Introduction
Current guidelines recommend endoscopic resection (ER) as the primary treatment option
for early esophageal neoplasia [1 ]
[2 ]
[3 ]. Next to the curative potential, ER may serve as an excellent diagnostic tool by
enabling accurate histopathological staging of the tissue specimen on which further
patient management is based. In the esophagus, the most commonly used resection methods
are endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD)
[4 ].
Despite both resection techniques having been proven safe and effective for early
esophageal neoplasia [5 ]
[6 ]
[7 ]
[8 ], patients with esophageal varices are considered a high-risk group owing to an increased
risk of massive bleeding [9 ]. However, most of these patients are also precluded from major esophageal surgery
because of portal hypertension resulting in collateral blood vessels [10 ]. In addition, a diagnosis of esophageal neoplasia may exclude selected patients
from liver transplantation [11 ]. Therefore, endoscopic therapy may be the only treatment option for this specific
patient category, provided that the prognosis of the underlying liver disease justifies
treatment of the early esophageal neoplasia.
Thus far, several approaches have been described in the literature to reduce the bleeding
risk and thereby enable ER in these patients. Ideally, varices are eradicated before
the planned ER through endoscopic band ligation (EBL), endoscopic injection sclerotherapy
(EIS), or transjugular intrahepatic portosystemic shunt (TIPS) [12 ]
[13 ]
[14 ]
[15 ]. Preventative EBL and EIS, however, is not possible if the neoplastic lesion is
located at the esophagogastric junction. In addition to the treatment of varices prior
to ER, bleeding may also be prevented by periprocedural interventions, such as direct
coagulation [16 ].
Although the currently available literature may support the idea that the presence
of esophageal varices is not an absolute contraindication to endoscopic management
of early esophageal neoplasia, these studies are limited by a small sample size and/or
short follow-up. Moreover, many clinicians still regard esophageal varices as a contraindication
to endoscopic therapy and thereby withhold patients from referral to a highly specialized
center for consideration of endoscopic therapy. In addition, there is no clear overview
of all possible preventative measures to mitigate the risk of bleeding. Therefore,
we performed a systematic review to evaluate the effectiveness and safety of ER for
early esophageal neoplasia in patients with esophageal varices, and to assess the
various prophylactic measures taken in these patients. Furthermore, based on the results
from the literature and our own clinical experiences, we have formulated several practical
recommendations to enable endoscopic treatment in this specific patient category.
Methods
Search strategy
A literature search was conducted in Cochrane Library, Embase, and Medline on August
20, 2024. The search strategy consisted of key terms and standardized index terms
related to esophageal varices and ER. No restrictions on publication language or publication
date were used. The full search algorithm is provided in Table 1s in the online-only Supplementary material. Reference lists of all included studies
were checked manually to identify additional relevant articles.
Study selection
After removal of duplicates, titles and abstracts were screened for eligibility by
two
independent authors (C.N.F. and L.S.B.). Any discrepancies regarding study inclusion
were
resolved through a consensus meeting. Thereafter, full-text review was independently
conducted by the same authors using the following predefined inclusion criteria: (1)
studies
including patients with esophageal varices and concurrent early esophageal neoplasia,
defined as high grade dysplasia or T1 cancer, and (2) studies reporting on the clinical
outcomes of EMR and/or ESD. Studies were included irrespective of design or follow-up
time.
Studies were excluded if they were (1) studies performed in the pediatric population
(age
<18 years), and (2) review articles, comments, or studies reported as conference
abstracts only. Studies that partly met the inclusion criteria were discussed individually
with a third assessor (B.L.A.M.W.) to resolve any disagreement on final inclusion.
As an addition to the available studies found through the literature search, we have
included unpublished data from an observational cohort, consisting of patients treated
between January 2014 and December 2023 at three Dutch centers with a tertiary referral
function for endoscopic management of early esophageal neoplasia. Please refer to
Text 1s for further details on this additional retrospective cohort.
Data extraction
Data from all included studies were extracted independently by two members of the
research team (C.N.F. and L.S.B.). All relevant end points were collected on a standardized
data collection form designed prior to data extraction. For each study, the following
variables were collected: first author, year of publication, country of origin, study
design, sample size, patient characteristics including age and Child–Pugh score, size
of varices, ER method, worst histopathology of ER specimen, and adverse event rate.
In addition, data on prophylactic measures to prevent bleeding were collected.
Quality assessment
The methodological quality of the studies was evaluated using the Joanna Briggs Institute
critical appraisal tools for case series and case reports (Table 2s ) [17 ]
[18 ]. This tool consists of 10 and 8 criteria for case series and case reports, respectively.
Eligible cohort studies were assessed using the appraisal tool for case series. The
quality of case series or cohort studies was classified as good when ≥9 criteria were
met, medium when 6–8 criteria were met, and poor quality when ≤5 criteria were met.
Similarly, case reports were classified as good, medium, or poor when meeting ≥7,
5–6, or ≤4 criteria, respectively.
Study end points
Based on the definitions shown in Table 3s , we defined several outcomes. The study end points regarding effectiveness included
rates of (1) en bloc resection, (2) radical resection, and (3) curative resection.
The study end points regarding safety included (1) the incidence of prophylactic measures
to prevent variceal bleeding, (2) the adverse event rate, including periprocedural
or delayed bleeding, perforation, and esophageal stricture requiring endoscopic dilation,
and (3) procedure-related mortality.
Ethics
This systematic review was conducted in accordance with the Preferred Reporting Items
for Systematic Review and Meta-Analyses (PRISMA) guidelines (Table 4s ) [19 ], and was registered in the international prospective register of systematic reviews
(PROSPERO CRD42024529958).
Statistical analysis
No statistical analysis was performed owing to the considerable heterogeneity of the
available literature. Instead, the results from the literature are presented in a
descriptive manner, using proportions to report categorical variables and medians
with minimum and maximum values (min–max) to describe the distribution of continuous
data.
Results
Search results
The initial search yielded 3382 records, of which 2371 records were screened for title
and abstract after removal of duplicates. A total of 51 potentially relevant publications
were retrieved for full-text analysis. After assessment for eligibility, 10 publications
were excluded based on the inclusion and exclusion criteria [20 ]
[21 ]
[22 ]
[23 ]
[24 ]
[25 ]
[26 ]
[27 ]
[28 ]
[29 ]. Finally, 41 publications were included ([Fig. 1 ]) [9 ]
[12 ]
[13 ]
[14 ]
[15 ]
[16 ]
[30 ]
[31 ]
[32 ]
[33 ]
[34 ]
[35 ]
[36 ]
[37 ]
[38 ]
[39 ]
[40 ]
[41 ]
[42 ]
[43 ]
[44 ]
[45 ]
[46 ]
[47 ]
[48 ]
[49 ]
[50 ]
[51 ]
[52 ]
[53 ]
[54 ]
[55 ]
[56 ]
[57 ]
[58 ]
[59 ]
[60 ]
[61 ]
[62 ]
[63 ]
[64 ]. Reference lists of all included studies were manually checked, but did not identify
additional eligible articles. In addition to the available studies found through the
literature search, the analysis hereafter also includes unpublished retrospective
data from our own clinical practice (Text 1s ).
Fig. 1 Flow diagram depicting study selection process through a literature search in
accordance with the Preferred Reporting Items for Systematic Review and Meta-Analyses
guidelines [19 ]. In addition to the six retrospective cohort studies retrieved through the
literature search, our own unpublished cohort was also included in the analysis
(supplementary material, Text 1).
Study characteristics and quality assessment
The included publications consisted of 1 prospective cohort study, 7 retrospective
cohort studies (including our own unpublished cohort), nine case series, and 25 case
reports
([Fig. 1 ]). The methodological quality of the case reports was mostly evaluated as good
(17/25; 68%), although eight case reports (8/25; 32%) were scored as medium quality
(Table 5s ). While all included cohort studies (8/8; 100%) were assessed
to be of good quality, the quality of the cases series varied widely, with three
publications (3/9; 33%) evaluated as good, five articles (5/9; 56%) as medium, and
one case
series as poor (1/9; 11%) (Table 6s ). The studies included a total
of 186 patients, with individual sample sizes ranging between 1 and 30 patients (Table 7s ). The Child–Pugh score was reported inconsistently, especially
in the case reports. Esophageal varices were predominantly classified as small, defined
as a
size of <5mm (128/186; 69%).
Prophylactic measures
As shown in Table 7s , the most widely adopted prophylactic measure was the eradication of varices prior
to ER through EBL in 72/186 patients (39%) ([Fig. 2 ]). Despite prior EBL, direct varix coagulation was still necessary in 12/72 patients
(17%) as a result of persistent varices. Other frequently described prophylactic measures
included EIS (22/186; 12%), direct varix coagulation (18/186; 10%), TIPS (9/186; 5%),
and a combination of EBL and EIS (5/186; 3%). The ligate-and-let-fall-off approach,
typically used when the lesion was located on top of a varix, was adopted in 12/186
patients (6%), preceded by EBL in three (3/12; 25%) of these cases. With this strategy,
the lesion and underlying varix are ligated using rubber bands while subsequent snaring
is omitted [12 ]. Eventually, the lesion will become necrotic and fall off spontaneously. In two
patients (1%), a temporary metal stent was placed after submucosal varices started
to bleed during the resection ([Fig. 3 ]). In the remaining 46/186 patients (25%), no preventative measures were described.
Fig. 2 Endoscopic resection after eradication of esophageal varices. a Endoscopic images of a patient with a C1M5 Barrett esophagus containing a
Paris 0-IIa lesion of 30 mm covering a third of the esophageal circumference. b As a result of post-alcoholic liver cirrhosis classified as
Child–Pugh A, small esophageal varices were detected and simultaneously treated with
endoscopic band ligation (EBL). c EBL was repeated after 2
weeks, during which multiple scars of the previous EBL session were seen. d Two weeks later, endoscopic submucosal dissection under
octreotide administration was technically successful, with histopathology revealing
a
moderately differentiated T1sm1 adenocarcinoma with no lymphovascular invasion, but
a
tumor-positive lateral resection margin. e The patient returned
1 week later with delayed bleeding due to an ulcer with a visible vessel at the
resection site, which was treated with an adrenaline injection and coagulation. f Three months later, the resection scar was completely regenerated
with squamous epithelium, without any signs of local recurrence. The patient then
entered strict endoscopic follow-up considering his comorbidities.
Fig. 3 Endoscopic resection in the presence of esophageal varices. a Endoscopic images of a patient with a C3M7 Barrett esophagus containing a
Paris 0-Is/IIa lesion of 30 mm covering half of the esophageal circumference. b During the same endoscopy, large esophageal varices were seen as
a result of post-alcoholic liver cirrhosis classified as Child–Pugh A. c Despite successful placement of a transjugular intrahepatic portosystemic
shunt, varices were still present during the endoscopic submucosal dissection (ESD)
procedure 3 months later. d During dissection under octreotide
administration, multiple collateral vessels were encountered and prophylactically
coagulated using the ESD knife and a hemostatic forceps. e As
an additional safety precaution, a fully covered metal stent was temporarily placed
to
prevent delayed variceal bleeding. f Three months after the
radical and curative resection of a moderately differentiated T1m3 adenocarcinoma
with
no lymphovascular invasion, the resection scar was almost completely regenerated with
squamous epithelium, without any signs of local recurrence.
ER outcomes
While 34/186 patients (18%) underwent EMR for esophageal neoplasia, the majority of
patients were treated with ESD (152/186; 82%) (Table 7s ). As shown in Table 8s , most patients were diagnosed with early squamous cell neoplasia (130/186; 70%).
In all studies, no procedure-related mortality was observed. Adverse events were seen
in 27/186 patients (14%) with one patient experiencing two separate complications
(Table 7s ). The most common adverse event was periprocedural bleeding (12/186; 6%), followed
by esophageal stricture requiring dilation (5/186; 3%), delayed bleeding (5/186; 3%),
infection (4/186; 2%), and perforation (2/186; 1%). Periprocedural bleeding was resolved
endoscopically in nine cases [51 ]
[58 ]
[60 ]
[62 ]
[64 ], but resulted in a prematurely terminated (n = 2) or macroscopically incomplete
(n = 1) resection in three patients [9 ]
[54 ]
[61 ]. In the five patients with delayed bleeding, the hemorrhage was resolved without
(n = 2) or with (n = 3) endoscopic intervention [57 ]
[58 ]
[63 ].
Histopathological outcomes
The histopathological characteristics of the ER specimens were reported inconsistently
throughout the included studies (Table 8s ). Based on the available data, the radical and curative resection rates were 137/160
(86%) and 59/82 (72%), respectively.
Discussion
Although the presence of esophageal varices is still considered by many endoscopists
to be an absolute contraindication to endoscopic therapy for early esophageal neoplasia,
the currently available literature evaluated in our systematic review demonstrates
that ER is safe and effective in this subset of patients, as reflected by the relatively
low adverse event rate and substantial curative resection rate. These favorable outcomes
of ER highlight that patients with early stage cirrhosis or noncirrhotic portal hypertension
should be considered for endoscopic therapy. Naturally, the risks of ER should be
balanced against the severity of comorbidities and the risk of dying from causes other
than esophageal cancer, including the complications of portal hypertension.
Unfortunately, the studies included in our review were inconsistent in their reporting
of
the histopathological outcomes after ER. Nevertheless, the curative resection rate
of 72%
derived from the available data is in line with the curative resection rates in the
general
Western population, which range from 72% to 96% for Barrett esophagus-related neoplasia,
and
from 46% to 74% for squamous cell neoplasia [5 ]
[65 ]
[66 ]
[67 ]
[68 ], demonstrating the effectiveness of ER in patients with esophageal varices. The
results from our systematic review are also comparable to those from other large studies
on
patients with liver cirrhosis or portal hypertension, which demonstrate a curative
resection
rate of 63%–76% [21 ]
[22 ].
Regarding safety, the available literature reported a combined periprocedural and
delayed bleeding risk of 9%. This incidence is slightly higher when compared with
the bleeding risk after ER in the general population, which varies from 2% to 5% [6 ]
[7 ]
[69 ]
[70 ]. Nonetheless, the majority of bleedings could be managed endoscopically, except
for three periprocedural bleedings, which resulted in an incomplete resection. When
combining the data from our systematic review, this is a mere 2% (3/186) of all ER
procedures. More importantly, the common message in all studies is that none of the
patients suffered life-threatening bleeding or death related to the ER procedure.
As a direct consequence of the low number of cases available in the literature, we
included all studies irrespective of ER technique. Even though this may have introduced
a certain level of heterogeneity, both EMR and ESD are currently advised for different
types of esophageal lesions and may therefore be complementary to each other [1 ]
[2 ]. Nevertheless, the present literature on patients with esophageal varices suggests
that ESD should be preferred over EMR. In our systematic review, only two cases of
delayed bleeding were observed after ESD (2/152; 1%), whereas EMR seemed to carry
a higher periprocedural and delayed bleeding risk (9% [3/34] and 3% [1/34], respectively).
This observed difference is in line with a recent, large observational study on patients
with liver cirrhosis or portal hypertension, which reported a lower bleeding risk
after ESD [21 ], but these results should be interpreted with caution given the limited number of
patients treated with EMR in both studies. Nevertheless, ESD holds the practical advantage
that any visible perforating varices can be directly coagulated during resection.
An important requirement to enable ER in this specific patient category is the application
of adequate prophylactic measures to prevent variceal bleeding. All studies provide
a variety of possibilities, making it hard to choose which prophylactic measure is
most suitable for each individual patient. Therefore, based on the available data
and expert opinion, we present a flow chart with a number of clinical scenarios to
help guide clinical decision making ([Fig. 4 ]). Primarily, irrespective of the size of the varices, we would advocate reducing
the splanchnic blood flow by initiating nonselective beta blockers prior to ER, assuming
no contraindications exist. In addition, we advise administration of a somatostatin
analog at the start of the ER procedure for at least 3 days, along with prophylactic
antibiotics to prevent peritonitis. Besides these drug options, we believe that repeated
EBL sessions should generally be the preferred method to completely eradicate the
varices prior to the ER procedure, under the condition that there is enough space
distally from the neoplastic lesion to ligate the varices. Otherwise, TIPS can be
considered in patients with large varices, while periprocedural coagulation of visible
varices remains an option in all cases. Although EIS is described in several case
reports as a possible preventative measure, we believe that the risk of inducing fibrosis
and thereby potentially complicating a re-resection is too high, especially considering
the adequate alternatives. In addition to the flow chart, we have formulated several
practical recommendations that may be helpful for endoscopists who perform ESD in
the presence of esophageal varices ([Table 1 ]).
Table 1 Practical recommendations for endoscopic submucosal dissection in the presence of
esophageal varices, with the aim of preventing variceal bleeding. These practical
recommendations are based on expert opinion derived from our own clinical experience.
Procedure step
Tips and tricks
Mucosal lifting
Start ejecting the lifting fluid just before the injection needle punctures the mucosa,
preferably using a colloid solution to decrease the blood flow in the varices by counter
pressure.
Inject a large volume of lifting fluid directly adjacent to the varices to ensure
adequate lifting of the mucosa above the varix, and to avoid cutting through a varix
during the mucosal incision.
Mucosal incision
Make a careful but complete mucosal incision outside the direct range of the varices to expose the varices but
prevent variceal bleeding at the same time.
Prophylactically coagulate and cut every exposed varix with either the endoscopic
submucosal dissection knife or a hemostatic forceps before proceeding to the submucosal
dissection.
Submucosal dissection
Dissect deeply through the submucosa by staying close to the muscular layer with the
aim of avoiding the varices by dissecting underneath them.
Fig. 4 Flow chart to help guide clinical decision making on prophylactic measures to mitigate
the bleeding risk of endoscopic resection (ER) in patients with concurrent early
esophageal neoplasia and esophageal varices. This flow chart is based on the limited
available literature as well as expert opinion. Besides the possible prophylactic
measures
indicated in green, administration of a somatostatin analog during ER should be considered
to reduce the splanchnic blood flow.
To the best of our knowledge, this is the first systematic review to evaluate the
safety and effectiveness of ER in patients with esophageal varices, and provides a
complete and thorough overview of the available literature on this subject according
to PRISMA guidelines [19 ]. Additionally, our synopsis of possible measures to prevent variceal hemorrhage
enabled us to formulate tips and tricks to help guide clinical decision making. Another
strength is the addition of our own unpublished observational cohort, for which data
were collected by a dedicated research fellow to ensure high quality. The expert setting
of this cohort may have contributed to the observed low adverse event rate, endorsing
the need to treat these challenging cases only in dedicated centers with the necessary
expertise.
This study also has several limitations that need to be acknowledged. The literature
search did not allow for a meta-analysis owing to the substantial heterogeneity of
the available studies, which mainly consisted of case reports and case series. The
between-study heterogeneity was partly the result of differences in ER techniques,
histopathological subtypes, applied prophylactic measures to mitigate bleeding, study
populations, and outcome parameters. In addition, the retrospective nature of most
included studies holds an inevitable risk of selection bias. As ER might have been
omitted in some patients due to the presence of varices or expected poor prognosis,
the currently available literature may possibly include a more favorable subset of
patients. Another limitation is the lack of an international, standardized protocol
for reducing the risk of variceal bleeding, resulting in a wide variety of prophylactic
measures initiated. On the other hand, this also highlights the multimodal character
of the “toolbox” available for facilitation of endoscopic therapy in this challenging
subset of patients. Regarding prophylactic measures, it should also be mentioned that
the periprocedural bleeding risk may have been underestimated as we considered the
use of direct varix coagulation as a preventative measure when, in reality, this option
may have also been applied to stop severe variceal bleeding. Finally, apart from four
publications that included more than 20 patients, most studies had a small sample
size, which is a direct consequence of the low prevalence of this specific patient
category.
To conclude, ER appears to be a safe and effective option in selected patients with
concurrent early esophageal neoplasia and esophageal varices, provided that adequate
prophylactic measures to prevent bleeding are tailored to the patient and applied
by an experienced endoscopist. Therefore, we strongly recommend referring patients
with concurrent early esophageal neoplasia and varices to a highly specialized center
for consideration of endoscopic therapy.
