Depletion of body solute can result from any significant losses of ECF. Whether via renal or nonrenal routes, body fluid losses by themselves rarely cause hypoosmolality because excreted or secreted body fluids are usually isotonic or hypotonic relative to plasma and therefore tend to increase plasma osmolality. Consequently, when hypoosmolality accompanies ECF losses it is generally the result of replacement of body fluid losses by more hypotonic solutions, thereby diluting the remaining body solutes. This often occurs when patients drink water or other hypotonic fluids in response to ongoing solute and water losses, and also when hypotonic intravenous fluids are administered to hospitalized patients.³⁶ When the solute losses are marked, these patients can show all of the obvious signs of volume depletion (e.g., addisonian crisis). However, such patients often have a more deceptive clinical presentation because their volume deficits may be partially replaced by subsequently ingested or infused fluids. Moreover, they may not manifest signs or symptoms of cellular dehydration because osmotic gradients will draw water into the relatively hypertonic ICF. Therefore, clinical evidence of hypovolemia strongly supports solute depletion as the cause of plasma hypoosmolality, but absence of clinically evident hypovolemia never completely eliminates this as a possibility. Although ECF solute losses are responsible for most cases of depletion-induced hypoosmolality, ICF solute loss can also cause hypoosmolality as a result of osmotic water shifts from the ICF into the ECF.³³ This mechanism likely contributes to some cases of diuretic-induced hypoosmolality in which depletion of total body K⁺ often occurs.³⁷,³⁸

Despite the obvious importance of solute depletion in some patients, most cases of clinically significant hypoosmolality are caused by increases in total body water rather than by primary loss of extracellular solute. This can occur because of either impaired renal free water excretion or excessive free water intake. However, the former accounts for most hypoosmolar disorders because normal kidneys have sufficient diluting capacity to allow excretion of up to 20 to 30 L per day of free water (see Chapter 4). Intakes of this magnitude are occasionally seen in a subset of psychiatric patients³⁹,⁴⁰ but not in most patients, including patients with SIADH in whom fluid intakes average 2 to 3 L per day.⁴¹ Consequently, dilutional hypoosmolality usually is the result of an abnormality of renal free water excretion. The renal mechanisms responsible for impairments in free water excretion can be subgrouped according to whether the major impairment in free water excretion occurs in proximal or distal parts of the nephron, or both (see Table 70.1).

Any disorder that leads to a decrease in glomerular filtration rate (GFR) causes increased reabsorption of both Na⁺ and water in the proximal tubule. As a result, the ability to excrete free water is limited because of decreased delivery of tubular fluid to the distal nephron. Disorders causing solute depletion through nonrenal mechanisms (e.g., gastrointestinal fluid losses) also produce this effect. Disorders that cause a decreased GFR in the absence of significant ECF losses are, for the most part, edema-forming states associated with decreased effective arterial blood volume (EABV) and secondary hyperaldosteronism.⁴²,⁴³ Although these conditions are typified by increased proximal reabsorption of both Na⁺ and fluid, it is now clear that in most cases water retention also results from increased distal reabsorption caused by nonosmotic baroreceptor-mediated stimulated increases in plasma AVP levels,⁴⁴,⁴⁵ with the possible exception of hypothyroidism.

Distal nephron impairments in free water excretion are characterized by an inability to dilute tubular fluid maximally. These disorders are usually associated with abnormalities in the secretion of AVP from the posterior pituitary. However, just as depletion-induced hypoosmolar disorders usually include an important component of secondary impairments of free water excretion, so do most dilution-induced hypoosmolar disorders involve significant degrees of secondary solute depletion. This was recognized even before the first clinical description of SIADH from studies of the effects of posterior pituitary extracts on water retention, which demonstrated that renal salt wasting was predominantly a result of the ECF volume expansion produced by the retained water.⁴⁶ Therefore, after sustained increases in total body water secondary to inappropriately elevated AVP levels, sufficient secondary solute losses, predominantly as Na⁺, occur and can result in further lowering of plasma osmolality. The actual contribution of Na⁺ losses to the hypoosmolality of SIADH is variable and depends in part on both the rate and volume of water retention.⁴⁷ The major factor responsible for secondary Na⁺ losses appears to be renal hemodynamic effects, and specifically the phenomenon of pressure natriuresis and diuresis induced by the volume expansion.⁴⁸ However, volume-stimulated hormones such as atrial natriuretic peptide (ANP) are also elevated in response to the water retention of patients with SIADH,⁴⁹,⁵⁰ and it seems likely that these factors also contribute to the secondary natriuresis, possibly via interactions with intrarenal hemodynamic effects.⁵¹ Regardless of the actual mechanisms involved, the solute losses that occur secondary to water retention can be best understood in the context of volume regulation of the ICF and ECF compartments in response to induced hypoosmolality, which is discussed in the next section.

Some dilutional disorders do not fit particularly well into either category. Chief among these is the hyponatremia that sometimes occurs in patients who ingest large volumes of beer with little food intake for prolonged periods, called “beer potomania.”⁵²,⁵³ Although the volume of fluid ingested may not seem sufficiently excessive to overwhelm renal diluting mechanisms, in these cases free water excretion is limited by very low urinary solute excretion thereby causing water retention and dilutional hyponatremia. A reported case in which hyponatremia occurred in an ovolactovegetarian with a very low protein intake but no beer ingestion is consistent with this pathophysiologic mechanism.⁵⁴ However, because most such patients have very low salt intakes as well, it is likely that relative depletion of body Na⁺ stores also is a contributing factor to the hypoosmolality in at least some cases.⁵⁵

Many studies have indicated that the combined effects of water retention plus urinary solute excretion cannot adequately explain the degree of plasma hypoosmolality observed in patients.²,⁵⁶,⁵⁷ This observation originally led to the theory of cellular inactivation of solute.² Simply stated, this theory suggested that as ECF osmolality falls, water moves into cells along osmotic gradients, thereby causing the cells to swell. At some point during this volume expansion, the cells osmotically “inactivate” some of their intracellular solutes as a defense mechanism to prevent continued cell swelling with subsequent detrimental effects on cell function and survival. As a result of this decrease in intracellular osmolality, water then shifts back out of the ICF into the ECF, but at the expense of further worsening the dilution-induced hypoosmolality. Despite the appeal of this theory, its validity has never been demonstrated conclusively in either human or animal studies.

An appealing alternative theory has been suggested by studies of cell volume regulation, in which cell volume is maintained under hypoosmolar conditions by extrusion of potassium rather than by osmotic inactivation of cellular solute.⁵⁸,⁵⁹ Whole brain volume regulation via similar types of electrolyte losses was first described by Yannet in 1940,⁶⁰ and has long been recognized as the mechanism by which the brain was able to adapt to hyponatremia and limit brain edema to sublethal levels.⁶¹,⁶²,⁶³ Following the recognition that low molecular weight organic compounds, called organic osmolytes, also constituted a significant osmotic component of a wide variety of cell types, studies demonstrated the accumulation of these compounds in response to hyperosmolality in both kidney⁶⁴ and brain⁶⁵ tissue. Multiple groups have now shown that the brain loses organic osmolytes in addition to electrolytes during the process of volume regulation to hypoosmolar conditions in experimental animals⁶⁶,⁶⁷,⁶⁸,⁶⁹ and human patients.⁷⁰ These losses occur relatively quickly (within 24 to 48 hours in rats) and can account for as much as one third of the brain solute losses during hyponatremia.⁷¹ Such coordinate losses of both electrolytes and organic osmolytes from brain tissue enables very effective regulation of brain volume during chronic hyponatremia (Fig. 70.1).⁷² Consequently, it is now clear that cell volume regulation in vivo in brain tissue occurs predominantly through depletion, rather than intracellular osmotic inactivation, of a variety of intracellular solutes. Ongoing experimental studies are better defining the complex cellular and molecular mechanisms that underlie this profound adaptation to hypoosmolality.

Most studies have focused on volume regulation in the brain during hyponatremia, but all cells volume regulate to varying degrees,⁵⁸ and there is little question that this process occurs throughout the body as whole organisms adapt to hypoosmolar conditions. Unexplained components of hyponatremia that led to previous speculation about cellular inactivation of solute are now better explained by cellular losses of both electrolyte and organic solutes as cells throughout the body undergo volume regulation during hypoosmolar conditions. However, volume regulatory processes are not limited to cells. Although most cases of hyponatremia clearly result from initial water retention induced by stimulated antidiuresis, it has always seemed likely that the resulting natriuresis served the purpose of regulating the volumes of the ECF and intravascular spaces. Many experimental and clinical observations are consistent with ECF volume regulation via secondary solute losses. First, dilutional decreases in concentrations of most blood constituents other than Na⁺ and Cl^– do not occur in patients with SIADH,⁷³ suggesting that their plasma volume is not nearly as expanded as would be predicted simply by the measured decreases in serum [Na⁺]. Second, an increased incidence of hypertension has never been observed in patients with SIADH,⁷⁴ again arguing against significant expansion of the arterial blood volume. Third, results of animal studies in both dogs⁷⁵ and rats⁷⁶ have clearly indicated that a significant component of chronic hyponatremia is attributable to secondary Na⁺ losses rather than water retention. Furthermore, the relative contributions from water retention versus sodium loss vary with the duration and severity of the hyponatremia: water retention was found to be the major cause of decreased serum [Na⁺] in the first 24 hours of induced hyponatremia in rats, but Na⁺ depletion then became the predominant etiologic factor after longer periods (7-14 days) of sustained hyponatremia, particularly at very low (<115 mEq per L) serum [Na⁺] levels.⁷⁶ Finally, multiple studies have attempted to measure body fluid compartment volumes in hyponatremic patients, but without consistent results that indicate either plasma or ECF volume expansion.¹,⁵⁷,⁷⁷,⁷⁸ In particular, a report of body fluid space measurements using isotope dilution techniques in hyponatremic and normonatremic patients with small cell lung carcinoma showed no differences between the two groups with regard to exchangeable sodium space, ECF volume by ³⁵SO₄ distribution, or total body water.⁷⁹ Such results have traditionally been explained by the relative insensitivity of isotope dilution techniques for measurement of body fluid compartment spaces, but an equally plausible possibility is that in the chronically adapted hyponatremic state body fluid compartments have regulated their volumes back toward normal via a combination of extracellular (predominantly electrolyte) and intracellular (electrolyte and organic osmolyte) solute losses.⁸⁰ Figure 70.2 schematically illustrates some of the volume regulatory processes that likely occur in response to water retention induced by inappropriate antidiuresis. The degree to which solute losses versus water retention contribute to the resulting hyponatremia will vary in association with many different factors, including the etiology of the hyponatremia, the rapidity of development of the hyponatremia, the chronicity of the hyponatremia, the volume of daily water loading and subsequent volume expansion, and undoubtedly some degree of individual variability as well. It therefore hardly seems surprising that studies of hyponatremic patients have failed to yield uniform results regarding the pathogenesis of hyponatremia in view of the marked diversity of hyponatremic patients and their presentation at different times during the process of adaptation to hypoosmolality via volume regulatory processes.

The presence of clinically detectable hypovolemia nearly always signifies total body solute depletion. A low urine [Na⁺] indicates a nonrenal cause of solute depletion. If the urine [Na⁺] is high despite hypoosmolality, renal causes of solute depletion are likely responsible. Therapy with thiazide diuretics is the most common cause of renal solute losses,³⁸ particularly in the elderly,⁸¹,⁸² but mineralocorticoid deficiency as a result of adrenal insufficiency⁸³ or mineralocorticoid resistance⁸⁴ must always be considered as well. Less commonly, renal solute losses may be the result of a salt-wasting nephropathy (e.g., polycystic kidney disease,⁸⁵ interstitial nephritis,⁸⁶ or chemotherapy⁸⁷).

The presence of clinically detectable hypervolemia usually signifies total body Na⁺ excess. In these patients, hypoosmolality results from an even greater expansion of total body water caused by a marked reduction in the rate of water excretion (and sometimes an increased rate of water ingestion). The impairment in water excretion is secondary to a decreased EABV,⁴²,⁴³ which increases the reabsorption of glomerular filtrate not only in the proximal nephron but also in the distal and collecting tubules by stimulating AVP secretion.⁴⁴,⁴⁵ These patients generally have a low urine [Na⁺] because of secondary hyperaldosteronism, which is also a product of decreased EABV. However, under certain conditions urine [Na⁺] may be elevated, usually secondary to concurrent diuretic therapy but also sometimes because of a solute diuresis (e.g., glucosuria in diabetics) or after successful treatment of the underlying disease (e.g., ionotropic therapy in patients with congestive heart failure). An additional disorder that can produce hypoosmolality and hypervolemia is acute or chronic renal failure with fluid overload¹² (although in early stages of renal failure polyuria from AVP resistance is more likely⁸⁸). Urine [Na⁺] in these cases is usually elevated, but it can be variable depending on the stage of renal failure. It is important to remember that primary polydipsia will not be accompanied by signs of hypervolemia because water ingestion alone, in the absence of Na⁺ retention, does not typically produce clinically apparent degrees of ECF volume expansion.

Many different hypoosmolar disorders can potentially present clinically with euvolemia, in large part because it is difficult to detect modest changes in volume status using standard methods of clinical assessment; in such cases, measurement of urine [Na⁺] is an especially important first step.⁸⁹ A high urine [Na⁺] in euvolemic patients usually implies a distally mediated, dilution-induced hypoosmolality such as SIADH. However, glucocorticoid deficiency can mimic SIADH so closely that these two disorders are often indistinguishable in terms of water balance.⁹⁰,⁹¹ Hyponatremia from diuretic use also can present without clinically evident hypovolemia, and the urine [Na⁺] will often be elevated in such cases because of the renal tubular effects of the diuretics.³⁸ Recent studies have suggested that the fractional excretion of uric acid may provide a better measure of ECF volume status in hyponatremic patients on diuretics.⁹² A low urine [Na⁺] suggests a depletion-induced hypoosmolality from ECF losses with subsequent volume replacement by water or other hypotonic fluids. The solute loss often is generally nonrenal in origin, but an important exception is recent cessation of diuretic therapy, because urine [Na⁺] can quickly decrease to low values within 12 to 24 hours after discontinuation of the diuretic. The presence of a low serum [K⁺] is an important clue to diuretic use, because few of the other disorders that cause hypoosmolality are associated with significant hypokalemia. However, even in the absence of hypokalemia, any hypoosmolar, clinically euvolemic patient taking diuretics should be assumed to have solute depletion and treated accordingly; subsequent failure to correct the hypoosmolality with isotonic saline administration and persistence of an elevated urine [Na⁺] after discontinuation of diuretics then requires reconsideration of a diagnosis of dilutional hypoosmolality. A low urine [Na⁺] also can also be seen in some cases of hypothyroidism, in the early stages of decreased EABV before the development of clinically apparent sodium retention and fluid overload, or during the recovery phase from SIADH. Hence, a low urine [Na⁺] is less meaningful diagnostically than is a high value.

Because euvolemic causes of hypoosmolality represent the most challenging etiologies of this disease, both in terms of differential diagnosis as well as with regard to the underlying pathophysiology, the subsequent sections will discuss the major causes of euvolemic hypoosmolality and hyponatremia in greater detail.

One of the most common associations of SIADH remains with tumors. Although many different types of tumors have been associated with SIADH (Table 70.4), bronchogenic carcinoma of the lung has been uniquely associated with SIADH since the first description of this disorder in 1957.¹ In virtually all cases, the bronchogenic carcinomas causing this syndrome have been of the small cell variety; a few squamous cell types have been described, but these are rare. Incidences of hyponatremia as high as 11% of all patients with smallcell carcinoma,¹⁰³ or 33% of cases with more extensive disease,¹⁰⁴ have been reported. The unusually high incidence of small cell carcinoma of the lung in patients with SIADH, together with the relatively favorable therapeutic response of this type of tumor, makes it imperative that all patients presenting with an otherwise unexplained SIADH be investigated thoroughly and aggressively for a possible tumor. The evaluation should include a computed tomography (CT) or magnetic resonance imaging (MRI) scan of the thorax. In cases with a high degree of suspicion (e.g., hyponatremia in a young smoker) bronchoscopy with cytologic analysis of bronchial washings should be considered even if the results of routine chest radiography are normal, since several studies have reported hypoosmolality that predated any radiographically evident abnormality in patients who then were found to harbor bronchogenic carcinomas 3 to 12 months later.¹⁰⁵,¹⁰⁶ Head and neck cancers account for another group of malignancies associated with relatively higher incidences of SIADH,¹⁰⁷ and some of these tumors have clearly been shown to be capable of synthesizing AVP ectopically.¹⁰⁸ A report from a large cancer hospital showed an incidence of hyponatremia for all malignancies combined of 3.7%, with approximately one third of these due to SIADH.²⁰

The second major etiologic group of disorders causing SIADH has its origins in the central nervous system (CNS). Despite the large number of different CNS disorders associated with SIADH, there is no obvious common denominator linking them. However, this is actually not surprising when one considers the neuroanatomy of neurohypophysial innervation. The magnocellular AVP neurons receive excitatory inputs from osmoreceptor cells located in the anterior hypothalamus, but also have a major innervation from brainstem cardiovascular regulatory and emetic centers (Fig. 70.4). Although various components of these pathways have yet to be fully elucidated, many of them appear to have inhibitory as well as excitatory components.¹⁰⁹ Consequently, any diffuse CNS disorder can potentially cause AVP hypersecretion either by nonspecifically exciting these pathways via irritative foci, or alternatively by disrupting them and thereby decreasing the level of inhibition impinging upon the AVP neurons in the neurohypophysis. The wide variety of diverse CNS processes that can potentially cause SIADH stands in contrast to CNS causes of diabetes insipidus, which are for the most part limited to lesions localized to the hypothalamus and/or posterior pituitary that destroy the magnocellular vasopressin neurons (see Chapter 71).

Pulmonary disorders represent a relatively common but frequently misunderstood cause of SIADH. A variety of pulmonary disorders have been associated with this syndrome, but other than tuberculosis,¹²⁸ acute pneumonia,¹²⁹,¹³⁰,¹³¹ and advanced chronic obstructive lung disease,¹³² the occurrence of hypoosmolality has been noted mainly in sporadic case reports. Some bacterial infections appear to be associated with a higher incidence of hyponatremia, particularly Legionella pneumoniae.¹³³ Although one case of pulmonary tuberculosis has been reported that suggested the possibility that tuberculous lung tissue might synthesize AVP ectopically,¹³⁴ several other studies have reported that advanced pulmonary tuberculosis is associated with the reset osmostat form of SIADH,⁹⁶,¹²⁸ presumably from nonosmotic stimulation of posterior pituitary AVP secretion. Most cases of pulmonary SIADH not associated with either tuberculosis or pneumonitis have occurred in the setting of respiratory failure. Although hypoxia has clearly been shown to stimulate AVP secretion in animals,¹³⁵,¹³⁶ it appears to be less effective as a stimulus in humans,¹³⁷ in whom the stimulus to abnormal water retention appears to be hypercarbia more so than hypoxia.¹³⁸,¹³⁹ When such patients were evaluated serially, the inappropriate AVP secretion was found to be limited to the initial days of hospitalization, when respiratory failure was most marked.¹⁴⁰ Even cases of tubercular SIADH generally have occurred in patients with far advanced, active, pulmonary tuberculosis, although interestingly hyponatremia was also found in 74% of a series of patients with miliary tuberculosis.¹⁴¹ Therefore, SIADH in non-tumor-related pulmonary disease generally conforms to the following characteristics: (1) the pulmonary disease will always be obvious as a result of severe dyspnea or extensive radiographically evident infiltrates, and (2) the inappropriate antidiuresis will usually be limited to the period of respiratory failure— once clinical improvement has begun, free water excretion generally improves rapidly. Mechanical ventilation can cause inappropriate AVP secretion, or it can worsen any SIADH caused by other factors. This phenomenon has been associated most often with continuous positive pressure ventilation,¹⁴² but it can also occur to a lesser degree with the use of positive end expiratory pressure.

One of the most recently described causes of hypoosmolality is the acquired immunodeficiency syndrome (AIDS) or AIDS-related complex (ARC), in patients with human immunodeficiency virus (HIV) infection, with incidences of hyponatremia reported as high as 30% to 38% in adults¹⁴³,¹⁴⁴,¹⁴⁵ and children.¹⁴⁶ Although there are many potential etiologies for hyponatremia in patients with AIDS/ARC, including dehydration, adrenal insufficiency, and pneumonitis, from 12% to 68% of AIDS patients who develop hyponatremia appear to meet criteria for a diagnosis of SIADH.¹⁴³,¹⁴⁴,¹⁴⁵ Not unexpectedly, reports have implicated some of the medications used to treat these patients as the cause of the hyponatremia, either via direct renal tubular toxicity or SIADH.¹⁴⁷,¹⁴⁸

A recent series of reports have documented a surprisingly high incidence of hyponatremia during endurance exercise events such as marathon¹⁴⁹ and ultramarathon¹⁵⁰ foot races, triathlons,¹⁵¹ forced marching,¹⁵² and hiking.¹⁵³ Occasionally, this has caused fatal outcomes associated with hyponatremic encephalopathy from acute brain edema.¹⁵⁴,¹⁵⁵ Most studies support excess drinking during the exercise as the major cause of the induced hyponatremia,¹⁵⁶,¹⁵⁷ but it now appears that water retention under such conditions is also contributed to by SIADH as a result of multiple potential nonosmotic stimuli (e.g., volume depletion, nausea, increased cytokine levels).¹⁵⁸,¹⁵⁹

Unexplained or idiopathic causes account for a relatively small proportion of all cases of SIADH. Although the etiology of the syndrome may not be diagnosed initially in many cases, the numbers of patients in whom an apparent cause cannot be established after consistent follow-up over time are relatively few. However, an exception to this appears to be elderly patients who sometimes develop SIADH without any apparent underlying etiology.¹⁶⁰,¹⁶¹,¹⁶² Coupled with the significantly increased incidence of hyponatremia in geriatric patients,⁷,¹³,¹⁴,⁹³,¹⁶³,¹⁶⁴ this suggests that the normal aging process may be accompanied by abnormalities of regulation of AVP secretion that predispose to SIADH. Such an effect could potentially account for the fact that virtually all causes of drug-induced hyponatremia occur much more frequently in elderly patients.⁸²,¹⁶⁵,¹⁶⁶ In several series of elderly patients meeting criteria for SIADH, 40% to 60% remained idiopathic despite rigorous evaluation,¹⁶⁷,¹⁶⁸,¹⁶⁹ leading some to conclude that extensive diagnostic procedures were not warranted in such elderly patients if routine history, physical examination, and laboratory evaluation failed to suggest a diagnosis.¹⁶⁷

Some well-known stimuli of AVP secretion are notable primarily because of their exclusion from Table 70.4. Despite unequivocal stimulation of AVP secretion by nicotine,¹⁷⁰ cigarette smoking has only rarely been associated with SIADH, and primarily in psychiatric patients who have several other potential causes of inappropriate AVP secretion.³⁹,¹⁷¹,¹⁷² This is in part because of chronic adaptation to the effects of nicotine, but also because the short half-life of AVP in plasma (approximately 15 min in humans¹⁷³) limits the duration of antidiuresis produced by relatively short-lived stimuli such as smoking. Although nausea remains the most potent stimulus to AVP secretion known in man,¹⁷⁴ chronic nausea is rarely associated with hypoosmolality unless accompanied by vomiting with subsequent ECF solute depletion followed by ingestion of hypotonic fluids.¹⁷⁵ Similar to smoking, this is probably attributable to the short half-life of AVP, but also to the fact that most such patients are not inclined to drink fluids under such circumstances. However, hyponatremia can occur when such patients are infused with high volumes of hypotonic fluids. This is likely a factor contributing to the hyponatremia that often occurs in cancer patients who are receiving chemotherapy.¹⁰³ Finally, a causal relation between stress and SIADH has often been suggested, but never conclusively established. This underscores the fact that stress, independent of associated nausea, dehydration, or hypotension, is not a major stimulus causing sustained elevations of AVP levels in humans.¹⁷⁶