uncovered SEMS group (10 vs 21 patients, P Z .035), and the frequency of episodes was lower (P Z .022). Additionally, the median stent patency SD was significantly longer in the partially covered ARMS group (13.0 3.4 vs 10.0 1.2 months, P Z .044). Although there was a limitation in that they compared stent patency of a partially covered ARMS and an uncovered SEMS, the study showed that an antireflux function could improve the duration of SEMS patency and reduce the occurrence of cholangitis after insertion of ARMS. For this trial, we developed a newly designed SEMS with an ARV. This stent has a relatively long valve length (20 mm) on the distal end of the metal stent. The valve is similar in appearance to a windsock and is made of e-polytetrafluoroethylene. Because 1 membrane covers the body and valve portions on the distal end of the stent, the distal end of the metal stent during stent deployment cannot be seen by the endoscope. However, there is a purple marker on the covered membrane at the distal end of the metal stent to indicate the start of the valve portion, such that the position of the distal end can be determined during stent deployment without difficulty. In this study, technical success of the procedure by using the ARVMS was achieved in all patients. If the endoscopic revision of the ARVMS was required due to stent dysfunction, stent Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 7
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
TABLE 3. Long-term clinical outcomes of patients in the ARVMS and cSEMS groups Type of stent ARVMS (n [ 39) Follow-up duration, median (range), days
cSEMS (n [ 38)
156 (20–507) 189.5 (25–610)
P value .318
Stent patency, median (SD), days
407 (92)
220 (37)
.013
Stent dysfunction, no. (%)
7 (17.9)
14 (36.8)
.077
Sludge occlusion
4 (10.3)
8 (21.1)
.224
Stent migration
3 (7.7)
3 (7.9)
1.00
Valve dysfunction
2 (5.1)
0
d
Tumor ingrowth
1 (2.6)
3 (7.9)
.358
0
2 (5.3)
.240
Causes of stent dysfunction, no. (%)*
Tumor overgrowth Hemobilia
1 (2.6)
0
1.00
Unknown
1 (2.6)
1 (2.6)
1.00
Adverse event, no. (%)
3 (7.7)
3 (7.9)
1.00
Mild pancreatitis
2 (5.1)
2 (5.3)
0
1 (2.6)
Cholecystitis Hemorrhage Survival, median (SD), days
1 (2.6)
0
289 (33)
340 (36)
.661
ARVMS, Antireflux valve metal stent; cSEMS, covered self-expandable metal stent. *Stent dysfunction may have occurred for more than 1 reason; thus, the total does not add up to the number of occurrences of stent dysfunction in each group.
removal was tried by using snare at first. When stent removal failed and/or insertion of the guidewire inside the ARVMS was needed, a guidewire could be inserted through the distal end of valve portion or a hole made on the valve by using a needle-knife. In this study, the impact of the antireflux function of the ARVMS was evaluated by barium meal examination. Reflux of barium in the ARVMS group was seen in only 3 patients, whereas all patients in the cSEMS group showed free reflux of barium (7.7% vs 100%, respectively; P < .001). Complete reflux in the cSEMS group was also significantly greater than that in the ARVMS group (2.6% vs 55.3%, respectively; P < .001). Additionally, complete duodenobiliary reflux was an independent risk factor for stent dysfunction on multivariate analysis (OR, 5.7; P Z .004). The ARVMS showed significantly longer stent patency than that of cSEMS (median SD, 407 92 vs 220 37 days, P Z .013) and was safe to use. However, our data did not show significant evidence of the ability to reduce stent dysfunction on multivariate analysis (OR, 0.57; P Z .859). Nevertheless, the efficacy of ARVMS for reducing stent dysfunction was suggested because the ARVMS showed reduced complete duodenobiliary reflux, which is an independent risk factor for stent dysfunction. Ampullary cancer (OR, 8.98; P Z .012) was also an inde8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2015
pendent risk factor for stent dysfunction. In patients with ampullary cancer, stent migration was observed in 3 of 4 (75%) cases of stent dysfunction. It seemed to be caused by a short stricture, which is limited to the short-segment ampullary lesion, which was insufficient to anchor the stent in situ.17 During the follow-up period, cholangitis due to valve dysfunction developed in 2 patients 402 and 81 days after stent insertion, respectively. In those patients, increased hardness of the valve was observed after stent removal. However, it was unclear what affected the hardness of the ARV: the material of the ARV itself, the types of food content, and/or the duodenal environment of the patients. Further studies and evaluations are required to address this. Our study had several limitations. First, there were no significant differences between the ARVMS group and the cSEMS group in terms of frequency of SEMS dysfunction, patient survival, and patients lost to follow-up. Nevertheless, the cumulative duration of stent patency in the ARVMS group was significantly longer than that in the cSEMS group (median SD, 407 92 vs 220 37 days, P Z .013). This discrepancy is considered to be caused by the limitation that it was a pilot study that began without a formal sample size calculation. However, although there was no significant difference, the stent dysfunction in the cSEMS group occurred about twice as often as in the ARVMS group (36.8% vs 17.9%, P Z .077). The duration until the development of stent dysfunction in the ARVMS group tended to occur later than that in the cSEMS group, but there was no significant difference (median, 181 vs 106 days; P Z .172). The significantly longer stent patency of ARVMS is already a clinically important result because ARVMS may be able to reduce the need for reintervention in patients with palliatively treated MBO. Therefore, further studies involving larger cohorts are needed to confirm and strengthen our results. Second, the current results were obtained by 2 experienced endoscopists using endoscopic biliary drainage with a new ARVMS. Although the technical success rate of the procedure was 100% in our study, it remains to be determined whether the promising results of this study are reproducible by less-experienced endoscopists. A third limitation was the nonblinding of the physicians to the stents, both before and after the procedure. However, the radiologists and patients were blinded to the stents. The final limitations are the relatively short follow-up period and the 9.1% of patients lost to follow-up. These might affect the relatively low incidence of stent dysfunction in our study compared with previous studies. In conclusion, our prospective, randomized study performed in 77 patients with unresectable distal MBO showed that the newly developed ARVMS has a significantly longer duration of stent patency and comparable safety compared with cSEMS. Additionally, the duodenobiliary reflux related to stent dysfunction can be effectively prevented by use of the ARVMS. However, the survival of www.giejournal.org
Lee et al
patients was not different between ARVMS and cSEMS placement in these patients. Further randomized, controlled trials involving larger cohorts are required to confirm the benefits of SEMSs with antireflux functions. ACKNOWLEDGMENTS We thank Dr Sung Kwon Kang, Mr Young Jae Lee, and Mr Seung Hwan Jegal of S&G Biotech, Seongnam, South Korea, for the development of the new SEMS with an antireflux valve. We thank Professor Joo Ha Hwang, Seattle, Washington, USA, Dr Hans-Ulrich Laasch, Manchester, UK, and Mr Derek W. Edwards for proofreading our manuscript. We thank A Ri Song, RN, Song Ah Jeong, RN, and Sun Yeong Moon, RN, and the rest of the nursing staff for assistance with the procedures. Professor Jong Ho Moon is the inventor of the EGIS Biliary Stent M-Valve. This work was supported, in part, by the SoonChunHyang University Research Fund.
REFERENCES 1. Davids PH, Groen AK, Rauws EA, et al. Randomised trial of selfexpanding metal stents versus polyethylene stents for distal malignant biliary obstruction. Lancet 1992;340:1488-92. 2. Prat F, Chapat O, Ducot B, et al. A randomized trial of endoscopic drainage methods for inoperable malignant strictures of the common bile duct. Gastrointest Endosc 1998;47:1-7. 3. Soderlund C, Linder S. Covered metal versus plastic stents for malignant common bile duct stenosis: a prospective, randomized, controlled trial. Gastrointest Endosc 2006;63:986-95. 4. Kaassis M, Boyer J, Dumas R, et al. Plastic or metal stents for malignant stricture of the common bile duct? Results of a randomized prospective study. Gastrointest Endosc 2003;57:178-82.
www.giejournal.org
Effectiveness of a newly designed antireflux valve metal stent 5. Loew BJ, Howell DA, Sanders MK, et al. Comparative performance of uncoated, self-expanding metal biliary stents of different designs in 2 diameters: final results of an international multicenter, randomized, controlled trial. Gastrointest Endosc 2009;70:445-53. 6. Misra SP, Dwivedi M. Reflux of duodenal contents and cholangitis in patients undergoing self-expanding metal stent placement. Gastrointest Endosc 2009;70:317-21. 7. Kim DU, Kwon CI, Kang DH, et al. New antireflux self-expandable metal stent for malignant lower biliary obstruction: in vitro and in vivo preliminary study. Dig Endosc 2013;25:60-6. 8. Hamada T, Isayama H, Nakai Y, et al. Novel antireflux covered metal stent for recurrent occlusion of biliary metal stents: a pilot study. Dig Endosc 2014;26:264-9. 9. Hu B, Wang TT, Wu J, et al. Antireflux stents to reduce the risk of cholangitis in patients with malignant biliary strictures: a randomized trial. Endoscopy 2014;46:120-6. 10. Yeoh KG, Zimmerman MJ, Cunningham JT, et al. Comparative costs of metal versus plastic biliary stent strategies for malignant obstructive jaundice by decision analysis. Gastrointest Endosc 1999;49:466-71. 11. Arguedas MR, Heudebert GH, Stinnett AA, et al. Biliary stents in malignant obstructive jaundice due to pancreatic carcinoma: a costeffectiveness analysis. Am J Gastroenterol 2002;97:898-904. 12. Leung JW, Liu Y, Chan RC, et al. Early attachment of anaerobic bacteria may play an important role in biliary stent blockage. Gastrointest Endosc 2000;52:725-9. 13. Leung JW, Liu YL, Chan RC, et al. Effects of adherence factors and human bile on bacterial attachment and biliary stent blockage: an in vitro study. Gastrointest Endosc 2002;56:72-7. 14. Zhang R, Luo H, Pan Y, et al. Rate of duodenal-biliary reflux increases in patients with recurrent common bile duct stones: evidence from barium meal examination. Gastrointest Endosc 2015;82:660-5. 15. Lee KJ, Chung MJ, Park JY, et al. Clinical advantages of a metal stent with an S-shaped anti-reflux valve in malignant biliary obstruction. Dig Endosc 2013;25:308-12. 16. Hu B, Wang TT, Shi ZM, et al. A novel antireflux metal stent for the palliation of biliary malignancies: a pilot feasibility study (with video). Gastrointest Endosc 2011;73:143-8. 17. Park SB, Kim HW, Kang DH, et al. Metallic or plastic stent for bile duct obstruction in ampullary cancer? Dig Dis Sci 2012;57:786-90.
Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 9
Effectiveness of a newly designed antireflux valve metal stent to reduce duodenobiliary reflux in patients with unresectable distal malignant biliary obstruction: a randomized, controlled pilot study (with videos) Yun Nah Lee, MD,1,* Jong Ho Moon, MD, PhD,1* Hyun Jong Choi, MD, PhD,1 Moon Han Choi, MD,1 Tae Hoon Lee, MD, PhD,1 Sang-Woo Cha, MD, PhD,1 Young Deok Cho, MD, PhD,1 Seo-Youn Choi, MD,2 Hae Kyung Lee, MD, PhD,2 Sang-Heum Park, MD, PhD1 Bucheon, Seoul, South Korea
Background and Aims: In patients with unresectable distal malignant biliary obstruction (MBO), endoscopic biliary drainage by using self-expandable metal stents (SEMSs) is an established palliative treatment. However, the placement of a SEMS across the major duodenal papilla prompts reflux of duodenal contents. In this study, we evaluated stent patency and duodenobiliary reflux caused by a newly developed SEMS with an antireflux valve (ARV) of the windsock type, compared with a conventional covered SEMS (cSEMS) in patients with MBO. Methods: Between January 2013 and September 2014, 77 patients with unresectable distal MBO were assigned randomly to groups treated with an ARV metal stent (ARVMS) group (39 patients) or a conventional cSEMS group (38 patients). In all patients, a barium meal examination was performed to evaluate reflux of barium within the SEMS and intrahepatic bile ducts. The primary outcome was stent patency duration. Secondary outcomes were the rates of technical and clinical success, duodenobiliary reflux on barium meal examination, factors causing stent dysfunction, overall patient survival, and adverse events. Results: Stent placement was technically successful in all patients. The clinical success rates were not statistically significantly different between the ARVMS and cSEMS groups (97.4% vs 97.4%, P Z 1.000). Overall reflux of barium was significantly lower in the ARVMS group than the cSEMS group (7.7% vs 100%, P < .001). The cumulative duration of stent patency was significantly longer in the ARVMS group than in the cSEMS group (median SD, 407 92 vs 220 37 days; P Z .013). On multivariate analysis, complete duodenobiliary reflux (odds ratio, 5.7, P Z .004) and ampullary cancer (odds ratio, 8.98, P Z .012) were identified as independent risk factors for stent dysfunction. There was no significant difference between the 2 groups in overall patient survival or in the incidence of adverse events. Conclusions: The newly developed ARVMS seemed to have a superior duration of stent patency and comparable safety compared with the cSEMS. In addition, the duodenobiliary reflux related to stent dysfunction can be prevented effectively by ARVMS. Further randomized, controlled trials using large numbers of subjects are required to confirm the benefit of SEMSs with antireflux function. (Clinical trial registration number: UMIN000012734.) (Gastrointest Endosc 2015;-:1-9.)
Abbreviations: ARMS, antireflux metal stent; ARV, antireflux valve; ARVMS, antireflux valve metal stent; CI, confidence interval; cSEMS, covered self-expandable metal stent; ePTFE, e-polytetrafluoroethylene; MBO, malignant biliary obstruction; OR, odds ratio; SEMS, self-expandable metal stent. DISCLOSURE: All authors disclosed no financial relationships relevant to this publication. *Authors Lee and Moon contributed equally to this article. Copyright ª 2015 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2015.08.084
www.giejournal.org
Received March 21, 2015. Accepted August 30, 2015. Current affiliations: Digestive Disease Center and Research Institute, Departments of Internal Medicine (1) and Radiology (2), SoonChunHyang University School of Medicine, Bucheon and Seoul, South Korea. Reprint requests: Jong Ho Moon, MD, PhD, SoonChunHyang University School of Medicine, Digestive Disease Center and Research Institute, SoonChunHyang University Bucheon Hospital, 170 Jomaru-ro, Wonmigu, Bucheon 420-767, South Korea. If you would like to chat with an author of this article, you may contact Dr Moon at [email protected].
Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 1
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
In patients with unresectable distal malignant biliary obstruction (MBO), endoscopic biliary drainage by using self-expandable metal stents (SEMSs) is an established palliative treatment for the relief of obstructive jaundice. Many previous studies have indicated that SEMSs are superior to plastic stents for maintaining patency in patients with MBO because of their larger diameter.1-4 However, although effective, a SEMS does not maintain patency for extremely long periods of time; the stent usually becomes obstructed as a result of tumor ingrowth and/or overgrowth, epithelial hyperplasia, biofilm deposition, or sludge. In particular, placement of an SEMS across the major duodenal papilla promotes reflux of duodenal contents and ascending infection.5 In a previous study using a barium examination, free barium reflux through a transpapillary SEMS was observed in all patients, and enteric biliary reflux was suggested as a major cause of cholangitis and as playing an important role in stent occlusion.6 Thus, especially in patients at high risk of duodenobiliary reflux, reduction of duodenobiliary reflux can be an important factor in prolonging stent patency, and several antireflux metal stents (ARMSs) have been developed.7-9 However, the studies were mostly preliminary, and objective results showing the prevention of duodenobiliary reflux by ARMSs have not been reported. In this study, we evaluated stent patency and duodenobiliary reflux of a newly developed SEMS with an antireflux valve (ARV) of the windsock type compared with a “conventional” covered SEMS (cSEMS) in patients with MBO.
METHODS Patients Consecutive patients with histologically confirmed unresectable MBO were enrolled prospectively in this study. The inclusion criteria were (1) obstructive jaundice, (2) MBO revealed by CT and/or magnetic resonance imaging, (3) typical radiological appearance of MBO on ERCP, (4) MBO proximal margin 2 cm or more from the hepatic confluence, (5) not a candidate for curative surgical resection due to tumor stage or operative risks, (6) older than 18 years of age, and (7) ability to provide informed consent. The exclusion criteria were (1) presence of any contraindication to ERCP, (2) previous biliary SEMS placement, surgery, or drainage procedure, and (3) active hepatitis or other hepatic diseases that may cause jaundice. Our institutional review board approved this study. Written informed consent was obtained from all enrolled patients. This study was registered on the UMIN Clinical Trial Registry (UMIN000012734).
Figure 1. A, The newly developed self-expandable metal stent with an antireflux valve (*). B, An 8F modified delivery system having a “bumper” for safe deployment and prevention of creasing of the valve.
designed with an ARV (Fig. 1A). The ARV metal stent (ARVMS) is 10 mm in diameter, and the metal portion is available in 60-, 70-, 80-, 90-, and 100-mm total lengths. The length of the valve portion is 20 mm. The stent is made of nitinol wire and consists of a bare portion with a flared proximal end of 7 mm, a covered body portion, and an antireflux valve portion at the distal end. The antireflux stent consists of an SEMS partially covered by an e-polytetrafluoroethylene (ePTFE) membrane that extends as a pliable tube beyond the distal end of the stent creating a windsock-type ARV. The stent is contained in a standard 8F, pull-back delivery system. The system was slightly modified to ensure safe deployment of the valve and to prevent it from creasing. Attached to the distal core of the delivery system is a novel expandable metal stent– like “bumper” consisting of 20-mm segment of an 8-mm nitinol stent (Fig. 1B). This is situated within the valve and expands and straightens during deployment. A purple marker on the outside of the stent indicates the transition point from the distal stent end to the valve. During deployment of the stent, the purple line is 5 to 10 mm below the papilla in the duodenal lumen, and release of the distal stent end and the valve is performed under endoscopic control. After successful placement of the valve in the duodenal lumen, a delivery system with the metal bumper is retracted and removed from the lumen of the valve, leaving the valve correctly positioned within the duodenal lumen (Video 1, available online at www.giejournal.org).
A newly developed SEMS with a windsocktype ARV
Study design
A new SEMS having an ARV (EGIS Biliary Stent M-Valve; S&G Biotech, Seongnam, South Korea) was modified from a cSEMS (EGIS Biliary Stent; S&G Biotect) and specially
This study was conducted as an open-label, randomized trial at a single center. Once informed consent was obtained and if the patients met the inclusion criteria and
2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2015
www.giejournal.org
Lee et al
Effectiveness of a newly designed antireflux valve metal stent
Figure 2. Endoscopic drainage by using a new self-expandable metal stent (SEMS) with an antireflux valve for malignant biliary obstruction in a patient with pancreatic cancer. A, Cholangiogram showing narrowing of the mid- to distal common bile duct with dilation of the proximal bile duct. B, A new SEMS with an antireflux valve was inserted and C, deployed within the bile duct. D, Fully deployed new SEMS. E, Endoscopic view showing a purple marker indicating the beginning of the valve portion and a bumper on the distal end of the delivery system for protection of the antireflux valve in the introducer. F, The valve portion of fully deployed new SEMS on the duodenum.
demonstrated none of the exclusion criteria, they were assigned randomly to receive an ARVMS (ARVMS group) or a conventional partially cSEMS (cSEMS group) at an allocation ratio of 1:1. Blocked randomization (block size 4) was performed by drawing sealed opaque envelopes. www.giejournal.org
Each envelope contained a card marked with a stent type, and the randomization process was performed by 1 of the participating endoscopists in the ERCP room; after the guidewire was placed correctly and the stricture was passed, the stent was then inserted immediately. Blinding Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 3
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
Figure 3. Barium meal examination for evaluation of duodenobiliary reflux. A, Complete reflux. B, Partial reflux. C, No reflux.
of the endoscopists to stent type was not attempted. The radiologists conducting the interpretation of barium study were blinded to stent assignment.
Endoscopic interventions All procedures were performed by using standardized protocols by 2 experienced investigators using a standard duodenoscope (TJF-260V; Olympus, Tokyo, Japan). An endoscopic sphincterotomy was performed to facilitate stent insertion in all patients. Standard techniques were used to cannulate the biliary tract, and contrast was injected to identify the location and length of the stricture. For patients randomly assigned to the cSEMS group, a conventional partially cSEMS (EGIS biliary stent; S&G Biotect) was used. The new SEMS with an ARV (EGIS Biliary Stent M-Valve; S&G Biotech) was placed in patients assigned to the ARVMS group. The length of the stents was determined according to the length of the bile duct stricture. Positioning of the stent in the center of the stricture was encouraged, and a 5- to 10-mm length of the SEMS should be visible in the duodenum. A stent was inserted 10 to 20 mm proximal to the end of the stricture (Fig. 2).
Evaluation of duodenobiliary reflux Two or 3 days after endoscopic SEMS insertion, a barium meal examination was performed by 2 experienced radiologists in all patients to evaluate the reflux of barium within the SEMS and intrahepatic bile ducts. The duodenobiliary reflux was classified as complete (Fig. 3A), partial (Fig. 3B), or absent (Fig. 3C). Complete reflux was defined as reflux of barium above the proximal end of the SEMS, and partial reflux was defined as a reflux of barium under the proximal end of the SEMS (Video 2, available online at www.giejournal.org).
Follow-up and event definitions To confirm a successful drainage procedure, liver function tests were performed before and 2 to 7 days after stent 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2015
insertion. The patients were scheduled to visit the hospital after 1 month, and liver function tests and a physical examination were performed. Then, clinical and liver function follow-up was conducted at 1-month intervals with a hospital visit. Patients lost to follow-up were analyzed on an intention-to-treat basis and were censored at the time of their last follow-up interview. Technical success was defined as successful stent insertion, based on radiological confirmation on ERCP that the stent was in an appropriate position. Clinical success was defined as at least a 50% decrease in bilirubin levels during the first 7 days after stent insertion. All procedure-related adverse events were recorded according to normal practice.10 Early adverse events were defined as those that occurred during the procedure or within 30 days of stent insertion, and late adverse events were those occurring after this period. Stent dysfunction was considered to occur if the patient had obstructive jaundice and a finding of biliary dilation on imaging studies (CT, magnetic resonance imaging, or US) with or without cholangitis. Endoscopic reintervention was then arranged to confirm and assess the cause of stent dysfunction. If the patients showed recurrent jaundice with high-grade fever, urgent reintervention was also performed, although biliary dilation was not confirmed by imaging studies. Tumor overgrowth was defined as stent occlusion by a tumoral mass over the stent. Tumor ingrowth was defined as stent occlusion by the tumor inside the stent. Migration, proximally or distally, was defined as SEMS movement over the MBO from the original post-placement position, as documented by the final spot radiograph obtained at the initial ERCP.
Outcome measurements The primary outcome was stent patency duration, defined as the time of endoscopic stent deployment until stent dysfunction requiring reintervention, death, loss to follow-up, or patients alive at the end of the study with no occlusion. Secondary outcomes were the rates of www.giejournal.org
Lee et al
Effectiveness of a newly designed antireflux valve metal stent
Assessed for eligibility (n=141) Excluded (n=64) – Surgical treatment: 23 – Multiple stricture: 6 – Close to hilum: 13 – Previous Billroth II operation: 3 – Refused to participate: 19 Randomized (n= 77) Intention-to-treat follow-up
ARVMS (n=39)
cSEMS (n=38)
Lost to follow up (n=3)
Lost to follow up (n=4)
Stent dysfunction (n=7) Died without stent dysfunction (n=26) Alive without stent dysfunction (n=3)
Stent dysfunction (n=14) Died without stent dysfunction (n=19) Alive without stent dysfunction (n=1) Analysis
Figure 4. Flowchart showing the results of inclusion, randomization, and follow-up of the patients. ARVMS, antireflux valve with metal stent; cSEMS, covered self-expandable metal stent.
technical and clinical success, the duodenobiliary reflux on barium meal examination, risk factors causing stent dysfunction, overall patient survival, and adverse events. Patient survival was measured from endoscopic stent deployment to death or the day of last follow-up.
with a multivariate analysis. All statistical analyses were performed using the SPSS software (version 18.0; SPSS Inc, Chicago, Ill). P values <.05 were considered to indicate statistical significance.
RESULTS Statistical analysis This study was intended as a pilot study; therefore, no formal power calculation was performed. The stent patency time and overall patient survival were evaluated by using the Kaplan-Meier method and compared by using the log-rank test. Categorical parameters were expressed as frequencies and proportions and compared by using the c2 test or Fisher exact test. Continuous data are expressed as medians with ranges or SDs and compared by using the Mann-Whitney U test. Univariate and multivariate analyses were performed to identify risk factors for stent dysfunction by using logistic regression analysis. The model included age (<60 vs 60 years), sex, previous plastic stent insertion, stent type (cSEMS vs ARVMS), length of stent, tumor etiology, duodenobiliary reflux on barium meal examination (no or partial vs complete), and adjuvant therapy. For each factor considered to have a potential association with stent dysfunction, the odds ratio (OR) with its 95% confidence interval (CI) was calculated. Factors with P values <.10 on univariate analysis were further analyzed www.giejournal.org
Between January 2013 and September 2014, 141 patients were screened; 64 were excluded for the following reasons: surgical treatment (n Z 23), close to the hilum (n Z 13), multiple strictures of the bile duct (n Z 6), previous Billroth II operation (n Z 3), and refusal to participate (n Z 19). The remaining 77 patients constituted the study cohort; 39 patients were randomly assigned to the ARVMS group and 38 patients to the cSEMS group. Loss to follow-up in the ARVMS and cSEMS groups occurred in 3 patients (7.7%) and 4 patients (10.5%), respectively (P Z .711) (Fig. 4). Baseline characteristics were not significantly different between the ARVMS and cSEMS groups. The causes of MBO were pancreatic cancer (n Z 45), cholangiocarcinoma (n Z 21), ampullary cancer (n Z 9), and metastatic nodes (n Z 2). There was no significant difference between the groups in terms of the underlying diagnoses. In total, 52 patients underwent adjuvant therapy after stent insertion: 29 (74.4%) in the ARVMS group and 23 (60.5%) in the cSEMS group (Table 1). Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 5
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
Stent placement and early clinical outcomes Stent placement was technically successful in all patients. The clinical success rates were not statistically significantly different between the ARVMS and cSEMS groups (97.4% vs 97.4%, P Z 1.00). No stent dysfunction was observed in either group within 30 days after stent placement. Early adverse events were observed in 3 patients (2 cases of mild pancreatitis and 1 case of cholangitis) and 6 patients (4 cases of mild pancreatitis and 2 cases of cholangitis) in the ARVMS group and cSEMS groups, respectively. The early adverse events in the ARVMS group were all considered to be procedure-related events. In 1 patient in the cSEMS group, cholangitis was diagnosed 25 days after stent placement; however, stent dysfunction was not observed by ERCP. There was no significant difference in 30-day mortality between the 2 groups (Table 2).
Stent patency and patient survival The median follow-up durations in the ARVMS and cSEMS groups were 156 days (range 20-507 days) and 189.5 days (range 25-610 days), respectively (P Z .318). The cumulative duration of stent patency in the ARVMS group was significantly longer than that in the cSEMS group (median SD, 407 92 vs 220 37 days; P Z .013, log-rank test) (Fig. 6). During follow-up, the occurrence of stent dysfunction did not differ between the ARVMS and cSEMS groups (P Z .077). Additionally, the detailed causes of stent dysfunction, including sludge occlusion (P Z .224), did not differ between the groups. In the ARVMS group, in total, 3 late adverse events (2 cases of mild pancreatitis and 1 case of hemobilia) were observed, with no significant difference from the cSEMS group (P Z 1.00). The 1 hemobilia case of unknown cause in the ARVMS group occurred 35 days after stent insertion and was treated by reinsertion of a metal stent during ERCP. There was no significant difference in overall patient survival, with the median SD survival being 289 33 days in the ARVMS group and 340 36 days in the cSEMS group (P Z .661, log-rank test) (Table 3).
Risk factors for stent dysfunction Univariate and multivariate analyses of the risk factors for stent dysfunction were performed. Univariate analysis No.
-
Age, y Sex, male/female, no. Total bilirubin, mg/dL Previous plastic stent, no. (%) Duration of plastic stent, days Length of stricture, mm
: 2015
ARVMS (n [ 39)
cSEMS (n [ 38)
P value
74 (46–91)
73 (44–101)
.672
22/17
15/23
.173
4.4 (0.3–20.1) 4.1 (1.4–35.7) 28 (73.7)
1.00
6 (3–150)
7 (3–152)
.317
19 (10–44)
23 (7–700)
.296 .462
6
1 (2.6)
4 (10.5)
7
37 (94.9)
29 (76.3)
1 (2.6)
5 (15.1)
18 (46.2)
16 (42.1)
Tumor etiology, no. (%)
.820 .223
Pancreatic cancer
22 (56.4)
23 (60.5)
Cholangiocarcinoma
10 (25.6)
11 (28.9)
Ampullary cancer
5 (12.8)
4 (10.5)
Metastatic nodes
2 (5.2)
0
29 (74.4)
23 (60.5)
Adjuvant therapy, no. (%)
.790
29 (74.4)
Length of stent, cm, no. (%)
Distant metastasis, no. (%)
A barium meal examination was performed in all 77 patients. In the cSEMS group, the free reflux of barium occurred through the SEMSs in all patients. Overall reflux (complete or partial) of barium was seen in 3 patients of the ARVMS group, which was significantly different from the cSEMS group (7.7% vs 100%, P < .001). Complete reflux in the cSEMS group was also significantly greater than in the ARVMS group (2.6% vs 55.3%, P < .001) (Fig. 5).
-,
Type of stent
8
Duodenobiliary reflux on barium meal examination
6 GASTROINTESTINAL ENDOSCOPY Volume
TABLE 1. Baseline characteristics of the patients in the ARVMS and cSEMS groups
.230
Quantitative data are expressed as median (range). ARVMS, Antireflux valve metal stent; cSEMS, covered self-expandable metal stent.
TABLE 2. Early technical and clinical outcomes of patients in the ARVMS and cSEMS groups Type of stent ARVMS (n [ 39) Technical success, no.
cSEMS (n [ 38)
P value
39
38
d
Clinical success, no. (%)
38 (97.4)
37 (97.4)
1.00
Adverse events, no. (%)
3 (7.7)
6 (15.8)
.310
Mild pancreatitis
2
4
.431
Cholangitis 30-day mortality, no. (%)
1
2
.615
2 (5.1)
1 (2.6)
1.00
ARVMS, Antireflux valve metal stent; cSEMS, covered self-expandable metal stent.
revealed that complete duodenobiliary reflux on barium meal examination (odds ratio [OR], 6.13; P Z .001) and ampullary cancer (OR, 9.13; P Z .006) were risk factors for stent dysfunction. ARVMS (OR, 0.38; P Z.067) and adjuvant therapy (OR, Z 0.53; P Z .236) were not significant factors on univariate analyses. On multivariate analyses, complete duodenobiliary reflux (OR, 5.7; P Z .004) and ampullary cancer (OR, 8.98; P Z .012) were identified as independent risk factors for stent dysfunction. A post hoc power analysis based on a log-rank test yielded a power of approximately 75% (1-b Z 0.7457) to www.giejournal.org
Lee et al
P < .001 100%
P < .001 ARVMS cSEMS
80
P = .013 80
60 55.3% 40 20 0
censored
100
7.7% Reflux
2.6% Complete reflux
Figure 5. Duodenobiliary reflux of patients in the ARVMS and cSEMS groups by using a barium meal examination. ARVMS, antireflux valve with metal stent; cSEMS, covered self-expandable metal stent.
detect estimated differences in median survival time of stent patency, which was 407 days for the 39 patients in the ARVMS group and 220 days for the 38 patients in the cSEMS group under the assumption that the survival function follows the exponential distribution. This power was computed by using STATA/SE 12.0 version with the stpower log-rank command (StataCorp, College Station, Tex).
DISCUSSION In patients with unresectable distal MBO, SEMS placement is regarded as the standard treatment, with a larger luminal diameter and longer patency than those of plastic stents.10,11 However, most SEMSs for distal MBO should be placed at the distal end of the duodenum. As a result, sphincter function is lost, allowing free duodenobiliary reflux of duodenal residue.6 In previous studies, the initial event leading to stent occlusion has been suggested to be the formation of biofilm by the adherence of host proteins and bacteria to the inner wall of the stent.12,13 This biofilm presumably protects the bacteria against antimicrobial agents. Bacterial glucuronidases and phospholipases act on bile components, leading to biliary sludge formation composed of bacterial products, calcium bilirubinate, and calcium fatty acid soaps. In addition, a study recently revealed that duodenobiliary reflux confirmed by using barium meal examination was an independent risk factor associated with CBD stone recurrence.14 Thus, reducing duodenobiliary reflux has been an attractive idea for prolonging the duration of stent patency, and several newly designed stents with antireflux function have been developed. Most of them showed efficacy and safety for palliation of jaundice caused by distal MBO, with a remarkably low risk of cholangitis and with improved stent patency. However, only small numbers of patients were enrolled in studies with nonrandomized www.giejournal.org
Cumulative patency, %
100
Effectiveness of a newly designed antireflux valve metal stent
ARVMS (n=39) 60
40
cSEMS (n=38)
20
0 0
100
200
300
400
500
600
Time, days Figure 6. Kaplan-Meier curve showing the cumulative stent patency of the stent in the antireflux valve metal stent (ARVMS) group and covered self-expandable metal stent (cSEMS) group (P Z .013, log-rank test).
designs.8,15,16 Recently, Hu et al9 reported results of a prospective, randomized, controlled trial in which a new partially covered ARMS and a standard uncovered SEMS were used. In a total of 112 patients with MBO, fewer patients experienced cholangitis in the partially covered ARMS group than in the uncovered SEMS group (10 vs 21 patients, P Z .035), and the frequency of episodes was lower (P Z .022). Additionally, the median stent patency SD was significantly longer in the partially covered ARMS group (13.0 3.4 vs 10.0 1.2 months, P Z .044). Although there was a limitation in that they compared stent patency of a partially covered ARMS and an uncovered SEMS, the study showed that an antireflux function could improve the duration of SEMS patency and reduce the occurrence of cholangitis after insertion of ARMS. For this trial, we developed a newly designed SEMS with an ARV. This stent has a relatively long valve length (20 mm) on the distal end of the metal stent. The valve is similar in appearance to a windsock and is made of e-polytetrafluoroethylene. Because 1 membrane covers the body and valve portions on the distal end of the stent, the distal end of the metal stent during stent deployment cannot be seen by the endoscope. However, there is a purple marker on the covered membrane at the distal end of the metal stent to indicate the start of the valve portion, such that the position of the distal end can be determined during stent deployment without difficulty. In this study, technical success of the procedure by using the ARVMS was achieved in all patients. If the endoscopic revision of the ARVMS was required due to stent dysfunction, stent Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 7
Effectiveness of a newly designed antireflux valve metal stent
Lee et al
TABLE 3. Long-term clinical outcomes of patients in the ARVMS and cSEMS groups Type of stent ARVMS (n [ 39) Follow-up duration, median (range), days
cSEMS (n [ 38)
156 (20–507) 189.5 (25–610)
P value .318
Stent patency, median (SD), days
407 (92)
220 (37)
.013
Stent dysfunction, no. (%)
7 (17.9)
14 (36.8)
.077
Sludge occlusion
4 (10.3)
8 (21.1)
.224
Stent migration
3 (7.7)
3 (7.9)
1.00
Valve dysfunction
2 (5.1)
0
d
Tumor ingrowth
1 (2.6)
3 (7.9)
.358
0
2 (5.3)
.240
Causes of stent dysfunction, no. (%)*
Tumor overgrowth Hemobilia
1 (2.6)
0
1.00
Unknown
1 (2.6)
1 (2.6)
1.00
Adverse event, no. (%)
3 (7.7)
3 (7.9)
1.00
Mild pancreatitis
2 (5.1)
2 (5.3)
0
1 (2.6)
Cholecystitis Hemorrhage Survival, median (SD), days
1 (2.6)
0
289 (33)
340 (36)
.661
ARVMS, Antireflux valve metal stent; cSEMS, covered self-expandable metal stent. *Stent dysfunction may have occurred for more than 1 reason; thus, the total does not add up to the number of occurrences of stent dysfunction in each group.
removal was tried by using snare at first. When stent removal failed and/or insertion of the guidewire inside the ARVMS was needed, a guidewire could be inserted through the distal end of valve portion or a hole made on the valve by using a needle-knife. In this study, the impact of the antireflux function of the ARVMS was evaluated by barium meal examination. Reflux of barium in the ARVMS group was seen in only 3 patients, whereas all patients in the cSEMS group showed free reflux of barium (7.7% vs 100%, respectively; P < .001). Complete reflux in the cSEMS group was also significantly greater than that in the ARVMS group (2.6% vs 55.3%, respectively; P < .001). Additionally, complete duodenobiliary reflux was an independent risk factor for stent dysfunction on multivariate analysis (OR, 5.7; P Z .004). The ARVMS showed significantly longer stent patency than that of cSEMS (median SD, 407 92 vs 220 37 days, P Z .013) and was safe to use. However, our data did not show significant evidence of the ability to reduce stent dysfunction on multivariate analysis (OR, 0.57; P Z .859). Nevertheless, the efficacy of ARVMS for reducing stent dysfunction was suggested because the ARVMS showed reduced complete duodenobiliary reflux, which is an independent risk factor for stent dysfunction. Ampullary cancer (OR, 8.98; P Z .012) was also an inde8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2015
pendent risk factor for stent dysfunction. In patients with ampullary cancer, stent migration was observed in 3 of 4 (75%) cases of stent dysfunction. It seemed to be caused by a short stricture, which is limited to the short-segment ampullary lesion, which was insufficient to anchor the stent in situ.17 During the follow-up period, cholangitis due to valve dysfunction developed in 2 patients 402 and 81 days after stent insertion, respectively. In those patients, increased hardness of the valve was observed after stent removal. However, it was unclear what affected the hardness of the ARV: the material of the ARV itself, the types of food content, and/or the duodenal environment of the patients. Further studies and evaluations are required to address this. Our study had several limitations. First, there were no significant differences between the ARVMS group and the cSEMS group in terms of frequency of SEMS dysfunction, patient survival, and patients lost to follow-up. Nevertheless, the cumulative duration of stent patency in the ARVMS group was significantly longer than that in the cSEMS group (median SD, 407 92 vs 220 37 days, P Z .013). This discrepancy is considered to be caused by the limitation that it was a pilot study that began without a formal sample size calculation. However, although there was no significant difference, the stent dysfunction in the cSEMS group occurred about twice as often as in the ARVMS group (36.8% vs 17.9%, P Z .077). The duration until the development of stent dysfunction in the ARVMS group tended to occur later than that in the cSEMS group, but there was no significant difference (median, 181 vs 106 days; P Z .172). The significantly longer stent patency of ARVMS is already a clinically important result because ARVMS may be able to reduce the need for reintervention in patients with palliatively treated MBO. Therefore, further studies involving larger cohorts are needed to confirm and strengthen our results. Second, the current results were obtained by 2 experienced endoscopists using endoscopic biliary drainage with a new ARVMS. Although the technical success rate of the procedure was 100% in our study, it remains to be determined whether the promising results of this study are reproducible by less-experienced endoscopists. A third limitation was the nonblinding of the physicians to the stents, both before and after the procedure. However, the radiologists and patients were blinded to the stents. The final limitations are the relatively short follow-up period and the 9.1% of patients lost to follow-up. These might affect the relatively low incidence of stent dysfunction in our study compared with previous studies. In conclusion, our prospective, randomized study performed in 77 patients with unresectable distal MBO showed that the newly developed ARVMS has a significantly longer duration of stent patency and comparable safety compared with cSEMS. Additionally, the duodenobiliary reflux related to stent dysfunction can be effectively prevented by use of the ARVMS. However, the survival of www.giejournal.org
Lee et al
patients was not different between ARVMS and cSEMS placement in these patients. Further randomized, controlled trials involving larger cohorts are required to confirm the benefits of SEMSs with antireflux functions. ACKNOWLEDGMENTS We thank Dr Sung Kwon Kang, Mr Young Jae Lee, and Mr Seung Hwan Jegal of S&G Biotech, Seongnam, South Korea, for the development of the new SEMS with an antireflux valve. We thank Professor Joo Ha Hwang, Seattle, Washington, USA, Dr Hans-Ulrich Laasch, Manchester, UK, and Mr Derek W. Edwards for proofreading our manuscript. We thank A Ri Song, RN, Song Ah Jeong, RN, and Sun Yeong Moon, RN, and the rest of the nursing staff for assistance with the procedures. Professor Jong Ho Moon is the inventor of the EGIS Biliary Stent M-Valve. This work was supported, in part, by the SoonChunHyang University Research Fund.
REFERENCES 1. Davids PH, Groen AK, Rauws EA, et al. Randomised trial of selfexpanding metal stents versus polyethylene stents for distal malignant biliary obstruction. Lancet 1992;340:1488-92. 2. Prat F, Chapat O, Ducot B, et al. A randomized trial of endoscopic drainage methods for inoperable malignant strictures of the common bile duct. Gastrointest Endosc 1998;47:1-7. 3. Soderlund C, Linder S. Covered metal versus plastic stents for malignant common bile duct stenosis: a prospective, randomized, controlled trial. Gastrointest Endosc 2006;63:986-95. 4. Kaassis M, Boyer J, Dumas R, et al. Plastic or metal stents for malignant stricture of the common bile duct? Results of a randomized prospective study. Gastrointest Endosc 2003;57:178-82.
www.giejournal.org
Effectiveness of a newly designed antireflux valve metal stent 5. Loew BJ, Howell DA, Sanders MK, et al. Comparative performance of uncoated, self-expanding metal biliary stents of different designs in 2 diameters: final results of an international multicenter, randomized, controlled trial. Gastrointest Endosc 2009;70:445-53. 6. Misra SP, Dwivedi M. Reflux of duodenal contents and cholangitis in patients undergoing self-expanding metal stent placement. Gastrointest Endosc 2009;70:317-21. 7. Kim DU, Kwon CI, Kang DH, et al. New antireflux self-expandable metal stent for malignant lower biliary obstruction: in vitro and in vivo preliminary study. Dig Endosc 2013;25:60-6. 8. Hamada T, Isayama H, Nakai Y, et al. Novel antireflux covered metal stent for recurrent occlusion of biliary metal stents: a pilot study. Dig Endosc 2014;26:264-9. 9. Hu B, Wang TT, Wu J, et al. Antireflux stents to reduce the risk of cholangitis in patients with malignant biliary strictures: a randomized trial. Endoscopy 2014;46:120-6. 10. Yeoh KG, Zimmerman MJ, Cunningham JT, et al. Comparative costs of metal versus plastic biliary stent strategies for malignant obstructive jaundice by decision analysis. Gastrointest Endosc 1999;49:466-71. 11. Arguedas MR, Heudebert GH, Stinnett AA, et al. Biliary stents in malignant obstructive jaundice due to pancreatic carcinoma: a costeffectiveness analysis. Am J Gastroenterol 2002;97:898-904. 12. Leung JW, Liu Y, Chan RC, et al. Early attachment of anaerobic bacteria may play an important role in biliary stent blockage. Gastrointest Endosc 2000;52:725-9. 13. Leung JW, Liu YL, Chan RC, et al. Effects of adherence factors and human bile on bacterial attachment and biliary stent blockage: an in vitro study. Gastrointest Endosc 2002;56:72-7. 14. Zhang R, Luo H, Pan Y, et al. Rate of duodenal-biliary reflux increases in patients with recurrent common bile duct stones: evidence from barium meal examination. Gastrointest Endosc 2015;82:660-5. 15. Lee KJ, Chung MJ, Park JY, et al. Clinical advantages of a metal stent with an S-shaped anti-reflux valve in malignant biliary obstruction. Dig Endosc 2013;25:308-12. 16. Hu B, Wang TT, Shi ZM, et al. A novel antireflux metal stent for the palliation of biliary malignancies: a pilot feasibility study (with video). Gastrointest Endosc 2011;73:143-8. 17. Park SB, Kim HW, Kang DH, et al. Metallic or plastic stent for bile duct obstruction in ampullary cancer? Dig Dis Sci 2012;57:786-90.
Volume
-,
No.
-
: 2015 GASTROINTESTINAL ENDOSCOPY 9