Physiologic Rationale for Use of AVP Antagonists

AVP antagonists block the V₂ receptor located on the basolateral surface of the collecting duct thereby antagonizing the ability of AVP to cause insertion of the aquaporin‐2 water channels into the luminal membrane. The increase in urine output is similar in quantity to diuretics but differs in content. V₂ receptor antagonists increase water excretion with little to no change in urinary electrolytes. As a result, lowering of the serum K⁺ level, metabolic alkalosis, and increases in the serum creatinine and blood urea nitrogen concentration are avoided in contrast to diuretics such as furosemide and hydrochlorothiazide. In addition, orthostatic hypotension and activation of neurohumoral effectors such as angiotensin II, circulating catecholamines, and aldosterone are not features of V₂ receptor blockade. These differences have lead to V₂ receptor antagonists being characterized as aquaretic agents so as to distinguish them from diuretics.

The physiologic rationale for use of V₂ receptor antagonists is best exemplified by considering the relationship between the serum Na⁺ concentration and the total body content of Na⁺, K⁺, and water approximated by the equation: 2 receptor antagonist) is approved by the Food and Drug Administration (FDA) for euvolemic and hypervolemic hyponatremia. The effectiveness of the drug was demonstrated in a randomized placebo controlled trial of 84 hospitalized patients with euvolemic or hypervolemic hyponatremia in which the serum Na⁺ ranged from 115 to <130 mEq/L.5 The patients were given a 20‐mg loading dose followed by a continuous infusion of either 40 or 80 mg daily for 4 days. Conivaptan significantly raised the serum Na⁺ concentration by 6.3 and 9.4 in the 40 and 80 mg/day arms respectively as compared to 0.8 mEq/L in the placebo group.

Conivaptan is also active as an oral formulation but its distribution has been restricted to parenteral use for short‐term (4 days maximum) in‐hospital administration only. This restriction is due to potent inhibitory effects of the drug on the hepatic cytochrome P₄₅₀ 3A4 enzyme system and the potential for untoward drug interactions. The inhibitory effects of other members of this class on the cytochrome P450 3A4 (CYP3A4) system are more limited, allowing for oral formulations to be used in a clinical setting.

Tolvaptan is the only other AVP antagonist currently approved by the FDA. Unlike conivaptan, tolvaptan is an oral agent with effects confined to the V₂ receptor. The drug is indicated for the treatment of clinically significant hypervolemic and euvolemic hyponatremia.

The efficacy of tolvaptan was evaluated in 2 simultaneously conducted multicenter, randomized, double blind trials called Study of Ascending Levels of Tolvaptan in hyponatremia 1 and 2 (SALT 1 and SALT 2).6 The 2 studies together randomized 225 patients with euvolemic or hypervolemic hyponatremia to outpatient treatment with the study drug for 30 days and 223 patients to placebo. Patients with a serum Na⁺ of <120 mEq/L in association with neurologic impairment were excluded from the trial. Serum Na⁺ increased more in the tolvaptan group than in the placebo during the first 4 days (3.62 vs. 0.25 mEq/L). After 30 days of therapy serum Na⁺ concentrations were 6.22 higher in the tolvaptan group compared with 1.66 mEq/L in the placebo group. When tolvaptan was discontinued at study end, the serum Na⁺ fell over a seven day period to a value similar to that in the placebo treated patients. Four patients (1.8%) exceeded the study goal of limiting correction of the hyponatremia to <12 mEq/L in the first 24 hours of treatment, but none of these patients developed adverse clinical sequelae.

As previously mentioned conivaptan and tolvaptan are indicated for the treatment of hyponatremia in the setting of euvolemia or hypervolemia. These drugs should not be used in hypovolemic states since the increase in renal water excretion can potentially predispose to worsening hemodynamics in the setting of volume depletion. To be sure, any volume of water removed from the body is principally derived from the intracellular compartment (two‐thirds) and would not be expected to affect blood pressure to a major extent. However, one‐twelfth of any water volume loss is derived from the circulating compartment and could potentially aggravate a borderline low blood pressure in the setting of volume depletion.

This concern is particularly true for conivaptan and its blocking effects on the V_1a receptor. AVP can cause peripheral vasoconstriction by stimulating the V_1a receptor on the peripheral vasculature.7 However, circulating concentrations observed in euvolemic and hypervolemic conditions are not typically of a magnitude to elicit this effect thus explaining the lack of clinically significant hypotension in clinical trials. By contrast, AVP may reach a high enough concentration and play a contributory role in the maintenance of blood pressure under conditions of significant depletion of extracellular fluid volume (hemorrhage) or in states of generalized vasodilation such as sepsis or advanced cirrhosis. In these settings blockade of the V_1a receptor may result in significant hypotension.

Another theoretical concern of blocking the V_1a receptor with conivaptan is the potential to cause further sequestration of fluid in the splanchnic vascular bed and theoretically increase the risk of hepatorenal syndrome.8 The importance of splanchnic vasodilation in the genesis of renal hypoperfusion has been indirectly illustrated by the response to ornipressin, an analog of AVP that is a preferential splanchnic vasoconstrictor. The administration of ornipressin to patients with advanced cirrhosis leads to correction of many of the systemic and renal hemodynamic abnormalities that are present. These include an elevation in mean arterial pressure, reductions in plasma renin activity and norepinephrine concentration, and increases in renal blood flow, glomerular filtration rate, and urinary sodium excretion and volume. V_1a receptor blockade has the potential to increase the degree of arterial vasodilation in the splanchnic arteriolar bed. Increasing degrees of splanchnic vasodilation contribute to a fall in mean arterial pressure and unloading of baroreceptors in the central circulation. As a result, central afferent sensors signal the activation of neurohumoral effectors which in turn decrease perfusion of other organs, particularly the kidney.

AVP Antagonists in Severe Symptomatic Hyponatremia

Symptoms of hyponatremia include nausea and malaise, which can be followed by headache, lethargy, muscle cramps, restlessness, confusion and disorientation.9 Life threatening symptoms are those of impending brain herniation and include seizures, decreased levels consciousness, and obtundation. Hypertonic saline remains the treatment of choice in those patients who are clinically determined to be severely symptomatic.10 Vaptan therapy does result in a brisk and relatively prolonged water diuresis. Indeed, in the first 6 hours following the parenteral administration of conivaptan there is a several mEq/L increase in the serum Na⁺ concentration.5 Whether this time course of correction is sufficient to abort fatal hyponatremic encephalopathy is simply not known. Such patients to date have been excluded from clinical studies since randomization in a placebo‐controlled trial would be clearly unethical in such subjects, and not using hypertonic saline might also be considered unsafe

Although not well studied, one could coadminister conivaptan along with hypertonic saline and anticipate a more rapid initial rise in the serum Na⁺. Once stabilized, the hypertonic saline could be discontinued and the remainder of correction be accomplished by use of the receptor antagonist alone. This strategy would help to minimize the likelihood of volume overload due to the use of hypertonic saline but would require frequent monitoring of the serum Na⁺ to ensure the rate of correction was within accepted guidelines.

AVP Antagonists in Patients With Mild to Moderate Symptoms of Hyponatremia

Hospitalized patients with mild to moderate symptoms of hyponatremia can be considered ideal candidates for the use of V₂ receptor antagonists. Both conivaptan and tolvaptan can be expected to increase the serum Na⁺ concentration to a greater extent and more predictably than fluid restriction alone. The superiority of these drugs would be particularly evident in those who require ongoing fluid administration for any number of reasons such as parenteral administration of antibiotics or proton pump inhibitors.11

In hyponatremic patients with decompensated heart failure there may be changes in mental status in which it is difficult to separate out the contribution of hyponatremia from the decrease in cerebral perfusion. Vaptan therapy offers a predictable way to improve the serum Na⁺ contribution over an acceptable period of time thereby removing one of the variables. Even if hyponatremia is not the direct cause of symptoms, it may lower the threshold for mental status changes resulting from poor cerebral perfusion. A similar argument for vaptan therapy can be made in hyponatremic patients with cirrhosis. In this population mental status changes can be due to the hyponatremia, the underlying liver disease, or both. If no improvement occurred following the correction of the hyponatremia, therapy could then be focused on the underlying liver disease. In fact, in any delirious patient with hyponatremia, correction of the underlying electrolyte disorder can help to clarify the degree to which the hyponatremia is exacerbating the altered mental status.

There are no head to head studies comparing conivaptan and tolvaptan in patients with mild to moderate symptoms of hyponatremia. Presumably both drugs would work with similar efficacy. Conivaptan therapy can be complicated by phlebitis when administered through a peripheral vein due to the irritative effects of polypropylene glycol which serves as a diluent for the drug.12 For this reason the drug is often given through a central vein. The theoretical risk of conivaptan's V_1a receptor blocking effects in patients with cirrhosis has already been discussed. Tolvaptan is given orally starting at 15 mg daily. Depending on the response the dose can be increased to 30 mg and ultimately to 60 mg daily. Few patients in the SALT 1 and 2 trials required the 60 mg dose.

Measurement of the urine osmolality may be useful in predicting how responsive a patient will be to the administration of a vaptan. In this regard, urine osmolality can be thought as a biomarker of the effect of AVP on the renal collecting duct.13 The higher the urine osmolality, presumably the higher the serum AVP level or the greater the effect of AVP is on the tubule. By contrast, those with a lower urine osmolality presumably have less AVP effect on the tubule and therefore will manifest a less robust response to a V₂ receptor blocker.

There are circumstances unique to the hospitalized patient where one may consider use of a V₂ receptor antagonist. Consider a patient admitted with pneumonia complicated by hyponatremia (Na⁺ = 128 mEq/L) due to syndrome of inappropriate antidiuretic hormone secretion (SIADH) who is treated with intravenous antibiotics, stabilized, and is now being considered for discharge. Despite fluid restriction the serum Na⁺ concentration is currently 126 mEq/L and the admitting physician is reluctant to discharge the patient for fear the Na⁺ may fall further and result in symptomatic hyponatremia. The hospital discharge is postponed and fluid restriction is intensified. Two days later the serum Na⁺ concentration has risen to 131 mEq/L and the patient is discharged. While one can question the wisdom of postponing the original planned discharge based on the lab value alone, this type of scenario is common in clinical practice. Fluid restriction is poorly tolerated, difficult to enforce, and often unpredictable in response. The administration of a single dose of a V₂ receptor antagonist is a reliable way to increase the serum Na⁺ concentration over a 24 hour period and in this case could have potentially shortened the hospital stay.

Consider another patient admitted with a hip fracture who is found to have a serum Na⁺ concentration of 126 meq/L attributed to ongoing use of an antidepressant agent. The patient is otherwise medically stable but the anesthesiologist is unwilling to accept the patient for surgery until the serum Na⁺ is at least >130 mEq/L. Fluid restriction is prescribed and only 2 to 3 days later has the serum Na⁺ reached a value of 131 mEq/L. Once again, administration of a vaptan would have almost certainly increased the Na⁺ concentration to the desired threshold over a 24 hour period, allowing the patient to undergo the surgical procedure without experiencing the undue delay.

In each of these scenarios short term use of a vaptan would provide for a more rapid and predictable increase in the serum Na⁺ concentration and potentially decrease the length of hospital stay in comparison to fluid restriction alone. In addition to being less predictable and often poorly tolerated, fluid restriction is often ineffective due to the obligatory fluid administration hospitalized patients require for other therapies or nutrition. Demeclocycline and lithium have been used to antagonize the effects of AVP on the tubule but these drugs take several days before any demonstrable increase in renal water excretion is seen. Demeclocycline antagonizes the effect of AVP through nephrotoxic effects on the tubular cell whereas vaptan's are competitive inhibitors of the V₂ receptor and are not associated with nephrotoxicity. Lithium interferes in the intracellular signaling pathways by which AVP causes insertion of water channels into the apical membrane. The doses required to illicit an increase in renal water excretion are near those which can result in lithium levels sufficient to cause neurotoxicity. The primary side effect of vaptan therapy are those one would predict from inducing an aquaretic effect and include thirst, increased urinary frequency, and increased urinary volume.

The decision to use a vaptan on a more prolonged basis is made on a case by case basis. When the underlying cause of increased AVP is deemed to be chronic and irreversible then therapy can be extended into the outpatient setting. Consideration for more chronic therapy would be appropriate for patients with SIADH due to underlying cancer, or those with severe chronic congestive heart failure or advanced cirrhosis. If the cause of the increased AVP is transient, then 1 to 2 doses of a vaptan while in the hospital may be all that is required. Transient causes of increased AVP would include drug related causes, acute pneumonia, hypoxia associated with respiratory failure, and acute decompensated heart failure. Once again vaptan therapy should not be used when the cause of increased AVP is due to total body salt depletion.

AVP Antagonists in Asymptomatic Hyponatremic Patients

One remaining question concerning the use of vaptan therapy is whether they are helpful in hyponatremic patients who are asymptomatic. There are several observations which raise the possibility that patients deemed to be otherwise asymptomatic do in fact have subtle abnormalities attributable to the low Na⁺ concentration and therefore could benefit if the hyponatremia was corrected. First, some patients with a serum Na⁺ in the range of 120 mEq/L to 129 mEq/L have subtle neurologic changes that improve when the serum Na⁺ concentration is increased.14 These include scores on tests of mental and social functioning. In the SALT 1 and 2 trials a general health survey filled out by the patients showed tolvaptan therapy was associated with improvements in the mental health component of the instrument assessing parameters such as vitality, social functioning, emotionally limited accomplishments, calmness, and sadness.6 Response time is delayed and the error rate is increased in response to various stimuli while patients are hyponatremic, suggesting hyponatremia is associated with reversible impairment in attention. It is certainly reasonable to speculate that correction of hyponatremia in an elderly patient with SIADH might improve that individual's quality of life by allowing them to better enjoy any number of activities such as reading or completing a crossword puzzle. The presence of concomitant neuropsychiatric disturbance (delirium and dementia in particular) may also be relevant, since even mild hyponatremia may increase vulnerability to further alterations in mental status in patients who are impaired at baseline.

Second, asymptomatic hyponatremic patients exhibit subtle disturbances in gait that improve following correction of the serum Na⁺ concentration.14, 15 In this regard, case control studies have shown an association between hyponatremia and risk of falls and fractures particularly in the elderly population.15, 16 Chronic hyponatremia has been shown to cause a reduction in bone mass in an experimental model of SIADH and hyponatremia is associated with osteoporosis in cross‐sectional human data.17 Fall‐related injury is associated with substantial adverse psychological and physical outcomes and is a cause of both death and disability in this population. It would be of considerable socioeconomic benefit if treatment of asymptomatic hyponatremia were able to reduce the risk of this complication.

Lastly, hyponatremia is associated with increased morbidity and mortality but generally the low Na⁺ concentration is merely thought to be marker of the severity of the underlying disease and not a direct contributor to the adverse outcome.18, 19 For example in decompensated heart failure there is an inverse relationship between the degree of hyponatremia and the extent of neurohormonal activation. Adverse effects resulting from persistent activation of the renin‐angiotensin‐aldosterone system and sympathetic nerve activity are thought to be the mechanism underlying the increase in mortality in hyponatremic patients with heart failure. Nevertheless, it has been suggested that hyponatremia itself, the associated hypotonicity, and/or elevated levels of AVP might exert adverse effects on the cardiovascular system or other organ systems and therefore play a contributory role in patient morbidity and mortality. While it remains speculative as to whether correction of the hyponatremia per se will improve patient outcomes, the V₂ receptor antagonists offer an opportunity to test this uncertainty in patients with euvolemic and hypervolemic hyponatremia.0, 0

Physiologic Rationale for Use of AVP Antagonists

AVP antagonists block the V₂ receptor located on the basolateral surface of the collecting duct thereby antagonizing the ability of AVP to cause insertion of the aquaporin‐2 water channels into the luminal membrane. The increase in urine output is similar in quantity to diuretics but differs in content. V₂ receptor antagonists increase water excretion with little to no change in urinary electrolytes. As a result, lowering of the serum K⁺ level, metabolic alkalosis, and increases in the serum creatinine and blood urea nitrogen concentration are avoided in contrast to diuretics such as furosemide and hydrochlorothiazide. In addition, orthostatic hypotension and activation of neurohumoral effectors such as angiotensin II, circulating catecholamines, and aldosterone are not features of V₂ receptor blockade. These differences have lead to V₂ receptor antagonists being characterized as aquaretic agents so as to distinguish them from diuretics.

The physiologic rationale for use of V₂ receptor antagonists is best exemplified by considering the relationship between the serum Na⁺ concentration and the total body content of Na⁺, K⁺, and water approximated by the equation: 2 receptor antagonist) is approved by the Food and Drug Administration (FDA) for euvolemic and hypervolemic hyponatremia. The effectiveness of the drug was demonstrated in a randomized placebo controlled trial of 84 hospitalized patients with euvolemic or hypervolemic hyponatremia in which the serum Na⁺ ranged from 115 to <130 mEq/L.5 The patients were given a 20‐mg loading dose followed by a continuous infusion of either 40 or 80 mg daily for 4 days. Conivaptan significantly raised the serum Na⁺ concentration by 6.3 and 9.4 in the 40 and 80 mg/day arms respectively as compared to 0.8 mEq/L in the placebo group.

Conivaptan is also active as an oral formulation but its distribution has been restricted to parenteral use for short‐term (4 days maximum) in‐hospital administration only. This restriction is due to potent inhibitory effects of the drug on the hepatic cytochrome P₄₅₀ 3A4 enzyme system and the potential for untoward drug interactions. The inhibitory effects of other members of this class on the cytochrome P450 3A4 (CYP3A4) system are more limited, allowing for oral formulations to be used in a clinical setting.

Tolvaptan is the only other AVP antagonist currently approved by the FDA. Unlike conivaptan, tolvaptan is an oral agent with effects confined to the V₂ receptor. The drug is indicated for the treatment of clinically significant hypervolemic and euvolemic hyponatremia.

The efficacy of tolvaptan was evaluated in 2 simultaneously conducted multicenter, randomized, double blind trials called Study of Ascending Levels of Tolvaptan in hyponatremia 1 and 2 (SALT 1 and SALT 2).6 The 2 studies together randomized 225 patients with euvolemic or hypervolemic hyponatremia to outpatient treatment with the study drug for 30 days and 223 patients to placebo. Patients with a serum Na⁺ of <120 mEq/L in association with neurologic impairment were excluded from the trial. Serum Na⁺ increased more in the tolvaptan group than in the placebo during the first 4 days (3.62 vs. 0.25 mEq/L). After 30 days of therapy serum Na⁺ concentrations were 6.22 higher in the tolvaptan group compared with 1.66 mEq/L in the placebo group. When tolvaptan was discontinued at study end, the serum Na⁺ fell over a seven day period to a value similar to that in the placebo treated patients. Four patients (1.8%) exceeded the study goal of limiting correction of the hyponatremia to <12 mEq/L in the first 24 hours of treatment, but none of these patients developed adverse clinical sequelae.

As previously mentioned conivaptan and tolvaptan are indicated for the treatment of hyponatremia in the setting of euvolemia or hypervolemia. These drugs should not be used in hypovolemic states since the increase in renal water excretion can potentially predispose to worsening hemodynamics in the setting of volume depletion. To be sure, any volume of water removed from the body is principally derived from the intracellular compartment (two‐thirds) and would not be expected to affect blood pressure to a major extent. However, one‐twelfth of any water volume loss is derived from the circulating compartment and could potentially aggravate a borderline low blood pressure in the setting of volume depletion.

This concern is particularly true for conivaptan and its blocking effects on the V_1a receptor. AVP can cause peripheral vasoconstriction by stimulating the V_1a receptor on the peripheral vasculature.7 However, circulating concentrations observed in euvolemic and hypervolemic conditions are not typically of a magnitude to elicit this effect thus explaining the lack of clinically significant hypotension in clinical trials. By contrast, AVP may reach a high enough concentration and play a contributory role in the maintenance of blood pressure under conditions of significant depletion of extracellular fluid volume (hemorrhage) or in states of generalized vasodilation such as sepsis or advanced cirrhosis. In these settings blockade of the V_1a receptor may result in significant hypotension.

Another theoretical concern of blocking the V_1a receptor with conivaptan is the potential to cause further sequestration of fluid in the splanchnic vascular bed and theoretically increase the risk of hepatorenal syndrome.8 The importance of splanchnic vasodilation in the genesis of renal hypoperfusion has been indirectly illustrated by the response to ornipressin, an analog of AVP that is a preferential splanchnic vasoconstrictor. The administration of ornipressin to patients with advanced cirrhosis leads to correction of many of the systemic and renal hemodynamic abnormalities that are present. These include an elevation in mean arterial pressure, reductions in plasma renin activity and norepinephrine concentration, and increases in renal blood flow, glomerular filtration rate, and urinary sodium excretion and volume. V_1a receptor blockade has the potential to increase the degree of arterial vasodilation in the splanchnic arteriolar bed. Increasing degrees of splanchnic vasodilation contribute to a fall in mean arterial pressure and unloading of baroreceptors in the central circulation. As a result, central afferent sensors signal the activation of neurohumoral effectors which in turn decrease perfusion of other organs, particularly the kidney.

AVP Antagonists in Severe Symptomatic Hyponatremia

Symptoms of hyponatremia include nausea and malaise, which can be followed by headache, lethargy, muscle cramps, restlessness, confusion and disorientation.9 Life threatening symptoms are those of impending brain herniation and include seizures, decreased levels consciousness, and obtundation. Hypertonic saline remains the treatment of choice in those patients who are clinically determined to be severely symptomatic.10 Vaptan therapy does result in a brisk and relatively prolonged water diuresis. Indeed, in the first 6 hours following the parenteral administration of conivaptan there is a several mEq/L increase in the serum Na⁺ concentration.5 Whether this time course of correction is sufficient to abort fatal hyponatremic encephalopathy is simply not known. Such patients to date have been excluded from clinical studies since randomization in a placebo‐controlled trial would be clearly unethical in such subjects, and not using hypertonic saline might also be considered unsafe

Although not well studied, one could coadminister conivaptan along with hypertonic saline and anticipate a more rapid initial rise in the serum Na⁺. Once stabilized, the hypertonic saline could be discontinued and the remainder of correction be accomplished by use of the receptor antagonist alone. This strategy would help to minimize the likelihood of volume overload due to the use of hypertonic saline but would require frequent monitoring of the serum Na⁺ to ensure the rate of correction was within accepted guidelines.

AVP Antagonists in Patients With Mild to Moderate Symptoms of Hyponatremia

Hospitalized patients with mild to moderate symptoms of hyponatremia can be considered ideal candidates for the use of V₂ receptor antagonists. Both conivaptan and tolvaptan can be expected to increase the serum Na⁺ concentration to a greater extent and more predictably than fluid restriction alone. The superiority of these drugs would be particularly evident in those who require ongoing fluid administration for any number of reasons such as parenteral administration of antibiotics or proton pump inhibitors.11

In hyponatremic patients with decompensated heart failure there may be changes in mental status in which it is difficult to separate out the contribution of hyponatremia from the decrease in cerebral perfusion. Vaptan therapy offers a predictable way to improve the serum Na⁺ contribution over an acceptable period of time thereby removing one of the variables. Even if hyponatremia is not the direct cause of symptoms, it may lower the threshold for mental status changes resulting from poor cerebral perfusion. A similar argument for vaptan therapy can be made in hyponatremic patients with cirrhosis. In this population mental status changes can be due to the hyponatremia, the underlying liver disease, or both. If no improvement occurred following the correction of the hyponatremia, therapy could then be focused on the underlying liver disease. In fact, in any delirious patient with hyponatremia, correction of the underlying electrolyte disorder can help to clarify the degree to which the hyponatremia is exacerbating the altered mental status.

There are no head to head studies comparing conivaptan and tolvaptan in patients with mild to moderate symptoms of hyponatremia. Presumably both drugs would work with similar efficacy. Conivaptan therapy can be complicated by phlebitis when administered through a peripheral vein due to the irritative effects of polypropylene glycol which serves as a diluent for the drug.12 For this reason the drug is often given through a central vein. The theoretical risk of conivaptan's V_1a receptor blocking effects in patients with cirrhosis has already been discussed. Tolvaptan is given orally starting at 15 mg daily. Depending on the response the dose can be increased to 30 mg and ultimately to 60 mg daily. Few patients in the SALT 1 and 2 trials required the 60 mg dose.

Measurement of the urine osmolality may be useful in predicting how responsive a patient will be to the administration of a vaptan. In this regard, urine osmolality can be thought as a biomarker of the effect of AVP on the renal collecting duct.13 The higher the urine osmolality, presumably the higher the serum AVP level or the greater the effect of AVP is on the tubule. By contrast, those with a lower urine osmolality presumably have less AVP effect on the tubule and therefore will manifest a less robust response to a V₂ receptor blocker.

There are circumstances unique to the hospitalized patient where one may consider use of a V₂ receptor antagonist. Consider a patient admitted with pneumonia complicated by hyponatremia (Na⁺ = 128 mEq/L) due to syndrome of inappropriate antidiuretic hormone secretion (SIADH) who is treated with intravenous antibiotics, stabilized, and is now being considered for discharge. Despite fluid restriction the serum Na⁺ concentration is currently 126 mEq/L and the admitting physician is reluctant to discharge the patient for fear the Na⁺ may fall further and result in symptomatic hyponatremia. The hospital discharge is postponed and fluid restriction is intensified. Two days later the serum Na⁺ concentration has risen to 131 mEq/L and the patient is discharged. While one can question the wisdom of postponing the original planned discharge based on the lab value alone, this type of scenario is common in clinical practice. Fluid restriction is poorly tolerated, difficult to enforce, and often unpredictable in response. The administration of a single dose of a V₂ receptor antagonist is a reliable way to increase the serum Na⁺ concentration over a 24 hour period and in this case could have potentially shortened the hospital stay.

Consider another patient admitted with a hip fracture who is found to have a serum Na⁺ concentration of 126 meq/L attributed to ongoing use of an antidepressant agent. The patient is otherwise medically stable but the anesthesiologist is unwilling to accept the patient for surgery until the serum Na⁺ is at least >130 mEq/L. Fluid restriction is prescribed and only 2 to 3 days later has the serum Na⁺ reached a value of 131 mEq/L. Once again, administration of a vaptan would have almost certainly increased the Na⁺ concentration to the desired threshold over a 24 hour period, allowing the patient to undergo the surgical procedure without experiencing the undue delay.

In each of these scenarios short term use of a vaptan would provide for a more rapid and predictable increase in the serum Na⁺ concentration and potentially decrease the length of hospital stay in comparison to fluid restriction alone. In addition to being less predictable and often poorly tolerated, fluid restriction is often ineffective due to the obligatory fluid administration hospitalized patients require for other therapies or nutrition. Demeclocycline and lithium have been used to antagonize the effects of AVP on the tubule but these drugs take several days before any demonstrable increase in renal water excretion is seen. Demeclocycline antagonizes the effect of AVP through nephrotoxic effects on the tubular cell whereas vaptan's are competitive inhibitors of the V₂ receptor and are not associated with nephrotoxicity. Lithium interferes in the intracellular signaling pathways by which AVP causes insertion of water channels into the apical membrane. The doses required to illicit an increase in renal water excretion are near those which can result in lithium levels sufficient to cause neurotoxicity. The primary side effect of vaptan therapy are those one would predict from inducing an aquaretic effect and include thirst, increased urinary frequency, and increased urinary volume.

The decision to use a vaptan on a more prolonged basis is made on a case by case basis. When the underlying cause of increased AVP is deemed to be chronic and irreversible then therapy can be extended into the outpatient setting. Consideration for more chronic therapy would be appropriate for patients with SIADH due to underlying cancer, or those with severe chronic congestive heart failure or advanced cirrhosis. If the cause of the increased AVP is transient, then 1 to 2 doses of a vaptan while in the hospital may be all that is required. Transient causes of increased AVP would include drug related causes, acute pneumonia, hypoxia associated with respiratory failure, and acute decompensated heart failure. Once again vaptan therapy should not be used when the cause of increased AVP is due to total body salt depletion.

AVP Antagonists in Asymptomatic Hyponatremic Patients

One remaining question concerning the use of vaptan therapy is whether they are helpful in hyponatremic patients who are asymptomatic. There are several observations which raise the possibility that patients deemed to be otherwise asymptomatic do in fact have subtle abnormalities attributable to the low Na⁺ concentration and therefore could benefit if the hyponatremia was corrected. First, some patients with a serum Na⁺ in the range of 120 mEq/L to 129 mEq/L have subtle neurologic changes that improve when the serum Na⁺ concentration is increased.14 These include scores on tests of mental and social functioning. In the SALT 1 and 2 trials a general health survey filled out by the patients showed tolvaptan therapy was associated with improvements in the mental health component of the instrument assessing parameters such as vitality, social functioning, emotionally limited accomplishments, calmness, and sadness.6 Response time is delayed and the error rate is increased in response to various stimuli while patients are hyponatremic, suggesting hyponatremia is associated with reversible impairment in attention. It is certainly reasonable to speculate that correction of hyponatremia in an elderly patient with SIADH might improve that individual's quality of life by allowing them to better enjoy any number of activities such as reading or completing a crossword puzzle. The presence of concomitant neuropsychiatric disturbance (delirium and dementia in particular) may also be relevant, since even mild hyponatremia may increase vulnerability to further alterations in mental status in patients who are impaired at baseline.

Second, asymptomatic hyponatremic patients exhibit subtle disturbances in gait that improve following correction of the serum Na⁺ concentration.14, 15 In this regard, case control studies have shown an association between hyponatremia and risk of falls and fractures particularly in the elderly population.15, 16 Chronic hyponatremia has been shown to cause a reduction in bone mass in an experimental model of SIADH and hyponatremia is associated with osteoporosis in cross‐sectional human data.17 Fall‐related injury is associated with substantial adverse psychological and physical outcomes and is a cause of both death and disability in this population. It would be of considerable socioeconomic benefit if treatment of asymptomatic hyponatremia were able to reduce the risk of this complication.

Lastly, hyponatremia is associated with increased morbidity and mortality but generally the low Na⁺ concentration is merely thought to be marker of the severity of the underlying disease and not a direct contributor to the adverse outcome.18, 19 For example in decompensated heart failure there is an inverse relationship between the degree of hyponatremia and the extent of neurohormonal activation. Adverse effects resulting from persistent activation of the renin‐angiotensin‐aldosterone system and sympathetic nerve activity are thought to be the mechanism underlying the increase in mortality in hyponatremic patients with heart failure. Nevertheless, it has been suggested that hyponatremia itself, the associated hypotonicity, and/or elevated levels of AVP might exert adverse effects on the cardiovascular system or other organ systems and therefore play a contributory role in patient morbidity and mortality. While it remains speculative as to whether correction of the hyponatremia per se will improve patient outcomes, the V₂ receptor antagonists offer an opportunity to test this uncertainty in patients with euvolemic and hypervolemic hyponatremia.0, 0

Physiologic Rationale for Use of AVP Antagonists

AVP antagonists block the V₂ receptor located on the basolateral surface of the collecting duct thereby antagonizing the ability of AVP to cause insertion of the aquaporin‐2 water channels into the luminal membrane. The increase in urine output is similar in quantity to diuretics but differs in content. V₂ receptor antagonists increase water excretion with little to no change in urinary electrolytes. As a result, lowering of the serum K⁺ level, metabolic alkalosis, and increases in the serum creatinine and blood urea nitrogen concentration are avoided in contrast to diuretics such as furosemide and hydrochlorothiazide. In addition, orthostatic hypotension and activation of neurohumoral effectors such as angiotensin II, circulating catecholamines, and aldosterone are not features of V₂ receptor blockade. These differences have lead to V₂ receptor antagonists being characterized as aquaretic agents so as to distinguish them from diuretics.

The physiologic rationale for use of V₂ receptor antagonists is best exemplified by considering the relationship between the serum Na⁺ concentration and the total body content of Na⁺, K⁺, and water approximated by the equation: 2 receptor antagonist) is approved by the Food and Drug Administration (FDA) for euvolemic and hypervolemic hyponatremia. The effectiveness of the drug was demonstrated in a randomized placebo controlled trial of 84 hospitalized patients with euvolemic or hypervolemic hyponatremia in which the serum Na⁺ ranged from 115 to <130 mEq/L.5 The patients were given a 20‐mg loading dose followed by a continuous infusion of either 40 or 80 mg daily for 4 days. Conivaptan significantly raised the serum Na⁺ concentration by 6.3 and 9.4 in the 40 and 80 mg/day arms respectively as compared to 0.8 mEq/L in the placebo group.

Conivaptan is also active as an oral formulation but its distribution has been restricted to parenteral use for short‐term (4 days maximum) in‐hospital administration only. This restriction is due to potent inhibitory effects of the drug on the hepatic cytochrome P₄₅₀ 3A4 enzyme system and the potential for untoward drug interactions. The inhibitory effects of other members of this class on the cytochrome P450 3A4 (CYP3A4) system are more limited, allowing for oral formulations to be used in a clinical setting.

Tolvaptan is the only other AVP antagonist currently approved by the FDA. Unlike conivaptan, tolvaptan is an oral agent with effects confined to the V₂ receptor. The drug is indicated for the treatment of clinically significant hypervolemic and euvolemic hyponatremia.

The efficacy of tolvaptan was evaluated in 2 simultaneously conducted multicenter, randomized, double blind trials called Study of Ascending Levels of Tolvaptan in hyponatremia 1 and 2 (SALT 1 and SALT 2).6 The 2 studies together randomized 225 patients with euvolemic or hypervolemic hyponatremia to outpatient treatment with the study drug for 30 days and 223 patients to placebo. Patients with a serum Na⁺ of <120 mEq/L in association with neurologic impairment were excluded from the trial. Serum Na⁺ increased more in the tolvaptan group than in the placebo during the first 4 days (3.62 vs. 0.25 mEq/L). After 30 days of therapy serum Na⁺ concentrations were 6.22 higher in the tolvaptan group compared with 1.66 mEq/L in the placebo group. When tolvaptan was discontinued at study end, the serum Na⁺ fell over a seven day period to a value similar to that in the placebo treated patients. Four patients (1.8%) exceeded the study goal of limiting correction of the hyponatremia to <12 mEq/L in the first 24 hours of treatment, but none of these patients developed adverse clinical sequelae.

As previously mentioned conivaptan and tolvaptan are indicated for the treatment of hyponatremia in the setting of euvolemia or hypervolemia. These drugs should not be used in hypovolemic states since the increase in renal water excretion can potentially predispose to worsening hemodynamics in the setting of volume depletion. To be sure, any volume of water removed from the body is principally derived from the intracellular compartment (two‐thirds) and would not be expected to affect blood pressure to a major extent. However, one‐twelfth of any water volume loss is derived from the circulating compartment and could potentially aggravate a borderline low blood pressure in the setting of volume depletion.

This concern is particularly true for conivaptan and its blocking effects on the V_1a receptor. AVP can cause peripheral vasoconstriction by stimulating the V_1a receptor on the peripheral vasculature.7 However, circulating concentrations observed in euvolemic and hypervolemic conditions are not typically of a magnitude to elicit this effect thus explaining the lack of clinically significant hypotension in clinical trials. By contrast, AVP may reach a high enough concentration and play a contributory role in the maintenance of blood pressure under conditions of significant depletion of extracellular fluid volume (hemorrhage) or in states of generalized vasodilation such as sepsis or advanced cirrhosis. In these settings blockade of the V_1a receptor may result in significant hypotension.

Another theoretical concern of blocking the V_1a receptor with conivaptan is the potential to cause further sequestration of fluid in the splanchnic vascular bed and theoretically increase the risk of hepatorenal syndrome.8 The importance of splanchnic vasodilation in the genesis of renal hypoperfusion has been indirectly illustrated by the response to ornipressin, an analog of AVP that is a preferential splanchnic vasoconstrictor. The administration of ornipressin to patients with advanced cirrhosis leads to correction of many of the systemic and renal hemodynamic abnormalities that are present. These include an elevation in mean arterial pressure, reductions in plasma renin activity and norepinephrine concentration, and increases in renal blood flow, glomerular filtration rate, and urinary sodium excretion and volume. V_1a receptor blockade has the potential to increase the degree of arterial vasodilation in the splanchnic arteriolar bed. Increasing degrees of splanchnic vasodilation contribute to a fall in mean arterial pressure and unloading of baroreceptors in the central circulation. As a result, central afferent sensors signal the activation of neurohumoral effectors which in turn decrease perfusion of other organs, particularly the kidney.

AVP Antagonists in Severe Symptomatic Hyponatremia

Symptoms of hyponatremia include nausea and malaise, which can be followed by headache, lethargy, muscle cramps, restlessness, confusion and disorientation.9 Life threatening symptoms are those of impending brain herniation and include seizures, decreased levels consciousness, and obtundation. Hypertonic saline remains the treatment of choice in those patients who are clinically determined to be severely symptomatic.10 Vaptan therapy does result in a brisk and relatively prolonged water diuresis. Indeed, in the first 6 hours following the parenteral administration of conivaptan there is a several mEq/L increase in the serum Na⁺ concentration.5 Whether this time course of correction is sufficient to abort fatal hyponatremic encephalopathy is simply not known. Such patients to date have been excluded from clinical studies since randomization in a placebo‐controlled trial would be clearly unethical in such subjects, and not using hypertonic saline might also be considered unsafe

Although not well studied, one could coadminister conivaptan along with hypertonic saline and anticipate a more rapid initial rise in the serum Na⁺. Once stabilized, the hypertonic saline could be discontinued and the remainder of correction be accomplished by use of the receptor antagonist alone. This strategy would help to minimize the likelihood of volume overload due to the use of hypertonic saline but would require frequent monitoring of the serum Na⁺ to ensure the rate of correction was within accepted guidelines.

AVP Antagonists in Patients With Mild to Moderate Symptoms of Hyponatremia

Hospitalized patients with mild to moderate symptoms of hyponatremia can be considered ideal candidates for the use of V₂ receptor antagonists. Both conivaptan and tolvaptan can be expected to increase the serum Na⁺ concentration to a greater extent and more predictably than fluid restriction alone. The superiority of these drugs would be particularly evident in those who require ongoing fluid administration for any number of reasons such as parenteral administration of antibiotics or proton pump inhibitors.11

In hyponatremic patients with decompensated heart failure there may be changes in mental status in which it is difficult to separate out the contribution of hyponatremia from the decrease in cerebral perfusion. Vaptan therapy offers a predictable way to improve the serum Na⁺ contribution over an acceptable period of time thereby removing one of the variables. Even if hyponatremia is not the direct cause of symptoms, it may lower the threshold for mental status changes resulting from poor cerebral perfusion. A similar argument for vaptan therapy can be made in hyponatremic patients with cirrhosis. In this population mental status changes can be due to the hyponatremia, the underlying liver disease, or both. If no improvement occurred following the correction of the hyponatremia, therapy could then be focused on the underlying liver disease. In fact, in any delirious patient with hyponatremia, correction of the underlying electrolyte disorder can help to clarify the degree to which the hyponatremia is exacerbating the altered mental status.

There are no head to head studies comparing conivaptan and tolvaptan in patients with mild to moderate symptoms of hyponatremia. Presumably both drugs would work with similar efficacy. Conivaptan therapy can be complicated by phlebitis when administered through a peripheral vein due to the irritative effects of polypropylene glycol which serves as a diluent for the drug.12 For this reason the drug is often given through a central vein. The theoretical risk of conivaptan's V_1a receptor blocking effects in patients with cirrhosis has already been discussed. Tolvaptan is given orally starting at 15 mg daily. Depending on the response the dose can be increased to 30 mg and ultimately to 60 mg daily. Few patients in the SALT 1 and 2 trials required the 60 mg dose.

Measurement of the urine osmolality may be useful in predicting how responsive a patient will be to the administration of a vaptan. In this regard, urine osmolality can be thought as a biomarker of the effect of AVP on the renal collecting duct.13 The higher the urine osmolality, presumably the higher the serum AVP level or the greater the effect of AVP is on the tubule. By contrast, those with a lower urine osmolality presumably have less AVP effect on the tubule and therefore will manifest a less robust response to a V₂ receptor blocker.

There are circumstances unique to the hospitalized patient where one may consider use of a V₂ receptor antagonist. Consider a patient admitted with pneumonia complicated by hyponatremia (Na⁺ = 128 mEq/L) due to syndrome of inappropriate antidiuretic hormone secretion (SIADH) who is treated with intravenous antibiotics, stabilized, and is now being considered for discharge. Despite fluid restriction the serum Na⁺ concentration is currently 126 mEq/L and the admitting physician is reluctant to discharge the patient for fear the Na⁺ may fall further and result in symptomatic hyponatremia. The hospital discharge is postponed and fluid restriction is intensified. Two days later the serum Na⁺ concentration has risen to 131 mEq/L and the patient is discharged. While one can question the wisdom of postponing the original planned discharge based on the lab value alone, this type of scenario is common in clinical practice. Fluid restriction is poorly tolerated, difficult to enforce, and often unpredictable in response. The administration of a single dose of a V₂ receptor antagonist is a reliable way to increase the serum Na⁺ concentration over a 24 hour period and in this case could have potentially shortened the hospital stay.

Consider another patient admitted with a hip fracture who is found to have a serum Na⁺ concentration of 126 meq/L attributed to ongoing use of an antidepressant agent. The patient is otherwise medically stable but the anesthesiologist is unwilling to accept the patient for surgery until the serum Na⁺ is at least >130 mEq/L. Fluid restriction is prescribed and only 2 to 3 days later has the serum Na⁺ reached a value of 131 mEq/L. Once again, administration of a vaptan would have almost certainly increased the Na⁺ concentration to the desired threshold over a 24 hour period, allowing the patient to undergo the surgical procedure without experiencing the undue delay.

In each of these scenarios short term use of a vaptan would provide for a more rapid and predictable increase in the serum Na⁺ concentration and potentially decrease the length of hospital stay in comparison to fluid restriction alone. In addition to being less predictable and often poorly tolerated, fluid restriction is often ineffective due to the obligatory fluid administration hospitalized patients require for other therapies or nutrition. Demeclocycline and lithium have been used to antagonize the effects of AVP on the tubule but these drugs take several days before any demonstrable increase in renal water excretion is seen. Demeclocycline antagonizes the effect of AVP through nephrotoxic effects on the tubular cell whereas vaptan's are competitive inhibitors of the V₂ receptor and are not associated with nephrotoxicity. Lithium interferes in the intracellular signaling pathways by which AVP causes insertion of water channels into the apical membrane. The doses required to illicit an increase in renal water excretion are near those which can result in lithium levels sufficient to cause neurotoxicity. The primary side effect of vaptan therapy are those one would predict from inducing an aquaretic effect and include thirst, increased urinary frequency, and increased urinary volume.

The decision to use a vaptan on a more prolonged basis is made on a case by case basis. When the underlying cause of increased AVP is deemed to be chronic and irreversible then therapy can be extended into the outpatient setting. Consideration for more chronic therapy would be appropriate for patients with SIADH due to underlying cancer, or those with severe chronic congestive heart failure or advanced cirrhosis. If the cause of the increased AVP is transient, then 1 to 2 doses of a vaptan while in the hospital may be all that is required. Transient causes of increased AVP would include drug related causes, acute pneumonia, hypoxia associated with respiratory failure, and acute decompensated heart failure. Once again vaptan therapy should not be used when the cause of increased AVP is due to total body salt depletion.

AVP Antagonists in Asymptomatic Hyponatremic Patients

One remaining question concerning the use of vaptan therapy is whether they are helpful in hyponatremic patients who are asymptomatic. There are several observations which raise the possibility that patients deemed to be otherwise asymptomatic do in fact have subtle abnormalities attributable to the low Na⁺ concentration and therefore could benefit if the hyponatremia was corrected. First, some patients with a serum Na⁺ in the range of 120 mEq/L to 129 mEq/L have subtle neurologic changes that improve when the serum Na⁺ concentration is increased.14 These include scores on tests of mental and social functioning. In the SALT 1 and 2 trials a general health survey filled out by the patients showed tolvaptan therapy was associated with improvements in the mental health component of the instrument assessing parameters such as vitality, social functioning, emotionally limited accomplishments, calmness, and sadness.6 Response time is delayed and the error rate is increased in response to various stimuli while patients are hyponatremic, suggesting hyponatremia is associated with reversible impairment in attention. It is certainly reasonable to speculate that correction of hyponatremia in an elderly patient with SIADH might improve that individual's quality of life by allowing them to better enjoy any number of activities such as reading or completing a crossword puzzle. The presence of concomitant neuropsychiatric disturbance (delirium and dementia in particular) may also be relevant, since even mild hyponatremia may increase vulnerability to further alterations in mental status in patients who are impaired at baseline.

Second, asymptomatic hyponatremic patients exhibit subtle disturbances in gait that improve following correction of the serum Na⁺ concentration.14, 15 In this regard, case control studies have shown an association between hyponatremia and risk of falls and fractures particularly in the elderly population.15, 16 Chronic hyponatremia has been shown to cause a reduction in bone mass in an experimental model of SIADH and hyponatremia is associated with osteoporosis in cross‐sectional human data.17 Fall‐related injury is associated with substantial adverse psychological and physical outcomes and is a cause of both death and disability in this population. It would be of considerable socioeconomic benefit if treatment of asymptomatic hyponatremia were able to reduce the risk of this complication.

Lastly, hyponatremia is associated with increased morbidity and mortality but generally the low Na⁺ concentration is merely thought to be marker of the severity of the underlying disease and not a direct contributor to the adverse outcome.18, 19 For example in decompensated heart failure there is an inverse relationship between the degree of hyponatremia and the extent of neurohormonal activation. Adverse effects resulting from persistent activation of the renin‐angiotensin‐aldosterone system and sympathetic nerve activity are thought to be the mechanism underlying the increase in mortality in hyponatremic patients with heart failure. Nevertheless, it has been suggested that hyponatremia itself, the associated hypotonicity, and/or elevated levels of AVP might exert adverse effects on the cardiovascular system or other organ systems and therefore play a contributory role in patient morbidity and mortality. While it remains speculative as to whether correction of the hyponatremia per se will improve patient outcomes, the V₂ receptor antagonists offer an opportunity to test this uncertainty in patients with euvolemic and hypervolemic hyponatremia.0, 0

Pathogenesis

The development of hyponatremia in cirrhosis is intimately related to the pathophysiology of portal hypertension and the non‐osmotic release of AVP3, 15 (Figure 1). In the early phases of cirrhosis, portal hypertension is the result of an increase in intrahepatic resistance. With the development of porto‐systemic collaterals, a hyperdynamic splanchnic circulation develops as a result of splanchnic arterial vasodilatation and increased vascular capacity. Nitric oxide, an endothelial derived relaxing factor, is the critical mediator of this process, and upregulation of its expression is pivotal in the pathogenesis of portal hypertension.

Figure 1

Multiple factors are related to the development of DH in cirrhosis. A reduction of effective central blood volume due to the development of porto‐venous collaterals and arterial splanchnic vasodilation, leading to baroreceptor‐mediated nonosmotic release of AVP, is considered the initiating and most important factor. Patients with cirrhosis and DH have higher plasma and urine vasopressin levels, higher plasma renin activity, and decreased plasma levels of atrial natriuretic factor than those with normal serum sodium concentrations, findings consistent with the presence of a decreased effective plasma volume.16 Arterial underfilling is sensed by baroreceptors located in the left ventricle, aortic arch, carotid sinus and renal afferent arterioles. Decreased activation leads to neurohumoral compensatory responses which include non‐osmotic release of vasopressin from the neurohypophysis and increased levels. Impaired catabolism of AVP that has been correlated with the severity of liver dysfunction may further contribute to increased levels.17 Initially, the increased AVP maintains arterial circulatory integrity by inducing splanchnic, peripheral and renal arterial vasoconstriction through its action on the V1a receptors and expansion of blood volume through renal water retention by its action on the V2 receptors located on the collecting ducts.

The initial adaptive response which leads to increased central blood volume can chronically result in detrimental effects, including the development of fluid overload with ascites, edema, and hyponatremia.16, 18 Additional factors that contribute to hyponatremia include decreased glomerular filtration rate (GFR) and/or increased proximal reabsorption of sodium (that reduce the distal delivery of filtrate and the potential for water reabsorption) and decreased cardiac function that further impairs effective central blood volume.19 In addition, urinary levels of AQP2 are increased in cirrhotic patients, especially those with decompensated disease with higher Child‐Pugh scores and ascites, and provide another potential mechanism to increase water reabsorption.20

Prevalence and Prognostic Significance

Hyponatremia in cirrhosis is a common finding. In a survey of 997 cirrhotic patients with ascites from 28 centers in Europe, North and South America, the prevalence of serum sodium concentration 135, 130, 125, 120 meq/L were 49.4%, 21.6%, 5.7%, and 1.2%, respectively.21 In a retrospective analysis of 188 inpatients, the prevalence of DH of 135, 130, and 125 were 20.8%, 14.9%, and 12.2%, respectively.22 The development of hyponatremia is a manifestation of increasing portal hypertension. In a natural history study of 263 patients hospitalized for first episode of significant ascites, 74 patients developed DH (Na level < 130 mEq/L), including 11 patients in whom it appeared during the first episode and 63 cases during follow‐up (mean period of 40 3 months) with a 5‐year incidence of 37.1%.23

The presence of hyponatremia carries significant adverse prognostic significance. It is strongly associated with severity of liver function impairment as assessed by Child‐Pugh and model for end‐stage liver disease (MELD) scores.22 Even mild hyponatremia is associated with severe complications such as massive ascites, severe hepatic encephalopathy, spontaneous bacterial peritonitis (SBP), and hepatic hydrothorax, and the severity of hyponatremia is directly related to the severity of these complications.21, 22 (Figure 2). In a natural history study of patients presenting with large volume ascites, 1‐year survival after its development was reduced to only 25.6%.230

Figure 2

Figure 3

Hyponatremia is an especially poor prognostic sign for a hospitalized cirrhotic patient. In a retrospective analysis of 156 cirrhotic patients, hyponatremiapresent in 57 (29.8%) of admissionswas associated with increased hospital mortality (26.3% vs. 8.9% among those with normal Na levels), and the mortality rate was even higher (48%) among the 25 patients who developed severe hyponatremia during the hospital stay.24 In hospitalized patients, hyponatremia is predictive of the development of acute renal failure which is associated with substantially increased mortality (73% vs. 13%).25 Similarly, a low serum sodium level in critically ill cirrhotic patients admitted to the ICU is associated with complications, in‐hospital mortality, and poor short‐term prognosis.26

Whether hyponatremia should impact liver transplant prioritization remains an area of controversy. The United Network for Organ Sharing (UNOS) contracted by the Organ Procurement and Transplant Network (OPTN) to optimize the efficient use of deceased organs through fair and timely allocation, currently uses the MELD score, a formula that calculates the risk of death within three months from the bilirubin, creatinine, and International Normalized Ratio (INR) levels. Hyponatremia is an earlier and more sensitive marker than serum creatinine to detect renal impairment and/or circulatory dysfunction in patients with advanced cirrhosis and adds to MELD in predicting waitlist mortality.2729 In patients with a MELD score of <21, only low serum sodium and persistent ascites are independent predictors of mortality.28 To account for the importance of hyponatremia on survival, both modification of the MELD score in which the Na level is incorporated (MELD‐Na model) and the MELD to serum sodium ratio (MESO) have been developed. Adding hyponatremia to the MELD score is a better predictor of death than MELD alone, particularly in patients with low MELD scores.27, 2931 The OPTN/UNOS Liver and Intestinal Organ Transplantation Committee has discussed updating the liver allocation system to include the Na level. However, it was concluded that implementation of MELD‐Na would change the allocation status of only 4% of candidates. Further, based on the concerns about the ability to manipulate serum sodium levels and the utility of employing resources to change the system for a relatively small number of patients, it was decided to defer incorporating the Na level pending further analysis (Report of the OPTN/UNOS Liver and Intestinal Organ Transplantation Committee To the Board of Directors, Los Angeles, California, September 17‐18, 2007). At this time, the use of Na is a regional decision.32 However, the OPTN/UNOS Liver and Intestinal Organ Transplantation Committee has recently solicited feedback from the transplant community about including Na in allocation for review at a forum in April 2010.

Precipitating Factors

The most important factor related to development of hyponatremia in cirrhosis is increasing severity of portal hypertension that is associated with impaired central blood volume as a result of progressive splanchnic vasodilatation. In a study in which 170 patients with decompensated alcoholic cirrhosis were prospectively followed for 33.9 27.9 months, the initial hepatic venous pressure gradient (HVPG) was an independent predictive factor for the 20 patients who developed hyponatremia.22

Cirrhotic patients with ascites with hyponatremia have increased AVP secretion, higher levels of plasma renin activity, and higher serum concentrations of aldosterone and norepinephrine compared to those with normal Na levels.33 Diuretic therapy is associated with the development of DH by inducing volume depletion and arterial underfilling, further activating the renin‐angiotensin system (RAS) and increasing the non‐osmotic release of AVP.34 Although diuretics block the salt retention associated with the RAS activation, the water‐retaining effects of AVP persist, and DH develops. The process is further exacerbated by a low sodium intake and a frequent uncontrollable thirst. As a result, diuretic therapy is commonly associated with the development of hyponatremia in patients with ascites.24, 35 Similarly, paracentesis (particularly when performed without albumin) is often associated with an increase in blood urea nitrogen and marked elevations in plasma renin activity and plasma aldosterone concentration, which may exacerbate this physiology, leading to further reduction in serum sodium concentration.36 Tense ascites can contribute to DH by increasing baroreceptor mediated AVP release by increasing intrathoracic pressure.37 Finally, non‐steroidal anti‐inflammatory drugs (NSAIDs) can cause DH by inhibiting the synthesis of renal prostaglandins (which normally function to antagonize the tubular action of AVP and are important in the maintenance of appropriate renal tubular transport of fluid and electrolytes in states of renal hypoperfusion).38

Medical Impact of Hyponatremia: Marker of Severe Disease or Direct Pathophysiologic Role?

Hyponatremia is associated with severe ascites, impaired renal function, hepatic encephalopathy, SBP, and hepatorenal syndrome.3, 20 Because hyponatremia is frequently present in advanced liver failure, it is unclear whether it is only a marker of advanced disease or whether it plays a direct pathophysiologic role, or both. Until recently, it has not been possible to address this issue due to the inability to easily and rapidly correct the hyponatremia. However, there is increasing evidence that hyponatremia has direct impact on the severity of hepatic encephalopathy (see Hepatic Encephalopathy section). The recent introduction of tolvaptan for the treatment of hyponatremia in cirrhosis (discussed below) will allow this question to be directly answered.

Management

Most patients with mild hypervolemic hyponatremia are asymptomatic. The initial recommended approach is fluid restriction and an Na‐restricted diet. For those with severe or progressive hyponatremia, diuretics should be minimized or discontinued to avoid intravascular volume depletion.49 For patients with tense ascites and severe DH, therapeutic paracentesis with plasma expanders is safe.33 Unfortunately, fluid restriction is limited in efficacy and often poorly tolerated. The use of hypertonic saline is generally not recommended unless severe neurologic symptoms are present as it leads to increased ascites and edema. When administered, it is important to avoid a rapid correction of the hyponatremia to prevent the development of central pontine myelinolysis and the osmotic demyelination syndrome.

Due to the pivotal role of AVP in the pathogenesis of DH, antagonism of its action has long been proposed to be the most rational approach, but until recently, effective and specific antagonism of AVP has remained elusive. Approaches that have been attempted include interference with its secretion and actions. Intravenous albumin has been reported to improve hyponatremia in patients with cirrhosis, ascites, and hyponatremia, presumably by decreasing AVP release by plasma volume expansion.50 An attempt at inhibition of central AVP release with the use of a kappa‐opioid receptor agonist, niravoline, was limited by loss of efficacy and potential adverse effects.51 Use of demeclocycline and lithium (which induce renal resistance to AVP and lead to a modest increase in urine volume with decreased urine osmolality and a corresponding rise in serum sodium) is limited by nephrotoxicity and hepatotoxicity.7, 52 Because of the important role played by prostaglandins in the maintenance of renal hemodynamics and water excretion in cirrhosis, oral misoprostol has also been evaluated but determined to be ineffective in inducing significant changes in free water clearance in patients with functional renal failure and/or DH.53

The recent introduction of vaptans, vasopressin receptor antagonists that block the physiologic action of vasopressin, represents a revolutionary and highly effective approach to the treatment of hyponatremia. Vaptans are antagonists of the V2 receptors of AVP in the principal cells of the collecting ducts. In healthy subjects, vaptans cause a dose‐dependent increase in urine volume and produce a dilute urine without causing natriuresis. To date, 2 AVP antagonists, conivaptan and tolvaptan, have been Food and Drug Administration (FDA)‐approved for the treatment of DH. Conivaptan, the first to be approved in 2005, is a mixed vasopressin V1a and V2 receptor antagonist that is administered intravenously for up to 4 days. In a randomized placebo‐controlled study of patients with euvolemic or hypervolemic hyponatremia, intravenous conivaptan treatment increased serum Na levels by >6 mEq/l or to a serum Na >135 mEq/l in 69 to 88.5% of subjects compared to 20.7% of those receiving placebo (Zeltser D, Rosansky S, Van Rensburg H, et al. Assessment of the efficacy and safety of intravenous conivaptan in euvolemic and hypervolemic hyponatremia. American J Nephrology 2007;27:447457). In a pilot study involving 24 patients with end‐stage liver disease, an infusion of conivaptan over 1 to 4 days was associated with an increase of serum sodium by >5 mmol/L in 60% of patients not receiving diuretics and 67% of patients on concomitant diuretic therapy by the end of treatment (O'Leary and Davis, 2009). Despite a concern about the potential for conivaptan to increase portal hypertension due to inhibition of splanchnic V1a receptors, the brief treatment appeared to be well tolerated without significant changes in systolic blood pressure, serum creatinine, variceal bleeding or worsening of ascites during the infusion period. However, approval for only 4 days of therapy and requirement for intravenous use eliminate any potential for chronic use.54

Tolvaptan is an orally available, selective V2 receptor antagonist whose efficacy was assessed in two multicenter, prospective, randomized, placebo‐controlled trials, Study of Ascending Levels of Tolvaptan in Hyponatremia 1 and 2 (SALT‐1 and SALT‐2).55 In these trials, clinically stable patients with DH (Na <135meq/l) associated with cirrhosis (22.4% in SALT‐1, 30.5% in SALT‐2), CHF or SIADH were randomized in a hospital setting to receive tolvaptan 15mg daily or placebo. Repeat Na levels were obtained at 8 hours, 2, 3, and 4 days and then weekly at days 11, 18, 25 and 30 after which study drug was discontinued and follow‐up Na level was determined 7 days later. The dose was adjusted to 30 mg and then 60 mg in an attempt to achieve a Na level >135 in those in whom hyponatremia persisted. During the initial day of the titration phase, fluid restriction was not maintained, and the patients were encouraged to respond to thirst with increased water ingestion.

Tolvaptan use was associated with a prompt increase in Na level as early as 8 hours after administration of the first dose. Serum Na increased more among those receiving tolvaptan than among those receiving placebo during the first 4 days and throughout the study period regardless of baseline Na level but returned to baseline within 1 week after discontinuation (Figure 3). The main side effects were increased thirst, dry mouth and increased urination. Importantly, an increased incidence of renal failure was not observed. Based on these results, FDA approval for tolvaptan in patients with hyponatremia was obtained in May 2009 for patients with DH‐associated with cirrhosis, CHF or SIADH for patients with Na levels <125 or symptomatic patients with Na levels between 125 and 135 that have not responded to fluid restriction.

Evaluation

Determination of volume status is paramount but frequently problematic in the hospitalized cirrhotic patient. Due to the vasodilated state present in severe portal hypertension that is characterized by a relative hypotension and resting tachycardia, the usual hemodynamic parameters of blood pressure and heart rate can be difficult to interpret. Although significant extravascular volume in the form of ascites and edema may be present, patients may be intravascularly depleted due to previous diuretic use and extra‐renal losses due to impaired oral intake, vomiting, lactulose‐induced diarrhea, and gastrointestinal bleeding. Infection is a commonly associated condition, and endotoxin mediated splanchnic vasodilatation, especially in the setting of SBP, can adversely effective central blood volume status in the presence of severe ascites. Also, due to the Na avidity of the kidney and previous diuretic use, renal electrolytes can be difficult to interpret.

For patients in whom there is strong clinical concern about intravascular depletion (history of impaired oral intake, excessive vomiting and/or diarrhea, rapid weight loss, small volume ascites with history of large volume, azotemia), administration of limited intravenous normal saline (0.5‐1 L) should be considered. Patients with severe neurological symptoms should receive normal saline or hypertonic saline. Unless severe neurologic symptoms associated with profound hyponatremia is present, however, intravenous normal saline should not be administered for the hyponatremia alone. Administration of salt poor albumin (25%), especially for those with marked fluid overload and ascites, is an effective means to expand the central blood volume without exacerbating ascites and edema. After evaluation for and/or treatment of hypovolemia, all patients should receive a Na restricted diet (<2000 mg daily) and placed on fluid restriction (see below for liberalization of fluid restriction upon initiation of tolvaptan therapy).

Diagnostic paracentesis should be performed for those with ascites to rule out the presence of SBP, and antibiotics administered to those with evidence of infection. High dose intravenous salt poor albumin should also be administered, especially to those at high risk of renal failure as determined by the presence of azotemia (Cr > 1.0 mg/dL) or severe liver insufficiency (TBili > 4.0 mg/dL).56, 57 Finally, all medications should be reviewed, and those associated with hyponatremia (diuretics, selective serotonin reuptake inhibitors, opiates, proton‐pump inhibitors) discontinued if possible.

Tolvaptan for DH

Pathogenesis

The development of hyponatremia in cirrhosis is intimately related to the pathophysiology of portal hypertension and the non‐osmotic release of AVP3, 15 (Figure 1). In the early phases of cirrhosis, portal hypertension is the result of an increase in intrahepatic resistance. With the development of porto‐systemic collaterals, a hyperdynamic splanchnic circulation develops as a result of splanchnic arterial vasodilatation and increased vascular capacity. Nitric oxide, an endothelial derived relaxing factor, is the critical mediator of this process, and upregulation of its expression is pivotal in the pathogenesis of portal hypertension.

Figure 1

Multiple factors are related to the development of DH in cirrhosis. A reduction of effective central blood volume due to the development of porto‐venous collaterals and arterial splanchnic vasodilation, leading to baroreceptor‐mediated nonosmotic release of AVP, is considered the initiating and most important factor. Patients with cirrhosis and DH have higher plasma and urine vasopressin levels, higher plasma renin activity, and decreased plasma levels of atrial natriuretic factor than those with normal serum sodium concentrations, findings consistent with the presence of a decreased effective plasma volume.16 Arterial underfilling is sensed by baroreceptors located in the left ventricle, aortic arch, carotid sinus and renal afferent arterioles. Decreased activation leads to neurohumoral compensatory responses which include non‐osmotic release of vasopressin from the neurohypophysis and increased levels. Impaired catabolism of AVP that has been correlated with the severity of liver dysfunction may further contribute to increased levels.17 Initially, the increased AVP maintains arterial circulatory integrity by inducing splanchnic, peripheral and renal arterial vasoconstriction through its action on the V1a receptors and expansion of blood volume through renal water retention by its action on the V2 receptors located on the collecting ducts.

The initial adaptive response which leads to increased central blood volume can chronically result in detrimental effects, including the development of fluid overload with ascites, edema, and hyponatremia.16, 18 Additional factors that contribute to hyponatremia include decreased glomerular filtration rate (GFR) and/or increased proximal reabsorption of sodium (that reduce the distal delivery of filtrate and the potential for water reabsorption) and decreased cardiac function that further impairs effective central blood volume.19 In addition, urinary levels of AQP2 are increased in cirrhotic patients, especially those with decompensated disease with higher Child‐Pugh scores and ascites, and provide another potential mechanism to increase water reabsorption.20

Prevalence and Prognostic Significance

Hyponatremia in cirrhosis is a common finding. In a survey of 997 cirrhotic patients with ascites from 28 centers in Europe, North and South America, the prevalence of serum sodium concentration 135, 130, 125, 120 meq/L were 49.4%, 21.6%, 5.7%, and 1.2%, respectively.21 In a retrospective analysis of 188 inpatients, the prevalence of DH of 135, 130, and 125 were 20.8%, 14.9%, and 12.2%, respectively.22 The development of hyponatremia is a manifestation of increasing portal hypertension. In a natural history study of 263 patients hospitalized for first episode of significant ascites, 74 patients developed DH (Na level < 130 mEq/L), including 11 patients in whom it appeared during the first episode and 63 cases during follow‐up (mean period of 40 3 months) with a 5‐year incidence of 37.1%.23

The presence of hyponatremia carries significant adverse prognostic significance. It is strongly associated with severity of liver function impairment as assessed by Child‐Pugh and model for end‐stage liver disease (MELD) scores.22 Even mild hyponatremia is associated with severe complications such as massive ascites, severe hepatic encephalopathy, spontaneous bacterial peritonitis (SBP), and hepatic hydrothorax, and the severity of hyponatremia is directly related to the severity of these complications.21, 22 (Figure 2). In a natural history study of patients presenting with large volume ascites, 1‐year survival after its development was reduced to only 25.6%.230

Figure 2

Figure 3

Hyponatremia is an especially poor prognostic sign for a hospitalized cirrhotic patient. In a retrospective analysis of 156 cirrhotic patients, hyponatremiapresent in 57 (29.8%) of admissionswas associated with increased hospital mortality (26.3% vs. 8.9% among those with normal Na levels), and the mortality rate was even higher (48%) among the 25 patients who developed severe hyponatremia during the hospital stay.24 In hospitalized patients, hyponatremia is predictive of the development of acute renal failure which is associated with substantially increased mortality (73% vs. 13%).25 Similarly, a low serum sodium level in critically ill cirrhotic patients admitted to the ICU is associated with complications, in‐hospital mortality, and poor short‐term prognosis.26

Whether hyponatremia should impact liver transplant prioritization remains an area of controversy. The United Network for Organ Sharing (UNOS) contracted by the Organ Procurement and Transplant Network (OPTN) to optimize the efficient use of deceased organs through fair and timely allocation, currently uses the MELD score, a formula that calculates the risk of death within three months from the bilirubin, creatinine, and International Normalized Ratio (INR) levels. Hyponatremia is an earlier and more sensitive marker than serum creatinine to detect renal impairment and/or circulatory dysfunction in patients with advanced cirrhosis and adds to MELD in predicting waitlist mortality.2729 In patients with a MELD score of <21, only low serum sodium and persistent ascites are independent predictors of mortality.28 To account for the importance of hyponatremia on survival, both modification of the MELD score in which the Na level is incorporated (MELD‐Na model) and the MELD to serum sodium ratio (MESO) have been developed. Adding hyponatremia to the MELD score is a better predictor of death than MELD alone, particularly in patients with low MELD scores.27, 2931 The OPTN/UNOS Liver and Intestinal Organ Transplantation Committee has discussed updating the liver allocation system to include the Na level. However, it was concluded that implementation of MELD‐Na would change the allocation status of only 4% of candidates. Further, based on the concerns about the ability to manipulate serum sodium levels and the utility of employing resources to change the system for a relatively small number of patients, it was decided to defer incorporating the Na level pending further analysis (Report of the OPTN/UNOS Liver and Intestinal Organ Transplantation Committee To the Board of Directors, Los Angeles, California, September 17‐18, 2007). At this time, the use of Na is a regional decision.32 However, the OPTN/UNOS Liver and Intestinal Organ Transplantation Committee has recently solicited feedback from the transplant community about including Na in allocation for review at a forum in April 2010.

Precipitating Factors

The most important factor related to development of hyponatremia in cirrhosis is increasing severity of portal hypertension that is associated with impaired central blood volume as a result of progressive splanchnic vasodilatation. In a study in which 170 patients with decompensated alcoholic cirrhosis were prospectively followed for 33.9 27.9 months, the initial hepatic venous pressure gradient (HVPG) was an independent predictive factor for the 20 patients who developed hyponatremia.22

Cirrhotic patients with ascites with hyponatremia have increased AVP secretion, higher levels of plasma renin activity, and higher serum concentrations of aldosterone and norepinephrine compared to those with normal Na levels.33 Diuretic therapy is associated with the development of DH by inducing volume depletion and arterial underfilling, further activating the renin‐angiotensin system (RAS) and increasing the non‐osmotic release of AVP.34 Although diuretics block the salt retention associated with the RAS activation, the water‐retaining effects of AVP persist, and DH develops. The process is further exacerbated by a low sodium intake and a frequent uncontrollable thirst. As a result, diuretic therapy is commonly associated with the development of hyponatremia in patients with ascites.24, 35 Similarly, paracentesis (particularly when performed without albumin) is often associated with an increase in blood urea nitrogen and marked elevations in plasma renin activity and plasma aldosterone concentration, which may exacerbate this physiology, leading to further reduction in serum sodium concentration.36 Tense ascites can contribute to DH by increasing baroreceptor mediated AVP release by increasing intrathoracic pressure.37 Finally, non‐steroidal anti‐inflammatory drugs (NSAIDs) can cause DH by inhibiting the synthesis of renal prostaglandins (which normally function to antagonize the tubular action of AVP and are important in the maintenance of appropriate renal tubular transport of fluid and electrolytes in states of renal hypoperfusion).38

Medical Impact of Hyponatremia: Marker of Severe Disease or Direct Pathophysiologic Role?

Hyponatremia is associated with severe ascites, impaired renal function, hepatic encephalopathy, SBP, and hepatorenal syndrome.3, 20 Because hyponatremia is frequently present in advanced liver failure, it is unclear whether it is only a marker of advanced disease or whether it plays a direct pathophysiologic role, or both. Until recently, it has not been possible to address this issue due to the inability to easily and rapidly correct the hyponatremia. However, there is increasing evidence that hyponatremia has direct impact on the severity of hepatic encephalopathy (see Hepatic Encephalopathy section). The recent introduction of tolvaptan for the treatment of hyponatremia in cirrhosis (discussed below) will allow this question to be directly answered.

Management

Most patients with mild hypervolemic hyponatremia are asymptomatic. The initial recommended approach is fluid restriction and an Na‐restricted diet. For those with severe or progressive hyponatremia, diuretics should be minimized or discontinued to avoid intravascular volume depletion.49 For patients with tense ascites and severe DH, therapeutic paracentesis with plasma expanders is safe.33 Unfortunately, fluid restriction is limited in efficacy and often poorly tolerated. The use of hypertonic saline is generally not recommended unless severe neurologic symptoms are present as it leads to increased ascites and edema. When administered, it is important to avoid a rapid correction of the hyponatremia to prevent the development of central pontine myelinolysis and the osmotic demyelination syndrome.

Due to the pivotal role of AVP in the pathogenesis of DH, antagonism of its action has long been proposed to be the most rational approach, but until recently, effective and specific antagonism of AVP has remained elusive. Approaches that have been attempted include interference with its secretion and actions. Intravenous albumin has been reported to improve hyponatremia in patients with cirrhosis, ascites, and hyponatremia, presumably by decreasing AVP release by plasma volume expansion.50 An attempt at inhibition of central AVP release with the use of a kappa‐opioid receptor agonist, niravoline, was limited by loss of efficacy and potential adverse effects.51 Use of demeclocycline and lithium (which induce renal resistance to AVP and lead to a modest increase in urine volume with decreased urine osmolality and a corresponding rise in serum sodium) is limited by nephrotoxicity and hepatotoxicity.7, 52 Because of the important role played by prostaglandins in the maintenance of renal hemodynamics and water excretion in cirrhosis, oral misoprostol has also been evaluated but determined to be ineffective in inducing significant changes in free water clearance in patients with functional renal failure and/or DH.53

The recent introduction of vaptans, vasopressin receptor antagonists that block the physiologic action of vasopressin, represents a revolutionary and highly effective approach to the treatment of hyponatremia. Vaptans are antagonists of the V2 receptors of AVP in the principal cells of the collecting ducts. In healthy subjects, vaptans cause a dose‐dependent increase in urine volume and produce a dilute urine without causing natriuresis. To date, 2 AVP antagonists, conivaptan and tolvaptan, have been Food and Drug Administration (FDA)‐approved for the treatment of DH. Conivaptan, the first to be approved in 2005, is a mixed vasopressin V1a and V2 receptor antagonist that is administered intravenously for up to 4 days. In a randomized placebo‐controlled study of patients with euvolemic or hypervolemic hyponatremia, intravenous conivaptan treatment increased serum Na levels by >6 mEq/l or to a serum Na >135 mEq/l in 69 to 88.5% of subjects compared to 20.7% of those receiving placebo (Zeltser D, Rosansky S, Van Rensburg H, et al. Assessment of the efficacy and safety of intravenous conivaptan in euvolemic and hypervolemic hyponatremia. American J Nephrology 2007;27:447457). In a pilot study involving 24 patients with end‐stage liver disease, an infusion of conivaptan over 1 to 4 days was associated with an increase of serum sodium by >5 mmol/L in 60% of patients not receiving diuretics and 67% of patients on concomitant diuretic therapy by the end of treatment (O'Leary and Davis, 2009). Despite a concern about the potential for conivaptan to increase portal hypertension due to inhibition of splanchnic V1a receptors, the brief treatment appeared to be well tolerated without significant changes in systolic blood pressure, serum creatinine, variceal bleeding or worsening of ascites during the infusion period. However, approval for only 4 days of therapy and requirement for intravenous use eliminate any potential for chronic use.54

Tolvaptan is an orally available, selective V2 receptor antagonist whose efficacy was assessed in two multicenter, prospective, randomized, placebo‐controlled trials, Study of Ascending Levels of Tolvaptan in Hyponatremia 1 and 2 (SALT‐1 and SALT‐2).55 In these trials, clinically stable patients with DH (Na <135meq/l) associated with cirrhosis (22.4% in SALT‐1, 30.5% in SALT‐2), CHF or SIADH were randomized in a hospital setting to receive tolvaptan 15mg daily or placebo. Repeat Na levels were obtained at 8 hours, 2, 3, and 4 days and then weekly at days 11, 18, 25 and 30 after which study drug was discontinued and follow‐up Na level was determined 7 days later. The dose was adjusted to 30 mg and then 60 mg in an attempt to achieve a Na level >135 in those in whom hyponatremia persisted. During the initial day of the titration phase, fluid restriction was not maintained, and the patients were encouraged to respond to thirst with increased water ingestion.

Tolvaptan use was associated with a prompt increase in Na level as early as 8 hours after administration of the first dose. Serum Na increased more among those receiving tolvaptan than among those receiving placebo during the first 4 days and throughout the study period regardless of baseline Na level but returned to baseline within 1 week after discontinuation (Figure 3). The main side effects were increased thirst, dry mouth and increased urination. Importantly, an increased incidence of renal failure was not observed. Based on these results, FDA approval for tolvaptan in patients with hyponatremia was obtained in May 2009 for patients with DH‐associated with cirrhosis, CHF or SIADH for patients with Na levels <125 or symptomatic patients with Na levels between 125 and 135 that have not responded to fluid restriction.

Evaluation

Determination of volume status is paramount but frequently problematic in the hospitalized cirrhotic patient. Due to the vasodilated state present in severe portal hypertension that is characterized by a relative hypotension and resting tachycardia, the usual hemodynamic parameters of blood pressure and heart rate can be difficult to interpret. Although significant extravascular volume in the form of ascites and edema may be present, patients may be intravascularly depleted due to previous diuretic use and extra‐renal losses due to impaired oral intake, vomiting, lactulose‐induced diarrhea, and gastrointestinal bleeding. Infection is a commonly associated condition, and endotoxin mediated splanchnic vasodilatation, especially in the setting of SBP, can adversely effective central blood volume status in the presence of severe ascites. Also, due to the Na avidity of the kidney and previous diuretic use, renal electrolytes can be difficult to interpret.

For patients in whom there is strong clinical concern about intravascular depletion (history of impaired oral intake, excessive vomiting and/or diarrhea, rapid weight loss, small volume ascites with history of large volume, azotemia), administration of limited intravenous normal saline (0.5‐1 L) should be considered. Patients with severe neurological symptoms should receive normal saline or hypertonic saline. Unless severe neurologic symptoms associated with profound hyponatremia is present, however, intravenous normal saline should not be administered for the hyponatremia alone. Administration of salt poor albumin (25%), especially for those with marked fluid overload and ascites, is an effective means to expand the central blood volume without exacerbating ascites and edema. After evaluation for and/or treatment of hypovolemia, all patients should receive a Na restricted diet (<2000 mg daily) and placed on fluid restriction (see below for liberalization of fluid restriction upon initiation of tolvaptan therapy).

Diagnostic paracentesis should be performed for those with ascites to rule out the presence of SBP, and antibiotics administered to those with evidence of infection. High dose intravenous salt poor albumin should also be administered, especially to those at high risk of renal failure as determined by the presence of azotemia (Cr > 1.0 mg/dL) or severe liver insufficiency (TBili > 4.0 mg/dL).56, 57 Finally, all medications should be reviewed, and those associated with hyponatremia (diuretics, selective serotonin reuptake inhibitors, opiates, proton‐pump inhibitors) discontinued if possible.

Tolvaptan for DH

Pathogenesis

The development of hyponatremia in cirrhosis is intimately related to the pathophysiology of portal hypertension and the non‐osmotic release of AVP3, 15 (Figure 1). In the early phases of cirrhosis, portal hypertension is the result of an increase in intrahepatic resistance. With the development of porto‐systemic collaterals, a hyperdynamic splanchnic circulation develops as a result of splanchnic arterial vasodilatation and increased vascular capacity. Nitric oxide, an endothelial derived relaxing factor, is the critical mediator of this process, and upregulation of its expression is pivotal in the pathogenesis of portal hypertension.

Figure 1

Multiple factors are related to the development of DH in cirrhosis. A reduction of effective central blood volume due to the development of porto‐venous collaterals and arterial splanchnic vasodilation, leading to baroreceptor‐mediated nonosmotic release of AVP, is considered the initiating and most important factor. Patients with cirrhosis and DH have higher plasma and urine vasopressin levels, higher plasma renin activity, and decreased plasma levels of atrial natriuretic factor than those with normal serum sodium concentrations, findings consistent with the presence of a decreased effective plasma volume.16 Arterial underfilling is sensed by baroreceptors located in the left ventricle, aortic arch, carotid sinus and renal afferent arterioles. Decreased activation leads to neurohumoral compensatory responses which include non‐osmotic release of vasopressin from the neurohypophysis and increased levels. Impaired catabolism of AVP that has been correlated with the severity of liver dysfunction may further contribute to increased levels.17 Initially, the increased AVP maintains arterial circulatory integrity by inducing splanchnic, peripheral and renal arterial vasoconstriction through its action on the V1a receptors and expansion of blood volume through renal water retention by its action on the V2 receptors located on the collecting ducts.

The initial adaptive response which leads to increased central blood volume can chronically result in detrimental effects, including the development of fluid overload with ascites, edema, and hyponatremia.16, 18 Additional factors that contribute to hyponatremia include decreased glomerular filtration rate (GFR) and/or increased proximal reabsorption of sodium (that reduce the distal delivery of filtrate and the potential for water reabsorption) and decreased cardiac function that further impairs effective central blood volume.19 In addition, urinary levels of AQP2 are increased in cirrhotic patients, especially those with decompensated disease with higher Child‐Pugh scores and ascites, and provide another potential mechanism to increase water reabsorption.20

Prevalence and Prognostic Significance

Hyponatremia in cirrhosis is a common finding. In a survey of 997 cirrhotic patients with ascites from 28 centers in Europe, North and South America, the prevalence of serum sodium concentration 135, 130, 125, 120 meq/L were 49.4%, 21.6%, 5.7%, and 1.2%, respectively.21 In a retrospective analysis of 188 inpatients, the prevalence of DH of 135, 130, and 125 were 20.8%, 14.9%, and 12.2%, respectively.22 The development of hyponatremia is a manifestation of increasing portal hypertension. In a natural history study of 263 patients hospitalized for first episode of significant ascites, 74 patients developed DH (Na level < 130 mEq/L), including 11 patients in whom it appeared during the first episode and 63 cases during follow‐up (mean period of 40 3 months) with a 5‐year incidence of 37.1%.23

The presence of hyponatremia carries significant adverse prognostic significance. It is strongly associated with severity of liver function impairment as assessed by Child‐Pugh and model for end‐stage liver disease (MELD) scores.22 Even mild hyponatremia is associated with severe complications such as massive ascites, severe hepatic encephalopathy, spontaneous bacterial peritonitis (SBP), and hepatic hydrothorax, and the severity of hyponatremia is directly related to the severity of these complications.21, 22 (Figure 2). In a natural history study of patients presenting with large volume ascites, 1‐year survival after its development was reduced to only 25.6%.230

Figure 2

Figure 3

Hyponatremia is an especially poor prognostic sign for a hospitalized cirrhotic patient. In a retrospective analysis of 156 cirrhotic patients, hyponatremiapresent in 57 (29.8%) of admissionswas associated with increased hospital mortality (26.3% vs. 8.9% among those with normal Na levels), and the mortality rate was even higher (48%) among the 25 patients who developed severe hyponatremia during the hospital stay.24 In hospitalized patients, hyponatremia is predictive of the development of acute renal failure which is associated with substantially increased mortality (73% vs. 13%).25 Similarly, a low serum sodium level in critically ill cirrhotic patients admitted to the ICU is associated with complications, in‐hospital mortality, and poor short‐term prognosis.26

Whether hyponatremia should impact liver transplant prioritization remains an area of controversy. The United Network for Organ Sharing (UNOS) contracted by the Organ Procurement and Transplant Network (OPTN) to optimize the efficient use of deceased organs through fair and timely allocation, currently uses the MELD score, a formula that calculates the risk of death within three months from the bilirubin, creatinine, and International Normalized Ratio (INR) levels. Hyponatremia is an earlier and more sensitive marker than serum creatinine to detect renal impairment and/or circulatory dysfunction in patients with advanced cirrhosis and adds to MELD in predicting waitlist mortality.2729 In patients with a MELD score of <21, only low serum sodium and persistent ascites are independent predictors of mortality.28 To account for the importance of hyponatremia on survival, both modification of the MELD score in which the Na level is incorporated (MELD‐Na model) and the MELD to serum sodium ratio (MESO) have been developed. Adding hyponatremia to the MELD score is a better predictor of death than MELD alone, particularly in patients with low MELD scores.27, 2931 The OPTN/UNOS Liver and Intestinal Organ Transplantation Committee has discussed updating the liver allocation system to include the Na level. However, it was concluded that implementation of MELD‐Na would change the allocation status of only 4% of candidates. Further, based on the concerns about the ability to manipulate serum sodium levels and the utility of employing resources to change the system for a relatively small number of patients, it was decided to defer incorporating the Na level pending further analysis (Report of the OPTN/UNOS Liver and Intestinal Organ Transplantation Committee To the Board of Directors, Los Angeles, California, September 17‐18, 2007). At this time, the use of Na is a regional decision.32 However, the OPTN/UNOS Liver and Intestinal Organ Transplantation Committee has recently solicited feedback from the transplant community about including Na in allocation for review at a forum in April 2010.

Precipitating Factors

The most important factor related to development of hyponatremia in cirrhosis is increasing severity of portal hypertension that is associated with impaired central blood volume as a result of progressive splanchnic vasodilatation. In a study in which 170 patients with decompensated alcoholic cirrhosis were prospectively followed for 33.9 27.9 months, the initial hepatic venous pressure gradient (HVPG) was an independent predictive factor for the 20 patients who developed hyponatremia.22

Cirrhotic patients with ascites with hyponatremia have increased AVP secretion, higher levels of plasma renin activity, and higher serum concentrations of aldosterone and norepinephrine compared to those with normal Na levels.33 Diuretic therapy is associated with the development of DH by inducing volume depletion and arterial underfilling, further activating the renin‐angiotensin system (RAS) and increasing the non‐osmotic release of AVP.34 Although diuretics block the salt retention associated with the RAS activation, the water‐retaining effects of AVP persist, and DH develops. The process is further exacerbated by a low sodium intake and a frequent uncontrollable thirst. As a result, diuretic therapy is commonly associated with the development of hyponatremia in patients with ascites.24, 35 Similarly, paracentesis (particularly when performed without albumin) is often associated with an increase in blood urea nitrogen and marked elevations in plasma renin activity and plasma aldosterone concentration, which may exacerbate this physiology, leading to further reduction in serum sodium concentration.36 Tense ascites can contribute to DH by increasing baroreceptor mediated AVP release by increasing intrathoracic pressure.37 Finally, non‐steroidal anti‐inflammatory drugs (NSAIDs) can cause DH by inhibiting the synthesis of renal prostaglandins (which normally function to antagonize the tubular action of AVP and are important in the maintenance of appropriate renal tubular transport of fluid and electrolytes in states of renal hypoperfusion).38

Medical Impact of Hyponatremia: Marker of Severe Disease or Direct Pathophysiologic Role?

Hyponatremia is associated with severe ascites, impaired renal function, hepatic encephalopathy, SBP, and hepatorenal syndrome.3, 20 Because hyponatremia is frequently present in advanced liver failure, it is unclear whether it is only a marker of advanced disease or whether it plays a direct pathophysiologic role, or both. Until recently, it has not been possible to address this issue due to the inability to easily and rapidly correct the hyponatremia. However, there is increasing evidence that hyponatremia has direct impact on the severity of hepatic encephalopathy (see Hepatic Encephalopathy section). The recent introduction of tolvaptan for the treatment of hyponatremia in cirrhosis (discussed below) will allow this question to be directly answered.

Management

Most patients with mild hypervolemic hyponatremia are asymptomatic. The initial recommended approach is fluid restriction and an Na‐restricted diet. For those with severe or progressive hyponatremia, diuretics should be minimized or discontinued to avoid intravascular volume depletion.49 For patients with tense ascites and severe DH, therapeutic paracentesis with plasma expanders is safe.33 Unfortunately, fluid restriction is limited in efficacy and often poorly tolerated. The use of hypertonic saline is generally not recommended unless severe neurologic symptoms are present as it leads to increased ascites and edema. When administered, it is important to avoid a rapid correction of the hyponatremia to prevent the development of central pontine myelinolysis and the osmotic demyelination syndrome.

Due to the pivotal role of AVP in the pathogenesis of DH, antagonism of its action has long been proposed to be the most rational approach, but until recently, effective and specific antagonism of AVP has remained elusive. Approaches that have been attempted include interference with its secretion and actions. Intravenous albumin has been reported to improve hyponatremia in patients with cirrhosis, ascites, and hyponatremia, presumably by decreasing AVP release by plasma volume expansion.50 An attempt at inhibition of central AVP release with the use of a kappa‐opioid receptor agonist, niravoline, was limited by loss of efficacy and potential adverse effects.51 Use of demeclocycline and lithium (which induce renal resistance to AVP and lead to a modest increase in urine volume with decreased urine osmolality and a corresponding rise in serum sodium) is limited by nephrotoxicity and hepatotoxicity.7, 52 Because of the important role played by prostaglandins in the maintenance of renal hemodynamics and water excretion in cirrhosis, oral misoprostol has also been evaluated but determined to be ineffective in inducing significant changes in free water clearance in patients with functional renal failure and/or DH.53

The recent introduction of vaptans, vasopressin receptor antagonists that block the physiologic action of vasopressin, represents a revolutionary and highly effective approach to the treatment of hyponatremia. Vaptans are antagonists of the V2 receptors of AVP in the principal cells of the collecting ducts. In healthy subjects, vaptans cause a dose‐dependent increase in urine volume and produce a dilute urine without causing natriuresis. To date, 2 AVP antagonists, conivaptan and tolvaptan, have been Food and Drug Administration (FDA)‐approved for the treatment of DH. Conivaptan, the first to be approved in 2005, is a mixed vasopressin V1a and V2 receptor antagonist that is administered intravenously for up to 4 days. In a randomized placebo‐controlled study of patients with euvolemic or hypervolemic hyponatremia, intravenous conivaptan treatment increased serum Na levels by >6 mEq/l or to a serum Na >135 mEq/l in 69 to 88.5% of subjects compared to 20.7% of those receiving placebo (Zeltser D, Rosansky S, Van Rensburg H, et al. Assessment of the efficacy and safety of intravenous conivaptan in euvolemic and hypervolemic hyponatremia. American J Nephrology 2007;27:447457). In a pilot study involving 24 patients with end‐stage liver disease, an infusion of conivaptan over 1 to 4 days was associated with an increase of serum sodium by >5 mmol/L in 60% of patients not receiving diuretics and 67% of patients on concomitant diuretic therapy by the end of treatment (O'Leary and Davis, 2009). Despite a concern about the potential for conivaptan to increase portal hypertension due to inhibition of splanchnic V1a receptors, the brief treatment appeared to be well tolerated without significant changes in systolic blood pressure, serum creatinine, variceal bleeding or worsening of ascites during the infusion period. However, approval for only 4 days of therapy and requirement for intravenous use eliminate any potential for chronic use.54

Tolvaptan is an orally available, selective V2 receptor antagonist whose efficacy was assessed in two multicenter, prospective, randomized, placebo‐controlled trials, Study of Ascending Levels of Tolvaptan in Hyponatremia 1 and 2 (SALT‐1 and SALT‐2).55 In these trials, clinically stable patients with DH (Na <135meq/l) associated with cirrhosis (22.4% in SALT‐1, 30.5% in SALT‐2), CHF or SIADH were randomized in a hospital setting to receive tolvaptan 15mg daily or placebo. Repeat Na levels were obtained at 8 hours, 2, 3, and 4 days and then weekly at days 11, 18, 25 and 30 after which study drug was discontinued and follow‐up Na level was determined 7 days later. The dose was adjusted to 30 mg and then 60 mg in an attempt to achieve a Na level >135 in those in whom hyponatremia persisted. During the initial day of the titration phase, fluid restriction was not maintained, and the patients were encouraged to respond to thirst with increased water ingestion.

Tolvaptan use was associated with a prompt increase in Na level as early as 8 hours after administration of the first dose. Serum Na increased more among those receiving tolvaptan than among those receiving placebo during the first 4 days and throughout the study period regardless of baseline Na level but returned to baseline within 1 week after discontinuation (Figure 3). The main side effects were increased thirst, dry mouth and increased urination. Importantly, an increased incidence of renal failure was not observed. Based on these results, FDA approval for tolvaptan in patients with hyponatremia was obtained in May 2009 for patients with DH‐associated with cirrhosis, CHF or SIADH for patients with Na levels <125 or symptomatic patients with Na levels between 125 and 135 that have not responded to fluid restriction.

Evaluation

Determination of volume status is paramount but frequently problematic in the hospitalized cirrhotic patient. Due to the vasodilated state present in severe portal hypertension that is characterized by a relative hypotension and resting tachycardia, the usual hemodynamic parameters of blood pressure and heart rate can be difficult to interpret. Although significant extravascular volume in the form of ascites and edema may be present, patients may be intravascularly depleted due to previous diuretic use and extra‐renal losses due to impaired oral intake, vomiting, lactulose‐induced diarrhea, and gastrointestinal bleeding. Infection is a commonly associated condition, and endotoxin mediated splanchnic vasodilatation, especially in the setting of SBP, can adversely effective central blood volume status in the presence of severe ascites. Also, due to the Na avidity of the kidney and previous diuretic use, renal electrolytes can be difficult to interpret.

For patients in whom there is strong clinical concern about intravascular depletion (history of impaired oral intake, excessive vomiting and/or diarrhea, rapid weight loss, small volume ascites with history of large volume, azotemia), administration of limited intravenous normal saline (0.5‐1 L) should be considered. Patients with severe neurological symptoms should receive normal saline or hypertonic saline. Unless severe neurologic symptoms associated with profound hyponatremia is present, however, intravenous normal saline should not be administered for the hyponatremia alone. Administration of salt poor albumin (25%), especially for those with marked fluid overload and ascites, is an effective means to expand the central blood volume without exacerbating ascites and edema. After evaluation for and/or treatment of hypovolemia, all patients should receive a Na restricted diet (<2000 mg daily) and placed on fluid restriction (see below for liberalization of fluid restriction upon initiation of tolvaptan therapy).

Diagnostic paracentesis should be performed for those with ascites to rule out the presence of SBP, and antibiotics administered to those with evidence of infection. High dose intravenous salt poor albumin should also be administered, especially to those at high risk of renal failure as determined by the presence of azotemia (Cr > 1.0 mg/dL) or severe liver insufficiency (TBili > 4.0 mg/dL).56, 57 Finally, all medications should be reviewed, and those associated with hyponatremia (diuretics, selective serotonin reuptake inhibitors, opiates, proton‐pump inhibitors) discontinued if possible.

Tolvaptan for DH

Post‐Operative Hyponatremia

The postoperative patient is particularly prone to developing hyponatremia. AVP levels are increased for several days following surgical procedures due to baroreceptor and nonbaroreceptor‐mediated mechanisms. These patients typically have subtle or overt decreases in effective arterial blood volume due to prolonged preoperative fasting combined with intraoperative and postoperative blood loss and third spacing of fluid. In addition to these factors which unload baroreceptors, postoperative pain, stress, anxiety, nausea, and administration of morphine can further stimulate the release of AVP. In some instances nonsteroidal antiinflmmatory drugs are given which have the effect of augmenting the hydroosmotic actions of AVP.6 In this setting of compromised ability to excrete a water load, administration of hypotonic fluid can precipitate acute iatrogenic hyponatremia.

Postoperative hyponatremia has been a major problem in pediatric populations due to the widespread practice of using hypotonic fluids for maintenance therapy. This approach is based on guidelines developed 50 years ago which were based on calculations linking energy expenditures to water and electrolyte losses.7 More recently, several groups have argued that isotonic rather than hypotonic fluids should be the routine maintenance fluid in such patients.8, 9 The pediatric community has been slow to embrace this approach out of the concern that excessive administration of Na⁺ would increase the risk of hypernatremia. A systematic meta‐analysis of studies comparing isotonic and hypotonic fluids in hospitalized children found the odds of developing hyponatremia following hypotonic solutions was 17.2 times greater than with isotonic fluids.10 The concern that isotonic maintenance fluids carry a risk of hypernatremia was not supported in the review. Some studies actually reported a decrease in serum Na⁺ concentration, presumably due to the desalination phenomenon in which hypertonic urine is excreted in volume expanded subjects with persistent vasopressin secretion.

Endocrine Disorders

Heart Failure

Hyponatremia is a common complication of left‐sided heart failure and several studies have shown that it is an independent predictor of mortality.13, 14 A similar association between reduced survival and hyponatremia is present in advanced right‐sided heart failure in patients with pulmonary arterial hypertension.15 Patients with heart failure who are hyponatremic have higher circulating levels of neurohormones (catecholamines, renin, angiotensin II, aldosterone, and AVP) than normonatremic subjects and are more likely to have prerenal azotemia. In addition to being a marker for the extent of neurohumoral activation, hyponatremia may play a more direct role in adverse outcomes through maladaptive volume regulatory responses of cardiac myocytes and by direct effects of AVP on cardiac and coronary V_1a receptors.

Heart failure is associated with arterial underfilling leading to arterial baroreceptor‐mediated activation of the neurohumoral axis. This underfilling is due to decreased cardiac output in low‐output heart failure and decreased systemic vascular resistance in high‐output heart failure. Activation of the sympathetic nervous system along with the renin‐angiotensin‐aldosterone system leads to renal salt retention while the increase in AVP is associated with water retention and hyponatremia.

Cirrhosis

Hyponatremia is a common electrolyte abnormality in patients with cirrhosis and occurs with a frequency that tends to parallel the severity of liver disease.16 Patients with a serum Na⁺ 130 mEq/L are more likely to have refractory ascites and require therapeutic paracentesis. Hepatic encephalopathy, hepatorenal syndrome, and spontaneous bacterial peritonitis are also more common in patients with a serum Na⁺ 130 mmol/L than in patients with a normal serum Na⁺ concentration. Hyponatremia increases morbidity and mortality from hepatic transplantation and is associated with osmotic demyelination in the postoperative period because of the large increase in Na⁺ concentration associated with the procedure.17

Hepatic cirrhosis is characterized by a decreased effective arterial blood volume and activation of neurohumoral effectors. Reduced effective circulating volume due to generalized and specifically to arterial splanchnic vasodilation leads to baroreceptor‐mediated nonosmotic stimulation of AVP release and an impaired ability to excrete electrolyte‐free water.18 Reduced Na⁺ delivery to the distal tubule because of a low glomerular filtration rate and increased proximal Na⁺ reabsorption adds to the susceptibility of cirrhotic patients to hyponatremia.

Hyponatremic‐Hypertensive Syndrome

The development of hyponatremia in patients with severe hypertension associated with renal artery stenosis has been called the hyponatremic‐hypertensive syndrome.19, 20 Patients with this syndrome present with a variety of signs and symptoms that include headache, confusion, postural dizziness, polyuria, polydipsia, and salt craving. In a retrospective review of 32 patients with this syndrome, most of the subjects were thin, elderly, women smokers who had atherosclerotic renal vascular disease.19 Biochemical abnormalities included not only hyponatremia, but hypokalemia and increased plasma renin activity. The mean serum Na⁺ concentration was 129.7 mmol/L (range 120‐135).

The precise mechanism of this syndrome is not known. Given the available data, angiotensin‐mediated thirst coupled with nonosmotic release of AVP provoked by angiotensin II and/or hypertensive encephalopathy are likely. Na⁺ depletion due to pressure natriuresis, and K⁺ depletion due to hyperaldosteronism are also likely to play a role in the pathogenesis of hyponatremia.

Pneumonia

An association between pneumonia and hyponatremia has been known for quite some time but has been poorly defined. Of the various etiologic agents, legionella is more commonly associated with hyponatremia as compared to other types of community acquired disease. As has been true for many disorders, hyponatremia is associated with longer hospital stays and hospital mortality, most likely reflecting the severity of the pneumonia rather than morbidity from the usually mild and asymptomatic hyponatremia.

Central Nervous System Disease

Hyponatremia is a frequent complication of central nervous system disease to include bacterial meningitis and traumatic brain injury. Potential mechanisms include the development of SIADH, cerebral salt wasting (CSW), or hypotonic fluid administration in the setting of impaired renal water excretion as in patients with low effective volume. Of these various causes, hyponatremia is frequently attributed to SIADH. As previously mentioned, this syndrome is characterized by hyponatremia in the setting of an inappropriately concentrated urine, increased urine Na⁺ concentration, and evidence of normal or slightly increased intravascular volume.

However there are patients with intracranial disease who develop hyponatremia with similar characteristics but differ in that there is clinical evidence of a contracted ECF volume. This form of hyponatremia is due to excessive renal Na⁺ excretion resulting from a centrally mediated process and is termed CSW. The onset of this disorder is typically seen within the first ten days following a neurosurgical procedure or after a definable event, such as a subarachnoid hemorrhage or stroke. CSW has also been described in other intracranial disorders, such as carcinomatous or infectious meningitis and metastatic carcinoma.21 The distinction between SIADH and CSW is of considerable clinical importance given the divergent nature of the treatments. Fluid restriction is the treatment of choice in SIADH, whereas the treatment of CSW comprises vigorous Na⁺ and volume replacement.

Post‐Operative Hyponatremia

The postoperative patient is particularly prone to developing hyponatremia. AVP levels are increased for several days following surgical procedures due to baroreceptor and nonbaroreceptor‐mediated mechanisms. These patients typically have subtle or overt decreases in effective arterial blood volume due to prolonged preoperative fasting combined with intraoperative and postoperative blood loss and third spacing of fluid. In addition to these factors which unload baroreceptors, postoperative pain, stress, anxiety, nausea, and administration of morphine can further stimulate the release of AVP. In some instances nonsteroidal antiinflmmatory drugs are given which have the effect of augmenting the hydroosmotic actions of AVP.6 In this setting of compromised ability to excrete a water load, administration of hypotonic fluid can precipitate acute iatrogenic hyponatremia.

Postoperative hyponatremia has been a major problem in pediatric populations due to the widespread practice of using hypotonic fluids for maintenance therapy. This approach is based on guidelines developed 50 years ago which were based on calculations linking energy expenditures to water and electrolyte losses.7 More recently, several groups have argued that isotonic rather than hypotonic fluids should be the routine maintenance fluid in such patients.8, 9 The pediatric community has been slow to embrace this approach out of the concern that excessive administration of Na⁺ would increase the risk of hypernatremia. A systematic meta‐analysis of studies comparing isotonic and hypotonic fluids in hospitalized children found the odds of developing hyponatremia following hypotonic solutions was 17.2 times greater than with isotonic fluids.10 The concern that isotonic maintenance fluids carry a risk of hypernatremia was not supported in the review. Some studies actually reported a decrease in serum Na⁺ concentration, presumably due to the desalination phenomenon in which hypertonic urine is excreted in volume expanded subjects with persistent vasopressin secretion.

Endocrine Disorders

Heart Failure

Hyponatremia is a common complication of left‐sided heart failure and several studies have shown that it is an independent predictor of mortality.13, 14 A similar association between reduced survival and hyponatremia is present in advanced right‐sided heart failure in patients with pulmonary arterial hypertension.15 Patients with heart failure who are hyponatremic have higher circulating levels of neurohormones (catecholamines, renin, angiotensin II, aldosterone, and AVP) than normonatremic subjects and are more likely to have prerenal azotemia. In addition to being a marker for the extent of neurohumoral activation, hyponatremia may play a more direct role in adverse outcomes through maladaptive volume regulatory responses of cardiac myocytes and by direct effects of AVP on cardiac and coronary V_1a receptors.

Heart failure is associated with arterial underfilling leading to arterial baroreceptor‐mediated activation of the neurohumoral axis. This underfilling is due to decreased cardiac output in low‐output heart failure and decreased systemic vascular resistance in high‐output heart failure. Activation of the sympathetic nervous system along with the renin‐angiotensin‐aldosterone system leads to renal salt retention while the increase in AVP is associated with water retention and hyponatremia.

Cirrhosis

Hyponatremia is a common electrolyte abnormality in patients with cirrhosis and occurs with a frequency that tends to parallel the severity of liver disease.16 Patients with a serum Na⁺ 130 mEq/L are more likely to have refractory ascites and require therapeutic paracentesis. Hepatic encephalopathy, hepatorenal syndrome, and spontaneous bacterial peritonitis are also more common in patients with a serum Na⁺ 130 mmol/L than in patients with a normal serum Na⁺ concentration. Hyponatremia increases morbidity and mortality from hepatic transplantation and is associated with osmotic demyelination in the postoperative period because of the large increase in Na⁺ concentration associated with the procedure.17

Hepatic cirrhosis is characterized by a decreased effective arterial blood volume and activation of neurohumoral effectors. Reduced effective circulating volume due to generalized and specifically to arterial splanchnic vasodilation leads to baroreceptor‐mediated nonosmotic stimulation of AVP release and an impaired ability to excrete electrolyte‐free water.18 Reduced Na⁺ delivery to the distal tubule because of a low glomerular filtration rate and increased proximal Na⁺ reabsorption adds to the susceptibility of cirrhotic patients to hyponatremia.

Hyponatremic‐Hypertensive Syndrome

The development of hyponatremia in patients with severe hypertension associated with renal artery stenosis has been called the hyponatremic‐hypertensive syndrome.19, 20 Patients with this syndrome present with a variety of signs and symptoms that include headache, confusion, postural dizziness, polyuria, polydipsia, and salt craving. In a retrospective review of 32 patients with this syndrome, most of the subjects were thin, elderly, women smokers who had atherosclerotic renal vascular disease.19 Biochemical abnormalities included not only hyponatremia, but hypokalemia and increased plasma renin activity. The mean serum Na⁺ concentration was 129.7 mmol/L (range 120‐135).

The precise mechanism of this syndrome is not known. Given the available data, angiotensin‐mediated thirst coupled with nonosmotic release of AVP provoked by angiotensin II and/or hypertensive encephalopathy are likely. Na⁺ depletion due to pressure natriuresis, and K⁺ depletion due to hyperaldosteronism are also likely to play a role in the pathogenesis of hyponatremia.

Pneumonia

An association between pneumonia and hyponatremia has been known for quite some time but has been poorly defined. Of the various etiologic agents, legionella is more commonly associated with hyponatremia as compared to other types of community acquired disease. As has been true for many disorders, hyponatremia is associated with longer hospital stays and hospital mortality, most likely reflecting the severity of the pneumonia rather than morbidity from the usually mild and asymptomatic hyponatremia.

Central Nervous System Disease

Hyponatremia is a frequent complication of central nervous system disease to include bacterial meningitis and traumatic brain injury. Potential mechanisms include the development of SIADH, cerebral salt wasting (CSW), or hypotonic fluid administration in the setting of impaired renal water excretion as in patients with low effective volume. Of these various causes, hyponatremia is frequently attributed to SIADH. As previously mentioned, this syndrome is characterized by hyponatremia in the setting of an inappropriately concentrated urine, increased urine Na⁺ concentration, and evidence of normal or slightly increased intravascular volume.

However there are patients with intracranial disease who develop hyponatremia with similar characteristics but differ in that there is clinical evidence of a contracted ECF volume. This form of hyponatremia is due to excessive renal Na⁺ excretion resulting from a centrally mediated process and is termed CSW. The onset of this disorder is typically seen within the first ten days following a neurosurgical procedure or after a definable event, such as a subarachnoid hemorrhage or stroke. CSW has also been described in other intracranial disorders, such as carcinomatous or infectious meningitis and metastatic carcinoma.21 The distinction between SIADH and CSW is of considerable clinical importance given the divergent nature of the treatments. Fluid restriction is the treatment of choice in SIADH, whereas the treatment of CSW comprises vigorous Na⁺ and volume replacement.

Post‐Operative Hyponatremia

The postoperative patient is particularly prone to developing hyponatremia. AVP levels are increased for several days following surgical procedures due to baroreceptor and nonbaroreceptor‐mediated mechanisms. These patients typically have subtle or overt decreases in effective arterial blood volume due to prolonged preoperative fasting combined with intraoperative and postoperative blood loss and third spacing of fluid. In addition to these factors which unload baroreceptors, postoperative pain, stress, anxiety, nausea, and administration of morphine can further stimulate the release of AVP. In some instances nonsteroidal antiinflmmatory drugs are given which have the effect of augmenting the hydroosmotic actions of AVP.6 In this setting of compromised ability to excrete a water load, administration of hypotonic fluid can precipitate acute iatrogenic hyponatremia.

Postoperative hyponatremia has been a major problem in pediatric populations due to the widespread practice of using hypotonic fluids for maintenance therapy. This approach is based on guidelines developed 50 years ago which were based on calculations linking energy expenditures to water and electrolyte losses.7 More recently, several groups have argued that isotonic rather than hypotonic fluids should be the routine maintenance fluid in such patients.8, 9 The pediatric community has been slow to embrace this approach out of the concern that excessive administration of Na⁺ would increase the risk of hypernatremia. A systematic meta‐analysis of studies comparing isotonic and hypotonic fluids in hospitalized children found the odds of developing hyponatremia following hypotonic solutions was 17.2 times greater than with isotonic fluids.10 The concern that isotonic maintenance fluids carry a risk of hypernatremia was not supported in the review. Some studies actually reported a decrease in serum Na⁺ concentration, presumably due to the desalination phenomenon in which hypertonic urine is excreted in volume expanded subjects with persistent vasopressin secretion.

Endocrine Disorders

Heart Failure

Hyponatremia is a common complication of left‐sided heart failure and several studies have shown that it is an independent predictor of mortality.13, 14 A similar association between reduced survival and hyponatremia is present in advanced right‐sided heart failure in patients with pulmonary arterial hypertension.15 Patients with heart failure who are hyponatremic have higher circulating levels of neurohormones (catecholamines, renin, angiotensin II, aldosterone, and AVP) than normonatremic subjects and are more likely to have prerenal azotemia. In addition to being a marker for the extent of neurohumoral activation, hyponatremia may play a more direct role in adverse outcomes through maladaptive volume regulatory responses of cardiac myocytes and by direct effects of AVP on cardiac and coronary V_1a receptors.

Heart failure is associated with arterial underfilling leading to arterial baroreceptor‐mediated activation of the neurohumoral axis. This underfilling is due to decreased cardiac output in low‐output heart failure and decreased systemic vascular resistance in high‐output heart failure. Activation of the sympathetic nervous system along with the renin‐angiotensin‐aldosterone system leads to renal salt retention while the increase in AVP is associated with water retention and hyponatremia.

Cirrhosis

Hyponatremia is a common electrolyte abnormality in patients with cirrhosis and occurs with a frequency that tends to parallel the severity of liver disease.16 Patients with a serum Na⁺ 130 mEq/L are more likely to have refractory ascites and require therapeutic paracentesis. Hepatic encephalopathy, hepatorenal syndrome, and spontaneous bacterial peritonitis are also more common in patients with a serum Na⁺ 130 mmol/L than in patients with a normal serum Na⁺ concentration. Hyponatremia increases morbidity and mortality from hepatic transplantation and is associated with osmotic demyelination in the postoperative period because of the large increase in Na⁺ concentration associated with the procedure.17

Hepatic cirrhosis is characterized by a decreased effective arterial blood volume and activation of neurohumoral effectors. Reduced effective circulating volume due to generalized and specifically to arterial splanchnic vasodilation leads to baroreceptor‐mediated nonosmotic stimulation of AVP release and an impaired ability to excrete electrolyte‐free water.18 Reduced Na⁺ delivery to the distal tubule because of a low glomerular filtration rate and increased proximal Na⁺ reabsorption adds to the susceptibility of cirrhotic patients to hyponatremia.

Hyponatremic‐Hypertensive Syndrome

The development of hyponatremia in patients with severe hypertension associated with renal artery stenosis has been called the hyponatremic‐hypertensive syndrome.19, 20 Patients with this syndrome present with a variety of signs and symptoms that include headache, confusion, postural dizziness, polyuria, polydipsia, and salt craving. In a retrospective review of 32 patients with this syndrome, most of the subjects were thin, elderly, women smokers who had atherosclerotic renal vascular disease.19 Biochemical abnormalities included not only hyponatremia, but hypokalemia and increased plasma renin activity. The mean serum Na⁺ concentration was 129.7 mmol/L (range 120‐135).

The precise mechanism of this syndrome is not known. Given the available data, angiotensin‐mediated thirst coupled with nonosmotic release of AVP provoked by angiotensin II and/or hypertensive encephalopathy are likely. Na⁺ depletion due to pressure natriuresis, and K⁺ depletion due to hyperaldosteronism are also likely to play a role in the pathogenesis of hyponatremia.

Pneumonia

An association between pneumonia and hyponatremia has been known for quite some time but has been poorly defined. Of the various etiologic agents, legionella is more commonly associated with hyponatremia as compared to other types of community acquired disease. As has been true for many disorders, hyponatremia is associated with longer hospital stays and hospital mortality, most likely reflecting the severity of the pneumonia rather than morbidity from the usually mild and asymptomatic hyponatremia.

Central Nervous System Disease

Hyponatremia is a frequent complication of central nervous system disease to include bacterial meningitis and traumatic brain injury. Potential mechanisms include the development of SIADH, cerebral salt wasting (CSW), or hypotonic fluid administration in the setting of impaired renal water excretion as in patients with low effective volume. Of these various causes, hyponatremia is frequently attributed to SIADH. As previously mentioned, this syndrome is characterized by hyponatremia in the setting of an inappropriately concentrated urine, increased urine Na⁺ concentration, and evidence of normal or slightly increased intravascular volume.

However there are patients with intracranial disease who develop hyponatremia with similar characteristics but differ in that there is clinical evidence of a contracted ECF volume. This form of hyponatremia is due to excessive renal Na⁺ excretion resulting from a centrally mediated process and is termed CSW. The onset of this disorder is typically seen within the first ten days following a neurosurgical procedure or after a definable event, such as a subarachnoid hemorrhage or stroke. CSW has also been described in other intracranial disorders, such as carcinomatous or infectious meningitis and metastatic carcinoma.21 The distinction between SIADH and CSW is of considerable clinical importance given the divergent nature of the treatments. Fluid restriction is the treatment of choice in SIADH, whereas the treatment of CSW comprises vigorous Na⁺ and volume replacement.

Incidence of Hyponatremia in Heart Failure Patients

Determining the exact incidence of hyponatremia in heart failure patients has been challenging due to differences in the populations that have been studied and in the criteria used to define hyponatremia. The Acute and Chronic Therapeutic Impact of a Vasopressin Antagonist in Congestive Heart Failure (ACTIV in CHF) study reported that 21.3% of the cohort of hospitalized patients had a serum sodium below 136 mmol/L.17 Higher incidences were reported in the Prospective Trial of Intravenous Milrinone for Exacerbations of Chronic Heart Failure (OPTIME‐CHF) which found that 27% of patients had serum sodium concentrations between 132 mmol/L to 135 mmol/L.18 The Evaluation Study of CHF and Pulmonary Artery Catheter Effectiveness (ESCAPE) trial reported that 18% of patients had hyponatremia defined as serum sodium concentration below 134 mmol/L.19 Data from OPTIMIZE‐HF, a registry that captured data from a large cohort of representative patients who were hospitalized with decompensated heart failure indicate that such patients have a wide distribution of admission sodium values (Figure 1).16 Overall, 19.7% of the OPTIIMIZE‐HF patients had values below 135 mmol/L.

Figure 1

Risk Associated With Hyponatremia

There is considerable evidence that hyponatremia is associated with increased risk for poor outcomes in heart failure patients. One of the first reports that related hyponatremia to a poor prognosis came from Lee and Packer who analyzed 30 clinical, hemodynamic, and biochemical variables in outpatients with severe heart failure.20 Their results showed that serum sodium was the most powerful predictor of CV mortality. Similar findings have been reported by other investigators.21 In patients hospitalized for decompensated heart failure, the presence of hyponatremia has been shown to be an independent predictor of longer duration of stay.22 Results from the OPTIMIZE‐HF Registry confirmed the adverse impact of hyponatremia on length of stay during a heart failure hospitalization and also suggested that low serum sodium was associated with significantly higher in‐hospital and post‐discharge mortality rates.16 In this large cohort of hospitalized patients with decompensated heart failure, each 3 mmol/L decrease in serum sodium below 140 mmol/L increased the risk of in‐hospital and follow‐up mortality by 19.5% and 10%, respectively. This association is depicted in Figure 2. A similar adverse impact of hyponatremia on post‐discharge outcomes was seen in the results of the ACTIV in CHF study.17 Overall, 69 patients out of a total of 319 who were included in the study (21.6%) had a serum sodium that was <136 mmol/L. Mortality over a 60 day period of follow‐up was 14.5% in the hyponatremic patients compared to 4% in the 250 patients whose serum sodium was 136 mmol/L.

Figure 2

Although these data present a powerful argument that hyponatremia is a potent risk factor for poor outcomes in heart failure patients both during and following hospitalization, they do not determine whether or not the relationship is causal. It is possible that the poor outcomes that have been observed in hyponatremic patients are related to an association between low serum sodium levels and the more profound alterations in hemodynamics, neurohormonal activation and renal function that are seen in patients with advanced heart failure. Alternatively, increased edema in vital organs including the heart in hyponatremic patients could further impair already tenuous function and contribute to a downward spiral in the clinical course. The impact of hyponatremia in limiting the use of loop diuretics and spironolactone may also be associated with a less favorable long‐term outcome. The presence of hyponatremia may also contribute to a poor outcome as a result of effects on cognitive and neuromuscular function.23, 24 Impaired cognition could adversely affect compliance with the medical regimen while neuromuscular problems related to hyponatremia could contribute to an increased incidence of falls and other traumatic injuries that older patients with chronic diseases are already at high risk of experiencing.

Pathophysiology of Hyponatremia

The pathophysiology of hyponatremia in heart failure patients involves several different processes.25 Overall, heart failure is characterized by retention of excessive amounts of salt and water in the body. Sodium retention is related to decreased renal perfusion that is caused by the effects of reduced cardiac output, decreased renal perfusion pressure and increased afferent glomerular arteriolar resistance. A reduction in glomerular filtration and increased reabsorption of sodium and water in the proximal renal tubules leads to a reduction in the delivery of water and solute to the diluting segment of the nephron. In patients with heart failure the renin‐angiotensin system (RAS) is activated, an event that is further stimulated by the administration of loop diuretics.26 Angiotensin II (Ang II), a key effector molecule of the RAS, increases tone in renal efferent arterioles. This tends to enhance both sodium and water reabsorption both through an increase in the glomerular filtration fraction and by direct effects on the distal tubule.27 Ang II also stimulates the thirst center of the brain both directly and through stimulation of antidiuretic hormone to promote the ingestion of excessive amounts of hypotonic fluids.25 Water reabsorption in the distal portion of the nephron is governed by arginine vasopressin (AVP). High levels of AVP are seen in patients with heart failure26, 28 and there is a significant association between serum levels of this peptide and the symptomatic state of the patient. There is evidence that in heart failure patients AVP levels are elevated disproportionally to plasma osmolality and serum sodium concentrations.28, 29 Even when serum osmolality is increased in this setting by infusion of sodium, AVP levels fail to demonstrate an appropriate reduction suggesting that mechanisms other than activation of osmoreceptors are involved. The effects of AVP in the pathogenesis of hyponatremia are significant. This peptide binds to the vasopressin‐2 (V2) receptor in the collecting duct of the kidney stimulating an increase in the second messenger cyclic AMP.30 Downstream signaling initiated by cyclic AMP leads to an increase in the number and activation of aquaporin‐2 water channels on the luminal surface of epithelial cells in the collecting tubule.31 The presence of these activated pores is necessary for water permeability in the collecting duct and leads to an increase in the reabsorption of free water. Finally, the use of diuretics in the treatment of heart failure has been implicated in the development and worsening of the hyponatremic state.32

Treatment of Hyponatremia

Treatment options for dealing with hyponatremia in heart failure patients have been limited until recently. Since low cardiac output and/or diminished renal perfusion are involved, interventions which improve cardiac and renal function can reverse hyponatremia. While this can be accomplished by the use of inotropic agents, the use of drugs such as dobutamine, milrinone, and other inotropes in either stable or decompensated heart failure patients with adequate tissue perfusion is not routinely recommended due to a well‐documented increase in adverse effects, particularly in patients with coronary artery disease.33 The use of hypertonic saline is also not recommended since it may worsen the extent of volume overload in decompensated patients. Fluid restriction can be used to treat hyponatremia, particularly in hospitalized patients where stricter control on input is possible. Hyponatremic patients, however, often experience excessive thirst and restriction to less than 1000 cc to 1500 cc is rarely, if ever, successful for more than a brief period of time. The use of angiotensin converting enzyme (ACE) inhibitors has been associated with an improvement in serum sodium levels. Packer et al. treated a cohort of heart failure patients who were receiving a stable dose of diuretic with an ACE inhibitor captopril and found that over a 2‐week period the serum sodium increased from 131.2 0.5 to 135.9 0.5 mmol/L; P < 0.001).34 These investigators concluded that the RAS was involved in the pathogenesis of hyponatremia and that ACEIs increased sodium levels in hyponatremic patients, though the mechanism through which this occurs has not been delineated.

Incidence of Hyponatremia in Heart Failure Patients

Determining the exact incidence of hyponatremia in heart failure patients has been challenging due to differences in the populations that have been studied and in the criteria used to define hyponatremia. The Acute and Chronic Therapeutic Impact of a Vasopressin Antagonist in Congestive Heart Failure (ACTIV in CHF) study reported that 21.3% of the cohort of hospitalized patients had a serum sodium below 136 mmol/L.17 Higher incidences were reported in the Prospective Trial of Intravenous Milrinone for Exacerbations of Chronic Heart Failure (OPTIME‐CHF) which found that 27% of patients had serum sodium concentrations between 132 mmol/L to 135 mmol/L.18 The Evaluation Study of CHF and Pulmonary Artery Catheter Effectiveness (ESCAPE) trial reported that 18% of patients had hyponatremia defined as serum sodium concentration below 134 mmol/L.19 Data from OPTIMIZE‐HF, a registry that captured data from a large cohort of representative patients who were hospitalized with decompensated heart failure indicate that such patients have a wide distribution of admission sodium values (Figure 1).16 Overall, 19.7% of the OPTIIMIZE‐HF patients had values below 135 mmol/L.

Figure 1

Risk Associated With Hyponatremia

There is considerable evidence that hyponatremia is associated with increased risk for poor outcomes in heart failure patients. One of the first reports that related hyponatremia to a poor prognosis came from Lee and Packer who analyzed 30 clinical, hemodynamic, and biochemical variables in outpatients with severe heart failure.20 Their results showed that serum sodium was the most powerful predictor of CV mortality. Similar findings have been reported by other investigators.21 In patients hospitalized for decompensated heart failure, the presence of hyponatremia has been shown to be an independent predictor of longer duration of stay.22 Results from the OPTIMIZE‐HF Registry confirmed the adverse impact of hyponatremia on length of stay during a heart failure hospitalization and also suggested that low serum sodium was associated with significantly higher in‐hospital and post‐discharge mortality rates.16 In this large cohort of hospitalized patients with decompensated heart failure, each 3 mmol/L decrease in serum sodium below 140 mmol/L increased the risk of in‐hospital and follow‐up mortality by 19.5% and 10%, respectively. This association is depicted in Figure 2. A similar adverse impact of hyponatremia on post‐discharge outcomes was seen in the results of the ACTIV in CHF study.17 Overall, 69 patients out of a total of 319 who were included in the study (21.6%) had a serum sodium that was <136 mmol/L. Mortality over a 60 day period of follow‐up was 14.5% in the hyponatremic patients compared to 4% in the 250 patients whose serum sodium was 136 mmol/L.

Figure 2

Although these data present a powerful argument that hyponatremia is a potent risk factor for poor outcomes in heart failure patients both during and following hospitalization, they do not determine whether or not the relationship is causal. It is possible that the poor outcomes that have been observed in hyponatremic patients are related to an association between low serum sodium levels and the more profound alterations in hemodynamics, neurohormonal activation and renal function that are seen in patients with advanced heart failure. Alternatively, increased edema in vital organs including the heart in hyponatremic patients could further impair already tenuous function and contribute to a downward spiral in the clinical course. The impact of hyponatremia in limiting the use of loop diuretics and spironolactone may also be associated with a less favorable long‐term outcome. The presence of hyponatremia may also contribute to a poor outcome as a result of effects on cognitive and neuromuscular function.23, 24 Impaired cognition could adversely affect compliance with the medical regimen while neuromuscular problems related to hyponatremia could contribute to an increased incidence of falls and other traumatic injuries that older patients with chronic diseases are already at high risk of experiencing.

Pathophysiology of Hyponatremia

The pathophysiology of hyponatremia in heart failure patients involves several different processes.25 Overall, heart failure is characterized by retention of excessive amounts of salt and water in the body. Sodium retention is related to decreased renal perfusion that is caused by the effects of reduced cardiac output, decreased renal perfusion pressure and increased afferent glomerular arteriolar resistance. A reduction in glomerular filtration and increased reabsorption of sodium and water in the proximal renal tubules leads to a reduction in the delivery of water and solute to the diluting segment of the nephron. In patients with heart failure the renin‐angiotensin system (RAS) is activated, an event that is further stimulated by the administration of loop diuretics.26 Angiotensin II (Ang II), a key effector molecule of the RAS, increases tone in renal efferent arterioles. This tends to enhance both sodium and water reabsorption both through an increase in the glomerular filtration fraction and by direct effects on the distal tubule.27 Ang II also stimulates the thirst center of the brain both directly and through stimulation of antidiuretic hormone to promote the ingestion of excessive amounts of hypotonic fluids.25 Water reabsorption in the distal portion of the nephron is governed by arginine vasopressin (AVP). High levels of AVP are seen in patients with heart failure26, 28 and there is a significant association between serum levels of this peptide and the symptomatic state of the patient. There is evidence that in heart failure patients AVP levels are elevated disproportionally to plasma osmolality and serum sodium concentrations.28, 29 Even when serum osmolality is increased in this setting by infusion of sodium, AVP levels fail to demonstrate an appropriate reduction suggesting that mechanisms other than activation of osmoreceptors are involved. The effects of AVP in the pathogenesis of hyponatremia are significant. This peptide binds to the vasopressin‐2 (V2) receptor in the collecting duct of the kidney stimulating an increase in the second messenger cyclic AMP.30 Downstream signaling initiated by cyclic AMP leads to an increase in the number and activation of aquaporin‐2 water channels on the luminal surface of epithelial cells in the collecting tubule.31 The presence of these activated pores is necessary for water permeability in the collecting duct and leads to an increase in the reabsorption of free water. Finally, the use of diuretics in the treatment of heart failure has been implicated in the development and worsening of the hyponatremic state.32

Treatment of Hyponatremia

Treatment options for dealing with hyponatremia in heart failure patients have been limited until recently. Since low cardiac output and/or diminished renal perfusion are involved, interventions which improve cardiac and renal function can reverse hyponatremia. While this can be accomplished by the use of inotropic agents, the use of drugs such as dobutamine, milrinone, and other inotropes in either stable or decompensated heart failure patients with adequate tissue perfusion is not routinely recommended due to a well‐documented increase in adverse effects, particularly in patients with coronary artery disease.33 The use of hypertonic saline is also not recommended since it may worsen the extent of volume overload in decompensated patients. Fluid restriction can be used to treat hyponatremia, particularly in hospitalized patients where stricter control on input is possible. Hyponatremic patients, however, often experience excessive thirst and restriction to less than 1000 cc to 1500 cc is rarely, if ever, successful for more than a brief period of time. The use of angiotensin converting enzyme (ACE) inhibitors has been associated with an improvement in serum sodium levels. Packer et al. treated a cohort of heart failure patients who were receiving a stable dose of diuretic with an ACE inhibitor captopril and found that over a 2‐week period the serum sodium increased from 131.2 0.5 to 135.9 0.5 mmol/L; P < 0.001).34 These investigators concluded that the RAS was involved in the pathogenesis of hyponatremia and that ACEIs increased sodium levels in hyponatremic patients, though the mechanism through which this occurs has not been delineated.

Incidence of Hyponatremia in Heart Failure Patients

Determining the exact incidence of hyponatremia in heart failure patients has been challenging due to differences in the populations that have been studied and in the criteria used to define hyponatremia. The Acute and Chronic Therapeutic Impact of a Vasopressin Antagonist in Congestive Heart Failure (ACTIV in CHF) study reported that 21.3% of the cohort of hospitalized patients had a serum sodium below 136 mmol/L.17 Higher incidences were reported in the Prospective Trial of Intravenous Milrinone for Exacerbations of Chronic Heart Failure (OPTIME‐CHF) which found that 27% of patients had serum sodium concentrations between 132 mmol/L to 135 mmol/L.18 The Evaluation Study of CHF and Pulmonary Artery Catheter Effectiveness (ESCAPE) trial reported that 18% of patients had hyponatremia defined as serum sodium concentration below 134 mmol/L.19 Data from OPTIMIZE‐HF, a registry that captured data from a large cohort of representative patients who were hospitalized with decompensated heart failure indicate that such patients have a wide distribution of admission sodium values (Figure 1).16 Overall, 19.7% of the OPTIIMIZE‐HF patients had values below 135 mmol/L.

Figure 1

Risk Associated With Hyponatremia

There is considerable evidence that hyponatremia is associated with increased risk for poor outcomes in heart failure patients. One of the first reports that related hyponatremia to a poor prognosis came from Lee and Packer who analyzed 30 clinical, hemodynamic, and biochemical variables in outpatients with severe heart failure.20 Their results showed that serum sodium was the most powerful predictor of CV mortality. Similar findings have been reported by other investigators.21 In patients hospitalized for decompensated heart failure, the presence of hyponatremia has been shown to be an independent predictor of longer duration of stay.22 Results from the OPTIMIZE‐HF Registry confirmed the adverse impact of hyponatremia on length of stay during a heart failure hospitalization and also suggested that low serum sodium was associated with significantly higher in‐hospital and post‐discharge mortality rates.16 In this large cohort of hospitalized patients with decompensated heart failure, each 3 mmol/L decrease in serum sodium below 140 mmol/L increased the risk of in‐hospital and follow‐up mortality by 19.5% and 10%, respectively. This association is depicted in Figure 2. A similar adverse impact of hyponatremia on post‐discharge outcomes was seen in the results of the ACTIV in CHF study.17 Overall, 69 patients out of a total of 319 who were included in the study (21.6%) had a serum sodium that was <136 mmol/L. Mortality over a 60 day period of follow‐up was 14.5% in the hyponatremic patients compared to 4% in the 250 patients whose serum sodium was 136 mmol/L.

Figure 2

Although these data present a powerful argument that hyponatremia is a potent risk factor for poor outcomes in heart failure patients both during and following hospitalization, they do not determine whether or not the relationship is causal. It is possible that the poor outcomes that have been observed in hyponatremic patients are related to an association between low serum sodium levels and the more profound alterations in hemodynamics, neurohormonal activation and renal function that are seen in patients with advanced heart failure. Alternatively, increased edema in vital organs including the heart in hyponatremic patients could further impair already tenuous function and contribute to a downward spiral in the clinical course. The impact of hyponatremia in limiting the use of loop diuretics and spironolactone may also be associated with a less favorable long‐term outcome. The presence of hyponatremia may also contribute to a poor outcome as a result of effects on cognitive and neuromuscular function.23, 24 Impaired cognition could adversely affect compliance with the medical regimen while neuromuscular problems related to hyponatremia could contribute to an increased incidence of falls and other traumatic injuries that older patients with chronic diseases are already at high risk of experiencing.

Pathophysiology of Hyponatremia

The pathophysiology of hyponatremia in heart failure patients involves several different processes.25 Overall, heart failure is characterized by retention of excessive amounts of salt and water in the body. Sodium retention is related to decreased renal perfusion that is caused by the effects of reduced cardiac output, decreased renal perfusion pressure and increased afferent glomerular arteriolar resistance. A reduction in glomerular filtration and increased reabsorption of sodium and water in the proximal renal tubules leads to a reduction in the delivery of water and solute to the diluting segment of the nephron. In patients with heart failure the renin‐angiotensin system (RAS) is activated, an event that is further stimulated by the administration of loop diuretics.26 Angiotensin II (Ang II), a key effector molecule of the RAS, increases tone in renal efferent arterioles. This tends to enhance both sodium and water reabsorption both through an increase in the glomerular filtration fraction and by direct effects on the distal tubule.27 Ang II also stimulates the thirst center of the brain both directly and through stimulation of antidiuretic hormone to promote the ingestion of excessive amounts of hypotonic fluids.25 Water reabsorption in the distal portion of the nephron is governed by arginine vasopressin (AVP). High levels of AVP are seen in patients with heart failure26, 28 and there is a significant association between serum levels of this peptide and the symptomatic state of the patient. There is evidence that in heart failure patients AVP levels are elevated disproportionally to plasma osmolality and serum sodium concentrations.28, 29 Even when serum osmolality is increased in this setting by infusion of sodium, AVP levels fail to demonstrate an appropriate reduction suggesting that mechanisms other than activation of osmoreceptors are involved. The effects of AVP in the pathogenesis of hyponatremia are significant. This peptide binds to the vasopressin‐2 (V2) receptor in the collecting duct of the kidney stimulating an increase in the second messenger cyclic AMP.30 Downstream signaling initiated by cyclic AMP leads to an increase in the number and activation of aquaporin‐2 water channels on the luminal surface of epithelial cells in the collecting tubule.31 The presence of these activated pores is necessary for water permeability in the collecting duct and leads to an increase in the reabsorption of free water. Finally, the use of diuretics in the treatment of heart failure has been implicated in the development and worsening of the hyponatremic state.32

Treatment of Hyponatremia

Treatment options for dealing with hyponatremia in heart failure patients have been limited until recently. Since low cardiac output and/or diminished renal perfusion are involved, interventions which improve cardiac and renal function can reverse hyponatremia. While this can be accomplished by the use of inotropic agents, the use of drugs such as dobutamine, milrinone, and other inotropes in either stable or decompensated heart failure patients with adequate tissue perfusion is not routinely recommended due to a well‐documented increase in adverse effects, particularly in patients with coronary artery disease.33 The use of hypertonic saline is also not recommended since it may worsen the extent of volume overload in decompensated patients. Fluid restriction can be used to treat hyponatremia, particularly in hospitalized patients where stricter control on input is possible. Hyponatremic patients, however, often experience excessive thirst and restriction to less than 1000 cc to 1500 cc is rarely, if ever, successful for more than a brief period of time. The use of angiotensin converting enzyme (ACE) inhibitors has been associated with an improvement in serum sodium levels. Packer et al. treated a cohort of heart failure patients who were receiving a stable dose of diuretic with an ACE inhibitor captopril and found that over a 2‐week period the serum sodium increased from 131.2 0.5 to 135.9 0.5 mmol/L; P < 0.001).34 These investigators concluded that the RAS was involved in the pathogenesis of hyponatremia and that ACEIs increased sodium levels in hyponatremic patients, though the mechanism through which this occurs has not been delineated.