Posterior Pituitary Dysfunction Associated with Pituitary Tumors
Alan G. Robinson, M.D., UCLA School of Medicine
An initial starting point might be to determine the total 24-hour urine output. To do this the patient need not save all the urine but only to keep a record of the time and volume of each voided urine over 24 hours. The first consideration is to document whether there is an osmotic diuresis caused by glucose, renal disease, etc. A hypotonic polyuria may be caused by primary excess drinking of water in which the large urine output is a normal physiologic response; lack of vasopressin, diabetes insipidus, in which the excess urine output occurs due to the lack of action of vasopressin on the collecting duct to produce water retention; or, nephrogenic diabetes insipidus in which vasopressin levels are high but the kidney is unresponsive because of a defect in vasopressin receptor response which may be congenital or caused by drugs such as lithium, or demeclocycline.
To diagnose diabetes insipidus one must determine whether a patient with polydipsia and hypotonic polyuria is able to concentrate the urine in response to normal physiologic input. Physiologic input to the posterior pituitary is from osmoreceptors and baro (volume) receptors. If, when the patient first appears, they are dehydrated with an elevated serum sodium, one need only check the urine osmolality to determine whether the possibility of diabetes insipidus exists. Patients with diabetes insipidus who are conscious will usually have sufficient thirst to maintain normal serum sodium in spite of polyuria. In this situation a standard dehydration test should be performed. This is well described in several textbooks and consists of having the patient under a controlled situation in which they are allowed no fluid to drink while continuing to record their polyuria and weight. When two consecutive voided urines differ by less than 10% in osmolality and the patient has lost 2% of the body weight, the patient is given intravenously 2 micrograms of desmopressin. Normal patients will have a less than 5% increase in urine osmolality over the next 2 hours, whereas patients with diabetes insipidus will have an increase usually greater than 50% of the plateau urine osmolality. Patients with nephrogenic diabetes insipidus and partial central diabetes insipidus may not be distinguished by this test because both may have a modest concentration of the urine osmolality with dehydration and a response of greater than 10% increase in urine osmolality to the administered desmopressin. These two disorders can be distinguished by measuring vasopressin at the end of the dehydration test. Plasma vasopressin levels in patients with nephrogenic diabetes insipidus is by most commercial labs markedly elevated, often as high as 10-20 pg/mL (normal, 0-5 pg/mL), at the end of dehydration. It should be noted that none of the patients under consideration will have a maximum urine osmolality either at the end of dehydration or in response to the administered desmopressin because excessive urine output regardless of etiology washes out the medullary gradient.
In the last few years there have been numerous studies on the vasopressin-sensitive water channel in the kidney, aquaporin-2, which, under the stimulation of vasopressin, transports water across the plasma membrane of the renal collecting duct. Abnormalities of aquaporin2 explain one type of autosomal recessive hereditary nephrogenic diabetes insipidus. Aquaporin-2 is both synthesized in the kidney and excreted in the urine in response to vasopressin. Patients with diabetes insipidus show no increase in Aquaporin-2 with dehydration, but then increased excretion in response to administered vasopressin. The main value of Aquaporin-2 in the differential diagnosis of diabetes insipidus would be to diagnose nephrogenic diabetes insipidus in which there was no increase in Aquaporin-2 excretion in response to administered vasopressin.
Often of note in the history is the relatively rapidly onset of symptoms and the desire for cold liquids. One reason for the rapid onset of symptoms is that while there is a linear correlation between vasopressin, plasma osmolality, and urine osmolality, the relation of those to urine volume is logrhythmic. Thus, a regular slow progressive loss of ability to secrete vasopressin will produce a regular decrease in maximum urine concentration, but cause little change in urinary volume until the last few vasopressin cells are exhausted. At this time there is a rapid increase in urine volume. Most subjects will not note polyuria until the urine is in excess of 4 liters per day. As few as 10% of the vasopressin cells may be sufficient to keep urine volume under 4 L/d. As these last few cells are lost urine volume rapidly increases. The desire for cold liquids has been documented to be secondary to thirst receptors in the oropharynx which are responsive to cold liquids.
It should also be noted that maximum urine output in the absence of vasopressin is never greater than approximately 18 L/d, the approximate volume delivered to the distal nepron. Urine volumes in excess of this indicates some abnormality of thirst.
The primary disorder in diabetes insipidus is disregulation of osmotically stimulated release of vasopressin. Regulation of osmoreceptor sensation, synthesis of vasopressin and release of vasopressin is in a discreet area in the hypothalamus. This small area encompasses the paraventricular nuclei lateral to the third ventricle, the supraoptic nuclei overlying the lateral optic chiasm, the track of these two vasopressin synthesizing groups of neurons through the pituitary stalk to the storage pool in the posterior pituitary. To cause diabetes insipidus greater than 90% of the vasopressin cells in the hypothalamus must be destroyed. Therefore, a lesion which destroys these must be a large lesion in the basal hypothalamus or must be specifically located where the tracks from the four nuclear groups converge to form the supraopticohypophyseal track just above the diaphragm sella. Diabetes insipidus is never caused by a discreet lesion in the posterior pituitary because vasopressin will continue to be synthesized in the hypothalamus and secreted from newly formed terminals above the diaphragm sella.
MRI studies have been shown to localize the normal posterior pituitary as a high intensity signal on T~ weighted images. Research reports indicate the bright spot is related to the content of stored hormone. This high intensity signal is present in most (but not all) normal subjects and is absent in most (but not all) patients with diabetes insipidus. Occasionally the bright spot is present early in the disease and then decreases as the disease progresses. In reported cases of primary polydypsia the bright spot of the posterior pituitary has been present indicating persistent pituitary store of vasopressin. Widening of the posterior pituitary stalk has been reported with a variety of diseases which cause diabetes insipidus, and may also be an early sign of inflammation of the neurohypophysis. A thickened stalk with absence of the posterior pituitary bright spot should prompt a thorough search for systemic diseases known to cause diabetes insipidus. Occasionally the widened stalk is due to inflammation of the posterior pituitary which may regress in follow-up similar to anterior pituitary hypophysitis.
The most common causes of diabetes insipidus are idiopathic diabetes insipidus in which there is no anatomic etiology, tumors in this area which are usually readily identified, trauma to the area due to head injury or pituitary surgery, infiltrative diseases such as granulomatous diseases (histiocytosis or infectious processes) or a hereditary disease occurring in childhood. If no suprasellar mass or systemic disease is found after four years of follow-up, the patient probably has idiopathic diabetes insipidus.
Triphasic Diabetes Insipidus--Triphasic diabetes insipidus is that disorder which occurs after an acute injury to the neurohypophysis as secondary to motor vehicle accidents (commonly with basal skull fracture) or during surgery in the suprasellar area. The pathology is caused by section or contusion of the posterior pituitary stalk. With damage to the stalk there is disruption of function of the vasopressinergic neurons and decreased release of vasopressin. This results immediately in diabetes insipidus. With section of the stalk and disruption of blood flow to the posterior pituitary the axon terminals in the posterior pituitary later become necrotic and begin to release their stored vasopressin in an uncontrolled manner. The posterior pituitary contains a remarkable store of vasopressin. This is sufficient to cause maximum antidiuresis for a period of 5-10 days in humans. Thus, the patients go into a period of uncontrolled release of vasopressin. If excess water is taken during this time, they will develop the syndrome of inappropriate secretion of antidiuretic hormone and hyponatremia (see below). Once the hormone is completely depleted from the posterior pituitary, there is often no further function of the neurohypophyseal neurons and a return of diabetes insipidus, completing the three phases.
Treatment of Diabetes Insipidus
Water--Water taken in sufficient quantities will prevent any pathophysiologic abnormality in diabetes insipidus. However, this is with a large urine output. The therapy of diabetes insipidus is to decrease the urine output and assuage thirst such that the patient is comfortable.
Desmopressin-1-desamino-8-D-arginine vasopressin is the treatment of choice in diabetes insipidus. A dropper bottle with a rhinal catheter can deliver doses of 5-20 micrograms. A metered spray bottle delivers 10 micrograms per spray. Desmopressin is made in tablets of 0.1 and 0.2 mg for oral administration. New patients may first want to try the tablets. The duration of response should be determined in each patient because there is considerable individual variation. This is best done by having the patient record the time and volume of each voided urine after an administered dose. An effect of decreasing urine volume is usually noted within 30-60 minutes and reaches a maximum in 1-2 hours after administration. The return of polyuria is usually rapid as the dose wears off, and the duration of dose can be determined. The duration of effect is about 8 hours for tablets and 8-24 for nasal doses. Control is achieved with 1 dose taken 24 times (usually 2 for intranasal and 3 for tablets) per day. One advantage of the desmopressin tablets is that they are stable at room temperature, whereas the intranasal preparation should be refrigerated. In initiating therapy with tablets doses of 0.05 mg (one-half tablet), 0.1 mg and 0.2 mg can be tried sequentially. Increasing the dose above 0.2 mg results in little additional duration of action. For intranasal it is easiest to start with the spray of 10 micrograms, but the rhinal catheter will allow lower doses, e.g., 5 micrograms. With any of the desmopressin treatments, a smaller dose taken 3 times a day may be more cost effective than a larger dose taken twice a day. When the response to individual doses is determined, a daily schedule can be worked out which allows the patient to sleep through the night and work through the day without excess urine output. It may be advantageous to have the patient delay an administered dose of desmopressin at least once a week to ensure that they do not become water intoxicated. Recurrent episodes of hyponatremia should stimulate re-evaluation of the diagnosis. Desmopressin is also available as a solution of 4 micrograms/mL which can be given parenterally. The usual dose is 0.5 to 2 micrograms. This may be given during surgical procedures or if the patient develops an allergy to intranasally administered desmopressin.
Drugs--Chlorpropamide in doses of 100-500 mg per day enhance the action of vasopressin on the collecting duct. This is most useful in patients with partial diabetes insipidus. Diuretics are useful in treating nephrogenic diabetes insipidus and cause some decrease in urine volume by causing dehydration. Carbamazepine and clofibrate both act centrally to cause increased release of vasopressin. Prostaglandin inhibitors decrease the action of renal prostaglandins of the E series which locally inhibit the action of vasopressin in the kidney. Nonsteroidal anti-inflammatory agents may decrease the urine output in patients with diabetes insipidus and prolong the effect of administered desmopressin.
Patients with diabetes insipidus from trauma or tumor may eventually have hypertrophy of the remaining vasopressinergic neurons and re-establishment of axons adjacent to capillaries in the hypothalamic suprasellar area. If this occurs, there may be return of ability to secrete vasopressin (at least at a low level) sufficient to maintain urine volume in an acceptable range. Such patients may then be changed from desmopressin to chlorpropamide or eventually tapered completely from treatment for diabetes insipidus.
Excess Secretion of Vasopressin Syndrome of Inappropriate Secretion of Anti-Diuretic Hormone (SIADH)
Differential Diagnosis of Hyponatremia
The first requirement is to determine that there is not artifactual lowering of the serum sodium due to the presence of lipids or high levels of protein in the plasma. These agents produce expansion of volume of plasma independent of water. As the serum sodium is determined on a given volume, there is an artifactual lowering of the measured serum sodium. Plasma osmolality can be calculated from the following equation:
1.86 x p[Na+] (meqtL) + glucose/18(mg%) + BUN/2.8(mg%)
This can be compared to an osmolality measured by freezing point depression. The temperature of freezing of the water component of plasma is independent of protein or lipids and, therefore, will be close to the calculated osmolality in the event of true hyponatremia.
Clinical assessment includes evaluation of the extracellular fluid status, neurologic evaluation, examination for cardiovascular or pulmonary abnormalities, and evidence of malignancy. A chest x-ray, serum electrolytes, glucose, BUN, creatinine and uric acid levels with simultaneous measurement of urine sodium and urine osmolality are done.
As indicated in Table 1, the volume status and urine sodium will divide diagnostic categories into four groups of patients. With dehydration and low urine sodium the diagnosis is a normal response to fluid and electrolyte loss, and the appropriate treatment is replacement with physiologic saline. With dehydration and excess urine sodium there is some abnormality of the kidney either due to diuretic use, intrinsic renal disease, or inappropriate hormonal signal as with hypoaldosteronism (usually associated with adrenal insufficiency). The treatment is again replacement with physiologic saline and hormonal replacement as indicated. Normal or expanded (edema) volume with decreased urinary sodium indicates hyperaldosteronism secondary to the perceived inadequate volume as occurs with congestive heart failure or excess accumulation of volume outside the plasma space as may occur in cirrhosis. Treatment of the underlying disease is indicated. Normal volume with excess urinary sodium suggests the diagnosis of inappropriate secretion of antidiuretic hormone which will be discussed in detail.
It should be noted that although measurement of urine osmolality is often described as helpful in making the diagnosis of SIADH, it is of no help. In each of these 4 categories of hyponatremia urine osmolality will be elevated, and comparison of urine osmolality to plasma osmolality is not useful diagnostically. Aquaporin-2 has also been studied in hyponatremia but would appear to have little value in the differential diagnosis. All states of hyponatremia have elevated vasopressin and concentrated urine. Aquaporin-2 is elevated in SIADH but also in cirrhosis in which the mechanism is not the pathophysiology of inappropriate secretion of antidiuretic hormone.
In the case above and in many cases of hyponatremia in elderly subjects it may be difficult to determine the state of hydration. In such patients a cautious diagnostic trial of normal saline may be indicated while following serum sodium and urine osmolality. The urine sodium of 62 meq/L per liter is elevated, but may have been due to recent administration of diuretics or to intrinsic renal disease. With the administration of normal saline, if there is a dilution of the urine and an increase in serum sodium within a few hours, this would confirm that the patient was dehydrated and the normal saline should be continued. If, however, the urine sodium increased and the serum sodium remained the same or even decreased with the administration of physiologic saline, this would suggest that the diagnosis was an inappropriate secretion of antidiuretic hormone and therapy for that disorder should be initiated.
Note that with plasma vasopressin at a fixed level, the urine volume which can be excreted is (as described earlier) related logrhythmically to that level of plasma vasopressin. Thus, even modest levels of fixed secretion of vasopressin will markedly inhibit the ability to maximally excrete the volume of free water. For example, a plasma vasopressin of only 2 pg/mL might limit the ability to dilute the urine to no less than 300 mOsm/kg and limit the ability to excrete more than 34 liters of free water a day. If fluid intake is less than 34 liters per day, there will be no manifestations of SIADH in spite of the fixed secretion of vasopressin. If, however, fluid intakes exceeds the maximum ability to excrete free water, then excess fluid will be retained. With retention of water there is a dilution of serum sodium and some decrease in the serum sodium. As the plasma volume expands, the body undergoes physiologic responses in an attempt to bring extracellular fluid volume back to normal. This is done by excretion of sodium isotonically in the urine. The physiologic responses which induce the excretion of sodium include increased glomerular filtration rate, suppression of aldosterone and initiation of natriuretic factors most notable atrial natriuretic peptide. If the natriuresis succeeds in bringing the extracellular fluid volume back to normal, the stimulus to natriuresis will cease and the patient will be in sodium balance at a low serum sodium. Further increase of volume with sodium or water will again initiate the sequence above and excretion of sodium.
Excretion of sodium results only in correction of the extracellular fluid volume and the intracellular fluid volume will remain expanded. Subsequently, the body attempts to return intracellular fluid volume to normal by excreting intracellular potassium and other osmoles (creatinine, glutamate, glutamine, taurine, myoinositol, and glycerophosphorylcholine). It requires about 48 hours of sustained hyponatremia to initiate intracellular fluid volume accommodation. When these intracellular osmoles are lost, the cells are less able to buffer sudden increases in osmolality. This has important therapeutic implications for neurons as discussed later.
Etiology of SIADH
Excess vasopressin may come from ectopic secretion by a tumor or from the normal source of vasopressin in the hypothalamus. When the disorder is due to abnormal secretion of vasopressin from the hypothalamus, it is probably caused by neurotransmitter input to the magnocellular neurons and generally involves the pathway of the baro (volume) receptors. The baro receptors and volume receptors are in the chest (aorta, large vessels and atria) and signals are carried via the 9th and 10th cranial nerve to synapse in the brain stem, and subsequently transmit messages to the magnocellular neurons where the input is primarily inhibitory. In contrast to the discreet and localized anatomy regulating osmotic release of vasopressin, volume control of vasopressin in more diffuse. Any disruption of the volume receptor system in the chest or central nervous system may lead to decreased tonic inhibition and increased secretion of vasopressin. Thus, a variety of drugs which act on the central nervous system; tumors, infections, trauma, etc., of the central nervous system; and diseases in the chest may lead to excess secretion of vasopressin and SIADH.
Just as administration of water is the primary treatment for diabetes insipidus, restriction of water is the primary treatment of SIADH. This will cause decrease in the extracellular volume and stop the ongoing stimulus to natriuresis. However, as total serum sodium may be depleted due to the natriuresis which attempted to correct the excess volume, there will have to be adequate replacement of sodium.
In profoundly hyponatremic patients (less than 120 meq/L) and especially in those who have CNS symptomatology, a more rapid correction of serum sodium may be required. However, the possibility of inducing CNS damage due to overzealous treatment of hyponatremia should be avoided. As described above, neurons accommodate to chronic hyponatremia by extruding potassium and intracellular osmoles. This allows shrinkage of the neurons back to normal, but also inhibits the ability to respond to a rapid increase in extracellular fluid osmolality which may be caused by administered sodium. In this case neurons may shrink and cause damage including a syndrome known as osmotic myelinolysis. As noted, intracellular volume correction only occurs after approximately 48 hours of hyponatremia. Thus, in the case of severe acute hyponatremia as may occur during fluid administration for a surgical procedure, where a patient is markedly symptomatic, and it is known that hyponatremia occurred over a few hours, there would not yet be accommodation of intracellular osmoles and it may be life-saving to rapidly bring the sodium back up to a level of approximately 120 meq/L.
In more chronic situations slower treatment should be considered. Symptomatology should determine how rapidly to correct patients with chronic hyponatremia. The most conservative therapy is to correct at a rate no greater 0.5 meq/hour, not to exceed 12 meq per 24 hours. Total body sodium can be calculated and a rate of administration of 3 % saline determined (if this is used). As a general guideline, about 10 microliters per kg body weight per minute should result in an increase of approximately 0.5 milequivalents per hour. However, the physician should make this calculation for their patient. An initial rapid replacement may be desirable to correct any cerebral edema. The brain can only enlarge about 10% in the rigid skull, and a rapid correction of 5-10% of the measured sodium may be desirable to decrease cerebral edema. After that the slower rate of correction described above may be continued.
There are cases in which controlling the degree of correction of serum sodium is difficult. Adrenal insufficiency produces a defect in ability to excrete water as a part of the glucocorticoid deficiency. If glucocorticoids are replaced and physiologic saline administered to correct volume deficiency, serum sodium may rise rapidly. If the patient has been chronically hyponatremic, one might consider more modest glucocorticoid replacement and saline replacement.
Patients with diabetes insipidus may become chronically hyponatremic due to over administration of desmopressin. In these patients, if the desmopressin is withheld, the patients will rapidly excrete water and correct their serum sodium. While this is only a concern for severe chronic hyponatremia, in such cases it may be more appropriate to continue the desmopressin and to correct the excess fluid balance by water restriction.
Maintenance Therapy for SIADH
Although fluid restriction is the most physiologic treatment, this may not be possible in all situations. One might then consider demeclocycline which causes a nephrogenic diabetes insipidus by inhibiting the action of vasopressin at the kidney. Doses of 600-1200 mg per day are effective with the maximum clinical response noted about four days after adjusting the dosage. Patients must have serum sodium followed to determine that they do not become hypernatremic, and the potential side effects of this drug need to be considered, especially in the case of liver disease.
Bononi PL, Robinson AG: Central diabetes insipidus: management in the postoperative period. Endocrinologist 1:180-185, 1991.
Buonocore CM, Robinson AG: The diagnosis and management of diabetes insipidus during medical emergencies. Endocrinol Metab Clin North Am 2:411-423, 1993.
Robinson AG: Disorders of antidiuretic hormone secretion. In Ney R1 (ed): Clinics in Endocrinology and Metabolism Investigations of Endocrine Disorders, vol. 14. WB Saunders, 1985, pp 55-85.
Kanno K, Sasaki S, Hirata Y, Ishikawa S, Fushimi K, Nakanishi S, Bichet DG, Marumo F: Urinary excretion of aquaporin-2 in patients with diabetes insipidus. New England Journal of Medicine 1995 Jun 8, 332(23):1540-5.
Nielsen S, Marples D, Frokiaer J, Knepper M, Agre P: The aquaporin family of water channels in kidney:an update on physiology and pathophysiology of aquaporin-2. Kidney International 1996 Jun, 49(6):1718-23.
Maghnie M, Villa A, Arico M et al: Correlation between magnetic resonance imaging of posterior pituitary and neurohypophyseal function in children with diabetes insipidus. J Clin Endocrinol metabol 74:795-800, 1992.
Lindheimer MD, Barron WM, Davison JM: Osmoregulation of thirst and vasopressin release in pregancy. Am J Phvsiol 257:F159-F169, 1989.
Verbalis JG: An experimental model of syndrome of inappropriate antidiuretic hormone secretion in the rat. Am J PhYsio1 Endocrinol Metabol 247:E540-E553, 1984.
Davison JM, Sheills EA, Philips PR et al: Metabolic clearance of vasopressin and an analogue resistant to vasopressinase in human pregnancy. Am J Phvsiol 264:F348-F353, 1993.
Robinson AG, Amico JA: "Non-sweet" diabetes of pregnancy. N Engl J Med 324:556-558, 1991.
Sterns RH. Severe Symptomatic Hyponatremia: Treatment and Outcome A Study of 64 Cases. Annals of Internal Medicine. 107:656-664, 1987.
Bert T. Treating hyponatremia: Damned if we do and damned if we dont. Nephrology Forum, Kidney Internal 37: 1006-1018, 1990.
Robinson, AG and Verbalis JG: Diabetes Insipidus. In: Current Therapv in Endocrinolo~ and Metabolism, Wayne Bardin (ed), Mosby-Year Book, Inc. 1997 In press.