Vaptans may correct hyponatraemia and mobilise ascites through an increased excretion of water. The effect on clinical outcomes is debated.
Vaptans may correct hyponatraemia and mobilise ascites through an increased excretion of water. The effect on clinical outcomes is debated.
To determine the effects of vaptans (tolvaptan, satavaptan and lixivaptan) on patients with cirrhosis and hyponatraemia or ascites.
Systematic review of randomised controlled trials. The primary outcome measure was mortality. Electronic and manual searches were combined (April 2012). Data were extracted from published reports, online information from the Food and Drug Administration website or obtained through correspondence with authors and pharmaceutical companies. The primary meta-analyses were performed using random effects models due to an expected clinical diversity.
Twelve trials with a total of 2266 patients were included. Randomisation was adequate in all trials. Eight trials were double-blind. Random effects meta-analyses found no clear differences between vaptans and control groups regarding mortality (RR = 1.06, 95% CI = 0.90–1.26, I2 = 0%), variceal bleeding, hepatic encephalopathy, spontaneous bacterial peritonitis, hepatorenal syndrome, or renal failure. Vaptans increased serum sodium levels (WMD = 1.8 mmol/L, 95% CI = 0.79–2.96) and lead to reductions in weight and the time to the first paracenthesis. Vaptans increased the risk of adverse events (RR = 3.97, 95% CI = 1.78–8.83), including an excessive urine volume (RR = 9.96, 95% CI = 1.38–71.68).
Vaptans have a small beneficial effect on hyponatraemia and ascites, but do not affect mortality, complications to cirrhosis or renal failure. The data do not support the routine use of vaptans in cirrhosis.
Ascites is a serious complication to cirrhosis, occurring in 50% of patients within 10 years of the diagnosis. The mortality rates within the first 2 years after the development of ascites are about 50%.[1, 2] The recommended treatment for ascites is diuretics (spironolactone and furosemide), but about 15% of patients are nonresponders. Patients with refractory ascites have reduced survival and an increased risk of complications to cirrhosis.[4, 5]
The haemodynamic changes seen in cirrhosis lead to an increased secretion of arginine vasopressin resulting in water retention and dilutional hyponatraemia. Use of traditional diuretics also reduces serum sodium. Clinically, significant hyponatraemia is a frequent problem in patients with ascites. Vaptans are nonpeptide vasopressin receptor antagonists, which block the V2 receptors in the renal collecting ducts. The result is an increased free water clearance with hypotonic diuresis. Vaptans are approved for hyponatraemia in patients with inappropriate antidiuretic hormone secretion. Vaptans may also reduce hyponatraemia and ascites in cirrhosis, but the clinical effects of vaptans in this patient group remain unknown.
This systematic review is based on a written protocol prepared according to the recommendations specified in the Cochrane Handbook for Intervention Reviews. The main objective was to analyse the beneficial and harmful effects of vaptans (tolvaptan, satavaptan and lixivaptan). Randomised controlled trials were included irrespective of publication status or language. Patients diagnosed with cirrhosis and hyponatraemia or ascites were eligible for inclusion (using the diagnostic criteria specified in the included trials). The intervention comparisons were vaptans vs. no intervention, placebo or other diuretics (furosemide or spironolactone). No specific criteria were made regarding the dose or duration of treatment.
Eligible trials were identified through electronic and manual searches (last update April 2012). Electronic searches were performed in Medline (through Pubmed 1966–2012), Embase (through Ovid 1928–2012) and Web of Science (1900–2012). The searches in Pubmed were performed using the terms vaptans OR tolvaptan OR satavaptan OR lixivaptan OR conivaptan OR mozavaptan OR nonpeptide arginine vasopressin antagonist.
Manual searches included scanning of reference lists in studies on vaptans and conference proceedings from the meetings of the European Association for the Study of Liver Disease and the American Association for the Study of Liver Disease. Additional trials were identified through correspondence with experts and pharmaceutical companies, and trial registers (http://www.who.int/ictrp). Potentially eligible trials were listed. Excluded trials were listed with the reason for exclusion.
The primary outcome measure was mortality. Secondary outcomes were complications to cirrhosis (hepatic encephalopathy, variceal bleeding, hepatorenal syndrome and spontaneous bacterial peritonitis), renal failure, serum sodium levels, mobilisation of ascites (as specified by authors of included trials, including changes in body weight after treatment, number of patients undergoing large volume paracentesis, time to the first large volume paracentesis, number of necessary large volume paracentesis, number of patients with reduced ascites and clinical severity of ascites), vasoactive hormones (vasopressin, renin and aldosterone levels), systolic and diastolic blood pressure, quality of life and adverse events (as defined by authors of included trials).
Authors extracted data independently. Data were extracted from the published trials retrieved through correspondence with authors, drug companies (Otsuka and Sanofi-aventis), the US Food and Drug Administration (FDA) homepage (http://www.fda.gov/), and the European Medicines Agency (EMEA) homepage (http://www.ema.europa.eu/ema/).[8-10]
On the basis of previous evidence, the randomisation methods (allocation sequence generation and concealment) were defined as the primary measure of bias control. The allocation sequence generation was classed as adequate if based on a table of random numbers, computer-generated random numbers, or similar. The allocation concealment was classed as adequate if based on central randomisation, identically appearing coded drug containers, serially numbered opaque sealed envelopes, or similar. We also extracted blinding (whether the trial was described as double-blind or single-blind, the method of blinding, whether patients, investigators, outcome assessors or other subjects involved in the trial were blinded, and whether the adequacy of blinding was assessed), the risk of reporting bias (based on the definition and reporting of clinically relevant outcome measures), the risk of attrition bias (numbers and reasons for dropouts and withdrawals and whether all patients were accounted for in the report and analysis of the trial) and other biases (whether sample size calculations were performed, and whether the preset sample size was achieved. For trials terminated prematurely, we registered whether this was based on predefined criteria. For all trials, we also retrieved data on whether the trials were registered in a trial database (e.g. in ClinicalTrials.gov) before the trial was terminated.
Data analyses were performed in Review Manager version 5 (The Nordic Cochrane Centre, Copenhagen, Denmark) and stata version 11 (Stata Corp, College Station, TX, USA). The primary meta-analyses were performed using random effects models due to an expected clinical diversity. Fixed effect model meta-analyses were performed as sensitivity analyses and reported when the conclusions differed from the random effects model meta-analyses. Data on all patients randomised were sought for all outcome measures. The results were expressed as relative risks (RR) or weighted mean differences (WMD) with 95% CI and I2 values as a marker of intertrial heterogeneity (with values >50% defined as considerable heterogeneity).
The risk of bias and sources of intertrial heterogeneity were assessed in subgroup, sensitivity and regression analyses. The subgroup analyses evaluated the influence of the interventions (type of vaptan and concomitant use of traditional diuretics), patient characteristics (inclusion of patients with ascites or hyponatremia at baseline) and the control of bias (excluding trials without adequate randomisation). The risk of small study effects (including publication bias and other biases) were analysed in Egger's regression analyses of funnel plot asymmetry (Egger's test).
A total of 1969 potentially eligible references were identified in the electronic and manual searches (Figure 1). After excluding duplicates, observational studies and trials that did not fulfil our inclusion criteria, 28 references were retrieved for further evaluation. Fifteen references referred potentially eligible studies.[12-25],  Ten of these references referred to 12 trials[12-16], [22-25],  that fulfilled our inclusion criteria.
Three of the included trials were described in one full paper article. Six of the remaining trials were published as full paper articles. Three trials were published in abstract form. We retrieved additional data on trial characteristics through correspondence with Sanofi-Aventis (satavaptan) and Otsuka Pharmaceutical (tolvaptan).
All trials were multicentered, multinational and industry funded. The trials were performed in Europe, the Americas, Singapore, Taiwan, South Africa, Russia, Turkey and Israel. Most trials were designed to assess changes in ascites and correction of hyponatraemia (Table 1).
|Trial||Interventions||Additional diuretics allowed||Maximum treatment duration||Reported primary outcomes||Maximum follow-up|
|Gines 2010||Satavaptan 5, 12.5 or 25 mg/day vs. placebo||Yes||14 days||Changes in ascites||21 days|
|Wong 2010||Satavaptan 5, 12.5 or 25 mg/day vs. placebo||Yes||12 weeks||Changes in ascites||12 weeks|
|Gines 2008||Satavaptan 5, 12.5 or 25 mg/day vs. placebo||Yes||14 days||Changes in ascites and serum sodium||14 days|
|Wong 2003||Lixivaptan 50, 250 or 500 mg/day vs. placebo||Yes||8 days||Safety, efficacy and serum sodium changes in one trial||8 days|
|Gerbes 2003||Lixivaptan 100 or 200 mg/day vs. placebo||NR||7 days||Changes in serum sodium||7 days|
|Wong 2012a||Satavaptan 5–10 mg/day vs. placebo||Yes||52 weeks||Changes in ascites and serum sodium||52 weeks|
|Wong 2012b||Satavaptan 5–10 mg/day vs. placebo||Yes||52 weeks||Changes in ascites and serum sodium||52 weeks|
|Wong 2012c||Satavaptan 5–10 mg/day vs. placebo||No||52 weeks||Changes in ascites and serum sodium||52 weeks|
|Wong 2009||Satavaptan 5–50 mg/day vs. placebo||Yes||52 weeks||Complications to cirrhosis and change in serum sodium or ascites||52 weeks|
|Gines 2008a||Satavaptan 5–50 mg/day vs. placebo||Yes||52 weeks||Complications to cirrhosis and change in serum sodium or ascites||52 weeks|
|Gines 2007||Satavaptan 5–50 mg/day vs. placebo||Yes||52 weeks||Complications to cirrhosis and change in serum sodium or ascites||12 months|
|Cardenas 2012||Tolvaptan 15–60 mg/day vs. placebo||Yes||52 weeks||Changes in serum sodium|
|Trial||Child–Pugh score (mean)||Proportion with ascites||Baseline serum sodium||No. pts vaptan/control|
|Gines 2010||8.2||100%||130–142 mmol/L||113/35|
|Wong 2010||8.9||100%||<135 mmol/L (40%) and 135–142 mmol/L (60%)||115/36|
|Gines 2008||9.9||100%||<130 mmol/L||82/28|
|Wong 2003||9.9||100%||<130 mmol/L||25/8|
|Gerbes 2003||NR||100%||115–132 mmol/L||40/20|
|Wong 2012a||7.9||100%||Mean 137 mmol/L||232/230|
|Wong 2012b||8.8||100%||Mean 136 mmol/L||328/168|
|Wong 2012c||9.0||92.1%||Mean 135 mmol/L||160/80|
|Wong 2009||NR||100%||Mean 135 mmol/L||186/48|
|Gines 2008a||NR||100%||<130 mmol/L||92/47|
|Gines 2007||NR||100%||Mean 132 mmol/L||47/26|
|Cardenas 2012||NR||NR||<135 mmol/L||63/57|
The total number of patients was 2266 (Table 1).The mean age ranged from 53 to 60 years and the proportion of men from 63% to 78%. All trials included patients with clinical and histological cirrhosis. Most trials included patients with ascites. The proportion of patients with hyponatraemia at baseline was 100% in six trials[13-15], , ,  and 40% in one trial. The remaining five trials did not report the proportion of patients with hyponatraemia at baseline. The number of patients randomised to vaptans was 1483. Control groups included 783 patients. The trials assessed satavaptan (n = 9), lixivaptan (n = 2) and tolvaptan (n = 1). The treatment duration ranged from 1 week to 12 months. The dose ranged from 5 to 25 mg/day for satavaptan, 50 to 500 mg/day for lixivaptan and 15 to 60 mg/day for tolvaptan.
All trials used adequate randomisation methods with computer-generated random numbers and central randomisation. Only one trial reported baseline differences between the intervention and the control group. Eight trials were double-blind.[12, 13, 15, 16, 20, 21, 25, 28] Four trials reported blinding, but did not report whether patients, investigators or other subjects involved in the trial were blinded or how blinding was achieved.[14, 22-24] None of the included trials tested the effect of blinding. Mortality was reported in nine trials and cirrhosis complications in ten trials. Evidence of reporting bias was found in the remaining trials. Follow-up was clearly described for in all trials and no evidence of attrition bias was identified. Eight trials reported sample size calculations.[12, 13, 15, 16, 25, 28] Two trials (described in one publication) were terminated prematurely due to increased mortality or adverse events. Seven trials were registered in clinical trial databases (http://www.clinicaltrials.gov). Two trials were registered before the study start,[21, 25] two trials after the first patients were enrolled[25, 28] and trials were registered after the trial termination date.[12, 15, 16]
Nine trials reported mortality (seven trials as crude mortality rates and two trials as Kaplan–Meier estimated rates). In total, 294 of 1329 patients (22%) randomised to vaptans and 147 of 721 patients (20%) randomised in the control group died (Figure 2). Random effects meta-analysis found no difference in mortality between the vaptan and control group (RR = 1.06, 95% CI = 0.90–1.26, I2 = 0%). We were able to retrieve data on complications to cirrhosis from 10 trials. No difference between the vaptan and control groups were identified when analysing the outcomes hepatic encephalopathy (262 of 1314 vs. 156 of 709 patients, RR = 0.87, 95% CI = 0.73–1.03, I2 = 0%, 9 trials) or variceal bleeding (85 of 1329 vs. 42 of 721 patients, RR = 1.04, 95% CI = 0.56–1.92, I2 = 46%, 9 trials, Figure 3). There was no clear effect of vaptans on spontaneous bacterial peritonitis (121 of 1266 vs. 81 of 642 patients, random effects model RR = 0.73, 95% CI = 0.53–1.01, fixed effect model RR = 0.76, 95% CI = 0.58–0.99, I2 = 21%, 8 trials). We were unable to extract data on hepatorenal syndrome. Eight trials reported renal impairment defined as an increase in creatinine of at least 50% from baseline or serum creatinine >133 mmol/L. On the basis of these definitions, renal impairment was registered for 271 of 1237 patients in the vaptan group vs. 126 of 674 control group patients (RR = 1.16, 95% CI = 0.88–1.53, I2 = 28%).
Meta-analysis of four trials[12-14, 16] showed that post treatment serum sodium levels were 2.02 mmol/L higher in the vaptan compared with the control groups (95% CI = 0.99–3.06, I2 = 0%). Two trials[15, 28] found that satavaptan was associated with a greater reduction in serum sodium compared with placebo (WMD = 3.33 mmol/L, 95% CI = 1.88–4.78, I2 = 0%).
Random effect meta-analyse of vaptans vs. control groups found that vaptans had beneficial effect on changes in body weight (WMD = −1.82 kg, 95% CI = −2.86–0.79, I2 = 0%, 2 trials), the time to the first large volume paracentesis (RR = 0.76, 95% CI = 0.63–0.90, I2 = 0%, 3 trials) and the clinical severity of ascites (RR = 0.71, 95% CI = 0.60–0.83, I2 = 0%, 2 trials). The number of required of large volume paracenteses was not different in the intervention and control groups (RR = 0.88, 95% CI = 0.72–1.03, I2 = 0%, 2 trials).
Three trials reported changes in vasopressin, renin and aldosterone levels and post-treatment blood pressure and heart rate.[12, 15, 16] Vaptans increased vasopressin (WMD = 1.76 ng/L, 95% CI = 1.53–1.98) and renin levels (WMD = 51.98 mU/L, 95% CI = 27.46–76.50). There was no clear difference in aldosterone levels (random effects WMD = −17.90 pM, 95% CI = −702.32–666.53 and fixed effect WMD = 336.93 pM, 95% CI = 231.10–442.76). Vaptans had no effect on the systolic blood pressure (WMD = 0.51 mmHg, 95% CI = −3.66–4.67), diastolic blood pressure (WMD = 0.47 mmHg, 95% CI = −2.14–3.08) or heart rate (WMD = 0.40 beats/minute, 95% CI = −3.72–4.51).
Vaptans did not increase the total number of adverse events (RR = 1.01, 95% CI = 0.97–1.05), serious adverse events (RR = 1.04, 95% CI = 0.94–1.15) or dropouts and withdrawals (RR = 0.78, 95% CI = 0.56–1.10). Vaptans increased the number of patients with thirst (RR = 3.97, 95% CI = 1.78–8.83) and excessive (above 5 L/day) urine volume (RR = 9.96, 95% CI = 1.38–71.68). There were no differences between vaptans and control interventions regarding the number of patients with a rapid increase (>8 mmol/L/day) in serum sodium (RR = 1.35, 95% CI = 0.65–2.80), serum creatinine (WMD = 1.01 mmol/L, 95% CI = −7.20–9.23) or serum potassium >5.5 mmol/L (RR = 1.14, 95% CI = 0.91–1.43).
In subgroup analyses, there was no difference between the intervention and control groups in trials on satavaptan (RR = 1.05, 95% CI = 0.88–1.25) or tolvaptan (RR = 1.13, 95% CI = 0.32–1.25), trials in which concomitant use of traditional diuretics was allowed (RR = 1.06, 95% CI = 0.87–1.29), trials on patients with ascites (RR = 1.05, 95% CI = 0.88–1.25) and trials that only included patients with hyponatraemia (RR = 0.82, 95% CI = 0.56–1.20). Exclusion of trials without adequate randomisation and double blinding did not influence the overall result. Regression analysis found no evidence small study effects (publication bias and other biases) in the assessment (Egger's test P = 0.351).
This systematic review found that vaptans increase serum sodium and mobilise ascites, but have no beneficial or detrimental effects on mortality or morbidity in patients with cirrhosis and ascites or hyponatraemia. Relatively frequent nonserious adverse events, including excessive thirst and urine volume, were identified. The combined evidence does not support the routine use of vaptans in patient with cirrhosis.
It may be argued that the inclusion of trials published in abstract form increases the risk of bias in meta-analyses. When trials are not available as full paper articles, the methods may be difficult to assess increasing. The risk of hidden biases may limit the strength of the overall conclusions. However, the exclusion of trials published as abstracts increases the risk of dissemination bias due to the selective publication of trials with statistically significant results.
The safety of vaptans has been debated. Satavaptan is not approved for use outside clinical trials due to safety concerns. Our meta-analysis did not show increased mortality or complications to cirrhosis or an increased risk of serious adverse events. However, the statistical power of the included trials and the duration of follow-up in included trials limit the strength of our findings. Two of the included trials were terminated due to adverse events. In one trial, an interim analysis found that mortality was 31% in the satavaptan and 22% in the placebo group. The increased mortality reflected an increased risk of serious cirrhosis complications and adverse events, including hypovolaemia, thromboembolic disorders, cardiac events, respiratory disorders, cancer and variceal bleeding. The second trial found a threefold increase in serum bilirubin in the satavaptan group, but no changes in liver function tests. Satavaptan was also associated with increased creatinine and prolonged QTcF. These adverse events did not increase mortality. The trial was terminated after the interim analyses found a poor risk benefit ratio.
In May 2009, the FDA approved tolvaptan for hypervolaemic or euvolaemic hyponatriaemia. The FDA underlined that special caution was necessary when treating patients with cirrhosis. Preclinical data suggested that tolvaptan depletes vitamin-K-dependent clotting factors and inhibits platelet aggregation. These changes may explain why one trial found higher rates of gastrointestinal bleeding, haematomas and ecchymoses in the tolvaptan (18%) than the placebo groups (2%) after 30 days of treatment.
Patients with cirrhosis and hyponatraemia may develop hepatic encephalopathy.[4, 5, 27, 28] Whether correction of hyponatraemia reverts the cerebral changes remains unknown. The EMEA withdrawal assessment report of satavaptan described concerns on hepatic encephalopathy, elevated potassium levels and renal failure. The EMEA report was based on a mixed population of patients with chronic heart failure, syndrome of inappropriate antidiuretic hormone and cirrhosis. We only included patients with cirrhosis and found no clear effect of satavaptan on these outcomes. This may reflect insufficient statistical power or the duration of follow as well as our patient inclusion criteria.
At present, tolvaptan and conivaptan are approved in the USA and Europe. The current prices are high. The cost of tolvaptan 15 mg is 88 Euro, whereas conivaptan 20 mg costs 393.75 USD. Considering the low price of traditional diuretics, considerable clinical benefits should be evident to achieve an acceptable cost-benefit ratio. The combined evidence does not support the routine use of vaptans for patients with cirrhosis.
Declaration of personal and funding interests: None.