Research Article

Effects of a selective vasopressin V2 receptor antagonist, satavaptan,on ascites recurrence after paracentesis in patients with cirrhosisq

Florence Wong1,*, Pere Gines2, Hugh Watson3, Yves Horsmans4, Paolo Angeli5, Paul Gow6,Pascal Minini3, Mauro Bernardi7

1Department of Medicine, University of Toronto, Canada; 2Hospital Clinic, University of Barcelona, Institut d’Investigacions BiomèdiquesAugust Pi-Sunyer (IDIBAPS), CIBER de Enfermedades Hepáticas y Digestivas (CIBEREHD), Barcelona, Catalunya, Spain; 3Internal

Medicine, Clinical Development, Sanofi-Aventis R&D, Chilly-Mazarin, France; 4Saint-Luc University Hospital, Brussels, Belgium; 5University ofPadova, Italy; 6Austin Hospital, University of Melbourne, Australia; 7Department of Internal Medicine, University of Bologna, Italy

See Editorial, pages 225–227

Background & Aims: Cirrhotic patients with recurrent ascites rienced major adverse events in all groups. Increases in serum

frequently require paracentesis despite diuretic therapy. Vaso-pressin receptor antagonists, by increasing free water clearance,may reduce the recurrence of ascites. To investigate the effectsof the addition of a vasopressin V2 receptor antagonist, satavap-tan, to 100 mg spironolactone on ascites recurrence after a largevolume paracentesis in patients with liver cirrhosis irrespectiveof the presence of hyponatraemia.Methods: One hundred and fifty one cirrhotic patients withrecurrent ascites with or without hyponatraemia, and normalto mildly abnormal renal function were randomised in a dou-ble-blind study to receive either 5 mg (n = 39), 12.5 mg (n = 36),25 mg (n = 40) of satavaptan or placebo (n = 36) for 12 weeks.Their Child–Pugh scores were 9.2 ± 1.3, 8.7 ± 1.7, 8.8 ± 1.3, and9.0 ± 1.5, respectively.Results: Median time to first paracentesis was 23, 26, and17 days with satavaptan 5, 12.5, and 25 mg, respectively, versus14 days with placebo (ns for all doses). The frequency of para-centeses was decreased significantly (p < 0.05) in all satavaptangroups versus placebo. Mean increase in ascites was2.82 ± 0.48 L/week for placebo versus 2.12 ± 0.40, 2.14 ± 0.33,and 2.06 ± 0.40 L/week for the 5, 12.5, and 25 mg of satavaptan,respectively (ns for all doses). Similar numbers of patients expe-

Journal of Hepatology 20

Keywords: Cirrhosis; Ascites; V2 receptor antagonist; Large volume paracentesis.Received 17 January 2010; received in revised form 22 February 2010; accepted 23February 2010; available online 24 May 2010DOI of original article: 10.1016/j.jhep.2010.04.003.q The study was registered on a public clinical trial registry website, www.Clin-icalTrials.org, number NCT 00501384.* Corresponding author. Address: 9th floor, North Wing, Room 983, TorontoGeneral Hospital, 200 Elizabeth Street, Toronto, Ontario, Canada M5G2C4. Tel.: +1416 340 3834; fax: +1 416 340 5019.E-mail address: florence.wong@utoronto.ca (F. Wong).Abbreviations: AST, aspartate transaminase; ALT, alanine transaminases; ALP,alkaline phosphatase; CI, confidence interval; dDAVP, 1-desamino-8-D-argininevasopressin; INR, international normalized ratio; ITT, intention to treat; LVP, largevolume paracentesis; MELD, model for end-stage liver disease; QTcF, QT intervalcorrected by the Fridericia formula; TIPS, transjugular intrahepatic portosystemicstent shunt.

creatinine, orthostatic changes in systolic pressure and thirstwere more common with satavaptan.Conclusions: Satavaptan has the potential to reduce recurrenceof ascites after a large volume paracentesis at doses from 5 to25 mg in cirrhotic patients with ascites.� 2010 European Association for the Study of the Liver. Publishedby Elsevier B.V. All rights reserved.

Introduction

The development of ascites in the natural history of cirrhosisheralds a worsening of the prognosis to 50% survival at 2 years[1], and this deteriorates to 30–50% at 1 year when the ascitesbecomes refractory to medical therapy [2,3]. Traditionally,the management of ascites consists of dietary sodium restric-tion, judicious use of diuretics, and large volume paracentesis(LVP) [4]. Although repeated LVPs have been shown to be safe[5,6], frequent LVPs involve significant medical manpower andinconvenience to the patients. Other treatment options, includ-ing the insertion of a transjugular intrahepatic portosystemicstent shunt, are only suitable for carefully selected patients [7].

Vasopressin V2 receptor antagonists are aquaretic agents. Byantagonising the antidiuretic effects of vasopressin at the V2

receptor located in the renal collecting duct, they increase freewater clearance, and thus may be helpful in mobilising excesswater in conditions associated with water retention including cir-rhosis [8–10]. The use of V2 receptor antagonists in cirrhosis withascites has be shown to be safe and efficacious, with a dose-dependent increase in urinary volume and a reduction in urinaryosmolality [8,9]. Furthermore, while cirrhotic patients with asci-tes on placebo gained weight, those on aquaretic agents main-tained their weight with the lower doses, and clearly lostweight with the higher doses [8,9].

Satavaptan is a selective V2 receptor antagonist, whichincreased urine output when administered with spironolactone incirrhotic patients with ascites and hyponatraemia [11], associated

10 vol. 53 j 283–290

Research Article

with improved ascites control over a 14-day period [11]. It is pos-tulated that satavaptan, together with spironolactone, is able toenhance ascites reduction by increasing free water excretion inpatients with cirrhosis irrespective of the presence of hyponatra-emia over a longer time period.

Therefore, the aim of this study was to assess the efficacy ofsatavaptan at three fixed doses on ascites recurrence after anLVP in cirrhotic patients with recurrent ascites being treated withspironolactone over a 12-week period.

Patients and methods

Patients

Cirrhotic patients with ascites, with or without hyponatraemia, who had at leastone LVP in the previous 3 months, as well as undergoing a LVP on the day ofentry, each of P 4 L, were assessed for enrolment.

Exclusion criteria were serum bilirubin >135 lmol/L, international normal-ized ratio (INR) >3.0, serum creatinine >175 lmol/L, serum sodium >142 mmol/L, serum potassium P 5.5 mmol/L, or systolic arterial pressure of <80 mmHg.Clinical exclusion criteria were significant cardiac diseases such as recent myo-cardial infarction (61 month), or prolonged QT interval corrected by the Frider-icia formula (QTcF) of P500 ms; complications of liver cirrhosis such as recentinfection including spontaneous bacterial peritonitis, or gastrointestinal bleed(610 days from randomisation), or ongoing hepatic encephalopathy of >grade1 [12], or known hepatocellular carcinoma of >5 cm in maximal diameter for1 lesion or <3 cm for each of 63 lesions. Patients who had a liver transplant,or a portosystemic shunt were not eligible, as were patients with non-cirrhoticascites.

Methods

The study was designed and developed by the Internal Medicine team at Sanofi-Aventis, together with Drs. Wong and Gines. Dr. Wong prepared the first and sub-sequent drafts of the manuscript with Dr. Bernardi. The data are held and ana-lyzed by Sanofi-Aventis, but accessible to Drs. Wong and Gines. Decisionsrelating to the final draft were made by Drs. Wong and Watson.

Ethics approval was obtained from all participating institutions. All patientsgave written informed consent. The study was registered on a public clinical trialregistry website, www.ClinicalTrials.org, number NCT 00501384.

This was a double-blind, randomised, placebo-controlled, parallel-groupstudy assessing the efficacy of three fixed doses of satavaptan, 5, 12.5, and25 mg versus placebo, plus low-dose spironolactone, in the prevention of ascitesrecurrence after a LVP of P4 L. On day �1, whilst inpatients, all patients under-went a complete physical examination, and a LVP of P4 L, plus albumin infusionat 6–8 g/L of ascitic fluid removed [4], together with complete biochemistry, acomplete blood count and an electrocardiograph to exclude prolonged QTc. A24-h urine collection was done to determine urinary volume, osmolality and elec-trolyte excretion. The following day (day 1), eligible patients were randomised toreceive a single daily dose of either 5, 12.5 or 25 mg of satavaptan or placebo, plus100 mg spironolactone per day for a total of 12 weeks. A central randomisationlist was generated electronically at Sanofi-Aventis with each site receiving a blockof equally distributed sealed treatment groups. At randomization, the lowestavailable treatment number at that particular site was selected for the patient,thus ensuring a random allocation of patients to each treatment group.

On days 1–3, patients were assessed clinically daily, as well as for serum elec-trolytes, renal function, serum and urinary osmolality, urinary volume and uri-nary electrolyte excretion. Patients were discharged on day 3, if possible.Outpatient visits occurred on days 7, 14, 28, 56, and 84, when the same laboratoryparameters that were assessed during the first 3 days were repeated. 24-h urinecollections were only done on days 28 and 84. Plasma vasopressin, supine aldo-sterone and supine plasma renin levels were measured on day 1 prior to studymedication, and again on days 7 and 84. An electrocardiograph was repeatedon days 2, 7, 28, and 84 of the study. LVP was permitted, plus albumin infusionat a dose of 6–8 g/L of ascitic fluid removed if patients had gained P4 kg inweight, accompanied by tense ascites. Patients were maintained on a sodiumrestriction of 688 mmol/day and were instructed to drink water as required bythirst. Throughout the study, patients were assessed for electrolyte abnormalities,renal dysfunction, dehydration and thirst.

The co-primary endpoints were time to the first paracentesis, and increase inascites as assessed by the increase in body weight and cumulative volume of asci-tes removed during the 12-week study period. Secondary endpoint was frequency

284 Journal of Hepatology 201

of paracentesis. Subgroup analysis of patients with refractory ascites wasperformed.

Laboratory assays

Serum and urinary electrolytes, complete blood count, INR, and liver function testswere performed using standard automated laboratory techniques. Blood samplesfor vasopressin, plasma renin, and aldosterone concentrations were collected onice. Plasma was separated by refrigerated centrifugation and stored at�70 �C untilassay. Serum and urinary osmolality were measured with a freezing point osmom-eter. Plasma renin was measured using an immuno-chemiluminometric assay(Nichols Kit & Instrument, San Clemente, CA). Plasma aldosterone (Coat-A-CountAldosterone kit, Diagnostic Products Corporation, Los Angeles, CA), and vasopres-sin (Quest Diagnostics, Madison, NJ) were assayed by radioimmunoassay.

Sample size calculation and statistical analysis

Sample size calculations were based on the two co-primary endpoints. For thefirst primary endpoint, it was assumed that the time to first paracentesis was>12 weeks in 20% of placebo patients, versus 60% of satavaptan patients, the num-ber required for each group would be 33 patients, or 132 patients for the fourstudy groups. For the second primary endpoint, it was assumed that placebopatients would gain 4 kg more ascites over the 12-week period versus satavaptanpatients. The number required for each group would be 35 patients, or 140patients for the four study groups. The primary analysis was conducted on theintent-to-treat (ITT) population. Cumulative mean number of paracenteses as afunction of time was estimated using the Nelson–Aalen estimator (an extensionof the Kaplan–Meier estimator for recurrent events). Comparisons between treat-ment groups were performed using a Cox model for recurrent events with robustestimate of the variance (sandwich estimate) [13]. Change from baseline in 24-hurine volume was analysed using an analysis of covariance with treatment groupas main factor and baseline value as covariate. Categorical variables were com-pared using the Fisher’s exact test. Calculations were performed with the statis-tical program SAS 8.2 (Cary, NC, USA). Results are presented as mean ± SD.Median values with ranges were presented for nonparametric variables.p < 0.05 was considered statistically significant.

Results

Of 173 potentially suitable patients who were consented andscreened between April 2004 and February 2005, 20 patients didnot meet inclusion/exclusion criteria. One patient withdrew con-sent, and one patient did not return for randomisation. Therefore,151 patients were entered into the study, randomised to receiveplacebo (n = 36), 5 mg (n = 39), 12.5 mg (n = 36) or 25 mg(n = 40) of satavaptan (Fig. 1). Patient demographics, baseline lab-oratory parameters are presented in Tables 1 and 2, respectively.

Control of ascites

Ascites was previously present in all patients for a mean period of>1 year (placebo: 16.7 ± 17.3 months, satavaptan 5 mg: 19.5 ±20.2 months, 12.5 mg: 27.6 ± 30.9 months, 25 mg: 33.1 ± 36.0months), with a previous median frequency of LVP of every 15 days(Table 1). The median time to the first LVP was increased to 23 days[95% confidence interval (CI): 16, 31 days], 26 days [95% CI, 16,56 days] and 17 days [95% CI: 14, 28 days] with 5, 12.5, and25 mg of satavaptan, respectively, versus 14 days with placebo[95% CI: 8, 29 days] (NS for all doses). The mean increase in asciteswas 2.82 ± 0.48 L/week for placebo versus 2.12 ± 0.40, 2.14 ± 0.33,and 2.06 ± 0.40 L/week for the 5, 12.5, and 25 mg of satavaptanrespectively (NS for all doses), with respective relative risks for afirst repeat LVP of 0.69, 0.60, and 0.63 for the three satavaptan doses.

Over the 12-week study period, the total number of LVPs,adjusted for the duration of assessment, was significantlyreduced at each dose of satavaptan: 3.11 LVPs (p = 0.026), 2.95LVPs (p = 0.018) and 2.72 LVPs (p = 0.017) for 5, 12.5, and

0 vol. 53 j 283–290

Screened patientsn = 173

Screened failuresn = 22

Randomized patientsn = 151

Patients who receiveddouble blind treatment

n=151

Placebon = 36

Satavaptan(5 mg/day)

n = 39

Satavaptan(12.5 mg/day)

n = 36

Satavaptan(25 mg/day)

n = 40

Completed20 (56%)

Completed19 (49%)

Completed27 (75%)

Completed28 (70%)

Discontinued = 16: Lack of efficacy = 3: Adverse event = 5: Poor compliance = 2: Subject's request = 4: Other reasons = 2

Discontinued = 12: Lack of efficacy = 2: Adverse event = 9: Poor compliance = 0: Subject's request = 1: Other reason = 0

Discontinued = 20: Lack of efficacy = 4: Adverse event = 13: Poor compliance = 1: Subject's request = 0: Other reason = 2

Discontinued = 9: Lack of efficacy = 3: Adverse event = 3: Poor compliance = 1: Subject's request = 1: Other reason = 1

Fig. 1. Distribution of patients who were screened for enrollment into the study.

Table 1. Patient demographics.

Satavaptan

Placebo 5 mg 12.5 mg 25 mg

n 36 39 36 40Age (y) 60 ± 9 59 ± 9 59 ± 9 60 ± 11M/F 26/10 32/7 28/8 29/11Aetiology of cirrhosis

Alcohol 26 30 29 26Hepatitis B 2 4 4 2Hepatitis C 9 8 6 9Others 4 2 2 5

Child–Pugh score 9.0 ± 1.5 9.2 ± 1.3 8.7 ± 1.7 8.8 ± 1.3MELD score 14.2 ± 4.3 15.6 ± 4.2 14.0 ± 3.9 14.2 ± 3.2Ascites classification

Refractory [4] 22 24 26 22Recidivant 13 14 10 17Neither 1 1 0 1

Median interval between paracenteses (days) 15 13 15 16

Recidivant ascites: recurrent ascites requiring large volume paracentesis for >3 times per year.

JOURNAL OF HEPATOLOGY

Journal of Hepatology 2010 vol. 53 j 283–290 285

Table 2. Baseline laboratory data of study patients.

Satavaptan

Placebo 5 mg 12.5 mg 25 mg

n 36 39 36 40Serum sodium (N = 135–145 mmol/L) 136 ± 4 133 ± 6 135 ± 5 134 ± 5Serum potassium (N = 3.5–4.5 mmol/L) 4.2 ± 0.7 4.1 ± 0.6 4.2 ± 0.6 4.5 ± 0.6Serum creatinine (N <100 lmol/L) 97 ± 30 94 ± 28 100 ± 27 100 ± 28Haemoglobin (N = 8.1–11.2 mmol/L for men 7.4–9.9 mmol/L for women) 6.9 ± 1.3 6.7 ± 1.4 7.2 ± 1.0 6.9 ± 3.7INR (N = 1.0–1.2) 1.5 ± 0.4 1.5 ± 0.3 1.4 ± 0.3 1.4 ± 0.3AST (N <35 IU/L) 49 ± 31 53 ± 24 44 ± 22 54 ± 38ALT (N <40 IU/L) 29 ± 17 33 ± 24 24 ± 11 33 ± 30ALP (N <110 IU/L) 149 ± 98 166 ± 96 195 ± 189 184 ± 174Serum albumin (N = 35–50 g/L) 30 ± 6 29 ± 6 30 ± 5 30 ± 5Serum bilirubin (N <17 lmol/L) 39 ± 28 49 ± 30 35 ± 21 33 ± 20Urinary volume (ml) 1015 ± 455 1214 ± 784 1202 ± 713 1283 ± 749QTcF (ms) 412 ± 31 417 ± 27 420 ± 27 420 ± 26

INR: international normalized ratio; AST: aspartate transaminase; ALT: alanine transaminase; ALP: alkaline phosphatase; QTcF: QT interval corrected by Fridericia’sformula.

Research Article

25 mg, respectively versus 4.32 LVPs for placebo. The relativerisks for any paracentesis were 0.69, 0.66, and 0.63, respectivelyfor the three satavaptan doses (Fig. 2). Allowing for drop-outs, theKaplan–Meier analysis estimated that 20–25% of satavaptanpatients versus 11% of placebo patients had no further paracent-eses during the 12-week study period.

For the subpopulation of patients with refractory ascites, themean number of paracenteses over the 12-week period was alsosignificantly reduced for the 5 mg (p = 0.015) and for the 12.5 mggroup (p = 0.017), though not for the 25 mg group (p = 0.103). Inaddition, the relative risks for requiring a first paracentesis in thissubpopulation of patients were 0.51 (p = 0.049), 0.47 (p = 0.023),

Table 3. Systemic haemodynamics, renal function, and sodium and water homeosta

Placebo

n 36Systolic blood pressure (N = 110–140 mm Hg)

Baseline 116 ± 14Month 1 116 ± 14Month 3 117 ± 13

Diastolic blood pressure (N = 60–90 mm Hg)Baseline 69 ± 10Month 1 71 ± 9Month 3 71 ± 9

Serum [Na] (N = 135–145 mmol/L)Baseline 136 ± 4Month 1 136 ± 4Month 3 135 ± 5

Urinary [Na] (mmol/hr)Baseline 2.2 ± 2.3Month 1 3.7 ± 5.5Month 3 2.9 ± 3.2

Serum osmolality (N = 285–298 mOsmol/kg of H2O)Baseline 291 ± 9Month 1 290 ± 14Month 3 285 ± 23

Urinary osmolality (mOsmol/kg of H2O)Baseline 522 ± 215Month 1 533 ± 188Month 3 519 ± 198

Serum creatinine (N <110 lmol/L)Baseline 97 ± 30Month 1 93 ± 29Month 3 94 ± 32

a p <0.01 compared to baseline.

286 Journal of Hepatology 201

and 0.70 (ns) for the three satavaptan doses, respectively, whencompared with placebo.

Renal function and sodium and water homeostasis

Throughout the study, there were no significant changes in themean serum creatinine, serum sodium or urinary sodium excre-tion (Table 3). At entry, there were 61 patients who had hypona-traemia (Placebo: 14, 5 mg: 17, 12.5 mg: 11, 25 mg: 19 patients).All hyponatraemic patients on satavaptan, irrespective of dose,had a significant increase in serum sodium by day 7 comparedto the placebo group. Since the majority of patients in this study

sis in all study patients.

Satavaptan

5 mg 12.5 mg 25 mg

39 36 40

116 ± 16 119 ± 17 119 ± 19122 ± 10 118 ± 14 121 ± 18126 ± 18 121 ± 15 119 ± 14

68 ± 11 73 ± 11 71 ± 1372 ± 7 72 ± 12 71 ± 1073 ± 10 75 ± 13 73 ± 9

133 ± 6 135 ± 5 134 ± 5134 ± 5 136 ± 6 137 ± 4136 ± 4 136 ± 4 137 ± 4

2.6 ± 3.5 2.7 ± 2.9 2.4 ± 2.83.2 ± 3.1 3.2 ± 3.1 3.2 ± 3.42.8 ± 3.2 4.2 ± 3.1 3.3 ± 3.1

287 ± 18 289 ± 11 283 ± 11289 ± 17 288 ± 12 294 ± 14290 ± 12 292 ± 12 297 ± 14

544 ± 184 461 ± 185 420 ± 175364 ± 158a 285 ± 138a 203 ± 101a

452 ± 201 299 ± 134a 227 ± 106a

94 ± 28 100 ± 27 100 ± 2893 ± 36 101 ± 27 102 ± 3595 ± 36 99 ± 38 98 ± 37

0 vol. 53 j 283–290

N at risk

Mea

n cu

mul

ativ

e nu

mbe

r of

par

acen

tesi

s

5

4

3

2

1

0

36393640

Placebo5 mg

12.5 mg25 mg

36393638

31343636

31323433

28293433

24263330

23243229

22223228

22213228

22202828

22192828

20192828

17152524

0 7 14 21 28 35 42 49 56 63 70 77 84

p <0.05

Time since treatment start (days)

Placebo5 mg12.5 mg25 mg

Fig. 2. Cumulative number of large volume paracentesis required per patientwith time. p <0.05 for all satavaptan groups compared with placebo.

25 mgPlacebo 5 mg 12.5 mg

24 h

r urin

ary

volu

me

in m

l

5000

4000

3000

2000

1000

0

End of Treatment

Baseline

Fig. 3. Twenty-four hour urinary volume at the beginning and at the end ofthe study.

JOURNAL OF HEPATOLOGY

were normonatraemic at entry, there was no significant increasein serum sodium overall (Table 3).

However, renal failure, when defined by a rise in serum creat-inine of >50% from baseline with a value P133 lmol/L, occurredat least once in 14.8% of satavaptan patients versus 2.7% for pla-cebo. These patients, especially in the 25 mg satavaptan group,generally had poorer baseline renal function as indicated byserum creatinine and estimated GFR compared with those notexperiencing such an event. The peak serum creatinine exceeded200 lmol/L in two patients on satavaptan 5 mg, two patients on12.5 mg and four patients on 25 mg.

There was a significant dose-dependent increase in urinaryvolume at the end of the study compared with baseline inpatients receiving satavaptan (Fig. 3), associated with asignificant reduction in urinary osmolality. The serum osmolality,however, did not rise significantly throughout the study (Table 3).

Hormonal levels

At baseline, plasma renin and aldosterone concentrations weresignificantly elevated, but these levels did not change signifi-cantly throughout the study. Plasma vasopressin levels decreased

Table 4. Hormonal levels in all study patients.

Placebo

n 36Plasma vasopressin (N = 0.9–12.3 pmol/L)

Baseline 2.7 (1.94, 3.35)Day 7 2.2 (1.51, 3.08)a

Month 3 1.9 (1.57, 2.33)a

Plasma renin (N = 2–29 mU/L)Baseline 229 (81, 553)Day 7 322 (101, 530)Month 3 177 (92, 501)

Plasma aldosterone (N = <0.440 nmol/L)Baseline 1.29 (0.75, 3.41)Day 7 1.91 (1.00, 4.72)Month 3 1.52 (1.00, 2.33)

Results are expressed as median and 25th and 75th percentiles.a p <0.05 compared with baseline.

Journal of Hepatology 201

in the placebo group, but significant increases were observed inall satavaptan groups, and appeared to be dose-related (Table 4).

Other adverse events

The overall incidence of adverse events did not differ markedlybetween placebo and the satavaptan groups (Table 5).

Orthostatic hypotension (�20 mm Hg), although more fre-quent (16.7%) (Table 5) in satavaptan patients, were mostlyasymptomatic. Two symptomatic patients (one each in the12.5 mg and 25 mg groups) recovered spontaneously withoutbeing reported as adverse events. One of these patients was tak-ing nadolol concomitantly. All patients with orthostatic hypoten-sion completed the 12-week study without further hypotensiveepisodes. Thirst was also a frequently reported adverse event insatavaptan patients (13.9%), but in none of the placebo patients.Other common adverse events included hyperkalaemia, thedevelopment of hepatic encephalopathy and muscle cramps(Table 5). The majority of patients, except 1 patient in the12.5 mg group, recovered from their hyperkalaemia followingwithdrawal of spironolactone, and withholding study medica-tions with or without the administration of sodium polystyrenesulfonate, or insulin or albumin. He had concomitant renal fail-

Satavaptan

5 mg 12.5 mg 25 mg

39 36 40

2.2 (1.84, 2.81) 2.2 (1.51, 3.13) 2.1 (1.51, 3.02)3.3 (2.38, 4.54)a 3.5 (2.38, 4.54)a 4.0 (3.02, 5.94)a

2.9 (2.27, 3.24)a 3.8 (2.05, 5.83)a 3.7 (2.70, 7.88)a

298 (62, 582) 169 (72, 456) 201 (88, 489)322 (77, 843) 230 (57, 575) 239 (68, 645)301 (112, 697) 178 (72, 734) 211 (96, 554)

1.02 (0.45, 4.22) 1.11 (0.58, 2.22) 1.80 (0.89, 4.69)1.94 (0.86, 4.94) 1.83 (1.14, 2.30) 2.05 (1.08, 6.05)1.80 (0.92, 3.33) 1.36 (0.97, 2.03) 1.97 (1.14, 4.33)

0 vol. 53 j 283–290 287

Table 5. Adverse events overview and incidence of selected types of adverse events reported 67 days after last dose of study medication.

Satavaptan

Placebo 5 mg 12.5 mg 25 mg

n 36 39 36 40Any adverse event 28 34 26 28Any serious adverse event 13 18 13 16Death 3 2 3 4Episodes of hepatic encephalopathy 3 7 4 4Bacterial peritonitis 3 0 1 2Esophageal variceal bleeding 1 3 2 0Muscle cramps 1 4 2 6Thirst 0 5 6 5Hyperkalaemia (serum [K] >5.5 mmol/L) 7 10 11 7Orthostatic hypotension (>20 mmHg) 1 9 7 3

Research Article

ure. Hyperkalaemia led to the discontinuation of study medica-tion in three patients (placebo: 2, 25 mg satavaptan: 1) Threecases of hepatic encephalopathy were associated with other com-plications of liver cirrhosis, and they all succumbed to theiradvanced liver disease. Otherwise, the majority of patients recov-ered with the administration of lactulose. Patients with musclecramps all recovered with symptomatic treatment. There wasno clear evidence of a dose-relationship in the frequency of anyevents.

There were no QTcF values of >500 ms observed either in theplacebo or the satavaptan groups throughout the study.

Several patients (placebo: 2; satavaptan 5 mg: 1; 12.5 mg: 2;25 mg: 2) had their study medication withheld for 1 day becauseof a >8 mmol/L rise in serum sodium, which quickly returned tonormal following temporary discontinuation of study medica-tion. All patients were able to resume study medication withoutfurther problems. No withdrawal from the study occurredbecause of persistent hypernatraemia.

Patient outcome

Overall, 57 patients discontinued the study early, for reasonslisted in Fig. 1. The median duration of study was 81, 68, 84,and 84 days for placebo, 5, 12.5, and 25 mg satavaptan groups,respectively.

Twelve patients died during the study period (placebo: 3,5 mg: 2, 12.5 mg: 3 and 25 mg: 4) from sepsis (n = 3), liver failure(n = 2), hepatorenal syndrome (n = 2), gastrointestinal hemor-rhage (n = 1), cerebral vascular accident (n = 1), pancreatic carci-noma (n = 1), biliary carcinoma (n = 1) and refractory anemia(n = 1). There was no increase in mortality in the satavaptan ver-sus placebo patients in this study.

Discussion

This study, in cirrhotic patients with recurrent ascites with orwithout hyponatraemia, found that satavaptan, a vasopressinV2 receptor antagonist, in conjunction with low-dose spironolac-tone, reduced the frequency of LVP over a 3-month period. Thetwo co-primary endpoints of time to first LVP and accumulatedvolume of ascites did not achieve statistical significance, butthe relative risks for first LVP were very consistent with thosefor all LVPs over 3 months, and all were within the range of0.60–0.69. The difference in statistical significance is explainedby the greater statistical power achieved by counting all LVPsinstead of only the first event. It might also be argued that the

288 Journal of Hepatology 201

reduction of all LVPs over 3 months is of greater clinical signifi-cance than the delay of only the first LVP.

Excess sodium and water retention is pivotal in the patho-physiology of ascites formation in cirrhosis [14]. The bindingof satavaptan, a selective vasopressin V2 receptor antagonist,to the V2 receptor at the renal collecting tubule, blocks the ser-ies of reactions that ultimately make the collecting tubuleimpermeable to water [15], thereby inducing a diuresis. Intui-tively, it seems unlikely that an aquaretic agent such as satavap-tan, could effect a reduction in ascites accumulation. However,the loss of a significant volume of water induced by satavaptancould lead to reduction in the intravascular volume. This couldbe compensated for by an increase in oral fluid intake and/oractivation of various physiological mechanisms to replenishthe effective intravascular volume. The fact that we did notobserve any significant change in the systemic haemodynamics,in serum osmolality, nor in plasma renin levels, suggests thatthe intravascular volume was maintained despite large diuresis,even with the 25 mg dose of satavaptan. We postulate that thiswas achieved partly through an increased reabsorption of asciticfluid, hence the reduction in ascites accumulation and adecrease in the frequency of paracentesis in the satavaptangroups, and partly through an increased movement of waterfrom the interstitial and/or intracellular compartment to thevascular compartment. It should be noted that despite higherdoses of satavaptan inducing a higher urine output, and pre-sumably a greater degree of intravascular volume contraction,there was not a dose response in the reduction in ascites. Thismay be related to the fact that reabsorption of ascitic fluid isrelatively fixed [16], maximally at approximately 400 mL/day.Since these patients had free access to fluid intake, the deficitin the intravascular volume induced by satavaptan presumablywas replenished by increasing oral fluid intake. The corollaryis that increasing aquaretic doses may not yield a greater reduc-tion of ascites despite a greater urinary water excretion. Fur-thermore, patients receiving aquaretic agents should not beplaced on fluid restriction.

It is possible that satavaptan was able to effect a reduction inascites by influencing other pathogenetic mechanisms of ascitesformation. Circulating vasopressin stimulates sodium reabsorp-tion in the collecting duct by increasing the activity of the amil-oride-sensitive epithelial sodium channel. 1-Desamino-8-D-arginine vasopressin, a potent vasopressin receptor agonist, hasbeen shown to markedly reduce urine flow rate and sodiumexcretion in healthy individuals [17], and the use of satavaptanin normal Wistar rats was associated with an increase in urinarysodium excretion [18]. The urinary sodium excretion in our studycohort remained relatively unchanged throughout the study.

0 vol. 53 j 283–290

JOURNAL OF HEPATOLOGY

However, the increased sodium excretion with satavaptan istransient [18], and therefore may not have been captured byintermittent urinary collections. It remains a possibility that thesodium excretory effects of satavaptan may have contributed tothe reduction in ascites in these patients.

Irrespective of which mechanisms are involved, the fact thatsatavaptan, when added to low-dose spironolactone over a 3-month period, was able to effect a reduction in ascites accumula-tion as measured by LVP frequency, suggests that it could be usedas adjunctive therapy for the management of ascites in cirrhosis,including patients with refractory ascites [4]. This has significantclinical implications, as cirrhotic patients with ascites of varyingseverity may potentially have their ascites better managed withthe addition of an aquaretic agent irrespective of serum sodiumconcentration. It should be noted that the fixed low-dose of spiro-nolactone, which was necessary to run a controlled clinical trialwithout introducing potentially confounding effects related tothe administration of various types and doses of diuretics, doesnot represent the most appropriate approach to the treatment ofcirrhotic ascites. Thus, future studies will need to test whetherthe addition of satavaptan to individually optimised regimens ofnatriuretic agents will yield the same results.

The safety profile of satavaptan in this study also suggests thatit may be an attractive adjunctive therapy. Although a significantnumber of patients died during the study period, most of thesepatients died from complications of cirrhosis, with mortality ratenot significantly different from what has been observed inpatients with advanced cirrhosis [1]. Furthermore, patients onvarious doses of satavaptan did not have significantly more majoradverse events compared with placebo. Thirst was clearly treat-ment related. However, as reported above, despite increasedurine volume with excess water, none of the patients takingsatavaptan experienced dehydration, or hypovolaemia. This isbecause all patients had free access to water, and were advisedto drink when thirsty. Therefore, only a few patients had theirsatavaptan dose temporarily withheld because of hypernatra-emia, unlike another aquaretic [9], thereby improving patientcompliance. However, one must be cognizant of the fact thatthere is a potential for dehydration with satavaptan, as theproportion of patients with increased serum creatinine washigher in the satavaptan patients than in the placebo group.Therefore, careful monitoring of renal function is required withaquaretic therapy.

There were several episodes of hyperkalaemia with the com-bination of spironolactone and satavaptan, although this was notsignificantly higher than in the placebo group. This may berelated to the fact that vasopressin regulates the Na–K–ATPaseand Na–K–Cl cotransporter activity [19], which is crucial forsodium reabsorption and potassium excretion in the distal neph-ron [20]. The use of a V2 receptor antagonist has been shown todecrease Na–K–ATPase expression and activity in experimentalanimals [19]. Since potassium excretion in the collecting ductusually occurs by active uptake across the basolateral membraneby Na–K–ATPase, and passive diffusion across the apical mem-brane into the lumen [21], decreasing Na–K–ATPase expressionby satavaptan would effectively reduce potassium secretion andhave the same clinical effects as aldosterone blockade. Therefore,patients on a combination of aldosterone antagonist and V2

receptor antagonist should have their electrolytes monitored reg-ularly. However, in the doses studied, the combination appears tobe safe. There was an increased incidence of muscle cramps or

Journal of Hepatology 201

hepatic encephalopathy, unrelated to hyperkalaemia, in the sat-avaptan patients. It would be prudent to monitor for theseadverse events and administer prompt corrective therapies.

In summary, the addition of fixed doses of a V2 receptor antago-nist, satavaptan, to low-dose spironolactone in cirrhotic patientswith ascites requiring frequent LVPs may improve the control ofascites, as the total number of LVPs over 3 months was significantlydecreased at all doses tested. The safety of satavaptan appearedacceptable at all of the doses tested and any increase in renal dys-function appeared to be mild. Careful and regular monitoring ofelectrolytes and renal function is advisable. In conclusion, the V2

receptor antagonist, satavaptan, added to low-dose spironolactone,has the potential to reduce recurrence of ascites at doses in the rangeof 5–25 mg in cirrhotic patients with recurrent or refractory ascites.

Competing interests

Florence Wong and Pere Gines are consultants for Sanofi-Aventis,Hugh Watson and Pascal Minini are employees of Sanofi-Aventis,Yves Horsmans, Paolo Angeli, Paul Gow and Mauro Bernardireport no conflict of interest.

Funding

The study was sponsored by Sanofi-Aventis.

Acknowledgment

CIBEREHD is supported by the Instituto de Salud Carlos III, Spain.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.jhep.2010.02.036.

References

[1] D’Amico G, Garcia-Tsao G, Pagliaro L. Natural history and prognosticindicators of survival in cirrhosis: a systematic review of 118 studies. JHepatol 2006;44:217–231.

[2] Planas R, Montoliu S, Balleste B, Rivera M, Miquel M, Masnou H, et al. Naturalhistory of patients hospitalized for management of cirrhotic ascites. ClinGastroenterol Hepatol 2006;4:1385–1394.

[3] Moreau R, Delegue P, Pessione F, Hillaire S, Durand F, Lebrec D, et al. Clinicalcharacteristics and outcome of patients with cirrhosis and refractory ascites.Liver Int 2004;24:457–464.

[4] Moore KP, Wong F, Gines P, Ochs A, Bernardi M, Salerno F, et al. Themanagement of ascites – report on the consensus conference of theInternational Ascites Club. Hepatology 2003;38:258–266.

[5] Gines P, Arroyo V, Quintero E, Planas R, Bory F, Cabrera J, et al. Comparison ofparacentesis and diuretics in the treatment of cirrhotics with tense ascites.Results of a randomized study. Gastroenterology 1987;93:234–241.

[6] Salerno F, Badalamenti S, Incerti P, Tempini S, Restelli B, Bruno S, et al.Repeated paracentesis and i.v. albumin infusion to treat ‘tense’ ascites incirrhotic patients. A safe alternative therapy. J Hepatol 1987;5:102–108.

[7] Wong F. The use of TIPS in chronic liver disease. Ann Hepatol 2006;5:5–15.[8] Gerbes AL, Gulberg V, Gines P, Decaux G, Gross P, Gandjini H, et al. Therapy

of hyponatremia in cirrhosis with a vasopressin receptor antagonist: arandomized double-blind multicenter trial. Gastroenterology 2003;124:933–939.

[9] Wong F, Blei AT, Blendis LM, Thuluvath PJ. A vasopressin receptor antagonist(VPA-985) improves serum sodium concentration in patients with hypona-

0 vol. 53 j 283–290 289

Research Article

tremia: a multicenter, randomized, placebo-controlled trial. Hepatology2003;37:182–191.

[10] Udelson JE, McGrew FA, Flores E, Ibrahim H, Katz S, Koshkarian G, et al.Multicenter, randomized, double-blind, placebo-controlled study on theeffect of oral tolvaptan on left ventricular dilation and function in patientswith heart failure and systolic dysfunction. J Am Coll Cardiol2007;49:2151–2159.

[11] Ginès P, Wong F, Watson H, Milutinovic S, Ruiz del Arbol L, Olteanu D, et al.Effects of satavaptan, a selective vasopressin V2 receptor antagonist, onascites and serum sodium concentration in patients with cirrhosis andhyponatremia. Hepatology 2008;48:204–213.

[12] Ferenci P, Lockwood A, Mullen K, Tarter R, Weissenborn K, Blei AT, et al.Hepatic encephalopathy-definition, nomenclature, diagnosis, and quantifi-cation: final report of the working party at the 11th World Congresses ofGastroenterology, Vienna, 1998. Hepatology 2002;35:716–721.

[13] Lawless JF, Nadeau C. Some simple robust method for the analysis ofrecurrent events. Technometrics 1995;37:158–168.

[14] Cárdenas A, Arroyo V. Mechanisms of water and sodium retention incirrhosis and the pathogenesis of ascites. Best Pract Res Clin EndocrinolMetab 2003;17:607–622.

290 Journal of Hepatology 201

[15] Bonilla-Felix M. The development of water transport in the collecting duct.Am J Physiol 2004;287:F1093–F1101.

[16] Shear L, Ching S, Gabuzda GJ. Compartmentalization of ascites and edema inpatients with hepatic cirrhosis. N Engl J Med 1970;282:1391–1396.

[17] Bankir L, Fernandes S, Bardoux P, Bouby N, Bichet DG. Vasopressin-V2

receptor stimulation reduces sodium excretion in healthy humans. J Am SocNephrol 2005;16:1920–1928.

[18] Perucca J, Bichet DG, Bardoux P, Bouby N, Bankir L. Sodium excretion inresponse to vasopressin and selective vasopressin receptor antagonists. J AmSoc Nephrol 2008;19:1721–1731.

[19] Bertuccio CA, Ibarra FR, Toledo JE, Arrizurieta EE, Martin RS. Endogenousvasopressin regulates Na–K–ATPase and Na(+)–K(+)–Cl(�) cotransport-er rbsc-1 in rat outer medulla. Am J Physiol Renal Physiol 2002;282:F265–F270.

[20] Rozansky DJ. The role of aldosterone in renal sodium transport. SeminNephrol 2006;26:173–181.

[21] Giebisch G. Renal potassium transport: mechanisms and regulations. Am JPhysiol 1998;274:F817–F833.

0 vol. 53 j 283–290