Heart 98:1761-1762 doi:10.1136/heartjnl-2012-302854
  • Editorials

Hyponatremia and outcomes in patients with heart failure

  1. Steven R Goldsmith
  1. Correspondence to Dr Steven R Goldsmith, Department of Cardiology, Hennepin County Medical Center, 701 Park Avenue, Minneapolis, MN 55415, USA;{at}

Studies spanning several decades have identified hyponatremia as common and significantly associated with poor outcomes in patients with acutely decompensated heart failure (ADHF). Despite the development and wide application of neurohormonal inhibitory therapy, which might have been expected to reduce the incidence of hyponatremia, recent studies from both randomised controlled trials and registries continue to identify hyponatremia in up to 27% of patients with ADHF. Also, despite the use of contemporary therapies, hyponatremia remains independently and closely correlated with poor outcomes, including death.1 A key question which is unresolved is whether hyponatremia contributes directly to poor outcomes in ADHF or is simply a marker for disease severity or, possibly, for other factors which might influence disease progression.

Hyponatremia could, in theory, directly influence morbidity and mortality in HF. The brain undergoes adaptation to a slow fall in serum osmolality by losing osmotically active substances, including various neurotransmitters. This loss may lead to adverse effects on neurocognitive function.2 Whether myocytes, particularly dysfunctional myocytes, are affected by hypotonicity has never been studied. It is certainly possible that either mechanical or electrical function could be adversely affected.

Hyponatremia in heart failure is largely due to the inappropriate secretion of the antidiuretic hormone arginine vasopressin (AVP). Free water reabsorption is mediated by the renal V2 receptor for AVP. However, the V1a effects of AVP include vasoconstriction and myocardial hypertrophy mediated by signalling pathways similar to those stimulated by angiotensin II. In theory, high AVP levels could adversely affect outcome in heart failure via V1a signalling.3 If this were true, hyponatremia may be a marker for high AVP levels, which in turn could affect outcome via excessive V1a stimulation.

Hyponatremia is often exacerbated by loop diuretic therapy, and clinicians often reduce diuretics when hyponatremia occurs, potentially leaving patients in a congested state. This scenario is undesirable since persistent congestion confers very poor outcome in HF.4 If this were common, the adverse effect of hyponatremia on outcome could be explained by the impact of hyponatremia on needed therapy for congestion. And finally, until recently, other treatments available for hyponatremia in HF have been either largely ineffective, such as water restriction, or potentially toxic, such as demeclocylcine. The lack of safe and effective treatment for hyponatremia has significantly hampered investigation into the nature of the relationship between this condition and outcomes in HF.

While many studies have examined the prognostic importance of an isolated measurement of serum sodium, few have attempted to assess the importance of a change in serum sodium during treatment. The study by Lee et al 5 adds significantly to our knowledge in this area. This study was a registry of 2888 Korean patients hospitalised with ADHF and followed for an average of 1.7 years. The investigators assessed the outcomes in patients stratified by whether they had normonatremia or hyponatremia at the time of hospitalisation, and on the basis of persistent or improved hyponatremia when reassessed prior to discharge. The key findings were that while baseline hyponatremia conferred a very poor prognosis, outcomes were not significantly altered by whether or not serum sodium was normal when assessed later on. Similar observations had been made in analyses from previous smaller studies, such as ESCAPE and OPTIME-HF.6 ,7 However, studies by Madan et al, cited in the article by Lee, and Rossi et al in a report from the ACTIV study, not cited in the current report8, have shown that improved serum sodium on followup did in fact correlate with improved survival.

The major strengths of the current study are its size, duration of follow-up and sophisticated statistical analysis. Its major weaknesses include the fact that it is a registry, not a prospective randomised trial, and that the assessment of the change in serum sodium was made only once, prior to or at discharge from the hospital. The authors do note, that in the study by Madan et al9 in which there was a positive correlation with outcome and improved serum sodium, that the followup time was considerably longer.9 Also, the aetiology of heart failure is not discussed in the present report, and the fact that the average ejection fraction was 39% means that the population was somewhat skewed to either patients having valvular heart disease, or to those with the syndrome of ‘HFPEF’ (Heart Failure with Preserved Ejection Fraction). Most prior investigations into hyponatremia and outcomes in heart failure have focused on patients with systolic heart failure, that is, those with a low ejection fraction. And finally, the study provides no information as to whether there was any attempt to treat hyponatremia in these patients, and if so, by what means. Treatment would not have included V2 antagonists since the study was conducted prior to the introduction of these agents.

This last point is important, since as noted above, treatment for hyponatremia prior to the introduction of V2 antagonists has been either ineffective, or when effective, potentially counterproductive. One of the studies which did show improved outcome with correction of serum sodium utilized a V2 antagonist although the patients in that trial were not randomized on the basis of serum sodium8. The lack of impact on outcomes in ADHF when hyponatremia normalised in the Korean registry could certainly mean that there truly was no outcomes benefit attributable to an improved serum sodium. In the milieu in which hyponatremia occurs in patients with heart failure, treatment of this electrolyte abnormality by any means, conceivably, would not produce benefit. These patients often have serious cardiac dysfunction and are prone to arrhythmias, as well as significant problems with lung and kidney function. It may be naïve to think that simply raising the serum sodium would in and of itself lead to benefit. On the other hand, as noted, it is possible that the impact of the treatments for hyponatremia in this registry, assuming attempts at treatment were made, had offsetting negative effects on the heart failure syndrome which mitigated any potential benefit from correcting hyponatremia per se. One would need considerably more information about whether or not treatment for hyponatremia was employed and if so, what treatment, to even begin to speculate about these matters. It is also possible, as already noted, that a longer time period of observation and/or repeated measurements of serum sodium would be needed to more completely understand the relationship between correcting hyponatremia and outcomes.

Therefore, hyponatremia may simply be a marker for disease severity, or a marker for other factors which could affect outcome in patients with ADHF. But the relationship between hyponatremia and outcome is highly significant, and consistently so. The largest single reported experience cites an essentially linear relationship between serum sodium and mortality between serum sodium of 125 and 138 meq/l.1 One should beware of simply assuming that such a tight correlation simply reflects association.

It would be tempting to conclude from the report by Lee et al 5, that hyponatremia is not a reasonable target of therapy in the patient with ADHS. In my opinion, the documentation in yet another large series of ADHF patients on modern background therapy that hyponatremia confers an extremely poor prognosis should be taken very seriously. The fact that a normalised serum sodium was not associated with better outcomes using the available methods during the time of the study means to me, that if we want to learn more about the relationship between hyponatremia and outcomes, we need to perform a prospective randomised trial using safe, effective and physiologically rational treatment with appropriate follow-up assessments. In the USA, both conivaptan, an intravenous, balanced V1a/V2 antagonist, and tolvaptan, an orally effective selective V2 antagonist, have been approved by the Federal Food and Drug Administration for the treatment of euvolemic and hypervolemic hyponatremia, including the hyponatremia seen in patients with HF.10 ,11 Either or both of these agents could be utilised in a study designed to test the hypothesis that correction of hyponatremia would improve outcomes in patients with HF. Both agents directly address the cause of the problem, that is, excessive V2 signalling leading to excessive free water reabsorption. Both agents produce an incremental diuresis when added to loop diuretics, an additional benefit if patients are hypervolemic at the time hyponatremia develops, which is usually the case. Also, both agents have been established to be safe in patients with heart failure, both with and without hyponatremia. With tolvaptan, the agent most useful for a chronic study, the long-term safety profile in patients with HF is excellent, as shown in the hyponatremic subset of the EVEREST trial.12 Indeed, a retrospective analysis of these patients in EVEREST has suggested improvement in the combined endpoint of cardiovascular morbidity and mortality in patients with serum sodium under 130 when treated with tolvaptan (Paul Hauptman, personal communication, data submitted for publication).

In sum, hyponatremia remains a significant, often challenging clinical problem in patients with ADHF. Hyponatremia unequivocally is associated with poor outcomes in these patients. The report by Lee et al does not offer hope that a prospective trial of treating hyponatremia by methods other than V2 antagonism would be of much use, assuming that attempts to treat hyponatremia in this registry were made. However, now that safe and effective therapy based on V2 antagonism is available, I believe that a properly powered multicenter study of the treatment of hyponatremia in HF is in order. The recent studies by Madan et al 9 and Rossi et al 8 which did link an improvement in serum sodium to improved survival, further support the rationale for such a study. The availability of V2 antagonists certainly makes hyponatremia a much more easily treatable target in patients with ADHF, and I do not believe we are justified in assuming, without proof, that the relationship between hyponatremia and poor outcomes is just an association.


  • Competing interests The author has received consulting income, research support and speaking fees from both Astellas and Otsuka, which manufacture Conivaptan and Tolvaptan respectively.

  • Provenance and peer review Commissioned; internally peer reviewed.