Introduction
Bleeding from gastric varices is more severe and life-threatening than bleeding from
esophageal varices and is associated with high mortality and morbidity rates [1 ]. Management of bleeding from gastric varices is challenging and requires expertise
because aggressive rebleeding can occur [1 ]
[2 ].
Treatment of varices by injection of cyanoacrylate via standard gastroscopy is associated
with a higher hemostasis rate and a lower rebleeding rate than either band ligation
or sclerotherapy [3 ], but has been associated with adverse events, such as pulmonary embolization, bleeding,
fever, chest pain, and even death [4 ]
[5 ]
[6 ]. In addition, the endoscopic injection of cyanoacrylate has been shown to result
in damage to the working channel of the endoscope [7 ]
[8 ]. Moreover, complete variceal obliteration is difficult to confirm during the procedure
and may require additional therapeutic sessions.
The use of endoscopic ultrasonography (EUS) to guide the cyanoacrylate injection has
several advantages over guidance by visible-light endoscopy. These advantages include
the precise targeting of the feeder vessel of the gastric varix and the requirement
for a lesser amount of cyanoacrylate for obliteration, which is an important consideration
because the risk of glue embolization is dependent on the volume of cyanoacrylate
injected [9 ]
[10 ]. In addition, EUS evaluation can confirm obliteration of the varix via Doppler imaging
after the procedure [11 ].
The EUS-guided deployment of embolization coils within gastric varices is associated
with fewer adverse events than EUS-guided injection of cyanoacrylate, but has the
disadvantages of a lower obliteration rate and a higher cost [12 ]
[13 ]. The combined injection of coils and cyanoacrylate was first described in 30 patients
and resulted in a 95.8 % obliteration rate [14 ]. In a larger cohort of 152 patients, this combined approach showed a 93 % obliteration
rate with a 3 % rebleeding rate; however, in this retrospective study, only 65.7 %
of patients had an EUS Doppler study performed to evaluate complete obliteration [15 ].
There is a lack of prospective data comparing EUS-guided coiling and cyanoacrylate
injection versus coiling alone, with a focus on the obliteration, rebleeding, and
adverse event rates, to define the most appropriate obliteration technique. The objective
of the present study was to compare the efficacy and safety of EUS-guided combined
deployment of coils and cyanoacrylate injection with EUS-guided coil deployment alone
for the management of patients with gastroesophageal varices type II (GOV II) and
isolated gastric varices type I (IGV I) in a prospective randomized controlled trial.
Methods
Study design
An interventional, parallel-randomized controlled trial was performed at the Instituto
Ecuatoriano de Enfermedades Digestivas (IECED), a tertiary referral center in Ecuador,
from March 2016 to October 2018. Patients were randomly allocated in a 1:1 manner
to either EUS-guided coil deployment or EUS-guided coil deployment and cyanoacrylate
injection for the management of gastric varices types GOV II and IGV I.
The study protocol and consent form were approved by the Institutional Review Board,
and the study was conducted in accordance with the Declaration of Helsinki. Written
informed consent was obtained from all participants or from designated relatives.
The study was registered at ClinicalTrials.gov under the code NCT03155256.
Population selection and inclusion and exclusion criteria
Patients fulfilling the following criteria were considered to be candidates for inclusion:
age ≥ 18 years, history of liver cirrhosis with endoscopic evidence of GOV II or IGV
I in accordance with the Sarin classification [2 ], and patient preference for EUS-guided therapy. Liver cirrhosis was defined in accordance
with the clinical and imaging findings; all patients had a transient elastography
(Fibroscan, EchoSens, Paris) consistent with liver cirrhosis. The study population
included patients with active bleeding, a history of previous bleeding secondary to
gastric varices (secondary prophylaxis), and those eligible for primary prophylaxis
in accordance with the Baveno VI consensus [16 ].
Patients were excluded if they had concurrent hepatorenal syndrome and/or multiorgan
failure, were pregnant or nursing, or had suspected splenic or portal vein thrombosis,
a platelet count < 50 000 /mL or an international normalized ratio (INR) ≥ 2, esophageal
stricture, or a known allergy to iodine.
Study outcomes
The primary end point of the study was a comparison of the efficacy of the two EUS-guided
procedures in terms of their technical and clinical success rates. Technical success
was defined as completion of the intended procedure. Clinical success was considered
to be complete and immediate obliteration of the varix, evaluated via esophagogastroduodenoscopy
(EGD) and the absence of flow during EUS Doppler evaluation.
The secondary end points of the study were the reappearance of gastric varices during
follow-up, rebleeding, the need for reintervention, and the complication and survival
rates in the two arms of the study.
Endoscopic procedure and technique
All procedures were performed in a hospital-based interventional endoscopy unit by
one endoscopist (C.R.-M.), who was blinded to each patient’s medical history. Procedures
were performed with the patient under general anesthesia, with tracheal intubation,
and in the supine position. All patients received antibiotic prophylaxis with 2 g
of ceftriaxone intravenously. The EUS-guided procedure was performed with a linear-array
therapeutic echoendoscope (3.8-mm working channel; EG3870UTK; Pentax, Hamburg, Germany)
attached to an ultrasonography console (Avius Hitachi, Tokyo, Japan). The echoendoscope
was positioned in the distal esophagus at the level of the cardia to enable visualization
of the gastric fundus and intramural varices.
After the echoendoscope had been positioned, water was instilled into the gastric
fundus to improve acoustic coupling and visualization of the gastric varices. Direct
visualization of the blood flow in the varices was evaluated via color Doppler on
the ultrasonography console. The offending vessel or vessels were followed from the
cardia to the proximal part of the esophagus, 2 – 3 cm above the cardia, to detect
the feeding vessel, which was considered to be the convergence of all offending vessels.
An EUS-guided fine-needle puncture was performed with a 19-gauge needle (Expect Flexible;
Boston Scientific, Marlborough, Massachusetts, USA) to access the feeder vessel; the
stylet was withdrawn and a 20-ml negative-pressure syringe was used to evaluate blood
return, thereby confirming the intravascular location. To prevent blood clotting in
the needle tip, 5 mL of saline solution was instilled. The endoscopist was able to
confirm the flow in the gastric varices with B-mode ultrasonography. Additionally,
EUS-guided varicealography was performed to confirm the target vessel and the flow
trajectory, as previously described [9 ].
Immediately following the EUS therapy, EGD was performed to evaluate the disappearance
of the gastric varices.
EUS-guided coil embolization and cyanoacrylate injection
Coils were deployed under EUS guidance, followed (in selected individuals) by injection
of 2-octyl-cyanoacrylate (Dermabond; Ethicon, Piscataway, New Jersey, USA). The 0.035-inch
diameter Nester Embolization Coils (Cook Medical, Bloomington, Indiana, USA) were
used for intravascular embolization. The coils were 10 – 16 mm in diameter and 12 – 20 cm
in straight length. Coils were delivered into the vessels through the EUS needle using
the stylet to advance them. To minimize the risks of perforation, bleeding, and coil
extrusion, and to provide enough space for the coils to curl, care was taken to avoid
placing the needle tip on the opposite wall of the varix. The size of the coil chosen
was > 120 % of the varix diameter according to the diameter measured during EUS; coils
were deployed in the direction of the portal vein (blood-outflow trajectory).
The cyanoacrylate injection was performed slowly after coil deployment. The volume
of the desired cyanoacrylate injection was measured in accordance with the diameter
of the vessel (< 2.5 mL), and 1 mL of saline solution was used to flush the glue completely
through the needle under EUS guidance. After 90 – 120 seconds, the cyanoacrylate had
solidified and the risk of bleeding at the puncture site had decreased, and the needle
was withdrawn. [Fig. 1 ] summarizes the procedure for combined coil deployment and cyanoacrylate injection.
The steps for the EUS-guided combined coil deployment and cyanoacrylate injection
are described in a short procedure video ([Video 1 ]).
Fig. 1 Procedure for endoscopic ultrasonography (EUS)-guided deployment of embolization
coils and cyanoacrylate injection in gastric varices. a Endoscopic view of the gastric varix (type IGV I) before treatment. b EUS evaluation of the gastric varices prior to therapy. c EUS Doppler evaluation of the feeder vessel for the gastric varices (left panel),
showing the flow amplitude wave of the feeder vessel (right panel). d EUS-guided deployment of coils and cyanoacrylate injection into the feeder vessel.
e Fluoroscopic evaluation of the deployed EUS-guided coils. f Endoscopic view after combined therapy showing disappearance of the varix.
Video 1 Endoscopic ultrasonography (EUS)-guided combined therapy with coil deployment and
cyanoacrylate injection with the addition of EUS-guided varicealography. (a ) Endoscopic evaluation of gastroesophageal varices type II. (b ) Endoscopic ultrasound evaluation of gastric varices and the feeder vessel. (c ) EUS-guided varicealography with evaluation of the flow trajectory. (d ) EUS-guided deployment of embolization coils with combined cyanoacrylate injection,
followed by EUS-guided Doppler evaluation of the feeder vessel and gastric varices
for obliteration confirmation. (e ) Endoscopic visualization of gastric varices after EUS-guided combined therapy.
EUS-guided coil embolization alone
In patients allocated to undergo EUS-guided coil embolization alone, coil insertion
was performed as described above, again using Nester Embolization Coils > 120 % of
the varix diameter.
Follow-up
Follow-up EGD and an EUS Doppler evaluation were performed 3 months after the initial
procedure by a second endoscopist (M.V.), who was blinded to the initial procedure.
The reappearance of gastric varices, need for reintervention, and survival rates were
evaluated. Patients were followed-up until the date of death or up to 12 months after
enrollment via visits to the clinic and review of medical, EGD, and EUS records.
If patients required reintervention secondary to varix reappearance or recurrent bleeding,
a second EUS-guided procedure was performed, subject to the endoscopist’s preference.
All reinterventions were performed in the same endoscopy unit as the index procedure.
Statistical analysis
A sample size of 30 participants per study group was calculated through a sample size
formula to compare two proportions (two-samples, one-sided), on the basis of a 5 %
α error, a 20 % β error, κ = 1, and success rates of 82 % for complete obliteration with EUS-guided coil embolization
and 53 % for EUS-guided cyanoacrylate injection, as described by Romero-Castro et
al. [13 ].
Categorical variables were described as frequencies or proportions (%). Numerical
variables were described as mean and standard deviation, or median and minimum to
maximum range, according to statistical distribution (Kolmogorov–Smirnov test). Differences
between the characteristics of the study groups were established through corresponding
hypothesis tests: chi-squared or Fisher’s exact test for categorical variables, Student’s
t test for normally distributed numerical variables, Mann–Whitney U test for non-normally distributed numerical variables. Hypothesis testing was performed
as a one-tailed analysis.
For primary outcome analysis, the effect of combined therapy on immediate disappearance
of the varix was estimated through relative risk (RR). For secondary outcome analysis,
survival time and reintervention-free time between the two study groups were compared
through Long-rank and Gray’s test, respectively. The effect of combined therapy on
overall survival and reintervention-free survival were estimated through the hazard
ratio (HR). A P value < 0.05 was considered statistically significant.
Data analysis was performed using R version 3.4.2 (R Foundation for Statistical Computing;
Vienna, Austria).
Results
Demographics
A total of 60 patients were enrolled, with 30 assigned to each group ([Fig. 2 ]). There were no significant differences in the baseline characteristics between
the two groups ([Table 1 ]). None of the patients in either group had previous interventions for the management
of the gastric varices.
Fig. 2 Flowchart of the management of the study patients. EUS, endoscopic ultrasonography.
Table 1
Demographic and clinical characteristics of the 60 patients with gastric varices who
were treated either by coils and cyanoacrylate injection or by coils alone.
Coils + CYAn = 30)
Coils alonen = 30)
P value
Age, mean (SD), years
61.8 (7.8)
61.6 (12.3)
0.95[1 ]
Sex, female, n (%)
14 (46.7)
11 (36.7)
0.43[2 ]
Cirrhosis etiology, n (%)
7 (23.3)
10 (33.3)
0.39[2 ]
23 (76.7)
20 (66.7)
Cirrhosis severity
Child–Pugh score, median (range)
6 (5 – 9)
6 (5 – 11)
0.29[3 ]
Child – Pugh score, n (%)
28 (93.3)
26 (86.7)
0.53[2 ]
2 (6.7)
3 (10.0)
0
1 (3.3)
MELD score, median (range)
9.5 (6 – 13)
9.5 (6 – 30)
0.46[3 ]
Indication, n (%)
Primary prophylaxis
3 (10.0)
4 (13.3)
0.50[4 ]
Active bleeding
1 (3.3)
5 (16.7)
0.09[4 ]
Secondary prophylaxis
26 (86.7)
21 (70.0)
0.09[4 ]
CYA, 2-octyl-cyanoacrylate; SD, standard deviation; MELD, model of end-stage liver
disease.
1 Student’s t test.
2 Chi-squared test.
3 Mann – Whitney U test.
4 Fisher’s exact test.
The median number of coils placed per patient was two in the combined treatment group
and three in the single treatment group (P = 0.006), and there was no significant difference between the distributions of coils
of different sizes in the two groups. After enrollment, patients were evaluated for
the initiation of beta-blocker therapy, which was tolerated in 83 % and 80 % of patients
in the combined and coiling alone groups, respectively (P = 0.74). The characteristics of the gastric varices and procedure, along with the
primary outcomes of the study are summarized in [Table 2 ].
Table 2
Comparison of variceal and procedural characteristics and primary outcomes of the
study.
Coils + CYAn = 30)
Coils alonen = 30)
P value
Variceal and procedural characteristics
Type, n (%)
19 (63.3)
12 (40.0)
0.07[1 ]
11 (36.7)
18 (60.0)
Diameter, median (range), mm
21 (10 – 32)
25 (10 – 38)
0.15[2 ]
Number of coils placed, median (range)
2 (1 – 3)
3 (1 – 7)
0.006[1 ]
Size of coils placed, n
15
20
0.19[1 ]
12
14
0.60[1 ]
11
13
0.59[1 ]
12
13
0.79[1 ]
CYA volume, median (range), mL
1.8 (1.2 – 2.4)
–
n/a
Primary outcomes
Technical success, n (%)
30 (100.0)
30 (100.0)
n/a
Complete obliteration, n (%)
30 (100.0)
27 (90.0)
0.12[3 ]
Immediate varix disappearance, n (%)
26 (86.7)
4 (13.3)
< 0.001[3 ]
Adverse events, n (%)
2 (6.7)
1 (3.3)
0.50[3 ]
Type of adverse event, n (%)
1 (50.0)
1 (100.0)
1 (50.0)
0
CYA, 2-octyl-cyanoacrylate; GOV II, gastroesophageal varices type II; IGV I, isolated
gastric varices type I.
1 Chi-squared test.
2 Mann–Whitney U test.
3 Fisher’s exact test.
Outcome results
Primary end points
Technical success of the intended procedure was achieved in 100 % of individuals in
both study arms. Complete obliteration as evidenced by EUS Doppler results was observed
in 100 % of participants with the combined technique and in 90 % of those with EUS-guided
coil embolization alone (P = 0.12). Immediate disappearance of the varix under direct visualization by EGD was
observed in 26 of the 30 individuals with the combined treatment, but in only four
of the 30 individuals with the single treatment (RR 6.5, 95 % confidence interval
[CI] 2.6 – 16.4; P < 0.001).
Secondary end points
The secondary end points were rebleeding, reappearance of the gastric varices during
follow-up, the need for reintervention, and survival rates in the two study arms.
The median follow-up duration was 14.5 months (range 0.6 – 31.2 months). The rebleeding,
varix reappearance, and reintervention rates were all significantly lower in the combined
treatment group than in the single treatment group ([Table 3 ]).
Table 3
Comparison of the secondary outcomes of the study.
Coils + CYA
Coils alone
P value
Survival time, median (range), months
16.4 (0.6 – 31.2)
14.2 (0.8 – 28.2)
0.90[1 ]
Mortality rate, n (%)
9/30 (30.0 %)
8/30 (26.7)
0.84[2 ]
Cause of death, n
Liver failure
1
0
Hepatocellular carcinoma
3
2
Uncontrolled hemorrhage
5
5
Acute coronary syndrome
0
1
Rebleeding, n (%)
1 (3.3)
6 (20.0)
0.04[2 ]
Varix reappearance, n (%)
4 (13.3)
14 (46.7)
< 0.001[2 ]
Reintervention-free time, median (range)
15.8 (0.3 – 31.2)
12.5 (0.1 – 20.2)
0.01[3 ]
Reintervention, n (%)
5 (16.7)
12 (40.0)
0.045[2 ]
Type of reintervention, n (%)
0.09[2 ]
Coils
0
5 (41.7)
Coils + CYA
4 (80.0)
7 (58.3)
CYA, 2-octyl-cyanoacrylate.
1 Long-rank test.
2 Chi-squared test.
3 Gray’s test.
In the combined treatment group, a first reintervention was required in five individuals,
four patients were treated with combined coil deployment and cyanoacrylate injection
within 3, 6, 14, and 15 months after enrollment. In the group treated with EUS-guided
coil embolization alone, 12 patients required a first reintervention for the management
of gastric varices: six reinterventions (50 %) occurred within 3 months of enrollment,
four by completion of the 1-year follow-up period, and two reinterventions were required
13 and 14 months after enrollment. Of these 12 patients, seven were treated for the
reappearance of the gastric varix with EUS-guided coil embolization and CYA injection,
of these two required a second reintervention with coil embolization and CYA injection.
Five individuals were treated with EUS-guided coil embolization alone as the first
reintervention; however, two required a second reintervention with coil embolization
and CYA injection.
The overall mortality rates were 30 % in the combined treatment group and 26.7 % in
the coil-only group (HR 0.95, 95 %CI 0.361 – 2.532; P = 0.90). The causes of death in both groups are listed on [Table 3 ]. Five patients in the combined treatment group died from uncontrolled hemorrhage
during follow-up (gastric varices [n = 2], esophageal varices [n = 3]); however, two
of these five patients had a concomitant hepatocellular carcinoma and all of the deceased
patients were Child – Pugh class C at the date of their death. Five patients in the
coiling-alone group also died from uncontrolled hemorrhage during follow-up (gastric
varices [n = 2], esophageal varices [n = 2], and gastric ulcer [n = 1]); one patient
had a concomitant hepatocellular carcinoma and again all of the deceased patients
were Child – Pugh class C at the date of death. The median overall survival period
was 16.4 months (0.6 – 31.2 months) for combined treatment and 14.2 months (0.8 – 28.2
months) for EUS-guided coil embolization alone (P = 0.90)
In the combined treatment group, 83.3 % of individuals were free from reintervention
during follow-up, compared with 60 % in the single treatment group (HR 0.27, 95 %CI
0.095 – 0.797; P = 0.01). The median reintervention-free period was 15.8 months (range 0.3 – 31.2
months) for those who underwent EUS-guided coil embolization with cyanoacrylate injection,
compared with 12.5 months (0.1 – 20.2 months) for those who underwent EUS-guided coil
embolization alone (P = 0.01). A cumulative incidence curve demonstrating the difference in the reintervention
rates between the two study groups is shown in [Fig. 3 ].
Fig. 3 Cumulative incidence curve of reintervention among all reinterventions performed.
Five patients (16.7 %) in the combined treatment group and 12 patients (40 %) in the
single treatment group underwent reinterventions (hazard ratio 0.27, 95 % confidence
interval 0.095 – 0.797; P = 0.01). Some patients had more than one reintervention.
Outcomes subanalysis
In a subanalysis excluding those patients treated with actively bleeding gastric varices,
we found a non-statistically significant difference in the rebleeding and reintervention
rates; however, we found a higher rate of gastric variceal reappearance in the EUS-guided
coiling-alone group (Table 1 s , available in online-only Supplementary Material).
In contrast, in a subanalysis excluding those patients treated for primary prophylaxis
(n = 7), we found a statistically significant difference in the EUS-guided coiling
group, with higher rebleeding, reappearance, and reintervention rates (Table 2 s ).
Discussion
In our study population of patients with GOV II and IGV I, the combined therapy of
EUS-guided coil deployment and cyanoacrylate injection resulted in a higher rate of
variceal disappearance, a significantly lower rebleeding rate, and a significantly
lower level of reintervention compared with EUS-guided coil deployment alone.
Studies of EUS-guided treatment of gastric varices using cyanoacrylate, coil embolization,
and combined cyanoacrylate injection and coil embolization have been conducted previously
[9 ]
[12 ]
[14 ]
[15 ]
[17 ]. EUS-guided injection of cyanoacrylate alone is associated with a higher risk of
pulmonary embolization or migration than EUS-guided coil embolization alone [13 ]. The combined approach, in which EUS-guided coil deployment is combined with cyanoacrylate
injection with additional endosonographic varicealography, has been shown to be safe
and effective in targeting the feeding vessel for complete variceal obliteration [9 ]. This method has the benefits of the use of a small volume of glue, with a higher
procedural success rate and a high rate of complete gastric variceal disappearance
[9 ]
[14 ].
Here, we compared the combined approach with EUS-guided coil embolization in a parallel-randomized,
controlled trial to determine the efficacy, safety, reappearance rate, recurrence
of gastric variceal bleeding, and survival rates of both techniques. Previously, N-butyl-cyanoacrylate
and 2-octyl-cyanoacrylate have been compared for the treatment of gastric varices,
and we used the latter because it eliminates the need to dilute the cyanoacrylate
with Lipiodol, which is viscous and makes injection more difficult, and has a longer
polymerization time, which reduces the risk of endoscope damage because of glue impaction
in the working channel [9 ]
[10 ].
In our population, the obliteration rate of combined coil and cyanoacrylate embolization
was 100 %, compared with 90 % after coil embolization alone, when confirmation was
defined by EUS Doppler; whereas, when evaluating the immediate varix disappearance
by endoscopic view, the combined therapy achieved an 86.6 % disappearance rate compared
with 13.3 % in the EUS-guided coiling group. Therefore, EUS Doppler evaluation confirmed
a higher obliteration rate for EUS-guided combined therapy more accurately than evaluation
of varix disappearance by endoscopic view. EUS Doppler evaluation for defining varix
obliteration should therefore be used instead of the endoscopic disappearance of gastric
varices during EGD.
Despite the confirmation of gastric variceal obliteration, absence of flow during
EUS Doppler was not significantly different between the study groups (100 % for combined
therapy vs 90 % for coil embolization alone); EUS Doppler confirmation of obliteration
also did not necessarily predict the need for further reinterventions. However, a
higher need for reintervention and a higher rebleeding rate were noted in those patients
treated with EUS-guided coiling alone in whom immediate varix disappearance evaluated
via EGD was statistically inferior compared with that in the combined therapy group.
Notably, 40 % of patients in the single treatment group required additional coil embolization
or combined coil and cyanoacrylate embolization during the follow-up period. The high
level of reintervention following coil embolization without cyanoacrylate will affect
the relative cost-effectiveness of the two techniques, considering that patients initially
allocated to combined therapy required fewer reinterventions than those in the EUS-guided
coiling-alone group.
A limitation of the present study is the generalizability of these findings because
all procedures were performed by a single endoscopist at a single tertiary center.
Even though there was no statistically significant difference in the number of patients
treated for active bleeding in the two groups, a higher number of patients were treated
with EUS-guided coiling alone compared with the combined therapy (16.7 % vs. 3.3 %;
P = 0.09) and this might have a role in some of the difference in outcomes. Despite
the promising results described herein, the generalizability of these findings may
depend on the availability of endoscopists trained to offer these EUS interventional
procedures. Therefore, a multicenter, superiority, randomized controlled trial should
be conducted to clarify the potential real-world clinical impact of EUS-guided placement
of coils and cyanoacrylate.
In conclusion, combined EUS-guided coil embolization and cyanoacrylate injection achieved
excellent clinical success, with low rates of rebleeding and reintervention, and high
reintervention-free time, in patients with GOV II and IGV I.
