Scandinavian Journal of Gastroenterology.

2012; Early Online, 18

ORIGINAL ARTICLE

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

Carvedilol or propranolol in portal hypertension? A randomized

comparison
LISE HOBOLTH1,2, SREN MLLER2, HENNING GRNBK3, KLAUS ROELSGAARD3,
FLEMMING BENDTSEN1 & ERIK FELDAGER HANSEN1
1

Department of Gastroenterology, Hvidovre University Hospital, Faculty of Health Sciences, Copenhagen University,
Hvidovre, Denmark, 2Department of Clinical Physiology and Nuclear Medicine, Hvidovre Hospital, Faculty of Health
Sciences, University of Copenhagen, Hvidovre, Denmark, and 3Department of Medicine V (Hepatology and
Gastroenterology), Aarhus University Hospital, Aarhus, Denmark

Abstract
Objectives. Carvedilol is a non-selective b-blocker with intrinsic anti-a1-adrenergic activity, potentially more effective than
propranolol in reducing hepatic venous pressure gradient (HVPG). We compared the long-term effect of carvedilol and
propranolol on HVPG and assessed whether the acute response to oral propranolol predicted the long-term HVPG response
on either drug. Material and methods. HVPG was measured in 38 patients with cirrhosis and HVPG 12 mm Hg at baseline
and then again 90 min after an oral dose of 80 mg propranolol. Patients were double-blinded randomized to either carvedilol
(21 patients) or propranolol (17 patients) and after 90 days of treatment HVPG measurements were repeated. Results. HVPG
decreased by 19.3 16.1% (p < 0.01) and by 12.5 16.7% (p < 0.01) in the carvedilol and propranolol groups, respectively,
with no signicant difference between treatment regimens (p = 0.21). Although insignicant, an acute decrease in HVPG of
12% was the best cut-off value to predict long-term HVPG response to propranolol when using ROC curve analysis.
Conclusions. This randomized study showed that carvedilol is at least as effective as propranolol on HVPG after longterm administration. Furthermore, a predictive value of an acute propranolol test on HVPG could not be conrmed.

Key Words: acute HVPG response, carvedilol, portal pressure, propranolol, randomized trial

Introduction
Bleeding from esophageal varices is a serious complication of cirrhosis and portal hypertension and is
associated with a high early mortality [1,2]. The
risk of variceal bleeding depends on the degree of
portal pressure and bleeding rarely occurs in patients
with a hepatic venous pressure gradient (HVPG)
below 1012 mm Hg [3]. Studies have shown that
reducing the HVPG with the non-selective b-blocker
propranolol to values below 12 mm Hg or by 20%
from baseline signicantly decreases the risk of bleeding and mortality [46]. However, up to 60% of
patients do not achieve such reductions in HVPG
after treatment with propranolol and face an increased

risk of bleeding [68]. Accordingly, other pharmacological treatments have been approached. Carvedilol
is a non-selective b-blocker with intrinsic anti-a1adrenergic activity. The effect on portal pressure is
induced by b1-blockade that reduces cardiac output
and b2-blockade that elicits splanchnic vasoconstriction, and a reduction in intrahepatic resistance due to
the a1-blocking effect [9]. Carvedilol effectively
decreases portal pressure after both acute and longterm administration [1018]. It remains to be established, whether the long-term effects of carvedilol on
portal pressure is superior to the effects of propranolol. One study has shown carvedilol to be superior to
propranolol [15] in reducing HVPG after longterm administration, whereas another study could

Correspondence: Lise Hobolth, Department of Gastroenterology, Hvidovre University Hospital, Faculty of Health Sciences, Kettegaard All 30,
DK-2650 Hvidovre, Denmark. Tel: +4536326540. Fax: +4536473311. E-mail: [email protected]

(Received 4 December 2011; revised 1 February 2012; accepted 2 February 2012)

ISSN 0036-5521 print/ISSN 1502-7708 online  2012 Informa Healthcare
DOI: 10.3109/00365521.2012.666673

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

L. Hobolth et al.

not demonstrate any signicant difference in effect on

HVPG [16].
It has been suggested that the acute HVPG response to b-blockers predicts the long-term response;
this acute response could therefore be helpful in selecting patients who may benet from pharmacological
treatment, but the results are conicting [1923].
The aims of this randomized study were to compare
the long-term effects of carvedilol with those of propranolol on HVPG in patients with cirrhosis and
portal hypertension and to assess whether the acute
response to an oral dose of propranolol predicts the
long-term effect of either drug on portal pressure.
Materials and methods
The study was performed as a double-blind, randomized, controlled trial with patient recruitment from
two specialized centers between September 2003 and
August 2009.
Patients with cirrhosis and clinical suspicion and
endoscopic signs of portal hypertension were evaluated for inclusion. Inclusion criteria were age
>18 years and HVPG 12 mm Hg. Exclusion criteria
were a) ChildPugh score >12; b) hepatic encephalopathy > grade II; c) hepatorenal syndrome or
serumcreatinine 2.26 g/dl; d) contraindications
to b-blockers such as: atrioventricular block, insulindependent diabetes mellitus, asthma or chronic
obstructive pulmonary disease; e) treatment with
vasoactive drugs or blood transfusion during the
week before inclusion; f) malignancy or life expectancy
less than 3 months. Informed consent was obtained
from all participants. Treatment allocation was by
block randomization of four, with an equal number
for carvedilol and propranolol. The encapsulation and
blinding of the medication was done by the pharmacy
at Aarhus University Hospital. The medication was
encapsulated in gelatine to ensure identical appearance, smell and taste and stored in numbered containers. The sequence was concealed until termination
of the study. The treatment codes were stored in
sealed opaque envelopes. Each capsule contained
either 3.125 mg carvedilol, Dimetone, Roche,
Basel, Switzerland or 40 mg propranolol DAK,
Nycomed Denmark A/S, Roskilde, Denmark.
Patient population
We screened 61 patients for inclusion in the study;
14 patients did not reach randomization due to either
HVPG < 12 mm Hg (n = 10), development of insulindependent diabetes between screening and baseline
hemodynamic measurements (n = 2), death before
baseline hemodynamic measurements (n = 1) or

unsuccessful catheterization because of obesity

(n = 1). We randomized 47 patients: 24 to treatment
with carvedilol and 23 to treatment with propranolol.
In the carvedilol group, three patients were withdrawn
after randomization due to skin rash (n = 1), chest
pain (n = 1) and early death from hematemesis, day
13 (n = 1). This patient had resumed alcohol abuse
after baseline catheterization with lack of compliance
to study medication. In the propranolol group, six
patients were withdrawn or dropped out after randomization because of variceal bleeding, day
53 (n = 1), dizziness (n = 2), impotence and noncompliance (n = 1), headache (n = 1) and one patient
was excluded after termination of the study due to
concurrent intake of an angiotensin-converting
enzyme (ACE) inhibitor. All together, this left
38 patients with cirrhosis and portal hypertension
to be analyzed: 21 randomized to carvedilol and
17 randomized to propranolol. Clinical and biochemical characteristics of the patients in the individual
groups are shown in Table I.
Hemodynamic investigations
All patients underwent a liver vein catheterization
with hemodynamic investigation to assess the degree
of portal hypertension. A Swan-Ganz 7F (n = 28) or
Cournand (n = 10) catheter was guided via the femoral vein under local analgesia to the hepatic veins.
The HVPG was determined as the wedged hepatic
venous pressure (WHVP) minus free hepatic venous
pressure (FHVP). The hepatic blood ow (HBF) was
determined by the indocyanine green constant infusion technique and the post-sinusoidal resistance as
Table I. Patient characteristics at baseline.

Age (years)
Sex (male/female)
MELD score
ChildPugh score
Child class A/B/C
Cirrhosis etiology (n)
Alcohol
Hepatitis B/C
Alcohol + hepatitis C
Idiopathic
Variceal status (n)
Grade 0
Grade 1
Grade 2
Grade 3
Former variceal bleeding (n)

Carvedilol
(n = 21)

Propranolol
(n = 17)

58.2 6.8
12/9
10.7 4.3
82
8/7/6

56.2 6.1
12/5
10.5 4.0
82
6/6/4

18
0/2
1
0

12
1/0
3
1

7
7
3
4
5

5
4
7
1
5

Abbreviations: MELD = model for end-stage liver disease;

SD = standard deviation.
Mean SD.

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

Carvedilol or propranolol in cirrhosis

trial drug was administered in the evening before the

nal hemodynamic measurements.
Response to treatment was dened as a decrease in
HVPG of 20% or to <12 mm Hg [4].
The study was conducted in accordance with the
rules of good clinical practice and approved by the
local research ethics committee (KF-02-049/03) and
the Danish Medicine Agency (2612-2341). The study
is registered at ClinicalTrials.gov (NCT00493480).

HVPG/HBF. Galactose elimination capacity (GEC)

was determined as previously described [24].
After baseline measurements, an oral dose of 80 mg
propranolol was administered to the patient and after
90 min, at the time of expected maximal plasma
concentration of propranolol, the hemodynamic measurements were repeated to establish the acute effect
of b-blockade [25]. An intravenous formula of propranolol would have been preferred as it can be
administered to sedated patients and potential differences in absorption would thereby be avoided, however this formulation is unavailable in Denmark.
Hemodynamic measurements were performed in
all patients at baseline and at the end of treatment. In
two patients in the carvedilol group and one patient in
the propranolol group, the acute effect of propranolol
was not assessed for technical reasons.

Statistics
A previous study [17] testing the acute effect of carvedilol compared with propranolol showed a signicant
difference in HVPG of mean 1.4 mm Hg in favor of
carvedilol. With a standard deviation (SD) of 1.2, we
calculated the sample size of 17 patients in each group
with a power of 90% and a signicance level of 5%.
Data are shown as mean SD. KolmogorovSmirnov
Test tested data for normality. The paired t-test was
used for comparison within each group; for comparison between groups the unpaired t-test was applied.
Categorical data are expressed as percentage and compared using c2 or Fishers exact test as appropriate.
The diagnostic value of the acute response of propranolol to predict the long-term treatment response
was evaluated using receiver operating characteristic
(ROC) curve analysis.
The SPSS version 17.0 (SPSS Inc., Chicago, IL,
USA) was used for statistical analyses.

Trial protocol
The day after baseline hemodynamic evaluation,
pharmacological therapy was started with either
carvedilol at a dosage of 3.125 mg twice daily or
propranolol at a dosage of 40 mg twice daily. Drug
titration was performed weekly, aiming at a pulse
reduction of 25% with a heart rate (HR) not below
55 beats per minute and a systolic blood pressure not
below 90 mm Hg. If these aims were not met, the dose
was doubled. The maximum dose for carvedilol was
25 mg/day (mean dose 14 7 mg) and for propranolol
320 mg/day (mean dose 122 64 mg). After stabilization of the dosage, patients were followed in the
outpatient clinic with clinical examination and registration of weight, pulse, blood pressure and side
effects after 30 days and 60 days. Compliance was
assessed by calculation of residual tablets at each visit.
After 92.7 13.6 days, clinical and hemodynamic
measurements were repeated. The last dose of the

Results
Long-term effects of carvedilol and propranolol
The mean baseline HVPG for the total population of
17.9 3.9 mm Hg decreased signicantly after an oral
dose of 80 mg propranolol by -19.3 12.8% (p < 0.01).

Table II. Portal and systemic hemodynamics in the two groups at baseline, at 90 min after propranolol and at end of treatment.
Carvedilol (n = 21)

WHVP (mm Hg)

FHVP (mm Hg)
HVPG (mm Hg)
HBF (l/min)
GEC (mmol/min)
HR (beats/min)
Systolic blood pressure
(mm Hg)
Diastolic blood pressure
(mm Hg)
MAP (mm Hg)

Baseline

90 min (n = 19)

23.7 4.9*
11.0 4.3
12.7 2.8*

27.6
10.0
17.6
1.23
1.54
82
134

5.4
4.1
4.2
0.51
0.29
15
21

61 6*

Propranolol (n = 17)

End of treatment
23.8
9.7
14.1
1.30
1.53
67
124

4.9*
4.0
4.0*
0.81 (n = 13)
0.35 (n = 19)
11*
19*

Baseline

90 min (n = 16)

26.6 2.5*
10.5 3.4
16.1 2.6*

27.9
9.5
18.4
1.81
1.53
81
135

4.5
3.9
3.6
1.23
0.35
14
k14

66 9*

End of treatment
25.1
9.0
16.1
1.16
1.46
63
123

4.9*
3.5
4.3*
0.57** (n = 13)
0.44 (n = 15)
9*
18*

79 13

76 11

79 9

73 10*

97 13

92 11

98 8

90 11*

Abbreviations: FHVP = free hepatic venous pressure; GEC = galactose elimination capacity; HVPG = hepatic venous pressure gradient;
HBF = hepatic blood ow; HR = heart rate; MAP = mean arterial pressure; WHVP = wedge hepatic venous pressure.
Compared with baseline*0.05, **p = 0.07.

L. Hobolth et al.

Table III. Side effects observed during the study.

Carvedilol Propranolol

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

Side effects
leading to
withdrawal

Hematemesis
Dizziness
Impotence
Headache
Chest pain
Skin rash

Side effects not

leading to
Cold extremities
withdrawal
Orthostatic hypotension
Shortness of breath
Impotence
Dreaming/nightmares
Diarrhea
Increasing diuretics
Headache

n
1
0
0
0
1
1

n
1
2
1
1
0
0

5
4
6
0
1
0
7
1

3
4
3
1
3
2
4
0

A more pronounced decrease in HVPG was seen in the

group later randomized to carvedilol (-24.4 12.5 vs.
-13.2 10.6%, p < 0.01) (Table II). At the end of the
treatment period, HVPG had decreased by -19.3
16.1% (p < 0.01) in the carvedilol group and by -12.5
16.7% (p < 0.01) in the propranolol group with no
signicant difference between the two treatment
groups (p = 0.21) (Table II). The number of patients
with a HVPG response 20% or to HVPG <12 mm Hg
after 90 days was not signicantly different between the
two groups: (13/21 (62%) in the carvedilol group
versus 7/17 (41%) in the propranolol group, p = 0.20).
The number of patients in the carvedilol and
propranolol groups with a drop in HVPG of 20%
after 90 days of treatment was 10/21 (48%) versus
7/17 (41%), p = 0.69, and 6/21(28.6%) versus
2/17 (11.8%), p = 0.26, reached a HVPG < 12 mm
Hg. No signicant effect of carvedilol was seen on
HBF, whereas propranolol tended to decrease HBF
(p = 0.07). GEC remained unchanged after longterm treatment with both carvedilol and propranolol
(Table II). Mean arterial pressure (MAP) was not
signicantly affected in the carvedilol group
(p = 0.23), while it decreased from 97.7 8.2 to
89.5 10.8 mm Hg in the propranolol group
(p 0.05) (Table II). No difference was seen in
DMAP between the two treatment regimens -5.1
14.9 versus -8.2 12.5 (p = 0.50). In the carvedilol
group, the HR decreased by -18% compared
with -23% in the propranolol group (p = 0.23). We
found no correlation between reduction in HR and
reduction in HVPG, this holds for the whole
group (r = -0.09, p = 0.60) as well as the
separate groups (r = -0.05, p = 0.84, carvedilol group
and r = -0.24, p = 0.24, propranolol group).

Table IV. Cross-tabulation of the acute hemodynamic response to

propranolol with the long-term hemodynamic response to
carvedilol.
Long-term hemodynamic response to carvedilol
Acute hemodynamic
response 20% or
HVPG <12 mm Hg

Yes
No
Total

Yes
6
6
12

No
6
1
7

Total
12
7
19

Abbreviation: HVPG = hepatic venous pressure gradient.

Fishers exact test p = 0.17, sensitivity 50%, specicity 14%.

Table V. Cross-tabulation of patients receiving propranolol.

Long-term hemodynamic response to propranolol
Acute hemodynamic
response 12% or
HVPG <12 mm Hg

Yes
No
Total

Yes
5
2
7

No
2
7
9

Total
7
9
16

Abbreviations: HVPG = hepatic venous pressure gradient;

ROC = receiver operating characteristic.
The acute HVPG response of 12% to propranolol was suggested
by ROC curve analysis to be the best cut-off value to predict longterm HVPG response to propranolol (HVPG 20% or
HVPG < 12 mm Hg).
Fishers exact test p = 0.13, sensitivity 71%, specicity 78%.

Creatinine clearance remained unchanged in both

groups. Body weight increased by 1.9 kg in the carvedilol group (from 76.9 15.7 kg to 78.8 15.2 kg,
p = 0.06), and decreased by 0.5 kg in the propranolol
group, but the difference between the groups was not
signicant (p = 0.12). Reduction in study medication
was needed in three patients: orthostatic hypotension
(carvedilol group), headache (carvedilol group) and
orthostatic hypotension (propranolol group). Adverse
events are shown in Table III.
Acute response to propranolol as a predictor of long-term
hemodynamic response
Data were available for effects of both acute and longterm response in 35 patients, 19 in the carvedilol
group and 16 in the propranolol group (Table IV
and Table V)
The ROC curve on the acute HVPG response to
oral propranolol (n = 19) as a prediction of the longterm HVPG response to carvedilol showed a
very poor discriminative value with an area under
the curve (AUC) of 0.53 (95% condence interval
(CI) 0.270.79, p = 0.83). From these gures it
was not possible to suggest an optimal cut-off value
(gures not shown).
The ROC curve on the acute HVPG response to
propranolol as a prognostic test to predict the longterm HVPG response to propranolol (n = 16) had an

Carvedilol or propranolol in cirrhosis

The long-term effect of chronic treatment with

carvedilol has previously been found to signicantly
decrease portal pressure [11,12,1416]. Our study
also demonstrated that carvedilol reduced the portal
pressure after 90 days of treatment in patients with
portal hypertension and cirrhosis. In comparison with
propranolol we did not nd any signicant difference
in the lowering effect of carvedilol on HVPG, which is
in accordance with the study by De et al. [16], who
followed 36 patients for 7 days. However, there was a
tendency although insignicant to a more pronounced effect of carvedilol than of propranolol on
HVPG (19.3% versus 12.5%). This nding is supported by the study by Banares et al., who found
carvedilol to be superior to propranolol in 51 patients
treated for 11 weeks [15].
In our study, we used a mean titrated dose of 14 mg
carvedilol compared with 12.5 mg and 31 mg in the
studies by De et al. [16] and Banares et al. [15],
respectively. In a pilot study, a dose of 25 mg carvedilol decreased HVPG signicantly whereas 12.5 mg
did not [17]. In long-term studies, the dose of
12.5 mg has later been shown to reduce HVPG
signicantly [11,14,16] and even to be superior to
banding in preventing the rst variceal bleeding [26].
The higher dose applied in the Spanish study [15] did

Discussion
In this randomized trial comparing carvedilol with
propranolol in patients with cirrhosis and portal hypertension, we found no signicant difference in the effect
on HVPG between the two treatments after 90 days.
Furthermore, in our study the acute response to an oral
dose of propranolol could not signicantly discriminate between long-term hemodynamic responders and
non-responders. The more pronounced acute effect of
propranolol on HVPG in the group randomized to
carvedilol is most likely due to a type-1 error; however,
it cannot be excluded that this group of patients was
more sensitive to b-blocker treatment than the group
randomized to propranolol.

ROC curve, propranolol group

1.0

Sensitivity

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

AUC of 0.73 (95% CI 0.470.99, p = 0.13) (Figure 1).

Although insignicant, the optimal cut-off value was a
decrease in HVPG 12% with a sensitivity of 71% and
a specicity of 77% (Table V). When using this cutoff value in the cross-tabulation, the short-term HVPG
response to propranolol predicted the long-term
response in 75% of the cases, with 5 out of 16 patients
being both acute and long-term hemodynamic responders, and 7 out of 16 being both acute and longterm hemodynamic non-responders (Table V).

0.8

0.6

0.4

0.2

0.0
0.0

0.2

0.4

0.6

0.8

1-Specificity
Figure 1. ROC curve, propranolol group.

1.0

L. Hobolth et al.

Table VI. Outcome of randomized long-term studies comparing carvedilol with propranolol.

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

Banares et al. [15]

Number of patients
Follow-up time (days)
Carvedilol/propranolol dose (mg)
Baseline HVPG (mm Hg)
End-of-treatment HVPG (mm Hg)
Reduction HVPG (%)
Response rate 20% or to <12 mm Hg (%)
Baseline HR
End-of-treatment HR
Reduction HR (%)
Baseline MAP (mm Hg)
End-of-treatment MAP (mm Hg)
Reduction MAP (%)

De et al. [16]

Carvedilol

Propranolol

24
77
31
19.0
15.2
19*
54
79.9
65.6
18*
91.4
81.2
11*

22
77
73
20.3
17.6
12*
23
77.3
58.2
25*
88.6
83.8
5

4
1.1
0.8
3.7
2.0
2.5
2.9

10
0.9
0.7
2.6
1.0
4.5
3.1

Carvedilol
18
7
12.5
19.0
13.6
28*
61
86.9
73.5
15*
97.3
82.2
16*

3.8
5.4
13.3
9.3
10.3
12.6

Present study

Propranolol
18
7
80
16.6
13.1
21*
65
92.6
69.6
25*
91.9
86.2
6*

4.0
5.3
11.9
8.0
16.0
13.3

Carvedilol
21
90
14
17.6
14.1
19*
62
82.4
67.4
18*
97.1
92.1
5

7
4.2
4.0
15.2
10.5
13.4
11.2

Propranolol
17
90
122
18.4
16.1
13*
41
81.2
62.6
23*
97.7
89.5
8*

64
3.6
4.3
14.4
9.2
8.2
10.8

Abbreviations: HVPG = hepatic venous pressure gradient; HR = heart rate; MAP = mean arterial pressure.
*p < 0.05, compared with baseline.

not lead to a higher response rate nor a higher percentage in HVPG reduction compared with our study
and that by De et al. (Table VI) [16]. The titrated
dose of propranolol was higher in our study compared
with the doses in the other studies [15,16]. It has
previously been observed by Abraldes et al. that the
dose of the non-selective b-blocker in secondary prophylaxis was an independent predictor of the patient
being a responder [27] and the high dose of propranolol in our study could be the reason for the higher
proportion of propranolol responders seen in our
study compared with the study by Banares et al.
[15]. The study by De et al. [16] had a very high
response rate to propranolol, which could partly be
explained by their lower baseline HVPG.
To determine whether carvedilol is superior to
propranolol, long-term comparative trials of carvedilol and propranolol with both bleeding and mortality
as outcome measures are required. Such trials should
include HVPG measurements.
The lack of correlation between reductions in HR
and HVPG in our study as seen in previous studies
[25,28] may lead to the conclusion that titration
should be guided by clinical tolerability rather than
by reductions in HR or optimally by repeated HVPG
measurements.
The weight gain in the carvedilol group is supposed
to be due to the a-blocking effect with a decrease in
MAP leading to a compensatory increase in uid
retention because of an increased production of aldosterone and renin. The drop in MAP has been a
matter of concern in cirrhotic patients due to renal
complications in particular at higher doses of carvedilol [12,15,29]. However in our study with slow
titration of the study medication, there was no difference in renal function, and MAP was left unaffected

in the carvedilol group, while it decreased signicantly

in the propranolol group, probably owing to the larger
doses of propranolol affecting the cardiac output.
Nevertheless, an insignicant increase in body
weight was seen in the carvedilol-treated patients
with the need for slightly increased diuretic doses
compared with the propranolol group as shown
in Table III.
We could not predict the long-term effect of carvedilol nor propranolol by measuring the acute
HVPG response to propranolol in our study.
Although the discriminative value of the propranolol
test was insignicant (p = 0.13), the ROC analysis
found the optimal cut-off value to be an acute reduction in HVPG of 12% with a sensitivity of 71% and a
specicity of 78%. These results are based on the
acute response to oral propranolol in the propranolol
group with only 16 patients included; therefore, the
lack of a signicant nding might have been overlooked due to a type 2 error.
Nevertheless, the cut-off value of 12% is in close
agreement with two recent, larger studies. One study
investigated the acute HVPG response to intravenous
propranolol and found a reduction of >10% to be
predictive of rst bleeding [21] and the other study
found a reduction of >12% to be predictive of rebleeding [20], both after 2 years follow-up.
The inability of the acute propranolol test to discriminate between long-term HVPG responders and
non-responders in our study is in line with two other
studies: one evaluating the acute effect of oral propranolol to predict the long-term effect with HVPG
reductions [22] and one evaluating the long-term effect
with variceal bleeding as end point [23].
Six out of seven patients failing an acute response to
oral propranolol showed long-term response after

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

Carvedilol or propranolol in cirrhosis

90 days of carvedilol (Table IV). This nding suggests
that non-responders to propranolol may benet from
treatment with carvedilol either alone or as add-ontherapy as an alternative to band-ligation.
In conclusion, this randomized comparison showed
that the portal pressure-lowering effects of carvedilol
and propranolol are at least equal after 90 days of
treatment. Slow titration to a mean daily dose of
14 mg carvedilol was safe and well tolerated. The
acute response to propranolol could not signicantly
discriminate between long-term responses on HVPG
when using either the previous suggested criterion of
an acute reduction in HVPG of 20% or the optimal
cut-off value of 12%.
Declaration of interest: This study received funding support from Jacob & Olga Madsens Foundation,
Roche, the Research Foundation of Copenhagen
Council, Hvidovre Hospital Foundation for Liver
Disease, L.F. Foghts Foundation, DPHG and the
Novo Nordisk Foundation.

[12]

[13]

[14]

[15]

[16]

[17]

References
[18]
[1] Hobolth L, Krag A, Bendtsen F. The recent reduction in
mortality from bleeding oesophageal varices is primarily
observed from days 1 to 5. Liver Int 2010;30:45562.
[2] Carbonell N, Pauwels A, Serfaty L, Fourdan O, Levy VG,
Poupon R. Improved survival after variceal bleeding in
patients with cirrhosis over the past two decades. Hepatology
2004;40:6529.
[3] Garcia-Tsao G, Groszmann RJ, Fisher RL, Conn HO,
Atterbury CE, Glickman M. Portal pressure, presence of
gastroesophageal varices and variceal bleeding. Hepatology
1985;5:41924.
[4] DAmico G, Garcia-Pagan JC, Luca A, Bosch J. Hepatic vein
pressure gradient reduction and prevention of variceal bleeding in cirrhosis: a systematic review. Gastroenterology 2006;
131:161124.
[5] DAmico G, Pagliaro L, Bosch J. Pharmacological treatment
of portal hypertension: an evidence-based approach. Semin
Liver Dis 1999;19:475505.
[6] Bosch J, Garcia-Pagan JC. Prevention of variceal rebleeding.
Lancet 2003;361:9524.
[7] Albillos A, Banares R, Gonzalez M, Ripoll C, Gonzalez R,
Catalina MV, et al. Value of the hepatic venous pressure
gradient to monitor drug therapy for portal hypertension:
a meta-analysis. Am J Gastroenterol 2007;102:111626.
[8] Feu F, Garcia-Pagan JC, Bosch J, Luca A, Teres J,
Escorsell A, et al. Relation between portal pressure response
to pharmacotherapy and risk of recurrent variceal haemorrhage in patients with cirrhosis. Lancet 1995;346:10569.
[9] Pedersen ME, Cockcroft JR. The vasodilatory beta-blockers.
Curr Hypertens Rep 2007;9:26977.
[10] Forrest EH, Bouchier IA, Hayes PC. Acute haemodynamic
changes after oral carvedilol, a vasodilating beta-blocker, in
patients with cirrhosis. J Hepatol 1996;25:90915.
[11] Tripathi D, Therapondos G, Lui HF, Stanley AJ, Hayes PC.
Haemodynamic effects of acute and chronic administration
of low-dose carvedilol, a vasodilating beta-blocker, in patients

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

with cirrhosis and portal hypertension. Aliment Pharmacol

Ther 2002;16:37380.
Stanley AJ, Therapondos G, Helmy A, Hayes PC. Acute and
chronic haemodynamic and renal effects of carvedilol in
patients with cirrhosis. J Hepatol 1999;30:47984.
Sekiyama T, Komeichi H, Nagano T, Ohsuga M, Terada H,
Katsuta Y, et al. Effects of the alpha-/beta-blocking agent
carvedilol on hepatic and systemic hemodynamics in patients
with cirrhosis and portal hypertension. Arzneimittelforschung 1997;47:3535.
Bruha R, Vitek L, Petrtyl J, Lenicek M, Urbanek P,
Zelenka J, et al. Effect of carvedilol on portal hypertension
depends on the degree of endothelial activation and inammatory changes. Scand J Gastroenterol 2006;41:145463.
Banares R, Moitinho E, Matilla A, Garcia-Pagan JC,
Lampreave JL, Piera C, et al. Randomized comparison of
long-term carvedilol and propranolol administration in the
treatment of portal hypertension in cirrhosis. Hepatology
2002;36:136773.
De BK, Das D, Sen S, Biswas PK, Mandal SK,
Majumdar D, et al. Acute and 7-day portal pressure response
to carvedilol and propranolol in cirrhotics. J Gastroenterol
Hepatol 2002;17:1839.
Banares R, Moitinho E, Piqueras B, Casado M, GarciaPagan JC, de DA, et al. Carvedilol, a new nonselective
beta-blocker with intrinsic anti-alpha1-adrenergic activity,
has a greater portal hypotensive effect than propranolol in
patients with cirrhosis. Hepatology 1999;30:7983.
Lin HC, Yang YY, Hou MC, Huang YT, Lee FY, Lee SD.
Acute administration of carvedilol is more effective than
propranolol plus isosorbide-5-mononitrate in the reduction
of portal pressure in patients with viral cirrhosis. Am J
Gastroenterol 2004;99:19538.
Vorobioff J, Picabea E, Villavicencio R, Puccini V, Rossi O,
Bordato J, et al. Acute and chronic hemodynamic effects of
propranolol in unselected cirrhotic patients. Hepatology
1987;7:64853.
La MV, Abraldes JG, Raffa S, Retto O, Berzigotti A, GarciaPagan JC, et al. Prognostic value of acute hemodynamic
response to i.v. propranolol in patients with cirrhosis and
portal hypertension. J Hepatol 2009;51:27987.
Villanueva C, Aracil C, Colomo A, Hernandez-Gea V,
Lopez-Balaguer JM, varez-Urturi C, et al. Acute hemodynamic response to beta-blockers and prediction of long-term
outcome in primary prophylaxis of variceal bleeding. Gastroenterology 2009;137:11928.
Merkel C, Bolognesi M, Sacerdoti D, Bombonato G,
Cavasin L, Gatta A. Disagreement between acute and
chronic haemodynamic effects of nadolol in cirrhosis:
a pathophysiological interpretation. Aliment Pharmacol
Ther 2005;22:4339.
Valla D, Jiron MI, Poynard T, Braillon A, Lebrec D. Failure
of haemodynamic measurements to predict recurrent gastrointestinal bleeding in cirrhotic patients receiving propranolol.
J Hepatol 1987;5:1448.
Tygstrup N. Determination of the hepatic elimination capacity (Lm) of galactose by single injection. Scand J Clin Lab
Invest Suppl 1966;18:11825.
Bendtsen F, Henriksen JH, Srensen TI. Propranolol and
haemodynamic response in cirrhosis. J Hepatol 1991;13:
1448.
Tripathi D, Ferguson JW, Kochar N, Leithead JA,
Therapondos G, McAvoy NC, et al. Randomized controlled
trial of carvedilol versus variceal band ligation for the prevention of the rst variceal bleed. Hepatology 2009;50:82533.

L. Hobolth et al.

Scand J Gastroenterol Downloaded from informahealthcare.com by Statsbiblioteket Tidsskriftafdeling on 03/13/12

For personal use only.

[27] Abraldes JG, Tarantino I, Turnes J, Garcia-Pagan JC,

Rodes J, Bosch J. Hemodynamic response to pharmacological treatment of portal hypertension and long-term prognosis
of cirrhosis. Hepatology 2003;37:9028.
[28] Garcia-Tsao G, Grace ND, Groszmann RJ, Conn HO,
Bermann MM, Patrick MJ, et al. Short-term effects of

propranolol on portal venous pressure. Hepatology 1986;6:

1016.
[29] Albillos A, Garcia-Pagan JC, Iborra J, Bandi JC, Cacho G,
Perez-Paramo M, et al. Propranolol plus prazosin compared
withpropranololplus isosorbide-5-mononitrate inthetreatment
of portal hypertension. Gastroenterology 1998;115:11623.