Tolvaptan, an Oral Vasopressin V2 Receptor Antagonist for Heart Failure?

What Did These Important Studies Show? 

In the EVEREST trials, HF patients hospitalized because of worsening symptoms or signs of congestion received either a fixed, once daily 30-mg dose of tolvaptan or a placebo, beginning no later than 48 hours after admission and continuing through the end of the outcome study. The study was conducted at 359 sites in the Americas and Europe. Because of the regulatory obligation to have 2 separate trials, the short-term study was divided into 2 identically conducted trials enrolling 2,048 and 2,085 patients, respectively. After hospital discharge, patients were pooled into the single long-term study comprising 4,133 patients. Subjects were equally divided between active drug and placebo groups.

Eligibility requirements included chronic HF, hospitalization primarily for worsening congestive symptoms, and left ventricular ejection fraction of 40% or less. Principal exclusion criteria were recent heart surgery, acute myocardial infarction as the cause for HF, refractory end-stage HF, systolic blood pressure less than 90 mm Hg, or serum creatinine greater than 3.5 mg/dL (309 μmol/L). The full range of conventional pharmacologic therapies for HF including diuretics, digoxin, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, aldosterone blockers, β-blockers, nitrates, and hydralazine could be prescribed at the discretion of each patient’s treating physician.

Tolvaptan was well tolerated; only 2.5% of patients discontinued it during the index hospitalization. In short-term trial A, there was a statistically significant difference between the tolvaptan-treated patients and controls in the predetermined primary endpoint, a composite of weight decrease at the earlier of hospital discharge or day 7 and change in patient-assessed global clinical status measured with a visual analog scale. However, the difference was driven entirely by the weight change induced by the aquaretic effect of the active drug since improvement in visual analog scale did not differ between the 2 groups. The respective weight losses were 3.35 ± 3.27 kg versus 2.73 ± 3.34 kg (P < 0.001). Among the roughly half of subjects reporting dyspnea at baseline, 76.74% of tolvaptan-treated individuals reported improvement at day 1, compared to 70.61% in the control group (P < 0.001). Results in trial B were similar, except that only in trial B was there a difference in the percentage of patients showing at least a 2-point (on a 5-point scale) change in clinical edema score (73.67% for tolvaptan v 70.81% for placebo, P < 0.02). In a pooled analysis of both trials, there were small differences in physician-assessed dyspnea, orthopnea, fatigue, jugular venous distension, rales, and edema on days 1 or 2, but the differences did not persist to day 4 for dyspnea, orthopnea, or jugular venous distension.

Mean reduction in furosemide dosage from baseline was slightly greater in the pooled tolvaptan group, 55.8 mg/d versus 42.9 mg/d (P = 0.002). As would be expected from the primary effect of the drug, the change in serum sodium from baseline was higher with tolvaptan, 2.01 ± 4.55 mEq/L (mmol/L) versus −1.06 ± 3.90 mEq/L (P < 0.001). The change in serum creatinine from baseline was clinically negligible, 0.08 ± 0.30 mg/dL (7 ± 27 μmol/L) in the tolvaptan group versus 0.02 ± 0.27 mg/dL (2 ± 24 μmol/L; P < 0.001) in controls. Blood urea nitrogen rose less in the tolvaptan group, 2.55 ± 11.0 mg/dL (0.9 ± 3.9 mmol/L) versus 3.22 ± 10.86 mg/dL (1.1 ± 3.9 mmol/L; P < 0.01). Overall, the investigators concluded that oral tolvaptan safely improved many of the congestive symptoms of hospitalized HF patients.

The long-term outcome trial followed the patients after hospital discharge while they continued to receive tolvaptan 30 mg or placebo. During this time, a median of 9.9 months (range, 2-24 months), 22% of tolvaptan and 21% of placebo patients withdrew from the trial, roughly half because of patient request and a quarter due to adverse events. Thirst (16.0% v 2.1%) and hypernatremia (1.7% v 0.5%) were more common with tolvaptan. The primary study endpoints were all-cause mortality and the composite endpoint was defined as cardiovascular death or HF hospitalization. In the tolvaptan group, all-cause mortality was 25.9%, versus 26.3% in the controls, corresponding to a hazard ratio of 0.96 (95% confidence interval, 0.87-1.11). The composite endpoint was reached in 42.0% of tolvaptan-treated patients and 40.2% of controls, with a hazard ratio of 1.04 (95% confidence interval, 0.95-1.14). Mortality at 1 year was 25.0% in the tolvaptan patients and 26.0% in the placebo group. Kaplan-Meier curves for the 2 treatment groups for each of the 2 primary endpoints were superimposable. These findings highlight the very high severity of illness in this population. There was no difference in questionnaire-assessed quality of life. The authors’ conclusion was that tolvaptan produced no long-term mortality or morbidity benefit in HF patients.

Taken together, these studies, which included formal noninferiority testing, show that long-term selective V2R blockade with tolvaptan in HF patients is neither beneficial nor harmful. The lack of benefit is disappointing, but the proof of lack of harm is important since other promising HF treatments have been shown in outcome trials to be deleterious.10, 11 Tolvaptan was associated with statistically significant, but clinically very modest, improvements in acute congestive symptoms and a reduction in diuretic dose. Its major side effect was thirst. On the basis of the short-term EVEREST findings, tolvaptan may be useful in selected HF patients in whom the addition of even a small aquaresis may be beneficial. Although 8% of the patients were hyponatremic and serum sodium improved in the treatment group, neither EVEREST study addressed the role of tolvaptan for treating hyponatremia per se.

How Do These Studies Compare With Prior Studies? 

The Acute and Chronic Therapeutic Impact of a Vasopressin 2 Antagonist (Tolvaptan) in Congestive Heart Failure (ACTIV in CHF) trial was a dose-ranging trial of tolvaptan in HF. Its results were similar to the short-term EVEREST trial except that in a post-hoc analysis, 60-day mortality was lower in the tolvaptan group than controls in the subset of patients with the most severe congestive symptoms or with prerenal azotemia.12 The study was not powered to evaluate mortality, and the survival difference was not confirmed in the general population of HF patients studied in EVEREST. Interestingly, a recent small trial comparing left ventricular volume parameters in HF patients receiving 30 mg tolvaptan or placebo for a year showed a significant reduction in the combined endpoint of mortality or HF hospitalization in a nonprespecified analysis. This study demonstrated no difference in its primary outcomes, left ventricular end diastolic volume or several related left ventricular volume measures.5

Although not detailed in EVEREST, there are good data on the use of tolvaptan to treat hyponatremia in HF. Both the ACTIV in CHF and the SALT trials showed that tolvaptan was effective in raising serum sodium in hyponatremic patients with CHF. The SALT trials also showed that correction of hyponatremia was associated with improved cognition.6, 12

What Should Clinicians and Researchers Do? 

The 2 large-scale EVEREST trials were designed to determine whether tolvaptan affects the short- or long-term outcome of HF. While statistically significant effects were seen in the short-term study, the clinical effects were very modest. Nephrologists discussing fluid balance often stress that administration of electrolyte-free water only minimally affects intravascular volume. As the small incremental aquaresis produced by tolvaptan was associated with an improvement in congestive symptoms, a more circumspect view may be appropriate, at least in patients with severe HF. It seems doubtful that tolvaptan therapy of all acutely hospitalized patients with decompensated HF is justified. Length of stay data were not reported in EVEREST, and how tolvaptan will be priced once it is approved by the US Food and Drug Administration has not been disclosed. Whether a reduced length of stay might mitigate the cost of adding tolvaptan for short-term treatment is not known. As the EVEREST trials provide clear evidence that long-term treatment of HF patients with tolvaptan confers no survival benefit, it is not yet time to add routine V2R blockade to the arsenal of HF treatments. However, the study proved the drug to be safe, providing reassurance about its eventual use to lessen fluid restriction and potentially improve cognitive function in HF patients who are hyponatremic. It is likely that tolvaptan will be marketed for oral use, making it the preferred agent for long-term and outpatient use. In the meantime, and for patients in whom intravenous therapy is preferred, the nonselective V1a/V2R antagonist conivaptan is available to clinicians now. It is approved by the Food and Drug Administration for short-term intravenous administration to treat euvolemic and hypervolemic hyponatremia, including hyponatremia with HF.

Several questions remain unanswered. Animal data indicate that stimulation of the V1aR promotes myocardial cell protein synthesis and hypertrophy in addition to its well-known pressor effects.5 Endogenous vasopressin levels are elevated during chronic selective V2R blockade, exposing treated patients to increased activation of the V1aR. We do not know how chronic blockade of the V1aR might affect HF outcomes. Although it may simply be that V2R blockade has no effect in HF, it is also possible that a favorable effect of V2R blockade by tolvaptan in EVEREST was offset by a detrimental myocardial effect of unopposed and perhaps even augmented V1aR activation. Long-term trials of both selective V1aR blockade and combined V1a/V2R blockade with conivaptan or another congener are essential. We know little about the hemodynamic effects of V1aR or V2R blockade. In a short-term study, conivaptan lowered pulmonary capillary wedge and right atrial pressures in patients with HF, although no alteration in cardiac output or in pulmonary or systemic blood pressure or vascular resistance was observed.13 Whether the changes were due to direct vascular effects of receptor blockade or just due to the aquaresis is unknown. More data on hemodynamic effects of selective V1aR or V2R blockade as well as nonselective vasopressin receptor blockade would be welcome. Although further subgroup analysis may be forthcoming, as reported to date, the EVEREST trials give no information on subsets of patients with HF in whom tolvaptan therapy might produce a survival improvement. The finding of improved outcomes in 2 smaller trials suggest that tolvaptan may be beneficial in more severe HF.5, 12 This possibility also needs more scrutiny.