Intravenous Therapies in Acute Heart Failure – Lack of Effect or Lack of Well Powered Studies?

‘But he hasn't got anything on,’ a little child said. ‘Did you ever hear such innocent prattle?’ said its father. And one person whispered to another what the child had said, ‘He hasn't anything on. A child says he hasn't anything on.’ ‘But he hasn't got anything on!’ the whole town cried out at last. (Hans Christian Andersen, The Emperor's New Clothes) Most intravenous (IV) therapies were introduced into the treatment of acute heart failure (AHF), and continue to be used, based on very limited studies (IV loop diuretics, IV nitrates), no data at all (IV morphine) or studies showing harm (IV inotropes). Despite these severe limitations, some of these interventions are recommended for use in the current guidelines of both the American College of Cardiology (ACC)/American Heart Association (AHA) and European Society of Cardiology (ESC). For example, the recent ESC guideline for the treatment of AHF1 states ‘Most patients with dyspnoea caused by pulmonary oedema obtain rapid symptomatic relief from administration of an i.v. diuretic, as a result of both an immediate venodilator action and subsequent removal of fluid’ without a single reference to back this statement. It proceeds to recommend this therapy as level 1B based on the results of the Diuretic Optimization Strategies Evaluation (DOSE) study (addressed below) in which no placebo arm was included and high-dose IV loop diuretics and very high dose IV loop diuretic strategies were compared. Similar recommendations were made in the ACC/AHA guideline based on similar evidence.2 When addressing the use of IV nitrates in AHF the authors of the current ESC AHF guidelines1 recommend this therapy as level II A/B and suggest that ‘An i.v. infusion of a nitrate should be considered in patients with pulmonary congestion/oedema and a systolic blood pressure >110 mmHg… to reduce pulmonary capillary wedge pressure and systemic vascular resistance. Nitrates may also relieve dyspnoea and congestion… ’. The authors reference two studies to support this recommendation. In the first study,3 the guideline-recommended dosing strategy was found to be ineffective compared with a higher dose based on boluses, which is not commonly used. In addition, that study was limited to the first hours of presentation, included 100 patients, and its primary endpoint was improved oxygenation. The second study is the Vasodilation in the Management of Acute Congestive Heart Failure (VMAC) study,4 designed primarily to assess the efficacy of nesiritide in the treatment of AHF; only 143 patients were treated with IV nitrates, in a dose range similar to the one recommended in the guidelines. At the pre-specified time-points (3 h from randomization), no significant effect of IV nitrates vs. placebo was found on wedge pressure, cardiac index or systemic vascular resistance, or on dyspnoea or general wellbeing. Moreover, in the VMAC study4 IV nitrates were found to be inferior to nesiritide therapy, which was shown in the later large clinical study Acute Study of Clinical Effectiveness of Nesiritide in Decongested Heart Failure (ASCEND HF);5 to have a minimal clinical effect on dyspnoea and no effect on outcomes. The ACC/AHA guidelines recommend the use of IV nitrates based on the same references, at level II b/A.2 These recommendations for use are made, based on a total of less than 1000 patients prospectively randomized in all the nitrates and furosemide studies combined, and despite evidence from other human studies suggesting potential harm.6, 7 The ACC/AHA guideline also quotes a small study showing rapid tachyphylaxis in response to IV nitrates over 24 h.8 Many large, retrospective analyses of the use of IV loop diuretics in AHF show neurohormonal activation and possible renal dysfunction,6 the latter was confirmed in the DOSE study (see below). A plethora of studies suggest that the administration of IV inotropes such as dobutamine, milrinone and levosimendan is potentially associated with harm and, in some studies, increased mortality.9 To address our lack of knowledge of the efficacy and safety of currently used and newly proposed therapies for patients with AHF, two studies were recently published: the DOSE and Renal Optimization Strategies Evaluation (ROSE) studies.10, 11 In the DOSE study, two aspects of IV loop diuretic therapy, administered for up to 72 h, were compared in 308 patients using a factorial design. The first aspect compared continuous infusion with bolus administration and the second compared a higher and lower dose. Results of DOSE suggest that higher doses of IV loop diuretics induce more diuresis and decongestion associated with more dyspnoea relief but also more creatinine elevations compared with lower doses while continuous and bolus administration do not differ. In the ROSE study, 360 patients were randomized to low-dose IV nesiritide, low-dose IV dopamine, or placebo; all were treated with IV loop diuretic. Results of ROSE suggest that the addition of low-dose nesiritide or low-dose dopamine did not differ from the addition of placebo to IV loop diuretic therapy. Unfortunately, the results of these studies add little to our knowledge of the effects of IV therapies in AHF. First, the primary efficacy endpoint of ROSE and the main effect observed in DOSE is diuresis. However, diuresis is not a direct measure of clinical benefit and, as shown in the large Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study with Tolvaptan (EVEREST) study, diuresis and weight loss do not necessarily translate into beneficial effects on either patients' symptoms or clinical outcomes.12 Second, in the DOSE study the ‘low’ and ‘high’ IV loop diuretic doses studied were substantially higher than doses reported in AHF registries and clinical studies. The ‘low-dose’ arm received a median of 358 mg of IV furosemide over 72 h. In the Acute Heart Failure National Registry (ADHERE) the vast majority of patients received <160 mg of IV furosemide equivalents,13 the median dose administered over 7 days was 280 mg in the Placebo Controlled Study of the Selective A1 Adenosine Receptor Antagonist Rolofylline (PROTECT) trial14 and a mean of 187 mg was administered over 5 days in the Relaxin in Acute Heart Failure (RELAX-AHF) study,15, (both studies of patients with AHF and renal impairment). Hence, the comparison of high-dose IV loop diuretics with very high doses in the DOSE study, which may have been conducted in diuretic-dependent patients with significant HF, does not elucidate the role of usual doses of IV loop diuretics in the general population of patients with AHF. Most importantly, both the DOSE and ROSE studies were small and thus not powered to detect meaningful effects on clinically relevant endpoints. The studies' small size and the variability of the endpoints has resulted in a low ‘signal to noise ratio’; in other words, the true effects of each strategy and therapy could vary between significant benefit to significant harm. The ‘true effects’ on dyspnoea, worsening of heart failure, renal impairment, readmission, and death could, within the confidence intervals of the treatment effects, range from a significant relative benefit of 30–50% to 30–50% harm despite the mostly neutral observations in the study. Thus, these studies do not rule out huge benefit or harm to the patients on any clinically meaningful measure of effect by high-dose furosemide, low-dose nesiritide or low-dose dopamine. In Table 1 we provide estimated power for a hypothetical study with two 100-patient arms or 150-patient arms to detect small, medium or large effects on symptom (dyspnoea) improvement, worsening heart failure, readmission or death, death, and changes in renal function based on comparative effects seen in other chronic and AHF studies where clinical effects were detected such as RELAX-AHF,15, Randomized Evaluation of Intravenous Levosimendan Efficacy ( REVIVE),16 VMAC,4 and other HF studies17, 18 and accepted clinically meaningful effect sizes. Clearly, studies the size of DOSE and ROSE are not powered to exclude very significant positive or negative effects on any of these endpoints. Therefore, studies of this size should be regarded as hypothesis-generating and not definitive ones on which robust recommendations for administration of (or withholding) therapy should be made. In conclusion, despite a generation of clinical research we continue to treat patients with AHF with IV therapies during the first days of admission based on little or no evidence. Recent attempts to improve our knowledge base by examining the effects of such therapies in small, underpowered studies only add to this lack of certainty by reporting mostly equivalent results within wide confidence intervals. Hence, our clinical practice continues to be uninformed and we may very well be undertreating patients by denying them effective therapies simply because we do not know they if are effective or administering therapies that cause harm because we do not know they do cause harm. It is incumbent on us to undertake well-designed and adequately powered studies in which the true effects of currently used IV therapies for AHF are examined, making the use of beneficial therapies truly evidence-based. Conflict of interest: G.C. and B.C.D. are employees of Momentum Research Inc., which has received remuneration for consulting on clinical studies from Novartis, Corthera, Travena, Cardio3 Biosciences, Sorbent Therapeutics, Amgen, Merck, Anexon.