Expert Opinion

Case Report: Primary Polydipsia vs. Central Diabetes Insipidus (DI) After Head Trauma Finding the cause of polyuria and polydipsia is important to avoid inappropriate vasopressin therapy and resultant water overload.

By Ivica Boban, MD, Jeffrey Miller, MD, and Steven Mandel, MD

ifferentiation of primary polydipsia from dia- Discussion betes insipidus (DI) as a cause of polyuria and Our case is noteworthy since the clinical presenta- Dpolydipsia is important to avoid water overload, tion suggests a diagnosis of central DI, due to onset which can result from inappropriate vasopressin ther- of polyuria and polydipsia after head injury, posi- apy. We present a report of such a scenario: tive family history, and loss of posterior bright spot A 57-year-old male was referred for ambulatory on pituitary MRI. Detailed history along with evaluation of polyuria and polydipsia which he appropriate interpretation of routine laboratory had been experiencing since the age of 17, when studies, however, suggested a diagnosis of primary he suffered closed head trauma from a minor polydipsia. motor vehicle accident. His brother also has a his- Polyuria in adults is defined as urine output tory of polyuria, but less intense. Medical history more than 3L/day (>40ml/kg/day)1. In the absence is unremarkable except for hyperlipidemia, for of the most common causes of polyuria, including which he takes atorvastatin. Patient also reports uncontrolled diabetes mellitus or iatrogenic increased water intake as he is frightened about (furosemide, lithium, mannitol, intravenous fluids) becoming dehydrated. or post-obstructive polyuria, three major categories The physical examination was unremarkable. remain: primary polydipsia, central DI, and Previous workup was extensive, including brain nephrogenic DI. imaging, overnight water deprivation test, and a The sustained increase in water intake in trial of desmopressin. MRI of the brain showed patients with primary polydipsia may be due to loss of posterior pituitary bright spot, but no other hypothalamic lesions of the thirst center (tumors, abnormalities. Overnight water deprivation test infiltrative disease) or more commonly due to revealed low normal plasma sodium level at 137 underlying psychiatric illness. Polydipsia is report- mmol/l, urine specific gravity of 1.009, and urine ed in at least 20 percent of psychiatric in-patients,2 osmolality of 176mOsml/kg. The patient was sub- and frequently these patients are very difficult to sequently started on desmopressin, which was dis- restrain from their habitual fluid intake problem. continued after a few weeks due to iatrogenic In patients with central DI, vasopressin secre- hyponatremia. The plasma sodium concentration tion is decreased or absent. Anatomical causes of was 117 mmol/l requiring hospitalization. central DI can be at the level of the hypothalamus, We were consulted at this time. Given this clini- where vasopressin is synthesized and where hypo- cal scenario, what is the most likely diagnosis? thalamic osmoreceptors are located (supraoptic or

September/October 2011 | Practical Neurology | 35 Expert Opinion

paraventricular nuclei), or at the level of the The likelihood that pituitary stalk thickening is supraopticohypophyseal tract. The posterior pitu- associated with central DI is 80 percent in young itary consists only of the distal axons of the hypo- females with other endocrine autoimmune diseases thalamic neurons. Lesions of the posterior pitu- in the presence of circulating antibodies to vaso- itary rarely cause permanent DI, as hypothalamic pressin secreting cells.6 In children with idiopathic nuclei can still produce and secrete vasopressin central DI, pituitary stalk thickening is commonly directly into the circulation. Thus, injury of the associated with growth hormone deficiency,7 pituitary stalk, which is very close to the hypothal- whereas children with normal MRI and other ante- amus above the level of the median eminence, will rior pituitary endocrinopathies have a higher cause central DI due to denervation of the posteri- chance of having a structural suprasellar8 abnor- or pituitary from the hypothalamus.3 mality that may not be visible on initial imaging; During head trauma, the pituitary stalk is the therefore regular endocrine and neuroradiological most vulnerable, since the pituitary gland is well follow-up is needed. protected within its bony cage, the sella turcica. A rare cause of central DI is hypoxic Depending on the severity of the pituitary stalk encephalopathy (due to shock or cardiopulmonary injury, the degree of central DI can significantly arrest). The least common cause is familial central vary. DI may complicate the postoperative course DI, an autosomal dominant disease caused by in as many as 30 percent of patients who have mutations in the arginine-vasopressin gene with undergone transsphenoidal surgery.4 This is usual- onset of symptoms several months to years after ly temporary. birth.9 Some of these patients who sustained hypothala- Gestational DI,10 with an incidence of approxi- mic or pituitary stalk injury during neurosurgery mately four in every 100,000 pregnancies, is also or trauma will develop the classic triphasic pattern worth mentioning. As it name implies, it is a tran- of central DI.4 The initial phase is polyuria starting sitory form of DI in pregnancy caused by placental soon after the injury, lasting a few days, followed liberation of oxytocinase and thus resolves after by a second phase, in which stored vasopressin is removal of the placenta. Oxytocinase is a placental released from the degenerating posterior pituitary enzyme that indiscriminately breaks down oxy- and therefore patients will have either transitory tocin as well as vasopressin due to their similar normal water balance or will develop hyponatrem- molecular structures. Desmopressin, a synthetic ia in cases of excess water intake. The third phase analog of endogenous vassopresin, is resistant to or permanent central DI develops when vaso- oxytocinase and might be required in some pressin stores are depleted, though case reports of patients with gestational DI. recovery after prolonged time periods have been Nephrogenic diabetes insipidus is characterized described. by genetic or acquired renal unresponsiveness to Roughly half the cases of central DI are caused vasopressin and therefore decreases urinary con- by trauma (surgery or head trauma), tumors (lung centrating ability. Vasopressin acts on two types of cancer, leukemia or lymphoma) or infiltrative dis- peripheral receptors: V1, mostly found on blood eases (Langerhans cell histiocytosis, sarcoidosis, vessels, and V2, mostly found in the renal collect- Wegener’s granulomatosis). The other half are due ing tubules. The activation of V1 causes vasocon- to idiopathic DI,5 where an autoimmune process is striction, and activation of V2 receptors produces the most common culprit, especially in young antidiuresis; hence two names for the same hor- adults. Among patients with idiopathic central DI, mone, vasopressin or antidiuretic hormone (ADH). the most common finding on brain imaging is loss Vascular endothelial cells also have V2 receptors, of the posterior pituitary hyperintense signal and where their activation induces secretion of von pituitary stalk thickening. Willebrand factor.

36 | Practical Neurology | September/October 2011 Expert Opinion

A mild form of nephrogenic DI is frequently seen drawn at the conclusion of the dehydration test in elderly patients due to chronic renal insufficiency. when a desmopressin challenge test is performed. In adults it can also be acquired due to drugs, A patient is given 2Ìg of desmopressin intravenous- among which lithium is the most common culprit.1 ly or intramuscularly. Urine output and osmolality Other reversible causes include hypercalcemia or are recorded initially then hourly for an additional hypokalemia. In children, hereditary causes are two hours. Patients with central DI will concen- more common (90 percent of cases of hereditary trate urine after receiving vasopressin, while nephrogenic DI have X-linked inheritance). patients with nephrogenic DI will not concentrate urine and have high vasopressin levels. Diagnosis Prolonged polyuria and polydipsia, regardless of The diagnosis of DI depends on a careful and com- the cause, can decrease maximal urine concentrat- plete clinical evaluation, followed by random basic ing ability, which can make primary polydipsia metabolic profile, plasma and urine osmolality, and indistinguishable from partial nephrogenic DI. a provocative test for confirmation. In a typical The preferred imaging test in the work-up of case of DI, random plasma sodium is above 143, central DI is an MRI of the hypothalamus-pitu- serum osmolality is above 295mOsm/kg, and urine itary. On T1 images a bright spot in the posterior osmolality is below 150mOsm/kg. Patients with pituitary is useful for excluding a diagnosis of cen- primary polydipsia also have low urine osmolality, tral DI. However, this finding is present in only but they may also have low normal plasma sodium about 80 percent of normal subjects. The bright and osmolality. These typical laboratory findings, spot is an area of hyperintense signal on T1 however, will be present only in patients who images, which corresponds to the portion of the don’t have access to water or patients with an posterior pituitary where neurosecretory granules impaired thirst center, like unconsciousness containing vasopressin are stored. Therefore the patients. Another laboratory finding, albeit nonspe- intensity varies with the amount of stored vaso- cific, is the serum uric acid level. Patients with DI pressin and will be low in severely dehydrated usually have serum uric acid levels above 5mg/dl, patients11 (due to uncontrolled diabetes mellitus or while patients with primary polydipsia have levels nephrogenic DI) or can be completely absent in below 5mg/dl. patients with central DI. The dehydration test is a universal provocative Ideally, treatment of polyuria and polydipsia test that can be used to clinch a diagnosis of DI. should always be directed toward the underlying Dehydration results in elevation of plasma osmo- cause. Unfortunately, that may not be possible in lality, stimulating vasopressin secretion, which the majority of the patients with central DI. Drug then causes appropriate concentration of urine. In therapy, however, is not necessary in patients with patients with large volume polyuria and high clini- mild DI if they have free access to fluids. In cal suspicion for DI, the test must be performed patients with central DI and intact thirst center under medical supervision in a controlled setting where polyuria significantly interferes with their to avoid the potential for extreme dehydration and daily activities or sleep, desmopressin 5-10mcg its sequelae. This is because severe dehydration once or twice daily is administered. In contrast to with resultant hypernatremia can develop within a vasopressin, desmopressin has a much longer few hours with potential serious sequelae. Failure duration of action, is devoid of pressor effects, and to concentrate urine osmolality greater than plas- in the ambulatory setting may be administered ma in response to adequate dehydration is diagnos- orally or intra-nasally. Thiazide is the only avail- tic of DI. able treatment for nephrogenic DI. Psychiatric In patients who cannot concentrate urine, a assessment and intervention are frequently needed plasma osmolality and plasma vasopressin are for a patient with primary psychogenic polydipsia.

September/October 2011 | Practical Neurology | 37 Expert Opinion

Assessing Our Diagnosis tral DI. However, if the desmopressin does not Based on this information, why does the patient in improve symptoms and/or the patient develops our clinical scenario more likely have primary hyponatremia, then the patient likely has some polydipsia rather than DI? abnormality of the thirst center and primary poly- The acute onset of polyuria and polydipsia after dipsia. a closed head injury does suggest a diagnosis of Finally, although checking random ADH levels central DI (due to complete or partial loss of vaso- is almost intuitive, rarely does it prove useful in pressin). The prolonged duration of his symptoms, the diagnostic process. Falsely high or falsely low low normal plasma sodium levels, and his state- levels can be seen due to unreliable assays or ments that he’s “frightened about dehydration,” physiologic suppression of ADH in patients with however, favors a diagnosis of primary polydipsia primary polydipsia. (due to psychological issues or hypothalamic thirst center injury). Additionally, patients with DI (cen- Conclusion tral or nephrogenic) usually present with plasma The diagnosis of DI is often made clinically, while sodium levels above 142, while a level less than laboratory tests provide confirmation. Patients 137 strongly suggests primary polydipsia. with primary or psychogenic polydipsia are distin- Other misleading findings in this case are a sug- guished by longstanding history of polyuria in the gestive family history and abnormal imaging. setting of increased water intake. Symptoms of Recall that hereditary hypothalamic DI starts in underlying psychiatric problems commonly seen in infancy or early childhood, and up to 20 percent of these patients can be very subtle, and in most of normal subjects can have an absent bright spot in the cases routine plasma sodium levels, urine the pituitary. osmolality, and overnight dehydration test are The goal of the overnight water deprivation test enough to make a correct diagnosis. ■ is to assess pituitary secretion of vasopressin in 1.Garofeanu CG, Weir M, Rosas-Arellano MP, et al. Causes of reversible nephrogenic dia- response to water deprivation and the kidney’s betes insipidus: a systematic review. Am J Kidney Dis 2005; 45:626. response to vasopressin. Urine osmolality above 2.De Leon J. Polydipsia--a study in a long-term psychiatric unit Eur Arch Psychiatry Clin 600mOsm/kg excludes DI (indicates normal vaso- Neurosci (2003) 253 : 37–39 3.Robinson AG, Verbalis JG. Posterior pituitary, Kronenberg HM, Melmed S, Polonsky KS: pressin secretion and normal kidney’s concentra- Williams Textbook of Endocrinology 11th ed, Saunders, 2007. 263-286 tion ability in response to dehydration). Urine 4.Nemergut EC, Zuo Z, Jane JA Jr, Laws ER Jr. Predictors of diabetes insipidus after osmolality below plasma osmolality can be seen in transsphenoidal surgery: a review of 881 patients. J Neurosurg 2005; 103:448. primary polydipsia (indicates a normal kidney’s 5. Maghnie M, Cosi G, Genovese E et al. Central diabetes insipidus in children and young diluting ability in response to excess water inges- adults.N Engl J Med. 2000;343(14):998. 6.Pivonello R, De Bellis A, Faggiano A, et al. Central diabetes insipidus and autoimmunity: tion) or in DI (central due to inappropriate pitu- relationship between the occurrence of antibodies to arginine vasopressin-secreting cells itary secretion of ADH or nephrogenic due to inap- and clinical, immunological, and radiological features in a large cohort of patients with propriate kidney response to ADH). Therefore, in central diabetes insipidus of known and unknown etiology. J Clin Endocrinol Metab 2003; 88:1629. this patient the low urine osmolality could repre- 7.Maghnie M, Cosi G, Genovese E, et al. Central diabetes insipidus in children and young sent DI or primary polydipsia. The concurrent adults. N Engl J Med 2000; 343:998. plasma sodium level of 137mmol/l makes the diag- 8.Charmandari E, Brook CG. 20 years of experience in idiopathic central diabetes insipidus. nosis of DI unlikely. Lancet 1999; 353:2212. In some patients, it can be difficult to differenti- 9.De Fost M, Van Trotsenburg ASP, Van Santen HM Familial neurohypophyseal diabetes insipidus due to a novel mutation in the arginine vasopressin–neurophysin II gene. ate partial hypothalamic DI from primary polydip- European Journal of Endocrinology (2011) 165 161–165 sia, in which case a trial of desmopressin may be 10.Kalelioglu I, Kubat Uzum A, Yildirim A et all. Transient gestational diabetes insipidus considered. If a standard dose of desmopressin diagnosed in successive pregnancies: review of pathophysiology, diagnosis, treatment and management of delivery. Pituitary (2007) 10:87–93 improves symptoms without reduction of sodium, 11.Lee MH, Choi HY, Sung YA. High signal intensity of the posterior pituitary gland on T1- then the patient almost certainly has partial cen- weighted MR images: Chorrelation with plasma vasopressin concentration to water depri- vation. Acta Radiol March 2001 42:129-134

38 | Practical Neurology | September/October 2011