Introduction
Gastric outlet obstruction (GOO) is very common in the late stage of many malignant
tumors of the digestive system, such as gastric, pancreatic, periampullary, and duodenal
cancers [1 ]. GOO has a serious impact on patient quality of life because it is often associated
with symptoms such as nausea, vomiting, and loss of appetite [2 ]
[3 ].
Palliative therapies for malignant GOO include gastroenterostomy, endoscopic placement
of luminal stent, and total parenteral nutrition (TPN). TPN not only reduces patient
quality of life, it also increases medical expenses. Endoscopic placement of an enteric
self-expanding metal stent (SEMS) across the malignant stricture is an alternative
treatment option. With this procedure, patients can resume oral intake of food, but
it may be complicated by recurrent obstruction [1 ]
[2 ]
[3 ]. Patients can also return to oral intake after the gastroenterostomy procedure.
Gastroenterostomy includes surgical gastroenterostomy and endoscopic ultrasound (EUS)-guided
gastroenterostomy (EUS-GE). Surgical gastroenterostomy has become the standard palliative
therapy for malignant gastric outlet obstruction caused by carcinoma of the stomach,
duodenum or pancreas. However, many patients tend to be poor surgical candidates due
to malnutrition [2 ]. The surgical approach is invasive and can be associated with significant surgical
mortality [4 ]
[5 ]
[6 ]
[7 ]
[8 ]. With the advent of EUS skills and luminal stents, EUS-GE has been accepted as a
palliative treatment for malignant GOO because it is a less invasive and long-lasting
luminal patency method [3 ]. Since Fritscher [9 ] introduced the EUS-GE procedure in pigs, many studies have researched it as a treatment
option malignant GOO [10 ]
[11 ]
[12 ]. According to these studies, EUS-GE results in recovery of oral intake without risk
of tumor ingrowth, while also avoiding the potential morbidity of surgery.
EUS-GE has not been commonly used in Mainland China. However, GOO is very common in
China because of the high incidence of gastric and pancreatic cancer. The aim of this
study was to demonstrate the feasibility, safety, and short- and long-term prognosis
of EUS-GE.
Patients and methods
Patients
This was a single-center study. We retrospectively reviewed electronic endoscopy records
from March 2017 to June 2019 and 36 consecutive patients with malignant GOO who underwent
EUS-GE at the Affiliated Drum Tower Hospital of Nanjing University Medical School.
Malignant GOO was diagnosed based on clinical symptoms, laboratory examinations, and
imaging (including transabdominal ultrasound, computed tomography scan of the abdomen,
magnetic resonance cholangiopancreatography, and EUS). Inclusion criteria consisted
of primary or metastatic malignancies that resulted in symptomatic GOO rated with
the GOO Scoring System (GOOSS). The GOOSS is based on a score of 0 to 4 (0 (no oral
intake), 1 (liquid only), 2 (half fluid only), 3 (almost normal diet) and 4 (normal
diet)). Contraindications included presence of severe ascites, benign GOO, gastric
body tumor, or complete GOO thath could not be passed by the guidewire. We reviewed
each patient’s endoscopic and medical records and collected information on demographics,
etiology of GOO, symptoms, signs, time of tumor diagnosis, time of GOO occurrence,
GOOSS scores before treatment, technical success rate, clinical success rate, procedure-related
adverse events (AEs) and post-procedural length of hospital stay. As described by
Suzanne M et al. [13 ] previously, food intake was measured daily during the first 30 days after treatment
and then weekly by using patient diaries. GOOSS score was used to evaluate food intake.
Clinical success was defined as patient ability to tolerate oral intake without vomiting.
Health-related quality of life (HRQoL) was assessed with a standardized QoL, the EuroQol-5 D
(EQ-5 D).
The endpoint of observation was June 20, 2019. Follow-up continued from the procedure
to the death of patients or to the end of the study. Patients received telephone follow-up
by specially trained research nurses at 14 days, 1 month, and then monthly following
the procedure, to inquire about complications and record symptoms, signs, recurrent
obstruction, GOOSS scores after treatment, and survival time. Depending on their conditions,
patients were referred for outpatient or inpatient treatment as indicated. Data collected
during follow-up included laboratory tests for liver and kidney functions and imaging
results. Follow-up data were collected prospectively. Informed consent was obtained
from all patients prior to the endoscopic procedure. The study protocol was approved
by the Ethics Committee of the Nanjing Drum Tower Hospital.
Definitions
AEs were defined as clinically important complications that were considered to be
definitely or probably related to EUS-GE. Delayed bleeding was defined as hemorrhage
that occurred 24 hours after the procedure. The metal stent stripping the guidewire
coating was defined as the metal stent inserting under the guidewire and stripping
the guidewire, causing the insulating layer of the guidewire to fall off and influence
stent placement.
Procedure
Patients were not given antibiotics, but were instructed to fast for 2 days before
their scheduled EUS-GE. The procedure was performed by two experienced endoscopists
(Xiaoping Zou and Lei Wang), as previously described by Khashab et al. and Itoi et
al. [14 ]
[15 ] ([Fig. 1 ]). Both experts had independently performed EUS-guided interventions in more than
300 cases. EUS-GE was performed using a double balloon-assisted technique to build
a gastroenterostomy bypass. First, a 100-cm enteroscopy overtube was placed outside
the upper gastrointestinal endoscope. The endoscope was advanced to the stenosis of
the duodenum. A 0.025-inch guidewire (Olympus Medical Systems, Tokyo, Japan) was then
maximally advanced. The endoscope was removed, leaving the guidewire and the enteroscopy
overtube in place. The guidewire was then inserted across the obstruction to the jejunum.
A double balloon with triple cavities (Tokyo Medical University type; Create Medic,
Yokohama, China) was inserted over the guidewire to the distal duodenum or the deep
small bowel under fluoroscopic assistance. The enteroscopy overtube was subsequently
removed. The double balloons were filled with saline and contrast to let the small
intestine open. Saline and contrast with methylene blue were injected into the space
between the two balloons. The fluid-filled distal duodenum or the proximal jejunum
adjacent to the gastric body was identified by the curved linear array echo endoscope
(GF-UCT260, Olympus, Tokyo, Japan). When the deep distended jejunum was identified,
a 19-gauge fine aspiration needle (EchoTip Ultra; Cook Medical) was used to puncture
the jejunum under EUS guidance. The blue fluid was aspirated to demonstrate that the
needle was in the jejunum. A 0.035-inch guidewire was advanced through the 19-gauge
needle. A lumen-apposing self-expanding metal stent with electrocautery (Micro-Tech
(Nanjing) Co., Ltd.) was inserted through the stomach wall into the jejunum wall.
The distal anchor flange of the stent was deployed under EUS and fluoroscopic guidance.
Then the proximal anchor flange was deployed under endoscopic guidance. The stent
lumen was dilated with a 10-mm dilating balloon.
Fig. 1 Endoscopic ultrasound (EUS)-guided gastroenterostomy for malignant gastric outlet
obstruction. a Fluoroscopic view of guidewire inserting across the obstruction. b Fluoroscopic view of a double balloon with triple cavities inserting into the jejunum
and filling with saline and contrast. c The distal anchor flange of the stent was deployed under the EUS view. d The distal anchor flange of the stent was deployed under the fluoroscopic view. e Proximal flange was deployed under endoscopic imaging. f Fluoroscopic view of the completely deployed stent.
Statistical analysis
Statistical analysis was performed with SPSS statistical program (SPSS 20.0, Chicago,
Illinois, United States). Clinical, endoscopic, and histopathologic characteristics
of GOO and patient outcomes were analyzed. Continuous data were expressed as mean
± standard deviation (SD). Statistical analysis was performed using the student’s
t -test and the Mann Whitney U for normally distributed and non-normally continuous
variables, respectively. For categorical variables, data weres also presented as value
(%). The χ2 test or Fisher's exact test was used to evaluate differences in proportions. P < 0.05 was considered statistically significant.
Results
Demographics
As shown in [Table 1 ], the mean age of GOO patients was 69.0 years (range 45 – 88 years). The male-to-female
ratio was 0.89. The mean time of tumor history was 4 months (range 1 – 36 months).
The mean time of GOO occurrence before EUS-GE was 2 months (range 0.25 – 3 months).
Reported symptoms at the time for EUS-GE included abdominal pain in 24 (66.7 %), nausea
in 31 (86.1 %), vomiting in 31 (86.1 %), jaundice in nine (25.0 %), fever in four
(11.1 %) and weight loss in 33 (91.7 %). The most common sign was abdominal tenderness
in 16 (44.4 %). Seven patients (19.4 %) had a history of surgery. The most common
cause of GOO in this study was pancreatic cancer (15/36, 41.7 %), followed by cholangiocarcinoma
(8/36, 22.2 %), duodenal cancer (5/36, 13.9 %), metastatic cancer (4/36, 11.1 %),
and gastric cancer (4/36, 11.1 %). The average body mass index was 17.9 (range 17 – 27).
The most common obstruction site was the second part of the duodenum (23/36, 63.9 %),
followed by the pylorus (7/36, 19.4 %). For 10 patients, enteral stenting (ES) was
performed one to three times and the mean time without obstruction was 78 days.
Table 1
Baseline characteristics of 36 patients with malignant GOO undergoing EUS-GE.
Age (y), mean ± SD
69.0 ± 12.8
Sex, n (%)
17(47.2)
Time of tumor history (months), median
4.0(range 1~36)
Time of GOO before treatment (months), median
2.0(range 0.25~3)
GOOSS score, median
0.2(0~1)
Symptoms, n(%)
24(66.7 %)
31(86.1 %)
31(86.1 %)
33(91.7 %)
4(11.1 %)
9(25.0 %)
Signs, n (%)
16(44.4 %)
12(33.3 %)
Etiology of malignant gastric outlet obstruction, n (%)
4(11.1 %)
15(41.7 %)
8(22.20 %)
5(13.9 %)
4(11.1 %)
Site of the obstruction, n (%)
7(19.4 %)
4(11.1 %)
23(63.9)
2(5.6 %)
concomitance's disease
14(38.9 %)
12(33.3 %)
GOO, gastric outlet obstruction; EUS-GE, endoscopic ultrasound-guided gastroenterostomy
Outcomes of the EUS-GE
[Table 2 ] shows outcomes of 36 patients who underwent EUS-GE. All patients (36/36, 100 %)
underwent double balloon-assisted EUS-GE. The site of puncture was the posterior wall
of gastric body in 36 patients (36/36, 100 %). The mean number of punctures was 1.3
(range 1 ~ 3 times). Nine patients received more than one puncture and the most likely
cause was most likely low position of obstruction. Median total procedure time was
52 minutes (range 34 ~ 156 min). Median time for determination of puncture site was
20 minutes (range 15~28 min). Median time from puncture to successful delivery of
the stent was 38 minutes (range 19 ~ 128 min). Significantly less time was spent on
determination of puncture site than for stent placement (P = 0.001). The technical success rate was 100 % (36/36). The clinical success rate
was 94.4 % (34/36).
Table 2
Outcomes for 36 patients that underwent EUS-GE.
Site of puncture, n (%)
36(100 %)
Number of punctures, mean±SD
1.3 ± 0.7
Total procedure time(minutes), median ± SD
52.0 (34 – 156)
Time for determination of puncture site (minutes), median±SD
20 (15 – 28)
Time between puncture to stent delivery (minutes), median±SD
38 (19 – 128)
Technical success, n (%)
36(100 %)
Clinical success, n (%)
34(94.4 %)
Adverse events during procedure, n (%)
2(5.6 %)
1(2.8 %)
3(8.3 %)
Adverse events after procedure, n (%)
2(5.6 %)
3(8.3 %)
2(5.6 %)
EUS-GE, endoscopic ultrasound-guided gastroenterostomy
AEs were seen in nine patients (9/36, 25.0 %) and 13 total AEs occurred. During the
procedure, three patients (3/36, 8.3 %) had the metal stent stripped of guidewire
coating, which affected the ability of the cutting current to penetrate the metal
stent through the gastrointestinal wall. Replacement of guidewire or additional puncture
became a remedy. The total procedure time was 109 minutes per patient. One patient
had stent misdeployment during the procedure, and the stent was pulled out and a new
stent was deployed successfully. After the procedure, stent migration was seen in
two patients (2/36, 5.6 %), including one who had the metal stent stripped of guidewire
coating ([Fig. 2 ]) during the procedure, and the new bare metal stent was inserted into the covered
metal stent. After the new stent insertion, one patient had an AE and bled severely.
Endoscopy showed that blood oozed from the puncture site and the patient died from
failure of endoscopic hemostasis and conservative treatment. For the other patient,
when the new bare-metal stent was unable to control the bleeding, the two stents,
including the bare and the covered one, were pulled out. Then the guidewire was inserted
through the puncture site into the jejunum, and the new covered metal stent was placed,
leading to resolution of the bleeding. In seven patients, postoperative pain occurred
in seven patients, which required readmittance within 1 to 4 days after conservative
treatment. No peritonitis occurred. No bleeding occurred during the procedure.
Fig. 2 The mental stent stripping the guidewire coating.
Quality of life and prognosis
[Fig. 3 ] shows food intake and EQ-5 D scores before EUS-GE and after EUS-GE. Only seven patients
ate a little fluid food and the other 29 patients were unable to eat any food. The
GOOSS score was 0.2 before EUS-GE. Food intake improved rapidly after EUS-GE. After
stent placement, ability to eat was restored in a median of 2 days. The GOOSS score
was 2.2 at 15 days after EUS-GE, which was higher than before EUS-GE (P = 0.001). During the follow-up period, the GOOSS was still higher than 2 at 30, 60,
90, 120, 150, and 180 days. EQ-5 D scores showed a slight increase after EUS-GE within
60 days, and then they gradually went down.
Fig. 3 Assessment of quality of life. a Gastric outlet obstruction (GOO) Scoring System (GOOSS) scores of patients with GOO
before endoscopic ultrasound-guided gastroenterostomy (EUS-GE) and after EUS-GE (GOOSS
score of 0, no oral intake; 1, liquid only; 2, half fluid only; 3, almost normal diet;
4, normal diet). b EuroQol-5 D (EQ-5 D) scores of patients with GOO before EUS-GE and after EUS-GE.
Post-procedure length of hospital stay was 5.8 ± 4.7 days. Median follow-up time was
89 days. The survival curve for patients undergoing EUS-GE is shown in [Fig. 4 ]. Median survival time was 103 days. The rate of GOO recurrence was 2.7 % (1/36,
2.7 %). The endoscopic views showed that the stent was patent and iodine water radiography
showed that the iodine water flowed from the stomach to the jejenum. However, symptoms
of vomiting occurred again and the patient was not able to eat 56 days after the procedure.
Reexamination of computed tomography showed that the tumor had metastasized to the
abdominal cavity, resulting in downstream intestinal congestion.
Fig. 4 Survival curves of patients undergoing endoscopic ultrasound-guided gastroenterostomy.
Discussion
GOO is very common in China because of the high incidence of malignant tumors of the
the digestive system. Most of these patients have advanced, unresectable tumors and
thus, symptom palliation and improvement in quality of life are the usual goals of
any intervention [16 ]. In 2002, Fritscher-Ravens et al. first reported on EUS-guided gastrojejunostomy
in pigs. In recent years, EUS-GE has been used as a palliative treatment for malignant
GOO. Several clinical and animal studies have shown that the technical success rate
with it is 86.7 % to 100 %. In clinical studies, success rates range from 83.3 % to
100 % [9 ]
[10 ]
[11 ]
[12 ]
[15 ]
[17 ]
[18 ]
[19 ]
[20 ].
This trial used palliative EUS-GE to treat late malignant GOO. In this study, 36 patients
with malignant GOO were included and the most common etiology for GOO was pancreatic
cancer. All patients underwent balloon-assisted EUS-GE. We selected the metal covered
stent with electrocautery delivery system, which supports electrocautery dilation
and stent deployment without the need for primary dilation. The site of puncture was
at the posterior wall of gastric body in 36 patients (100 %). The technical success
rate was 100 %, and the clinical success rate was 94.4 %. Food intake improved after
EUS-GE. GOOSS scores increased from 0.2 to 2.2. Most patients could eat soft food
without vomiting. If the stent was patent, the GOOSS score increased to more than
2 at 180 days after the procedure. Overall, our data suggest that EUS-GE is a very
effective treatment for symptomatic GOO in China. A long-term cohort study has shown
that the rate of reintervention is low with the EUS-GE approach [21 ].
EUS-GE has some complications, including bleeding, stent misdeployment, and recurrence
of obstruction [20 ]
[22 ]. Chen YI et al. reported a rate of severe AEs of 10 % in the EUS-GE group. However, overall rates
of AEs were not significantly different between the EUS-GE group and the endoscopic
ES group [20 ]. Previous studies showed that stent occlusion was very common in the ES group [3 ]
[23 ]
[24 ]. Compared to ES placement, EUS-GE has a lower rate of stent failure requiring repeat
intervention [25 ]. However, the covered enteral stent has higher rates of stent migration and tumor
overgrowth remains a significant problem in ES [26 ]
[27 ]
[28 ]. Mouen A et al. reported three misdeployments of the stent first flange in the peritoneum and two
episodes of abdominal pain requiring hospitalization. Stent misdeployment prolonged
hospitalization, but conservative treatment with antibiotics was safe [22 ]. Tyberg A et al. also reported one death in a severely debilitated patient, which was associated
with procedure-related peritonitis [10 ]. Our data show that procedure-associated complications are metal stent stripping
the guidewire coating and stent misdeployment. The insulated outer sheath of the guidewire
was cut down and covered on the front end of the electrocoagulation part of the metal
stent, which affected cutting current penetration into the gastrointestinal wall.
In one patient, the metal stent cut the guidewire, prolonging the procedure time.
Stent misdeployment occurred in one patient during the procedure. We pulled out the
stent and successfully deployed a new stent, and no severe AE occurred. Post-procedure
AEs included stent migration in two patients, and the new bare metal stent was inserted
into the primary stent. One patient experienced severe bleeding and died due to failure
of endoscopic and conservative treatment and patient refusal of surgical treatment.
In the other patient, when the second stent was found to be ineffective, the two stents
were pulled out quickly. We deployed a new metal-covered stent, which controlled the
bleeding. The patient recovered with conservative treatment. In this study, we had
only one case of severe bleeding after EUS-GE. Overall, we found that the procedure
was safe for most of our patients. However, the reasons for stent migration need to
be evaluated carefully before a new bare stent is deployed into the primary stent,
to avoid fatal complications.
In this study we used a LA-FSEMS with cautery tip. The distal end of the stent delivery
system was inserted through the stomach wall into the jejunal lumen by applying electrocautery.
This delivery system does not require balloon catheters or endoscopic removal before
stent insertion. This type of stent, with the same function, is manufactured by several
different companies [10 ]
[11 ]. Our results showed that a metal stent cutting the guidewire could affect electrocoagulation
and prolong the procedure time. During the procedure, cutting of the guidewire by
the stent is confirmed, the guidewire should be reinserted into the jejunum and the
damaged guidewire kept far away from the puncture area. The alternative is to pull
out the guidewire and stents and do the puncture again. In our study, two patients
had stent migration after the procedure. In the future, research on stent choice to
avoid migration should be undertaken, including size and width of biflanges, and apposing
properties.
During EUS-GE, it is difficult to puncture the jejunum because the jejunum is highly
mobile and can be easily moved away from the stomach. We undertook EUS-guided double-balloon-occluded
gastroenterostomy and chose to puncture the small bowel around the ligament of Treitz
because that part of the intestine is adjacent to the stomach. In this study, puncture
was successfully in 36 patients. A 19-gauge FNA needle was used to puncture the jejunum
(between the two balloons) rather than the balloon with filled saline and contrast
because the balloon could have affected the electrocoagulation. It was not appropriate
to perform EUS-GE when the tumor had expanded to the jejunum around the ligament [29 ].
Mean length of hospital stay for EUS-GE in a comparative trial of endoscopic it versus
surgical gastrojejunostomy was 11.6 days. However, mean length of hospital stay for
EUS-GE was 12 days [22 ]. The first US clinical study of EUS-GE demonstrated that mean length of hospital
stay was 2.2 days [11 ]. Comparatively, our data showed that mean length of hospital stay was 5.8 days.
Although the other study demonstrated that there was no symptom recurrence during
a mean follow-up period of 150 days [11 ], previous studies have reported symptom recurrence during the follow-up period because
of stent obstruction with food, which was successfully treated with endoscopic extraction
[22 ]. One patient in our study had recurrence of symptoms. Neither stent migration nor
obstruction could be confirmed with endoscopy or iodine water. However, the patient
had symptoms of vomiting and nausea 56 days after the procedure. Median survival time
after EUS-GE was 103 days in two previous studies [20 ]
[22 ]. It was the same in our study, which showed that EUS-GE is valuable for many patients
with advanced malignant GOO. Overall, EUS-GE has many advantages, such as minimal
hospital stay after the procedure, lower rate of symptom recurrence, and improved
overall survival.
There were several limitations to our study. First, it was single-center and retrospective.
Second, surgical gastroenterostomy and endoscopic enteral stenting were not compared.
Third, only one type of stent was used and there was one fatal AE. In the future,
potential for AEs during and after the procedure should be controlled. Multicenter,
prospective, and comparative studies should be performed of endoscopic versus surgical
gastroenterostomy and endoscopic enteral stenting for malignant GOO.
Conclusion
In conclusion, our results indicated that EUS-GE was safe and effective. EUS-GE has
many advantages for patients with advanced malignant GOO, such as its minimally invasive
nature, requirement for only a short postoperative stay, lower rate of symptom recurrence,
and improved overall survival rates. However, the procedure is challenging procedure
and should only be performed by experienced physicians. Moreover, appropriate patient
selection for EUS-GE is important. Overall, EUS-GE is a promising treatment option
for malignant GOO and proper care should be used when considering patients for this
procedure.
