ARTICLE IN PRESS

G Model
YDLD-1934; No. of Pages 6

Digestive and Liver Disease xxx (2011) xxx–xxx

Contents lists available at ScienceDirect

Digestive and Liver Disease

journal homepage: www.elsevier.com/locate/dld

Liver, Pancreas and Biliary Tract

Prevention of paracentesis-induced circulatory dysfunction in cirrhosis: Standard

vs half albumin doses. A prospective, randomized, unblinded pilot study夽
Carlo Alessandria ∗ , Chiara Elia, Lavinia Mezzabotta, Alessandro Risso, Alida Andrealli,
Maurizio Spandre, Anna Morgando, Alfredo Marzano, Mario Rizzetto
Division of Gastroenterology and Hepatology, San Giovanni Battista Hospital – University of Turin, Turin, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Background: Paracentesis-induced circulatory dysfunction is a well-known complication of large volume
Received 17 March 2011 paracentesis. Albumin infusion (8 g of albumin/L of ascites removed) is effective in preventing it, but high
Accepted 1 June 2011 costs and scant availability limit its use.
Available online xxx
Aim: To compare standard vs half albumin doses.
Methods: Seventy cirrhotic patients treated with large volume paracentesis were randomized to receive
Keywords:
intravenous albumin as prevention of paracentesis-induced circulatory dysfunction: group 1 (35 patients)
Albumin
received 4 g/L of ascites removed, group 2 (35 patients) received 8 g/L of ascites removed.
Ascites
Circulatory dysfunction
Results: The incidence of paracentesis-induced circulatory dysfunction (14% vs 20% in group 1 and group
Cirrhosis 2, respectively; p = ns), hyponatremia (9% vs 6%, p = ns) and renal impairment (0% in both groups) on the
Paracentesis 6th day from paracentesis was similar between the two groups. After 6 months of follow-up, rates of
survival and of recurrence of ascites requiring large volume paracentesis were not different between the
two groups.
Conclusions: This unblinded, randomized, pilot study suggests that treatment with half doses of albumin
is effective in the prevention of paracentesis-induced circulatory dysfunction and its related clinical
complications in cirrhotic patients with tense ascites treated by large volume paracentesis. If confirmed,
these results could support a significant costs reduction in the management of ascites in cirrhotic patients.

© 2011 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.

1. Introduction probably multiple and still not completely known. Dynamics of

Total paracentesis is the first-line therapy for refractory ascites
in patients with cirrhosis [1,2] and it has been shown to be as
effective as standard therapy with diuretics in the management
of tense ascites, with significantly faster resolution and lower rate
of complications [3,4]. Paracentesis-induced circulatory dysfunc-
tion (PICD), a disorder characterized by marked activation of the
renin–angiotensin axis secondary to the further increase of an
already established arteriolar vasodilatation [5], is a frequent and
potentially harmful complication of large volume paracentesis [6].
It is associated with faster reaccumulation of ascites, renal impair-
ment and shorter survival [6]. The causes of this syndrome are
夽 The clinical trial was registered with the U.S. National Institute of Health database
(NCT00428506).
∗ Corresponding author at: Division of Gastroenterology and Hepatology, San Gio-
vanni Battista Hospital – University of Turin, Corso Bramante 88, 10126 Turin, Italy.
Tel.: +39 011 6335569/6335561; fax: +39 011 6335927.
E-mail address: carloalessandria@libero.it (C. Alessandria).
paracentesis (the rate of fluid extraction) [5,7], mechanical mod-
ifications (due to abdominal decompression) [7,8] and release of
vasodilator molecules, such as nitric oxide, from vascular endothe-
lium are thought to play a major role in development of PICD [5].
PICD appears in up to 80% of patients who are not infused with
plasma expanders after large volume paracentesis (LVP); plasma
volume expansion after paracentesis strongly reduces the inci-
dence of PIDC [9].
A controversial issue remains the kind of volume replacement
that should be given. Albumin is the most used plasma expander;
its safety and efficacy in preventing PICD is well demonstrated, as it
reduces the rate of development of PICD to about 15% [3,6,9]. How-
ever, costs and limited availability have prompted to the research
of alternatives. Many other options were tested, including different
plasma expanders (haemaccel [10], dextran-70 [6,11], polygeline
[6], dextran-40 [12], saline [13]) and vasoconstrictor agents (ter-
lipressin [14,15], noradrenalin [16], midodrine alone [17,18] or
combined with octreotide [19]): however, lower efficacy with
respect to albumin infusion [6,10–13], high costs [14,15], poten-
tial harmfulness [16] and controversial results [17–19] have made

1590-8658/$36.00 © 2011 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.dld.2011.06.001

Please cite this article in press as: Alessandria C, et al. Prevention of paracentesis-induced circulatory dysfunction in cirrhosis: Standard vs half
albumin doses. A prospective, randomized, unblinded pilot study. Dig Liver Dis (2011), doi:10.1016/j.dld.2011.06.001
 ARTICLE IN PRESS
G Model
YDLD-1934; No. of Pages 6

2 C. Alessandria et al. / Digestive and Liver Disease xxx (2011) xxx–xxx

albumin the plasma-expander of choice. It is recommended that
the albumin infusion consists of 8 g/L of ascites removed [1,2,20],
but there is no study comparing different doses of albumin in this
context.
The aim of this study was to compare standard (8 g/L of ascites
removed) vs half (4 g/L of ascites removed) albumin doses in the
prevention of PICD in patients with cirrhosis and ascites treated by
large volume paracentesis.
The study was approved by the local Investigation Committee
and written, informed consent was obtained from all patients.
3. Methods of measurement
Plasma renin activity and plasma aldosterone levels were mea-
sured by radioimmunoassay [21]. The normal values for PRA
and plasma aldosterone in our laboratory are 0.1–4 ng/mL/h and
12–150 pg/mL, respectively.

2. Patients and methods

4. Statistical analysis

This prospective study was conducted in the Gastro-Hepatology

The main end point of the study chosen to calculate the sample
Unit of our hospital between January 2004 and December 2008.
size was the incidence of paracentesis-induced circulatory dys-
During this five-year period all consecutive cirrhotic patients with
function. Because there are no data on the incidence of PICD after
tense ascites were investigated for inclusion. The diagnosis of
paracentesis and infusion of albumin 4 g/L of ascites removed, fol-
cirrhosis was based on clinical, laboratory and ultrasonographic
lowing the example of a previously published study [13] the sample
findings. Inclusion criteria were: cirrhosis with tense ascites sub-
size was calculated according to the study of Planas et al. [11], which
mitted to paracentesis >5 L; age between 18 and 75 years; written
showed an incidence of PICD of 15% when patients were treated
informed consent. Exclusion criteria were: multinodular hepato-
with albumin 8 g/L of ascites removed. Expecting no difference in
cellular carcinoma (>3 nodules), portal vein thrombosis, ongoing
the proportion of PICD after the infusion of albumin 4 g/L of ascites
bacterial infection, ongoing or recent (less than one week) bleeding,
removed, we calculated that with 35 patients per group the 95%
cardio-pulmonary failure, serum creatinine >2 mg/dL (176 ␮mol/L),
confidence interval of the difference would be between ±17%.
intrinsic renal disease, ongoing treatment with vasoactive drugs
Comparisons between the two groups were performed using
(including beta-blockers), recent use (within 30 days) of plasma
the Student’s t-test for continuous data and the 2 and Fisher test
expanders.
for categorical data. Comparisons of the variables in the same group
Patients were on low sodium diet (60–80 mEq/day) without
were performed using the Wilcoxon test and the 2 and Fisher test.
diuretic therapy for at least 3 days before paracentesis. On day 3
To identify variables predicting the development of PICD we
blood and 24-h urine samples were obtained for haematologic and
performed an univariate analysis by 2 test and Student’s test.
biochemical studies. Patients were in a bed rest supine position
Variables were then introduced in a multivariate analysis using a
for at least 45 min before blood samples were taken. Samples were
stepwise logistic regression model to identify independent predic-
immediately brought to the laboratory and centrifuged. The plasma
tors of development of PICD.
obtained was frozen (−80 ◦ C) until analysis.
Survival curves were calculated by the Kaplan–Meier method
Then a total paracentesis (>5 L) was performed under strict ster-
and compared with the log rank test. Patients submitted to TIPS
ile conditions as previously described [1]. Leukocytes count and
or liver transplantation during the follow-up were considered cen-
percentage of polymorphonuclear leukocytes were determined in
sored at the time of the procedures.
ascitic fluid.
Statistical analyses of the data were performed by using SPSS
All patients were given intravenous albumin (Uman Albumin
Statistical Software (SPSS Inc., Chicago, IL).
20 g of albumin/100 mL; Kedrion S.p.A., Barga, Lucca, Italy) within
Results are expressed as mean ± standard deviation. All p-values
the first hour after the procedure. Patients were randomly assigned
are 2 tailed, with values less than .05 considered statistically sig-
to treatment with albumin 4 g (group 1) or albumin 8 g (group
nificant.
2) per litre of ascites removed by paracentesis. Randomization
was made by using the sealed opaque envelopes method. Patients
with serum creatinine between 1.5 and 2 mg/dL were randomized 5. Definitions
independently from those with serum creatinine below 1.5 mg/dL
to ensure a similar number of cases with renal failure in both 5.1. Paracentesis-induced circulatory dysfunction
groups.
Patients did not receive diuretics during the 6 days following Increase in PRA of more than 50% of the preparacentesis value
paracentesis. At day 6 from paracentesis, blood and 24-h urine sam- to a level more than 4 ng/mL/h on the 6th day after paracentesis
ples were newly obtained for biochemical tests, including plasma [6].
renin activity and plasma aldosterone, following the same proce-
dure previously defined. 5.2. Renal failure
After day 6 from paracentesis diuretics were reintroduced as
needed and patients then discharged from hospital when medically Increase in serum creatinine ≥50% compared with the base-
fit. Patients were followed-up for 6 months or until transjugular line value to a level >1.5 mg/dL (132 ␮mol/L) on the 6th day after
intrahepatic portosystemic shunt (TIPS), liver transplantation or paracentesis [9]. Patients with a serum creatinine level >1.5 mg/dL
death. During follow-up patients were treated with diuretics and (132 ␮mol/L) at the inclusion in the study who showed an increase
large volume paracentesis followed by albumin infusions at stan- in serum creatinine ≥30% compared with the baseline values on
dard doses as needed. Patients with no or small varices did not the 6th day after paracentesis were also considered as patients
receive any kind of prophylactic treatment. Patients with medium developing renal failure.
or large varices and without any kind of prevention at inclusion
were treated with propranolol after day 6 from paracentesis and 5.3. Hyponatremia
continued this drug during the follow-up. Patients with contraindi-
cations to beta-blockers were treated with band ligation. Decrease in serum sodium ≥5 mEq/L to a level <130 mEq/L on
The study protocol conformed to the ethical guidelines of the the 6th day after paracentesis. Patients with a serum sodium level
1975 Helsinki declaration. <130 mEq/L at the inclusion in the study who showed a decrease

Please cite this article in press as: Alessandria C, et al. Prevention of paracentesis-induced circulatory dysfunction in cirrhosis: Standard vs half
albumin doses. A prospective, randomized, unblinded pilot study. Dig Liver Dis (2011), doi:10.1016/j.dld.2011.06.001
 ARTICLE IN PRESS
G Model
YDLD-1934; No. of Pages 6

C. Alessandria et al. / Digestive and Liver Disease xxx (2011) xxx–xxx 3

Fig. 1. Flow diagram of the study, with allocation to treatment and follow-up.

in serum sodium ≥5 mEq/L on the 6th day after paracentesis were Table 1
Demographic, clinical and laboratory data of all patients according to the group of
also considered as patients developing hyponatremia [9].
treatment at the time of inclusion.

Group 1 (4 g/L) Group 2 (8 g/L) p

6. Results
Age (years) 55 ± 12 57 ± 11 0.8
Sex (M/F) 30/5 27/8 0.5
During the study period 180 consecutive patients with cirrhosis
Alcohol (%) 34% 40% 0.8
who were admitted to the hospital for tense ascites were investi- N. with renal insufficiency* 3 2 0.9
gated for inclusion (Fig. 1). N. with hyponatremia# 3 5 0.9
One hundred and ten patients were not considered eligible for: Mean arterial pressure (mmHg) 84 ± 9 82 ± 11 0.7
Serum creatinine (mg/dL) 0.9 ± 0.3 0.8 ± 0.2 0.8
ongoing therapy with beta-blockers for the prevention of variceal
Serum creatinine (␮mol/L) 79 ± 24 70 ± 26 0.8
bleeding (29 patients, 16%), advanced hepatocellular carcinoma (25 Serum sodium (mEq/L) 136 ± 5 133 ± 4 0.7
patients, 14%), spontaneous bacterial peritonitis (SBP, 18 patients, Child–Pugh (points) 10 ± 2 10 ± 2 0.9
10%) or other bacterial infections (5 patients, 3%), of portal vein MELD score 16 ± 2 17 ± 2 0.8
thrombosis (13 patients, 7%), serum creatinine levels above 2 mg/dL Plasma renin activity (ng/mL h) 9.8 ± 7.4 7.6 ± 3.8 0.3
Plasma aldosterone (pg/mL) 278 ± 251 197 ± 157 0.2
(10 patients, 6%), gastrointestinal bleeding (2 patients, 1%), sig-
Plasma noradrenaline (pg/mL) 1019 ± 472 807 ± 872 0.3
nificant intrinsic renal disease (8 patients, 4%). The remaining 70
MELD: model for end-stage liver disease.
patients did not show any of the exclusion criteria, gave informed *
Renal insufficiency was defined as serum creatinine level greater than 1.5 mg/dL
written consent and were then enrolled in the study. They were ran- (132 ␮mol/L).
domly assigned to treatment with albumin 4 g/L of ascites removed #
Hyponatremia was defined as serum sodium level less than 130 mEq/L.
by paracentesis (35 patients; group 1) or albumin 8 g/L of ascites
removed (35 patients; group 2); they all received the allocated
rect index of reaccumulation of ascites, body weight was recorded
intervention. No patient was lost at follow-up. The final analysis
immediately after paracentesis and 6 days after paracentesis. The
of the trial included all the 70 patients.
increment of weight at day 6 was 4 ± 2.5 and 3.7 ± 3 kg in group 1
and 2, respectively (p = ns). The mean cost of the administered albu-
6.1. Characteristics at inclusion and data related to paracentesis min per paracentesis was 105 ± 42 Euros in group 1 and 227 ± 38
Euros in group 2, p < 0.0001. Duration of hospitalization (from ran-
6.1.1. Characteristics at inclusion domization to discharge from hospital) was also similar in both
Table 1 shows the characteristics of patients at study inclusion. groups (group 1, 9.4 ± 6 days; group 2, 8.6 ± 9 days; p = ns).
No significant differences were detected between the two groups Table 2 shows the systemic haemodynamic and laboratory vari-
regarding clinical data, liver and renal function tests, mean arterial ables before paracentesis as well as 6 days after the procedure. No
pressure, plasma renin activity and plasma aldosterone. significant variations in mean arterial pressure, activation of the
renin–angiotensin–aldosterone and sympathetic nervous systems,
6.1.2. Data related to paracentesis serum sodium and serum creatinine levels were observed in either
The volume of ascites removed was 7.4 ± 3.1 L (median: 8 L; group.
range 5.2–19 L) and 8 ± 3.7 L (median: 7.8 L; range: 5.4–16 L) in
group 1 and 2, respectively (p = ns). The amount of albumin infused 6.1.3. Complications
was significantly higher in group 2 than in group 1 according Table 3 shows the complications observed during the study
to the schedule (64.2 ± 23 vs 28.1 ± 14 g; p < 0.0001). As an indi- period (from paracentesis to day 6). Five of the 35 patients (14%) in

Please cite this article in press as: Alessandria C, et al. Prevention of paracentesis-induced circulatory dysfunction in cirrhosis: Standard vs half
albumin doses. A prospective, randomized, unblinded pilot study. Dig Liver Dis (2011), doi:10.1016/j.dld.2011.06.001
 ARTICLE IN PRESS
G Model
YDLD-1934; No. of Pages 6

4 C. Alessandria et al. / Digestive and Liver Disease xxx (2011) xxx–xxx

Table 2
Systemic haemodynamic and laboratory variables before and 6 days after paracentesis in each group of treatment.

Group 1 (4 g/L) Group 2 (8 g/L)

Day 0 Day 6 p Day 0 Day 6 p

Mean arterial pressure (mmHg) 84 ± 9 86 ± 10 0.8 82 ± 11 82 ± 6 0.9

Serum creatinine (mg/dL) 0.9 ± 0.3 0.8 ± 0.4 0.8 0.8 ± 0.2 0.9 ± 0.2 0.8
Serum creatinine (␮mol/L) 79 ± 24 70 ± 35 0.8 70 ± 19 79 ± 18 0.8
Serum sodium (mEq/L) 136 ± 5 135 ± 5 0.7 133 ± 4 134 ± 4 0.6
Plasma renin activity (ng/mL h) 9.8 ± 7.4 9 ± 9 0.7 7.6 ± 3.8 9 ± 6.1 0.4
Plasma aldosterone (pg/mL) 278 ± 251 316 ± 200 0.4 197 ± 157 146 ± 158 0.5
Plasma noradrenaline (pg/mL) 1019 ± 472 650 ± 604 0.3 807 ± 872 738 ± 740 0.7

Table 3
Complications observed during the study period (from paracentesis to day 6) in both groups of patients.

Group 1 (4 g/L) Group 2 (8 g/L) Differences (1–2)

N (%) 95%CI N (%) 95%CI % 95%CI

PICD 5/35 (14.3%) 4–30 7/35 (20%) 8–37 −5.7% −23.5 to 12.4
Renal impairment 0/35 (0%) 0–10 0/35 (0%) 0–10 0% −9.9 to 9.9
Hyponatremia 3/35 (9%) 2–23 2/35 (6%) 0–19 +2.9% −11.2 to 17.3
G-I bleeding 0/35 (0%) 0–10 0/35 (0%) 0–10 0% −9.9 to 9.9
Encephalopaty 0/35 (0%) 0–10 0/35 (0%) 0–10 0% −9.9 to 9.9
Infections 0/35 (0%) 0–10 0/35 (0%) 0–10 0% −9.9 to 9.9

PICD: paracentesis-induced circulatory dysfunction.

group 1 and 7 of the 35 patients (20%) in group 2 developed PICD
(p = ns). The number of complications other than PICD was similar
between the two groups. Three patients of group 1 (3/35, 9%) and 2
patients of group 2 (2/35, 6%) developed hyponatremia (p = ns). No
patient developed renal failure. We did not observe any episode of
gastrointestinal bleeding or hepatic encephalopathy.
In group 1, amongst the 5 patients who developed PICD none
developed hyponatremia (none of the 3 patients who developed
hyponatremia showed PRA levels variations consistent with the
diagnosis of PICD). In group 2, amongst the 7 patients with PICD
only one developed hyponatremia. No patient died during the 6-day
study period.
discharge. Readmission for ascites requiring large-volume para-
centesis occurred at a mean of 98 days (95% confidence intervals,
32–142 days) in patients of group 1 and at a mean of 112 days (95%
confidence intervals, 48–175 days) in group 2 (p = ns).
6.1.5.3. Hospital readmissions. The number of hospital readmis-
sions related to complications of cirrhosis and portal hypertension
was similar: 12 in group 1 (8 for tense ascites and 4 for hepatic
encephalopathy) and 14 in group 2 (9 for tense ascites, 4 for hepatic
encephalopathy and 1 for bleeding related to congestive gastropa-
thy).

6.1.4. Predictive factors for the development of PICD

6.1.5.4. Survival. Six patients died during the follow-up, 3 in each
All the variables recorded at study inclusion were analysed to
group. In 4 patients (2 in each group) liver failure was the main
evaluate predictive factors for the development of PICD. Patients
cause of death. The remaining 2 patients (1 in each group) died
were also analysed for Child score (A/B vs C) and MELD score (≥15 vs
for septic shock (due to Escherichia coli urinary tract infection in
< 15). Only mean arterial pressure was found to be significant in the
group 1 and to culture-negative spontaneous bacterial peritonitis
univariate analysis: 77 ± 9 vs 88 ± 10 mmHg in patients with and
in group 2). Amongst the remaining patients, 3 were transplanted (2
without PICD respectively; p .02. Mean arterial pressure was also
in group 1 and 1 in group 2) and 2 underwent TIPS (1 in each group).
found to have independent predictive value for the development
Patients were censored at the time of the procedures. Survival at
of PICD in the multivariate analysis (regression coefficient −0.125;
6 months (Fig. 2) was not significantly different between group 1
p .02).
(mean 148 ± 49 days) and group 2 (mean 158 ± 56 days; p = ns).

6.1.5. Follow-up
6.1.5.1. Varices and beta-blockers. Twenty-one patients (9 of group
1 and 12 of group 2; p = ns) had no or small varices and did not
receive any kind of prophylactic treatment. Forty-nine patients had
medium/large varices and any kind of prevention for variceal bleed-
ing at inclusion in the study. Amongst them, 45 started treatment
with propranolol after day 6 from paracentesis and continued this
drug during the follow-up. The remaining 4 patients (2 patients
in each group) had contraindications to beta-blockers and were
treated with band ligation.
6.1.5.2. Diuretics and paracentesis. The mean doses of spironolac-
tone during follow-up was 175 ± 96 vs 156 ± 76 mg/day for group
1 and 2 respectively (p = ns); furosemide: 62 ± 56 vs 53 ± 55 for
group 1 and 2 respectively (p = ns). There were no significant differ-
ences between the two groups in the frequency of paracentesis after
6.1.5.5. Beta-blockers vs non beta-blockers group. No differences
were found between patients who were and patients who were
not under beta-blockers. Amongst the 26 hospital readmissions, 17
were observed in patients under beta-blockers (17/45, 38%) and
9 in patients without beta-blockers (9/25, 36% (p = ns). Similarly,
amongst the six patients who died, 4 were under beta-blockers
(4/45, 9%) and 2 were not (2/25, 8%); p = ns.
6.1.5.6. Global costs. No significant differences were found regard-
ing the hospital costs. According to the Diagnosis Related Groups
system the costs for the hospital readmissions were 53.168 Euros in
group 1 (12 readmissions, mean cost for readmission: 4.430 ± 380
Euros) and 60.729 Euros in group 2 (14 readmissions, mean cost for
readmission: 4.338 ± .456 Euros); p = ns.

Please cite this article in press as: Alessandria C, et al. Prevention of paracentesis-induced circulatory dysfunction in cirrhosis: Standard vs half
albumin doses. A prospective, randomized, unblinded pilot study. Dig Liver Dis (2011), doi:10.1016/j.dld.2011.06.001
 ARTICLE IN PRESS
G Model
YDLD-1934; No. of Pages 6
C. Alessandria et al. / Digestive and Liver Disease xxx (2011) xxx–xxx
Fig. 2. Probability of survival in patients who received albumin 4 g/L of ascites evacuated (discontinuous line) and albumin 8 g/L of ascites evacuated (continuous line)
post-paracentesis.
7. Discussion
Paracentesis-induced circulatory dysfunction develops in up to
80% of cirrhotic patients with tense ascites treated by large volume
paracentesis if this is not followed by plasma volume expansion [9].
The development of PICD was associated with worsening of
renal function, reaccumulation of ascites, portal pressure increase
[6,22] and diminished survival [6,23].
Intravenous albumin infusion is effective in avoiding complica-
tions of paracentesis [9]; PICD development rate is reduced down to
about 15% [6,9]. Other plasma volume expanders were tested (syn-
thetic colloids, saline solution), but they were less effective than
albumin in the prevention of PICD [6,10–13].
The recommended dose of albumin is 8 g/L of ascites evacu-
ated. This schedule, proposed almost twenty years ago [3,20] to
best replace the amount of proteins removed with the ascitic fluid,
still remains a landmark in paracentesis [1,2]. It was calculated on
the basis of mean ascitic volume removed per tap and mean pro-
tein concentration in ascitic fluid in a series of cirrhotic patients (3).
However, following this schedule a high albumin amount is needed,
with remarkable costs.
To our knowledge no study was performed to compare the effi-
cacy of different doses of albumin in this context and ours is the
first randomized, controlled investigation comparing standard vs
reduced doses of albumin in the prevention of PICD. The dose
of 4 g of albumin for litre of ascites removed was chosen arbi-
trarily to halve the cost of treatment, as a first step towards the
definition of the best cost saving schedule to be used in this set-
ting.
At inclusion, no significant differences were detected between
patients treated with albumin 4 g/L of ascites removed and patients
treated with albumin 8 g/L of ascites removed regarding clinical
data, standard liver and renal function tests, and systemic haemo-
dynamics. The volume of ascitic fluid removed was similar. Thus,
the two groups were comparable in terms of liver failure, renal and
Please cite this article in press as: Alessandria C, et al. Prevention of paracentesis-induced circulatory dysfunction in cirrhosis: Standard vs half
albumin doses. A prospective, randomized, unblinded pilot study. Dig Liver Dis (2011), doi:10.1016/j.dld.2011.06.001
5
circulatory dysfunction and amount of fluid removed by paracen-
tesis.
PICD was prevented in most patients. The incidence of PICD dur-
ing hospitalization was not statistically different between group 1
(5/35 patients, 14%) and group 2 (7/35 patients, 20%) and was sim-
ilar to the rate of PICD reported in published studies that made use
of albumin infusion at the recommended dose [6,9]. However, the
cost of the administered albumin per paracentesis was markedly
reduced in group 1 (105 ± 42 Euros) compared to group 2 (227 ± 38
Euros). These data suggest that half doses of albumin are effective
in the prevention of PICD in cirrhotics, but at a distinctly diminished
cost.
After 6 days from paracentesis the effects on systemic haemo-
dynamic and laboratory variables were similar in the two groups.
Prevention of paracentesis-induced vasodilation explains the lack
of changes in effective circulating volume and in renal function in
both groups.
In patients receiving adequate plasma volume expansion after
large volume paracentesis, the reported incidence of development
of renal failure and of hyponatremia is low (5–15%) [6,9,13–15]. In
keeping with these data, 3 patients of group 1 (9%) and 2 patients of
group 2 (6%) developed hyponatriemia and none experienced renal
failure. We observed a poor correlation between the development
of hyponatremia and PICD. This apparently surprising finding is,
nevertheless, consistent with the literature [13] and may reflect
the fact that only a small proportion of patients with PICD develops
alterations of serum electrolytes [6,13] and that, on the contrary,
a spontaneous development of hyponatremia is often seen in cir-
rhotic patients with ascites, independently from PICD.
Of interest, our results confirm that variables estimating sys-
temic haemodynamics are very important in cirrhotic patients with
ascites [24], the mean arterial pressure representing the most reli-
able variable in predicting the development of PICD in our study.
It should be said that it is possible that some patients would
have required different amounts of albumin to be infused if we had
 ARTICLE IN PRESS
G Model
YDLD-1934; No. of Pages 6
6 C. Alessandria et al. / Digestive and Liver Disease xxx (2011) xxx–xxx
calculated them on the basis of albumin and proteins evacuated by
paracentesis. This kind of approach would be very interesting and
should be investigated.
Finally, a follow-up evaluation of 6 months showed that the
rate of survival and of recurrence of ascites and complications due
to portal hypertension requiring readmissions to hospital were
not different between the two groups. These data strengthen our
results, allowing to perform a rationale comparison of the effective-
ness of the two treatments and of the global cost of the management
of these patients.
We are aware that our study has flaws that somewhat limit its
value. The sample size is clearly low to demonstrate the equivalence
of the two treatments or the non-inferiority of the low dose albumin
schedule compared to the standard dose of albumin. Such a study
would require more than 100 patients per group.
Furthermore, our patients, although presenting with end-stage
liver disease and tense ascites, were not so compromised in terms
of circulatory and renal dysfunction. Consequently it must be
underlined that half doses of albumin seem to be effective in the
prevention of PICD in cirrhotics with stable clinical conditions and
relatively preserved cardiocirculatory function, whereas patients
with severe circulatory dysfunction and/or ongoing clinical decom-
pensations would maybe require the standard doses of albumin.
Therefore, the application of our results to more compromised
patients requires further investigations.
Taking into account all these defects we still think that our
study deserves consideration. It introduces an important element
of novelty as it explores a new approach to reduce the cost of
expansion with albumin. Furthermore, although the sample size
was frankly admitted to be low to demonstrate the equivalence of
the two schedules or the non-inferiority of the experimental treat-
ment compared to the standard one, differences and confidence
intervals of PICD rates and its related complications (Table 3) sug-
gest an effectiveness of the experimental arm with low albumin
doses.
In conclusion, the results of this unblinded, randomized, pilot
study suggest that treatment with half doses of albumin is effec-
tive and safe in the prevention of PICD and of its related clinical
complications in cirrhotic patients with tense ascites treated by
LVP, but with a significant cost reduction. These findings should
be considered as the first step in the investigation of the best cost
saving schedule of albumin to be used in cirrhotic patients after
large volume paracentesis and cannot be currently used to generate
formal recommendations. Larger appropriately powered equiva-
lence or non-inferiority studies are mandatory to confirm these
results.
Financial disclosure
None declared.
Conflict of interest statement
None declared.
List of abbreviations
LVP, large volume paracentesis; PICD, paracentesis-induced
circulatory dysfunction; PRA, plasma renin activity; SBP,
spontaneous bacterial peritonitis.
Please cite this article in press as: Alessandria C, et al. Prevention of paracentesis-induced circulatory dysfunction in cirrhosis: Standard vs half
albumin doses. A prospective, randomized, unblinded pilot study. Dig Liver Dis (2011), doi:10.1016/j.dld.2011.06.001
Acknowledgement
Supported by: Azienda Ospedaliero-Universitaria San Giovanni
Battista di Torino.
References
[1] Runyon BA, AASLD Practice Guidelines Committee. Management of
adult patients with ascites due to cirrhosis: an update. Hepatology
2009;49:2087–107.
[2] Moore KP, Wong F, Ginès P, et al. The management of ascites in cirrhosis: report
on the consensus conference of the International Ascites Club. Hepatology
2003;38:258–66.
[3] Gines P, Arroyo V, Quintero E, et al. Comparison between paracentesis and
diuretics in the treatment of cirrhotics with tense ascites. Results of randomized
study. Gastroenterology 1987;93:234–41.
[4] Garcia-Tsao G, Lim JK, Members of Veterans Affairs Hepatitis C Resource Cen-
ter Program. Management and treatment of patients with cirrhosis and portal
hypertension: recommendations from the Department of Veterans Affairs Hep-
atitis C Resource Center Program and the National Hepatitis C Program. Am J
Gastroenterol 2009;104:1802–29.
[5] Sola-Vera J, Such J. Understanding the mechanisms of paracentesis-induced
circulatory dysfunction. Eur J Gastroenterol Hepatol 2004;16:295–8.
[6] Ginès A, Fernández-Esparrach G, Monescillo A, et al. Randomized trial com-
paring albumin, dextran 70, and polygeline in cirrhotic patients with ascites
treated by paracentesis. Gastroenterology 1996;111:1002–10.
[7] Coll S, Vila MC, Molina L, et al. Mechanisms of early decrease in systemic
vascular resistance after total paracentesis: influence of flow rate of ascites
extraction. Eur J Gastroenterol Hepatol 2004;16:347–53.
[8] Cabrera J, Falcon L, Gorriz E, et al. Abdominal decompression plays a major role
in early postparacentesis haemodynamic changes in cirrhotic patients with
tense ascites. Gut 2001;48:384–9.
[9] Ginès P, Titó L, Arroyo V, et al. Randomized comparative study of therapeutic
paracentesis with and without intravenous albumin in cirrhosis. Gastroenterol-
ogy 1988;94:1493–502.
[10] Salerno F, Badalamenti S, Lorenzano E, et al. Randomized comparative study
of Hemaccel vs. albumin infusion after total paracentesis in cirrhotic patients
with refractory ascites. Hepatology 1991;13:707–13.
[11] Planas R, Ginès P, Arroyo V, et al. Dextran-70 versus albumin as plasma
expanders in cirrhotic patients with tense ascites treated with total paracen-
tesis. Results of a randomized study. Gastroenterology 1990;99:1736–44.
[12] García-Compean D, Blanc P, Larrey D, et al. Treatment of cirrhotic tense ascites
with Dextran-40 versus albumin associated with large volume paracentesis: a
randomized controlled trial. Ann Hepatol 2002;1:29–35.
[13] Sola-Vera J, Miñana J, Ricart E, et al. Randomized trial comparing albumin and
saline in the prevention of paracentesis-induced circulatory dysfunction in
cirrhotic patients with ascites. Hepatology 2003;37:1147–53.
[14] Moreau R, Asselah T, Condat B, et al. Comparison of the effect of terlipressin and
albumin on arterial blood volume in patients with cirrhosis and tense ascites
treated by paracentesis: a randomised pilot study. Gut 2002;50:90–4.
[15] Singh V, Kumar R, Nain CK, et al. Terlipressin versus albumin in
paracentesis-induced circulatory dysfunction in cirrhosis: a randomized study.
J Gastroenterol Hepatol 2006;21:303–7.
[16] Singh V, Kumar B, Nain CK, et al. Noradrenaline and albumin in paracentesis-
induced circulatory dysfunction in cirrhosis: a randomized pilot study. J Intern
Med 2006;260:62–8.
[17] Singh V, Dheerendra PC, Singh B, et al. Midodrine versus albumin in the
prevention of paracentesis-induced circulatory dysfunction in cirrhotics: a ran-
domized pilot study. Am J Gastroenterol 2008;103:1399–405.
[18] Appenrodt B, Wolf A, Grünhage F, et al. Prevention of paracentesis-induced
circulatory dysfunction: midodrine vs albumin. A randomized pilot study. Liver
Int 2008;28:1019–25.
[19] Karwa R, Woodis CB. Midodrine and octreotide in treatment of cirrhosis-related
hemodynamic complications. Ann Pharmacother 2009;43:692–9.
[20] Tito L, Gines P, Arroyo V, et al. Total paracentesis associated with intravenous
albumin management of patients with cirrhosis and ascites. Gastroenterology
1990;98:146–51.
[21] Gines P, Arroyo V, Vargas V, et al. Paracentesis with intravenous infusion of
albumin as compared with peritoneovenous shunting in cirrhosis with refrac-
tory ascites. N Engl J Med 1991;325:829–35.
[22] Ruiz-del-Arbol L, Monescillo A, Jimenez W, et al. Paracentesis-induced circula-
tory dysfunction: mechanism and effect on hepatic hemodynamics in cirrhosis.
Gastroenterology 1997;113:579–86.
[23] Cardenas A, Gines P, Runyon BA. Is albumin infusion necessary after large vol-
ume paracentesis? Liver Int 2009;29:636–40.
[24] Llach J, Ginès P, Arroyo V, et al. Prognostic value of arterial pressure,
endogenous vasoactive systems, and renal function in cirrhotic patients admit-
ted to the hospital for the treatment of ascites. Gastroenterology 1988;94:
482–7.