Original Article

Tolvaptan for improvement of hepatic edema:A phase 3, multicenter, randomized, double-blind,placebo-controlled trial

Isao Sakaida,1 Seiji Kawazoe,2 Kozo Kajimura,3 Takafumi Saito,4 Chiaki Okuse,5

Koichi Takaguchi,6 Mitsuru Okada7 and Kiwamu Okita8 for theASCITES-DOUBLEBLIND Study Group*1Yamaguchi University Graduate School of Medicine, Ube, Yamaguchi, 2Saga Prefectural Hospital Koseikan,Saga, 3Kishiwada City Hospital, Kishiwada, Osaka, 4Yamagata University Hospital, Yamagata, 5St MariannaUniversity School of Medicine Hospital, Kawasaki, Kanagawa, 6Kagawa Prefectural Central Hospital, Takamatsu,Kagawa, 7Otsuka Pharmaceutical, Osaka, and 8Social Insurance Alliance Shimonoseki Kosei Hospital,Shimonoseki, Yamaguchi, Japan

Aim: Hepatic edema is manifested by ascites, lower limbedema and intolerable symptoms. Some patients insuffi-ciently respond to the conventional diuretic therapy. There-fore, a novel therapeutic option is required. We conducted aphase 3 study to confirm therapeutic effect of tolvaptan onhepatic edema associated with liver cirrhosis.

Methods: In our multicenter, randomized, double-blind,placebo-controlled trial, liver cirrhosis patients who showedinsufficient response to conventional diuretics were randomlyassigned to 7-day administration of either tolvaptan at7.5 mg/day or placebo as an add-on therapy to conventionaldiuretics. The primary outcome was change in bodyweightfrom baseline.

Results: Of 164 eligible patients, 84 were assigned to tolvap-tan and 80 to placebo. Change in bodyweight from baselineon the final dosing day was -0.44 kg (standard deviation [SD],1.93) in the placebo group and -1.95 kg (SD, 1.77) in thetolvaptan group (P < 0.0001). Improvement rates for lower

limb edema and ascites-related clinical symptoms werehigher with tolvaptan than with placebo. Even in patients withlow serum albumin (<2.5 g/dL), decrease in bodyweight wasgreater with tolvaptan than with placebo (P = 0.0163). In addi-tion, tolvaptan significantly increased serum sodium concen-tration from baseline.

Conclusion: Add-on therapy with tolvaptan was effective forthe treatment of hepatic edema and ascites-related clinicalsymptoms. Furthermore, tolvaptan is expected to improvelow serum sodium concentration and to exert its effectregardless of serum albumin level. Add-on therapy withtolvaptan is therefore considered to be a novel therapeuticoption for hepatic edema.

Key words: add-on therapy, ascites-related clinicalsymptoms, conventional diuretics, hepatic edema, livercirrhosis, tolvaptan

Correspondence: Dr Isao Sakaida, Department of Gastroenterology and Hepatology, Yamaguchi University Graduate School of Medicine, 1-1-1Minami Kogushi, Ube, Yamaguchi 755-8505, Japan. Email: sakaida@yamaguchi-u.ac.jp*ASCITES-DOUBLEBLIND study group (ASCITES: Aquaretic Samsca, tolvaptan, in liver Cirrhosis patient Indicating a novel Therapeutic optionwith superior Effects): Akihide Masumoto (Iizuka Hospital); Akira Kaneko (NTT West Osaka Hospital); Atsuo Inoue (Osaka GeneralMedical Center); Chitomi Hasebe (Keiyukai Yoshida Hospital); Eiichi Takesaki (National Hospital Organization HigashihiroshimaMedical Center); Eijiro Hayashi (Kinki Central Hospital of the Mutual Aid Association of Public School Teachers); Hajime Ohta(National Hospital Organization Kanazawa Medical Center); Harumasa Yoshihara (Osaka Rosai Hospital); Hideo Morimoto (NationalHospital Organization Kanazawa Medical Center); Hideo Nishimura (National Hospital Organization Asahikawa Medical Center);Hideo Tsuruta (National Hospital Organization Ureshino Medical Center); Hiroaki Iwase (National Hospital Organization NagoyaMedical Center); Hirohisa Tanimura (Gokeikai Osaka Kaisei Hospital); Hiromitsu Kumada (Toranomon Hospital); Hiroshi Hasegawa(National Hospital Organization Mie Central Medical Center); Hiroshi Kohno (National Hospital Organization Kure Medical Centerand Chugoku Cancer Center); Hisashi Hidaka (Kitasato University East Hospital); Hisataka Moriwaki (Gifu University Hospital); IchizoKobayashi (Higashiosaka City General Hospital); Kazuya Shimoda (University of Miyazaki Hospital); Keiji Tsuji (Hiroshima City AsaHospital); Keisuke Ario (National Hospital Organization Ureshino Medical Center); Kenichi Yamamoto (Saga Memorial Hospital);Kinichi Yokota (Yoshida Hospital); Manabu Yamauchi (Nagoya Memorial Hospital); Masahiko Takahashi (National HospitalOrganization Tokyo Medical Center); Masami Inada (Toyonaka Municipal Hospital); Masanori Kawaguchi (Saiseikai Wakayama

bs_bs_banner

Hepatology Research 2014; 44: 73–82 doi: 10.1111/hepr.12098

© 2013 The Japan Society of Hepatology 73

INTRODUCTION

IN MOST PATIENTS with liver cirrhosis, hepaticedema is manifested by ascites due to portal hyper-

tension and impaired albumin synthesis.1,2 Persistenthepatic edema may lead to the development of varioussubjective and objective symptoms, resulting in deterio-ration of quality of life (QOL).3

The conventional pharmacotherapy for hepaticedema is combination use of an aldosterone antagonist,spironolactone, and a loop diuretic, such as furo-semide.4 However, there are some patients who do notshow sufficient response to this combination therapy.5

As serum albumin and sodium levels are usually lowin such patients, sufficient diuretic effect may not beexpected only by treatment with furosemide. Further-more, the dose of furosemide cannot be easily increaseddue to the risk of the adverse drug reactions, such ashyponatremia and renal impairment. As a result, it isdifficult to manage hepatic edema using conventionaldiuretics alone.6

Tolvaptan is a non-peptide orally available argininevasopressin V2 receptor antagonist.7,8 The drug has beenapproved for the treatment of hyponatremia in the USA,for the treatment of hyponatremia secondary to syn-drome of inappropriate antidiuretic hormone in the EU,and for the treatment of volume overload in patientswith heart failure in Japan.9 In addition to these indica-tions, add-on therapy of tolvaptan to conventionaldiuretics is expected to be useful for the treatment ofhepatic edema because tolvaptan induces diuresiswithout sodium excretion.10

As a series of clinical trials to obtain approval for theadditional indication, a phase 2 dose finding study oftolvaptan in liver cirrhosis patients with ascites wereconducted at a low dose of 7.5 mg/day, a middle doseof 15 mg/day (the recommended dose for volume over-load in patients with heart failure in Japan) and a highdose of 30 mg/day.11 Based on the results from thepreceding study, a dose of 7.5 mg/day was selected asthe optimal dose for liver cirrhosis patients, and a phase3 study was conducted to confirm the therapeutic effectof tolvaptan as add-on therapy to conventional diuret-ics at that dose on hepatic edema associated with livercirrhosis.

METHODS

THIS MULTICENTER, RANDOMIZED, double-blind,placebo-controlled, phase 3 study was conducted at

80 trial sites in Japan between February 2010 andAugust 2011. This trial consisted of a 3-day pretreatmentobservation period, a 7-day treatment period and a14-day post-treatment observation period.

This trial enrolled liver cirrhosis patients with asciteswho had been receiving combination therapies with aloop diuretic and an anti-aldosterone agent from at least7 days prior to acquisition of informed consent. Stan-dard daily dose of concomitant diuretics in Japan was asfollows: a loop diuretic at a daily dose equivalent tofurosemide 40 mg/day or higher and spironolactoneat 25 mg/day or higher, or a loop diuretic at a dailydose equivalent to furosemide 20 mg/day or higher andspironolactone at 50 mg/day or higher.12 Patients, aged

Hospital); Masatoshi Kudo (Kinki University Hospital); Masatoshi Tanaka (Kurume University Medical Hospital); Michiharu Mihara(Kumamoto City Hospital); Michihiro Suzuki (Kawasaki Municipal Tama Hospital); Michio Kato (National Hospital OrganizationMinamiwakayama Medical Center); Mikiya Kitamoto (Hiroshima Prefectural Hospital); Mitsuhiro Murata (Miyazaki Medical CenterHospital); Nobuhiko Higashi (Saiseikai Fukuoka General Hospital); Nobuyuki Chida (National Hospital Organization Sendai MedicalCenter); Osami Inoue (Nagasaki Rousai Hospital); Seigo Takano (Sagamihara Kyodo Hospital); Seiichi Nakayama (National HospitalOrganization Tochigi National Hospital); Seisyu Hayashi (Tokyo Metropolitan Komagome Hospital); Shigetoshi Fujiyama (NTT WestKyusyu Hospital); Shinichi Fujioka (Okayama Saiseikai General Hospital); Shinichi Shimizu (Okayama Rosai Hospital); ShinjiKatsushima (National Hospital Organization Kyoto Medical Center); Shinji Kitamura (Sakai City Hospital); Shinsuke Kira (HiroshimaRed Cross Hospital and Atomic Bomb Survivors Hospital); Shinya Nagaoka (National Hospital Organization Nagasaki Medical Center);Shuji Terai (Yamaguchi University Hospital); Sumio Watanabe (Juntendo University Hospital); Syoichi Takahashi (HiroshimaUniversity Hospital); Syuji Inoue (National Hospital Organization Kochi National Hospital); Takashi Yamaguchi (National HospitalOrganization Mito Medical Center); Takeaki Satoh (National Hospital Organization Kokura Medical Center); Takeshi Azuma (KobeUniversity Hospital); Takumi Igura (Ikeda Municipal Hospital); Tatsuji Komatsu (National Hospital Organization Yokohama MedicalCenter); Toru Tamura (Mazda Hospital); Toshifumi Ito (Osaka Koseinenkin Hospital); Toshiharu Ueki (Fukuoka University ChikushiHospital); Toyokichi Muro (National Hospital Organization Oita Medical Center); Tsubasa Morita (KKR Takamatsu Hospital); YoichiNishigaki (Gifu Municipal Hospital); Yoshifumi Ohnishi (National Hospital Organization Shizuoka Medical Center); Yoshiyasu Karino(Hokkaido P.W.F.A.C Sapporo Kosei General Hospital); Yoshiyuki Narahara (Nippon Medical School Hospital); Yukio Osaki (OsakaRed Cross Hospital); Yutaka Mano (National Hospital Organization Sendai Medical Center).Received 5 February 2013; revision 5 February 2013; accepted 17 February 2013.

74 I. Sakaida et al. Hepatology Research 2014; 44: 73–82

© 2013 The Japan Society of Hepatology

20–80 years, were required to be hospitalized or to beavailable for hospitalization during the trial period.

Main exclusion criteria were hepatic encephalopathy(coma scale, 3II),13 vascular invasive hepatocellular car-cinoma, requiring new treatment for varices esophagealor gastric varices, hemorrhoidal hemorrhage secondaryto rectal varices, and receiving blood products includingalbumin.

This trial was conducted in accordance with ethicalprinciples originating from the Declaration of Helsinkiand in compliance with good clinical practice guide-lines. The protocol was approved by the institutionalreview board at each trial site. All patients providedwritten informed consent. This trial is registered withClinicalTrials.gov, number NCT01050530.

Patients with ascites volume of 1000 mL or more ascalculated by computed tomography (CT)14 and inwhom bodyweight before breakfast on the second andthird days of the pretreatment observation period wasstable (11.0 kg) were eligible for advancement to thetreatment period by investigator’s judgment. If thepatient was confirmed to meet these criteria, the inves-tigator sent the Treatment Assignment Report Form tothe registration center. The registration center recon-firmed the eligibility of the patient.

Patients were randomized to the tolvaptan group orthe placebo group (1:1) for 7-day administration of7.5 mg tolvaptan or placebo once daily after breakfast asadd-on therapy to conventional diuretics. The dose ofconventional diuretics was to remain fixed from 7 daysprior to the start of trial drug administration. The regis-tration center assigned a trial drug code to the patient bydynamic allocation using the ascites volume as a ran-domization factor.

A randomization code was pre-assigned to each trialdrug, and each patient was assigned a treatment codecorresponding to each trial drug code by the trial drugallocation manager from the contract research organiza-tion for the registration center. All patients, trial person-nel, investigators and the sponsor were masked totreatment allocation throughout the trial. The trial drugallocation manager sealed the assignment list immedi-ately after assignment, and kept it sealed until the des-ignated time for unmasking.

For all variables, data obtained immediately beforethe start of trial drug administration were used as base-line data. The day that each patient completed or dis-continued the administration of trial drugs was definedas the final dosing day.

Placebo was used as a reference drug because tolvap-tan has a novel mechanism of action different from that

of conventional diuretics and no positive comparator isavailable. The treatment period was set at 7 days becausedifference in change in bodyweight between the tolvap-tan and placebo groups was evaluable by 7 days afterstart of the trial drug administration in the previoustrials, and also due to ethical consideration for patientsassigned to the placebo group.11

The primary endpoint was change in bodyweightfrom baseline considered to reflect improvement ofhepatic edema on the final dosing day.15 The secondaryendpoints included changes in abdominal circumfer-ence and ascites volume calculated by CT, and improve-ment rates in lower limb edema and ascites-relatedclinical symptoms (bloated feeling, loss appetite,malaise, sensation of pressure in the decubitus position,and breathing difficulty in patients with the symptom atbaseline) compared with baseline on the final dosingday, respectively.

Investigators assessed the severity of lower limbedema as “none”, “mild”, “moderate” and “severe”.Change in lower limb edema was assessed as “markedlyimproved” (complete resolution or improvement by 32grades), “improved” (improvement by 31 grade, exceptcomplete resolution), “unchanged” and “worsened”(worsened by 31 grade). Patients assessed as markedlyimproved or improved were counted as effectivecases. Investigators also asked patients if they hadascites-related clinical symptoms at baseline and ifthe symptoms changed by day 7. Changes in ascites-related clinical symptoms were assessed as “resolved”,“improved”, “unchanged” or “worsened”. Patientsassessed as resolved or improved were counted as effec-tive case. Both improvement rates were calculated bydividing the number of effective case by the number ofpatients with symptoms at baseline.

Day 1 was defined as the period from the first admin-istration until the second administration of trial drugs.Days 2–7 were similarly defined. Bodyweight was mea-sured before breakfast following urination at baselineand on days 1–7, and abdominal circumference wasmeasured before breakfast at baseline, on any dayduring days 2–4 and on day 7. Ascites volume wascalculated at baseline and on day 7. Lower limb edemawas evaluated before breakfast at baseline, on any dayduring days 2–4 and on day 7. Ascites-related clinicalsymptoms were assessed at baseline and on day 7.

Urine samples to determine cumulative daily urinevolume were collected at baseline and on days 1 and 7,and blood samples to determine serum sodium concen-tration were collected at baseline, 4–8 h and 24 h onday 1, on any day during days 2–4 and on day 7.

Hepatology Research 2014; 44: 73–82 Tolvaptan for improvement of hepatic edema 75

© 2013 The Japan Society of Hepatology

Safety assessments, including adverse events, clinicallaboratory tests, vital signs and 12-lead electrocardio-gram, were conducted during the trial period.

The required sample size was calculated assuming sta-tistically significant difference for change in bodyweightfrom baseline on the final dosing day using one-sidedpaired t-test at a significance level of 0.025 and 90%power. In the previous trial, changes in bodyweightwere -0.36 kg (standard deviation [SD], 2.06) in theplacebo group, -2.31 kg (SD, 2.35) in the 7.5 mg group,-1.88 kg (SD, 2.45) in the 15 mg group and -1.67 kg(SD, 1.46) in the 30 mg group.11 In this trial, it wasassumed that difference in change in bodyweightbetween two groups would be -1.31 kg (SD, 2.45),based on the minimum difference and the maximumSD among all treatment groups in the previous trial.Therefore, the required sample size was calculated to be75 patients per group, and we determined to enroll aminimum of 80 patients per group, considering thepossibility of a number of withdrawals.

Analyses were performed on the full analysis set(FAS). The FAS included all randomized patients whoreceived the trial drugs at least once. Missing data on thefinal dosing day were imputed by the last data obtainedafter the start of treatment (the last observation carriedforward method). If ascites volume calculated on day 7

was unavailable, its data was imputed by data obtainedbefore treatment. Patients, who received the trial drugsat least once, were included in safety assessments.

Continuous data were analyzed using Student’s t-testor paired t-test. Categorical data were analyzed usingFischer’s exact test or Kruskal–Wallis test. Statistical sig-nificance was defined as P < 0.05 (two-sided). Also, 95%confidence intervals (CI) were calculated. All statisticalanalyses using SAS software ver. 9.2 (SAS Institute, Cary,NC, USA), were performed by EPS (Tokyo, Japan).

RESULTS

OF THE ENROLLED 164 patients, 84 were assignedto the tolvaptan group and 80 to the placebo

group (Fig. 1). However, two were withdrawn due tophysicians’ decisions and withdrew consent in thetolvaptan group before the start of treatment. Thereafter,10 patients were withdrawn from the placebo groupand eight from the tolvaptan group. The reasons forearly discontinuation were one withdrew consent, threeadverse events (bleeding from esophageal varices withhepatic encephalopathy, old myocardial infarctionwith hepatic encephalopathy and liver disease-relatededema), and six physicians’ decisions in the placebogroup, and one protocol violation, six adverse events

Patients excluded not meet criteria (n = 32)

Patients included in pre-treatmentobservation period (n = 187)

Patients assessed for eligibility(n = 219)

Patients excluded (n = 23) not meet criteria (n = 19) adverse evens (n = 2) physician’s decision (n = 1) withdrew consent (n = 1)

Patients randomized(n = 164)

Patients excluded (n = 2) physician’s decision (n = 1) withdrew consent (n = 1)

Patients randomizedto the placebo group (n = 80)

Patients randomizedto the tolvaptan group (n = 82)

Figure 1 Patient flow chart.

76 I. Sakaida et al. Hepatology Research 2014; 44: 73–82

© 2013 The Japan Society of Hepatology

(hepatic encephalopathy, umbilical hernia, dehydra-tion, chronic renal failure, eruption and hyponatremia)and one physician’s decision in the tolvaptan group.There were no significant differences in patient demo-graphics and clinical characteristics between the twogroups (Table 1).

Change in bodyweight from baseline on the finaldosing day was -0.44 kg (SD, 1.93) in the placebogroup and -1.95 kg (SD, 1.77) in the tolvaptan group.Difference between the two groups (-1.51 kg) was sta-tistically significant (P < 0.0001). Tolvaptan showed sig-nificant decrease in bodyweight compared with baselineat each time point, and placebo showed from on day 5onward (Fig. 2).

Change in ascites volumes were -191.8 mL (SD,690.8) in the placebo group and -492.4 mL (SD, 760.3)

in the tolvaptan group. Difference between the twogroups (-300.6 mL) was statistically significant (P =0.0093, Fig. 3a). Change in abdominal circumferencewas -1.11 cm (SD, 3.67) in the placebo group and-3.38 cm (SD, 3.56) in the tolvaptan group. Differencebetween the two groups (-2.27 cm) was statisticallysignificant (P = 0.0001, Fig. 3b). Lower limb edemaimprovement rate was 28.3% in the placebo group and54.8% in the tolvaptan group. Difference between thetwo groups (26.5%) was statistically significant (P =0.0168, Fig. 3c). In evaluating lower limb edema, threepatients in the tolvaptan group who showed symptomsduring days 2–4 but not at baseline were included inanalysis, resulting in unchanged or worsened.

Urine volume in the tolvaptan group significantlyincreased from baseline on day 1 (1006 mL [SD, 763],

Table 1 Demographic and clinical characteristics at baseline

Characteristics Placebo group (n = 80) Tolvaptan group (n = 82) P-value

Age (years) 69 (46–78) 68 (31–80) 0.91†Sex (male) 49 (61.3%) 52 (63.4%) 0.87‡Bodyweight (kg) 59.2 (13.2) 59.4 (12.7) 0.92†Abdominal circumference (cm) 89.7 (10.8) 89.6 (11.1) 0.93†Ascites volume calculated by computed tomography (mL) 2841 (1424) 2846 (1398) 0.98†Severity of lower limb edema 0.25§

None 34 (42.5%) 43 (52.4%)Mild 23 (28.8%) 19 (23.2%)Moderate 13 (16.3%) 12 (14.6%)Severe 10 (12.5%) 8 (9.8%)

Presence of pleural effusion 20 (25.0%) 25 (30.5%) 0.49‡Cause of liver cirrhosis

Hepatitis B 7 (8.8%) 5 (6.1%)Hepatitis C 42 (52.5%) 48 (58.5%)Alcoholic cirrhosis 20 (25.0%) 27 (32.9%)Unknown and other 16 (20.0%) 10 (12.2%)

Child–Pugh classification 0.07§A 0 (0%) 0 (0%)B 54 (67.5%) 44 (53.7%)C 26 (32.5%) 38 (46.3%)

ComplicationsLiver cancer 16 (20.0%) 23 (28.0%) 0.27‡Varicose veins 65 (81.3%) 68 (82.9%) 0.84‡

Serum albumin concentration (g/dL) 2.88 (0.50) 2.81 (0.46) 0.39†Serum creatinine concentration (mg/dL) 1.05 (0.44) 0.99 (0.35) 0.27†Performed sodium restriction 29 (36.3%) 32 (39.0%) 0.75‡Concomitant diuretics used

Spironolactone (mg/day) 54.1 (24.3) 55.2 (26.3) 0.78†Furosemide (mg/day) 61.5 (28.8) 64.5 (40.4) 0.59†

Data are expressed as median (range), number (%) or mean (standard deviation).†Student’s t-test.‡Fischer’s exact test.§Kruskal–Wallis test.

Hepatology Research 2014; 44: 73–82 Tolvaptan for improvement of hepatic edema 77

© 2013 The Japan Society of Hepatology

P < 0.0001) and on day 7 (633 mL [SD, 644],P < 0.0001), but urine volume in the placebo groupshowed no significant change (Fig. 4).

Serum sodium concentration was significantlyincreased from baseline on the final dosing day inthe tolvaptan group (135.3 mEq/L [SD, 4.5] to

136.5 mEq/L [SD, 4.4], P = 0.0002), while the serumsodium concentration was significantly decreased frombaseline in the placebo group (135.7 mEq/L [SD, 4.1] to135.0 mEq/L [SD, 4.3], P = 0.0060) (Table 2).

Tolvaptan significantly improved ascites-related clini-cal symptoms, bloated feeling (P = 0.0090), malaise

–5

–4

–3

–2

–1

0

1

2

BL 1 2 3 4 5 6 7 FDD

Time point (day)Number of patients

PlaceboTolvaptan

8082

7882

7981

7481

7378

7176

7075

7175

8082

Change in b

ody w

eig

ht (k

g)

P < 0. 0001

***

******

******

****** ***

** ***

*

Figure 2 Time-course change in bodyweight from baseline.Data are expressed as mean 1 standard deviation. Compari-sons between each group and baseline were performed byusing paired t-test (*P < 0.05, **P < 0.01 and ***P < 0.001).Comparison between the tolvaptan group and the placebogroup was performed by using Student’s t-test. BL, baseline;FDD, final dosing day. , placebo; , tolvaptan.

0

10

20

30

40

50

60

Placebo(n = 46)

Tolvaptan(n = 42)

–1400

–1200

–1000

–800

–600

–400

–200

0

Placebo(n = 80)

Tolvaptan(n = 82)

(a) (b)

Ch

an

ge

in

ascite

s v

olu

me

(m

L)

P = 0.0093 P = 0.0001

P = 0.0168(c)

Ch

an

ge

in

ab

do

min

al cir

cu

mfe

ren

ce

(cm

)

Imp

rove

me

nt

rate

of lo

we

r lim

b e

de

ma

(%

)

–8

–7

–6

–5

–4

–3

–2

–1

0

Placebo(n = 79)

Tolvaptan(n = 81)

Figure 3 Effects of tolvaptan onhepatic edema. (a) Change in ascitesvolume (computed tomography) frombaseline on the final dosing day. (b)Change in abdominal circumferencefrom baseline on the final dosing day.Data are expressed as mean 1 standarddeviation. (c) Improvement rate inlower limb edema compared with base-line on the final dosing day in patientswith its symptom at baseline. Improve-ment rate = (number of effective case) /(number of patients with symptomat baseline) ¥ 100 (%). Comparisonsbetween the tolvaptan group and theplacebo group were performed by usingStudent’s t-test or Fischer’s exact test.

Placebo Tolvaptan

Change in u

rine v

olu

me (

mL)

P = 0.3352

P = 0.4595

P < 0.0001

P < 0.0001

0

500

1000

1500

2000

Day 1(n = 80)

Day 7(n = 71)

Day 1(n = 82)

Day 7(n = 75)

Figure 4 Change in urine volume from baseline on days 1 and7. Data are expressed as mean 1 standard deviation. Compari-sons between each group (on days 1 and 7) and baseline wereperformed by using paired t-test.

78 I. Sakaida et al. Hepatology Research 2014; 44: 73–82

© 2013 The Japan Society of Hepatology

(P = 0.0074), sensation of pressure in the decubitusposition (P = 0.0017) and breathing difficulty (P =0.0233) compared with placebo (Table 3).

Forty-eight patients showed adverse events in theplacebo group (60.0%) and 60 in the tolvaptan group(73.2%). Adverse events observed in the tolvaptangroup during the trial period at 3% or greater frequencywere thirst, constipation, renal impairment, diarrhea,pollakiuria, pyrexia, hepatic encephalopathy, vomiting,insomnia, stomatitis and pruritus (Table 4). Tenpatients showed serious adverse events in the placebogroup (12.5%) and seven in the tolvaptan group(8.5%). Serious adverse events observed in the tolvaptangroup were disseminated intravascular coagulation,liver cirrhosis, portal vein thrombosis, omphalitis, bileduct cancer, liver malignant neoplasm, hepatic encepha-lopathy, renal impairment, respiratory failure and

hemoperitoneum (Table 5). No marked abnormalitieswere clinically observed in clinical laboratory tests, vitalsigns and 12-lead electrocardiograms.

DISCUSSION

IN LIVER CIRRHOSIS patients with ascites, reductionin bodyweight from baseline was significantly greater

in the tolvaptan group than in the placebo group.Reductions in abdominal circumference (parameter forassessing ascites retention) and ascites volume weregreater in the tolvaptan group than in the placebogroup. In addition, improvement rates of ascites-relatedclinical symptoms and lower limb edema (a symptomassociated with hepatic edema) were higher in thetolvaptan group than in the placebo group.

Table 2 Changes in serum sodium concentration

Group Time point n Concentration Changes from baseline

mEq/L mEq/L 95% CI P-value†

Placebo Baseline 80 135.7 (4.1)4–8 h on day 1 79 134.7 (4.1) -1.0 (1.4) -1.3 to -0.6 <0.000124 h on day 1 80 135.1 (4.5) -0.6 (1.5) -0.9 to -0.2 0.0012Days 2–4 78 135.3 (4.0) -0.6 (1.8) -1.0 to -0.1 0.0097Day 7 71 135.3 (4.1) -0.7 (2.2) -1.2 to -0.2 0.0082Final dosing day 80 135.0 (4.3) -0.7 (2.2) -1.2 to -0.2 0.0060

Tolvaptan Baseline 82 135.3 (4.5)4–8 h on day 1 82 135.4 (4.8) 0.0 (1.9) -0.4 to 0.5 0.812124 h on day 1 82 137.6 (4.3) 2.3 (1.8) 1.9 to 2.7 <0.0001Days 2–4 80 137.9 (4.3) 2.5 (2.3) 1.9 to 3.0 <0.0001Day 7 75 136.9 (3.9) 1.2 (2.8) 0.5 to 1.8 0.0006Final dosing day 82 136.5 (4.4) 1.2 (2.7) 0.6 to 1.8 0.0002

Data are expressed as mean (standard deviation).†Paired t-test (relative to baseline).CI, confidence interval.

Table 3 Improvement rates of ascites-related clinical symptoms

Symptoms Placebo Tolvaptan P-value†% %

Bloated feeling 37.3 (22/59) 62.5 (35/56) 0.0090Loss appetite 16.7 (6/36) 38.9 (14/36) 0.0640Malaise 30.6 (15/49) 58.7 (27/46) 0.0074Sensation of pressure in decubitus position 26.7 (8/30) 65.8 (25/38) 0.0017Breathing difficulty 23.8 (5/21) 62.5 (10/16) 0.0233

Data are improvement rates. Improvement rate was calculated by dividing the number of effective case by the number of patients withsymptom at baseline.†Fischer’s exact test.

Hepatology Research 2014; 44: 73–82 Tolvaptan for improvement of hepatic edema 79

© 2013 The Japan Society of Hepatology

Decrease in bodyweight is considered to reflectimprovement of ascites and lower limb edema.15 Inthis trial, ascites was considered to be improved bytolvaptan, because abdominal circumference and ascitesvolume both decreased, and ascites-related clinicalsymptoms improved. As worsening of ascites, lowerlimb edema and ascites-related clinical symptoms

results in deterioration of QOL in patients with hepaticedema,3 tolvaptan may also improve QOL.

Liver cirrhosis patients are reported to have low serumsodium concentration.16 In this trial, mean serumsodium concentration in patients was also around thelower limit of the normal range at the start of treatment.Kim et al. reported that, in liver cirrhosis patients regis-tered on the waiting list for liver transplantation, thehazard ratio of death increased 1.05-fold (95% CI, 1.03to 1.08) per 1 mEq/L decrease in serum sodium concen-tration when serum sodium concentration was between125 and 140 mEq/L.17 In this trial, mean serum sodiumconcentration increased in the tolvaptan group anddecreased in the placebo group. When sufficient diuresisis not achieved by conventional diuretics, increasing thedose of the conventional diuretics may be considered,but there is a risk of further decrease in serum sodiumconcentration. However, add-on therapy of tolvaptanto conventional diuretics is expected to adequatelyimprove serum sodium concentration in liver cirrhosispatients. Add-on therapy of tolvaptan could thus bebeneficial to liver cirrhosis patients not only by overallimprovement of hepatic edema but also by improve-ment in serum sodium concentration.

Albumin preparation use is recommended in patientswith a serum albumin levels of less than 2.5 g/dL.18,19 Asfurosemide exerts its diuretic effect through its highbinding to albumin,20 furosemide cannot fully exert itsdiuretic effect in liver cirrhosis patients with low serumalbumin.21 There have been some reports that patientswith a serum albumin level of less than 2.5 g/dL do notsufficiently respond to furosemide.22–24 In the presenttrial, most of the patients enrolled had a serum albuminlevel below the lower limit of the normal range. Addi-tional stratified analysis of change in bodyweight wasconducted according to serum albumin level of 2.5 g/dLby Student’s t-test. In patients with serum albumin levelof less than 2.5 g/dL, tolvaptan changed bodyweight by-2.19 kg (SD, 2.03) relative to baseline compared with-0.45 kg (SD, 2.20) in the placebo group (P = 0.0163,Fig. 5), indicating that tolvaptan could improve hepaticedema independent of serum albumin level.

Numerically, incidence rates of adverse events in bothgroups were nearly equal. Adverse events predicted fromtolvaptan’s pharmacological action were observed inthis trial.

This trial confirmed the effects of 7-day administra-tion of tolvaptan on hepatic edema compared withplacebo. However, the appropriate dose for a long-termuse should be investigated and the effects of tolvaptanon QOL should be confirmed in the future trials.

Table 4 Incidence of adverse events

Adverse events Placebo(n = 80)

Tolvaptan(n = 82)

Total 48 (60.0) 60 (73.2)Observed in 33% of patients

Thirst 1 (1.3) 11 (13.4)Constipation 6 (7.5) 6 (7.3)Renal impairment 6 (7.5) 6 (7.3)Diarrhea 4 (5.0) 6 (7.3)Pollakiuria 0 (0.0) 6 (7.3)Pyrexia 6 (7.5) 4 (4.9)Hepatic encephalopathy 4 (5.0) 4 (4.9)Vomiting 2 (2.5) 4 (4.9)Insomnia 0 (0.0) 4 (4.9)Stomatitis 1 (1.3) 3 (3.7)Pruritus 0 (0.0) 3 (3.7)

Data are number of patients (%).

Table 5 Incidence of serious adverse events

Serious adverse events Placebo(n = 80)

Tolvaptan(n = 82)

Total 10 (12.5) 7 (8.5)Disseminated intravascular

coagulation0 (0.0) 1 (1.2)

Iron-deficiency anemia 1 (1.3) 0 (0.0)Esophageal varices hemorrhage 1 (1.3) 0 (0.0)Liver cirrhosis 0 (0.0) 1 (1.2)Portal vein thrombosis 0 (0.0) 1 (1.2)Omphalitis 0 (0.0) 1 (1.2)Pneumonia 1 (1.3) 0 (0.0)Femoral neck fracture 2 (2.5) 0 (0.0)Bile duct cancer 0 (0.0) 1 (1.2)Liver malignant neoplasm 0 (0.0) 1 (1.2)Hepatic coma 1 (1.3) 0 (0.0)Hepatic encephalopathy 3 (3.8) 1 (1.2)Renal failure 1 (1.3) 0 (0.0)Renal impairment 1 (1.3) 1 (1.2)Respiratory failure 0 (0.0) 1 (1.2)Hemoperitoneum 0 (0.0) 1 (1.2)

Data are number of patients (%).

80 I. Sakaida et al. Hepatology Research 2014; 44: 73–82

© 2013 The Japan Society of Hepatology

In conclusion, add-on therapy of tolvaptan was effec-tive in the treatment of hepatic edema and ascites-related clinical symptoms in liver cirrhosis patients whodid not respond sufficiently to conventional diuretics.Furthermore, tolvaptan is expected to improve lowserum sodium concentration, and the efficacy of tolvap-tan was not affected by serum albumin level. Add-ontherapy of tolvaptan would therefore be useful as anovel clinical therapeutic option for the treatment ofhepatic edema.

ACKNOWLEDGMENTS

THIS TRIAL WAS funded by Otsuka Pharmaceutical.We thank the ASCITES-DOUBLEBLIND Study

Group members for their commitment to this trial, andpatients who participated in this trial.

REFERENCES

1 D’Amico G, Garcia-Tsao G, Pagliaro L. Natural historyand prognostic indicators of survival in cirrhosis: a sys-tematic review of 118 studies. J Hepatol 2006; 44: 217–31.

2 Schuppan D, Afdhal NH. Liver cirrhosis. Lancet 2008; 371:838–51.

3 Solà E, Watson H, Graupera I et al. Factors related toquality of life in patients with cirrhosis and ascites: rel-evance of serum sodium concentration and leg edema. JHepatol 2012; 57: 1199–206.

4 European Association for the Study of the Liver. EASL clini-cal practice guidelines on the management of ascites, spon-taneous bacterial peritonitis, and hepatorenal syndrome incirrhosis. J Hepatol 2010; 53: 397–417.

5 Wong F. Management of ascites in cirrhosis. J GastroenterolHepatol 2012; 27: 11–20.

6 Cárdenas A, Arroyo V. Refractory ascites. Dig Dis 2005; 23:30–8.

7 Gassanov N, Semmo N, Semmo M, Nia AM, Fuhr U, Er F.Arginine vasopressin (AVP) and treatment with argininevasopressin receptor antagonists (vaptans) in congestiveheart failure, liver cirrhosis and syndrome of inappropriateantidiuretic hormone secretion (SIADH). Eur J Clin Phar-macol 2011; 67: 333–46.

8 Decaux G, Soupart A, Vassart G. Non-peptide arginine-vasopressin antagonists: the vaptans. Lancet 2008; 371:1624–32.

9 Zmily HD, Daifallah S, Ghali JK. Tolvaptan, hyponatremia,and heart failure. Int J Nephrol Renovasc Dis 2011; 4: 57–71.

10 Yi JH, Shin HJ, Kim HJ. V2 receptor antagonist; tolvaptan.Electrolyte Blood Press 2011; 9: 50–4.

11 Sakaida I, Kawazoe S, Hasebe C et al. Dose-finding trial oftolvaptan in liver cirrhosis patients with increased bodyfluid. Hepatol Res 2013; doi: 10.1111/hepr.12099.

12 Uemura M, Matsumoto M, Tsujii T et al. Effects of “bodycompression” on parameters related to ascites formation:therapeutic trial in cirrhotic patients. J Gastroenterol 1999;34: 75–82.

13 Shibata M, Ueno Y, Sumino Y et al. Clinico-pathologicalstudies of alcoholic liver disease, especially on the diagnos-tic significance of endogenous bile acid tolerance test.Arukoru Kenkyuto Yakubutsu Ison 1989; 24: 511–21.

14 Oriuchi N, Nakajima T, Mochiki E et al. A new, accurateand conventional five-point method for quantitative evalu-ation of ascites using plain computed tomography incancer patients. Jpn J Clin Oncol 2005; 35: 386–90.

15 Ginès P, Cárdenas A, Arroyo V, Rodés J. Management ofcirrhosis and ascites. N Engl J Med 2004; 350: 1646–54.

16 Habib S, Boyer TD. Vasopressin V2-receptor antagonists inpatients with cirrhosis, ascites and hyponatremia. TherapAdv Gastroenterol 2012; 5: 189–97.

17 Kim WR, Biggins SW, Kremers WK et al. Hyponatremia andmortality among patients on the liver-transplant waitinglist. N Engl J Med 2008; 359: 1018–26.

18 Moore KP, Aithal GP. Guidelines on the management ofascites in cirrhosis. Gut 2006; 55: vi1–12.

19 Laffi G, Gentilini P, Romanelli RG, La Villa G. Is the use ofalbumin of value in the treatment of ascites in cirrhosis?The case in favour. Dig Liver Dis 2003; 35: 660–3.

20 Assy N, Kayal M, Mejirisky Y, Gorenberg M, Hussein O,Schlesinger S. The changes in renal function after a single

< 2.5 g/dL Albumin level ≥ 2.5 g/dL

TolvaptanPlacebo TolvaptanPlacebo

P = 0.0163

P < 0.0001

Change in b

ody w

eig

ht (k

g)

–5

–4

–3

–2

–1

0(n = 15) (n = 24) (n = 65) (n = 58)

Figure 5 Stratified analysis of change in bodyweight frombaseline on the final dosing day based on serum albumin levelof 2.5 g/dL. Data are expressed as mean 1 standard deviation.Comparisons between the tolvaptan group and the placebogroup were performed by using Student’s t-test.

Hepatology Research 2014; 44: 73–82 Tolvaptan for improvement of hepatic edema 81

© 2013 The Japan Society of Hepatology

dose of intravenous furosemide in patients with compen-sated liver cirrhosis. BMC Gastroenterol 2006; 6: 39.

21 Doungngern T, Huckleberry Y, Bloom JW, Erstad B. Effectof albumin on diuretic response to furosemide in patientswith hypoalbuminemia. Am J Crit Care 2012; 21: 280–6.

22 Dharmaraj R, Hari P, Bagga A. Randomized cross-over trialcomparing albumin and frusemide infusions in nephroticsyndrome. Pediatr Nephrol 2009; 24: 775–82.

23 Péron JM, Bureau C, Gonzalez L et al. Treatment of hepa-torenal syndrome as defined by the international ascitesclub by albumin and furosemide infusion according to thecentral venous pressure: a prospective pilot study. Am JGastroenterol 2005; 100: 2702–7.

24 Na KY, Han JS, Kim YS et al. Does albumin preinfusionpotentiate diuretic action of furosemide in patients withnephrotic syndrome? J Korean Med Sci 2001; 16: 448–54.

82 I. Sakaida et al. Hepatology Research 2014; 44: 73–82

© 2013 The Japan Society of Hepatology