UEU-co logo

MD Consult: Books: Goldman: Cecil Medicine: Chapter 243 – POSTERIOR PITUITARY

Goldman: Cecil Medicine, 23rd ed.

Copyright © 2007 Saunders, An Imprint of Elsevier


Joseph G. Verbalis


The hormones of the posterior pituitary, vasopressin and oxytocin, are synthesized in specialized neurons in the hypothalamus, the neurohypophysial neurons. These neurons are specialized for synthesis and secretion of each hormone and are notable for their large size, hence their designation as magnocellular neurons. In the hypothalamus, the neurohypophysial magnocellular neurons are clustered in the paired paraventricular and supraoptic nuclei ( Fig. 243-1 ). Vasopressin and oxytocin are also synthesized in parvicellular (i.e., small cell) neurons of the paraventricular nuclei, and vasopressin (but not oxytocin) is also synthesized in the suprachiasmatic nucleus.

FIGURE 243-1  Sagittal view of the head demonstrating the position of the neurohypophysis. The magnocellular neurons are clustered in two paraventricular nuclei (PVN) and two supraoptic nuclei (SON). Only one nucleus of each pair is illustrated. The supraoptic nuclei are lateral to the edge of the optic chiasm, whereas the paraventricular nuclei are central along the wall of the third ventricle. The axons of the four nuclei combine to form the supraopticohypophysial tract as they course through the pituitary stalk to their storage terminals in the posterior pituitary. The osmostat (Osm) is in the hypothalamus anterior to the third ventricle; the thirst center (Thirst) is distributed across different brain areas. Ant. Pit. = anterior pituitary.  (From Buonocore CM, Robinson AG: Diagnosis and management of diabetes insipidus during medical emergencies. Endocrinol Metab Clin North Am 1993;22:411-423.)

Transcription of vasopressin and oxytocin mRNA and translation of the vasopressin and oxytocin prohormones occur entirely in the cell bodies of the neurohypophysial neurons. The pre-prohormones are cleaved from the signal peptide in the endoplasmic reticulum, and the prohormones provasopressin and pro-oxytocin are packaged along with processing enzymes into neurosecretory granules. The neurosecretory granules are transported out of the perikaryon of the neurohypophysial neurons through microtubules down the long axons that form the supraopticohypophysial tract to terminate in axon terminals in the posterior pituitary. The entire unit including the magnocellular neurons in the supraoptic and paraventricular nuclei, the supraopticohypophysial tract, and the axon terminals in the posterior pituitary is called the neurohypophysis. During transport, the processing enzymes cleave provasopressin into vasopressin (9 amino acids), vasopressin-neurophysin (95 amino acids), and vasopressin glycopeptide (39 amino acids). Pro-oxytocin is similarly cleaved to oxytocin (which differs from vasopressin by only two of nine amino acids) and oxytocin-neurophysin, but there is no glycopeptide contained in pro-oxytocin. Within the neurosecretory granules, neurophysins form neurophysin-hormone complexes that stabilize the hormones. Crystallography has demonstrated that tetramers of neurophysin form specific binding sites for five molecules of hormone, so the hormone within the granules is always bound. Stimulatory (e.g., glutamatergic, cholinergic, and angiotensin) neurotransmitter terminals and inhibitory (e.g., γ-aminobutyric acid and noradrenergic) neurotransmitter terminals control the release of vasopressin by the activity of synaptic contacts on the cell bodies. Physiologic release of vasopressin or oxytocin into the general circulation occurs at the level of the posterior pituitary, where, in response to an action potential, intracellular calcium is increased and causes the neurosecretory granules to fuse with the axon membrane, thereby releasing the entire contents of the granule through exocytosis into the pericapillary space. Once it is released, each hormone has no further association with its respective neurophysin, and each of the peptide products can be independently detected in the general circulation. Although vasopressin and oxytocin compose only small parts of their respective prohormones, they are the only known biologically active components of the prohormones. Factors that stimulate the release of neurohypophysial hormones also stimulate their synthesis; however, whereas release is instantaneous, synthesis requires a longer time. Because synthesis is delayed, maintenance of a large store of hormone in the posterior pituitary is essential for the instantaneous and massive release of each hormone that is necessary with acute hemorrhage (vasopressin) or during parturition (oxytocin). In most species, sufficient vasopressin is stored in the posterior pituitary to support maximum antidiuresis for several days and to maintain baseline levels of antidiuresis for weeks without ongoing synthesis of new hormone.

The axons of the parvicellular neurons of the paraventricular nuclei project to different areas within the brain rather than to the posterior pituitary. Some terminate in the median eminence of the basal hypothalamus, where, similar to other hypothalamic releasing factors, the hormones are secreted into the portal capillary system and where vasopressin serves as one of the regulators of secretion of adrenocorticotropic hormone. Other neurons project to the limbic system, the brain stem, and the spinal cord, where vasopressin and oxytocin serve neurotransmitter and neuromodulatory roles. Still other axons secrete hormones into the cerebrospinal fluid of the third ventricle, the function of which is unknown.


Vasopressin and Regulation of Osmolality

The primary physiologic action of vasopressin is its function as a water-retaining hormone. The central sensing system (osmostat) for control of release of vasopressin is anatomically discrete, located in a small area of the hypothalamus just anterior to the third ventricle that also includes the circumventricular organs the organum vasculosum of the lamina terminalis and the subfornical organ (see Fig. 243-1 ). The osmostat controls release of vasopressin to cause water retention and also stimulates thirst to cause water repletion. Osmotic regulation of vasopressin release and osmotic regulation of thirst are usually tightly coupled, but experimental lesions and some pathologic situations in humans demonstrate that each can be regulated independently. The primary extracellular osmolyte to which the osmoreceptor responds is sodium. Under normal physiologic conditions, glucose and urea readily traverse neuron cell membranes and do not stimulate release of vasopressin. Although basal osmolality in normal subjects lies between 280 and 295 mOsm/kg H2O, extracellular fluid osmolality for each individual is maintained within narrow ranges. Increases in plasma osmolality as small as 1% will stimulate the osmoreceptors to release vasopressin. Basal plasma levels of vasopressin are generally 0.5 to 2 pg/mL, which is sufficient to maintain urine osmolality above plasma osmolality and urine volume in the range of 2 to 3 L/day. When vasopressin levels are suppressed below 0.5 pg/mL, maximum urine osmolality decreases to less than 100 mOsm/kg H2O and a free water diuresis ensues to levels approaching 800 to 1000 mL/hr (18 to 24 L/day). Increases in plasma osmolality cause a linear increase in plasma vasopressin and a corresponding linear increase in urine osmolality. At a plasma osmolality of approximately 295 mOsm/kg H2O, urine osmolality is maximally concentrated to 1000 to 1200 mOsm/kg H2O. Thus, the entire physiologic range of urine osmolality is accomplished by relatively small changes in plasma vasopressin of 0 to 5 pg/mL, as illustrated in Figure 243-2 .

FIGURE 243-2  Idealized schematic of the normal physiologic relationships among plasma osmolality, plasma vasopressin (AVP), urine osmolality, and urine volume. The entire physiologic range of urine osmolality occurs with plasma vasopressin levels from 0 to 5 pg/mL. Increases in plasma osmolality above approximately 290 to 295 mOsm/kg H2O result in increases in plasma vasopressin but no further concentration of the urine, which is limited by the maximal osmolality in the inner medulla. The relation of volume (calculated on the basis of a constant osmolar load) is inversely exponential to the other parameters. Because of this relationship, urine volume does not change substantially until there is nearly absent vasopressin secretion, after which urine volume increases dramatically. The shaded area represents the normal range and the interrelationships among the various parameters.  (Calculated from formulae presented in Robertson GL, Shelton RL, Athar S: The osmoregulation of vasopressin. Kidney Int 1976;10:25-37. Figure drawn by J. G. Verbalis, Georgetown University, Washington, DC.)

To maintain fluid balance, water must be not just conserved but consumed as well to replace insensible water losses and obligate urine output. Most studies have indicated that thirst is not stimulated until a somewhat higher plasma osmolality (5 to 10 mOsm/kg H2O) than the threshold for release of vasopressin. During the course of a normal day, most humans derive sufficient water from habitual fluid intake and catabolism of food to maintain plasma osmolality below the threshold required to activate thirst. Therefore, under normal physiologic conditions, water balance (and hence plasma osmolality) is regulated more by secretion of vasopressin than by true thirst. However, with more severe degrees of dehydration, thirst is essential to restore body water deficits.

Vasopressin acts on V2 or antidiuretic receptors in the collecting duct cells of the kidney to cause water retention, or antidiuresis. Vasopressin V2 receptors are G protein–coupled receptors that activate adenylate cyclase with subsequent increased intracellular cyclic adenosine monophosphate (cAMP) levels upon ligand activation of the receptor. The increased cAMP initiates the movement of aquaporin-2 water channels from the cytoplasm to the apical (luminal) membrane of the collecting duct cells. Once they are inserted into the apical membrane, these channels allow facilitated rapid transport of water from the collecting duct lumen into the cell along osmotic gradients. The water then exits the cell through the basolateral membrane, into the kidney medullary circulation through aquaporin-3 and aquaporin-4 water channels, which are constitutively present in the basolateral membrane. This entire process is termed antidiuresis. In the absence of vasopressin, the aquaporin-2 channels are reinternalized from the apical membrane into subapical vesicles. This prevents active reabsorption of water from the collecting duct lumen, resulting in diuresis. In addition to this rapid “shuttling” of the aquaporin-2 channels to regulate water reabsorption on a minute-to-minute basis, vasopressin also acts through V2 receptors to regulate long-term stores of aquaporin-2; that is, increased vasopressin stimulates aquaporin-2 synthesis, and the absence of vasopressin results in decreased aquaporin-2 synthesis. The hypertonic medullary interstitium determines the maximum concentration of the urine, which is isotonic with the inner medulla of the kidney under conditions of maximal antidiuresis ( Chapter 116 ).

Vasopressin and Pressure and Volume Regulation

In contrast to the osmoregulatory system, volume regulation is anatomically more diffuse. High-pressure baroreceptors are located in the aorta and carotid sinus, and low-pressure baroreceptors are located in the right and left atria. Stimuli for pressure and volume receptors are carried through the glossopharyngeal (ninth) and vagal (tenth) cranial nerves to the nucleus tractus solitarius in the brain stem. Subsequent secondary and tertiary projections converge on the magnocellular neurons, where they provide inhibitory as well as excitatory inputs. Decreases in blood pressure or vascular volume stimulate vasopressin release, whereas situations that increase blood volume or left atrial pressure (e.g., negative-pressure breathing) decrease secretion of vasopressin. The release of vasopressin in response to changes in volume or pressure is less sensitive than the release in response to osmoreceptors, and generally a 10 to 15% reduction in blood volume or pressure is needed to stimulate release of vasopressin. However, once arterial pressure falls below this threshold, the stimulated response is exponential, and plasma levels of vasopressin achieved are markedly greater than those achieved by osmotic stimulation. Other nonosmotic stimuli, such as nausea and intestinal traction, probably act through similar nonosmotic neural pathways to release vasopressin.

The pressor effects of vasopressin are mediated through V1a receptors on vascular smooth muscle. For V1a and V1b receptors, the mechanism of action of vasopressin is to increase intracellular calcium rather than to stimulate adenylate cyclase. In intact animals, the pressor activity of vasopressin is weak because of compensatory vasodilatory systems that act to modulate vasopressor actions. The relatively insensitive regulation of vasopressin secretion by changes in volume and pressure and the modest role of vasopressin to regulate blood pressure are consistent with the notion that regulation of body sodium homeostasis by the renin-angiotensin-aldosterone system is more important for control of extracellular and blood volume than is regulation of water homeostasis. However, the action of vasopressin to regulate blood pressure can become prominent when other blood pressure regulatory systems are deficient (e.g., autonomic neuropathy or renin-angiotensin-aldosterone system blockade) or in states of pathologic vasodilation (e.g., liver cirrhosis, septic shock).

Vasopressin and Adrenocorticotropic Hormone

Vasopressin in the parvicellular neurons whose axons terminate in the median eminence is released into the pituitary portal capillaries and carried to the anterior pituitary. Anterior pituitary corticotrophs are stimulated through V1b receptors to release adrenocorticotropic hormone. Although the major regulator of adrenocorticotropic hormone secretion is corticotropin-releasing hormone, because vasopressin and corticotropin-releasing hormone activate different signal transduction systems in the corticotrophs, each hormone has synergistic effects on the actions of the other to release adrenocorticotropic hormone.

Interaction of Osmotic and Volume Regulation

The vasopressin system has been adapted to optimize mammalian drinking behavior. Water is consumed as available in the absence of stimulated thirst, then vasopressin regulates water excretion to maintain plasma osmolality; this allows extensive geographic movement without thirst that would produce time-consuming and potentially dangerous water-seeking behavior. Yet, thirst serves as a back-up if dehydration becomes excessive. Similarly, because pressure-volume regulation of vasopressin is less sensitive, modest changes in pressure or volume, which are exacerbated by upright posture, do not interfere with regulation of osmolality. Yet, the pressor effect of high vasopressin serves as a back-up if volume depletion or hypotension becomes excessive. Usually, the physiologic regulation of osmolality and pressure-volume are synergistic. Dehydration causes an increase in osmolality and a decrease in volume, both of which stimulate release of vasopressin. Excess administration of fluid causes both a decrease in osmolality and an expansion of volume, both of which inhibit vasopressin secretion. However, other pathologic situations can result in opposing signals to vasopressin secretion (e.g., hyponatremia resulting from diuretic use with a decreased extracellular fluid volume or from cardiac failure or cirrhosis due to a decreased effective arterial blood volume). In these situations, vasopressin secretion represents a balance between the excitatory and inhibitory inputs provided by the osmotic and volume-pressure stimuli.

In addition, other factors can modulate the osmotic release and action of vasopressin. With volume expansion, natriuretic factors such as atrial natriuretic peptide and brain natriuretic peptide are released from atrial myocytes and act at the kidney to induce natriuresis. Atrial natriuretic peptide is also synthesized in the hypothalamus, where it may act to decrease vasopressin secretion. During pregnancy, there is a decrease of plasma osmolality by approximately 10 mOsm/kg H2O as a result of a resetting of the osmostat, whereby vasopressin increases and decreases appropriately around the lower plasma osmolality; the osmostat for thirst is reset in parallel. Data suggest that this effect may be mediated by the placental hormone relaxin.

Abnormalities in water and electrolyte balance are common in elderly humans. This is in part due to age-related changes in body volume (as much as a 50% decrease in total body water above the age of 75 years) and renal function. However, elderly humans also have a decreased appreciation for thirst; and although there is a normal or even increased ability to secrete vasopressin with age, there is a decreased ability to achieve either maximum urine concentration to retain water or maximum dilution of urine to excrete water. Consequently, the elderly are particularly prone to hypernatremia or hyponatremia with diseases that affect water balance or from the drugs used as therapy for these diseases.


Oxytocin is also synthesized in the magnocellular neurons of the neurohypophysis and differs in structure from vasopressin by only two of nine amino acids. Oxytocin has similar concentrations in the posterior pituitary of both men and women, but to date a physiologic function for oxytocin has been described only in women.

Prolactin is the main hormone necessary for milk production, but oxytocin is essential for milk secretion. Suckling stimulates tactile receptors in the nipple, producing an afferent signal to the hypothalamus that causes release of oxytocin from the posterior pituitary. Oxytocin binds to oxytocin receptors in the breast and induces contraction of myoepithelial cells around the alveoli and ductules to eject milk. Oxytocin also participates in parturition, although this action is more complex and is variable among species. In all species, there is interaction of oxytocin with gonadal steroid hormones, prostaglandins, and relaxin. In addition, upregulation of the uterine receptors for oxytocin dramatically increases the response to oxytocin at the end of pregnancy. The interaction of these various hormones in a cross-stimulation feed-forward cascade to support parturition is important to ensure survival of the species, so it is understandable that during parturition, lack of any single hormone (including oxytocin) is generally not sufficient to inhibit delivery. The greatest release of oxytocin occurs with, not before, delivery of the infant, probably secondary to stretching of the vaginal wall. This Fergusson reflex may aid delivery of subsequent fetuses in animals with multiple births; in humans, oxytocin release may be more important to induce uterine contraction to inhibit blood loss after delivery than to initiate parturition.

No pathologic syndromes of either increased or decreased secretion of oxytocin have yet been defined. Women with diabetes insipidus secondary to traumatic damage of the magnocellular neurons often have normal pregnancy and delivery and can breast-feed their infants; but it may be that in these cases, oxytocin neurons survive better than vasopressin neurons. In animal studies, administration of oxytocin to males increases sperm transport, but this function has not been documented in humans. Although only a single receptor for oxytocin has been identified, similar to the receptors for vasopressin, the receptors for oxytocin in the breast and in the myometrium are independently regulated. Because of the structural similarity between vasopressin and oxytocin, at high plasma levels, oxytocin can activate vasopressin receptors and vasopressin can activate oxytocin receptors. One example of this is that administration of oxytocin to induce labor can stimulate V2 receptors of the kidney and cause water retention and hyponatremia if excess fluids are administered simultaneously.

Although peripheral effects of oxytocin have been described only in females, parvicellular oxytocin neurons in the paraventricular nuclei, like those of vasopressin, project to multiple different areas within the brain where oxytocin serves as a neurotransmitter and neuromodulator. Evidence in animals has strongly implicated brain oxytocin in the production of maternal behaviors in females and in affiliative (i.e., affection and bonding) behaviors in both sexes. These may prove to be of equal or greater importance than the peripheral actions of this hormone.


Excess secretion of vasopressin can be caused by abnormal secretion from the posterior pituitary or by ectopic synthesis and secretion of vasopressin by a tumor. The inappropriate secretion of vasopressin causes renal water retention and volume expansion of body fluids, with a secondary natriuresis and consequent hyponatremia. This disorder is called the syndrome of inappropriate antidiuretic hormone secretion and is discussed in Chapter 117 .



Diabetes insipidus is the excretion of a large volume of hypotonic, insipid (tasteless) urine, usually manifested by polyuria (increased urination) and polydipsia (increased thirst). The large volume, usually in excess of 50 to 60 mL/kg/day, must be distinguished from increased frequency of small volumes and from large volumes of isotonic or hypertonic urine, both of which have other clinical significance.


Four pathophysiologic mechanisms must be considered in the differential diagnosis of diabetes insipidus.

   1.    Central diabetes insipidus is caused by the inability to secrete (and usually to synthesize) vasopressin in response to increased osmolality. No concentration of the dilute glomerular filtrate takes place in the renal collecting duct, and consequently a large volume of urine is excreted. This produces an increase in serum osmolality with stimulation of thirst and secondary polydipsia. Levels of vasopressin in plasma are unmeasurable or inappropriately low for plasma osmolality.
   2.    Nephrogenic diabetes insipidus is caused by the inability of an otherwise normal kidney to respond to vasopressin. As in hypothalamic diabetes insipidus, the dilute glomerular filtrate entering the collecting duct is excreted as a large volume of hypotonic urine. The rise in plasma osmolality that occurs stimulates thirst and produces polydipsia. Unlike central diabetes insipidus, however, measured levels of vasopressin in plasma are high or appropriate for plasma osmolality.
   3.    Gestational diabetes insipidus is a rare condition produced by elevated levels or activity of placental cystine aminopeptidase (oxytocinase or vasopressinase) during pregnancy. The rapid destruction of vasopressin produces diabetes insipidus with polyuria and secondary stimulation of thirst with polydipsia. Because of the circulating vasopressinase, plasma vasopressin levels usually cannot be measured.
   4.    Primary polydipsia is a disorder of thirst stimulation rather than of vasopressin secretion or activity. Excessive ingested water produces a mild decrease in plasma osmolality that shuts off secretion of vasopressin. In the absence of vasopressin action on the kidney, urine does not become concentrated, and a large volume of dilute urine is excreted. The amount of vasopressin in plasma is unmeasurable or low but is appropriate for the low plasma osmolality.
Although the pathophysiologic mechanisms for each of these four disorders are distinct, patients in each category usually manifest polyuria and polydipsia. The serum sodium level is normal because the normal thirst mechanism is sufficiently sensitive to maintain water homeostasis in the first three disorders, and the normal kidney has sufficient capacity to excrete the excess water load in the fourth.

Clinical Manifestations

Central Diabetes Insipidus

The sudden appearance of hypotonic polyuria after transcranial surgery in the area of the hypothalamus or after head trauma with basal skull fracture and hypothalamic damage obviously suggests the diagnosis of central diabetes insipidus. In these situations, if the patient is unconscious and unable to recognize thirst, hypernatremia is a common accompaniment. However, even in patients with more insidious progression of a specific disease or in patients with idiopathic central diabetes insipidus, the onset of polyuria is often relatively abrupt and occurs during several days or weeks. The initial problem is the volume of urine and polydipsia, not the decrease in urine osmolality. Most patients do not report polyuria until urine volume exceeds 4 L/day, and as illustrated in Figure 243-2 , urine volume does not exceed 4 L/day until the ability to concentrate the urine is severely limited and plasma vasopressin is nearly absent. As few as 10% of the normal number of vasopressinergic neurons in the hypothalamus can maintain asymptomatic urine volume, but the further loss of these few neurons produces a rapid increase in urine volume and symptomatic polyuria. Urine volume seldom exceeds the amount of dilute fluid delivered to the collecting duct (about 18 to 24 L in humans), and in many cases urine volume is less because patients voluntarily restrict fluid intake, which causes some mild volume contraction and increased proximal tubular reabsorption of fluid. Patients often express a preference for cold liquids, which are more effective in assuaging thirst. Both thirst and increased urine output persist through the night. Patients with partial central diabetes insipidus have some ability to secrete vasopressin, but this secretion is markedly attenuated at normal levels of plasma osmolality. Therefore, these patients have symptoms and urine volume similar to those of patients with complete central diabetes insipidus. Because most patients with central diabetes insipidus have sufficient thirst to drink fluid to match urine output, few laboratory abnormalities are present at the time of initial evaluation. Serum sodium level may be in the high-normal range, whereas blood urea nitrogen level may be low secondary to large urine volume. Uric acid is relatively high due to the modest volume contraction and the lack of action of vasopressin on V1a receptors in the kidney, which stimulate clearance of uric acid. Uric acid levels greater than 5 mg/dL have been reported to distinguish diabetes insipidus from primary polydipsia.

A variant of central diabetes insipidus is the syndrome of osmoreceptor dysfunction. This syndrome has been referred to as essential hypernatremia and adipsic hypernatremia because such patients have a chronically increased serum sodium concentration and an absence of thirst. Physiologic maneuvers demonstrate that when the patients are euvolemic, an increase in plasma osmolality produces neither secretion of vasopressin nor a sensation of thirst. However, vasopressin is still synthesized by the hypothalamus and stored in the posterior pituitary because stimulation of baroreceptors by hypovolemia or hypotension results in prompt secretion of vasopressin; the kidney is responsive because vasopressin release by volume receptor stimulation causes urinary concentration. Because patients lack thirst, they are chronically dehydrated, with increased serum sodium levels. However, it is the dehydration-induced volume depletion, not the increased osmolality, that eventually stimulates secretion of vasopressin. The volume of urine output depends on the degree of dehydration-induced secretion of vasopressin. If sufficient fluid replacement is given to return extracellular fluid volume to normal, these patients are unable to regulate vasopressin by osmolality and then become markedly polyuric, thereby manifesting their underlying diabetes insipidus.

Central diabetes insipidus can be inherited as an autosomal dominant disease that is typically characterized by an asymptomatic infancy but onset later in childhood. Most genetic defects are either in the signal peptide of the pre-prohormone or in the neurophysin portion of the prohormone. Mutations involving the vasopressin sequence itself are few. Abnormal packaging of vasopressin and neurophysin might also produce increased degradation of both the mutant and the normal wild-type prohormone and hence inadequate secretion of vasopressin. However, most cases are believed to result from disruption of cleavage from the signal peptide or abnormal folding of the neurophysin, which slows trafficking of the mutant prohormone through the endoplasmic reticulum, thereby leading to neuronal cell dysfunction and death by virtue of accumulation of the unprocessed mutant protein within the magnocellular neurons. Because this is a cumulative process, this accounts for the later onset of central diabetes insipidus with these types of mutations.

Myxedema and adrenal insufficiency both impair the ability to excrete free water by renal mechanisms. The simultaneous occurrence of either of these diseases with central diabetes insipidus (as can occur with a tumor of the hypothalamus or pituitary) may decrease an otherwise large urine output, thereby masking the symptoms of the diabetes insipidus. Replacement treatment for the anterior pituitary deficiency, especially glucocorticoids, can then cause a sudden and massive excretion of dilute urine. Similarly, the onset of either hypothyroidism or adrenal insufficiency during the course of diabetes insipidus can decrease the need for vasopressin replacement and in some cases even cause hyponatremia. Central diabetes insipidus is extremely common in patients with severe brain ischemia and is usually indicative of brain death. Treatment of the diabetes insipidus along with any coexistent anterior pituitary hormone deficiencies has been advocated to preserve donor organs in such cases.

Nephrogenic Diabetes Insipidus

Nephrogenic diabetes insipidus is caused by mutations of the vasopressin V2 receptor or the vasopressin-induced water channel, aquaporin-2, or impairments in the signal transduction system linking the V2 receptor and aquaporin-2. Familial nephrogenic diabetes insipidus is a rare disease, most cases of which (>90%) are due to mutations of the V2 receptor. More than 100 different V2 receptor mutations have been described, which can be classified into four general categories on the basis of differences in transport to the cell surface and vasopressin binding or stimulation of adenylate cyclase: (1) the mutant receptor is not inserted in the membrane; (2) the mutant receptor is inserted in the membrane but does not bind or respond to vasopressin; (3) the mutant receptor is inserted in the membrane and binds vasopressin but does not activate adenylate cyclase; or (4) the mutant protein is inserted into the membrane and binds vasopressin but responds subnormally in terms of adenylate cyclase activation. Because the gene for the V2 receptor is located on the Xq28 region of the X chromosome, this is an X-linked recessive disease. Symptoms are noted only in affected males, who present with vomiting, constipation, failure to thrive, fever, and polyuria during the first week of life. Hypernatremia is found with a hypo-osmolar urine. The phenotype is similar in the less than 10% of patients with mutations of the aquaporin-2 water channel, but because the aquaporin-2 gene is located on chromosome 12, mutations cause autosomal recessive disease; consequently, consanguinity and a family history of the disease in men and women is common, and this disorder should be suspected when the proband is a girl.

Nephrogenic diabetes insipidus can also be acquired during treatment with certain drugs, such as demeclocycline (which is used to treat inappropriate secretion of vasopressin), lithium carbonate (used to treat bipolar disorders), and fluoride (previously used in fluorocarbon anesthetics), and from electrolyte abnormalities such as hypokalemia and hypercalcemia. All causes of acquired nephrogenic diabetes insipidus have in common decreased synthesis and function of aquaporin-2 due to impaired vasopressin signaling from V2 receptor binding and activation. Other diseases of the kidney produce polyuria and inability to concentrate the urine secondary to altered renal medullary blood flow or to other disorders that inhibit maintenance of the hyperosmolar concentrating gradient in the inner medulla. Renal manifestations of such disorders (e.g., sickle cell disease, sarcoidosis, pyelonephritis, multiple melanoma, analgesic nephropathy) are discussed in Chapter 123 .

Gestational Diabetes Insipidus

In pregnancy, there is an increased metabolism of vasopressin due to cystine aminopeptidase (oxytocinase or vasopressinase), an enzyme that is normally produced by the placenta to degrade circulating oxytocin and prevent premature uterine contractions. Because of the close structural similarity between vasopressin and oxytocin, this enzyme degrades both peptides. Normally, this can be overcome by increased synthesis of vasopressin. Rarely during pregnancy, women with normal regulation of vasopressin develop diabetes insipidus because of nonphysiologic, markedly elevated levels of vasopressinase. Some of these patients have been noted to have accompanying preeclampsia, acute fatty liver, and coagulopathies, but causal relations between the diabetes insipidus and these abnormalities have not been identified. In general, diabetes insipidus is not found after the pregnancy ends or in subsequent normal pregnancies.

Polyuria can also become manifest in patients who have limited vasopressin reserve because of a decreased ability either to secrete vasopressin (partial central diabetes insipidus) or to respond to vasopressin action (compensated nephrogenic diabetes insipidus). In these cases, vasopressin synthesis cannot keep up with even physiologic circulating levels of vasopressinase during pregnancy. Treatment may be required only during the pregnancy, and the patient often returns to previous baseline function without need for therapy when the pregnancy ends. Less commonly, central diabetes insipidus of another cause first becomes symptomatic during pregnancy and then persists with the usual course of the diabetes insipidus.

Primary Polydipsia

Excessive fluid intake also causes hypotonic polyuria and, by definition, polydipsia. Consequently, this disorder must be differentiated from the various causes of diabetes insipidus. It is apparent that despite normal pituitary and kidney function, nonetheless patients with this disorder share many characteristics of both central diabetes insipidus (i.e., vasopressin secretion is suppressed as a result of the decreased plasma osmolality) and nephrogenic diabetes insipidus (kidney aquaporin-2 expression is decreased as a result of the suppressed plasma vasopressin levels). Many different names have been used to describe patients with excessive fluid intake, but primary polydipsia remains the best descriptor because it does not presume any single etiology for the increased fluid intake.

Primary polydipsia is sometimes due to a severe mental illness such as schizophrenia, mania, or an obsessive-compulsive disorder, in which case it is called psychogenic polydipsia. These patients usually deny true thirst and attribute their polydipsia to bizarre motives, such as a need to cleanse their body of poisons. Series of patients in psychiatric hospitals have shown an incidence as high as 42% of patients with some form of polydipsia, and there is no obvious explanation for the polydipsia in most reported cases. However, primary polydipsia can also be caused by an abnormality in the osmoregulatory control of thirst, in which case it has been termed dipsogenic diabetes insipidus. These patients have no overt psychiatric illness and invariably attribute their polydipsia to a nearly constant thirst. Dipsogenic diabetes insipidus is usually idiopathic, but it can also be secondary to organic structural lesions in the hypothalamus identical to any of the disorders described as causes of central diabetes insipidus, such as neurosarcoidosis of the hypothalamus, tuberculous meningitis, multiple sclerosis, or trauma. Consequently, all polydipsic patients should be evaluated with a magnetic resonance imaging (MRI) scan of the brain before it is concluded that excessive water intake is due to an idiopathic or psychiatric cause. Primary polydipsia can also be produced by drugs that cause a dry mouth or by any peripheral disorder causing marked elevations of renin or angiotensin.

Finally, primary polydipsia is sometimes caused by physicians, nurses, lay practitioners, or health writers who recommend a high fluid intake for valid (e.g., recurrent nephrolithiasis) or unsubstantiated health reasons. These patients lack overt signs of mental illness but also deny thirst and usually attribute their polydipsia to habits acquired from years of adherence to their drinking regimen. Patients with primary polydipsia may drink even greater amounts of fluid (e.g., >20 L/day) during the day than patients with central diabetes insipidus do yet characteristically do not manifest nocturia and often sleep through the night with minimal disruption. The disorder may be exacerbated during times of stress, yet not be bothersome during normal intervals. Laboratory studies in these patients are generally normal, although serum sodium concentration can sometimes be at the low end of the normal range and the level of uric acid is lower than in patients with other forms of diabetes insipidus.


Physiologic Diagnosis

Diabetes insipidus should be considered in all patients presenting with significant polyuria (i.e., urine output of more than 50 mL/kg/day). Although osmotic diuresis secondary to hyperglycemia, intravenous contrast agents, or renal injury is a more common clinical cause of polyuria, the medical history, an isotonic urine osmolality, and routine clinical laboratory tests generally distinguish these disorders from diabetes insipidus. A diagnosis of diabetes insipidus can be made when urine osmolality is inappropriately low in the presence of an elevated plasma osmolality as a result of increased serum sodium concentration. These criteria are sometimes met at the initial examination, especially in cases of acute diabetes insipidus occurring after trauma or after surgery in which fluid replacement has not been adequate. In such patients with hypernatremia and hypotonic urine osmolality with normal renal function, one need only administer a vasopressin agonist to differentiate central diabetes insipidus, in which a renal response with decreased urine volume and increased urine osmolality occurs, from nephrogenic diabetes insipidus, in which a subnormal renal response is seen. Sometimes in the postoperative state, a water diuresis occurs as a result of water retention during the surgical procedure. Vasopressin is normally secreted in response to surgical stress, causing fluid administered intravenously during the procedure to be retained. During recovery, vasopressin levels fall, and a diuresis of the retained fluid occurs. In this case, however, the serum sodium level is almost always normal; but if further fluid is administered to match the urine output, persistent polyuria might be mistaken for diabetes insipidus. In this situation, the physician should decrease the rate of fluid administered and follow the urine output and serum sodium level. If urine output decreases and the serum sodium level remains normal, no treatment is necessary; if serum sodium rises above the normal range and the urine remains hypotonic, diabetes insipidus is likely, and the response to a vasopressin agonist will ascertain the type.

Most outpatients with diabetes insipidus will not be hypernatremic because the polydipsia produced by a normal thirst response is generally sufficient to maintain water homeostasis. Instead, they will present with polyuria, polydipsia, and a normal sodium level. In these patients, it is necessary to perform further testing to increase serum osmolality and then measure the plasma vasopressin level or the urinary response to an administered vasopressin agonist. The best described test is the water deprivation test ( Fig. 243-3 ). The test should be carried out under controlled observation in the hospital or an appropriately equipped outpatient area. The timing of the test depends on the symptoms of the patient. If the patient has marked polyuria during the night, it is best to begin the test during the day because the patient may become dehydrated overnight. However, if the patient has only two or three episodes of nocturia per night, it is best to begin the test in the evening so that the major part of the dehydration takes place when the patient is asleep. In either case, the patient is weighed at the beginning of the test, and all subsequent fluids are withheld. The volume and osmolality (measured directly by freezing point depression) of all excreted urine are measured and the patient is re-weighed after each liter of urine output. When three consecutive urine samples have an osmolality differing by no more than 10% and the patient has lost at least 2% of body weight, a blood sample is obtained for measurement of serum osmolality, sodium, and plasma vasopressin. The patient is then given 2 mg of desmopressin intravenously or intramuscularly and observed for an additional 2 hours.

FIGURE 243-3  Responses to the water deprivation test to differentiate various types of diabetes insipidus and primary polydipsia (as described by Miller and associates, Ann Intern Med 1970;73:721). The response to dehydration shows a plateau, and the subsequent change in urine osmolality in response to administered vasopressin is illustrated. See discussion in the text. DI = diabetes insipidus.

Adults with normal vasopressin secretion will be able to concentrate their urine to above 800 mOsm/kg H2O and will have less than a 10% increase in urine osmolality in response to administered desmopressin. Patients with complete central diabetes insipidus have minimal concentration of the urine with dehydration and a marked increase in urine osmolality (usually >50%) in response to administered desmopressin. Patients with nephrogenic diabetes insipidus usually have no increase in urine concentration in response to administered desmopressin, although in some cases of acquired nephrogenic diabetes insipidus, some increased urinary concentration (generally <10%) can occur. Nephrogenic diabetes insipidus is best distinguished from central diabetes insipidus by measurement of vasopressin in the plasma; plasma vasopressin levels are elevated normally in cases of nephrogenic diabetes insipidus, especially after dehydration.

In patients with partial central diabetes insipidus and patients with primary polydipsia, the urine is often somewhat concentrated in response to dehydration, but not to the maximum of a normal person because the chronically reduced level of vasopressin downregulates the synthesis of aquaporin-2 water channels, and the large urine volume, regardless of cause, washes out the medullary osmotic gradient that determines the maximum urine concentration. When desmopressin is administered, patients with partial central diabetes insipidus have a further increase (usually >10% but <50%) in urine osmolality, whereas most patients with primary polydipsia have no further increase (i.e., <10%). However, the reliability of distinguishing between these last two disorders by the water deprivation test is suboptimal. Some patients with primary polydipsia may not become sufficiently dehydrated with the test to secrete maximum vasopressin and hence have an increase in urine osmolality in response to administered desmopressin. Alternatively, some patients with partial central diabetes insipidus can become sufficiently dehydrated that their maximal concentration of urine is reached during the test and no further concentration is seen with administered desmopressin. Plasma vasopressin levels at the end of dehydration offer a better means to discriminate between these two disorders, but only at high serum sodium concentrations (i.e., >145 mmol/L). Consequently, some investigators advise a limited infusion of hypertonic (3%) NaCl solution to achieve these levels if they are not achieved by the water deprivation itself. It is important to maintain adequate follow-up of patients with suspected partial central diabetes insipidus to ensure that during treatment with vasopressin, a good therapeutic response is obtained. In some difficult cases, the response to treatment with a vasopressin agonist can be an adjunct to the diagnosis. If a decrease in polyuria and thirst with maintenance of normal serum sodium concentration occurs, a diagnosis of partial central diabetes insipidus is confirmed; however, if polydipsia persists and hyponatremia develops, a diagnosis of primary polydipsia is confirmed.

Etiologic Diagnosis

If the water deprivation test confirms that inadequate vasopressin secretion or function is responsible for the polyuria, the underlying cause must be determined. MRI of the hypothalamic-pituitary area is the most important diagnostic tool in these cases. The three areas of interest are the immediate suprasellar region of the hypothalamus, the pituitary stalk, and the posterior pituitary within the sella turcica. As noted in the discussion of anatomy, vasopressin is synthesized in the paired paraventricular nuclei high on the walls of the third ventricle and the paired supraoptic nuclei lateral to and above the optic chiasm and then transported to the posterior pituitary for storage and release. Transection or damage of these axons at the level of the posterior pituitary causes an accumulation of neurosecretory material in the axon proximal to the site of the injury as well as outgrowth of axons in the median eminence, where vasopressin can be secreted into the pituitary portal capillaries. Therefore, most slow-growing tumors that are confined to the sella do not cause diabetes insipidus. To cause central diabetes insipidus, tumors in the hypothalamic area immediately above the sella must be either sufficiently large to destroy 80 to 90% of the vasopressin cells or located where the paths of the four nuclear groups converge at the origin of the pituitary stalk just above the diaphragma sellae. Primary tumors, especially craniopharyngioma, suprasellar germinoma, metastatic tumors, and infiltrative diseases, can also cause diabetes insipidus by infiltration of the pituitary stalk, which is then thickened on MRI scanning. In addition, on T1-weighted MRI images, the vasopressin and oxytocin stored in neurosecretory granules in the posterior pituitary are visualized as a “bright spot” in the sella turcica. Most but not all normal subjects have this bright spot; and in most but not all patients with central diabetes insipidus, the bright spot is lost. Thickening of the stalk and absence of the bright spot are therefore especially suggestive of a hypothalamic disease process.

Tumors that cause central diabetes insipidus are most often benign primary intracranial tumors, such as craniopharyngioma, ependymoma (suprasellar germinoma), and pinealoma, that arise in the third ventricle. Primary tumors of the anterior pituitary cause diabetes insipidus only when substantial suprasellar extension is present. However, rapidly growing intrasellar lesions, such as metastases from carcinomas of the lung, breast, and melanoma or hemorrhage into pituitary adenomas, can cause diabetes insipidus because of insufficient time to allow vasopressin axons to adapt by releasing hormone into the vasculature of the median eminence. Metastasis to the hypothalamus can also lodge in the portal capillaries of the median eminence, thereby destroying the supraopticohypophysial tract and producing diabetes insipidus. Granulomatous diseases, such as Langerhans cell histiocytosis, sarcoidosis, and tuberculosis, and leukemic infiltrates and lymphomas of the hypothalamus can cause diabetes insipidus by destroying vasopressin cells in the hypothalamus. In such patients with diseases having peripheral manifestations, the diagnosis is usually suspected on the basis of general medical findings.

Lymphocytic infundibuloneurohypophysitis is an autoimmune disease similar to lymphocytic hypophysitis of the anterior pituitary in which lymphocytes infiltrate the neurohypophysis to produce diabetes insipidus. The hallmarks of this process are a thickened pituitary stalk and an absence of the pituitary bright spot in a patient with abrupt onset of polyuria and polydipsia, particularly in a postpartum female. The diagnosis was originally demonstrated by pituitary biopsy but now is generally made by regression of the thickened stalk with continued follow-up. When no cause is found, the diagnosis of exclusion is idiopathic diabetes insipidus, but most such cases probably also are caused by an autoimmune disease, and other autoimmune diseases are often recognized in affected patients. When central nervous system disease is suspected but not diagnosed by MRI or general physical examination, cerebrospinal fluid obtained by lumbar puncture may be helpful in identifying tumor cells or markers of tumors or inflammatory processes (e.g., elevated angiotensin-converting enzyme levels in neurosarcoidosis).


Because excess diuresis of water is the primary manifestation of diabetes insipidus, water replacement in adequate quantities avoids the metabolic complications of this disease. However, oral or intravenous administration of the volume of fluid required to replace urinary losses in diabetes insipidus is difficult and inconvenient. The goal of therapy is to reduce the amount of polyuria and polydipsia to a tolerable level while avoiding overtreatment, which can produce water retention and hyponatremia. The best therapeutic agent for treatment of hypothalamic diabetes insipidus is the vasopressin agonist desmopressin. Desmopressin is different from vasopressin in that the amino group of the N-terminal cystine residue has been removed to prolong the duration of action, and d-arginine has been substituted for l-arginine in position 8 to decrease vasopressor effects. At therapeutic dosages, this agent acts on V2 or antidiuretic receptors with minimal activity at V1a or pressor receptors. Desmopressin is available in tablets of 0.1 mg or 0.2 mg for oral administration and in either a spray bottle that delivers a fixed dose of 10 μg in 100 μL or in a bottle with a rhinal catheter that can deliver 50 to 200 μL (5 to 20 μg for intranasal administration). When therapy is initiated, it is generally best to begin with a low dose (e.g., half of a 0.1-mg tablet, 5 μg by the rhinal tube, or a single 100 μL spray of 10 μg) at bedtime to allow the patient to sleep through the night, then to determine the duration of action by quantifying the polyuria the next day. The duration of action of a single dose varies between patients from 6 to 24 hours; but in most patients, a dosage can be determined that provides a good therapeutic response on an every-12-hour schedule for the nasal spray and an 8- or 12-hour schedule for the tablets. If patients are never polyuric on a fixed schedule, it may be advisable to delay administration of a dose once a week to allow diuresis of any accumulated water. Desmopressin is also available for parenteral use in 1-mL vials of 4 μg/mL. Parenteral administration is especially useful postoperatively or when a patient is unable to take the nasal preparation. In hospitalized patients, some physicians add vasopressin directly to a crystalloid solution to infuse doses in the range of 0.25 to 2.7 mIU/kg/hr to cause modest but persistent urinary concentration as a treatment of diabetes insipidus. With any parenteral administration, serum sodium levels should be monitored frequently to prevent the development of hyponatremia.

Some orally administered pharmacologic agents are also useful in treating diabetes insipidus. Chlorpropamide in doses of 100 to 500 mg daily enhances the effect of vasopressin at the renal tubule and is especially useful in patients with partial central diabetes insipidus. An antidiuretic effect is noted in 1 to 2 days, but maximum antidiuresis may not be achieved until after several days of administration. Thiazide diuretics cause sodium depletion and volume contraction and decrease urine volume by increasing proximal tubular reabsorption of glomerular filtrate. Prostaglandin synthase inhibitors (e.g., indomethacin) block the action of prostaglandin E to inhibit the action of vasopressin on the kidney. Although use of a prostaglandin synthase inhibitor is not a primary treatment of central diabetes insipidus, it can alter the antidiuretic response of other agents. Chlorothiazide, amiloride, and prostaglandin synthase inhibitors are useful to reduce polyuria in nephrogenic diabetes insipidus. However, none of these agents has approval of the Food and Drug Administration for treatment of diabetes insipidus; therefore, the prescribing physician should be aware of potential toxicities and side effects. In cases of drug-induced nephrogenic diabetes insipidus, the most direct therapy is discontinuation of the offending agent, if possible. Symptomatic nephrogenic diabetes insipidus is usually treated with a thiazide diuretic, which is enhanced by the coadministration of the potassium-sparing diuretic amiloride. Amiloride can be especially beneficial in cases of nephrogenic diabetes insipidus induced by lithium because the drug decreases the entrance of lithium into cells in the distal tubule. When diuretics are used to treat nephrogenic diabetes insipidus, special attention should be paid to the possibility that the induced dehydration may increase the concentration of other drugs.

Some situations require special attention during therapy. Rarely, if patients with diabetes insipidus are unable to drink or are given a hypertonic solution, severe hypernatremia can develop acutely. Osmotic equilibrium with the intracellular water of neurons and glia produces shrinking of the brain. The brain is in a closed vault (i.e., the skull), and when the brain shrinks, traction on the vasculature of the central nervous system can cause rupture of blood vessels and subarachnoid or intracerebral hemorrhage. If the hypernatremia persists for a longer time, the neurons accommodate by production of organic osmolytes, previously called idiogenic osmoles, which limit the amount of brain shrinkage. Once this adaptation has occurred, too rapid lowering of osmolality in the extracellular fluid will produce a shift of water into the brain and cause cerebral edema. In this situation, desmopressin can be administered to produce constant antidiuresis, and the amount of water given can be regulated to decrease osmolality by no more than approximately 1 mEq every 2 hours. Postoperatively or after head trauma, diabetes insipidus can be transient (see prognosis in the next section), and the need for long-term maintenance therapy cannot be immediately established. During pregnancy, vasopressinase increases the metabolism of vasopressin but not of desmopressin, so desmopressin is the drug of choice. The vasopressinase activity subsides by a few weeks after delivery, and patients with the onset of partial diabetes insipidus during pregnancy may become asymptomatic after delivery. An additional advantage of desmopressin is that it has little action on the oxytocin receptors of the uterus. During pregnancy, normal plasma osmolality decreases by approximately 10 mOsm/kg H2O because of changes in serum sodium, so pregnant patients with diabetes insipidus require sufficient desmopressin only to maintain their serum sodium at this lower level.


The prognosis of properly treated diabetes insipidus is excellent. Historical complications of bladder hypertrophy and hydroureter secondary to voluntarily decreasing urine frequency are largely unseen with modern therapy. If nephrogenic diabetes insipidus is diagnosed and treated early, intracranial calcification and mental retardation do not occur. When the diabetes insipidus is secondary to a recognized disease process, it is that disease which generally determines the ultimate prognosis. In some specific clinical situations, the course is different and characteristic. The development of diabetes insipidus after surgical or traumatic injury to the neurohypophysis can follow any of several different well-defined patterns ( Fig. 243-4 ). In some patients, polyuria develops 1 to 4 days after injury and resolves spontaneously. Less often, the diabetes insipidus is permanent and continues indefinitely. Most interestingly, one can see a “triphasic” response that has been well described after pituitary stalk transection. The initial diabetes insipidus (first phase) is due to axon shock and lack of function of the damaged neurons. This phase lasts several hours to several days and then is followed by an antidiuretic phase (second phase) that is due to the uncontrolled release of vasopressin from the disconnected and degenerating posterior pituitary or from the remaining severed neurons. Overly aggressive administration of fluids during this second phase does not suppress the uncontrolled vasopressin release from the damaged neurohypophysis and can lead to hyponatremia. The antidiuresis can last 2 to 14 days, after which diabetes insipidus recurs after depletion of the vasopressin from the degenerating posterior pituitary gland (third phase). Transient hyponatremia without preceding or subsequent diabetes insipidus has been reported recently after transsphenoidal surgery for pituitary microadenomas. This generally occurs 5 to 10 days postoperatively, and the incidence may be as high as 30% when such patients are carefully observed, although the majority of cases are mild and self-limited. This is due to inappropriate vasopressin secretion by the same mechanism as the triphasic response, except that in these cases, only the second phase occurs (“isolated second phase”) because the neural lobe or pituitary stalk damage does not impair vasopressin secretion sufficiently to produce clinical manifestations of diabetes insipidus.

FIGURE 243-4  Diagrammatic summary of the major patterns of postoperative and post-traumatic diabetes insipidus. The abscissa represents time after the initial injury (arrow); the ordinate represents urinary volume relative to a hypothetical “normal” urine output of 2 to 3 L/24 hr (solid line). See discussion in text. During the triphasic response, uncontrolled release of vasopressin from the disconnected or damaged posterior pituitary gland causes an antidiuresis that can lead to water retention and a dilutional hyponatremia. Diabetes insipidus returns as the third phase after the stored hormone in the posterior pituitary has been depleted. DI = diabetes insipidus.  (From Verbalis JG, Robinson AG, Moses AM: Postoperative and post-traumatic diabetes insipidus. In Czernichow AP, Robinson A [eds]: Diabetes Insipidus in Man: Frontiers of Hormone Research. Basel, S Karger, 1985, p 247.)

Once a deficiency of vasopressin secretion has been present for more than a few days or weeks, it rarely improves even if the underlying cause of the neurohypophysial destruction is eliminated. The major exception to this is in patients with postoperative diabetes insipidus, when spontaneous resolution is the rule. Although recovery from diabetes insipidus that persists more than several weeks postoperatively is less common and is distinctly uncommon after 1 year of continued diabetes insipidus, well-documented cases of long-term recovery have been reported as long as 10 years after the initiating event. The reason for amelioration and resolution is apparent from pathologic and histologic examination of neurohypophysial tissue after pituitary stalk section in both animals and humans. Neurohypophysial neurons that have intact perikarya are able to regenerate axons and form new nerve terminal endings capable of releasing vasopressin into nearby capillaries. Potential return of function is another reason for occasionally withholding therapy during long-term treatment. Diabetes insipidus should not be considered idiopathic until at least 4 years of follow-up. During this interval, annual computed tomography or MRI scans are indicated to test for the appearance of a tumor or infiltrative process that may not have been detected at the initial examination.

Email to Colleague Print Version

Copyright © 2007 Elsevier Inc. All rights reserved. –

Leave a Reply

Time limit is exhausted. Please reload the CAPTCHA.


apply_now Pepperstone Group Limited