Diabetes insipidus (DI) is defined as the passage of large volumes (>3 L/24 hr) of dilute urine (< 300 mOsm/kg). It has the following 2 major forms:
Two other forms are gestational DI and primary polydipsia (dipsogenic DI); both are caused by deficiencies in AVP, but the deficiencies do not result from a defect in the neurohypophysis or kidneys.
The predominant manifestations of DI are as follows:
The most common form is central DI after trauma or surgery to the region of the pituitary and hypothalamus, which may exhibit 1 of the following 3 patterns:
In infants with DI, the most apparent signs may be the following:
In children, the following manifestations typically predominate:
If the condition that caused DI also damaged the anterior pituitary or hypothalamic centers that produce releasing factors, patients may present with the following:
Physical findings vary with the severity and chronicity of DI; they may be entirely normal or may include the following:
See Clinical Presentation for more detail.
If the clinical presentation suggests DI, laboratory tests must be performed to confirm the diagnosis, as follows:
Additional studies that may be indicated include the following:
See Workup for more detail.
Most patients with DI can drink enough fluid to replace their urine losses. When oral intake is inadequate and hypernatremia is present, provide fluid replacement as follows:
Pharmacologic therapeutic options include the following:
See Treatment and Medication for more detail.
Diabetes insipidus (DI) is defined as the passage of large volumes (>3 L/24 h) of dilute urine (< 300m Osm/kg). DI has 2 major forms: central and nephrogenic.
Central DI is characterized by decreased secretion of antidiuretic hormone (ADH)—also known as arginine vasopressin (AVP)—which gives rise to polyuria and polydipsia by diminishing the person’s ability to concentrate urine. Other terms for central DI are neurogenic, pituitary, and neurohypophyseal DI. (See Etiology, Presentation, and Workup.)
Nephrogenic DI is characterized by a decrease in the ability to concentrate urine because of resistance to ADH action in the kidney. Nephrogenic DI can be observed in chronic renal insufficiency, lithium toxicity, hypercalcemia, hypokalemia, glucosuria, and tubulointerstitial disease. (See Etiology, Presentation, and Workup.)
Two other forms of DI are gestational DI and primary polydipsia. Both are caused by deficiencies in AVP, but the deficiencies do not result from a defect in the neurohypophysis or kidneys. Gestational DI results from degradation of vasopressin by a placental vasopressinase. Primary polydipsia (dipsogenic DI) results from a primary defect in osmoregulation of thirst. The exact location of the lesion is not known, but structural lesions may exist, as dipsogenic DI has been reported in tuberculous meningitis, multiple sclerosis, and neurosarcoidosis.
Rarely, DI may be hereditary. Hereditary nephrogenic DI manifests in early infancy, often before the age of 1 week. Hereditary central DI typically manifests in childhood. For more information on DI in children, see Pediatric Diabetes Insipidus.
Pharmacologic treatment of DI generally involves the use of desmopressin (1-deamino-8-D-arginine vasopressin [DDAVP]), nonhormonal drugs, or both. Patients must be instructed in simple principles of water balance to avoid dehydration and water intoxication (if they are not carefully monitoring water intake). (See Treatment.)
The normal range of plasma osmolality is between 275 and 295 mOsm/kg. The ability of the kidneys to modify the concentration of urinary solutes ranges between 50–1200 mOsm/kg. Healthy adults on a normal diet excrete 800–1200 mOsm of solute daily. Thus, to excrete 1000 mOsm of solute, the obligate urinary water excretion would be 1000 mOsm per 1200 mOsm/kg water, which translates into 0.8 kg (0.8 L) of water per day. This urine is maximally concentrated and appears dark yellow or orange in color. If this requirement for obligate water excretion is not met, solutes accumulate, leading to uremia.
Conversely the maximum volume of urine (secondary to limits imposed by renal dilutional capacity) is 20 L of water per day (1000 mOsm per 50 mOsm/kg water). This maximally dilute urine is colorless.
The maintenance of water balance in healthy humans is principally accomplished through 3 robust, interrelated determinants: thirst, AVP, and the kidneys. In addition, recognition of a fourth factor, apelin, has emerged in recent years. Apelin is a bioactive peptide that is widely distributed throughout the body. In the brain, it is expressed in supraoptic and paraventricular nuclei, as well as in other sites, and has specific receptors located on vasopressinergic neurons. Apelin acts as a potent diuretic neuropeptide that inhibits ADH release.
AVP is the primary determinant of free water excretion in the body. Its main target is the kidney, where it acts by altering the water permeability of the cortical and medullary collecting tubules. Water is reabsorbed by osmotic equilibration with the hypertonic interstitium and returned to the systemic circulation. The actions of AVP are mediated through at least 2 receptors, as follows[4, 5] :
The vasoconstrictor effect of AVP is negligible in humans. No clinically significant defects in blood pressure regulation or cortisol secretion are apparent in patients with DI.
Diminished or absent ADH production can be the result of a defect in 1 or more sites in the neurohypophysis. These include the hypothalamic osmoreceptors, the supraoptic or paraventricular nuclei, and the supraopticohypophyseal tract.
Ordinarily, a decrease in the extracellular fluid (ECF) volume elicits the following simultaneous responses:
Volume depletion activates baroreceptor mechanisms that exert similar effects on aldosterone, thirst, and AVP, whereas osmoreceptor-mediated mechanisms impact thirst and AVP secretion only.
Osmoreceptors for thirst are solute specific, responding preferentially to increased sodium levels in the ECF. Thus, elevated glucose levels in diabetes mellitus do not induce thirst; rather, the increased thirst in uncontrolled diabetes mellitus is secondary to volume depletion from osmotic diuresis.
DI is usually an acquired disorder, with central DI having different causes than does nephrogenic DI. In rare cases, central or nephrogenic DI may be an inherited disorder.
Central DI has many possible causes. According to the literature, the principal causes of central DI and their oft-cited approximate frequencies are as follows:
Idiopathic central DI presumably develops when cells in the hypothalamus are damaged or destroyed. Identification of antibodies against AVP-secreting cells and advances in imaging techniques have made idiopathic cases less common than they previously were.
Increasingly, the role of inflammation and autoimmunity in DI is being recognized. Cases of lymphocytic hypophysitis were possibly classified as idiopathic prior to improved imaging studies. This disorder is characterized by lymphocytic infiltration of the stalk and posterior pituitary. Magnetic resonance imaging (MRI) may show abnormalities in these structures.
Antibodies directed against vasopressin cells have been found in patients with idiopathic central DI; however, these antibodies have also been found in patients with Langerhans cell histiocytosis (LCH) or germinomas, which indicates that this finding can not be considered a reliable marker of autoimmune etiology in central DI. Indeed, reliance on AVP antibodies may delay the diagnosis of LCH or germinoma.
Given the possible diagnostic confusion, close clinical and MRI follow-up is necessary. Serial contrast-enhanced brain MRIs (every 3-6 months for the first 2 years) in patients with central DI who have pituitary stalk thickening may shorten the time to diagnosis of germinoma by as much as 1 year.
The role of human chorionic gonadotropin (hCG) in the early diagnosis of germinoma is not fully established. A negative result for hCG in the cerebrospinal fluid (CSF) does not exclude germinoma.
Primary intracranial tumors causing DI include craniopharyngiomas, germinomas, and pineal tumors, among others. The appearance of other hypothalamic manifestations may be delayed for as long as 10 years in these cases.
Craniopharyngioma is a benign tumor that arises from squamous cell nests in the primitive Rathke pouch. It is the most frequent pediatric intracranial neoplasm, accounting for nearly 54% of cases. Central DI insipidus and multiple pituitary hormone deficiencies are common manifestations in childhood craniopharyngiomas. Surgery is the preferred treatment.
The frequency with which DI develops after neurosurgery varies with the surgery’s scope. Approximately 10-20% of patients experience DI after transsphenoidal removal of an adenoma, compared with 60-80% of those who have undergone excision of large tumors.
Not all cases of postoperative DI are permanent. In a German study of metabolic disturbances after transsphenoidal pituitary adenoma surgery, only 8.7% of DI cases persisted for more than 3 months.
Postoperative polyuria does not necessarily indicate DI. The most common causes of postoperative polyuria are excretion of excess fluid administered during surgery and an osmotic diuresis resulting from treatment for cerebral edema.
DI in head trauma
Central DI can be an acute or chronic complication of head injury or subarachnoid hemorrhage.[6, 9] Risk factors for acute DI include penetrating trauma and severe head trauma. Other forms of pituitary dysfunction (eg, adrenocorticotropic hormone deficiency) may accompany posttraumatic DI. The dysfunction may be transient or, less commonly, may develop gradually.
Hereditary central DI
Approximately 10% of central DI cases are familial (although some experts suggest that familial DI may be underdiagnosed). Most of these cases show autosomal dominant inheritance and result from a defect in the AVP-NP2 gene on chromosome 20p13. The defect results in the production of mutant prohormone that is toxic to the neuron and eventually destroys it.[12, 13, 14]
There are also autosomal recessive forms of DI, which result from defects in the AVP-NP2 (AVP neurophysin) gene, as well as in the WFS1 gene. The latter gene encodes for wolframin, a tetrameric protein that may serve as a novel endoplasmic reticular calcium channel in pancreatic beta cells and neurons. Mutations in WFS1 lead to Wolfram syndrome, which is also known by the acronym DIDMOAD (Diabetes Insipidus, Diabetes Mellitus, Optic Atrophy, Deafness).
Another recessive form of central DI results from the production of biologically inactive AVP. In addition, an X-linked form of neurohypophyseal DI exists. A specific genetic defect has not been identified.
Genetic testing to determine the specific etiology can obviate the search for another cause. Finding a genetic anomaly will also answer recurrence risk questions for the family, and may prove to be helpful with treatment options.
Other causes of central DI include the following:
In adults, nephrogenic DI most often develops as a result of lithium toxicity or hypercalcemia. Impairment of urinary concentration occurs in up to 20% of patients taking lithium, as a result of dysregulation of the aquaporin system in principal cells of the collecting duct.[16, 17] Prolonged elevation of serum calcium concentrations above 11 mg/dL (2.75 mmol/dL can also impair urinary concentrating ability.
Other causes of acquired nephrogenic DI include the following:
In addition to lithium, other drugs that can reduce urinary concentrating ability include the following:
Hereditary nephrogenic DI
Hereditary nephrogenic DI is relatively rare. The most common inherited form results from mutations in the AVP receptor 2 gene (AVPR2) on chromosome Xq28. Defects in the AVP receptor cause resistance to the antidiuretic effect of vasopressin. Because hereditary nephrogenic DI is an X-linked disorder, most cases occur in males; however, cases occasionally arise in females as a result of skewed X inactivation.
Approximately 1% of familial nephrogenic DI cases result from mutations in AQP2 (aquaporin 2), a gene on chromosome 12q13 that gives rise to a water channel that is expressed exclusively in the kidney’s collecting ducts. Autosomal recessive and autosomal dominant forms of nephrogenic DI from AQP2 mutations have been reported.
DI is uncommon in the United States, with a prevalence of 3 cases per 100,000 population. No significant sex-related differences in central or nephrogenic DI exist, with male and female prevalence being equal. Similarly, no significant differences in prevalence among ethnic groups have been found.
With both central and nephrogenic DI, inherited causes account for approximately 1-2% of all cases. An incidence of about 1 in 20 million births for nephrogenic DI caused by AQP2 mutations has been cited.
The prognosis for patients with DI is generally excellent, depending on the underlying illness. In nephrogenic DI caused by medication (eg, lithium), stopping the medication may help to restore normal renal function; after many years of lithium use, however, permanent nephrogenic DI may occur.
DI-related mortality is rare in adults as long as water is available. Severe dehydration, hypernatremia, fever, cardiovascular collapse, and death can ensue in children and elderly people, as well as in persons with complicating illnesses.
Polyuria, polydipsia, and nocturia are the predominant manifestations of diabetes insipidus (DI). The daily urine volume is relatively constant for each patient but is highly variable between patients, ranging from 3-20 L.
A patient’s history may indicate whether he/she has central or nephrogenic DI. The most common form of DI is central DI following trauma or surgery to the region of the pituitary and hypothalamus. It may exhibit 1 of 3 patterns: transient, permanent, or triphasic. The triphasic pattern is observed more often clinically. Whether improvements in surgical techniques and approaches have altered the frequency of the triphasic pattern is not well studied.
The first phase of the triphasic pattern is a polyuric one that lasts 4-5 days, caused by inhibition of antidiuretic hormone (ADH). An immediate increase in urine volume and a concomitant fall in urinary osmolality occur. The second phase is an antidiuretic one that lasts 5-6 days, resulting from the release of stored hormone; urinary osmolality rises. The third phase can be permanent DI, when stores of ADH are exhausted and the cells that produce ADH are absent or unable to produce more.
In infants with DI, the most apparent signs may be the following:
In children, the following manifestations typically predominate:
Pregnancy is associated with an increased risk of DI, but this form remits after delivery. In addition, pregnancy may unmask subclinical or mild central DI. (See Pituitary Disease and Pregnancy.)
In mentally intact patients, thirst is the most sensitive indicator of water balance. Many patients have a predilection for drinking cold liquids, often water. Neurologic symptoms vary with the patient’s access to water; patients with free access may have no neurologic symptoms at all. However, the clinical presentation depends on the cause and severity, as well as on the patient’s associated medical condition(s).
Rarely, patients present with adipsic DI. This usually implies a lesion affecting ADH production and the hypothalamic osmoreceptor. Patients with adipsic DI are at much higher risk of severe dehydration and may require prescriptive fluid replacement regimes.
If the condition that caused DI also damaged the anterior pituitary or hypothalamic centers that produce releasing factors, patients may present with other symptoms and signs of anterior pituitary dysfunction. These would include excessive fatigue, diminished libido or erectile dysfunction, headache, dry skin, and hair loss.
The physical examination findings vary with the severity and chronicity of DI. The examination findings may be entirely normal. Hydronephrosis, with pelvic fullness, flank pain or tenderness, or pain radiating to the testicle or genital area, may be present. Bladder enlargement occurs in some patients. Unless the thirst mechanism is impaired or access to fluid is restricted, dehydration is not seen. Aside from an enlarged bladder, no specific signs of DI exist.
In a patient whose clinical presentation suggests diabetes insipidus (DI), laboratory tests must be performed to confirm the diagnosis. A 24-hour urine collection for determination of urine volume is required. In addition, the clinician should measure the following:
Perform testing with the patient maximally dehydrated as tolerated—that is, at a time when ADH release is the highest and his/her urine is the most concentrated. Water deprivation testing may be useful in situations in which the diagnosis is uncertain.
A urinary specific gravity of 1.005 or less and a urinary osmolality of less than 200 mOsm/kg are the hallmark of DI. Random plasma osmolality generally is greater than 287 mOsm/kg. Suspect primary polydipsia when large volumes of very dilute urine occur with plasma osmolality in the low-normal range. Polyuria and elevated plasma osmolality despite a relatively high basal level of ADH suggests nephrogenic DI.
Water deprivation followed by the administration of vasopressin may help to differentiate central from nephrogenic DI. The result of this test must be interpreted with caution, however, because patients with partial nephrogenic DI or primary polydipsia may show a response similar to that seen in central DI.
Historically, diagnostic tests in DI can be traced back to the 1930s, when Gilman and Goodman first demonstrated recovery of an antidiuretic substance in the urine of rats following dehydration with hypertonic saline. When animals were provided free access to water, no antidiuretic activity was recovered from urine, and no antidiuretic activity could be recovered from the urine of hypophysectomized rats dehydrated with hypertonic saline.
The water deprivation test (ie, the Miller-Moses test), a semiquantitative test to ensure adequate dehydration and maximal stimulation of ADH for diagnosis, is typically performed in patients with more chronic forms of DI. The extent of deprivation is usually limited by the patient’s thirst or by any significant drop in blood pressure or related clinical manifestation of dehydration.
With mild polyuria, water deprivation can begin the night before the test. With severe polyuria, water restriction is carried out during the day to allow close observation.
All water intake is withheld, and urinary osmolality and body weight are measured hourly. When 2 sequential urinary osmolalities vary by less than 30 mOsm/kg or when the weight decreases by more than 3%, 5 U of aqueous ADH or desmopressin are administered subcutaneously. A final urine specimen is obtained 60 minutes later for osmolality measurement.
In healthy individuals, water deprivation leads to a urinary osmolality that is 2-4 times greater than plasma osmolality. Additionally, in normal, healthy subjects, administration of ADH produces an increase of less than 9% in urinary osmolality. The time required to achieve maximal urinary concentration ranges from 4-18 hours.
In central and nephrogenic DI, urinary osmolality will be less than 300 mOsm/kg after water deprivation. After the administration of ADH, the osmolality will rise to more than 750 mOsm/kg in central DI but will not rise at all in nephrogenic DI. In primary polydipsia, urinary osmolality be above 750 mOsm/kg after water deprivation. A urinary osmolality that is 300-750 mOsm/kg after water deprivation and remains below 750 mOsm/kg after administration of ADH may be seen in partial central DI, partial nephrogenic DI, and primary polydipsia.
Water deprivation test results may be misleading in patient with chronic primary polydipsia, who may experience partial washout of the medullary interstitial gradient and downregulation of ADH release. This would resemble nephrogenic DI, with an inability to concentrate urine. The combination of a plasma ADH assay with water deprivation testing can lead to greater accuracy in differentiating the different forms of DI from each other and from primary polydipsia.
On MRI, T1-weighted images of the healthy posterior pituitary yield a hyperintense signal. This signal is also invariably present in primary polydipsia. In patients with central DI, this signal is absent, except in a few children with the rare, familial form of the disorder. It is also absent in most patients with nephrogenic DI.
Measurement of circulating pituitary hormones other than ADH may be valuable after traumatic brain injury (TBI). In a study of 89 TBI patients, in which the patients’ hormonal function was evaluated at the time of injury and afterward (at 3, 6, and 12 months), Krahulik et al found primary hormonal dysfunction—including major deficits such as DI, growth hormone dysfunction, and hypogonadism—in 19 patients (21% of the cohort).
The major deficits tended to occur in patients with the worst Glasgow Outcome Scale scores. Moreover, the occurrence of empty sella syndrome, as revealed on MRI scans, was highest in patients with deficits. The authors recommended that pituitary hormone testing be routinely performed within 6 months and 1 year after injury in patients who have sustained a moderate to severe TBI.
Most patients with diabetes insipidus (DI) can drink enough fluid to replace their urine losses. When oral intake is inadequate and hypernatremia is present, replace losses with dextrose and water or an intravenous (IV) fluid that is hypo-osmolar with respect to the patient’s serum. Do not administer sterile water without dextrose intravenously, as it can cause hemolysis.
To avoid hyperglycemia, volume overload, and overly rapid correction of hypernatremia, fluid replacement should be provided at a rate no greater than 500-750 mL/h. A good rule of thumb is to reduce serum sodium by 0.5 mmol/L (0.5 mEq/L) every hour. The water deficit may be calculated on the basis of the assumption that body water is approximately 60% of body weight.
In patients with central DI, desmopressin is the drug of choice.[27, 28] A synthetic analogue of antidiuretic hormone (ADH), desmopressin is available in subcutaneous, IV, intranasal, and oral preparations. Generally, it can be administered 2-3 times per day. Patients may require hospitalization to establish fluid needs. Frequent electrolyte monitoring is recommended during the initial phase of treatment.
Alternatives to desmopressin as pharmacologic therapy for DI include synthetic vasopressin and the nonhormonal agents chlorpropamide, carbamazepine, clofibrate (no longer on the US market), thiazides, and nonsteroidal anti-inflammatory drugs (NSAIDs). Because of side effects, carbamazepine is rarely used, being employed only when all other measures prove unsatisfactory. NSAIDs (eg, indomethacin) may be used in nephrogenic DI, but only when no better options exist.
In central DI, the primary problem is a hormone deficiency; therefore, physiologic replacement with desmopressin is usually effective. Use a nonhormonal drug for central DI if response is incomplete or desmopressin is too expensive.
Monitor for fluid retention and hyponatremia during initial therapy. Follow the volume of water intake and the frequency and volume of urination, and inquire about thirst. Monitor serum sodium, 24-hour urinary volumes, and specific gravity. Request posthospitalization follow-up visits with the patient every 6-12 months. Patients with normal thirst mechanisms can usually self-regulate.
No specific dietary considerations exist in chronic DI, but the patient should understand the importance of an adequate and balanced intake of salt and water. A low-protein, low-sodium diet can help to decrease urine output.
Patients with DI must take special precautions, such as when traveling, to be prepared to treat vomiting or diarrhea and to avoid dehydration with exertion or hot weather.
After pituitary surgery, patients should undergo continuous monitoring of fluid intake, urinary output, and specific gravities, along with daily measurements of serum electrolytes. In patients who develop DI, administer parenteral desmopressin every 12-24 hours, along with adequate fluid to match losses.
Follow the specific gravity of the urine, and administer the next dose of desmopressin when the specific gravity has fallen to less than 1.008-1.005 with an increase in urine output. When the patient can tolerate oral intake, thirst can become an adequate guide.
In patients with DI who have undergone surgery of any kind, administer the usual dose of desmopressin and give (hypotonic) IV fluids to match urinary output.
In the setting of neurosurgery or head trauma, the diagnosis of DI may be obvious, and even expected. The intensivists and nurses who manage the patient acutely are in the best position to provide acute care.
In the more subtle forms of DI, and certainly in all chronic forms of DI for which therapy is expected to be indefinite, the clinical endocrinologist is an invaluable aid in establishing the diagnosis and designing therapy.
Medical genetics consultation is appropriate if there is a family history of DI and an inherited form is suspected.
Treatment for diabetes insipidus (DI) varies with the form of the disorder. In central DI and most cases of gestational DI, the primary problem is a deficiency of antidiuretic hormone (ADH)—also known as arginine vasopressin (AVP)—and therefore, physiologic replacement with desmopressin is usually effective. A nonhormonal drug can be used if response is incomplete or desmopressin is too expensive.
Desmopressin has no role in the treatment of nephrogenic DI or primary polydipsia. Nonhormonal drugs usually are more effective in treating nephrogenic DI.
Clinical Context: Desmopressin is a synthetic analogue of ADH with potent antidiuretic activity but no vasopressor activity.
Clinical Context: Vasopressin has vasopressor and ADH activity. It increases water resorption at collecting ducts (ADH effect). At high doses, it also promotes smooth muscle contraction throughout the vascular bed of renal tubular epithelium (vasopressor effects). However, vasoconstriction is also increased in splanchnic, portal, coronary, cerebral, peripheral, pulmonary, and intrahepatic vessels.
In patients with central DI, replacement of endogenous ADH with exogenous hormones prevents complications of DI and reduces morbidity.
Clinical Context: Chlorpropamide promotes renal response to ADH.
The hypoglycemic agent chlorpropamide helps to relieve diuresis in patients with DI.
Clinical Context: Carbamazepine possibly ameliorates DI by promoting the release of ADH. It is not useful in nephrogenic DI and generally is not a first-line drug.
Certain antiepileptic drugs, such as carbamazepine, have proven helpful in DI.
Clinical Context: Hydrochlorothiazide is a thiazide diuretic that decreases urinary volume in the absence of ADH. It may induce mild volume depletion and cause proximal salt and water retention, thereby reducing flow to the ADH-sensitive distal nephron. Its effects are additive to those of other agents.
Diuretics may reduce flow to the ADH-sensitive distal nephron.
Clinical Context: Inhibition of prostaglandin synthesis reduces the delivery of solute to distal tubules, reducing urine volume and increasing urine osmolality. Indomethacin is usually used in nephrogenic DI.
Clinical Context: Inhibition of prostaglandin synthesis reduces the delivery of solute to distal tubules, reducing urine volume and increasing urine osmolality. Ibuprofen is usually used in nephrogenic DI.
Clinical Context: Inhibition of prostaglandin synthesis reduces the delivery of solute to distal tubules, reducing urine volume and increasing urine osmolality.
The mechanism of action of NSAIDs is not known, but these agents may act by inhibiting prostaglandin synthesis.
Clinical Context: Amiloride is a potassium-sparing diuretic. Thus, the risk of hypokalemia is decreased when amiloride is used in combination with hydrochlorothiazide. In addition, the 2 agents are synergistic with respect to antidiuresis.
Diuretics may reduce flow to the ADH-sensitive distal nephron.