SPECIAL ARTICLE
The Management of Ascites in Cirrhosis:
Report on the Consensus Conference of the
International Ascites Club
Kevin P. Moore,1 Florence Wong,2 Pere Gines,3 Mauro Bernardi,4 Andreas Ochs,5 Francesco Salerno,6
Paolo Angeli,7 Michael Porayko,8 Richard Moreau,9 Guadelupe Garcia-Tsao,10 Wladimiro Jimenez,11
Ramon Planas,12 and Vicente Arroyo3
Ascites is a common complication of cirrhosis, and heralds a new phase of hepatic decompensation in
the progression of the cirrhotic process. The development of ascites carries a significant worsening of
the prognosis. It is important to diagnose noncirrhotic causes of ascites such as malignancy, tubercu-
losis, and pancreatic ascites since these occur with increased frequency in patients with liver disease.
The International Ascites Club, representing the spectrum of clinical practice from North America to
Europe, have developed guidelines by consensus in the management of cirrhotic ascites from the early
ascitic stage to the stage of refractory ascites. Mild to moderate ascites should be managed by modest
salt restrictionanddiuretic therapywithspironolactoneoranequivalent inthefirst instance.Diuretics
should be added in a stepwise fashion while maintaining sodium restriction. Gross ascites should be
treated with therapeutic paracentesis followed by colloid volume expansion, and diuretic therapy.
Refractory ascites is managed by repeated large volume paracentesis or insertion of a transjugular
intrahepaticportosystemicstentshunt(TIPS).SuccessfulplacementofTIPSresults in improvedrenal
function,sodiumexcretion,andgeneralwell-beingof thepatientbutwithoutprovensurvivalbenefits.
Clinicians caring for these patients should be aware of the potential complications of each treatment
modality and be prepared to discontinue diuretics or not proceed with TIPS placement should
complications or contraindications develop. Liver transplantation should be considered for all ascitic
patients, and this should preferably be performed prior to the development of renal dysfunction to
prevent further compromise of their prognosis. (HEPATOLOGY 2003;38:258-266.)
Ascites occurs in more than 50% of patients within10 years of the diagnosis of cirrhosis. Cirrhoticascites accounts for over 75% of patients who
present with ascites, with the remaining 25% being due to
malignancy (10%), cardiac failure (3%), pancreatitis
(1%), tuberculosis (2%), or other rarer causes.1 There
have been several changes in the clinical management of
ascites over recent years. The recommendations put for-
ward in this document were agreed on by an International
Ascites Club Consensus Meeting on the management of
ascites. This meeting was supported by an unconditional
educational grant from Searle, Spain. These recommen-
dations have been updated in line with subsequent recent
publications of controlled clinical studies.
Diagnosis and Investigation of Ascites
All patients need investigation of the causes of ascites,
even when cirrhosis is suspected. Ascitic fluid should be
sent for the determination of albumin or protein concen-
tration. To diagnose spontaneous bacterial peritonitis
(SBP), ascitic fluid should be examined by microscopy
and inoculated directly into blood culture bottles. An as-
citic fluid neutrophil count of�250 polymorphonuclear
cells/mm3 is diagnostic of SBP, but a Gram stain of the
ascitic fluid is usually uninformative.2 Ascitic fluid from
patients with suspected malignant or pancreatic ascites
should be sent for cytology or measurement of amylase.
Abbreviations: SBP, spontaneous bacterial peritonitis; HRS, hepatorenal syn-
drome; PPH, postparacentesis effective hypovolemia; TIPS, transjugular intrahe-
patic portosystemic shunt.
From the 1Centre for Hepatology, Royal Free and University College Medical School,
UCL, London, United Kingdom; 2Division of Gastroenterology, Toronto General Hos-
pital, University of Toronto, Toronto, Canada; 3Liver Unit, Hospital Clinic, Barcelona
and the University of Barcelona School of Medicine, Barcelona, Spain; 4Department
Medicina Interna, Cardioangiologia, Epatologia, Alma Mater Studiorum–Universita`
di Bologna, Italy; 5Department of Internal Medicine, University of Freiburg, Freiberg,
Germany; 6Department of Internal Medicine, IRCCS Policlinico, University of Milan,
Milan, Italy; 7Department of Clinical and Experimental Medicine, University of
Padua, Padova, Italy; 8Department of Liver Transplantation, Thomas Jefferson Uni-
versity, Philadelphia, PA; 9INSERM U-481 et Service d’Hepatologie, Hopital Beaujon,
Clichy, France; 10Digestive Diseases Section, Yale University School of Medicine, New
Haven, CT; 11Hormal Laboratory, Hospital Clinic, Barcelona and the University of
Barcelona School of Medicine, Barcelona, Spain; and 12Department of Gastroenterol-
ogy, Hospital Universitari Germans Trias i Pujol, Badalona, Spain.
Received October 31, 2002; accepted May 7, 2003.
Address reprint requests to: Kevin Moore, M.D., Centre for Hepatology, Royal
Free and University College Medical School, UCL, Rowland Hill St., London
NW3 2PF, England. E-mail: kmoore@rfc.ucl.ac.uk; fax: (44) 207-433-2877.
Copyright © 2003 by the American Association for the Study of Liver Diseases.
0270-9139/03/3801-0034$30.00/0
doi:10.1053/jhep.2003.50315
258
The use of ascitic fluid protein in the differential diag-
nosis of the causes of ascites is overrated and misinter-
preted. Conventionally, the type of ascites is divided into
exudates and transudates in which the ascitic fluid protein
concentration is �25 g/L or �25 g/L, respectively, to
help in the differential diagnosis of the causes of ascites.
However, many physicians assume that cardiac ascites will
have a low ascitic protein, but this is rarely the case.3
Moreover, �15% of cases of cirrhotic ascites have an
ascitic protein�25 g/L, and 20% of patients with malig-
nancy have a low ascitic protein.4 The serum-ascites albu-
min gradient (i.e., serum albumin� ascitic fluid albumin
concentration) is more specific and sensitive at distin-
guishing ascites due to portal (sinusoidal) hypertension
(gradient �11 g/L) from that occurring as a result of
different pathogenetic mechanisms, such as inflammation
or peritoneal malignancy (gradient �11 g/L).5,6 Thus,
ascitic fluid protein classified the causes of ascites correctly
in 55% of cases, whereas serum-ascitic albumin gradient
assigns the causes correctly 97% of the time.5
Definitions
Uncomplicated Ascites. Uncomplicated ascites is as-
cites that is not infected and that is not associated with the
development of the hepatorenal syndrome (HRS). Grade
1 ascites is mild ascites only detectable by ultrasound ex-
amination. Grade 2 ascites or moderate ascites is manifest
by moderate symmetrical distension of abdomen. Grade 3
ascites is large or gross ascites with marked abdominal
distension.
Refractory Ascites. In 1996 the International Ascites
Club defined “refractory ascites” as ascites that cannot be
mobilized or the early recurrence of which cannot be sat-
isfactorily prevented by medical therapy.7 It occurs in ap-
proximately 5% to 10% of all cases of ascites.8 Two
subgroups were identified as diuretic-resistant ascites and
diuretic-intractable ascites.7 Additional findings fre-
quently include type II HRS, dilutional hyponatremia,
and wasting. The diagnostic criteria for refractory ascites
have been slightly revised and are shown in Table 1.
Prognosis of Cirrhosis With Ascites
The development of ascites in patients with cirrhosis
indicates a poor prognosis. The probability of death in
cirrhotic patients hospitalized with ascites is �40% at 2
years.8 The prognosis is worse for those with refractory
ascites and those who develop SBP.2
Treatment of Ascites: An Evidenced Based
Approach
The aim of treatment is to improve sodium balance or
circulatory function until liver transplantation or until
the disease runs its natural course. Patients with alcohol-
induced cirrhosis who stop drinking may have a consid-
erable improvement of liver function with resolution of
ascites.
Bed Rest
In patients with cirrhosis and ascites, upright posture
activates sodium-retaining systems and impairs renal per-
fusion and sodium excretion. In one study, bed rest im-
proved the response to diuretics.9 However, no clinical
trials have shown that bed rest actually improves the effi-
cacy of medical treatment.
Sodium and Water Restriction
A negative sodium balance can be achieved by dietary
salt restriction or by increasing renal sodium excretion.
With dietary salt restriction, loss of ascites occurs in 10%
to 15% of patients.1 The use of low salt diets is almost
universally recommended. However, this approach is not
backed by the results of controlled clinical trials.
Trials on Sodium Restriction. Severe sodium re-
stricted diets are unpalatable, leading to poor patient
compliance and poor nutrition. Five studies have com-
pared the efficacy of different dietary regimes.10-14 Some
societies readily adapt to a lower salt intake, whereas oth-
ers are less compliant because of cultural differences.
When severe dietary salt restriction (22 mmol/d) was
compared with a less restricted diet, dietary salt restriction
was associated with a shorter time for resolution of ascites
but was associated with a higher incidence of diuretic-
induced renal impairment and hyponatremia.10-13 In one
controlled study, a slightly reduced salt diet (120
mmoles/d) was equally effective in patients with ascites
when compared with a low-salt diet (50 mmol/d).14
There are no significant differences in survival between
patients receiving salt-restricted or -unrestricted diets, al-
Table 1. Revised Diagnostic Criteria of Refractory Ascites
1. Treatment duration: Patients must be on intensive diuretic therapy
(spironolactone 400 mg/d and furosemide 160 mg/d) for at least 1 week
and on a salt-restricted diet of less than 90 mmoles or 5.2 g of salt/d.
2. Lack of response: Mean weight loss of �0.8 kg over 4 days and urinary
sodium output less than the sodium intake.
3. Early ascites recurrence: Reappearance of grade 2 or 3 ascites within 4
weeks of initial mobilization.
4. Diuretic-induced complications: Diuretic-induced hepatic encephalopathy is
the development of encephalopathy in the absence of any other
precipitating factor. Diuretic-induced renal impairment is an increase of
serum creatinine by �100% to a value �2 mg/dL in patients with ascites
responding to treatment. Diuretic-induced hyponatremia is defined as a
decrease of serum sodium by �10 mmol/L to a serum sodium of �125
mmol/L. Diuretic induced hypo- or hyperkalemia is defined as a change in
serum potassium to �3 mmol/L or �6 mmol/L despite appropriate
measures.
HEPATOLOGY, Vol. 38, No. 1, 2003 MOORE ET AL. 259
though the survival of patients with previous gastrointes-
tinal bleeding was better in the low-salt group.13 A
limitation of all such studies is patient compliance and
being able to verify compliance.
Water Restriction. Dilutional hyponatremia occurs
in patients with decreased free water clearance, which is
driven by nonosmotic baroceptor secretion of antidiuretic
hormone secondary to effective central hypovolemia.15
Treatment of dilutional hyponatremia classically consists
of water restriction. However, there have been no clinical
trials to assess the effects of water restriction in patients
with cirrhosis and dilutional hyponatremia, and indeed
this treatment may exacerbate the central hypovolemia.
Diuretics
Anti-Mineralocorticoids. Secondary hyperaldoste-
ronism is a major factor promoting renal sodium reten-
tion16 in the distal tubules and collecting ducts of the
nephron. Both controlled and uncontrolled clinical trials
have shown that spironolactone is the drug of choice for
the initial treatment.11-14,17,18 Spironolactone or canreno-
ate (not available in North America) achieves a better
natriuresis more often than “loop” diuretics such as furo-
semide.18,19 The recommended initial dose of spironolac-
tone is 100 to 200 mg once daily. When severe
hyperaldosteronism (i.e., severe sodium retention) is
present, the dosage may be increased to 400 mg/d.18,20,21
A therapeutic response is observed in �75% of patients.
Gynecomastia is the main side effect of spironolactone,
but metabolic acidosis with or without hyperkalemia may
occur with renal impairment.22
Other Potassium-Sparing Diuretics. Amiloride and
triamterene also act in the distal tubule. No controlled
clinical trials on triamterene are available. Amiloride
(20-60 mg/d) has been shown to be less effective than
potassium canrenoate (150-500 mg/d).19
Loop Diuretics. Loop diuretics, such as furosemide,
are frequently used as an adjunct to spironolactone ther-
apy. The initial oral dose of furosemide is usually 20 to 40
mg/d, and is generally adjusted upward every few days up
to a maximum of 160 mg/d. Furosemide may cause po-
tassium depletion, metabolic hypochloremic alkalosis,
and hyponatremia, as well as hypovolemia, leading to re-
nal dysfunction.23,24
Assessing the Response to Diuretics. The mobiliza-
tion of ascites is best assessed by daily weighing of the
patient using accurate standardized weighing protocols.
The rate of weight loss should not exceed �0.5 kg/d in
the absence of edema, or �1 kg/d when edema is
present.24 Medical treatment based on sodium-restricted
diets, anti-mineralocorticoids, and loop diuretics achieves
a response rate in up to 90% of patients without renal
failure in controlled clinical trials.14,18,20
Paracentesis
Paracentesis is used to treat ascites that has not re-
sponded to medical therapy, to resolve large-volume as-
cites rapidly, to enable easier ultrasound examination in
patients with massive ascites, and to periodically treat re-
fractory ascites.
Trials of Paracentesis Versus Diuretics. There have
been 5 randomized controlled trials comparing therapeu-
tic paracentesis with diuretics in cirrhotic patients with
ascites.25-29 These studies compared repeated large vol-
ume paracentesis (5 L/d) with albumin infusion (6-8 g/L
of ascites removed), and showed that paracentesis was
more effective than diuretics in eliminating ascites and
shortened the duration of hospitalization. Moreover,
there were significantly fewer complications in the para-
centesis-treated group compared with those treated with
diuretics alone. These data have been confirmed by other
studies.30 The issue of whether paracentesis should be
repeated daily with 5-liter paracentesis or a single total
paracentesis has been resolved.31,32 Tito` et al.31 showed
that total paracentesis was as effective and as safe as re-
peated partial paracentesis. Paracentesis causes an acute
increase of cardiac output and a lowering of systemic vas-
cular resistance, leading to a modest reduction of blood
pressure.32 Pulmonary capillary wedge pressure decreases
at �6 hours postparacentesis, whereas the right atrial
pressure falls acutely following the onset of paracentesis,
secondary to a reduction of intrathoracic pressure.32 Since
postparacentesis effective hypovolemia (PPH) can occur
hours or days after the procedure (see below), volume
expansion should commence once the paracentesis has
been completed. Total paracentesis shortens the period of
hospitalization and can be done as an outpatient proce-
dure. However, paracentesis does not obviate the need for
diuretics. In one study, ascites recurred within 4 weeks
postparacentesis in 18% of patients receiving diuretics
immediately postparacentesis, compared with 93% in pa-
tients receiving a placebo.33
Controversy of Volume Expansion. There is only
one controlled trial comparing therapeutic paracentesis
with and without volume expansion with follow-up over
a few weeks. In this study, Gines et al. randomized pa-
tients to receive repeated paracentesis (�5 L/d) plus albu-
min or to repeated paracentesis alone.30 Side effects
occurred in 30% of patients treated with paracentesis
without albumin compared with 16% in those treated
with albumin. Complications included a high incidence
of renal impairment and hyponatremia and a marked el-
evation of plasma renin and aldosterone concentrations.
260 MOORE ET AL. HEPATOLOGY, July 2003
If volume expansion is not given following paracentesis,
patients may develop postparacentesis hypovolemia, lead-
ing to hyponatremia and renal impairment. Postparacen-
tesis volume expansion is recommended in patients with
cirrhosis and ascites. The choice of fluid is, however, con-
troversial. Human albumin solution is expensive and car-
ries the risk of infection with noneradicated viruses or
prion-related diseases. The use of albumin has been con-
tested by the Practice Guidelines Committee of the Amer-
ican Association for the Study of Liver Diseases.34 There
have been 5 randomized controlled trials comparing vol-
ume expansion with albumin with other forms of plasma
expanders, including dextrans, collagen-based colloids,
and hydroxyethyl starch.35-39 All studies have shown that
synthetic plasma expanders are as effective as albumin at
preventing the clinical complications of paracentesis,
namely hyponatremia or renal impairment. However,
Gines et al. showed that PPH, as defined by an increase in
plasma renin activity or aldosterone concentrations, was
prevented more effectively by albumin than synthetic
plasma expanders.35
Contraindications and Complications of Paracen-
tesis. Depite the fact that all published studies on para-
centesis have excluded patients with SBP, renal failure,
severe hepatic encephalopathy, thrombocytopenia, low
blood pressure, or severe jaundice, there is no evidence
that these complications should be considered as contra-
indications for paracentesis in clinical practice. Thus,
some physicians carry out a total paracentesis in patients
with SBP to remove the infected fluid. However, there are
no data to support this approach, and controlled studies
are needed. There are no data to support the correction of
mild coagulopathy with blood products prior to thera-
peutic paracentesis,40 but caution is needed when severe
thrombocytopenia is present. Acute complications fol-
lowing paracentesis are sporadic. Bleeding occasionally
occurs and may be fatal. Leakage of ascitic fluid should be
managed by placing a purse-string suture around the
opening and instructing the patient to lie with the punc-
ture site uppermost. The most common complication is
PPH and renal impairment. To date, there are no studies
identifying factors that predict the development of post-
paracentesis hypovolemia and renal impairment.
Transjugular Intrahepatic Portosystemic Shunt
(TIPS)
Short-Term and Long-Term Effects of TIPS on
Circulatory Function and Renal Function. TIPS, in
which a self-expanding shunt is inserted to create a shunt
between the hepatic vein (low pressure) and portal vein
(high pressure), can lead to an improvement of renal func-
tion and sodium excretion and the resolution of ascites.41
In the longer term it can also improve nitrogen balance
and patient well being.42-44 It has largely replaced the use
of surgically placed shunts. Insertion of a TIPS shunt
leads to a marked increase of cardiac output, right atrial
pressure, and pulmonary artery pressure,45 with a second-
ary decrease of systemic vascular resistance and an increase
in pulmonary vascular resistance and effective arterial
blood volume. Sodium excretion and renal function im-
prove over 4 weeks.46 Thus, serum creatinine decreased
from 1.5 to 0.9 mg/dL in patients with refractory ascites
post-TIPS.47
Complications. Immediate complications include
capsule puncture and intra-abdominal bleeding. Late but
common complications include shunt thrombosis and
stenosis. The development of hepatic encephalopathy oc-
curs in�30% of patients post-TIPS, but the incidence is
higher in those patients with pre-TIPS encephalopathy
and in those greater than 60 years old.48 TIPS increases
the cardiac preload45 and may precipitate cardiac failure
in pat