American Journal of Gastroenterology
C ?2007 by Am. Coll. of Gastroenterology
Published by Blackwell Publishing
Banding Ligation Versus Beta-Blockers as Primary
Prophylaxis in Esophageal Varices: Systematic Review
of Randomized Trials
Lise L. Gluud, M.D., D.M.Sc.,1,2Sarah Klingenberg, M.Sc.,1Dimitrinka Nikolova, M.A.,1
and Christian Gluud, M.D., D.M.Sc.1
1Cochrane Hepato-Biliary Group, Copenhagen Trial Unit, Center for Clinical Intervention Research,
Rigshospitalet, Copenhagen, Denmark; and2Department of Internal Medicine F, Gentofte University Hospital,
OBJECTIVE: To compare banding ligation versus beta-blockers as primary prophylaxis in patients with
esophageal varices and no previous bleeding.
METHODS: Randomized trials were identified through electronic databases, reference lists in relevant articles,
and correspondence with experts. Three authors extracted data. Random effects meta-analysis and
metaregression were performed. The reported allocation sequence generation and concealment
were extracted as measures of bias control.
RESULTS:The initial searches identified 1,174 references. Sixteen trials were included. In 15 trials, patients
had high-risk varices. Three trials reported adequate bias control. All trials reported mortality for
banding ligation (116/573 patients) and beta-blockers (115/594 patients). Mortality in the two
treatment groups was not significantly different in the trials with adequate bias control (relative risk
1.22, 95% CI 0.84–1.78) or unclear bias control (RR 1.02, 95% CI 0.75–1.39). Trials with adequate
bias control found no significant difference in bleeding rates (RR 0.86, 95% CI 0.55–1.35). Trials
with unclear bias control found that banding ligation significantly reduced bleeding (RR 0.56, 95%
CI 0.41–0.77). Both treatments were associated with adverse events. In metaregression analyses,
the estimated effect of ligation was significantly more positive if trials were published as abstracts.
Likewise, the shorter the follow-up, the more positive the estimated effect of ligation.
CONCLUSIONS:Banding ligation and beta-blockers may be used as primary prophylaxis in high-risk esophageal
varices. The estimated effect of banding ligation in some trials may be biased and was associated
with the duration of follow-up. Further high-quality trials are still needed.
(Am J Gastroenterol 2007;102:2842–2848)
complications in portal hypertension (1, 2). During a 5-yr
period, about one in five patients with cirrhosis will die from
factors including the variceal size, wall tension, and local
varices and the risk of bleeding once the varices are present
are closely related to the portal pressure. Esophageal varices
develop once the hepatic venous pressure gradient exceeds
10 mmHg (6). A reduction in the hepatic venous pressure
gradient with at least 20% of the baseline pressure or to less
which decrease portal pressure may therefore help prevent
To access a continuing medical education exam for this article, please visit
Randomized trials and meta-analyses show that beta-
blockers reduce the number of bleeding events when used
as primary prevention in esophageal varices (8). Treatment
for primary prevention in medium-risk to high-risk varices
(9). However, the effect on mortality is debated and many
patients are unable to tolerate beta-blockers due to adverse
events (8–10). Banding ligation may be considered as an al-
ternative treatment option (11). A meta-analysis of 13 ran-
domized trials found that banding ligation reduced the risk
difference in mortality was found. Although these findings
are promising, the results of individual trials are conflicting
(12–14). Furthermore, the frequency of adverse events as-
sociated with banding ligation may be underestimated due
to the inclusion of selected patient groups and the fact that
only experienced operators participated in the trials (15–
17). A meta-analysis of eight trials found that beta-blockers
Banding Ligation Versus Beta-Blockers in Esophageal Varices 2843
were associated with a significantly higher number of ad-
verse events than banding ligation (18). If less experienced
endoscopists had participated in the trials, the frequency
of adverse events after banding ligation might have been
ing ligation versus beta-blockers for primary prevention in
versus any type of beta-blocker for patients with esophageal
varices and no bleeding history were included. The primary
outcome measure was all-cause mortality. Secondary out-
and adverse events. Three authors extracted data. Disagree-
ments were resolved through discussion before analyses.
Excluded trials were listed with the reason for exclusion. Pri-
mary authors of the included trials were contacted for addi-
in the published trial reports.
Electronic searches (19) for eligible trials were per-
formed in the following databases: The Cochrane Hepato-
Biliary Group Controlled Trials Register (March 2007), The
Cochrane Library (2007 Issue 1), MEDLINE (1950–March
2007), EMBASE (1960–March 2007), and Science Citation
Index Expanded (1945–March 2007). Additional trials were
identified through scanning of reference lists in relevant pa-
pers and conference proceedings. We also wrote to experts
and pharmaceutical companies to inquire about additional
Methodological quality was defined as the control of bias
that may be assessed through the reported randomization
methods (20). The reported randomization, which primarily
reflects selection bias, was classified according to previous
definitions based on the reported allocation sequence gener-
ation and concealment (21). The allocation sequence genera-
number table, or similar) or unclear (not described). The al-
location concealment was classified as adequate (central in-
dependent unit, serially numbered opaque sealed envelopes,
on-site locked computer, or identically appearing numbered
drug bottles or containers prepared by an independent phar-
macist) or unclear (not described). No quasi-randomized tri-
als or other trials with inadequate allocation sequence gener-
ation or allocation concealment were identified.
Double blinding may also be associated with bias, but we
expected that few (if any) trials would be double blind due
to the nature of the interventions. Accordingly, we did not
did, however, extract whether the trials reported blinded out-
come assessment. We also extracted data on the duration of
follow-up, the numbers and reasons for dropouts and with-
drawals, and whether preset sample size calculations were
reported and reached. We also gathered data on the charac-
teristics of included patients, treatments, and trials including
country of origin and publication status.
The analyses were performed in RevMan version 4.2 (The
Nordic Cochrane Center, Copenhagen, Denmark) and Stata
version 8.0 for Windows (Stata Corp., College Station, TX).
Due to expected clinical and methodological heterogeneity,
the primary meta-analyses were performed using random ef-
fects models stratified by the control of bias (i.e., with sep-
arate subgroups including trials with adequate or unclear
bias control). The results are presented as relative risks (RR)
with 95% confidence intervals (CI) and I2as a measure of
intertrial heterogeneity. All analyses were performed using
ized. For patients with missing data, carry forward of the
last observed response was used. We were unable to perform
meta-analyses on adverse events, as data were not presented
for both treatment groups. Adverse events are therefore pre-
sented as proportions. For the primary outcome measure, a
worst-case scenario analysis in which patients with missing
data were counted as treatment failures was performed to as-
sess potential attrition bias. Random effects metaregression
was performed to assess sources of intertrial heterogeneity.
The covariates included bias control (adequate or unclear),
publication status (full paper article or abstract), duration of
follow-up (in months), and mean dose of beta-blocker. Fixed
by the significant predictors of intertrial heterogeneity were
was dichotomized by the median duration of follow-up (20
hoc subgroup analyses assessing the influence of follow-up
and publication status on this outcome were performed. For
failures was performed to assess potential attrition bias. We
also performed two additional post hoc analyses to evaluate
the influence of trials with certain characteristics. In one of
these analyses, a trial on patients waiting for a liver trans-
plantation was excluded (22). In the second post hoc analysis
we excluded two trials that were terminated early (23, 24).
The electronic search strategy generated 1,174 potentially
relevant references. No additional references for published
with experts or pharmaceutical companies. After reading the
titles and abstracts, we excluded 1,125 references that were
duplicates or clearly did not refer to randomized trials and
retrieved 49 references for further assessment. Subsequently,
33 references assessing prevention of rebleeding (N = 14)
or interventions not included in the present review, e.g., scle-
published in 1998–2007 were included (22–37).
2844 Gluud et al.
Ten of the included trials were published as full paper ar-
ticles and six were published in abstract form. Eight trials
were conducted in the United States or Europe. The remain-
ico, and Taiwan. On average, the trials included 77 patients
(range 24–152, total number 1,167). All patients had cirrho-
sis diagnosed based on clinical, biochemical, or histological
did not specify whether the included patients had high-risk
esophageal varices. The criteria used for assessing the risk of
bleeding included red color signs (red wale markings, cherry
at least one third of the esophageal lumen, or pseudotumor-
ous varices. Other trials classified varices as high risk if the
diameter was at least 5 mm or at least 2 mm plus at least
one red color sign. The reported exclusion criteria were con-
patients ranged from 39 to 62 yr. The duration of follow-up
ranged from 10 to 40 months.
In total, 573 patients were randomized to banding ligation
and 594 patients were randomized to beta-blockers in the
included trials. Banding ligation was performed with con-
ventional or multiband ligators under benzodiazepine seda-
tion with or without local anesthesia. Banding ligation was
repeated with 3–4 wk intervals until the varices were erad-
icated. The proportion of patients who achieved complete
eradication ranged from 71–100% (reported in 11 trials). On
average, recurrence of the varices was identified in 13% of
patients who underwent routine control endoscopy.
All trials assessed nonselective beta-blockers (propanolol,
daily. The dose was adjusted to achieve a 20–25% reduction
in heart rate or to achieve a heart rate less than 60 beats per
minute. The mean daily dose after the dose adjustment was
73 mg (range 60–114 mg).
Three trials reported both adequate allocation sequence
generation and adequate allocation concealment. Four tri-
als reported that envelopes were used, but did not specify
whether the envelopes were serially numbered, opaque, or
sealed. None of the trials reported blinded outcome assess-
ment or data analyses. Seven trials reported the number and
reasons for losses to follow-up. In total, 11 patients random-
ized to banding and 13 patients randomized to beta-blockers
were lost to follow-up (reported in five trials). The remaining
trials gave the impression that no patients were lost, but did
not specifically describe that there were no losses to follow-
Two trials were terminated prematurely after interim anal-
yses. The trials did not report the number of preplanned in-
terim analyses, stopping rules, or whether an independent
data monitoring and safety committee was involved. In one
of the trials (23), interim analyses suggested that the trial had
inadequate statistical power. The observed failure rate was
10% instead of the expected 25%. The trial was planned to
include 180 patients but was terminated after inclusion of 31
patients. The other trial (24) was terminated after a planned
The trial was planned to include 104 patients but was termi-
nated after inclusion of 62 patients.
All 16 trials reported all-cause mortality. In trials with ade-
quate bias control, mortality was 34% among patients ran-
domized to banding ligation and 27% among patients ran-
domized to beta-blockers. In trials with unclear bias control
mortality was 15% for patients on banding ligation and 16%
for patients on beta-blockers. As shown in Figure 1, the mor-
tality in the two treatment groups was not significantly dif-
ferent in trials with adequate (RR 1.22, 95% CI 0.84–1.78,
I2= 13%) or unclear (RR 1.02, 95% CI 0.75–1.39, I2= 0%)
Upper Gastrointestinal Bleeding
and Bleeding-Related Mortality
In the trials with adequate bias control, incident upper gas-
trointestinal bleeding (variceal or nonvariceal) occurred for
18% of patients randomized to banding ligation versus 20%
of patients randomized to beta-blockers (N = 3 trials). In tri-
als with unclear bias control, the numbers were 12% versus
banding ligation versus beta-blockers for primary prophylaxis in
esophageal varices. The forest plot shows subgroups of trials with
adequate and unclear bias control.
Meta-analysis of mortality in randomized trials on
Banding Ligation Versus Beta-Blockers in Esophageal Varices2845
23% in the two treatment groups (N = 13 trials). Trials with
1.35, I2= 0%) whereas trials with unclear bias control found
95% CI 0.41–0.77, I2= 0%).
In the three trials with adequate bias control, the bleeding-
related mortality and variceal bleeding rates were not signif-
icantly different among patients randomized to banding lig-
ation versus beta-blockers (RR 1.29, 95% CI 0.61–2.72 and
RR 0.80, 95% CI 0.50–1.28). In trials with unclear bias con-
trol bleeding-related mortality was not significantly different
in the two treatment groups (RR 0.68, 95% CI 0.31–1.53,
N = 8 trials) whereas banding ligation significantly reduced
variceal bleeding (RR 0.57, 95% CI 0.38–0.85, N = 11 tri-
als). No significant intertrial heterogeneity was detected in
any of the meta-analyses on secondary outcome measures.
Ten trials reported one or more adverse events. Both banding
ligation and beta-blockers were associated with serious and
nonserious adverse events. Two trials using the conventional
banding devices reported cases of esophagus perforation and
severe bleeding during insertion of the overtube. On average,
leading to cases of anemia, dysphagia, or chest pain. Among
patients randomized to beta-blockers, treatment withdrawal
due to noncompliance or adverse events occurred for 12%
and dose reductions were necessary in 20%. Other adverse
events associated with beta-blockers included fatigue (21%),
hypotension (15%), dyspnea (13%), nausea (9%), dizziness
of prerenal azotemia, worsening of peripheral vascular dis-
ease, and exanthema were also described.
Sensitivity and Subgroup Analyses
In the worst-case scenario analysis, the all-cause mortality
was not significantly different among patients randomized to
banding ligation or beta-blockers in trials with adequate bias
control (RR 1.11, 95% CI 0.87–1.43) or trials with unclear
bias control (RR 1.00, 95% CI 0.75–1.33). Accordingly, the
worst-case scenario analyses confirmed the overall results of
our primary analyses.
The only significant predictors of the estimated treatment
effect in the metaregression analyses were the duration of
follow-up (regression coefficient [r] = 0.71, P = 0.032) and
publicationstatus(r=0.74, P =0.029).Noneoftheremain-
treatmenteffect(allocationconcealment P =0.23,allocation
sequencegeneration P =0.36,andmeandoseofbeta-blocker
P = 0.55).
In the subgroup analyses, trials with at least 20 months
of follow-up found that the mortality was 31% for patients
randomized to banding ligation versus 25% for patients ran-
domized to beta-blockers (RR 1.23, 95% CI 0.97–1.56, I2=
banding ligation versus beta-blockers for primary prophylaxis in
esophageal varices. The forest plot shows subgroups of trials with
at least 20 months follow-up and less than 20 months follow-up.
Meta-analysis of mortality in randomized trials on
0%, Fig. 2). The trials with less than 20 months of follow-
up found considerably lower mortality than trials with longer
follow-up, but found no significant difference between the
two treatments (8% vs 12%, RR 0.64, 95% CI 0.38–1.07, I2
= 0%). Likewise, mortality rates were not significantly dif-
ferent between the banding ligation versus beta-blockers in
N = 10 trials) or abstracts (8% vs 12%, RR 0.67, 95% CI
0.37–1.23, N = 6 trials).
In the trials with at least 20 months of follow-up, incident
among patients randomized to banding ligation versus beta-
blockers (18% vs 22%, RR 0.78, 95% CI 0.57–1.06). The
trials with less than 20 months of follow-up found that band-
ing ligation significantly reduced bleeding (8% vs 22%, RR
0.38, 95% CI 0.24–0.61). Banding ligation reduced bleeding
significantly compared with beta-blockers in trials published
as full paper articles (15% vs 22%, RR 0.68, 95% CI 0.51–
0.91, N = 10 trials) and abstracts (9% vs 22%, RR 0.43, 95%
CI 0.25–0.73, N = 6 trials).
The trial including patients waiting for a liver transplanta-
the two remaining trials with adequate bias control found no
significant difference in mortality between treatments (RR
1.32, 95% CI 0.75–2.32).
The two trials that were terminated early reported unclear
bias control. After these trials were excluded, the remaining
2846Gluud et al.
in mortality between banding ligation and beta-blockers (RR
1.01, 95% CI 0.73–1.38).
The present review compared the effects of banding liga-
tion versus nonselective beta-blockers as primary prophy-
laxis for patients with esophageal varices and no history
of bleeding. Three of the 16 trials included were classified
as having adequate bias control based on the reported ran-
domization methods. The remaining trials did not describe
randomization methods and were classified as having un-
clear bias control. Most of the included trials therefore had a
potential risk of selection bias (20). In the trials with ad-
equate bias control, no significant differences on mortal-
ity or bleeding rates were found. In the trials with unclear
bias control banding ligation significantly reduced bleed-
ing, but the effect did not translate to improved all-cause
or bleeding-related mortality rates. In subgroup analyses, tri-
als with less than 20 months of follow-up and published as
abstracts were more positive towards the benefit of band-
ing ligation than trials with longer follow-up and trials pub-
lished as full paper articles, respectively. The overall evi-
dence suggests that both banding ligation and beta-blockers
may be considered as primary prophylaxis for patients with
esophageal varices and no history of bleeding. The subgroup
analyses suggest that the beneficial effect of banding ligation
on bleeding in some trials may be influenced by selection or
with sufficient follow-up is still needed to determine which
of the two treatments should be recommended as standard.
in trials with the longest duration of follow-up. The trials
with the shortest duration of follow-up were more positive
towards the effect of banding ligation than the trials with
longer follow-up. The reported recurrence rates for varices
after banding ligation suggests that the effect of banding lig-
ation may diminish with time. However, the results may also
simply reflect random or systematic errors such as selection
or publication bias. An alternative explanation is that beta-
blockers may have long-term beneficial effects that are seen
after treatment for a certain time.
Theoretically, beta-blockers may reduce the risk of recur-
ring varices after banding ligation. However, a randomized
trial addressing this question found no significant difference
in mortality or bleeding between banding ligation plus pro-
pranolol versus banding ligation alone (38). The size and the
number of events in the trial suggest that a clinically relevant
difference may have been overlooked. Additional evidence is
therapy with banding ligation plus beta-blockers.
The assessment of adverse events in meta-analyses is of-
ten difficult. One of the reasons is the lack of consensus re-
garding the classification of serious and nonserious adverse
tion are classified as serious (39). None of the included trials
used these criteria. The number of reported adverse events
in the trials would have been higher if the guidelines had
been followed. The lack of agreement among trials is likely
to explain the differences in reported adverse event rates. For
first-degree heart block after treatment with beta-blockers as
serious adverse events (24). In the same trial development of
similar trial (33) registered five serious adverse events in the
form of bleeding occurring after banding ligation. The risk
ators that were used. Conventional banding devices involved
insertion of an overtube, which increased the risk of bleeding
new multiband devices are significantly safer to use than the
conventional device (40). The size of the varices may also
be related to the risk of complications. A randomized trial
comparing banding ligation with multiband ligators with no
intervention as primary prophylaxis had to be stopped early
due to unacceptably high bleeding rates (41). The complica-
tion rate might have been related to the inclusion of patients
with small esophageal varices. Similar results were found in
a large randomized trial showing that sclerotherapy should
not be used for prevention of bleeding from small varices
(42). The results suggest that banding ligation used as pri-
mary prophylaxis should only be considered for patient with
The fact that no significant difference is found between
two treatments does not necessarily mean that the treatments
are equally effective. In several trials, clinically important
treatment effects are overlooked due to inadequate statistical
experimental treatment (e.g., banding) is not worse than the
established treatment (e.g., beta-blockers) is the equivalence
(or noninferiority) trial. When calculating the required sam-
ple size, trials include a prespecified delta value (the equiva-
lence margin) for the acceptable difference between treat-
“as effective and safe” as beta-blockers (31, 32, 37). These
conclusions are debatable as none of the trials were designed
to establish noninferiority.
beta-blockers may be considered as primary prophylaxis for
patients with high-risk esophageal varices and no history of
this did not translate into reduced mortality. Evidence of se-
lection and publication bias favoring banding ligation was
identified and the beneficial effect of banding was primar-
ily seen in the trials with the shortest follow-up. Therefore,
high-quality randomized trials with sufficient follow-up are
still needed. In future trials, the outcome measures including
Banding Ligation Versus Beta-Blockers in Esophageal Varices 2847
ably following recommendations such as the BAVENO cri-
teria (44), ICH-GCP recommendations, and the CONSORT
A protocol on beta-blockers for primary prevention in
esophageal varices drafted by Wendong Chen was used as
part of the background material in the development of the
protocol for the present review. Wendong Chen was unable
article is a condensed version of a systematic review, which
will be published in The Cochrane Library.
Reprint requests and correspondence: Lise L. Gluud, M.D.,
D.M.Sc., Cochrane Hepato-Biliary Group, Copenhagen Trial Unit,
Centre for Clinical Intervention Research, Copenhagen University
Hospital, Panum Institute, 2200-DK, Copenhagen, Denmark.
Received June 15, 2007; accepted July 18, 2007.
1. Le Moine O, Adler M, Bourgeois N, et al. Factors related
to early mortality in cirrhotic patients bleeding from varices
and treated by urgent sclerotherapy. Gut 1992;33:1381–5.
in a university hospital in Sweden 1994–1999. Scand J Gas-
4. Gluud C, Henriksen JH, Nielsen G, et al. Prognostic indica-
tors in alcoholic cirrhotic men. Hepatology 1988;8:222–7.
5. MacMathuna P, Westaby D, Williams R. Taking tension
out of the portal system: An approach to the management
of portal hypertension in the 1990s. Scand J Gastroenterol
6. Arguedas MR. The critically ill liver patient: The variceal
bleeder. Semin Gastroint Dis 2003;14:34–8.
7. Turnes J, Garcia-Pagan JC, Abraldes JG, et al. Pharmaco-
logical reduction of portal pressure and long-term risk of
first variceal bleeding in patients with cirrhosis. Am J Gas-
effects on gastrointestinal hemorrhage in cirrhotic patients.
World J Gastroenterol 2003;9:1836–9.
9. Lebrec D, Vinel JP, Dupas JL. Complications of portal hy-
pertension in adults: A French consensus. Eur J Gastroen-
terol Hepatol 2005;17:403–10.
10. Garcia-Pagan JC, Villanueva C, Vila MC, et al. Isosorbide
mononitrate in the prevention of first variceal bleed in pa-
tients who cannot receive beta-blockers. Gastroenterology
11. Triantos C, Vlachogiannakos J, Manolakopoulos S, et al. Is
banding ligation for primary prevention of variceal bleed-
ing as effective as beta-blockers, and is it safe? Hepatology
12. Andreani T, Poupon RE, Balkau BJ, et al. Preventive ther-
apy of first gastrointestinal bleeding in patients with cir-
rhosis: Results of a controlled trial comparing propra-
nolol, endoscopic sclerotherapy and placebo. Hepatology
13. Avgerinos A, Armonis A, Manolakopoulos S, et al. Endo-
scopic sclerotherapy plus propranolol versus propranolol
alone in the primary prevention of bleeding in high risk
cirrhotic patients with esophageal varices: A prospective
multicenter randomized trial. Gastrointest Endosc 2000;51:
14. Vlachogiannakos J, Goulis J, Patch D, et al. Review arti-
cle: Primary prophylaxis for portal hypertensive bleeding in
cirrhosis. Al Pharmacol Ther 2000;14:851–60.
haemostasis in the management of bleeding peptic ulcers.
Postgrad Med J 2002;78:347–51.
16. Parente F, Anderloni A, Bargiggia S, et al. Outcome of non-
variceal acute upper gastrointestinal bleeding in relation to
17. Lim CH, Vani D, Shah SG, et al. The outcome of suspected
upper gastrointestinal bleeding with 24-hour access to up-
per gastrointestinal endoscopy: A prospective cohort study.
18. Khuroo MS, Khuroo NS, Farahat KL, et al. Meta-analysis:
Endoscopic variceal ligation for primary prophylaxis of
oesophageal variceal bleeding. Aliment Pharmacol Ther
19. Gluud LL, Klingenberg SLF, Nikolova D, et al. Endoscopic
therapy or beta-blockers for primary prevention of bleed-
20. Gluud LL. Bias in clinical intervention research. Am J Epi-
21. Kjaergard LL, Villumsen J, Gluud C. Reported method-
ological quality and discrepancies between large and
small randomised trials in meta-analyses. Ann Intern Med
22. Norberto L, Polese L, Cillo U, et al. A randomized study
comparing ligation with propranolol for primary prophy-
laxis of variceal bleeding in candidates for liver transplan-
tation. Liver Transpl 2007;13:1212–78.
23. Thuluvath PJ, Maheshwari A, Jagannath S, et al. A ran-
domized controlled trial of beta-blockers versus endoscopic
band ligation for primary prophylaxis: A large sample size
is required to show a difference in bleeding rates. Dig Dis
24. Jutabha R, Jensen DM, Martin P, et al. Randomized
study comparing banding and propranolol to prevent initial
variceal hemorrhage in cirrhotics with high-risk esophageal
varices. Gastroenterology 2005;128:870–81.
25. Chen CY, Sheu MZ, Su SY. Prophylactic endo-
scopic variceal ligation (EVL) with multiple band liga-
tor for oesophageal varices [Abstract]. Gastroenterology
26. De BK, Ghoshal UC, Das T, et al. Endoscopic variceal liga-
tion for primary prophylaxis of oesophageal variceal bleed:
Preliminary report of a randomized controlled trial. J Gas-
troenterol Hepatol 1999;14:220–4.
27. Sarin SK, Lamba GS, Kumar M, et al. Comparison of endo-
scopic ligation and propranolol for the primary prevention
of variceal bleeding. N Engl J Med 1999;340:988–93.
28. de la Mora JG, Farca B, Uribe M, et al. Ligation vs
propanolol for primary prophylaxis of variceal bleeding us-
ing a multiple band ligator and objective measurements of
treatment adequacy: Preliminary results [Abstract]. Gas-
29. Song JH, Shin JW, Kim IH, et al. Prophylactic banding
ligation of high-risk esophageal varices in patients on the
waiting list for liver transplantation: An interim analysis
[Abstract]. J Hepatol 2000;32:41.
2848 Gluud et al.
domized trial between the prophylactic endoscopic variceal
ligation and propranolol administration for prevention of
first bleeding in cirrhotic patients with high risk esophageal
varices [Abstract]. J Hepatol 2002;36:38A.
ing band ligation, propranolol, and isosorbide mononitrate.
32. Lo GH, Chen WC, Chen MH, et al. Endoscopic liga-
tion vs. nadolol in the prevention of first variceal bleed-
ing in patients with cirrhosis. Gastrointest Endosc 2004;59:
33. Schepke M, Kleber G, Nurnberg D, et al. Ligation versus
in cirrhosis. Hepatology 2004;40:65–72.
34. Drastich, Lata J, Petrtyl J, et al. Endoscopic variceal band
variceal bleeding in patients with liver cirrhosis [Abstract].
J Hepatol 2005;42:79.
35. Psilopoulos D, Galanis P, Goulas S, et al. Endoscopic
variceal ligation vs. propranolol for prevention of first
variceal bleeding: A randomized controlled trial. Eur J Gas-
troenterol Hepatol 2005;17:1111–7
36. Abdelfattah MH, Rashed MA, Elfakhry AA, et al. Endo-
scopic variceal ligation versus pharmacological treatment
for primary prophylaxis of variceal bleeding: A randomised
study [Abstract]. J Hepatol 2006;44:S83.
37. Lay CS, Tsai YT, Lee FY, et al. Endoscopic variceal lig-
ation versus propranolol in prophylaxis of first variceal
bleeding in patients with cirrhosis. J Gastroenterol Hepa-
38. Sarin SK, Wadhawan M, Agarwal SR, et al. Endoscopic
ligation alone in primary prophylaxis of variceal bleeding.
Am J Gastroenterol 2005;100:797–804.
Group. Media: Parexel Barnett, 1997.
40. Wong T, Pereira SP, McNair A, et al. A prospective, ran-
domized comparison of the ease and safety of variceal lig-
ationusing a multiband vs. a conventional ligation device.
41. Triantos C, Vlachogiannakos J, Armonis A, et al. Primary
prophylaxis of variceal bleeding in cirrhotics unable to take
beta-blockers: A randomized trial of ligation. Al Pharmacol
42. The PROVA Study Group. Prophylaxis of first hemorrhage
from esophageal varices by sclerotherapy, propranolol or
both in cirrhotic patients: A randomized multicenter trial.
43. Le Henanff A, Giraudeau B, Baron G, et al. Quality of re-
44. de Franchis R. What have we accomplished? Portal hyper-
tension III: Proceedings of the third Baverno international
consensus workshop on definitions, methodology and ther-
apeutic strategies 2001:1–12.
CONFLICT OF INTEREST
Guarantor of the article: Lise Lotte Gluud, M.D.,
Specific author contributions: Lise Lotte Gluud drafted the
review and performed the statistical analyses. Sarah Louise
Klingenberg wrote the search strategy and performed the
electronic searches. Lise Lotte Gluud, Sarah Louise Klin-
vision of the review, and have approved of the final version
Potential competing interests: None.