Climbing the Mountain of Acute Decompensated Heart Failure: Title and subTitle BreakThe EVEREST Trials

In this issue of JAMA, 2 articles report findings from the Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study With Tolvaptan (EVEREST) clinical trials program. Konstam et al¹ report the overall long-term effects of tolvaptan in terms of safety and clinical outcomes for patients with worsening heart failure (HF) requiring hospitalization. Gheorghiade et al² report short-term effects of tolvaptan for acute symptom relief in these same patients, but presented as 2 trials that were based on study center assignment following completion of patient enrollment and randomization in the long-term trial. The aggregate findings demonstrate that tolvaptan relieves some symptoms associated with acute decompensated heart failure (ADHF) and has no demonstrable evidence of harm such as worsening of renal failure, but importantly, does not reduce mortality or HF-related morbidity at 1 year.

In the context of ADHF, these are noteworthy findings. To date, no other therapeutic intervention has been demonstrated in large-scale randomized, placebo-controlled studies to positively influence symptoms in ADHF without generating a question of harm. Inotropes have been associated with increased mortality risk,³ and calcium sensitizers have been associated with increased cardiovascular mortality.⁴ Nesiritide was shown to relieve dyspnea and to reduce pulmonary congestion over the short term, but also was associated with increased intermediate-term mortality risks.⁵ Traditional vasodilator therapy relieves HF symptoms, but neither nitroglycerin nor nitroprusside has been subjected to prospective randomized controlled study, and effects on mortality are unknown. Even diuretics, which clearly relieve congestion, have uncertain effects on outcomes.^{6 - 7}

Addressing the challenge of ADHF has been daunting. Acute decompensated HF accounts for more than 1 million acute hospitalizations per year in the United States at an annual cost of more than $30 billion and is associated with significant mortality.⁸ The risk of inpatient death is approximately 4% (but ranges from 2%-22%),⁹ and the risk of death/rehospitalization at 60 to 90 days after an episode of ADHF is 36%.¹⁰ At the core of this challenge, however, is the limited understanding of the pathophysiology of ADHF. While hemodynamic disturbances are clearly responsible for the observed signs and symptoms, plausible causative mechanisms that trigger acute decompensation, eg, ventricular injury and augmented neurohormonal activation, remain uncertain. Thus, therapy for ADHF is relegated to correcting perturbed hemodynamics. Comparison is made with the evolution of therapy for chronic HF—targeting hemodynamic abnormalities did not yield meaningful clinical benefits, but elucidation of relevant pathophysiological mechanisms followed by appropriately focused interventions in large-scale trials resulted in salutary improvements. A similar approach is needed for ADHF.

Another challenge of ADHF involves clinical trial design and construct. The studies to date have been small to moderate short-term hemodynamic or symptom-focused designs, constructed primarily to meet regulatory requirements. Important questions, including mechanistic hypotheses and the effect of interventions on rehospitalizations/mortality, have been inadequately studied. Trials in ADHF have also been confounded by considerable heterogeneity of the patient phenotype. Patients with ADHF are older, represent a mix of new-onset HF and decompensated chronic HF, have both reduced and preserved ejection fraction HF, have equal gender representation, and frequently have a number of important comorbidities, especially renal insufficiency.¹¹

Moreover, “standard therapy” does not truly exist, as there are no evidence-based interventions proven to reduce morbidity and mortality associated with ADHF. Guidelines for ADHF have recently been put forward, but the treatment algorithms are sparse and focus on hemodynamic targets.¹² The strategy is to relieve symptoms preferentially with diuretic use, including high-dose loop diuretics and continuous diuretic infusions (the safety of which is not known in HF), and to add vasodilator therapy (nitroglycerin or nitroprusside) if adjunctive approaches are needed. Inotropes are reserved for impending or frank shock but are otherwise dissuaded. Ultrafiltration represents at best an emerging technology with intriguing preliminary data but limited applications.^{12 - 13}

In this context of absent clarity regarding the pathophysiology of ADHF, a significant clinical burden of excessive morbidity and mortality, a need to meet regulatory requirements that focus on symptom relief, consistently worrisome signals of risk associated with other therapies for ADHF, and a heterogeneous patient phenotype, emerge the EVEREST data.

Upregulated arginine vasopressin (AVP) stimulation represents a potentially clinically important maladaptive neurohormonal response observed in the setting of left ventricular dysfunction.¹⁴ Arginine vasopressin increases blood volume by promoting free water retention via its interface with the V₂ AVP receptor located on renal cortical collecting ducts. Arginine vasopressin also leads to vasoconstriction and possibly cardiac hypertrophy through the cardiac V_1a receptor.¹⁵ Prior small-scale studies demonstrated a reduction in congestion, decrease in body weight, and correction of hyponatremia with selective inhibition of the AVP V₂ receptor.^{16 - 17}

The EVEREST studies tested the benefit of tolvaptan, a V₂ selective AVP antagonist, given once a day (30 mg vs placebo) in 3 clinical trials: 2 identical short-term trials (needed to meet the regulatory requirement for 2 independent confirmatory trials) and a longer-term safety and outcome trial. A total of 4133 patients with ADHF were randomized in the long-term trial, and based on an algorithm for assigning the study centers, approximately 2000 patients constituted the cohorts for each of the 2 short-term trials. The studies were adequately powered for both short- and long-term end points. The primary outcome measure of the short-term trials was a composite of change in global clinical status and body weight at 7 days or the day of discharge, whichever was earlier. The primary outcome measure of the longer-term trial was a dual primary end point of all-cause mortality (superiority and noninferiority) and cardiovascular death or HF hospitalization (superiority only).

Patients enrolled in EVEREST had been hospitalized with worsening chronic HF and documented evidence of impaired left ventricular ejection fraction. Background medical therapy was quite good and reflected an improvement over that seen in registry data.¹¹ Inpatient risk of death was less than 3%, reflecting better than expected outcomes and likely consistent with recent data that evidence-based therapy for chronic HF improves short-term outcomes in ADHF.¹⁰ The short-term trials met the primary end point and showed greater, but modest, improvement with tolvaptan vs placebo. This benefit was largely driven by change in body weight, as the global clinical score was not different at day 7 or discharge. Secondary end points revealed modest improvements in dyspnea and edema. The longer-term trial demonstrated no reduction in risk of death or hospitalization and no significant worsening of renal function. Hyponatremia, when present, was improved. Taken together, these findings would suggest some short-term benefit of tolvaptan on certain acute symptoms of ADHF without evidence of harm and represent an important contribution for understanding the management of patients with ADHF.

However, enthusiasm for these findings must be tempered by several concerns: (1) the improvements in the global clinical status were not significantly different between the tolvaptan and placebo groups in the short-term studies; (2) the lack of an effect on recurrent HF hospitalizations and mortality supports an absence of harm but does not confirm expected long-term benefit of this neurohormonal antagonist—a disappointing observation; (3) the adverse event rates were high—more so for nuisance events (thirst and dry mouth) than serious events, thus limiting the use of tolvaptan for long-term administration; (4) despite earlier data that AVP antagonism effectively corrects hyponatremia, the number of patients with a serum sodium level of less than 134 mEq/L was too small to generate a signal of enhanced benefit; (5) tolvaptan was given in addition to, not in lieu of, diuretics; although the mean diuretic dose was lower in the tolvaptan-treated groups, there is no evidence that the lower diuretic dose resulted in better outcomes or that tolvaptan is a diuretic alternative; and (6) since all patients studied had a history of established HF with reduced ejection fraction, extrapolation of these data to other patient cohorts with ADHF is not warranted.

Given the results and concerns from the EVEREST trial for patients with ADHF, clinicians should be encouraged to continue to use diuretics judiciously as needed to relieve congestion, with concomitant vasodilators for patients with compelling symptoms at rest. Adjunctive short-term AVP antagonism may be considered for the patient with established low ejection fraction HF and perhaps continued until acute symptoms are relieved. There may be some utility to continue AVP antagonism longer if hyponatremia is present, but this requires further study. Inotropes should be reserved for patients with shock, and fluctuations in renal function should be minimized. Prior to discharge, all indicated evidence-based therapies for chronic HF should be initiated and careful follow-up should ensue.

The data from the EVEREST study suggest that a novel therapy for ADHF has been identified that may potentially have an impact on clinical practice. However, use of tolvaptan must be carefully considered, as evidence of long-term benefit is lacking. Moreover, given the characteristics of the patient cohort in EVEREST, the use of tolvaptan should not be extrapolated to patients who are dissimilar. More research is needed to better understand the role of AVP antagonism in HF, particularly regarding the use of nonselective vasopressin antagonists. Additional study also is required in other ADHF cohorts, especially those with HF and preserved ejection fraction, acute-onset HF, and advanced HF—perhaps as characterized by hyponatremia. As additional reports from EVEREST and from new clinical trials with both medical and device therapies become available, the hope is that progress toward reaching additional evidence-based therapies on the mountain of ADHF will continue.