LIVER FAILURE AND LIVER DISEASE

Propranolol Plus Placebo Versus Propranolol Plus

Isosorbide-5-Mononitrate in the Prevention of a First
Variceal Bleed: A Double-Blind RCT
Juan Carlos Garcı́a-Pagán,1 Rosa Morillas,2 Rafael Bañares,3 Agustin Albillos,4 Candido Villanueva,5 Carme Vila,6
Joan Genescà,7 Manuel Jimenez,8 Manuel Rodriguez,9 Jose Luis Calleja,10 Joaquin Balanzó,5 Fernando Garcı́a-Durán,3
Ramón Planas,2 Jaume Bosch,1 and the Spanish Variceal Bleeding Study Group

Nonselective ␤-blockers are very effective in preventing first variceal bleeding in patients with cirrho-
sis. Treatment with isosorbide-5-mononitrate (IS-MN) plus propranolol achieves a greater reduction
in portal pressure than propranolol alone. The present multicenter, prospective, double-blind, ran-
domized, controlled trial evaluated whether combined drug therapy could be more effective than
propranolol alone in preventing variceal bleeding. A total of 349 consecutive cirrhotic patients with
gastroesophageal varices were randomized to receive propranolol ⴙ placebo (n ⴝ 174) or proprano-
lol ⴙ IS-MN (n ⴝ 175). There were no significant differences in the 1- and 2-year actuarial probability
of variceal bleeding between the 2 groups (propranolol ⴙ placebo, 8.3% and 10.6%; propranolol ⴙ
IS-MN, 5% and 12.5%). The only independent predictor of variceal bleeding was a variceal size
greater than 5 mm. However, among patients with varices greater than 5 mm (n ⴝ 196), there were no
significant differences in the incidence of variceal bleeding between the 2 groups. Survival was also
similar. Adverse effects were significantly more frequent in the propranolol ⴙ IS-MN group due to a
greater incidence of headache. There were no significant differences in the incidence of new-onset or
worsening ascites or in impairment of renal function. In conclusion, propranolol effectively prevents
variceal bleeding. Adding IS-MN does not further decrease the low residual risk of bleeding in patients
receiving propranolol. However, the long-term use of this combination drug therapy is safe and may
be an alternative in clinical conditions associated with a greater risk of bleeding. (HEPATOLOGY 2003;37:
1260-1266.)

of those patients with large esophageal varices will bleed at

See Editorial on Page 1254
2 years.1 Nonselective ␤-blockers are an effective therapy
for the prophylaxis of first variceal bleeding in patients
with cirrhosis.2-4 This beneficial effect of ␤-blockers is

V
ariceal bleeding is one of the more severe compli-
cations of patients with cirrhosis and portal hy-
pertension. More than 40% of cirrhotic patients
already have esophageal varices at diagnosis. Nearly 30%
mainly due to its ability to reduce portal pressure.5,6 Com-
plete protection from variceal bleeding is achieved when
the portal pressure gradient is reduced to less than 12 mm

Abbreviation: IS-MN, isosorbide-5-mononitrate.

From the 1Hepatic Hemodynamic Laboratory, Liver Unit, Institut de Malaties Digestives, Hospital Clinic, Institut d’Investigacions Biomediques August Pi i Sunyer,
University of Barcelona, Barcelona; 2Department of Gastroenterology, Hospital Universitari Germans Trias i Pujol, Badalona; 3Hospital Gregorio Marañón, Madrid;
4Department of Gastroenterology, Hospital Ramón y Cajal, University of Alcalá de Henares, Madrid; 5Department of Gastroenterology, Hospital Sant Pau, Barcelona;
6Department of Gastroenterology, Hospital del Mar, Barcelona; 7Internal Medicine Department, Hospital Vall d’Hebró, Universitat Autonoma of Barcelona, Barcelona; 8Hospital

Virgen Macarena, Sevilla; 9Department of Gastroenterology, Hospital Central de Asturias, Oviedo; and 10Clı́nica Puerta de Hierro, Madrid, Spain.
Received January 14, 2003; accepted March 3, 2003.
Supported by grants 02/0692 and 02/0739 from Fondo de Investigaciones Sanitarias and SAF 2000-0219 and a grant from the Instituto de Salud Carlos III (CO 3/02).
Isosorbide-5-mononitrate and placebo were kindly supplied by Boehringer Mannhein (Barcelona, Spain).
For the Spanish Variceal Bleeding Study Group: Hospital Clı́nic, Barcelona: Juan G. Abraldes, Angels Escorsell, Eduardo Moitinho, Claudio Rodriguez, Josep Llach, and Joan Rodes;
Hospital Germans Trias i Pujol, Badalona: Jaume Boix; Hospital Gregorio Marañón, Madrid: Marta Casado; Clı́nica Puerta de Hierro, Madrid: Francisco Domper and Jose Luis Martinez;
Hospital de Sant Pau, Barcelona: Josep Miñana and German Soriano; Hospital del Mar, Barcelona: Judith Marquez, Maria Dolors Gimenez, and Ricard Solà; Hospital Vall
d⬘Hebró, Barcelona: Antonio González, Juan Carlos Lopez-Talavera, and Rafael Estebán; Hospital Virgen Macarena, Sevilla: Angel Piñar; Hospital Central de Asturias, Oviedo:
Nieves Sotorrio and Luis Rodrigo; Hospital Marques de Valdecilla, Santander: Joaquin de la Peña and Fernando Pons; Hospital Ramón y Cajal, Madrid: Luis Ruiz-del-Arbol.
Address reprint requests to: Juan Carlos Garcı́a-Pagán, M.D., Hepatic Hemodynamic Laboratory, Liver Unit, IMD, Hospital Clinic, Villaroel, 170, Barcelona 08036,
Spain. E-mail: jcgarcia@medicina.ub.es; fax: (34) 932279856.
Copyright © 2003 by the American Association for the Study of Liver Diseases.
0270-9139/03/3706-0009$30.00/0
doi:10.1053/jhep.2003.50211

1260
HEPATOLOGY, Vol. 37, No. 6, 2003
Hg and an almost complete protection with reductions of
20% or more from baseline values.7-9 The association of
isosorbide-5-mononitrate (IS-MN) with ␤-blockers en-
hances the reduction in portal pressure.10-13 An open trial
has suggested that this combined therapy may be more
effective than ␤-blockers alone in preventing first variceal
bleeding.14 The aim of the present study was to evaluate,
in a double-blind, randomized, multicenter clinical trial,
whether propranolol ⫹ IS-MN is more effective than
propranolol alone in the prevention of first variceal bleed-
ing in patients with cirrhosis and esophageal varices.
Patients and Methods
Selection of Patients. Cirrhotic patients aged be-
tween 18 and 70 years with endoscopy-proven esophageal
or gastric varices of any size and no previous variceal
bleeding referred to any of the 12 participating hospitals
between February 1996 and July 1998 were considered
eligible for the study. According to Baveno, esophageal
varices larger than 5 mm were considered large and oth-
erwise were considered small.15,16 Patients with contrain-
dications to ␤-blockers (severe chronic pulmonary
obstructive disease, asthma, severe insulin-dependent di-
abetes mellitus, heart failure, grade II atrioventricular
block, sinus bradycardia ⬍50 bpm, aortic stenosis, pe-
ripheral arterial disease, arterial hypotension with systolic
pressure ⬍85 mm Hg) or contraindications to long-act-
ing nitrates (glaucoma), a known hepatocellular carci-
noma, a Child-Pugh score greater than 12, or a
concomitant disease with reduced life expectancy were
not included in the study. According to these criteria, 377
patients were considered eligible for the study from a
group of 565 patients initially considered for the study. A
total of 105 of these patients had contraindications to
␤-blockers, and the remaining 83 patients had either a
hepatocellular carcinoma (n ⫽ 25), a Child-Pugh score
greater than 12 (n ⫽ 19), a severe nonhepatic disorder
(n ⫽ 22), or refused to enter the study (n ⫽ 17). In those
patients eligible for the study, oral propranolol therapy
was initiated. The dose was increased stepwise every 2 to 3
days to achieve a 25% reduction in heart rate or a heart
rate less than 55 bpm. Twenty-eight patients could not
tolerate ␤-blockers and were excluded, and 349 patients
tolerated the drug. These 349 patients were randomized
to receive either placebo or IS-MN in addition to pro-
pranolol (Fig. 1).
Randomization was performed by a central telephone
call to an independent center. The assigned treatments
were generated by computer. Double-blindness was
maintained until the study was completed and the data
were entered into a final database.
GARCI´A-PAGA´ N ET AL. 1261
Fig. 1. Enrollment of patients included in the study.
The protocol and procedures were scheduled to con-
form to the guidelines of good clinical practice in clinical
trials. The ethical committee of each participating hospi-
tal approved the study.
Clinical Assessment and General Management. Be-
fore randomization, a full clinical history, physical examina-
tion, electrocardiogram, chest radiograph, and laboratory
test results were obtained in all patients. Treatment with
IS-MN was then initiated by administering 20 mg
IS-MN (half pill) or placebo at night and increased step-
wise up to a maximum of 40 mg twice a day unless symp-
tomatic hypotension or severe headache occurred. After
titration, patients were visited at month 1, month 3, and
then every 3 months until the end of follow-up. Each visit
included anamnesis, physical examination, and blood test
including hematology as well as liver and renal biochem-
istry. At each visit, compliance was assessed by pill count-
ing and patients were dispensed medication to follow
treatment until next control. Patients were followed up
until variceal bleeding, death, or liver transplantation oc-
curred or up to May 1997.
End Points. The main end point of the study was the
development of first variceal bleeding. Secondary end
points were bleeding from any source, development of
adverse effects, and the effect of treatment on renal func-
tion, ascites, and survival. All patients were instructed to
go to the hospital whenever they experienced melena or
hematemesis. If upper gastrointestinal bleeding was con-
firmed, an emergency endoscopy was performed. Variceal
bleeding was diagnosed when varices were actively bleed-
ing or had stigmata of recent bleeding and/or if fresh
blood was observed in the stomach and varices were the
only potential source of bleeding. Variceal bleeding was
treated primarily by means of emergency sclerotherapy
associated or not with vasoactive drugs (somatostatin or
1262 GARCI´A-PAGA´ N ET AL. HEPATOLOGY, June 2003

terlipressin). For mortality analysis, patients were fol- Table 1. Clinical Characteristics of the 2 Groups of Patients
lowed up for 6 weeks after bleeding. at Inclusion in the Study
Calculation of Sample Size. According to meta- Propranolol Propranolol
ⴙ Placebo ⴙ IS-MN
analysis, patients with esophageal varices treated with (n ⴝ 174) (n ⴝ 175)
␤-blockers have a 16% chance of bleeding at 2 years of
Age (y) 56 ⫾ 11 57 ⫾ 10
follow-up.17,18 The sample size needed to detect a 10% Sex (M/F) 116/58 125/50
reduction in the risk of bleeding was calculated as 150 Etiology (alcohol/hepatitis C virus/
patients in each treatment group using a 2-sided test with others) (n) 54/84/36 68/85/22
Active alcoholism (n) 21 24
80% power at a significance level of .05.
Child’s grade (A/B/C) (n) 114/45/15 111/52/12
Statistical Analysis. The SPSS statistical package Ascites (n) 57 54
(SPSS Inc., Chicago, IL) was used for this analysis. Data Total bilirubin level (mg/dL) 1.9 ⫾ 1.5 2 ⫾ 1.5
are reported as means with SDs. Categorical variables Serum albumin level (mg/dL) 37 ⫾ 6 37 ⫾ 7
Prothrombin ratio (%) 72 ⫾ 20 70 ⫾ 18
were compared using Fisher’s exact test and continuous Creatinine level (mg/dL) 0.9 ⫾ 0.2 0.9 ⫾ 0.2
variables with the unpaired Student’s t test (or the non- Heart rate (bpm) 78 ⫾ 11 77 ⫾ 12
parametric Mann-Whitney rank sum test for unpaired Mean arterial pressure (mm Hg) 92 ⫾ 12 92 ⫾ 12
Variceal size ⬎5 mm (n) 92 104
data). Actuarial probability curves were constructed using Presence of red signs (n) 23 26
the Kaplan-Meier method and compared with the log- Dosage of propranolol (mg/d) 67 ⫾ 46 77 ⫾ 65
rank test. Stepwise Cox regression analysis was used to
NOTE. Data shown as mean ⫾ SD. All P values were nonsignificant (P ⬎ .10).
identify independent predictors for the first variceal
bleeding and death. Relative hazards and the 95% confi-
dence interval were given. Statistical significance was es-
tablished at a P value less than .05. An intent-to-treat placebo; NS]). According to attendance to the scheduled
strategy was used in the analysis of the results. visits and pill count, the remaining 302 patients had more
than 95% compliance to the study drug until the end of
follow-up.
Results Eleven patients in the propranolol ⫹ placebo group
and 20 in the propranolol ⫹ IS-MN group were lost to
Patients. A total of 174 patients were randomized to
receive propranolol ⫹ placebo and 175 patients to receive follow-up after a mean follow-up of 3.3 ⫾ 3 months
propranolol ⫹ IS-MN. A total of 196 patients had large (range, 1 day to 12 months) (after withdrawal because of
varices (92 receiving propranolol ⫹ placebo and 104 re- adverse effects, after withdrawal of consent, or without
ceiving propranolol ⫹ IS-MN). There were no significant any known cause: propranolol ⫹ placebo, 4, 2, and 5
differences at study entry in clinical or endoscopic char- patients, respectively; propranolol ⫹ IS-MN, 10, 2, and 8
acteristics (Table 1). The degree of ␤-blockade was similar patients, respectively). These 31 patients were censored at
in the 2 groups of patients, as shown by a similar and the time of the last visit. The remaining patients were
marked reduction in heart rate (propranolol ⫹ placebo, followed up until variceal bleeding, death, or liver trans-
from 78 ⫾ 11 to 61 ⫾ 10 bpm; propranolol ⫹ IS-MN, plantation occurred or until the end of the study. Mean
from 77 ⫾ 13 to 62 ⫾ 8 bpm; both P ⬍ .01) and by a follow-up was 16.2 ⫾ 8 months in the propranolol ⫹
similar dose of propranolol (propranolol ⫹ placebo, 67 ⫾ placebo group and 16 ⫾ 8.5 months in the propranolol ⫹
46 mg/d; propranolol ⫹ IS-MN, 76 ⫾ 65 mg/d; NS). IS-MN group (NS).
After titration, 87% of patients treated with IS-MN Bleeding. Fifteen patients in the propranolol ⫹ pla-
reached the full dose of the drug versus 92% of the pa- cebo group and 15 patients in the propranolol ⫹ IS-MN
tients treated with placebo (NS). However, the mean dose group had the first variceal bleeding during follow-up
of IS-MN was slightly lower than that of placebo (73 ⫾ (NS). There were no significant differences in the 1- and
19 vs. 77 ⫾ 12 mg/d; P ⫽ .05). The study drug was 2-year actuarial probability of first variceal bleeding (pro-
discontinued early in 32 patients because of adverse ef- pranolol ⫹ placebo, 8.3% and 10.6%; propranolol ⫹
fects (Fig. 1 and Table 1) (22 patients receiving propran- IS-MN, 5% and 12.5%; NS) (Fig. 2). On univariate anal-
olol ⫹ IS-MN vs. 10 patients receiving propranolol ⫹ ysis, the size of the varices and hematocrit at inclusion
placebo; P ⬍ .05). Fifteen patients withdrew consent (ear- were the only 2 variables significantly related to an in-
ly after beginning treatment in 5 patients [3 receiving creased risk of variceal bleeding. When these variables as
propranolol ⫹ IS-MN and 2 receiving propranolol ⫹ well as treatment received and Child score were entered
placebo] and 3-14 months later in 10 patients [5 receiving on a multivariate analysis using Cox’s model, only a
propranolol ⫹ IS-MN and 5 receiving propranolol ⫹ variceal size greater than 5 mm was independently associ-
HEPATOLOGY, Vol. 37, No. 6, 2003 GARCI´A-PAGA´ N ET AL. 1263

Fig. 2. Actuarial probability of remaining free of a first variceal bleed- Fig. 3. Actuarial probability of remaining free of a first variceal bleed-
ing. ing in patients with large varices (⬎5 mm).

Changes in Variceal Size. In 135 patients (69 in the

ated with an increased risk of variceal bleeding (hazard propranolol ⫹ placebo group and 66 in the proprano-
ratio, 7.5; 95% confidence interval, 2.2-24; P ⬍ .01). lol ⫹ IS-MN group), variceal size was endoscopically as-
Severity of Bleeding. Severity of bleeding was similar sessed at 12 months of follow-up. Changes in variceal size
in patients treated with propranolol ⫹ placebo as in those were similar in the 2 groups of patients (propranolol ⫹
receiving propranolol ⫹ IS-MN, as evaluated by hemat- placebo vs. propranolol ⫹ IS-MN; no change, 74% vs.
ocrit at admission (29% ⫾ 6% vs. 29% ⫾ 7%; NS), 77%; from large to small, 19% vs. 14%; from small to
transfusion requirements (2.3 ⫾ 1.8 vs. 2.8 ⫾ 2.7 packed large, 7% vs. 9%; NS).
red blood cell units; NS), shock at admission (3 of 15 vs. Adverse Effects. Adverse effects were significantly
3 of 14; NS), treatment required (endoscopic ⫾ vasoac- more frequent in the propranolol ⫹ IS-MN group (39%)
tive drugs, 10 vs. 10 [NS]; vasoactive drugs, 4 vs. 3 [NS]; than in the propranolol ⫹ placebo group (26%; P ⬍ .01)
transjugular intrahepatic portosystemic shunt, 1 vs. 2 due to a significantly greater incidence of headache (Table
[NS]), and death (2 of 15 vs. 4 of 15; NS). 2). There were no significant differences in the 1- and
Incidence of Bleeding in Patients With Large Var- 2-year actuarial probability of new or worsening ascites
ices. A total of 196 patients had large varices (92 treated (propranolol ⫹ placebo, 14% and 26%; propranolol ⫹
with propranolol ⫹ placebo and 104 with propranolol ⫹ IS-MN, 16.5% and 22%; NS) or in changes of renal
IS-MN). The incidence of first variceal bleeding was function (Table 3).
higher in these patients than in the overall population (27 At 12 months of follow-up, a mild but significant im-
of 196 in patients with large varices vs. 30 of 349 in the provement in Child-Pugh score was observed in both
overall population) but, again, there were no significant groups of patients (Table 3).
differences in the 1- and 2-year actuarial probability of Survival. There were no significant differences in the
first bleeding between patients treated with proprano- 1- and 2-year actuarial probability of survival (proprano-
lol ⫹ placebo or propranolol ⫹ IS-MN (propranolol ⫹ lol ⫹ placebo, 96% and 89%; propranolol ⫹ IS-MN,
placebo, 13% and 17%; propranolol ⫹ IS-MN, 7% and 95% and 88%; NS; Fig. 4).
20%; NS) (Fig. 3).
Bleeding From Any Cause. Six additional patients
Table 2. Adverse Effects Observed in Patients Treated With
bled from causes other than varices; 3 were in the pro- Propranolol ⴙ IS-MN or Propranolol ⴙ Placebo
pranolol ⫹ placebo group (2 secondary to portal hyper-
Propranolol Propranolol
tensive gastropathy and one to a peptic ulcer) and 3 in the ⴙ Placebo ⴙ IS-MN P
propranolol ⫹ IS-MN group (2 secondary to portal hy- Overall (%) 45/174 (26) 69/175 (39) ⬍.01
pertensive gastropathy and one of unknown origin). Headache 15 39 ⬍.01
There were no significant differences in the actuarial Mild hypotension/asthenia 15 15 NS
probability of gastrointestinal bleeding of any cause. Heart failure/dyspnea 7 9 NS
Bradycardia 2 3 NS
Results analyzing either first variceal bleeding or gas- Angina/stroke 2 3 NS
trointestinal bleeding of any cause were similar when the Cutaneous reaction 4 0 NS
analysis was repeated according to treatment received Adverse effect requiring withdrawal
of IS-MN or placebo (%) 10 (6) 22 (12) ⬍.05
(data not shown).
1264 GARCI´A-PAGA´ N ET AL. HEPATOLOGY, June 2003

Table 3. Effects of Propranolol ⴙ Placebo and Propranolol ⴙ IS-MN on Liver and Renal Function in the Control of Ascites
and on Systemic Hemodynamics in Patients With 12 Months of Follow-up
Propranolol ⴙ Placebo Propranolol ⴙ IS-MN

Baseline 12 Months Baseline 12 Months

Child-Pugh score 6 ⫾ 1.6 5.5 ⫾ 1.1* 6.2 ⫾ 1.5 5.7 ⫾ 1.1*

Blood urea nitrogen level (mg/dL) 16 ⫾ 8 17 ⫾ 8 16 ⫾ 7 17 ⫾ 81
Creatinine level (mg/dL) 0.9 ⫾ 0.2 1 ⫾ 0.3 0.9 ⫾ 0.2 0.9 ⫾ 0.3
New or worsening ascites (%) 17 15.5
Systolic arterial pressure (mm Hg) 130 ⫾ 20 125 ⫾ 22* 129 ⫾ 19 125 ⫾ 19*
Diastolic arterial pressure (mm Hg) 75 ⫾ 11 71 ⫾ 12* 73 ⫾ 10 72 ⫾ 14
Mean arterial pressure (mm Hg) 93 ⫾ 12 89 ⫾ 13* 92 ⫾ 11 89 ⫾ 14

*P ⬍ .01 vs. baseline.

There were also no significant differences in the causes
of death (Table 4). On multivariate analysis, only the
Child-Pugh score, variceal bleeding on follow-up, male
sex, and age were found to be independent predictors of
increased mortality, whereas treatment received was not
related to survival.
Discussion
Nonselective ␤-blockers such as propranolol or nado-
lol have been shown to significantly reduce the incidence
of variceal bleeding by 40% to 50% during follow-up.17
In addition, ␤-blockade is associated with an almost sig-
nificant reduction in mortality and with a significant re-
duction in bleeding-related deaths.17 Because of this,
␤-blockers are currently an accepted treatment for the
prevention of first variceal bleeding in patients with cir-
rhosis and esophageal varices.19 However, although the
residual risk of bleeding is relatively low (about 15% at 2
years), ␤-blockers do not protect all treated patients,2
probably due to an insufficient reduction in the hepatic
venous pressure gradient.7,20
The addition of IS-MN to nonselective ␤-blockers has
been suggested to improve its clinical efficacy14,21 because
of the ability of the combined therapy to promote a
greater reduction in portal pressure than ␤-blockers ad-
ministered alone.10,11 Several clinical studies suggest that
this is true for the secondary prevention of variceal bleed-
ing,9,22-26 and many centers use the combination of non-
selective ␤-blockers with IS-MN in patients receiving
drug therapy for the prevention of variceal rebleeding.
However, whether the combination of ␤-blockers plus
IS-MN enhances the efficacy of ␤-blockers administered
alone in preventing first variceal bleeding has only been
evaluated so far in a single nonblinded, randomized, con-
trolled trial.14,21
The present large double-blind, multicenter, random-
ized, controlled trial has failed to confirm a benefit of the
combined pharmacologic therapy in preventing first
variceal bleeding. The reason for such a discrepancy is not
clear. However, several differences between both studies
should be considered. As mentioned, contrary to the pre-
vious study, our study was performed as a double-blind
investigation, therefore minimizing the risk of investiga-
tor bias. In addition, the significantly greater proportion
of alcoholic patients in the study by Merkel et al. com-
pared with our study (54% vs. 35%), many of whom were
actively drinking at the time of randomization (20% vs.
13% in our study), may also have influenced the results.
The lack of beneficial effect of adding IS-MN cannot
be argued to be due to a low risk of first bleeding in the
propranolol-treated group. Thus, the 10.6% 2-year actu-
arial risk of first bleeding in our patients treated with
propranolol ⫹ placebo is well within the 95% confidence
interval (7%-23%) or the interquartile rate (4%-21%) of

Table 4. Causes of Death

Propranolol ⴙ Placebo Propranolol ⴙ IS-MN

Variceal bleeding 2 4
Liver failure 4 6
Hepatocellular carcinoma 3 2
Others 2 2
Total (%) 11/174 (6.3) 14/179 (8)

Fig. 4. Actuarial probability of survival. NOTE. All P values were nonsignificant.

HEPATOLOGY, Vol. 37, No. 6, 2003
first bleeding in the 11 previously reported studies (6 in-
cluding only large varices and 5 including varices of any
size).2 It is important to mention that, in the only previ-
ous double-blind randomized trial, the rate of first bleed-
ing in patients treated with propranolol was 4%. Again,
when only large varices were considered, the 2-year actu-
arial risk of first bleeding in the standard treatment group
was 17%, which is equal to the reported value of random-
ized, controlled trials performed in patients with large
varices.
It is important to note that the different results also
cannot be explained by the fact that we included patients
with varices of any size whereas only patients with large
varices were included in the study by Merkel et al., be-
cause no additional beneficial effect of the combined ther-
apy was observed when only patients with large varices
were analyzed. In addition, cirrhotic patients included in
our study had a slightly better liver function than those in
the study by Merkel et al., as indicated by a lower Child-
Pugh score and a significantly lower prevalence of ascites
at inclusion (32% vs. 42%). A greater reduction in he-
patic venous pressure gradient after administration of
propranolol has been observed in patients with good liver
function compared with those with poor liver function.1
We have previously shown that IS-MN only produces a
marginal portal hypotensive effect in patients with a good
response to propranolol.27 Unfortunately, this study did
not incorporate measurements in hepatic venous pressure
gradient verifying this point. However, a similar rate of
good response, defined as a more than 20% reduction in
hepatic venous pressure gradient or less than 12 mm Hg,
was observed after nadolol was administered alone or in
association with IS-MN (55% vs. 60%) in a recent hemo-
dynamic study performed in patients submitted to pri-
mary prophylaxis with a liver function similar to ours.7 If
patients with a greater risk of variceal bleeding can be
detected and treated, the possible beneficial effect of the
combined therapy should probably be reevaluated.
An additional finding from our study is the fact that, as
previously shown,10,12,28 propranolol combined with
IS-MN is safe and does not produce deleterious effects on
renal function. This is an important finding because the
combination of ␤-blockers and IS-MN is increasingly
used in the prevention of rebleeding.29
The clinical efficacy of nonselective ␤-blockers in pri-
mary prophylaxis has also recently been tested against
endoscopic band ligation in 5 randomized, controlled tri-
als: 3 published in full30-32 and 2 still as abstracts.33,34
Only one study showed a significant benefit with endo-
scopic band ligation.30 However, in this study, the bleed-
ing rate under propranolol was unusually high, which
questions the validity of its conclusions in other settings.
GARCI´A-PAGA´ N ET AL. 1265
No differences in the risk of bleeding or survival were
found in the meta-analysis performed excluding this out-
lier trial.35 Therefore, available evidence does not show a
superiority of endoscopic band ligation over ␤-blockers in
primary prophylaxis.
In summary, the present study confirms that propran-
olol is highly effective in preventing first variceal bleeding.
The low residual risk of first variceal bleeding observed in
patients receiving propranolol was not further decreased
by adding IS-MN, so this drug combination cannot be
advocated for the primary prophylaxis of variceal bleed-
ing. However, the long-term use of propranolol ⫹
IS-MN neither deteriorates renal function nor worsens
the control of ascites or increases mortality, indicating
that this therapeutic combination is safe and may be a
good alternative in other clinical conditions associated
with a higher risk of bleeding, such as in the prevention of
variceal rebleeding.
References
1. Bosch J, Garcia-Pagan JC. Complications of cirrhosis. I. Portal hyperten-
sion. J Hepatol 2000;32:141-156.
2. D’Amico G, Pagliaro L, Bosch J. Pharmacological treatment of portal
hypertension: an evidence-based approach. Semin Liver Dis 1999;19:475-
505.
3. de Franchis R, Pascal JP, Ancona E, Burroughs AK, Henderson M, Fleig
W, Groszmann R, et al. Definitions, methodology and therapeutic strate-
gies in portal hypertension. A Consensus Development Workshop,
Baveno, Lake Maggiore, Italy, April 5 and 6, 1990. J Hepatol 1992;15:
256-261.
4. Grace ND, Groszmann RJ, Garcia-Tsao G, Burroughs AK, Pagliaro L,
Makuch RW, Bosch J, et al. Portal hypertension and variceal bleeding: an
AASLD single topic symposium. HEPATOLOGY 1998;28:868-880.
5. Bosch J, Mastai R, Kravetz D, Bruix J, Gaya J, Rigau J, Rodes J. Effects of
propranolol on azygos venous blood flow and hepatic and systemic hemo-
dynamics in cirrhosis. HEPATOLOGY 1984;4:1200-1205.
6. Feu F, Bordas JM, Garcia-Pagan JC, Bosch J, Rodes J. Double-blind
investigation of the effects of propranolol and placebo on the pressure of
esophageal varices in patients with portal hypertension. HEPATOLOGY
1991;13:917-922.
7. Merkel C, Bolognesi M, Sacerdoti D, Bombonato G, Bellini B, Bighin R,
Gatta A. The hemodynamic response to medical treatment of portal hy-
pertension as a predictor of clinical effectiveness in the primary prophylaxis
of variceal bleeding in cirrhosis. HEPATOLOGY 2000;32:930-934.
8. Escorsell A, Bordas JM, Castaneda B, Llach J, Garcia-Pagan JC, Rodes J,
Bosch J. Predictive value of the variceal pressure response to continued
pharmacological therapy in patients with cirrhosis and portal hyperten-
sion. HEPATOLOGY 2000;31:1061-1067.
9. Bureau C, Peron JM, Alric L, Morales J, Sanchez J, Barange K, Payen JL,
et al. “A la carte” treatment of portal hypertension: adapting medical ther-
apy to hemodynamic response for the prevention of bleeding. HEPATOL-
OGY 2002;36:1361-1366.
10. Garcia-Pagan JC, Feu F, Bosch J, Rodes J. Propranolol compared with
propranolol plus isosorbide-5-mononitrate for portal hypertension in cir-
rhosis. A randomized controlled study. Ann Intern Med 1991;114:869-
873.
11. Garcia-Pagan JC, Navasa M, Bosch J, Bru C, Pizcueta P, Rodes J. En-
hancement of portal pressure reduction by the association of isosorbide-5-
mononitrate to propranolol administration in patients with cirrhosis.
HEPATOLOGY 1990;11:230-238.
1266 GARCI´A-PAGA´ N ET AL.
12. Merkel C, Gatta A, Donada C, Enzo E, Marin R, Amodio P, Torboli P, et
al. Long-term effect of nadolol or nadolol plus isosorbide-5-mononitrate
on renal function and ascites formation in patients with cirrhosis. GTIP
Gruppo Triveneto per l’Ipertensione Portale. HEPATOLOGY 1995;22:808-
813.
13. Merkel C, Sacerdoti D, Bolognesi M, Enzo E, Marin R, Bombonato G,
Angeli P, et al. Hemodynamic evaluation of the addition of isosorbide-5-
mononitrate to nadolol in cirrhotic patients with insufficient response to
the beta- blocker alone. HEPATOLOGY 1997;26:34-39.
14. Merkel C, Marin R, Sacerdoti D, Donada C, Cavallarin G, Torboli P,
Amodio P, et al. Long-term results of a clinical trial of nadolol with or
without isosorbide mononitrate for primary prophylaxis of variceal bleed-
ing in cirrhosis. HEPATOLOGY 2000;31:324-329.
15. de Franchis R. Updating consensus in portal hypertension: report of the
Baveno III Consensus Workshop on definitions, methodology and thera-
peutic strategies in portal hypertension. J Hepatol 2000;33:846-852.
16. de Franchis R. Developing consensus in portal hypertension. J Hepatol
1996;25:390-394.
17. D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a
meta-analytic review. HEPATOLOGY 1995;22:332-354.
18. Poynard T, Cales P, Pasta L, Ideo G, Pascal JP, Pagliaro L, Lebrec D.
Beta-adrenergic-antagonist drugs in the prevention of gastrointestinal
bleeding in patients with cirrhosis and esophageal varices. An analysis of
data and prognostic factors in 589 patients from four randomized clinical
trials. Franco-Italian Multicenter Study Group. N Engl J Med 1991;324:
1532-1538.
19. de Franchis R. Developing consensus in portal hypertension. J Hepatol
1996;25:390-394.
20. Groszmann RJ, Bosch J, Grace ND, Conn HO, Garcia-Tsao G, Navasa
M, Alberts J, et al. Hemodynamic events in a prospective randomized trial
of propranolol versus placebo in the prevention of a first variceal hemor-
rhage [see comments]. Gastroenterology 1990;99:1401-1407.
21. Merkel C, Marin R, Enzo E, Donada C, Cavallarin G, Torboli P, Amodio P,
et al. Randomised trial of nadolol alone or with isosorbide mononitrate for
primary prophylaxis of variceal bleeding in cirrhosis. Gruppo-Triveneto per
L’ipertensione portale (GTIP). Lancet 1996;348:1677-1681.
22. Villanueva C, Balanzo J, Novella MT, Soriano G, Sainz S, Torras X, Cusso
X, et al. Nadolol plus isosorbide mononitrate compared with sclerotherapy
for the prevention of variceal rebleeding. N Engl J Med 1996;334:1624-
1629.
23. Villanueva C, Minana J, Ortiz J, Gallego A, Soriano G, Torras X, Sainz S,
et al. Endoscopic ligation compared with combined treatment with nado-
lol and isosorbide mononitrate to prevent recurrent variceal bleeding.
N Engl J Med 2001;345:647-655.
HEPATOLOGY, June 2003
24. Gournay J, Masliah C, Martin T, Perrin D, Galmiche JP. Isosorbide
mononitrate and propranolol compared with propranolol alone for the
prevention of variceal rebleeding. HEPATOLOGY 2000;31:1239-1245.
25. Escorsell A, Banares R, Garcia-Pagan JC, Gilabert R, Moitinho E, Piqueras
B, Bru C, et al. TIPS versus drug therapy in preventing variceal rebleeding
in advanced cirrhosis: a randomized controlled trial. HEPATOLOGY 2002;
35:385-392.
26. Patch D, Sabin CA, Goulis J, Gerunda G, Greenslade L, Merkel C, Bur-
roughs AK. A randomized, controlled trial of medical therapy versus en-
doscopic ligation for the prevention of variceal rebleeding in patients with
cirrhosis. Gastroenterology 2002;123:1013-1019.
27. Albillos A, Garcia-Pagan JC, Iborra J, Bandi JC, Cacho G, Perez-Paramo
M, Escorsell A, et al. Propranolol plus prazosin compared with propranolol
plus isosorbide-5-mononitrate in the treatment of portal hypertension.
Gastroenterology 1998;115:116-123.
28. Morillas RM, Planas R, Cabre E, Galan A, Quer JC, Feu F, Garcia Pagan
JC, et al. Propranolol plus isosorbide-5-mononitrate for portal hyperten-
sion in cirrhosis: long-term hemodynamic and renal effects. HEPATOLOGY
1994;20:1502-1508.
29. Groszmann RJ, Garcia-Tsao G. Endoscopic variceal banding vs. pharma-
cological therapy for the prevention of recurrent variceal hemorrhage: what
makes the difference? Gastroenterology 2002;123:1388-1391.
30. Sarin SK, Lamba GS, Kumar M, Misra A, Murthy NS. Comparison of
endoscopic ligation and propranolol for the primary prevention of variceal
bleeding. N Engl J Med 1999;340:988-993.
31. De BK, Ghoshal UC, Das T, Santra A, Biswas PK. Endoscopic variceal
ligation for primary prophylaxis of oesophageal variceal bleed: preliminary
report of a randomized controlled trial. J Gastroenterol Hepatol 1999;14:
220-224.
32. Lui HF, Stanley AJ, Forrest EH, Jalan R, Hislop WS, Mills PR, Finlayson
ND, et al. Primary prophylaxis of variceal hemorrhage: a randomized
controlled trial comparing band ligation, propranolol, and isosorbide
mononitrate. Gastroenterology 2002;123:735-744.
33. Chen CY, Sheu MZ, Su SY. Prophylactic endoscopic variceal ligation for
esophageal varices [Abstract]. Gastroenterology 1998;114:1224A.
34. Song H, Shin JW, Kim HI, Choi J, Lim CY, Kim JW, Roe IH. A prospec-
tive randomized 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 HEPATOLOGY 2000;
32(Suppl 2):41A.
35. Bosch J, D’Amico G, Garcı́a-Pagán JC. Portal hypertension. In: Schiff ER,
Sorrell MF, Maddrey WC, eds. Schiff’s Diseases of The Liver. 9th ed.
Philadelphia: Lippincott Williams & Wilkins, 2002:429-486.