Hemodynamic Response to Pharmacological Treatment

of Portal Hypertension and Long-Term Prognosis

of Cirrhosis
Juan G. Abraldes, Ilaria Tarantino, Juan Turnes, Juan Carlos Garcia-Pagan, Juan Rode s, and Jaime Bosch

In cirrhotic patients under pharmacologic treatment for portal hypertension, a reduction in

hepatic venous pressure gradient (HVPG) of >20% of baseline or to <12 mm Hg markedly
reduces the risk of variceal rebleeding. This study was aimed at evaluating whether these
hemodynamic targets also prevent other complications of portal hypertension and improve
long-term survival. One hundred five cirrhotic patients included in prospective trials for the
prevention of variceal rebleeding were studied. Seventy-three of the patients had 2 separate
HVPG measurements, at baseline and under pharmacologic therapy with propranolol
isosorbide mononitrate. Patients were followed for up to 8 years. Survival and risk of
developing portal hypertension-related complications were compared between responders
and nonresponders. Twenty-eight patients showed a reduction of HVPG >20% of baseline
or to <12 mm Hg (responders), and 45 patients were nonresponders. Nonresponders had a
significantly greater risk of developing variceal rebleeding (P .013), ascites (P .025),
spontaneous bacterial peritonitis (P .003), hepatorenal syndrome (P .026), and hepatic
encephalopathy (P .024) than responders. Eight-year cumulative probability of survival
was significantly lower in nonresponders than in responders (52% vs. 95%, respectively, P
.003). At multivariate analysis, being a nonresponder was independently associated with the
risk of developing rebleeding, ascites, spontaneous bacterial peritonitis, and lower survival.
In conclusion, in cirrhotic patients receiving pharmacologic treatment for prevention of
variceal rebleeding, a decrease in HVPG >20% or to <12 mm Hg is associated with a
marked reduction in the long-term risk of developing complications of portal hypertension
and with improved survival. (HEPATOLOGY 2003;37:902-908.)

P
ortal hypertension is responsible for the more
frequent and severe complications of cirrhosis:
gastrointestinal bleeding from gastroesophageal
varices, ascites, renal dysfunction, and hepatic enceph-
alopathy. Because of the combined impact of these
complications, portal hypertension represents the main
Abbreviations: HVPG, hepatic venous pressure gradient; TIPS, transjugular
intrahepatic portosystemic shunt; ISMN, isosorbide mononitrate; CI, cardiac index.
From the Hepatic Hemodynamic Laboratory, Liver Unit, Institut de Malalties
Digestives, Hospital Clnic, Institut de Investigacions Biomediques August Pi i
Sunyer (IDIBAPS), Universitat de Barcelona, Barcelona, Spain.
Received September 30, 2002; accepted January 8, 2003.
J.G.A. and I.T. share first authorship of this report.
Supported in part by grants from Fondo de Investigaciones Sanitarias (FIS 00/
0444) (01/9356 to J.G.A.) and Comision Interministerial de Ciencia y Tecnologa
(SAF99/0007).
Address reprint requests to: Jaime Bosch, Hepatic Hemodynamic Laboratory,
Liver Unit, Hospital Clinic, Villarroel 170, 08036 Barcelona, Spain. E-mail:
jbosch@clinic.ub.es; fax: (34) 93 22 79856.
Copyright 2003 by the American Association for the Study of Liver Diseases.
0270-9139/03/3704-0026$30.00/0
doi:10.1053/jhep.2003.50133
902
cause of death and liver transplantation in patients
with cirrhosis.1
Previous cross-sectional studies have shown that a por-
tal pressure gradient (determined as the hepatic venous
pressure gradient; HVPG) above 12 mm Hg is required
for variceal bleeding to develop.2,3 Indeed, longitudinal
studies demonstrated that, when HVPG decreases below
this threshold, there is total protection from the risk of
bleeding.4-6 Subsequent studies showed that, even with-
out reaching this target, a reduction of HVPG of at least
20% from baseline values is associated with a very low
residual risk of rebleeding on follow-up.7-10 However, no
study so far has examined whether the hemodynamic re-
sponse to drug therapy correlates also with the risk of
developing other complications of portal hypertension
and survival over a long-term follow-up.
The present study was aimed at addressing these issues.
For that purpose, we investigated the relationship be-
tween the HVPG response to continued pharmacologic
therapy and occurrence of variceal rebleeding, ascites,
HEPATOLOGY, Vol. 37, No. 4, 2003
spontaneous bacterial peritonitis, hepatic encephalopa-
thy, hepatorenal syndrome, and death in a large cohort of
patients included in an 8-year follow-up study.
Patients and Methods
Study Cohort. The study included 105 cirrhotic pa-
tients with portal hypertension admitted to the Liver Unit
of the Hospital Clinic with bleeding from esophageal var-
ices that were included in 3 prospective single-center11 or
multicenter12,13 trials evaluating the efficacy of pharma-
cologic therapy versus endoscopic treatment, surgery, or
transjugular intrahepatic portosystemic shunt (TIPS). In-
clusion criteria were the following: diagnosis of cirrhosis
(based on liver biopsy and/or unequivocal clinical data
and compatible findings on imaging techniques); sinusoi-
dal portal hypertension with HVPG above 12 mm Hg;
index bleeding from esophageal varices confirmed by
emergency endoscopy; repeat HVPG measurements, the
first before starting therapy, the second 1 to 6 months
later, under continued therapy with propranolol
isosorbide mononitrate (ISMN); and minimal follow-up
of 6 months. Patients with hepatocellular carcinoma at
baseline, cholestatic liver disease, or portal vein thrombo-
sis were not considered for the study. Twenty-three pa-
tients were excluded because of intolerance to
pharmacologic treatment (n 7) or a terminating event
before the second hemodynamic study (8 rebleedings, 7
deaths, 1 hepatocellular carcinoma). Nine additional pa-
tients refused the second hemodynamic study. Therefore,
73 patients were available for final analysis. The day of the
second hemodynamic study was taken as time zero for the
follow-up.
Hemodynamic Measurements. The initial hemody-
namic measurement was performed a median time of 8
days (range 5-14 days) from the index bleeding. Under
local anesthesia, a venous introducer was placed in the
right femoral vein or internal jugular vein by the Seldinger
technique. Under fluoroscopy, a 7F balloon-tipped cath-
eter (Boston Scientific Medi-Tech, Natick, MA) was
guided into the main right hepatic vein for measurements
of wedged (occluded) and free hepatic venous pressures.
Portal pressure was measured as the HVPG, the difference
between wedged and free hepatic venous pressures, as pre-
viously described.14 Cardiopulmonary pressures and car-
diac output were measured by means of a Swan-Ganz
catheter (Abbott Laboratories, Chicago, IL) advanced
into the pulmonary artery. After this initial evaluation,
patients were started on oral propranolol, 20 mg twice
daily. This dose was increased until heart rate had fallen
by 25% or below 55/min, or systolic blood pressure was
below 90 mm Hg. In 45 patients,12,13 after stabilizing on
ABRALDES ET AL. 903
propranolol, ISMN was added at increasing doses up to
40 mg twice a day. The second hemodynamic evaluation
was done after a median of 111 days (range 34-164).
Permanent tracings of the hemodynamic studies were re-
viewed by an independent investigator unaware of clinical
data. Patients were considered responders if, at the sec-
ond hemodynamic evaluation, HVPG was 12 mm Hg
or had decreased at least by 20% from baseline values and
nonresponders if these criteria were not met.7
Follow-up. Patients were followed in the outpatient
clinic, initially at weekly intervals until stabilized on
maintenance doses of drug therapy, then at 3 months and
every 6 months thereafter. At each visit, medical history,
physical examination, biochemistry, hematologic tests,
abdominal ultrasound, continued alcohol abuse, and ful-
fillment of treatment were recorded. Compliance was as-
sessed by attendance at scheduled visits, anamnesis, and
heart rate and was considered good when these parameters
were met in 80% of visits, fair when met in 50% to 80%
of visits, and poor when met in 50% of visits. Follow-up
data were gathered for up to 8 years, death, or liver trans-
plantation. Median follow-up was 70 months. End points
evaluated were the following: survival, rebleeding from
gastroesophageal varices, ascites, spontaneous bacterial
peritonitis, hepatic encephalopathy, and hepatorenal syn-
drome. Patients undergoing liver transplantation were
censored as alive the day of the transplant. Patients requir-
ing derivative treatment were censored the day of the
procedure, except for survival analysis. Variceal bleeding
was defined following Baveno criteria.15 The ascites end
point was defined as de novo development of ascites or
worsening of preexisting ascites (patients requiring a sus-
tained increase in diuretic dose or admission for large-
volume paracentesis). De novo ascites was diagnosed
clinically and confirmed by ultrasound and/or paracente-
sis. Spontaneous bacterial peritonitis and hepatorenal
syndrome were defined as in accepted consensus.16,17 He-
patic encephalopathy was diagnosed on clinical basis. Pa-
tients lost to follow-up were censored the day of the last
visit. However, for survival analysis, the final status of the
patient was traced by contacting the patients, their rela-
tives, or, when this was not successful, by means of a
requirement to local civil registry to assess the exact date
and cause of death.
Statistical Analysis. Statistical analysis was done with
the SPSS 10.0 package (SPSS, Chicago, IL). Cumulative
risk of developing each of the 6 analyzed end points were
assessed by Kaplan-Meier curves. Comparisons between
responders and nonresponders were performed with the
log-rank test. The adjusted influence of hemodynamic
response on the risk of developing the events was analyzed
with Cox proportional hazards model by introducing co-
904 ABRALDES ET AL. HEPATOLOGY, April 2003

Table 1. Baseline Characteristics of Responders and response were identified by logistic regression. Signifi-
Nonresponders cance was established at P .05.
Responders Nonresponders P Value

Age 53 8 56 8 NS Results
Sex (% male) 64 67 NS
Etiology (% alcohol) 54 49 NS Hemodynamic Response. A total of 73 patients were
Active alcoholism (%) 37 29 NS
Child (% A/B/C) 50/46/4 40/51/9 NS
included in the study. Mean follow-up was 70 months.
Child (score) 6.6 1.3 7.0 1.6 NS Twenty-eight patients were HVPG responders and 45
Ascites (%) 32 31 NS nonresponders. Table 1 shows the main characteristics of
Basal heart rate (bpm) 82 14 80 13 NS the 2 groups of patients. Among responders, 11 patients
HVPG (mm Hg) 18.6 3.9 18.4 3.4 NS
Albumin (g/L) 35.6 6.0 32.7 6.0 .045 decreased HVPG to 12 mm Hg, and 17 patients de-
Bilirubin (mg/dL) 1.4 0.8 2.0 1.5 .020 creased HVPG by 20% of baseline, with final HVPG
Prothrombin rate (%) 70 17 65 15 NS 12 mm Hg. Nonresponders had greater bilirubin levels,
Variceal size (% large) 71 75 NS
Concomitant ISMN (%) 62 61 NS
lower albumin, and were treated with lower doses of pro-
Propranolol dose (mg/d) 150 97 89 56 .002 pranolol than responders (nonresponders 89 56 mg/d,
Decrease in HR (%) 19 17 21 14 NS vs. responders 150 97 mg/d; P .002). At multivariate
Decrease in CO (%) 20 14 19 14 NS
Decrease in HVPG (%) 28 10 5 12 .000
analysis, the only independent predictor of being a re-
Compliance (% good/fair/poor) 78/5/17 89/3/8 NS sponder was propranolol dose (P .005) (Table 1). The
concomitant use of ISMN was not. The reasons for not
NOTE. On logistic regression analysis, only the dose of propranolol predicted
response. increasing further the dose of propranolol in nonre-
Abbreviations: NS, not significant; HR, heart rate; CO, cardiac output; HVPG, sponders were the following: reaching a 25% decrease in
hepatic venous pressure gradient. heart rate (n 18), reaching the maximum allowed dose
of propranolol (320 mg/d, n 1), reaching a final heart
rate of 55 bpm (n 6), hypotension (n 8), and
variates that were related to these events in a univariate intolerance (n 12).
analysis (P .1). Tested covariates included demo- Rebleeding. Over the 8-year follow-up, responders
graphic data (age, gender), etiology of cirrhosis (alcoholic had a significantly lower actuarial probability of variceal
vs. nonalcoholic), presence of ascites, hepatic encephalop- rebleeding than nonresponders (72% vs. 43% free of re-
athy, serum sodium, creatinine, bilirubin (transformed bleeding, respectively; P .013) (Fig. 1). Total number
into its natural logarithm), albumin, prothrombin ratio, of rebleeders were 20/45 nonresponders and 6/28 re-
platelet count, endoscopic data (big vs. small varices), sponders (including one that had decreased HVPG to
hemodynamic data (HVPG, mean arterial pressure, car- 12 mm Hg). Among responders, 2 patients required
diac output) at baseline and at time of second hemody-
namic evaluation, treatment-related data (propranolol
dose, concomitant treatment with ISMN, use of diuretics,
compliance with treatment), and continued alcohol
abuse. The contribution of each significant variable to the
risk of reaching the end point was estimated by the relative
hazard with its 95% confidence intervals (CIs). Each
component of Child-Pugh score was individually forced
into the final models (if not already present) and main-
tained whenever important changes in regression coeffi-
cients or their standard errors were observed. For survival
end point, to better adjust the influence of hemodynamic
response by the degree of liver failure, we performed an
additional analysis using Cox model, including hemody-
namic response, age, and Child-Pugh score. MELD
score18 was not entered as a covariate because it was de-
signed to assess short-term prognosis in patients with
advanced liver disease. The predicting factors of hemody- Fig. 1. Kaplan-Meier plot showing cumulative probability of remaining
free of variceal rebleeding. Eight-year probability of remaining free of
namic response were analyzed by t test, Mann-Whitney rebleeding was higher in responders than in nonresponders (R, respond-
test, or 2 tests as appropriate. Independent predictors of ers; NR, nonresponders).
HEPATOLOGY, Vol. 37, No. 4, 2003 ABRALDES ET AL. 905

Table 2. Factors Independently Associated With the

Analyzed End Points
RH 95% CI P Value

Rebleeding
Lack of hemodynamic response 2.804 1.107-7.104 .028
Albumin (g/L) 0.928 0.865-0.995 .030
Ascites
Lack of hemodynamic response 3.483 1.265-9.587 .009
Platelet count ( 109/L) 0.985 0.971-0.999 .016
Spontaneous bacterial peritonitis
Lack of hemodynamic response 8.687 1.080-69.886 .016
Prothrombin rate (%) 0.950 0.904-0.998 .027
Hepatic encephalopathy
Bilirubin (mg/dL) 2.617 1.069-6.408 .008
Survival
Lack of hemodynamic response 12.387 1.555-98.690 .001
Age (y) 1.117 1.026-1.216 .008
Albumin (g/L) 0.903 0.824-0.990 .026

NOTE. Analyzed with Cox proportional hazards model. The value of relative
Fig. 2. Cumulative probability of not reaching the ascites end point
hazard indicates the relative risk or the strength of association of each variable
was higher in responders than in nonresponders (R, responders; NR,
adjusted by the other significant variables. Values above 1 indicate increased risk
nonresponders).
of reaching the end point. Values lower than 1 indicate protective effect.
Abbreviation: RH, relative hazard.

Hepatorenal Syndrome. Only 4 patients developed

hepatorenal syndrome during follow-up. All were nonre-
derivative treatment during follow-up (1 surgical shunt, 1 sponders. The risk of developing hepatorenal syndrome
TIPS), versus 8 patients in nonresponders (7 surgical was associated with albumin levels (P .003), prothrom-
shunts, 1 TIPS). At univariate analysis, lack of hemody- bin ratio (P .050), Child-Pugh score (P .022), and
namic response, albumin (P .013), and bilirubin levels lack of hemodynamic response (P .026). The low num-
(P .046) were significantly associated with rebleeding. ber of events did not permit multivariate analysis.
At multivariate analysis, albumin and lack of hemody- Hepatic Encephalopathy. Fifteen patients developed
namic response were independent predictors of rebleed- hepatic encephalopathy (12 nonresponders). No patient
ing (Table 2). Among nonresponders, rebleeding rates in whom HVPG decreased to 12 mm Hg developed
were similar in patients exhibiting no HVPG reduction as encephalopathy during follow-up. The eight-year proba-
compared with those with a moderate decrease in HVPG bility of developing encephalopathy was significantly
(between 10% and 20%). lower in responders than in nonresponders (84% vs. 58%
Ascites. The actuarial probability of not reaching the
ascites end point was significantly higher in responders
than in nonresponders (70% vs. 42% at 8 years, P
.025) (Fig. 2). The corresponding number of events were
16 in nonresponders versus 6 in responders (2 in patients
that decreased HVPG to 12 mm Hg). At univariate
analysis, hemodynamic response and platelet count (P
.031) were associated with the probability of ascites. Pro-
thrombin ratio, albumin, and bilirubin approached sta-
tistical significance (all P .01). Independent predictors
were lack of hemodynamic response and platelet count
(Table 2).
Spontaneous Bacterial Peritonitis. The probability
of being free of spontaneous bacterial peritonitis was
higher in responders than in nonresponders (94% vs.
58% at 8 years, P .003) (8 events in nonresponders vs.
1 in responders; none in patients with final HVPG 12
mm Hg) (Fig. 3). Prothrombin ratio was also associated
Fig. 3. Eight-year cumulative probability of remaining free of sponta-
with spontaneous bacterial peritonitis (P .019). Both neous bacterial peritonitis was higher in responders than in nonre-
were independent predictors (Table 2). sponders (R, responders; NR, nonresponders).
906 ABRALDES ET AL.
Fig. 4. Eight-year cumulative probability of remaining free of hepatic
encephalopathy was higher in responders than in nonresponders (R,
responders; NR, nonresponders).
free of encephalopathy; P .024) (Fig. 4). At univariate
analysis albumin (P .015), bilirubin (P .001), pro-
thrombin ratio (P .031), propranolol dose (P .032),
and lack of hemodynamic response (P .024) were asso-
ciated with the risk of developing encephalopathy. The
only independent predictor was bilirubin (Table 2).
Survival. Among responders, 1 patient died during
follow-up, and 4 (one of them that had decreased HVPG
to 12 mm Hg) underwent liver transplantation. In con-
trast, among nonresponders, 14 patients died, and 3 were
transplanted (Table 3). Eight-year survival probability
was higher in responders than in nonresponders (95% vs.
52%; P .003) (Fig. 5). This was still so if patients
undergoing transplantation were considered deceased the
day of transplantation (83% vs. 47%; P .028). Other
factors predicting survival at univariate analysis were age
(P .007), albumin (P .008), past encephalopathy
(P .028), and alcoholic etiology of cirrhosis (P .032).
Bilirubin approached significance (P .086). In a mul-
tivariate analysis age, hemodynamic response and albu-
min were independent predictors of survival (Table 2).
Table 3. Causes of Death in Responders and
Nonresponders
Nonresponders Responders P Value
Portal hypertension complications
Bleeding 4 0
Hepatic encephalopathy 2 0
Hepatorenal syndrome 4 0
Total 10 0 .019
Other causes 4 1
Total 14 1 .003
NOTE. P values are those of log-rank test.
HEPATOLOGY, April 2003
Fig. 5. Eight-year cumulative probability of survival was higher in
responders than in nonresponders. Survival difference at 8 years was
43% (95% CI: 21%-63%) (R, responders; NR, nonresponders).
To confirm that hemodynamic response retains its pre-
dictive value independent of liver function, this variable
was introduced in a model together with Child-Pugh
score, as a global test of liver function and age. The 3
variables shown to be independent predictors (P .0005,
P .010, P .015, respectively), being relative hazard
for hemodynamic response in this model 18.249 (95%
CI: 2.165-153.833).
Discussion
Several studies have demonstrated that a reduction of
portal pressure 20% of baseline values or to 12 mm
Hg provides effective protection from the risk of first or
recurrent variceal bleeding.7-10,19 It is important to re-
mark that these observations came from studies in differ-
ent laboratories and in different countries, indicating that
measurement of the HVPG response is a reproducible
and robust test. Such hemodynamic response has been
subsequently considered the target of drug therapy for
portal hypertension.7,8,15 One study further demon-
strated that patients decreasing HVPG to 12 mm Hg
had significantly longer actuarial survival.4
In a previous longitudinal study in patients treated
with TIPS, we demonstrated that the same portal pressure
reduction that protected against recurrent bleeding did
also protect from development or worsening of ascites,
suggesting that the beneficial effects afforded by decreas-
ing portal pressure may extend to other manifestations of
portal hypertension.20 On the other hand, because most
studies examining the beneficial effect of portal pressure
reduction had only a short-term follow-up, it is not
known whether protection against recurrent bleeding,
HEPATOLOGY, Vol. 37, No. 4, 2003
and/or from other complications of portal hypertension,
persist on long-term follow-up.
The present study addressed these issues in a large se-
ries of patients included in prospective studies on long-
term pharmacologic therapy, with an 8-year follow-up.
The study confirms that patients showing a decline in
HVPG of at least 20% and/or to 12 mm Hg had a
much lower risk of recurrent variceal bleeding than those
not achieving these targets. We further demonstrated that
the lower risk of recurrent bleeding in responders was
maintained for the whole extent of the follow-up period,
supporting the concept that these patients should receive
lifelong pharmacologic therapy.21
The most salient finding of the study is that, further to
protect from recurrent variceal bleeding, such target re-
duction in HVPG protected from all other relevant com-
plications of portal hypertension. Thus, responders had a
much lower risk of developing or worsening ascites, spon-
taneous bacterial peritonitis, and hepatic encephalopathy,
and this resulted in better actuarial survival.
The question that arises is to what extent the better
outcomes of responders are due to the reduction in portal
pressure or if lack of response reflects more advanced,
resistant liver disease? For most outcomes, our results
suggest that lack of response was indeed a major determi-
nant of the worse prognosis observed in nonresponders, as
compared with responders. This is supported by our ob-
servation that, on multivariate analysis, lack of hemody-
namic response was an independent predictor of a worse
outcome in terms of rebleeding, developing of ascites and
spontaneous bacterial peritonitis, and survival. However,
multivariate analysis selected additional variables influ-
encing prognosis. Some of them, such as albumin, biliru-
bin, or prothrombin ratio are clearly related with the
degree of liver failure. Therefore, both severity of liver
disease and ability to reduce HVPG are major indepen-
dent determinants of prognosis in patients with advanced
cirrhosis receiving pharmacologic therapy for portal hy-
pertension. It should be noted, however, that the rela-
tively low sample size and number of events results in low
statistical power to detect associations of moderate (but
potentially clinically important) strength, in wide confi-
dence intervals, and in limited ability to examine multiple
covariates and to control for multiple confounding vari-
ables in the Cox proportional hazards models. Still, max-
imum efforts were taken to adjust the influence of
hemodynamic response by the degree of liver function
markers, by testing the individual components of Child-
Pugh score into the final models, maintaining them if
relevant changes in relative hazards were noted.
Regarding what makes a patient more likely to respond
to drug therapy, our study suggests that the worse the liver
ABRALDES ET AL. 907
disease, the less likely to be a responder. However, this
appears of limited relevance because parameters of liver
function were not independently associated with response
on multivariate analysis. Interestingly, the only indepen-
dent predictor was the greater dose of propranolol re-
ceived by responders than by nonresponders. This finding
has potential practical implications. Up to now, the dose
of propranolol was titrated against reduction in heart rate,
aiming empirically at 25% decrease,9,10,22 which reflects
pronounced -1-adrenoceptor blockade. However, the
beneficial effects of -blockers in portal hypertension are
due to blockade of both -1 and -2 adrenoceptors,23,24
and the degree of -1 blockade does not correlate with the
fall in portal pressure.25,26 Titration against heart rate
might not be adequate and may prevent some patients
from being given greater doses. Indeed, the reason why
the dose was not further increased in 40% of nonre-
sponders was having achieved this 25% fall in heart rate.
Should these patients have received a greater dose (i.e.,
titrated against clinical tolerance), the number of re-
sponders might have been greater.26
Drug therapy for portal hypertension should be aimed
at achieving the target reduction in portal pressure, min-
imizing the number of nonresponders. In the past, the
only way of approaching this problem has been to associ-
ate nitrates to nonselective -blockers.22,27 The use of
greater doses of -blockers may result in an increased
number of adverse effects. The same holds true with other
drugs or combinations with greater potential for HVPG
reduction than -blockers. Carvedilol,28 prazosin,29 and
the association of prazosin propranolol30 caused greater
falls in portal pressure than propranolol but also caused
significantly more adverse effects.28-30
The only current way of assessing whether a patient
treated with -blockers is a responder or not is by re-
peated HVPG measurements. Of course, it would be de-
sirable to have noninvasive ways of identifying
responders. Unfortunately, this has not been possible so
far,1,31,32 which has been used as an argument against
HVPG monitoring. On the contrary, we believe that the
fact that HVPG monitoring is unique in providing such
relevant prognostic information strongly supports a wider
use of this measurement. The fact that it predicts not only
rebleeding but also the risk of other complications and
survival may help in making such a policy more attractive
and, perhaps, more cost-effective. However, the objective
benefits of tailoring the treatment of portal hypertension
against HVPG response remains to be investigated.
In summary, in patients under continued pharmaco-
logic therapy for secondary prophylaxis of variceal bleed-
ing, a reduction in HVPG of 20% from baseline or to
12 mm Hg is associated with lower risk of developing
908 ABRALDES ET AL.
all the complications of portal hypertension and with im-
proved survival. Thus, HVPG monitoring provides rele-
vant, unique information on the prognosis of these
patients.
Acknowledgment: The authors thank Laura Rocabert,
Angels Baringo, and Rosa Saez for their cooperation in these
studies.
References
1. Bosch J, Garcia-Pagan JC. Complications of cirrhosis. I. Portal hyperten-
sion. J Hepatol 2000;32:141-156.
2. Viallet A, Marleau D, Huet M, Martin F, Farley A, Villeneuve JP, Lavoie
P. Hemodynamic evaluation of patients with intrahepatic portal hyperten-
sion. Relationship between bleeding varices and the portohepatic gradient.
Gastroenterology 1975;69:1297-1300.
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;5:419-424.
4. 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 [Comments]. Gastroenterology 1990;99:1401-1407.
5. Vorobioff J, Groszmann RJ, Picabea E, Gamen M, Villavicencio R, Bor-
dato J, Morel I, et al. Prognostic value of hepatic venous pressure gradient
measurements in alcoholic cirrhosis: a 10-year prospective study. Gastro-
enterology 1996;111:701-709.
6. Armonis A, Patch D, Burroughs A. Hepatic venous pressure measurement:
an old test as a new prognostic marker in cirrhosis? HEPATOLOGY 1997;25:
245-248.
7. Feu F, Garcia-Pagan JC, Bosch J, Luca A, Teres J, Escorsell A, Rodes J.
Relation between portal pressure response to pharmacotherapy and risk of
recurrent variceal haemorrhage in patients with cirrhosis. Lancet 1995;
346:1056-1059.
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. 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.
10. 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.
11. Teres J, Bosch J, Bordas JM, Garcia Pagan JC, Feu F, Cirera I, Rodes J.
Propranolol versus sclerotherapy in preventing variceal rebleeding: a ran-
domized controlled trial. Gastroenterology 1993;105:1508-1514.
12. McCormick PA, Feu F, Sabrin C, Planas R, Anglo-Spanish variceal re-
bleeding group. Propranolol and isosorbide mononitrate versus sclerother-
apy or shunt surgery for the prevention of variceal rebleeding: a
randomized trial. HEPATOLOGY 1994;20:106A.
13. 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.
14. Bosch J, Mastai R, Kravetz D, Navasa M, Rodes J. Hemodynamic evalu-
ation of the patient with portal hypertension. Semin Liver Dis 1986;6:309-
317.
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.
HEPATOLOGY, April 2003
16. Arroyo V, Gines P, Gerbes AL, Dudley FJ, Gentilini P, Laffi G, Reynolds
TB, et al. Definition and diagnostic criteria of refractory ascites and hepa-
torenal syndrome in cirrhosis. International Ascites Club. HEPATOLOGY
1996;23:164-176.
17. Rimola A, Garcia-Tsao G, Navasa M, Piddock LJ, Planas R, Bernard B,
Inadomi JM. Diagnosis, treatment and prophylaxis of spontaneous bacte-
rial peritonitis: a consensus document. International Ascites Club. J Hepa-
tol 2000;32:142-153.
18. Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM,
Kosberg CL, DAmico G, et al. A model to predict survival in patients with
end-stage liver disease. HEPATOLOGY 2001;33:464-470.
19. 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.
20. Casado M, Bosch J, Garcia-Pagan JC, Bru C, Banares R, Bandi JC, Es-
corsell A, et al. Clinical events after transjugular intrahepatic portosystemic
shunt: correlation with hemodynamic findings. Gastroenterology 1998;
114:1296-1303.
21. Abraczinskas DR, Ookubo R, Grace ND, Groszmann RJ, Bosch J, Garcia-
Tsao G, Richardson CR, et al. Propranolol for the prevention of first
esophageal variceal hemorrhage: a lifetime commitment? HEPATOLOGY
2001;34:1096-1102.
22. Garcia-Pagan JC, Escorsell A, Moitinho E, Bosch J. Influence of pharma-
cological agents on portal hemodynamics: basis for its use in the treatment
of portal hypertension. Semin Liver Dis 1999;19:427-438.
23. Mills PR, Rae AP, Farah DA, Russell RI, Carter DC. Comparison of three
adrenoceptor blocking agents in patients with cirrhosis and portal hyper-
tension. Gut 1984;25:73-78.
24. Hillon P, Lebrec D, Mun oz C, Jungers M, Goldfarb G, Banhamou JP.
Comparison of the effects of a cardioselective and a nonselective -blocker
on portal hypertension in patients with cirrhosis. HEPATOLOGY 1982;2:
528-531.
25. 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.
26. Garcia-Tsao G, Grace ND, Groszmann RJ, Conn HO, Bermann MM,
Patrick MJ, Morse SS, et al. Short-term effects of propranolol on portal
venous pressure. HEPATOLOGY 1986;6:101-106.
27. 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.
28. Banares R, Moitinho E, Piqueras B, Casado M, Garcia-Pagan JC, de Diego
A, Bosch J. Carvedilol, a new nonselective -blocker with intrinsic anti-
Alpha1-adrenergic activity, has a greater portal hypotensive effect than
propranolol in patients with cirrhosis. HEPATOLOGY 1999;30:79-83.
29. Albillos A, Lledo JL, Rossi I, Perez-Paramo M, Tabuenca MJ, Banares R,
Iborra J, et al. Continuous prazosin administration in cirrhotic patients:
effects on portal hemodynamics and on liver and renal function. Gastro-
enterology 1995;109:1257-1265.
30. 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.
31. Luca A, Garcia-Pagan JC, Feu F, Lopez-Talavera JC, Fernandez M, Bru C,
Bosch J, et al. Noninvasive measurement of femoral blood flow and portal
pressure response to propranolol in patients with cirrhosis. HEPATOLOGY
1995;21:83-88.
32. Merkel C, Sacerdoti D, Bolognesi M, Bombonato G, Gatta A. Doppler
sonography and hepatic vein catheterization in portal hypertension: assess-
ment of agreement in evaluating severity and response to treatment.
J Hepatol 1998;28:622-630.