Introduction
Endoscopic ultrasound guided gastro-enterostomy (EUS-GE) is a novel procedure, recently
described for patients with gastric outlet obstruction (GOO) [1]
[2]. The reported technical and clinical success of this procedure is high [3]
[4]. Ascites is often observed in GOO patients due to progression of malignancy. EUS-GE
in presence of ascites carries theoretical risk of adverse events, especially perforation,
where the fallback option is often surgery. The feasibility and safety of EUS-GE in
patients with GOO with ascites has not been objectively addressed in literature. There
are several described methods of doing EUS-GE, whose preference is based on operator
experience and availability of devices [5]. The technical challenges and adverse events of EUS-GE in presence of ascites are
not well known. The objective of this study was to evaluate the feasibility and safety
of EUS-GE in patients with gastric outlet obstruction and ascites.
Patients and methods
Consecutive patients with symptomatic GOO who underwent EUS GE between January 2019
to March 2021 constituted the study population. The study was conducted at a tertiary
care hospital and approved by institutional ethical committee. Demographic details,
aetiology of GOO and presence of ascites during evaluation were recorded. Patients
with ascites were further evaluated by ascitic fluid analysis including the cytology
to establish the cause of ascites. In patients with malignant ascites, distal intestinal
or multi-level obstruction was excluded by contrast imaging. Ascites was managed according
to the standard of care. Patients with high SAAG (Serum Ascites Albumin Gradient)
ascites, were managed by diuretics and intravenous albumin when required. Patients
with malignant ascites were considered for paracentesis whenever they were symptomatic.
In patients with significant ascites, either large volume paracentesis or percutaneous
catheter (12 Fr, 25 cm drainage catheter, Cook Medical, USA) placement for drainage
was considered. Paracentesis was performed on the same day before the procedure. Catheter
drainage in patients with rapidly filling ascites was placed a day before the procedure.
Procedure
EUS-GE was performed either by double balloon assisted (EPASS technique) or free-hand
technique. All patients received intravenous antibiotics prior to the procedure. The
procedure was performed under anaesthetist supervised sedation without endotracheal
intubation.
EPASS technique
A clean stomach devoid of any food residue was ensured by prior endoscopy or lavage
through nasogastric tube. The patient was placed in either prone or supine position.
The steps in the procedure were as follows:
-
A standard gastroscope was passed through the over-tube of short enteroscope. The
whole assembly was inserted into the stomach as one unit. The tip of the over-tube
was kept close to the stricture (for gastric cancer) or in duodenal cap (for duodenal
stricture).
-
A long (450 cm) guidewire was passed through catheter and negotiated across the stricture
deep inside the proximal jejunum.
-
The gastroscope was removed, leaving the over-tube and guidewire in place. The purpose
of over-tube was to prevent looping of the guidewire and/or the flexible soft oro-enteric
tube having double balloon (EPASS) in stomach.
-
The EPASS balloon catheter was passed under fluoroscopy over the guidewire through
the over-tube, across the luminal stricture, deep into proximal jejunum. The radio-opaque
markers located at the tip and at the two balloons help in its accurate positioning
([Fig. 1a]).
-
The two balloons (front and back) were then distended with about 40cc of diluted contrast
to appose and anchor the small bowel loops ([Fig. 1a], [Fig. 2a]).
-
The intervening proximal small bowel (jejunal) segment between the 2 balloons was
filled and distended with 100–120 ml solution consisting of saline, contrast and a
colouring agent (indigo-carmine) ([Fig. 1b], [Fig. 2b]).
Fig. 1 EUS-guided gastroenterostomy using the EPASS technique. a Fluoroscopic image showing the EPASS catheter passing across the stricture with inflated
both proximal and distal balloons and the previously placed blocked enteral SEMS and
biliary SEMS. b Fluoroscopic image showing EUS scope targeting vertical limb of contrast filled jejunal
segment between the two balloons. c EUS image showing the distended jejunal segment. EPASS catheter can be seen in the
lumen. d Fluoroscopic image showing EUS scope and puncture of jejunal segment with Hot Axios
delivery system. e Fluoroscopic image showing deployed LAMS between jejunum and stomach. f Endoscopic image showing deployed LAMS draining coloured contrast fluid. EPASS, EUS-guided
double-balloon-occluded gastrojejunostomy bypass; SEMS, self expandable metal stent;
LAMS, lumen apposing metal stent.
Fig. 2 EUS-guided gastroenterostomy in a patient with ascites. a Fluoroscopic image showing EUS scope and EPASS catheter passing across the stricture
with inflated both proximal and distal balloons and the percutaneous catheter placed
to drain the ascites. b Fluoroscopic view of EUS scope and contrast-injected jejunal segment between the
two balloon. c Fluoroscopic image showing deployed successfully deployed LAMS. EPASS, EUS-guided
double-balloon-occluded gastrojejunostomy bypass; LAMS, lumen apposing metal stent.
The double balloon (EPASS) catheter has four dedicated channels (one for distal balloon,
one for proximal balloon, one for the guidewire, and one for the filling the bowel
segment between the two balloons. The aim was to place the proximal balloon, identified
by respective radio-opaque marker, near DJ flexure and distal balloon beyond duodeno-jejunal
flexure.
With the free-hand technique, either a standard feeding enteral tube (12 Fr, 120 cm,
Fresennius Kabi, Germany) or a biliary catheter (6 Fr, 250 cm, Meditech, India) was
passed through mouth over the guidewire across the luminal stricture segment into
proximal jejunum. The target jejunal loop was distended rapidly with normal saline
mixed with contrast and / or colouring agent under EUS and fluoroscopy vision. In
cases with Afferent loop obstruction, guidewire was used after puncture with 19 g
needle of the adherent distended afferent loop.
A linear echoendoscope (GF-UCT180; Olympus) was passed into the stomach at the time
of filling the target jejunum lumen. The selectively distended jejunal segment, preferably
the descending vertical limb, was identified under EUS and cross-checked on fluoroscopy
before puncture ([Fig. 1c]). EUS-GE was performed using a 20- or 15-mm lumen apposing metal stent (LAMS) (HOT
Axios, Boston Scientific, Marlborough, Massachusetts). The distal flange of the LAMS
was deployed first under EUS and fluoroscopic guidance ([Fig. 1d], [Fig. 1e], [Fig. 2c]). The proximal flange was then deployed either within the channel or under endoscopic
view. The appearance of blue colouring agent in gastric lumen confirms the accurate
placement of LAMS ([Fig. 1f]). Often the pink mucosa of the small bowel was visible through the stent. Balloon
dilation of the stent was not performed. The LAMS gradually expands over the next
few days.
Patients were kept nil by mouth on the day of the procedure. A screening oral contrast
study was performed at 24 to 48 hours to confirm the path of gastric emptying. Oral
liquids were allowed after the contrast study, followed by low-residue feeds on subsequent
day. Antibiotics were continued for 5 days in patients with ascites. Patients were
discharged when an oral semisolid diet was well tolerated and if they were clinically
stable. Rates of technical success, clinical success, and adverse events (AEs) were
evaluated. The outcomes were compared between patients with and without ascites. All
patients were followed every month, till death or last contact (Physical or by Telephone).
Definitions
Technical success was defined as successful deployment of LAMS between the proximal
small bowel (jejunum or distal duodenum) and the stomach. Clinical success was defined
as relief of symptoms of GOO and ability to tolerate diet.
Ascites was defined as mild, moderate, or large, according to definitions given by
International ascites club [6]. Ascites was mild if it was detectable only on ultrasound. Moderate ascites caused
moderate symmetrical distension of the abdomen. Large ascites caused marked abdominal
distension
Statistical analysis
Categorical and continuous data were expressed as proportion and mean (standard deviation
[SD]), respectively. All categorical variables were compared using a chi-square test
or Fisher exact test wherever required. Continuous data were compared using an independent
t test. The survival of patients with or without ascites was compared with Kaplan-Meier
analysis and log-rank test. P < 0.05 was considered statistically significant. Statistical analysis was done using
SPSS version 26 (IBM Corp., Armonk, New York, New York, United States)
Results
A total of 31 patients underwent EUS-GE. Approximately two-thirds (20 of 31) were
male. The mean age was 61.65 years ± 10.65 (SD). The etiology of GOO was malignancy
in 29 patients (93.5 %), two patients had benign causes (duodenal obstruction due
to chronic pancreatitis in one, retroperitoneal fibrosis in one patient). Among these,
prior enteral SEMS were in situ and blocked in three patients due to tumor ingrowth.
Prior biliary SEMS were present in six patients (ERCP 4, EUS guided choledocho-duodenostomy
2). Three patients underwent EUS-guided hepatico-gastrostomy (HGS) for biliary obstruction
following EUS-GE, at the same session.
Overall, ascites was observed in 12 of 31 patients (38.7 %) prior to EUS-GE. Among
the patients with malignant GOO, ascites was present in 41.3 % (12 of 29). The mean
age, sex, and etiology of GOO were not significantly different between patients with
or without ascites ( [Table 1]). Ascitic fluid evaluation revealed cytology positive for malignancy in four patients.
The remaining eight patients with negative cytology had high a high serum ascites
albumen gradient (chronic liver disease in 1, portal vein thrombosis in 1, liver metastasis
in 2 and hypo-albuminemia in 4 ). The ascites was mild to moderate in five patients.
The remaining seven patients had large ascites, three of whom underwent therapeutic
paracentesis once before the procedure. A per-cutaneous catheter was placed in the
other four patients. The catheter was removed in one patient 3 days after EUS-GE and
in the remaining three patients, catheters were left in situ.
Table 1
Comparison of clinical profile and outcomes between patients of gastric outlet obstruction
with and without ascites.
|
GOO with ascites
N = 12 (38.7 %)
|
GOO without
ascites N = 19 (61.3 %)
|
P value
|
|
Age in years, mean ± S.D.
|
62.2 ± 12.5
|
61.2 ± 9.6
|
0.806
|
|
Sex (male)
|
8 (66.6 %%)
|
12 (63.1 %)
|
0.842
|
|
Etiology of GOO, n (%)
|
|
|
0.481
|
|
Pancreatic cancer (15)
|
6 (50 %)
|
9 (47.3 %)
|
|
|
Periampullary cancer (2)
|
2 (16.6 %)
|
0 (0 %)
|
|
|
Gall bladder cancer (5)
|
2 (16.6 %)
|
3 (15.7 %)
|
|
|
Duodenal cancer (1)
|
0 (0 %)
|
1 (5.2 %)
|
|
|
Gastric cancer (4)
|
1 (8.3 %)
|
3 (15.7 %)
|
|
|
Postoperative recurrence in GJ anastomotic site (2)
|
1 (8.3 %)
|
1 (5.2 %)
|
|
|
Benign (2)
|
0 (0 %)
|
2 (10.5 %)
|
|
|
Procedural technique
|
|
|
0.546
|
|
EPASS
|
11 (91.6 %)
|
16 (84.2 %)
|
|
|
Free hand
|
1 (8.3 %)
|
3 (15.7 %)
|
|
|
Technical success
|
11 (91.6 %)
|
17 (89.4 %)
|
0.841
|
|
Procedure time in minutes (mean ± SD)
|
32 ± 12.5
|
31.6 ± 6.6
|
0.968
|
|
Clinical success
|
10 (83.3 %)
|
17 (89.4 %)
|
0.619
|
|
Adverse events
|
0 (0 %)
|
2 (10.5 %)
|
0.245
|
|
Median survival (days)
|
36
|
290
|
0.001
|
Technical success
EPASS technique was used in 27 patients (87 %. Four (13 %) underwent the free-hand
technique. Overall, EUS-GE was successfully performed in 28 of 31 patients (90.3 %).
The mean procedure time was 31.88 minutes ± 10.13 (SD). There was no difference in
technical success between patients with or without ascites (91.6 % vs. 89.4 %; P = 0.841). Also, the mean procedure time was not significantly different (32 vs. 31.6
min; P = 0.968).
LAMS placement was unsuccessful in three patients (2 with EPASS and one free-hand).
This occurred in two patients in our early phase of LAMS experience while using the
EPASS technique and both patients had only gastric (without jejunal) puncture. The
gastric defect was then closed with an over-the-scope-clip (OTSC) followed by enteral
SEMS placement in the same session. The third patient underwent LAMS placement using
the free-hand technique and developed colonic perforation requiring surgery.
During EUS-GE with EPASS, unexpected technical challenges were encountered in seven
patients during the procedure. In two patients, there was distal seepage of contrast
saline-colored solution trapped between the two anchoring balloons due to suboptimal
anchoring leading to incomplete occlusion of the lumen by the forward balloon. The
event was identified before the LAMS deployment, although it caused some delay. In
another two patients, even after full inflation, the balloons migrated proximally
due to friction caused by echoendoscope movement, making the procedure difficult.
In another patient, duodenal narrowing was located in the distal duodenum near the
DJ flexure, hence both balloons were positioned beyond the DJ flexure, causing an
atypical location of the target jejunal segment, leading to difficulty in deployment
of LAMS. Malrotation of the duodenum or location of the DJ flexure on the right side
was observed in two patients. This caused awkward positioning of the echoendoscope
at puncture. In spite of the above technical challenges, the EUS-GE was technically
successful in all.
Clinical success
Overall, clinical success was achieved in 27 patients (87 %). One patient with postoperative
recurrence of malignancy in a gastro-jejunostomy anastomotic site had persistent symptoms
after LAMS placement in an obstructed afferent limb. On reevaluation, endoscopy confirmed
efferent loop stoma narrowing due to the recurrence. Hence, an enteral SEMS was placed
into the efferent limb after 2 days and his overall symptoms improved. Clinical success
was similar in patients with and without ascites (83.3 % vs. 89.4 %; P = 0.619).
Adverse events
AEs were observed in two patients (6.4 %). One had colonic perforation requiring surgery
while using the free-hand technique. Another patient with benign duodenal stricture
due to chronic pancreatitis and portal vein thrombosis with peri-gastric collaterals
had gastrointestinal bleeding during LAMS deployment [7]. Gastric side bleeding was controlled with clips and jejunal side bleeding was controlled
by passing the gastroscope through the LAMS and applying the clip over jejunal puncture
side. Both these AEs occured in patients without ascites; however, the difference
was not statistically significant (0 % vs. 10.5 %; P = 0.245). None of the patient with ascites had infection or peritonitis after the
procedure.
Follow-up
Twenty two of 28 patients (78.5 %) with successful LAMS placement died at a median
follow-up of 66 days (range: 8 to 417 days). One patient with postoperative recurrence
of cancer in an anastomotic site died 8 days after the procedure due to underlying
cardiac disease. Of the two patients with benign GOO, one with retroperitoneal fibrosis
died at 85 days after EUS-GE due to multisystem involvement and renal failure. Overall,
the median survival time after the procedure was 51 days (range: 8 to 417 days). The
median survival time in patients with ascites was 36 days (range: 8 to 180 days),
whereas in patients without ascites, it was 290 days (range: 15 to 417 days). On Kaplan-Meier
analysis, the difference in median survival was statistically significant (36 vs.
290 days; P < 0.001) ([Fig. 3]).
Fig. 3 Kaplan-Meier analysis showing survival time of malignant GOO patients with and without
ascites. GOO, gastric outlet obstruction.
Discussion
Preferred palliative management for malignant GOO is surgical gastrojejunostomy in
suitable patients, which is associated with long-term patency. Presence of ascites
in these patients makes them unsuitable for surgical GJ. Endoscopic enteral stent
placement is an alternate option in surgically ineligible patients because of its
minimal invasive nature [8]. The major drawback of enteral SEMS is tumor ingrowth associated with recurrence
of GOO. EUS-GE is a recent, exciting alternate endoscopic option that appears to be
feasible and safe in expert hands with good technical and clinical success [3]
[4]. Philip and colleagues recently compared EUS GE with enteral SEMS and reported higher
clinical success and lower stent occlusion rates with EUS-GE [9]. The major benefit of EUS-GE over enteral SEMS is that the gastro-enterostomy stent
is physically away from the tumor, and hence, less susceptible to malignant tumor
ingrowth and occlusion. In the current study, we report that EUS-GE can be safely
performed in patients with GOO even in the presence of ascites, after taking reasonable
precautions.
Several techniques have been described for performing EUS-GE, including free-hand,
assisted (balloon, basket), and EPASS [10] double balloon. A recent retrospective
comparative study showed that the direct free-hand and balloon-assisted techniques
have comparable technical and clinical success rates [11]. However, there have been no direct comparative studies of those assisted techniques
with EPASS. In this study, we used EPASS in the majority of patients and the overall
technical and clinical success rateswere comparable to previous publications [4]
[10]
[11].
Ascites is not uncommon in malignant GOO and is generally associated with poor prognosis.
Ascites was observed in nearly 40 % of patients with malignant GOO in the current
study. Pathophysiology of ascites in malignant GOO is multifactorial, including peritoneal
involvement by malignancy, malnutrition, and portal hypertension due to underlying
cirrhosis or liver metastases. The presence of ascites poses challenges for the procedure
and anesthesia. Significant ascites precludes close apposition of the freely mobile
jejunum to the stomach, thus posing a technical challenge for EUS-GE. To reduce the
effect of ascites on EUS-GE, especially when there is large ascites, drainage of the
ascites prior to the procedure may be required. We carefully evaluated and assessed
ascites before performing EUS-GE. Wherever we felt significant ascites, we drained
it with paracentesis or a catheter. Such precautions prevented any interloop fluid
stagnation requiring aspiration during the EUS-GE. Irani et al described placement
of LAMS in presence of ascites before gallbladder drainage and afferent loop obstruction
[12].
Our study is the first to describe EUS-GE in the presence of ascites with comprehensive
details of the patients with GOO and ascites and their management.
In this study, EPASS was used in the majority of GOO patients. We believe that the
double balloon technique helps to anchor and fix the jejunal loop, thus allowing ease
of the final step of LAMS placement. The stability even in the presence of ascites
provides additional safety of the procedure. Almost all procedures in GOO with ascites
were performed exclusively using EPASS except one with afferent loop obstruction.
There were no technical failures or AEs due to ascites per se. Peritoneal infection
is an expected complication in these patients. However, none of the patients developed
any evidence of it immediately after the procedure or during follow-up, which could
be due to adequate prophylactic antibiotic coverage before the procedure. In this
study, the median survival was low in patients with ascites when compared to those
without ascites Presence of ascites generally indicates an advanced stage of malignancy.
Also, overall survival depends on many factors, including response to adjuvant treatments.
There are few strengths of this study. First, it showed the safety of EUS-GE using
EPASS in a high-risk group of patients with ascites and malignant GOO. Any procedural
AE in such patients is serious and life-threatening. Second, this study also highlighted
the various technical challenges encountered in detail, despite using EPASS. Third,
we systematically followed all the patients at regular intervals until death or last
contact to assess the impact of the procedure. The limitations of the study include
the relatively small sample size, which was mainly due to the rarity of an evolving
procedure from a single center. The second limitation was the retrospective analysis,
although from prospectively maintained data. We made sure to document all clinical
details systematically during the study period. Third, a control arm of other techniques
was lacking to clearly establish whether EPASS was better and safer. However, it may
be appropriate to say that fixing the jejunal segment to maintain close opposition
with the stomach using EPASS made the procedure relatively safer. Future prospective
multicenter studies comparing different techniques with large sample sizes are needed
to evaluate to the technical success and sustained clinical benefit of the procedure
in patients with ascites.
Conclusions
Ascites is common in patients with malignant GOO. EUS-GE is feasible in the presence
of ascites with EPASS. The technical and clinical success rates were similar to those
in patients without ascites.
