CHAPTER 26 DIABETES INSIPIDUS AND HYPEROSMOLAR SYNDROMES
Principles and Practice of Endocrinology and Metabolism
CHAPTER 26 DIABETES INSIPIDUS AND HYPEROSMOLAR SYNDROMES
PETER H. BAYLIS AND CHRISTOPHER J. THOMPSON
Thirst in Hypothalamic Diabetes Insipidus
Diabetes Insipidus and Pregnancy
Osmoregulatory Defects in Chronic Hypernatremia
Blood osmolality in healthy persons is maintained within narrow limits by a series of mechanisms that are described in detail in Chapter 25. Adjustments in water balance determine the constancy of blood osmolality, which is mediated by delicate alterations in thirst appreciation (with consequent promotion of drinking) plus the enormous capacity of the kidney to alter urine flow rates and urine osmolality in response to relatively small changes in the plasma vasopressin concentration (see Chap. 206). Thus, healthy humans are able to conserve their osmotic internal milieu despite extremes in climatic conditions, sustained severe exertion, or, to a certain degree, an inadequate supply of water.
Aberrations of the intricate mechanisms involved in maintaining osmoregulation can lead to the inappropriate accumulation of water, which is recognized as one of the hypoosmolar states, or to the loss of renal water, which usually is clinically apparent as polyuria but also may be manifested as one of the hyperosmolar syndromes. This chapter is concerned with clinical situations associated with polyuria and, on occasion, abnormalities of thirst appreciation.
Diabetes insipidus refers to the passage of copious volumes of dilute urine and is synonymous with polyuria. In adults, the urine volume exceeds 2.5 L per 24 hours (>40 mL/kg per 24 hours), while in children the output is greater (>100 mL/kg per 24 hours).
Three pathophysiologic conditions result in diabetes insipidus. An absolute or partial deficiency of vasopressin secretion from the neurohypophysis in response to normal osmotic stimulation is termed hypothalamic diabetes insipidus. This disorder is also known as cranial, central, or neurogenic diabetes insipidus. Patients with hypothalamic diabetes insipidus generally have normal thirst sensation. Their basic abnormality is insufficient circulating antidiuretic activity, which is the principal, but not the sole, cause of their polyuria. Diabetes insipidus secondary to decreased renal sensitivity to the antidiuretic effect of vasopressin circulating in normal or high concentrations is usually called nephrogenic diabetes insipidus. Again, these patients rely on normal thirst sensation to regulate water balance. The third mechanism leading to diabetes insipidus is the ingestion of excessive volumes of fluid, which results in suppression of vasopressin release and consequent polyuria. This condition is referred to as dipsogenic diabetes insipidus, sometimes termed primary polydipsia.
A decrease in maximal urine-concentrating ability occurs after prolonged periods of polyuria, regardless of the primary cause. The passage of large amounts of dilute urine through the distal nephron removes solute from the renal medullary interstitium, a process known as the washout phenomenon.1,2 The osmotic gradient across the collecting tubular cell, which is essential for the antidiuretic action of vasopressin, is decreased. Thus, any of the three pathophysiologic mechanisms responsible for diabetes insipidus may lead to an additional defect that complicates the interpretation of diagnostic tests based on indirect assessment of the antidiuretic action of vasopressin.
In contrast to hypothalamic diabetes insipidus, which is usually the result of a loss of neurosecretory neurons, the chronic hyperosmolar syndromes are frequently the consequence of a defective thirst mechanism. Thirst osmoreceptors may fail to respond to hypertonicity, which results in hypodipsia. Because the putative thirst osmoreceptors are believed to be in proximity to the osmoreceptors that regulate vasopressin secretion, a defect in osmotically mediated vasopressin release is often associated with hypodipsia. Polyuria is rarely a feature of hyperosmolar syndromes, because many patients secrete small amounts of vasopressin that are sufficient to concentrate urine to some extent, and nonosmotic factors regulating vasopressin secretion often remain intact. In view of this characteristic difference between diabetes insipidus and hyperosmolar syndromes, these conditions are discussed separately.
In theory, any of a series of defects in the vasopressin neurosecretory process can be implicated as the cause of hypothalamic diabetes insipidus.3 Abnormalities may arise in the osmoreceptor that controls vasopressin secretion, even when the thirst osmoreceptor is spared. Alternatively, abnormalities may involve the synthesis and packaging of vasopressin (including genetic defects), damage to the vasopressinergic neurons, or disorders of neurohypophyseal hormone release. Enhanced inactivation of vasopressin by circulating degrading enzymes or antibodies is another potential cause of decreased antidiuretic activity.
In practice, however, most cases of permanent hypothalamic diabetes insipidus are caused by damage to the hypothalamo-neurohypophyseal area. The most common causes of this condition are listed in Table 26-1.4
TABLE 26-1. Causes of Diabetes Insipidus
Studies have revealed exciting data regarding a variety of genetic abnormalities found on chromosome 20 in several kindreds with autosomal dominant hypothalamic diabetes insipidus. The first report on different families showed single nucleotide substitutions in the region coding for neurophysin (glycine to serine at position 57).5 This mutation is presumed to interfere with the normal vasopressin-neurophysin tetramer complex formation that occurs in the packaging and transport of vasopressin to the neurohypophysis. Two groups investigating another extended family discovered a nucleotide substitution in the signal peptide (alanine to threonine at position –1).6,7 The signal peptide directs the prohormone to the endoplasmic reticulum, where it is cleaved. Both groups speculate that the mutation alters the cleavage mechanism, resulting in abnormal processing of the prohormone. After the description of the first genetic abnormalities causing familial hypothalamic diabetes insipidus, more than 22 different kindreds with unique genetic mistakes have been documented8 (Fig. 26-1).
The genetic basis of the DIDMOAD (diabetes insipidus, diabetes mellitus, optic atrophy, deafness), or Wolfram, syndrome is less well understood, although some evidence exists that it is a disorder of mitochondrial DNA.9,10
Closed head trauma or frank damage to the pituitary stalk or hypothalamus as a result of surgical intervention is often the cause of a form of diabetes insipidus that usually presents within 24 hours of injury. In ~50% of cases of post-traumatic diabetes insipidus, the condition resolves spontaneously within a few days. Permanent diabetes insipidus develops in another 30% to 40% of these patients, and the remainder exhibit a triphasic response to injury. In the last group, the onset of polyuria is abrupt and the condition lasts a few days. It is followed by a period of antidiuresis that may last 2 to 14 days before permanent diabetes insipidus develops. This triple response to injury is believed to be attributable to release of the vasopressin that is stored in granules.11 Recognition of this entity by clinicians should help to prevent inappropriate treatment that would result in hyponatremia during the second of the three phases.
Tumors of the anterior pituitary rarely cause diabetes insipidus. In a series of >100 cases of hypothalamic diabetes insipidus, 13% were attributable to tumors, which included glioma, germinoma, and craniopharyngioma.12 In children, a central tumor is a frequent cause of hypothalamic diabetes insipidus, accounting for ~25% of cases; in this population, the most common intracranial tumor is germinoma.13 Metastatic deposits in the hypothalamus causing diabetes insipidus usually arise from carcinoma of the breast or bronchus.
FIGURE 26-1. Vasopressin gene, vasopressin precursor molecule, and vasopressin with its specific neurophysin. Three exons encode for the precursor molecule, which comprises a signal protein, vasopressin hormone, neurophysin, and a glycoprotein moiety coupled by amino acids. Mutations in the vasopressin gene have been located in all parts of the precursor molecule except vasopressin itself. (Vp, vasopressin.)
Granulomatous disease accounts for only a few cases of diabetes insipidus in adults (i.e., sarcoidosis, tuberculosis). However, in children with granulomatous disease, histiocytosis X may cause as many as 40% of pediatric cases.
Idiopathic hypothalamic diabetes accounts for ~25% of all cases.12 One-third of patients with apparent idiopathic disease have circulating antibodies to the vasopressin-producing cells in the hypothalamus, a finding which suggests an autoimmune origin for the disorder.14 Some patients have an acute lymphocytic infiltration of the infundibulum and neurohypophysis that can be demonstrated on open biopsy and subsequently resolves.15
NEPHROGENIC DIABETES INSIPIDUS
Mild forms of nephrogenic diabetes insipidus are relatively common (see Table 26-1). Mechanisms responsible for renal resistance to the antidiuretic effect of vasopressin may occur at one or more of the many different sites in the chain of biochemical responses to vasopressin.11,15a Chronic renal disease secondary to numerous conditions, many drugs (e.g., lithium15b), prolonged electrolyte disturbances from hypokalemia, and hypercalcemia account for most cases of nephrogenic diabetes insipidus. The inherited forms of nephrogenic diabetes insipidus are rare, and cause severe polyuria, dehydration, and failure to thrive in the young. With the identification of the V2 (antidiuretic) receptor gene on the X chromosome, a variety of substitutions, mutations, or premature stops have been isolated in kindreds with this disorder that cause defects in the trans-membrane V2 receptor.16 Studies involving three other families with congenital nephrogenic diabetes insipidus have shown autosomal inheritance of the disorder, in contrast to the more common X-linked form due to genetic mutations of the V2 receptor gene localized to the Xq28 region of the long arm of the X chromosome. The autosomal form is due to novel genetic mutations of the gene encoded for the vasopressin-sensitive water channel protein, aquaporin-2, which is located in the collecting tubules.17
DIPSOGENIC DIABETES INSIPIDUS
Dipsogenic diabetes insipidus, also called primary polydipsia or habitual water drinking, is often psychogenic in origin. The course of polyuria in psychotic patients is variable, with fluctuations in polydipsia and urine volumes occurring over the years. Occasional patients with hypothalamic diabetes insipidus who are treated with antidiuretic preparations continue to have polydipsia and, consequently, run the risk of developing hyponatremia. Whether these patients continue to drink because of habit or because of a hypothalamic lesion affecting the thirst osmoreceptor is unknown. A few structural abnormalities resulting in increased thirst have been reported. Drugs that cause dryness of the mouth (e.g., thioridazine hydrochloride) may increase drinking but do not result in true polydipsia, in contrast to lithium, which can stimulate thirst directly.
In adults, the major clinical manifestations of diabetes insipidus include the frequent passage of large volumes of dilute urine (often both day and night), excessive thirst, and increased fluid ingestion. Patients with mild degrees of diabetes insipidus may consider their symptoms to be so minimal that they fail to seek medical attention. However, the severity of diabetes insipidus varies widely, with 24-hour urine volumes ranging from 2.5 to 20 L. Even with the most extreme forms of the disorder, patients maintain their water balance as long as thirst appreciation remains intact and adequate volumes of fluid are ingested.
The onset of the disease occurs at any time from the neonatal period to old age, and the sex distribution is approximately equal, although one large review of adults with hypothalamic diabetes insipidus reported a slight male preponderance (60%:40%).12 In infants, diabetes insipidus usually presents with evidence of chronic dehydration, unexplained fever, vomiting, neurologic disturbance, and failure to thrive.13 Enuresis, sleep disturbances, and difficulties at school are the most common presenting complaints in older children. Usually, no growth retardation or failure to enter puberty occurs. Affected children from families with histories of diabetes insipidus often do not complain; they regard their polydipsia and polyuria as the norm. Once hypothalamic diabetes insipidus develops, it rarely goes into remission spontaneously.
Patients with hypothalamic diabetes insipidus also may have anterior pituitary dysfunction (see Chap. 17), particularly if their disorder resulted from trauma to or a tumor in the hypo-thalamo-neurohypophyseal area. Even patients with the idiopathic form of the disease frequently have endocrinologic evidence of anterior pituitary dysfunction, which suggests a more generalized hypothalamic disorder.12 Glucocorticoid deficiency secondary to impaired corticotropin secretion or to primary adrenal disease leads to impairment of the ability of the kidneys to excrete a water load and to dilute urine maximally. At least two mechanisms are responsible for this defect. One involves the distal nephron, which remains partially impermeable to water in the absence of glucocorticoid; the other involves persistent vasopressin secretion, possibly secondary to a resetting of the osmostat. A similar abnormality of water excretion has been reported with severe thyroid hormone deficiency (see Chap. 45). Thus, impairment of anterior pituitary function can mask hypothalamic diabetes insipidus, which becomes apparent only when the hypopituitarism is adequately treated.
Routine skull radiographs are rarely helpful in patients with hypothalamic diabetes insipidus but nuclear magnetic resonance imaging (MRI) can be extremely useful. T1-weighted magnetic resonance imaging of the neurohypophysis produces a characteristic hyperintense signal in healthy persons that disappears in most patients with hypothalamic diabetes insipidus.18 The infundibular stalk frequently is thickened in the early phase of the idiopathic form of the disorder, as demonstrated by both MRI and computed tomographic scanning.15
Polyuria may be defined as the excretion of >2.5 L of urine per 24 hours on two consecutive days, provided that patients are allowed free access to and drink water ad libitum. Once polyuria has been demonstrated, the clinician’s first responsibility is to establish the pathophysiologic mechanism—dipsogenic diabetes insipidus, hypothalamic diabetes insipidus, or nephrogenic diabetes insipidus. Defining the underlying disease process is then important.
Before the development of plasma assays that were capable of detecting low physiologic concentrations of vasopressin, indirect methods were used to assess the antidiuretic activity of the hormone.
The classic diagnostic approach is to measure the responses of urinary osmolality and flow rate to a period of dehydration and, subsequently, to the administration of an exogenous vasopressin preparation. Various dehydration tests have been described in which changes in plasma and urine osmolalities in patients with polyuria were compared to the responses of healthy persons.19,20 and 21 In theory, differentiation between hypothalamic diabetes insipidus, nephrogenic diabetes insipidus, and dipsogenic diabetes insipidus should be readily possible, but the disorders actually can be diagnosed correctly only in certain circumstances. Frequently, tests yield equivocal results. For example, although patients with primary polydipsia might be anticipated to have significantly lower plasma osmolalities and sodium concentrations than patients with hypothalamic or nephrogenic diabetes insipidus, such a distinction is useful diagnostically in only a few cases.22 The reason for the lack of differentiation is the wide variation in the setpoint of the osmoregulatory mechanisms for thirst and vasopressin secretion. Even after fluid deprivation and the administration of exogenous vasopressin, urine osmolality frequently fails to attain normal values (Fig. 26-2) because of the secondary nephrogenic diabetes insipidus induced by prolonged polyuria, as explained earlier. A similar ambiguity arises in the interpretation of results from patients with hypothalamic diabetes insipidus. In theory, these patients, who have low circulating concentrations of vasopressin, should demonstrate a substantial increase in urine osmolality to the normal range in response to exogenous vasopressin. In practice, however, they fail to do so (see Fig. 26-2). Again, the reason for the inadequate urinary response lies in the washout of solute from the renal medullary interstitium. The greater the 24-hour urine volume, the greater is the degree of renal resistance to antidiuretic activity.2 The large overlap in urine osmolality values with these tests is illustrated explicitly in Figure 26-2. The refinement of dehydration tests by calculation of free water clearance adds little to their ability to distinguish among the causes of diabetes insipidus. Other means of evaluating the osmoregulatory system using indirect methods to assess antidiuretic activity in an attempt to identify the cause of polyuria (i.e., infusion of hypertonic saline23) also fail to establish an unequivocal diagnosis for the reasons described earlier, as well as because the saline load induces a solute diuresis. Thus, indirect tests of vasopressin function are considerably limited in their ability to establish the cause of polyuria.
FIGURE 26-2. Urine osmolality is depicted under basal conditions after a period of fluid deprivation (hydropenia) designed to attain maximum urinary concentration, and after intramuscular injection of 5 U of vasopressin (Pitressin). The test group included healthy subjects (stippled area) and patients with dipsogenic diabetes insipidus (primary polydipsia, PP), hypothalamic diabetes insipidus (HDI), or nephrogenic diabetes insipidus (NDI). The bars represent the range of results and the closed circle indicates the mean value for the group. (Adapted from Robertson, GL. Diagnosis of diabetes insipidus. In: Czernichow P, Robinson AG, eds. Diabetes insipidus in man. Frontiers of hormone research, vol 13. Basel: S Karger, 1985:176.)
The introduction of sensitive and specific radioimmunoassays capable of detecting low physiologic concentrations of vasopressin in plasma not only has clarified and simplified the diagnosis of diabetes insipidus, but also has extended the understanding of its underlying pathophysiologic mechanisms. The measurement of plasma vasopressin, plasma osmolality, and urine osmolality under basal conditions affords little diagnostic discrimination. However, after osmotic stimulation by a period of fluid deprivation, an infusion of hypertonic saline, or both, the estimation of these indices provides a precise diagnosis.
PLASMA VASOPRESSIN IN THE DIAGNOSIS OF DIABETES INSIPIDUS
Hypothalamic Diabetes Insipidus. Hypothalamic diabetes insipidus is recognized by the subnormal plasma concentrations of vasopressin in relation to plasma osmolality that occur in affected persons (Fig. 26-3A). A clear distinction may be made between patients with hypothalamic diabetes insipidus, patients with nephrogenic diabetes insipidus or dipsogenic diabetes insipidus, and healthy persons by assessing plasma osmolality as it is increased by the infusion of hypertonic saline.24 Many patients with diabetes insipidus have detectable plasma vasopressin, which represents the partial form of the disorder. The ability to secrete vasopressin at high plasma osmolality levels partially explains the ability of some patients to generate a concentrated, if submaximal, urine after fluid deprivation. A few patients fail to exhibit detectable immunoreactive plasma vasopressin, despite marked increases in plasma osmolality. However, some of these patients still manage to concentrate their urine to some extent, suggesting that their kidneys are particularly sensitive to very low concentrations of vasopressin. Occasionally, patients with hypothalamic diabetes insipidus clearly demonstrate osmotically regulated vasopressin release (see Fig. 26-3A). In such patients, the theoretic threshold for vasopressin release, obtained from the abscissal intercept of the osmoregulatory line (the function relating plasma, vasopressin to plasma osmolality), is normal, but the slope that defines the sensitivity of osmotically regulated vasopressin release is significantly reduced. Thus, the osmoreceptor controlling vasopressin secretion is probably intact in this group of patients.
FIGURE 26-3. A, Plasma osmolality and vasopressin responses to infusion of 5% hypertonic saline solution in representative patients with hypothalamic diabetes insipidus (HDI, ·—·), nephrogenic diabetes insipidus (NDI, n—n), and dipsogenic diabetes insipidus (DDI,
). B, Plasma vasopressin and urine osmolality responses to a period of dehydration in patients with hypothalamic diabetes insipidus (·), nephrogenic diabetes insipidus (n), and dipsogenic diabetes insipidus (
). The responses of healthy subjects are indicated by the stippled areas. The limit of detection of the vasopressin assay (LD) was 0.3 pmol/L (1 pmol/L º 1.1 pg/mL).
Some patients who have been treated with parenteral neurohypophyseal extract (Pitressin) have developed antibodies to vasopressin. These patients exhibit “vasopressin” in their plasma as a direct result of assay interference, but this can be recognized by laboratory testing simply through detection of plasma-binding activity to synthetic vasopressin. Therefore, screening for vasopressin antibodies in all patients treated with Pitressin is wise to prevent spurious plasma results.
After hypertonic saline infusion, patients with primary polydipsia or nephrogenic diabetes insipidus have plasma vasopressin and plasma osmolality values that fall within the normal reference range (see Fig. 26-3A). A supranormal plasma vasopressin response to rising plasma osmolality has been demonstrated in a few patients. Whether this response is attributable to a state of chronic underhydration is not known. Thus, hypothalamic diabetes insipidus is clearly distinguishable from other forms of diabetes insipidus by relating plasma vasopressin to plasma osmolality after osmotic stimulation.
Nephrogenic Diabetes Insipidus Versus Primary Polydipsia. Analysis of the relationship between plasma vasopressin and urine osmolality after a period of fluid deprivation offers a potential means of differentiating nephrogenic diabetes insipidus from dipsogenic diabetes insipidus (see Fig. 26-3B). Patients with nephrogenic diabetes insipidus have plasma vasopressin concentrations that are inappropriately high in relation to urine osmolality. However, prolonged polyuria from any cause can induce renal resistance to vasopressin, with blunting of the maximal urinary concentration in response to vasopressin. This difficulty can be partially overcome by examining the basal values of plasma vasopressin and urine osmolality. Plasma vasopressin tends to be detectable or even elevated in nephrogenic diabetes insipidus, whereas immunoreactive vasopressin is generally undetectable in primary polydipsia2 (see Fig. 26-3B). The administration of exogenous vasopressin after a period of fluid deprivation does not help to discriminate further between the causes of polyuria.
Close examination of the data regarding plasma vasopressin and urine osmolality in patients with partial hypothalamic diabetes insipidus reveals an inappropriately high urine osmolality in relation to the low plasma vasopressin concentrations (see Fig. 26-3B). This observation confirms the earlier impression that the renal tubule may become extraordinarily sensitive to vasopressin. Because sustained polyuria from any cause induces a state of secondary partial nephrogenic diabetes insipidus, the administration of exogenous vasopressin to patients with partial hypothalamic diabetes insipidus fails to induce maximal urinary osmolality.
A systematic study to compare the diagnostic efficacy of indirect tests with direct measurement of osmotically stimulated vasopressin release has clearly demonstrated that direct plasma vasopressin measurement methods are superior.25
If the clinician does not have ready access to suitable assays for plasma vasopressin measurement, a satisfactory diagnosis may be established using a closely monitored, prolonged therapeutic trial with desmopressin, administered intramuscularly in daily dosages of 1 µg for as long as 7 days, preferably while patients are hospitalized. Patients with hypothalamic diabetes insipidus who undergo this regimen show an improvement in the degree of polyuria and a reduction in polydipsia. At the end of the trial, a standard water deprivation test, followed by the administration of exogenous vasopressin using indirect assessment methods, may demonstrate maximal urinary concentrations that are within the normal reference range. This is because, during the period of the trial, the renal medullary interstitial solute concentration was restored. Patients with primary polydipsia experience progressive hyponatremia and gain weight because they continue to drink fluid despite persistent antidiuresis. Some of these patients run the risk of neurologic disturbances, particularly seizures, secondary to the development of sudden, profound hyponatremia. Finally, patients with nephrogenic diabetes exhibit little, if any, improvement in thirst or polyuria.
Even the most carefully conducted therapeutic trial, however, can yield misleading results. For example, a few patients who have had severe hypothalamic diabetes insipidus for many years develop water intoxication when first treated with vasopressin because they initially fail to reduce their water intake appropriately, and therefore they appear to have primary polydipsia. Thus, with the currently available diagnostic investigations, the measurement of plasma vasopressin levels with plasma and urine osmolalities during osmotic stimulation provides the most reliable method for determining the cause of polyuria.
THIRST IN HYPOTHALAMIC DIABETES INSIPIDUS
The thirst mechanism in patients with hypothalamic and nephrogenic diabetes insipidus generally operates normally.26 Such patients rely on an intact thirst appreciation to maintain water balance. In one study that documented the osmolar threshold for the onset of thirst, no significant difference was found between the mean value for a group of 11 patients with hypothalamic diabetes insipidus and that for 11 healthy persons; however, the range of values was wider in the affected patients. Only a few patients who are treated with vasopressin and who have inappropriate persistent thirst develop episodes of hyponatremia. When thirst appreciation is blunted, hypernatremia develops.
DIABETES INSIPIDUS AND PREGNANCY
Normal human pregnancy is associated with subtle changes in osmoregulation. Plasma osmolality falls by 8 to 10 mOsm/kg due to a lowering of the osmolar thresholds for both thirst and vasopressin release,27 and a small reduction occurs in maximal urinary concentrating ability.
Established hypothalamic diabetes insipidus appears to have little effect on fertility, gestation, delivery, and lactation in humans.28 A few studies have demonstrated that the secretion of oxytocin is normal in patients with diabetes insipidus. By contrast, many pregnant patients with hypothalamic diabetes insipidus notice a worsening of their polyuria and polydipsia.29 The mechanisms responsible for this may include depression of the osmolar thirst threshold to levels at which vasopressin secretion is suppressed further; circulating vasopressin that is degraded by the placental enzyme cystine aminopeptidase; and renal resistance to vasopressin, which is increased. Some patients show no change in their polyuria, whereas a few improve.
Transient central and nephrogenic diabetes insipidus associated with pregnancy has been documented in some patients30,31 and has been found to recur in subsequent pregnancies. Whether this is attributable to an exaggeration of normal physiologic adaptation to pregnancy or represents a distinct disease entity remains unresolved.
Hypothalamic diabetes insipidus can be treated by the ingestion of adequate volumes of water, with patients relying on the quenching of thirst as the sole indicator of sufficient intake. Some patients who have had severe polyuria since childhood may prefer to manage their symptoms in this manner, thereby avoiding the need for medication. They organize their lives around the inconveniences of frequent micturition and copious drinking. However, good reasons exist why all patients with moderate to severe polyuria should be treated more actively. Prolonged, severe polyuria can result in distention and atonia of the bladder and hydroureter, and, eventually, in hydro-nephrosis with consequent renal damage. Susceptibility to potassium deficiency is another potential risk. Furthermore, untreated patients, if deprived of fluid for any reason, are at risk for the development of life-threatening hypernatremia and dehydration. In young children, withholding therapy may lead to failure to thrive. For these reasons, as well as for the relief of symptoms, antidiuretic treatment is advised for patients with 24-hour urine volumes exceeding 4 L.
Hormone replacement therapy using arginine vasopressin, a natural endogenous peptide, is inappropriate for most patients with hypothalamic diabetes, regardless of whether it is administered parenterally or intranasally. The peptide has a short half-life and may be associated with significant pressor side effects that render the use of aqueous vasopressin preparations impractical. However, during the last three decades, considerable advances have been made in the development of synthetic vasopressin analogs with various agonist and antagonist activities to the pressor (V1) and antidiuretic (V2) receptors. One class of analogs with minimal pressor activity but increased antidiuretic potency and some resistance to degradation in vivo has been developed to treat hypothalamic diabetes insipidus. The current drug of choice is desmopressin.32,33 For adults, it can be administered orally, 50 to 400 µg one to three times daily; intranasally, 5 to 40 µg once or twice daily; or parenterally, 0.5 to 2.0 µg daily. A wide variation is seen in individual desmopressin requirements for the control of polyuria. For children, the dosage is halved. Desmopressin is not associated with pressor agonist side effects but carries the potential hazard of dilutional hyponatremia if patients continue to drink inappropriately despite persistent antidiuresis.
If desmopressin proves to be too potent, it can be diluted. Alternatively, a shorter-acting preparation—lysine vasopressin—can be administered intranasally. However, because it possesses pressor activity, it may induce vasoconstriction, angina, or renal and intestinal colic when taken in excess. If desmopressin is unavailable, it still may be possible to obtain vasopressin tannate in oil, a crude extract of bovine neurohypophysis containing arginine vasopressin suspended in peanut oil. Before intramuscular injection, this preparation should be warmed and shaken vigorously until the extract is evenly distributed in the oil. A single dose of 5 to 10 IU provides as much as 72 hours of antidiuresis. However, this agent is associated with pressor side effects similar to those of lysine vasopressin, an erratic absorption rate, and the formation of sterile abscesses. Pitressin also has been administered as a nasal insufflation.
Now that desmopressin is established as the drug of choice, little need exists to prescribe the partially effective oral agents—chlorpropamide, carbamazepine, clofibrate, or thiazide diuretics—because all are associated with significant and sometimes dangerous side effects.
Rarely, direct treatment of the underlying cause of hypothalamic diabetes insipidus relieves the symptoms. Documented examples include corticosteroid therapy for hypothalamic sarcoidosis, cyclophosphamide therapy for Wegener granulomatosis, and radiotherapy for metastatic disease of the hypothalamus.
Effective treatment of nephrogenic diabetes insipidus still poses problems, except for the forms that are drug induced or related to metabolic disorders (see Table 26-1). The latter are frequently reversible after withdrawal of the drug or correction of the metabolic disturbance. Profound polyuria secondary to the familial forms of this disease is particularly difficult to treat. Restriction of sodium intake, combined with the administration of a thiazide diuretic, reduces urine output by almost 40% in infants. A similar reduction in urine flow may be achieved with the prostaglandin synthetase inhibitor indomethacin when it is administered in dosages of 1.5 to 3.0 mg/kg. The most promising results are achieved with the administration of a combined regimen of thiazide, indomethacin, and desmopressin, which reduces diuresis by as much as 80%.
Hyperosmolar or hypernatremic syndromes may be defined as plasma osmolality levels and sodium concentrations of >300 mOsm/kg and >145 mEq/L (145 mmol/L), respectively. Although rare, they constitute a major management challenge.
Transient hyperosmolality may occur after the ingestion of large amounts of salt,34 but most hypernatremic states occur after inadequate water intake. This can occur in any healthy individual in whom the combination of excess fluid loss—from skin, gastrointestinal tract, lungs, or kidneys—and inadequate access to water is found. This occurs most commonly in acute illness in which water intake is compromised by vomiting or impaired consciousness and most vividly in patients with diabetes insipidus, before treatment, or when access to water is denied. In other cases, however, hypernatremia reflects a primary disorder of thirst deficiency (hypodipsia).
A number of conditions are associated with hypodipsia (Table 26-2). One of the more common causes of hypodipsic hypernatremia that the authors have seen is ligation of the anterior communicating artery, after subarachnoid hemorrhage from a berry aneurysm. Other centers have reported that neoplasms account for 50% of such cases.35 Craniopharyngiomas are particularly associated with hypodipsic diabetes insipidus, sometimes in conjunction with other hypothalamus-related disorders, such as polyphagia, weight gain, and abnormal thermoregulation. Survivors of diabetic hyperosmolar coma have been shown to have impaired osmoregulated thirst,35 which suggests that hypodipsia contributes to the development of the hypernatremia, which is characteristic of the condition. In almost every case of hypodipsia, associated abnormalities of vasopressin secretion are seen, a finding that reflects the close anatomic proximity of the osmoreceptors for vasopressin secretion and thirst.
TABLE 26-2. Specific Causes of Hypodipsic Hypernatremia
In young children and the elderly, hypernatremia may be associated with significant degrees of dehydration.36 Infants are at particular risk, and the mortality is high. In this clinical situation, signs are seen of extracellular fluid loss, decreased skin turgor and elasticity, dry and shrunken tongue, tachycardia, and orthostatic hypotension. Affected infants have depressed fontanelles and tachypnea, and their respirations are deep and rapid. Fever is often present, and the temperature may be as high as 40.5°C (105°F). Adults with mild hypernatremia may have no symptoms, but as plasma sodium levels rise above 160 mEq/L, neurologic signs become apparent.36,37 Early symptoms include lethargy, nausea, and tremor, which progress to irritability, drowsiness, and confusion. Later features of muscular rigidity, opisthotonus, seizures, and coma reflect generalized cerebral and neuromuscular dysfunction. The most severe neurologic disturbances are seen at both ends of the age spectrum. The severity of such disturbances is also related to the rate at which hypernatremia develops, as well as to the absolute degree of hyperosmolality. Intracerebral vascular lesions are often the cause of death.
In contrast to patients with the life-threatening clinical features of hypernatremic dehydration, patients with long-standing, moderate hypernatremia (plasma sodium concentrations of 145 to 160 mEq/L) may have few manifestations of the disorder other than lack of thirst. Hypodipsia is the crucial symptom, but it is often overlooked in the clinical setting because patients fail to complain of lack of thirst. However, careful evaluation of these patients reveals that some have no desire to drink any fluid under any circumstances, which suggests a total loss of the thirst osmoreceptor function. Others have only minimal thirst with marked hypertonicity, whereas a third group eventually experiences a normal thirst sensation, but only at high plasma osmolality levels.
The key to recognizing subtle differences in thirst appreciation rests with a satisfactory measure of thirst. Visual analog scales for measuring thirst during dynamic tests of osmoregulation38,39 have been shown to produce highly reproducible results.40 When these scales are used in evaluating healthy persons, a linear increase is noted in the degree of thirst and fluid intake with increase in plasma osmolality, and an osmolar threshold for thirst is seen that is a few milliosmoles per kilogram higher than the osmolar threshold for vasopressin secretion.39 The application of these techniques to patients with chronic hypernatremia has disclosed numerous disorders of osmoregulation.
OSMOREGULATORY DEFECTS IN CHRONIC HYPERNATREMIA
Chronic hypernatremia is characterized by inappropriate lack of thirst despite increased plasma osmolality and mild hypovolemia. Plasma sodium concentrations are typically elevated (150–160 mmol/L) and may reach extremely high concentrations during intercurrent illnesses (e.g., gastroenteritis) in which body water deficits increase. Although adipsic hyper-natremia is uncommon, four distinct patterns of abnormal osmoregulatory function have been described.
Type 1 Adipsia. The characteristic abnormalities in type 1 adipsia are subnormal vasopressin levels and thirst responses to osmotic stimulation (Fig. 26-4). The sensitivity of the osmoreceptors is decreased, producing partial diabetes insipidus and relative hypodipsia. Because some capacity remains to secrete vasopressin and experience thirst, such patients are protected from extremes of hypernatremia, as they can produce near-maximal antidiuresis as plasma osmolality increases. Patients with this type of adipsia usually have normal vasopressin responses to hypotension and hypoglycemia, and show suppression of vasopressin secretion, with the development of hypotonic diuresis in response to water loading.
FIGURE 26-4. Thirst and vasopressin responses to osmotic stimulation in adipsic hypernatremia. Type 1: subnormal response of both thirst and vasopressin secretion. Type 2: total lack of response of thirst and vasopressin secretion. Type 3: reset of osmostat for vasopressin release and thirst to the right of normal. Shaded areas indicate the response ranges in healthy control subjects; the dotted lines are the mean regression lines. (pAVP, plasma arginine vasopressin; LD, limit of detection of the pAVP assay [0.3 pmol/L]).
Type 2 Adipsia. Total ablation of the osmoreceptors produces complete diabetes insipidus and absence of thirst in response to hyperosmolality. This is the pattern of osmoregulatory abnormality seen after surgical clipping of aneurysms of the anterior communicating artery,41,42 and despite the complete absence of osmoregulated thirst and vasopressin release, thirst and vasopressin responses to hypotension and apomor-phine are preserved.41,43 Some patients also develop this type of osmoregulatory dysfunction after surgery for large, suprasellar craniopharyngiomas. Interestingly, these patients also have absent baroregulated thirst and vasopressin secretion—presumably because the extent of surgical injury is such that both the osmoreceptors and the paraventricular and supraoptic nuclei are damaged. Patients with complete adipsic diabetes insipidus have no defense against dehydration, and unless they are closely supervised and trained to drink even in the absence of thirst, they can develop profound hypernatremic dehydration, even in the absence of intercurrent illness.
Interest has been shown in the concept that osmoreceptor activity is under bimodal control; that is, a specific stimulus is required to switch off vasopressin secretion in the same way that elevation of plasma osmolality stimulates vasopressin secretion. Patients with complete osmoreceptor ablation clearly are unable to respond to inhibitory inputs; this has been demonstrated in clinical studies in which complete suppression of the secretion of the small quantities of radioimmunoassayable vasopressin or the achievement of maximal free water clearance during water loading was impossible in a patient with this type of osmoregulatory dysfunction.44 Therefore, in some patients vasopressin secretion may not be entirely suppressed during fluid loads, resulting in significant hyponatremia.
Type 3 Adipsia. The osmostats for thirst and vasopressin release may be reset to the right of normal (type 3 in Fig. 26-4), such that vasopressin secretion and thirst do not occur until higher plasma osmolalities are reached. Thereafter, the slope of the osmoregulatory lines are normal. This pattern is found in conjunction with a number of cases of “essential” hypernatremia, although type 1 defects have also been reported.45,46 and 47 Patients also have intact nonosmotic release of vasopressin and increased renal sensitivity to vasopressin, so that renal concentrating ability may be reasonably well maintained.
Miscellaneous Causes of Adipsia. Osmoregulatory dysfunction has also been reported in elderly patients, who have diminished thirst in response to hypernatremia.38 Although the defect in thirst appreciation is similar to that in type 1 dysfunction, vasopressin responses have variously been reported as being subnormal, normal, or enhanced. Survivors of diabetic hyper-osmolar, nonketotic coma have also been reported to have hypodipsia with exaggerated vasopressin secretion.35 In addition, a single case has been reported of a young patient who had hypodipsia but a normal osmotically regulated vasopressin release.48 All of these reports lend support to the hypothesis that the osmoreceptors subserving vasopressin release are anatomically and functionally distinct from those controlling thirst.
Water replacement is the basic therapy for patients with hyper-osmolar states associated with dehydration. The oral route is preferred, but if the clinical situation warrants urgent treatment, the infusion of hypotonic solutions may be necessary. However, overzealous rehydration with hypotonic fluids may result in seizures, neurologic deterioration, coma, and even death secondary to cerebral edema.34,37 Therefore, the decision to treat with hypotonic intravenous fluids should not be made lightly, and rehydration to a euosmolar state should proceed cautiously over at least 72 hours. As plasma osmolality falls, polyuria indicative of hypothalamic diabetes insipidus may develop; this responds to administration of desmopressin.
For patients with chronic hyperosmolar syndromes (see Fig. 26-4), longer-term therapy must be considered. Patients with type 3 defects rarely need specific therapy because their osmoregulatory system is essentially intact but operates around a higher than normal plasma osmolality. Patients with type 1 defects (involving partial destruction of the osmoreceptor) should be treated with a regimen of increased water intake (2–4 L every 24 hours). If this leads to persistent polyuria, a small dose of desmopressin can be administered, but plasma osmolality or sodium levels must then be monitored regularly.
Considerable difficulties arise in treating patients who have complete destruction of their osmoreceptors (type 2 defect), because these patients cannot protect themselves from extremes of dehydration and overhydration. Most patients need between 2 and 4 L of fluid per day, but the precise amount varies according to seasonal climatic changes, and the body weight must be monitored daily to provide an index of fluid balance.44 Regular (usually weekly) measurements of plasma osmolality or sodium are needed to ensure that no significant fluctuations occur in body water, and constant supervision is required to make certain the requisite volume of water is consumed. Despite the most vigorous supervision, such patients are extremely vulnerable to swings in plasma osmolality and are particularly prone to severe hypernatremic dehydration.
De Wardener HE, Herxheimer A. The effect of high water intake on the kidneys’ ability to concentrate urine in man. J Physiol (Lond) 1957; 139:42.
Robertson GL. Diagnosis of diabetes insipidus. In: Czernichow P, Robinson AG, eds. Diabetes insipidus in man. Frontiers of hormone research, vol 13. Basel: S Karger, 1985:176.
Maffly RH. Diabetes insipidus. In: Andreoli TE, Grantham JJ, Rector FCJ, eds. Disturbances in body fluid osmolality. Bethesda, MD: American Physiology Society, 1977:285.
Robertson GL. Diabetes insipidus. Endocrinol Metab Clin North Am 1995; 24:549.
Ito M, Mori Y, Oiso Y, Saito H. A single base substitution in the coding region for neurophysin II associated with familial central diabetes insipidus. J Clin Invest 1991; 87:725.
Krishnamani MRS, Philips PA III, Copeland KC. Detection of a novel arginine vasopressin defect by dideoxy fingerprinting. J Clin Endocrinol Metab 1993; 77:596.
McLeod JF, Kovacs L, Gaskill MB, et al. Familial neurohypophyseal diabetes insipidus associated with a signal peptide mutation. J Clin Endocrinol Metab 1993; 77:599A.
Heppner C, Kotzka J, Bullmann C, et al. Identification of mutations of the arginine vasopressin-neurophsia II gene in two kindreds with familial central diabetes insipidus. J Clin Endocrinol Metab 1998; 83:693.
Rotig A, Cormier V, Chatelain P. Deletion of the mitochondrial DNA in a case of early-onset diabetes mellitus, optic atrophy and deafness, Wolfram syndrome (MIM 222300). J Clin Invest 1993; 91:1095.
Barrett TG, Bundey SE. Wolfram (DIDMOAD) syndrome. J Med Genet 1997; 34:838.
Verbalis JG, Robinson AG, Moses AM. Postoperative and post-traumatic diabetes insipidus. In: Czernichow P, Robinson AG, eds. Diabetes insipidus in man. Frontiers of hormone research, vol 13. Basel: S Karger, 1985:247.
Moses AM. Clinical and laboratory observations in the adult with diabetes insipidus and related syndromes. In: Czernichow P, Robinson AG, eds. Diabetes insipidus in man. Frontiers of hormone research, vol 13. Basel: S Karger, 1985:156.
Baylis PH, Cheetham T. Diabetes insipidus. Arch Dis Child 1998; 79:84.
Scherbaum WA, Bottazzo GF. Autoantibodies to vasopressin cells in idiopathic diabetes insipidus: evidence for an autoimmune variant. Lancet 1983; 1:897.
Imura H, Nakao K, Shimatsu A, et al. Lymphocytic infundibuloneurohypo-physitis as a cause of central diabetes insipidus. N Engl J Med 1993; 329:683.
Knoers NV, Monnens LL. Nephrogenic diabetes insipidus. Semin Nephrol 1999; 19:344.
Bendz H, Aurell M. Drug-induced diabetes insipidus. Drug Saf 1999; 21:449.
Bichet DG, Birnbaumer M, Louergan M, et al. Nature and recurrence of AVPR2 mutations in X-linked nephrogenic diabetes insipidus. Am J Hum Genet 1994; 55:278.
Hochberg Z, van Lieburg A, Even L, et al. Autosomal recessive nephrogenic diabetes insipidus caused by an aquaporin-2 mutation. J Clin Endocrinol Metab 1997; 82:686.
Sato N, Ishizaka H, Yagi H, et al. Posterior lobe of the pituitary in diabetes insipidus: dynamic MR imaging. Radiology 1993; 186:357.
Dashe AM, Cramm RE, Crist CA, et al. A water deprivation test for the differential diagnosis of polyuria. JAMA 1963; 185:699.
Miller MT, Dalakos T, Moses AM, et al. Recognition of partial defects in antidiuretic hormone secretion. Ann Intern Med 1970; 73:721.
Baylis PH. Diabetes insipidus. Medicine 1997; 25:9.
Robertson GL. The regulation of vasopressin function in health and disease. Recent Prog Horm Res 1977; 33:333.
Moses A, Streeten D. Differentiation of polyuric states by measurement of responses to changes in plasma osmolality induced by hypertonic saline infusions. Am J Med 1967; 42:368.
Baylis PH, Robertson GL. Vasopressin response to hypertonic saline infusion to assess posterior pituitary function. J R Soc Med 1980; 73:255.
Zerbe RL, Robertson GL. A comparison of plasma vasopressin measurement with a standard indirect test in the differential diagnosis of polyuria. N Engl J Med 1981; 305:1539.
Thompson CJ, Baylis PH. Thirst in diabetes insipidus: clinical relevance of quantitative assessment. Q J Med 1987; 65:853.
Davison JM, Gilmore EA, Dürr J, et al. Altered osmotic thresholds for vasopressin secretion and thirst in human pregnancy. Am J Physiol 1984; 246:F105.
Amico J. Diabetes insipidus in pregnancy. In: Czernichow P, Robinson AG, eds. Diabetes insipidus in man. Frontiers in hormone research, vol 13. Basel: S Karger, 1985:266.
Hime MC, Richardson JA. Diabetes insipidus and pregnancy: case report, incidence and review of the literature. Obstet Gynecol Surv 1978; 33:375.
Barron WM, Cohen LH, Ulland LA, et al. Transient vasopressin-resistant diabetes insipidus of pregnancy. N Engl J Med 1984; 310:442.
Hughes JM, Barron WM, Vance ML. Recurrent diabetes insipidus associated with pregnancy: pathophysiology and therapy. Obstet Gynecol 1989; 73:462.
Cobb WE, Spare S, Reichlin S. Diabetes insipidus: management with DDAVP (1-desamino-8-D-arginine vasopressin). Ann Intern Med 1978; 88:183.
Williams TDM, Dungar DB, Lyon CC, et al. Antidiuretic effect and pharma-cokinetics of oral 1-desamino-8-D-arginine vasopressin. 1. Studies in adults and children. J Clin Endocrinol Metab 1986; 63:129.
Ross EJ, Christie SBM. Hypernatremia. Medicine (Baltimore) 1969; 48:441.
McKenna K, Morris AM, Azam H, et al. Subnormal osmotically stimulated thirst and exaggerated vasopressin release in human survivors of hyperosmolar coma. Diabetologia May 1999; 42:538.
Robertson GL, Aycinena P, Zerbe RL. Neurogenic disorders of osmoregulation. Am J Med 1982; 72:339.
Arieff AL, Guisado R. Effects on the central nervous system of hypernatremic and hyponatremic states. Kidney Int 1976; 10:104.
Phillips PA, Rolls BJ, Ledingham JGG, et al. Reduced thirst after water deprivation in healthy elderly men. N Engl J Med 1984; 311:753.
Thompson CJ, Thompson J, Burd J, Baylis PH. The osmotic threshold for thirst and vasopressin release are similar in healthy men. Clin Sci 1986; 71:651.
Thompson CJ, Selby P, Baylis PH. Reproducibility of osmotic and nonosmotic tests of vasopressin secretion in men. Am J Physiol 1991; 260:R533.
Pearce SHS, Argent NB, Baylis PH. Chronic hypernatremia due to impaired osmoregulated thirst and vasopressin secretion. Acta Endocrinol (Copenh) 1991; 125:234.
McIver B, Connacher A, Whittle A, et al. Adipsic diabetes insipidus after clipping of anterior communicating artery aneurysm. BMJ 1991; 303:1465.
Teelucksingh S, Steer CR, Thompson CJ, et al. Hypothalamic syndrome and central sleep apnea associated with toluene exposure. Q J Med 1991; 286:185.
Ball SG, Vaidja B, Baylis PH. Hypothalamic adipsic syndrome: diagnosis and management. Clin Endocrinol 1997; 47:405.
De Rubertis FR, Michelis MF, Beck N, et al. “Essential” hypernatremia due to ineffective osmotic and intact volume regulation of vasopressin secretion. J Clin Invest 1971; 50:97.
Dunger DB, Seckl JR, Lightman SL. Increased renal sensitivity to vasopressin in two patients with essential hypernatremia. J Clin Endocrinol Metab 1987; 64:185.
Gill G, Baylis PH, Burn J. A case of “essential” hypernatremia due to resetting of the osmostat. Clin Endocrinol (Oxf) 1985; 22:545.
Hammond DN, Moll GW, Robertson GL, Chelmicks-Schorr E. Hypodipsic hypernatremia with normal osmoregulation of vasopressin. N Engl J Med 1986; 315:433.