Congestive Heart Failure

Hyponatremia in heart failure stems from reduced cardiac output and blood pressure which stimulate vasopressin, catecholamines, and the renin–angiotensin–aldosterone axis. Increased vasopressin levels have even been documented in patients with impaired left ventricular function before the onset of symptomatic heart failure. Hyponatremic patients with congestive heart failure have higher levels of plasma renin activity, norepinephrine, epinephrine, and lower renal and hepatic plasma flows than normonatremic patients with an apparently similar degree of heart disease. Hyponatremia in heart failure is associated with a poor prognosis.

Hepatic Cirrhosis

Cirrhosis is characterized by a hyperdynamic circulation with low blood pressure, low systemic vascular resistance, and high cardiac output. Systemic vasodilatation causes relative underfilling of the arterial vascular compartment and neurohumoral responses similar to those that occur in response to a low cardiac output. Activation of the renin–angiotensin–aldosterone axis and the sympathetic system, combined with non-osmotic release of vasopressin, results in renal water and sodium retention. Escape from the sodium-retaining effect of aldosterone does not occur, and there is renal resistance to atrial natriuretic peptide. Although the pathogenesis of the peripheral arterial vasodilation is incompletely understood, increased vascular nitric oxide by the endothelium may play a role. In a rat model of cirrhosis, normalization of vascular nitric oxide production with a nitric oxide synthetase inhibitor corrects the hyperdynamic circulation, improves sodium and water excretion, and decreases neurohumoral activation.

Peritoneovenous shunting of hyponatremic cirrhotic patients with refractory ascites improves cardiac output, renal plasma flow, and creatinine clearance, and results in an immediate diuresis and natriuresis with a decrease in urine osmolality and an increase in plasma sodium concentration associated with a small but significant decrease in plasma vasopressin levels.

Definitions

Non-osmotic release of vasopressin without a hemodynamic stimulus to account for it is considered “inappropriate”. When Bartter and Schwarz first described the syndrome of inappropriate antidiuretic hormone secretion (SIADH), they defined clinical criteria for the disorder which are still generally accepted: hypoosmolality and clinical euvolemia with a sodium-containing urine (>30 mmol/L) that is less than maximally dilute (>100 mOsm/kg) without recent diuretic use or impaired renal function. Schwarz and Bartter also excluded endocrine disorders–primary and secondary adrenal insufficiency and hypothyroidism–from this designation. We have not made this exclusion, because patients with undiagnosed endocrine disturbances may present with all the clinical features of SIADH. Indeed, the discovery of SIADH is often the presenting feature of a clinically important systemic disease. Abnormal vasopressin secretion may be caused by ectopic production of the hormone by tumors, from disordered secretion by the neurohypophysis or from increased sensitivity to the hormone ( Table 44.3 ).

Patients with the syndrome of inappropriate antidiuretic hormone secretion (SIADH) retain ingested water, but they have no evidence of volume depletion and no tendency to form edema. Because of water retention, SIADH causes mild, subclinical volume-expansion, which is reflected by high uric acid clearance, a low plasma uric acid concentration, and urine sodium excretion which matches or exceeds sodium intake. Clinicians make use of these characteristics to distinguish SIADH from hyponatremia caused by volume-depletion. The recently described syndrome of nephrogenic inappropriate antidiuresis (see section “Common Causes of SIADH”) is discussed in this section because it exhibits clinical features of SIADH; in physiological terms, however, it is a cause of hyponatremia that is independent of vasopressin, as plasma vasopressin levels are undetectable ( Table 44.2 ).

Patterns of Vasopressin Secretion

In most patients with SIADH, vasopressin secretion has followed one of two basic patterns: “reset osmostat” or “vasopressin leak”. In the reset osmostat variant of SIADH, seen in patients with chronic, debilitating illness and in normal pregnancy, the urine can be diluted maximally, but at a lower set-point than normal. Such patients are thus mildly hyponatremic, but unlike other patients with SIADH, their plasma sodium concentration is very stable and they do not require dietary water restriction or other measures used to treat chronic hyponatremia. In the vasopressin leak variant, the basal level of vasopressin is elevated and unresponsive to osmotic stimuli when the plasma osmolality is low, but the levels increase appropriately when the plasma osmolality increases above a threshold level. Less commonly, patients exhibit erratic vasopressin secretion which is unrelated to osmotic stimuli. In about 10% of patients who present with typical clinical manifestations of SIADH, plasma vasopressin levels are at a low basal level that fails to increase as plasma osmolality increases. Such a pattern would be expected if an antidiuretic factor other than vasopressin were produced or if the collecting tubules were hypersensitive to normal hormone levels.

Interplay of Water Retention and Cation Depletion in SIADH

In SIADH, increased intravascular volume decreases renin secretion and increases release of atrial natriuretic peptide. These volume and hormonal changes promote sodium excretion, despite a low serum sodium concentration. Natriuresis in SIADH blunts the increase in extracellular volume caused by water retention ( Figure 44.1 ), but it also exacerbates hyponatremia ( Eq. (44.1) ). Balance studies in a group of patients with SIADH showed that during a period of high water and low sodium intake, the plasma sodium concentration decreased by 8 mEq/L with no gain in weight and no increase in chloride space (a measure of extracellular fluid volume). Negative sodium and potassium balances accounted for the stability of extracellular volume, and for over 80% of the calculated solute loss. During a period of high sodium intake, over 600 mEq of sodium was retained (with only a small increase in weight and chloride space), fully accounting for the 11 mEq/L increase in serum sodium concentration. In this study, there was a strong negative correlation between water intake and sodium balance, and between water intake and aldosterone secretory rate. Similar findings have been reported in studies in which pituitary extract was administered chronically to normal subjects.

In an experimental model of SIADH produced by DDAVP and half-isotonic saline in the rat, hyponatremia was caused exclusively by negative balances of sodium and potassium; water balance, which was slightly negative, did not contribute to the decrease in sodium concentration. Despite negative balances for sodium, chloride, and water, the extracellular volume (measured by inulin space) was not contracted, suggesting that water shifted from the intracellular to the extracellular space in response to the loss of intracellular solute (potassium and phosphate).

Similarly, direct measurements of body composition in a rat model of chronic SIADH showed that after 14 days of severe hyponatremia, body water content had returned to control levels ( Figure 44.3 ). Body sodium and chloride levels were reduced after one day of hyponatremia and were sustained for 14 days, and body potassium was significantly decreased after 7 days. Acutely, water retention was the major cause of hyponatremia, but solute depletion was primarily responsible when the electrolyte disturbance was sustained.

Body water content in experimental SIADH.

Panel (a) depicts measurements of body water content in rats made hyponatremic with dDAVP and a liquid diet. The open bar at the far left of the figure represents measurements obtained in normonatremic controls given no dDAVP. The black bars represent measurements obtained in animals made severely hyponatremic (plasma sodium 106 to 112 mmol/l) by giving dDAVP at 5 ng/hr; the gray bars represent measurements obtained in animals made less hyponatremic (plasma sodium 119 to 124 mmol/l) by giving dDAVP at 1 ng/hr. Body water content initially increases after the first day of hyponatremia at both doses of dDAVP, but falls to control levels after 14 days of severe hyponatremia. Panel (b) depicts whole body Na, K, and Cl measurements in the same experiment. Cation and chloride losses occur, beginning on the first day of hyponatremia, and are more severe in animals given the higher dose of dDAVP (with lower plasma sodium levels). Depletion of Na and Cl most likely represent adaptive responses to extracellular fluid volume expansion caused by retained water. Depletion of body K most likely represents a cell volume adaptive response.

Urinary losses can directly lower the plasma sodium concentration when the concentration of sodium plus potassium in the urine is higher than the plasma sodium concentration. This can occur when high vasopressin levels (which concentrate the urine) and high rates of sodium and potassium excretion occur together. The excretion of hypertonic urine generates free water, in essence “desalinating” the plasma.

Serum Bicarbonate Concentration in SIADH

In SIADH, the serum sodium and chloride concentrations are lowered by dilution, but the serum bicarbonate concentration is typically normal. This finding has been explained by a direct effect of hyponatremia on the adrenal gland to increase aldosterone secretion, which then augments renal net acid excretion. Patients with hyponatremia due to hypopituitarism have many features in common with patients with non-endocrine SIADH, but their serum bicarbonate concentrations are about 5 mmol/l lower. Consistent with the hypothesis that hyponatremia-induced hyperaldosteronism is responsible for the normal serum bicarbonate in classic SIADH, aldosterone levels are much lower in patients with ACTH deficiency than in patients with non-endocrine SIADH.

Tumors

The first cases of SIADH were described in patients with lung cancer. Small-cell carcinoma of the lung remains a common cause of the syndrome; approximately 10 to 15% of these patients present with SIADH, whereas fewer than 1% of patients with non-small-cell lung cancer become hyponatremic. Ectopic production of vasopressin appears to be responsible for most cases hyponatremia associated with small-cell carcinoma. Arginine vasopressin, oxytocin, and neurophysins have been found by radioimmunoassay in tumors, and are produced by the vast majority of small-cell lung cancers; the quantity of vasopressin peptide is closely correlated with the presence of hyponatremia. Atrial natriuretic peptide mRNA has also been detected in a high percentage of cell lines from patients with small-cell cancer. Hyponatremia develops in approximately 3 to 7% of patients with head and neck cancer ; the mechanism for SIADH associated with these tumors is unknown. Although a number of other tumors can produce vasopressin, there are very few reports of hyponatremia associated with them. SIADH may also emerge in patients with acute tumor lysis syndrome.

Pulmonary Disease

Mild hyponatremia has long been recognized in patients with tuberculosis. The mechanism for vasopressin release has not been determined, but hormone levels fall in response to water-loading, reflecting the reset osmostat variant of SIADH. Ectopic secretion of antidiuretic hormone was the presumptive cause of SIADH in a patient with central diabetes insipidus who developed pulmonary tuberculosis. Tuberculosis-associated hyponatremia resolves within days to weeks of antituberculous therapy. Plasma vasopressin levels are typically elevated on admission in patients with bacterial pneumonia, and fall rapidly during treatment. Hyponatremia is common, and usually self-corrects relatively rapidly after a few days. Abnormal vasopressin secretion in simple pneumonia is not attributable to hypovolemia, hypotension or abnormal PO ₂ or PCO ₂ . An antidiuretic decapeptide (“pneumadin”), which rapidly increases arginine vasopressin (AVP) levels, has been isolated from rat and human lung. Further study is needed to determine if the decapeptide mediates SIADH associated with pneumonia and other lung disorders. Acute respiratory failure (hypoxia and hypercarbia) and severe asthma are also associated with SIADH.

Endocrine Disease

Hyponatremia is present in up to 88% of patients with Addison’s disease, and in 28% of patients with isolated ACTH deficiency. In Addison’s disease, vasopressin is released in response to volume-depletion caused by mineralocorticoid deficiency, and altered hemodynamics caused by glucocorticoid deficiency. Glucocorticoids may also directly inhibit vasopressin release. Unlike Addison’s disease, impaired water excretion in hypopituitarism is not associated with hyperkalemia, and it does not respond to volume replacement with isotonic saline so that hyponatremia in this condition has all the features of SIADH. However, hyponatremia in patients with hypopituitarism is associated with a lower serum bicarbonate concentration than in patients with hyponatremia from other causes of SIADH (see section “Serum Bicarbonate Concentration in SIADH”). Impaired water excretion in hypopituitarism and in adrenalectomized mineralocorticoid replaced subjects is associated with elevated vasopressin levels; administration of a vasopressin V2 receptor-antagonist normalizes urinary water excretion in adrenalectomized mineralocorticoid replaced rats. Inhibitory glucocorticoid receptors have been identified in the magnocellular neurons which secrete vasopressin, and these receptors are markedly increased under hypoosmolar conditions. Hyponatremia with clinical features of SIADH has been described frequently in hypothyroidism; impaired water excretion can be corrected rapidly by the administration of thyroid hormone. However, it is unclear whether hemodynamic or intrarenal factors, rather than vasopressin, is responsible for impaired water excretion in this condition.

Non-Diuretic Drugs

Arginine vasopressin used therapeutically in the treatment of gastrointestinal bleeding or in the management of septic shock, and the vasopressin analog desmopressin acetate (DDAVP), a pure V2 receptor-agonist used to treat diabetes insipidus, enuresis, and von Willebrand disease, may cause hyponatremia. In addition, a growing number of drugs that are unrelated to vasopressin have been reported to cause hyponatremia. Most published reports involve thiazide and thiazide-like diuretics, chlorpropamide, carbamazepine, oxcarbantipsychotics, antidepressants, and nonsteroidal anti-inflammatory drugs. Vincristine and vinblastine increase vasopressin release by unknown mechanisms, and hyponatremia associated with these agents is dose related.

Several drugs associated with hyponatremia appear to increase the response of the collecting duct to circulating vasopressin. Chlorpropamide has been the most thoroughly studied. In addition to augmenting release of vasopressin from the neurohypophysis, chlorpropamide increases the number of vasopressin receptors on collecting tubule cells and inhibits renal medullary synthesis of prostaglandin E ₂ , an agent which blunts the hydroosmotic effect of vasopressin by diminishing adenylate cyclase activity. Inhibition of PGE ₂ permits increased cAMP formation, enhancing the effect of the hormone. Nonsteroidal anti-inflammatory agents increase the hydroosmotic effect of vasopressin by a similar mechanism. Hyponatremia attributable solely to nonsteroidals has been rarely reported, but these commonly used drugs may exacerbate other causes of hyponatremia.

Carbamazepine most commonly causes hyponatremia when it is given to subjects who habitually drink large volumes of water. Carbamazepine increases water permeability of the distal inner medulary collecting duct in the absence of vasopressin, an effect that is cAMP-dependent and negated by a vasopressin V2 receptor-antagonist, implying an action on the V2 receptor–protein G complex. Oxcarbazepine, which is enjoying increased use because of fewer drug interactions than carbamazepine, causes hyponatremia in 9% of patients with epiliepsy who are treated with the drug.

Hyponatremia caused by cyclophosphamide may also be due to enhanced vasopressin action, but the mechanism for this effect has not yet been elucidated. Cyclophosphamide’s antidiuretic effect is delayed, with a time-course which parallels excretion of active metabolites of the drug. Cyclophosphamide’s antidiuretic effect may be enhanced by indomethacin. There are many reports of SIADH associated with psychotropic drugs, including phenothiazines, monoamine oxidase inhibitors, and tricyclic antidepressants; causality has been most convincingly demonstrated with tricyclics. More recently, a large number of cases of SIADH have been reported in patients taking serotonin reuptake inhibitors (SSRIs). Prospective series in elderly patients have shown that 12 to 40% became hyponatremic within two weeks of starting therapy with paroxetine. All of these agents act centrally, and could conceivably affect vasopressin release directly. There is evidence that SSRIs may have a direct effect on water permeability of the inner medullary collecting duct, increasing AQP2 without changing vasopressin levels. The recreational drug, 3,4-methylenedioxymethamphetamine (“Ecstasy”) has been associated with severe, and sometimes fatal, acute hyponatremia. Ecstasy induces vasopressin secretion, and users of the drug who become hyponatremic typically manifest marked polydipsia. Many cases of SIADH associated with amiodarone have been reported.

Post-Operative SIADH

Vasopressin levels are elevated after operative procedures, and remain elevated for several days. Administration of hypotonic fluid during this period of antidiuresis causes acute hyponatremia, with potentially disastrous consequences.

Urinary cation loss has been shown to play an important role in the pathogenesis of hyponatremia in patients with post-operative SIADH. In the post-operative period, it is common for physicians to infuse several liters of isotonic or hypotonic saline solutions, exceeding the intended replacement of third-space and external losses, and actually causing extracellular fluid volume-expansion. A balance study in women undergoing uncomplicated gynecological surgery showed that sodium plus potassium concentrations in the urine peaked at 295±9 mEq/L, and remained hypertonic to plasma for the first 16 hours after induction of anesthesia ( Figure 44.4 ). Because of the action of vasopressin and the natriuretic response to saline-induced volume-expansion, electrolyte-free water was generated, lowering the plasma sodium concentration despite infusion of isotonic saline. The infused saline was, in effect, “desalinated.” A similar phenomenon occurs when patients with subarachnoid hemorrhage and “cerebral salt-wasting” are given large volumes of isotonic saline to protect cerebral perfusion.

Urine electrolytes in post-operative SIADH.

The figure depicts the sum of urine sodium and potassium concentrations obtained post-operatively in 22 women undergoing uncomplicated gynecologic surgery and receiving infusions of 0.9% saline (sodium concentration 154 mmol/l) or lactated Ringer’s solution (sodium concentration 130 mmol/l). The dotted line represents a urine cation concentration of 150 mmol/l, isotonic to normal plasma water. During the first 1000 minutes after the induction of anesthesia, urine cation concentrations were uniformly hypertonic to plasma, contributing to the genesis of hyponatremia. In samples obtained after 1000 minutes (from 16 to 24 hours), the urine remained hypertonic in some patients and became hypotonic in others. Hypotonic urine (below the dotted line) contributes to the correction of hyponatremia.

Neurologic Disorders and Cerebral Salt-Wasting

An association between hyponatremia and intracranial disease has been recognized since the 1950s, and hyponatremia has been reported in a wide variety of systemic diseases involving the CNS, including systemic lupus erythematosus, Guillain-Barre syndrome, meningitis, encephalitis, brain tumors, and brain abscesses. Noting increased urinary sodium excretion in hyponatremic patients with neurologic disorders, Peters referred to the condition as “cerebral salt-wasting”. Once it was recognized that high rates of urinary sodium excretion could be caused by unregulated secretion of antidiuretic hormone, most investigators ascribed hyponatremia in brain disease to SIADH, a syndrome which has been associated a wide array of central nervous system disorders, consistent with the multiple anatomic pathways leading to vasopressin secretion by hypothalamic neurons ( Table 44.3 ).

Cerebral salt-wasting was generally a forgotten term until the early 1980s, when a more concerted effort was made to understand the pathogenesis of hyponatremia in patients with intracranial disease (especially subarachnoid hemorrhage). Reduced blood and plasma volume were found in most patients with intracranial disease who were presumed to have hyponatremia secondary to SIADH. Prospective studies of patients with subarachnoid hemorrhage given maintenance fluids documented negative sodium balance, decreasing plasma volume, and increasing BUN among patients who became hyponatremic within a week of presentation. A course consistent with cerebral salt-wasting was also observed in patients with head injury, brain metastases, and hydrocephalus.

Maintaining an adequate circulatory volume has important clinical implications, especially in subarachnoid hemorrhage, where volume depletion and fluid restriction have been reported to predispose to cerebral ischemia and infarction. This finding, and evidence that volume-expansion protects against cerebral ischemic events in subarachnoid hemorrhage, has led to a general acceptance of “hypertensive, hypervolemic, hemodilutional” therapy for the disorder. When such treatment is given, a high urine sodium concentration and hyponatremia are not reliable indicators of salt-wasting, because hyponatremia may be due to SIADH and the natriuresis may be a response to iatrogenic volume-expansion. In one study, a positive balance for sodium could be documented in most patients believed to have cerebral salt-wasting when calculations included all infusions from the time of first contact with medical or paramedical personnel. High levels of catecholamines which are often associated with brain injury decrease venous capacitance and raise blood pressure, potentially increasing “effective arterial blood volume,” and promoting a physiological natriuresis.

A valid diagnosis of cerebral salt-wasting requires proof of urinary sodium losses, despite reduced effective arterial blood volume. Attempts to establish a diagnosis of hypovolemia have included clinical impressions, central venous pressure, and measurements of plasma and blood volume; none of these can define hypovolemia definitively.

As in patients with SIADH, plasma vasopressin levels are increased and urine sodium concentrations are elevated, but the increased vasopressin secretion in cerebral salt-wasting has been attributed to volume-depletion caused by the primary salt-wasting. Investigation into the pathogenesis of cerebral salt-wasting has focused primarily on the relative roles of natriuretic hormones and vasopressin after subarachnoid hemorrhage. Atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) are both derived from cardiac tissue, and have natriuretic and aldosterone-inhibiting properties. Although BNP has been localized to the hypothalamus, it is primarily of cardiac origin. Cardiac release of ANP and BNP is regulated in part by the CNS, and brain injury may induce their release from cardiac tissue. Plasma ANP levels generally correlate with the presence and severity of blood in the ventricles and increased intracranial pressure. In one study, however, plasma levels of BNP, but not ANP or vasopressin, were correlated with urinary sodium excretion, and in a rat model of cerebral salt-wasting, ANP levels decreased and BNP levels did not change. Jugular venous sampling in suspected cerebral salt-wasting did not support cerebral release of BNP. It is thus unclear whether elevated natriuretic peptide levels are responsible for natriuresis in subarachnoid hemorrhage, and other factors, such altered sympathetic tone, depressed aldosterone levels, ouabain-like compound, and endothelin may contribute. In fact, several studies have indicated that fludrocortisone may be effective in the treatment of cerebral salt-wasting.

Regardless of the pathogenesis of renal salt-wasting in patients with subarachnoid hemorrhage, hyponatremia appears to be caused by vasopressin release that is independent of volume-depletion, and is thus “inappropriate.” In a large prospective study of acute aneurysmal subarachnoid hemorrhage treated aggressively with isotonic saline (between 3 and 8 liters daily), hyponatremia developed in one-third of the patients within 4 to 6 days, despite positive fluid balance, increased blood volume, and suppressed plasma renin and aldosterone levels. Plasma vasopressin was measured at concentrations of 1 to 4 pg/ml, despite plasma osmolalities at which the hormone should have been undetectable. Although plasma ANP levels were also increased in most patients, they did not correlate with serum sodium concentration. Thus, it is likely that both salt loss and SIADH contribute in varying degrees to the pathogenesis of hyponatremia after subarachnoid hemorrhage.

The true incidence of cerebral salt-wasting among patients with subarachnoid hemorrhage is probably low, and a clear distinction between this entity and SIADH in neurosurgical patients may not be important, because in either case the most effective treatment is intravenous hypertonic saline.