Diabetes Insipidus

Are You Sure the Patient Has Diabetes Insipidus?

Diabetes insipidus (DI) is defined as an uncontrolled solute-free water diuresis (which is also called “aquaresis”) due to an inability to maximally concentrate the urine. The clinical hallmark of DI is the excretion of a large volume of hypotonic, insipid (tasteless) urine, usually manifested by polyuria (increased urination) and polydipsia (increased thirst). Consequently, patients diagnosed with DI must have hypotonic polyuria: the 24-hour urine volume exceeds 50 mL/kg body weight and the urine is inappropriately dilute (i.e., specific gravity [S.G.] <1.010 and urine osmolality [Uosm] <300 mOsm/kg H2O), in the absence of a solute diuresis such as glucosuria. The essential criteria for diagnosing DI are summarized in Table I.

Table I.n
Essential Criteria for Diagnosing DI.

DI can result from impaired secretion of the antidiuretic hormone arginine vasopressin (AVP), impairments of AVP biological activity in the kidney, or increased destruction of circulating AVP:

– Insufficient secretion of AVP from the posterior pituitary gland in response to increased osmolality (central DI, or CDI). Because submaximal concentration of the dilute glomerular filtrate takes place in the renal collecting duct, a large volume of urine is excreted. This produces an increase in serum osmolality with stimulation of thirst and secondary polydipsia. Levels of AVP in plasma are unmeasurable or inappropriately low for plasma osmolality.

– Inability of an otherwise normal kidney to respond to AVP in response to increased osmolality (nephrogenic DI, or NDI). As in CDI, 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 CDI, however, measured levels of AVP in plasma are high or appropriate for plasma osmolality.

– Increased destruction of circulating AVP by elevated levels or activity of placental cystine aminopeptidase (oxytocinase or vasopressinase) during pregnancy (gestational DI, or GDI). The rapid destruction of AVP produces polyuria and secondary stimulation of thirst with polydipsia as in the other types of DI. Because of the circulating vasopressinase, plasma AVP levels usually cannot be measured.

What Else Could the Patient Have?

There are three general types of disorders that must be considered and excluded in the differential diagnosis of patients with polyuria and polydipsia before a diagnosis of DI should be further explored:

1. Other causes of frequent urination. Urinary frequency and urinary urgency are relatively common problems in adults that may be caused by a variety of disorders causing lower urinary tract symptoms (LUTS) such as overactive bladder (OAB) and benign prostatic obstruction (BPO). The key to distinguishing these disorders from DI is measurement of 24-hour urine output. Complaints of frequent urination with a 24-hour urine volume below 50 ml/kg H2O strongly suggests urinary frequency rather than polyuria, and referral to a urologist for further evaluation of the cause of urinary frequency is appropriate rather than further work-up for DI.

2. Other causes of hyperosmolality. Hyperosmolality generally indicates a state of dehydration, with the exception of sodium intoxication (e.g., sea water drowning, excessive infusion of hypertonic saline, increased ingestion of concentrated NaCl substances such as soy sauce). The normal physiological response to hyperosmolality is secretion of AVP to conserve water via renal concentration. Hence, most such patients will have a hypertonic urine (typically >800 mOsm/kg H2O) rather than the hypotonic urine necessary for a diagnosis of DI.

However, several situations can lead to a submaximal urine concentration despite dehydration-induced hyperosmolality: a) a solute diuresis as seen with glucosuria from poorly controlled diabetes mellitus (see below) can lead to excretion of an isotonic urine; b) elderly individuals often have decreased ability to concentrate their urine; and c) partial DI may lead to an increase in AVP secretion at very high plasma osmolalities that is sufficient to concentrate urine to levels above isotonicity (see Figure 1). Consequently, hyperosmolar patients with urine osmolalities <800 mOsm/kg H2O should be followed carefully as they are rehydrated to see if they develop a hypotonic urine (i.e., Uosm < 300 mOsm/kg H2O) before their plasma osmolality returns to normal ranges, in which case further evaluation for partial forms of DI should be entertained.

Figure 1.
Relation between plasma AVP levels, urine osmolality, and plasma osmolality in subjects with normal posterior pituitary function (100%) compared with patients with graded reductions in AVP-secreting neurons (to 50%, 25%, and 10% of normal).

3. Other causes of polyuria and polydipsia.

a. Solute diuresis. Excretion of large amounts of osmotically-active solutes in the urine will cause increased renal water excretion in order to clear the excreted solutes. As the solute excretion increases, urine concentration will approach isotonicity. This most often occurs with glucosuria from poorly controlled diabetes mellitus, but other solutes can have this effect as well (e.g., infusion of mannitol for treatment of increased intracerebral pressure, administration of urea for treatment of hyponatremia). One underappreciated example of this phenomenon is when a patient with severe dehydration and pre-renal azotemia is re-hydrated, the accumulated blood urea is mobilized and excreted in the urine resulting in a solute diuresis with increased water excretion and decreased Uosm.

b. Primary polydipsia. Primary polydipsia (PP) is a disorder of excessive drinking behavior rather than abnormal AVP secretion or activity. Excessive ingested water produces a mild decrease in plasma osmolality that shuts off secretion of AVP. In the absence of AVP action in the kidney, urine does not become concentrated and a large volume of dilute urine is excreted. The amount of AVP in plasma is unmeasurable or low, but is appropriate for the low plasma osmolality. Although the pathophysiologic mechanism for PP is distinct from DI, patients with PP manifest polyuria and polydipsia just as do patients with DI, and both have a normal serum [Na+] because the normal thirst mechanism is sufficiently sensitive to maintain water homeostasis with DI, and the normal kidney has sufficient capacity to excrete the excess water load with PP. Hence, differentiation of PP from the various types of DI requires more detailed testing as described in the subsequent section.

Key Laboratory and Imaging Tests

Once a patient is verified to have true hypotonic polyuria, the differential diagnosis follows a well-defined sequential series of laboratory and imaging studies, which results in a correct diagnosis in the majority of cases. However, the extent of testing necessary depends on the setting in which the hypotonic polyuria occurs.

Post-operative and post-traumatic polyuria

In a patient with onset of polyuria or polydipsia immediately after surgery in the hypothalamic/pituitary area or after head trauma (especially with skull fracture and loss of consciousness), the diagnosis of CDI is highly likely since such patients rarely are able to drink excessive amounts of fluids. However, some patients may simply be excreting an intraoperative fluid load via an appropriate post-operative diuresis. Therefore, it is best to confirm a diagnosis of CDI before beginning therapy. This is most easily done by withholding intravenous and oral fluids until the serum sodium concentration ([Na+]) increases to >145 mmol/L, which usually occurs within a few hours in the setting of true DI. Persistence of large volumes of hypotonic urine despite plasma hypertonicity confirms a diagnosis of DI and allows commencement of appropriate antidiuretic therapy. The criteria used for diagnosing DI in post-operative and post-traumatic settings are summarized in Table II.

Table II.n
Criteria for Diagnosis of Post-operative or Post-traumatic DI.

Polyuria during pregnancy

Two types of transient DI can occur in pregnancy, both caused by the enzyme, cysteine aminopeptidase, also called oxytocinase or vasopressinase because of its ability to degrade both oxytocin and vasopressin. In the first type, the activity of cysteine aminopeptidase is abnormally elevated, and sometimes occurs in association with pre-eclampsia, acute fatty liver and coagulopathies, e.g., the HELLP syndrome (hemolysis, elevated liver enzymes and low platelets). In the second type, the accelerated metabolic clearance of AVP produces DI in a patient with borderline AVP function from pre-existing diseases, e.g., mild nephrogenic DI or partial central DI. In such cases the neurohypophysis is unable to keep up with the increased demand caused by degradation of circulating AVP. There is the possibility of chronic and severe dehydration when GDI is unrecognized, which may pose a threat to a pregnant woman.

Definitive diagnosis of DI during pregnancy is difficult because of multiple factors: pregnant patients with polyuria may have serum [Na+] levels that would be in the normal range for a non-pregnant patient, but still may be indicative of DI in the pregnant patient because the expanded volume and decreased osmolality and serum [Na+] that typically occurs in normal pregnancy; 2) water deprivation tests are not recommended during pregnancy because induced hyperosmolality may precipitate premature labor as a result of increased AVP levels acting at uterine oxytocin receptors, and 3) measurement of serum oxytocinase levels is not performed by most labs, and a level that is “normal” for a given stage of pregnancy may still cause GDI in patients with compromised AVP function. Consequently, in patients with a high degree of suspicion, a careful trial of desmopressin, which is more resistant than AVP to the actions of cysteine aminopeptidase, is often the most pragmatic way to confirm a presumptive diagnosis.

Spontaneous polyuria

Patients with spontaneous polyuria represent the most difficult challenge in the differential diagnosis of DI. Although a careful history and physical exam are important for determining the etiology of diagnosed DI, they are of limited value to differentiate DI from PP as it is difficult to ascertain whether the polydipsia or the polyuria is the primary precipitant of the syndrome. However, some clues are helpful.

First, central DI often has a sudden onset of symptoms that occur when pituitary AVP secretion falls below levels necessary to concentrate the urine to isotonicity, whereas increased drinking from primary polydipsia generally has a more insidious onset over a longer period of time. Consequently, patients who can describe the onset of their symptoms within a several month window are more likely to have CDI. Second, the polyuria and polydipsia of DI is persistent throughout the day and night, whereas the polyuria of PP occurs predominantly during the daytime in association with increased drinking. Consequently, polyuria with no or minimal nocturia is unlikely to be DI. Third, the polydipsia of DI is associated with a desire for cold liquids, in contrast to PP where patients typically ingest most beverages at room temperature. This is because cold beverages better assuage physiological thirst due to dehydration, which is the underling stimulus to thirst in patients with DI. Consequently, patients with a clear preference for cold beverages are more likely to have DI. Beyond these few hints, the diagnosis of DI depends on careful laboratory evaluation.

Patients with DI usually have serum [Na+] in the higher range of normal while patients with PP have serum [Na+] in the lower range of normal. However, the overlap is so large that this is not diagnostic except in those cases with serum [Na+] outside of normal ranges: hypernatremia and hypoosmolality in a patient with hypotonic polyuria is diagnostic of DI because patients with primary polydipsia are volume expanded, and hyponatremia and hypoosmolality in a patient with hypotonic polyuria is diagnostic of PP because patients with DI will not overdrink fluids since their thirst is abolished once they reach normal levels of plasma osmolality.

Blood urea nitrogen (BUN) concentration is often low in both DI and in PP because of the high renal clearance, but serum uric acid is often elevated in DI both because of modest volume contraction and also the absence of the normal action of AVP on kidney AVP V1a receptors to increase renal urate clearance. Consequently, serum uric acid values >5 mcg/dl are more suggestive of DI than PP, but not diagnostic because of other potential causes of hyperuricemia. Urine volume greater than 18 liters/24h is highly suggestive of PP because this exceeds the amount of glomerular filtrate delivered to the collecting duct. Most patients with DI have modest dehydration, decreased glomerular filtration rate, and have urine volumes in the range of 6-12 liters/day.

Because few baseline studies can differentiate DI from PP, provocative testing is usually necessary for definitive diagnosis. This consists of a water deprivation test, which is the gold standard for differential diagnosis of hypotonic polyuria.

Use of water deprivation to diagnose DI has been in clinical practice since the first description of this procedure by Miller and Moses in 1970. It basically has the goal of making the patient hyperosmolar to see if they can concentrate their urine normally, followed by administration of vasopressin to see if this improves urine concentrating ability. Consequently, there is no indication to do this test in a patient who already presents with hyperosmolality, since they are essentially already at the desired end point of a water deprivation test.

Since the original description of the water deprivation test, many different variants have been proposed, such that there is no single protocol that is currently accepted as the proper way to perform this test. Perhaps the most important of these is the use of direct measurement of plasma AVP levels (so-called “direct test”) rather than the response to administered AVP or desmopressin (so-called “indirect test”) at the end of the period of water deprivation to distinguish among the various types of polyuria. It is important to understand that the direct test was shown to be superior to the indirect test mainly for people whose post-deprivation urine osmolality was >300 mOsm/kg, therefore indicating partial forms of CDI or NDI versus PP. While making an accurate diagnosis is important in such patients, it is not as critical since patients who can concentrate their urine to isotonicity do not in general have large volumes of polyuria due to the inverse relation between urine osmolality and urine volume (See Figure 2).

Figure 2.
Relationship of plasma osmolality, plasma AVP concentrations, urine osmolality, and urine volume in humans. Note that the osmotic threshold for AVP secretion defines the point at which urine concentration begins to increase, but the osmotic threshold for thirst is significantly higher and approximates the point at which maximal urine concentration has already been achieved. Note also that, because of the inverse relation between urine osmolality and urine volume, changes in plasma AVP concentrations have much larger effects on urine volume at low plasma AVP concentrations than at high plasma AVP concentrations. (Adapted from Robinson AG: Disorders of antidiuretic hormone secretion. J Clin Endocrinol Metab 14:55–88, 1985.)

Because both the indirect and direct tests have been shown to be useful in the differential diagnosis of polyuria, we have adopted a procedure that combines both elements by measuring plasma AVP levels at the end of the period of water deprivation followed by administration of desmopressin (Table III). Because the accuracy of interpretation of the plasma AVP level increases with plasma osmolality, it is important to ensure that the patient becomes hyperosmolar by the end of the period of water deprivation. As first described by Robertson, this is best achieved by a very brief period of administration of hypertonic (3%) NaCl to reach a serum [Na+] >145 mmol/L (Table III) in cases where this level is not achieved by the water deprivation itself. Omitting this step will not compromise the ability to interpret the indirect test, but it will potentially complicate interpretation of the direct test.

Table III.n
Fluid Deprivation Test for the Diagnosis of Diabetes Insipidus.

Recent studies of Fenske et al. have published results supporting specific algorithms for diagnosing DI. While these differ in some aspects from previously used criteria for interpretation of the water deprivation test, all authors generally agree on a set of common principles:

1. Patients who achieve a urine osmolality <300 mOsm/kg despite a 3% BW loss have DI. In such patients a ≥50% increase in urine osmolality following AVP or desmopressin administration indicates CDI, whereas <50% is more consistent with NDI.

2. Patients who achieve a urine osmolality ≥800 mOsm/kg despite a 3% BW loss have PP. However, given the nature of relation between urine osmolality and urine volume shown in Figure 2, patients who achieve a urine osmolality ≥600 mOsm/kg very likely have PP since patients who can achieve this urine osmolality do not have clinically significant polyuria.

3. Patients who achieve a urine osmolality between 300 to 600-800 mOsm/kg despite a 3% BW loss are indeterminate by the indirect test, and require additional evaluation. Ideally, this consists of the direct test if AVP levels were obtained at the end of water deprivation. If not, then other factors should be taken into account such as imaging of the posterior pituitary as discussed below.

Recent studies have identified copeptin (the C-terminal fragment of the vasopressin prohormone) as a surrogate measure of AVP and proposed its use as an alternative to measurement of AVP after water deprivation testing. This is appealing because copeptin has a greater molecular weight and is more stable in plasma making it easier to assay. Both copeptin and the ratio of copeptin to final serum [Na+] have been shown to be significantly higher in patients with PP than CDI, though obviously not with NDI since AVP levels are elevated in NDI. However, the sensitivity and specificity of this determination will need to be established before routine use, and at the current time this assay has not been approved by the FDA for diagnostic purposes in the United States.

For all of the above situations, it is appropriate to also consider the utility of imaging studies in order to ascertain the correct diagnosis. For post-operative DI, imaging is not necessary, but will generally be obtained for neurosurgical follow-up. For post-traumatic DI, brain imaging by CT is indicated for the evaluation of a possible basilar skull fracture. For gestational DI, a brain MRI would be useful to rule out a hypothalamic or pituitary lesion, but is relatively contraindicated in pregnant patients. Therefore, the only situation in which imaging is indicated for evaluation of posterior pituitary function is with spontaneous polyuria.

On T1-weighted images the MRI produces a bright spot in the sella caused by stored hormone in neurosecretory granules in the posterior pituitary. The bright spot is present in approximately 80% of normal subjects and is absent in most patients with DI. Some studies have reported a bright spot in patients with clinical evidence of DI. This can be due to stored oxytocin as a source of the pituitary bright spot in patients with a selective loss of vasopressin neurons. The posterior pituitary bright spot decreases with a prolonged stimulus to AVP secretion and has been variably reported to be absent in other polyuric disorders. In PP the bright spot usually is seen. In NDI the bright spot has been reported to be absent in some patients but present in others.

Patients with NDI have high levels of AVP in plasma and are chronically dehydrated, so the posterior pituitary can become depleted of AVP stores. Similarly, with the osmotic stress of untreated diabetes mellitus or transient GDI of pregnancy the posterior pituitary can be depleted of AVP and the bright spot lost, but then return with recovery. Consequently, neither the presence nor the absence of a posterior pituitary bright spot is diagnostic, but the presence of a bright spot is more useful to decrease the likelihood of CDI than is the absence of a bright spot to confirm the diagnosis.

Imaging of the hypothalamus is also an important diagnostic tool for diseases of the neurohypophysis. Because 80-90% of the vasopressin neurons must be destroyed to produce symptomatic DI, it apparent that for a mass lesion or a destructive lesion to produce DI it must either destroy a large area of the hypothalamus, and/or be located where the tracks of these four nuclei converge at the base of the hypothalamus at the top of the pituitary stalk. Notably, tumors confined to the sella do not usually cause DI. With section of the axons or pressure on the axons at the level of the posterior lobe there is a shifting of the site of release of AVP more superiorly above the site of injury, typically in the median eminence. Important exceptions to this are sellar lesions that increase rapidly in size (e.g., metastatic lesions, pituitary apoplexy, rapidly progressing lymphocytic hypophysitis), in which case there is not sufficient time to shift the site of AVP secretion more superiorly.

The pituitary stalk can also be readily identified on magnetic resonance imaging, and has been an additional tool in the differential diagnosis of diseases of the neurohypophysis. Enlargement of the stalk can be seen with inflammatory and infiltrative processes that cause CDI; consequently, the combination of CDI, thickening of the pituitary stalk and absence of the posterior pituitary bright spot mandates a search for systemic and CNS diseases such as sarcoidosis, histiocytosis and malignancies (germinoma, lymphoma). When the etiology is still in doubt repeat MRI should be done every 3-6 months for the first 2 years. When follow-up shows a decrease in size of the stalk a likely diagnosis is lymphocytic infundibuloneurohypophysitis, but an increase in size will mandate a biopsy to rule out a malignant process.

Management and Treatment of the Disease

Appropriate management of DI depends on both the type of DI as well as the clinical situation in which the patient presents, as summarized below.

Central DI

The goal of the treatment of CDI is to correct any existing water deficits and decrease the volume of urine output. Patients with DI generally do not become hypernatremic as long as they have free access to water since their thirst-stimulated polydipsia is able to match their polydipsia. Therefore, the major goal of therapy is to decrease the thirst and polyuria to a level that allows the patient to maintain a normal lifestyle. The timing and quantity of dosage should be individually prescribed and easy for the patient to accommodate. Safety of the prescribed agent and avoiding detrimental effects of overtreatment are primary considerations because of the relatively benign course of DI versus the potential adverse consequences of hyponatremia.

The therapeutic agents to treat DI are shown in Table IV. Water is considered a therapeutic agent because when taken in sufficient quantity there is no metabolic abnormality. Desmopressin, a synthetic analogue of AVP, is the treatment of choice for central and gestational DI. It has a much longer half-life than AVP and lacks the vasopressor activity of AVP at vascular V1aR. Desmopressin is generally administered intranasally (5-20 μg every 8-24 hours) or orally (0.1-0.2 mg every 6-8 hours). Most patients prefer desmopressin tablets, but its duration of action can have significant variability due to gastrointestinal destruction of >99% of the administered dose; to decrease this variability, patients should be instructed to take the tablet on an empty stomach, 1 hour before or 2 hours after meals. A sublingual melt form of desmopressin (60, 120 and 240 micrograms every 8-24 hours) is reported to be more acceptable in children, but is not currently available in the U.S. Parenteral administration (1-2 μg IV, SC, or IM every 8-24 hours) should be reserved for acute situations.

Table IV.n
Therapeutic Agents for the Treatment of Diabetes Insipidus.

Whatever form is employed, the dose should be titrated firstly to control nocturia, and only secondarily to control daytime urination, with the goal of maintaining urine output of <3 L/day. The onset of urinary frequency and thirst are accurate indicators of the need for additional desmopressin dosing. When a dose is sufficient to elicit a stable therapeutic response, further increasing the dose, for example doubling the dose, produces only a moderate increase in duration of a few hours consistent with the half-life of desmopressin in plasma. Consequently, if patients report long periods (>2-3 hours) of polyuria and polydipsia between doses, it is usually more effective to add additional doses rather than to increase existing doses.

Agents such as chlorpropamide or thiazide diuretics can be useful when only a modest decrease in urine volume will make the patients asymptomatic. However, they are rarely used today for primary treatment of DI given the superiority of desmopressin as a targeted therapeutic agent. Perhaps the most important aspect of the other agents listed in Table IV is the recognition that their antidiuretic actions can augment the effect of administered desmopressin, thereby exposing the patient to the possibility of excess water retention and hyponatremia. This is especially true of nonsteroidal anti-inflammatory drugs (NSAID), which inhibit the action of prostaglandin E2. Prostaglandin E2 has a limiting action on AVP-induced water reabsorption by enhancing the retrieval of aquaporin-2 (AQP2) water channels from the apical membrane of collecting duct principal cells and returning them to the intracellular pool. NSAIDs inhibit prostaglandin E2 synthesis and thereby prolong the time the AQP2 water channels remain in the apical membrane, thus extending the duration of action of administered desmopressin.

It must be remembered that unregulated water intake in a patient taking desmopressin can result in severe hyponatremia. However, this occurs relatively infrequently. There are two major reasons for this. First, thirst decreases as the plasma becomes hypo-osmolar, so normal individuals who develop hyponatremia on desmopressin generally limit their fluid intake to a sufficient degree that prevents further falls in the serum [Na+]. Second, the duration of action of desmopressin usually is shorter than the frequency of administration, which allows time for excretion of any retained water before the next dose. Patients at the most risk for hyponatremia are those who dose desmopressin frequently and maintain themselves in a continually antidiuretic state, and those who have high fluid intakes independently of thirst, whether habitually or because of palatability of the ingested fluids.

Recent studies have reported an increased antidiuretic response to desmopressin in females and the elderly. Because desmopressin is primarily cleared renally, the age-associated decrease in glomerular filtration rate leads to a longer half-life of desmopressin with increased risk of hyponatremia using standard dosing. Therefore, treatment of an elderly person with DI requires special attention to avoid hyponatremia. Because elderly persons may have an increased use of NSAIDs, they should be specifically informed of the risk of developing hyponatremia when taking a NSAID with desmopressin.

Females are more sensitive to desmopressin because of increased expression of the AVP V2 receptor owing to its location on the X-chromosome at a position that escapes X-inactivation. Based on the above considerations, it is not surprising that elderly females are most susceptible to developing hyponatremia, even with single doses taken at bedtime for treatment of nocturia. Serum [Na+] should be monitored frequently (i.e., every 1-2 weeks) in patients beginning desmopressin, and the intervals can then be lengthened to every 4-12 months in most patients who maintain normal electrolytes on therapy. Patients should be cautioned to avoid binge drinking and drink according to their thirst. In some cases, building in a weekly “escape” period during which the patient holds a dose of desmopressin until they develop polyuria and polydipsia can be effective in preventing cumulative water retention.

Treatment of infants requires special attention and expertise. Infants consume a large part of their calories as liquid formula or breast milk and have corresponding high volume dilute urine. Treatment with oral or intranasal desmopressin is reported to cause large variations in serum [Na+] and risk of symptomatic hyponatremia. In the U.S. pediatricians have sometimes used desmopressin subcutaneously or substituted a low solute formula with a thiazide diuretic.

Two unique situations require special considerations in the treatment of CDI:

1. Post-operative or post-traumatic CDI.

Sometimes diuresis after surgery is the result of water retention during the procedure. AVP release is stimulated during surgical procedures with subsequent retention of administered fluid. When the stress of surgery abates, the AVP levels fall and previously retained fluid is then excreted. If an attempt is made to match the urine output with further fluid infusion, persistent polyuria might be mistaken for DI. If in doubt, fluid can be withheld until there is a modest increase in serum [Na+]. If the urine output decreases and the serum [Na+] remains normal, the polyuria was due to excretion of physiologically retained fluid. If the serum [Na+] increases to >145 mmol/L while urine osmolality remains low, the diagnosis of DI can be established (Table II). Sometimes the duration of DI is quite transient and the surgeon may prefer to treat this only with fluid replacement parenterally or orally (if the patient is awake and able to respond to thirst).

To treat post-operative DI, desmopressin can be given parenterally 0.5-2 µg subcutaneously, intramuscularly or intravenously. The intravenous route is preferable because there is no question about the degree of drug absorption. Urine output will be reduced in 1-2 hours and the duration of effect is 6-24 hours. If the patient is alert, thirst is a good guide to fluid replacement. Care should be taken that intravenous fluids (especially hypotonic) are not given excessively after administering desmopressin as this can lead to profound hyponatremia. As the DI may be transient and some of these patients may develop the triphasic pattern as described below, it is desirable to allow polyuria to return before administering subsequent doses of desmopressin (Table II).

Postoperative CDI can follow several distinct patterns. The most common pattern is transient DI, which begins abruptly after surgery and usually lasts only 1-5 days. Permanent DI also has a precipitous onset, but it generally results from infundibular or hypothalamic damage leading to longer-term deficits in AVP secretion; this pattern is less common since 85-90% of the vasopressin neurons must be destroyed to cause permanent CDI. The triphasic response is the result of interruption or damage of the pituitary stalk, in which usually less than 15% of hypothalamic AVP neurons survive (See Figure 3 [3A]).

Figure 3.
Mechanisms underlying the pathophysiology of the triphasic pattern of diabetes insipidus and the isolated second phase. (A) In the triphasic response, the first phase of DI is initiated following a partial or complete pituitary stalk section, which severs the connections between the AVP neuronal cell bodies in the hypothalamus and the nerve terminals in the posterior pituitary gland, thus preventing stimulated AVP secretion (1°). This is followed in several days by the second phase of SIADH, which is caused by uncontrolled release of AVP into the bloodstream from the degenerating nerve terminals in the posterior pituitary (2°). After all of the AVP stored in the posterior pituitary gland has been released, the third phase of DI returns if greater than 85% to 90% of the AVP neuronal cell bodies in the hypothalamus have undergone retrograde degeneration (3°). (B) In the isolated second phase, the pituitary stalk is injured, but not completely severed. Although maximum AVP secretory response will be diminished as a result of the stalk injury, DI will not result if the injury leaves intact at least 10% to 15% of the nerve fibers connecting the AVP neuronal cell bodies in the hypothalamus to the nerve terminals in the posterior pituitary gland (1°). However, this is still followed in several days by the second phase of SIADH, which is caused by uncontrolled release of AVP from the degenerating nerve terminals of the posterior pituitary gland that have been injured or severed (2°). Because a smaller portion of the posterior pituitary is denervated, the magnitude of AVP released as the pituitary degenerates will be smaller and of shorter duration than with a complete triphasic response. After all of the AVP stored in the damaged part of the posterior pituitary gland has been released, the second phase ceases, but clinical DI will not occur if less than 85% to 90% of the AVP neuronal cell bodies in the hypothalamus undergo retrograde degeneration (3°). (From Loh JA, Verbalis JG: Disorders of water and salt metabolism associated with pituitary disease. Endocrinol Metab Clinics NA 37:213-234, 2008.).

In the first phase of the triphasic response (which can last up to 7 days), acute polyuria represents transient dysfunction of the hypothalamic neurons, during which time no AVP is released since electrical action potentials from the hypothalamus cannot reach the posterior pituitary to stimulate AVP release. During the second phase (lasting from 2-7 days), there is decreased urine output due to AVP leakage from the damaged or denervated nerve terminals in the posterior pituitary. Although this has been called a “normal interphase”, it actually represents the syndrome of inappropriate antidiuretic hormone secretion (SIADH), in which patients are at risk for the development of hyponatremia. After exhaustion of pituitary AVP stores, the third phase of polyuria resumes due to a permanent loss of a critical number of hypothalamic AVP neurons.

Perhaps the most common post-operative disorder of water metabolism does not involve DI at all, namely isolated postoperative hyponatremia. Transient hyponatremia without preceding or subsequent DI has been reported following transphenoidal surgery for pituitary microadenomas. This generally occurs 5-10 days postoperatively in as many as 30% of patients when they are carefully followed in some series. This scenario can be best understood within the framework of the pathophysiology of the triphasic response, except that in these cases only the second phase of inappropriate AVP secretion occurs because the neural lobe/pituitary stalk damage is not sufficient to cause a loss of >85-90% of AVP neurons (See Figure 3 [3B]). Consequently, this syndrome has been called “isolated second phase” of the triphasic response.

Acute DI after blunt trauma to the head can be treated similarly to postoperatively except that the patient with head injury is more likely to be comatose and unable to respond to thirst, so is more likely to develop hypernatremia. As a comatose patient must be given fluids parenterally some clinicians prefer to use a continuous infusion of low dose vasopressin. The vasopressin can either be added directly to the crystalloid solution that is being administered or infused separately to maintain a constant antidiuresis while adjusting the fluid intake appropriate to any persistent polyuria and to cover insensible water loss. Doses of 0.25-2.7 mU/kg/hr have been described. If this method is used there is a potential to produce hyponatremia and serum [Na+] must be checked regularly. In addition, with continuous replacement rather than intermittent dosing of vasopressin or desmopressin one will not know whether there has been a return of normal function nor whether a patient might be entering the second phase of a triphasic pattern, as described earlier.

2. Adipsic CDI (osmoreceptor dysfunction).

With lack of thirst and continuing polyuria these patients will often develop severe hypernatremia; if encouraged to drink and an antidiuretic agent is administered they are at risk for hyponatremia; so, these patients are subject to wide swings in osmolality but most characteristically persistent hypernatremia. Treatment consists of desmopressin and controlled fluid intake. The patients are not thirsty therefore it is difficult to balance water intake, thus a better treatment regimen is a fixed dose of desmopressin to maintain chronic antidiuresis and a prescribed quantity of water that must be drunk every 6-8 hours. Daily weight can be used to guide intake.

Regular follow-up with measurement of serum [Na+] is essential to assure that these patients do not develop water intoxication with hyponatremia or recurrent dehydration with hypernatremia. This may be especially difficult to manage in infants, but administration of desmopressin by injection combined with careful management of fluids and regular measurement of serum [Na+] has been described successfully. The availability of point-of-care testing for serum [Na+] at home can facilitate monitoring of patients with adipsic DI and allow compensatory changes in fluid administration to prevent ER visits and hospitalization. The high incidence of anterior pituitary deficiency in these patients should also be considered and evaluated along with treatment of DI.

Nephrogenic DI

In all forms of NDI, adequate water intake should always be maintained and may be lifesaving in children congenital NDI. By definition, NDI does not respond to AVP or desmopressin although there may be some mutations that retain some response to higher doses (e.g., 3-5 times standard doses). In congenital NDI therapy is aimed at reducing symptomatic polyuria. This is done primarily by causing volume contraction with a low sodium diet and a thiazide diuretic. The antidiuretic effect has been interpreted as due to extracellular fluid volume contraction, decreased glomerular filtration rate, increased proximal sodium and water reabsorption and decreased delivery of fluid to the collecting duct resulting in a decreased volume of urine. Studies have also suggested that thiazide diuretics can increase AQP2 membrane insertion independently of vasopressin signal transduction pathways. All the thiazide-type diuretics appear to have similar effects. Potassium replacement and/or coadministration of a potassium sparing antidiuretic is often necessary. There can be an added effect obtained by co-administration of NSAID drugs, but gastrointestinal hemorrhage can be an adverse side effect.

Drug-induced NDI should be treated by stopping the offending agent, if possible. Persistence of NDI can be similarly treated by hydrochlorothiazide and/or amiloride. With the induced volume contraction, these patients should be closely followed for the development of renal or other toxicities of the drug that caused the DI. For example, volume contraction produced by thiazide diuretics when used to treat lithium-induced NDI can decrease lithium excretion and thereby exacerbate lithium toxicity. The diuretic amiloride blocks Na+ channels in the apical membrane of the collecting duct cells and inhibits lithium reabsorption, a unique advantage in treating lithium-induced NDI. In animal studies of lithium-induced NDI, treatment with amiloride increased both the levels of AQP2 and urea transporters in the renal tubules.

Gestational DI

Desmopressin is the only therapy recommended for treatment of DI during pregnancy. Desmopressin has 2%-25% the oxytocic activity of lysine vasopressin or arginine vasopressin and can be used with minimal stimulation of the oxytocin receptors in the uterus. The physician must note the naturally occurring volume expansion and reset osmostat that occurs in pregnancy and give sufficient therapy to satisfy thirst and to maintain a serum [Na+] at the low level that is normal during pregnancy. Desmopressin is not destroyed by the cysteine aminopeptidase (oxytocinase) of pregnancy and is reported to be safe for both the mother and the child.

During delivery these patients can maintain adequate oral intake and continue administration of desmopressin. Physicians should be cautious about over administration of fluid parenterally during delivery because these patients will not be able to excrete the fluid and can develop water intoxication and hyponatremia. After delivery, oxytocinase decreases in plasma and the patients may recover completely or be asymptomatic with regard to volume of fluid intake and urine excretion, but the time for complete recovery of normal urine concentrating ability may take up to 2 weeks.

What’s the Evidence?/References

Adrogue, HJ, Madias, NE.. “Hypernatremia”. N Engl J Med. vol. 342. 2000. pp. 1493-1499.

Agha, A, Sherlock, M, Phillips, J, Tormey, W, Thompson, CJ.. “The natural history of post-traumatic neurohypophysial dysfunction”. Eur J Endocrinol. vol. 152. 2005. pp. 371-377.

Baylis, PH, Thompson, CJ.. “Osmoregulation of vasopressin secretion and thirst in health and disease”. Clin Endocrinol (Oxf). vol. 29. 1988. pp. 549-576.

Bichet, DG, Bockenhauer, D.. “Genetic forms of nephrogenic diabetes insipidus (NDI): Vasopressin receptor defect (X-linked) and aquaporin defect (autosomal recessive and dominant)”. Best Pract Res Clin Endocrinol Metab. vol. 30. 2016. pp. 263-276.

Christ-Crain, M, Morgenthaler, NG, Fenske, W.. “Copeptin as a biomarker and a diagnostic tool in the evaluation of patients with polyuria-polydipsia and hyponatremia”. Best Pract Res Clin Endocrinol Metab. vol. 30. 2016. pp. 235-247.

Cowen, LE, Hodak, SP, Verbalis, JG.. “Age-associated abnormalities of water homeostasis”. Endocrinol Metab Clin North Am. vol. 42. 2013. pp. 349-370.

Crowley, RK, Sherlock, M, Agha, A, Smith, D, Thompson, CJ.. “Clinical insights into adipsic diabetes insipidus: a large case series”. Clin Endocrinol.(Oxf). vol. 66. 2007. pp. 475-482.

Fenske, W, Quinkler, M, Lorenz, D, Zopf, K, Haagen, U, Papassotiriou, J, Pfeiffer, AF, Fassnacht, M, Stork, S, Allolio, B.. “Copeptin in the differential diagnosis of the polydipsia-polyuria syndrome–revisiting the direct and indirect water deprivation tests”. J Clin Endocrinol Metab. vol. 96. 2011. pp. 1506-1515.

Fenske, W, Allolio, B.. “Clinical review: Current state and future perspectives in the diagnosis of diabetes insipidus: a clinical review”. J Clin Endocrinol Metab. vol. 97. 2012. pp. 3426-3437.

Grunfeld, JP, Rossier, BC.. “Lithium nephrotoxicity revisited”. Nat Rev Nephrol. vol. 5. 2009. pp. 270-276.

Juul, KV, Klein, BM, Sandstrom, R, Erichsen, L, Norgaard, JP.. “Gender difference in antidiuretic response to desmopressin”. Am J Physiol Renal Physiol. vol. 300. 2011. pp. F1116-F1122.

Khanna, A.. “Acquired nephrogenic diabetes insipidus”. Semin Nephrol. vol. 26. 2006. pp. 244-248.

Liu, J, Sharma, N, Zheng, W, Ji, H, Tam, H, Wu, X, Manigrasso, MB, Sandberg, K, Verbalis, JG.. “Sex differences in vasopressin V(2) receptor expression and vasopressin-induced antidiuresis”. Am J Physiol Renal Physiol. vol. 300. 2011. pp. F433-F440.

Loh, JA, Verbalis, JG.. “Disorders of water and salt metabolism associated with pituitary disease”. Endocrinol Metab Clin North Am. vol. 37. 2008. pp. 213-34.

Miller, M, Dalakos, T, Moses, AM, Fellerman, H, Streeten, DH.. “Recognition of partial defects in antidiuretic hormone secretion”. Ann Intern Med. vol. 73. 1970. pp. 721-729.

Morgenthaler, NG, Struck, J, Jochberger, S, Dunser, MW.. “Copeptin: clinical use of a new biomarker”. Trends Endocrinol Metab. vol. 19. 2008. pp. 43-49.

Oiso, Y, Robertson, GL, Norgaard, JP, Juul, KV.. “Clinical review: Treatment of neurohypophyseal diabetes insipidus”. J Clin Endocrinol Metab. vol. 98. 2013. pp. 3958-3967.

Robertson, GL.. “Diabetes insipidus”. Endocrinol Metab Clin North Am. vol. 24. 1995. pp. 549-572.

Robinson, AG.. “DDAVP in the treatment of central diabetes insipidus”. N Engl J Med. vol. 294. 1976. pp. 507-511.

Robinson, AG, Verbalis, J.G., Melmed, S., Polonsky, K.S., Larsen, P.R., Kronenberg, H.M.. “The posterior pituitary”. Williams Textbook of Endocrinology. 2016. pp. 300-332.

Sands, JM, Bichet, DG.. “Nephrogenic diabetes insipidus”. Ann Intern Med. vol. 144. 2006. pp. 186-194.

Schreckinger, M, Szerlip, N, Mittal, S.. “Diabetes insipidus following resection of pituitary tumors”. Clin Neurol Neurosurg. vol. 115. 2013. pp. 121-126.

Smith, D, McKenna, K, Moore, K, Tormey, W, Finucane, J, Phillips, J, Baylis, P, Thompson, CJ.. “Baroregulation of vasopressin release in adipsic diabetes insipidus”. J Clin Endocrinol Metab. vol. 87. 2002. pp. 4564-4568.

Verbalis, JG.. “Brain volume regulation in response to changes in osmolality”. Neuroscience. vol. 168. 2010. pp. 862-870.

Verbalis, JG., Brenner, B.. “Disorders of water balance”. Brenner & Rector’s The Kidney. 2016. pp. 460-510.