Arch Dis Child 1998;79:431–434 431

Hypertonic saline test for the investigation of Arch Dis Child: first published as 10.1136/adc.79.5.431 on 1 November 1998. Downloaded from posterior pituitary function

Angelika Mohn, Carlo L Acerini, Timothy D Cheetham, StaVord L Lightman, David B Dunger

Abstract (ADH) concentrations. The test is well estab- The hypertonic saline test is a useful tech- lished in adults56 yet there is little reported nique for distinguishing partial diabetes experience of its use in children. insipidus from psychogenic polydipsia, We describe five cases in which the hyper- and for the diagnosis of complex disorders tonic saline test provided a clear diagnosis of of osmoreceptor and posterior pituitary various disorders of water balance where function. However, there is little infor- standard diagnostic procedures had failed or mation concerning its use in childhood. were deemed unreliable. The experience of using this test in five children (11 months to 18 years) who pre- Methods and materials sented diagnostic problems is reported. In All the tests were performed at the John two patients, in whom water deprivation RadcliVe Hospital, Oxford and informed con- tests were equivocal or impractical, an sent was obtained from the patients and/or inappropriately low antidiuretic hormone their parents. Before testing, each patient was (ADH) concentration (< 1 pmol/l) was assessed to exclude the presence of any demonstrated in the presence of an ad- confounding factor such as hypercalcaemia, equate osmotic stimulus (plasma osmola- hypokalaemia, glycosuria or any other cause of lity > 295 mosmol/kg). In two children— a solute diuresis. Free access to food and fluids, one presenting with adipsic hyper- to maintain adequate hydration, was allowed natraemia and the other with hyponatrae- until the morning of the tests. mia complicating desmopressin treat-

ment of partial diabetes insipidus— WATER DEPRIVATION TEST defects of osmoreceptor function were A seven hour water deprivation test was carried identified. Confirming a diagnosis of idio- out between 08:30 and 15:30 using a standard pathic syndrome of inappropriate ADH procedure.7 The test was terminated before secretion (SIADH) was possible in a seven hours if osmolality concentrated to patient with no other evidence of pituitary > 750 mosmol/kg, if weight loss exceeded 5% dysfunction. The hypertonic saline test of initial body weight, or if the patient http://adc.bmj.com/ was well tolerated, easy to perform, and developed intolerable thirst. diagnostic in all cases. (Arch Dis Child 1998;79:431–434) DESMOPRESSIN TEST Keywords: hypertonic saline test; disorders of water An eight hour desmopressin test (DDAVP; balance; diabetes insipidus Ferring, Middlesex, UK) was carried out between 08:00 and 16:00. The child was

weighed and was measured on September 23, 2021 by guest. Protected copyright. Department of The diVerential diagnosis of polyuric states before an intramuscular injection of desmo- Paediatrics, John using standard diagnostic procedures such as pressin (0.1 µg/kg, maximum 4 µg). For the RadcliVe Hospital, the short and long water deprivation tests following hours the child was encouraged to Oxford, UK A Mohn (WDT) are usually reliable in distinguishing void hourly. All urine was collected for C L Acerini severe central diabetes insipidus and nephro- measurement of urine osmolality. D B Dunger genic diabetes insipidus from primary polydip- sia. However, both central and nephrogenic HYPERTONIC SALINE TEST Department of diabetes insipidus may be partial,1 and polyuria Intravenous saline (0.85 mol/l) was adminis- Paediatrics, Royal associated with compulsive water drinking may tered continuously (0.05 ml/kg/min) through Victoria Infirmary, 2 Newcastle, UK limit the ability to concentrate urine. Diagnos- an indwelling catheter for up to a maximum of T D Cheetham tic confusion between these conditions may three hours, or until a plasma osmolality of arise as all are capable of producing a similar 300 mosmol/kg was achieved. The patient Department of rise in urine osmolality using standard diagnos- remained supine from 30 minutes before and Medicine, University tic procedures.3 Furthermore, complex disor- throughout the test. samples were taken of Bristol, Bristol, UK S L Lightman ders of osmoreceptor function such as essential 30 minutes before and at 30 minute intervals hypernatraemia are rarely diagnosed using from the start of the test into lithium heparin Correspondence to: WDTs, as a hypovolaemic stimulus may result containers for measurement of plasma sodium, Dr D B Dunger, Department in apparently normal urinary concentrating ADH concentrations, and osmolality. Urine of Paediatrics, Level 4, John 4 RadcliVe Hospital, Oxford ability. samples were also collected before the start of OX3 9DU, UK. The hypertonic saline test oVers an alterna- the test and at 60 minute intervals, where pos- email: david.dunger@ tive approach to the diagnosis of polyuric sible, for measurement of urinary sodium and paediatrics.ox.ac.uk states, as it defines the relation between plasma osmolality. Thirst behaviour and blood press- Accepted 22 April 1998 osmolality and plasma antidiuretic hormone ure were recorded at 30 minute intervals 432 Mohn, Acerini, Cheetham, Lightman, Dunger Arch Dis Child: first published as 10.1136/adc.79.5.431 on 1 November 1998. Downloaded from 6.0 ioural problems. He was diagnosed as having 5.5 insulin dependent diabetes mellitus at the age 5.0 of 8 years. His metabolic control had been poor 4.5 over the previous year (mean HbA1c 11%). At 4.0 15 years of age he developed a severe episode of 3.5 complicated by cerebral 3.0 oedema and cerebral infarction, which had 2.5 resulted in loss of short term memory. 2.0 Because of serious behavioural problems a 1.5 WDT was impossible and a hypertonic saline Plasma ADH (pmol/l) 1.0 test was performed. He had a small maximal 0.5 increase of ADH (0.4 pmol/l) during the test 0 (fig 1), which was inappropriately low for the 280 285 295290 300 305 310 320315 corresponding plasma osmolality (314 Plasma osmolality (mosmol/kg) mosmol/kg). A diagnosis of diabetes insipidus Figure 1 Plasma ADH related to plasma osmolality was made and treatment with intranasal during hypertonic saline infusion in two patients (triangles, desmopressin (10 µg bid) was started. case 1; squares, case 2) with diabetes insipidus. The shaded area represents the normal range. CASE 3 throughout. If the ADH response to a hypo- This boy presented at 11 months old with an tonic stimulus was to be assessed the hyper- afebrile, generalised tonic clonic seizure. tonic challenge was preceded by an intravenous On admission, hyponatraemia (sodium water load (20 ml/kg of 5% dextrose) given 129 mmol/l) was detected, which persisted over two to three hours. despite fluid restriction. All preliminary inves- Plasma sodium and urine sodium were tigations, including renal function, thyroid measured using standard laboratory methods. function (thyroid stimulating hormone 2 mU/l, Urine and plasma osmolality were measured by total thyroxine 123 nmol/l), and adrenal func- freezing point depression (Microosmometre; tion (random cortisol 426 nmol/l, adreno- Advanced Instruments, Vetech Scientific Ltd, corticotrophic hormone 29 ng/l, aldosterone West Sussex, UK). Blood for determination of 295 pmol/l) were normal. Random plasma plasma ADH was separated in a refrigerated osmolality and urine osmolality were centrifuge and stored at −20°C. Plasma ADH 259 mosmol/kg and 955 mosmol/kg, respec- was measured by radioimmunoassay as previ- tively, and plasma ADH was 2.98 pmol/l. He ously described.8 The limit of detection of the had a low plasma osmolality (272 mosmol/kg) immunoassay was 0.02 pg/ml (coeYcient of at the beginning of the hypertonic saline test variation intra-assay 9.7% and interassay with inappropriately high plasma ADH 15.3%). All samples from each patient were (3.01 pmol/l) and an exaggerated ADH re- measured in the same assay run. sponse (maximum 27.38 pmol/l) during the osmotic challenge (fig 2). The syndrome of http://adc.bmj.com/ Case reports inappropriate ADH secretion (SIADH) was CASE 1 diagnosed and treatment with water restriction A 14 year old girl presented with a three month (500 ml/day) and demeclocycline (150 mg history of polydipsia (up to 14 l/day) and bid) was started. However, the hyponatraemia polyuria. Significant emotional problems were was corrected only after oral sodium supple- identified. A frontal head injury with skull frac- ments were added. Magnetic resonance imag- ture had been sustained at the age of 13 years, ing of the brain remained normal but second-

and nine months before presentation a further ary hypothyroidism developed and he is now on September 23, 2021 by guest. Protected copyright. minor head injury had occurred. All prelimi- receiving treatment with thyroxine (50 µg/day). nary investigations were normal, except for a random prolactin concentration, which was CASE 4 raised to 2387 mU/l. The WDT resulted in A 4 year old boy presented with erratic maximum urinary and plasma osmolalities of drinking characterised by intermittent polydip- 287 and 289 mosmol/kg, respectively. Re- sia (up to 5 l/day) and polyuria. Episodic head- sponses to desmopressin were equivocal with aches and aggressive behaviour had been baseline urinary osmolality 135 mosmol/kg noted. He had been diagnosed three years ear- reaching a maximum of 331 mosmol/kg. lier with histiocytosis X with histologically During the osmotic challenge with hyper- confirmed lesions of the skin, ear, and lung, tonic saline (fig 1), a low maximal ADH which had been treated with prednisolone. A (1 pmol/l) with raised plasma osmolality WDT suggested partial central diabetes insip- (312 mosmol/kg) was detected and established idus (maximum urine osmolality 630 mosmol/ the diagnosis of partial diabetes insipidus. She kg) and he was treated with intranasal desmo- is currently being treated with intranasal pressin (5 µg/day). However, on this treatment desmopressin (15 µg bid), thyroxine (50 µg/ he developed progressive hyponatraemia and day), and hydrocortisone (15 mg/m2/day) as aggressive behaviour. A hypertonic saline test further evidence of anterior pituitary disease was performed (fig 3). At the beginning of the subsequently developed. test ADH concentration was 0.74 pmol/l with a plasma osmolality of 288 mosmol/kg. Subse- CASE 2 quently ADH secretion was not suppressed A 16 year old boy presented with extreme following the water load (0.95 pmol/l) despite a polydipsia (up to 12 l/day) and other behav- decreasing plasma osmolality (283 mosmol/kg) Hypertonic saline test for posterior pituitary function 433 Arch Dis Child: first published as 10.1136/adc.79.5.431 on 1 November 1998. Downloaded from 30.0 plasma osmolality 310 mosmol/kg. By the end of the test the values were 0.2 pmol/l and 27.0 299 mosmol/kg, respectively. During osmotic stimulation, no adequate ADH secretion could 24.0 be detected (maximum 0.6 pmol/l) despite increasing plasma osmolality (310 mosmol/kg) 21.0 confirming the diagnosis of essential hypernat- raemia associated with an osmoreceptor defect. 18.0

15.0 Discussion An infusion of hypertonic saline with measure- 12.0 ments of plasma osmolality and ADH is a sim- ple and well tolerated means of assessing

Plasma ADH (pmol/l) 9.0 posterior pituitary function. It provides reliable information regarding the relation between 6.0 plasma osmolality and ADH. In normal adults and children there is a linear relation between 3.0 those two parameters, which is described by a 0 threshold and a slope. When the threshold 270 275 280 285 290 295 300 305 310 315 320 (average 280 mosmol/kg) is exceeded, ADH Plasma osmolality (mosmol/kg) secretion rises sharply, which leads to antidiu- Figure 2 Plasma ADH related to plasma osmolality during hypertonic saline infusion in resis and to increased urinary concentration. a patient with SIADH (case 3). The shaded area represents the normal range. Below this threshold maximal urinary dilution occurs and ADH is undetectable with current 6.0 radioimmunoassay methods.4 There is a wide interindividual variation in both the slope and 5.5 threshold of the regression lines, but the 9 5.0 relation is individually constant. DiVerent pathological conditions alter this relation, and 4.5 the hypertonic saline test can be used as a 4.0 means of diVerentiating defects of water balance. 3.5 The standard WDT, relying on indirect assessment of ADH secretion, may not be suf- 3.0 ficiently sensitive to identify patients with par- 2.5 tial central diabetes insipidus from those with primary polydipsia.10 Incorporating plasma or 2.0 urine ADH measurements to the test may be Plasma ADH (pmol/l) useful11 12; however, unless the plasma osmolal- 1.5 http://adc.bmj.com/ ity rises above the individual threshold for 1.0 ADH release, a lack of ADH increase cannot be deemed pathological. In case 1 the WDT 0.5 was not able to raise the plasma osmolality 0 above 292 mosmol/kg, the lowest level at which 280 285 290 295 300 305 310 315 320 healthy subjects invariably have a rise in plasma Plasma osmolality (mosmol/kg) ADH.9 Although diVerent protocols for the 12–14 Figure 3 Plasma ADH related to plasma osmolality during water load and hypertonic WDT in children have been proposed it on September 23, 2021 by guest. Protected copyright. saline infusion in one patient (case 4; triangles) with osmoreceptor defect and one (case 5; always involves a long period of observation. squares) with essential hypernatraemia associated with an osmoreceptor defect. The shaded This needs careful monitoring of the patient, area represents the normal range. either because the child may drink while unat- and was subnormally stimulated with osmotic tended or because potentially hazardous dehy- challenge (5.54 pmol/l at plasma osmolality of dration may occur. The hypertonic saline test is 317 mosmol/kg). The diagnosis of osmorecep- shorter, and with modern vein cannulation tor defect was established and his fluid balance techniques well tolerated and easy to perform. is currently stable with desmopressin (5 µg The osmotic challenge given during the hyper- once a week). tonic saline test usually raises the plasma osmolality above the threshold of ADH release CASE 5 and permits diagnosis of partial diabetes insip- This boy was investigated at the age of 18 years idus as shown in our patients. because of adipsic hypernatraemia (sodium SIADH is a rare but well recognised 147–153 mmol/l) and increased plasma osmo- disorder15 and the clinical diagnosis is based on lality (297–305 mosmol/kg). Polyuria was not features first proposed by Bartter and Schwartz evident and daily fluid intake varied between 1 in 1967.16 Once suspected, SIADH requires and 1.5 litres. He experienced thirst only after careful management with water restriction and heavy exercise. He also had abnormal aggres- sometimes by pharmacologically induced ne- sive behaviour and had precocious puberty and phrogenic diabetes insipidus using demeclocy- gynaecomastia. cline. Patients with SIADH usually exhibit a After the water load, a hypertonic saline test characteristic response to water restriction: (fig 3) was performed. Plasma ADH concen- weight loss is accompanied by correction of tration before water loading was 0.6 pmol/l and hyponatraemia. The absence of such a re- 434 Mohn, Acerini, Cheetham, Lightman, Dunger

sponse can make the diagnosis of SIADH Standard diagnostic tests for posterior pitui- Arch Dis Child: first published as 10.1136/adc.79.5.431 on 1 November 1998. Downloaded from doubtful17 and requires other means of confir- tary disorders are usually reliable in most mation. In our patient (case 3) the hypertonic patients presenting with disordered water saline test provided clear evidence of failure to balance. Nevertheless, there are clear advan- suppress ADH completely when plasma osmo- tages to the direct osmotic stimulation of the lality fell below the threshold, followed by an posterior pituitary gland, and the hypertonic exaggerated ADH secretion with respect to saline test may be used in situations where plasma osmolality during the osmotic chal- diagnostic diYculties arise. lenge, diagnostic of SIADH. The hypertonic saline test provides a simple method for confirming persistence of SIADH at regular We thank Miss Karen Jukes for performing the ADH assay. intervals as the natural outcome for idiopathic SIADH is unknown. 1 Miller M, Dalakos T, Moses AM, Fellerman H, Streeten It is well known that histiocytosis X may DHP. Recognition of partial defects in antidiuretic hormone secretion. Ann Intern Med 1970;73:721–9. present with disorders of water balance, such as 2 de Wardener HE, Herxheimer H. The eVect of high water partial central diabetes insipidus, because of intake on the ’s ability to concentrate the urine in man. J Physiol 1957;139:45–52. infiltration of the posterior pituitary or hy- 3 Robertson GL. DiVerential diagnosis of polyuria. Annu Rev pothalamus resulting in local tissue Med 1988;39:425–42. 18 19 4 Baylis PH, Thompson CJ. of damage. However, more complex defects of secretion and thirst in health and disease. 20 Clin Endocrinol water balance have also been reported. (Oxford) 1988;29:549–76. Although the WDT was indicative of partial 5 Moses AM, Streeten DHP. DiVerentiation of polyuric states by measurement of responses to changes in plasma osmo- central diabetes insipidus in case 4, treatment lality induced by hypertonic saline infusion.AmJMed with modest doses of desmopressin was 1967;42:368–77. 6 Baylis PH, Robertson GL. Plasma vasopressin response to complicated by recurrent hyponatraemia. Dur- hypertonic saline infusion to assess posterior pituitary ing the hypertonic saline test, preceded by a function. JRSocMed1980;73:255–60. 7 Dashe AM, Cramm RE, Crist CA, Habener JF, Solomon hypotonic stimulus, this patient showed a per- DH. A water deprivation test for the diVerential diagnosis sistent secretion of small quantities of ADH of polyuria. JAMA 1963;185:699–703. when plasma osmolality fell below 284 8 Rooke P, Baylis PH. A new sensitive radioimmunoassay for plasma arginine vasopressin. J Immunoassay 1982;3:313– mosmol/kg and a subnormal ADH response 30. when plasma osmolality increased. This indi- 9 Robertson GL. Physiology of ADH secretion. Kidney Inter- national 1987;32(suppl 1):S20–6. cates a selective defect of the osmostat charac- 10 Zerbe RL, Robertson GL. A comparison of plasma terised by relative insensitivity to either hyper- vasopressin measurements with a standard indirect test in 4 the diVerential diagnosis of polyuria. N Engl J Med tonic or hypotonic influences. Patients with 1981;305:1539–46. osmoreceptor dysfunction must be managed 11 Milles JJ, Spruce B, Baylis PH. A comparison of diagnostic methods to diVerentiate diabetes insipidus from primary with particular care as they lack an eVective polyuria: a review of 21 patients. Acta Endocrinol Copenh defence against overhydration and underhydra- 1983;104:410–16. 12 Dunger DB, Seckl JR, Grant DB, Yeoman L, Lightman SL. tion. Free access to fluids or regular treatment A short water deprivation test incorporating urinary with desmopressin can lead to water intoxica- arginine vasopressin estimations for the investigation of tion as shown by our patient, while lack of fluid posterior pituitary function in children. Acta Endocrinol 21 Copenh 1988;117:13–18. http://adc.bmj.com/ intake may result in hypernatraemia. Stable 13 Frasier SD, Kutnik LA, Schmidt RT, Smith FG. A water water balance can be achieved using once deprivation test for the diagnosis of diabetes insipidus for 22 children. Am J Dis Child 1967;114:157–60. weekly treatment with desmopressin. 14 Richman RA, Post EM, Notman DD, Hochberg Z, Moses Essential hypernatraemia in association with AM. Simplifying the diagnosis of diabetes insipidus in chil- dren. Am J Dis Child 1981;135:839–41. defective osmoreceptor function, as shown by 15 Haycock GB. The syndrome of inappropriate secretion of case 5, can produce a diYcult diagnostic prob- antidiuretic hormone. Pediatr Nephrol 1995;9:375–81. 16 Bartter FC, Schwartz WB. The syndrome of inappropriate lem that can only be resolved by the hypertonic secretion of antidiuretic hormone. Am J Med 1967;42:790– 23

saline test. Patients with selective osmorecep- 806. on September 23, 2021 by guest. Protected copyright. 17 Verbalis JG. Hyponatraemia. Baillieres Clin Endocrinol Metab tor dysfunction may have normal ADH re- 1989;3:313–30 sponses to hypotensive and other non-osmotic 18 Matus-Ridely M, Raney RB Jr, Thawerani H, Meadows AT. stimuli, but refractory responses to osmotic Histiocytosis X in children: patterns of disease and results 24 of treatment. Med Pediatr Oncol 1983;11:99–105. stimuli. This dissociation suggests normal 19 French Langerhans’ Cell Histiocytosis Study Group. A haemodynamic control over ADH release. On multicentre retrospective survey of Langerhans’ cell histiocytosis: 348 cases observed between 1983 and 1993. the other hand, patients with partial central Arch Dis Child 1996;75:17–24. diabetes insipidus have a subnormal response 20 Dunger DB, Broadbent V, Yeoman E, et al. The frequency and natural history of diabetes insipidus in children with of ADH not only to osmotic stimulation but Langerhans-cell histiocytosis. N Engl J Med 1989;321: also to non-osmotic stimulation.4 The dehydra- 1157–62. 21 Robertson GL, Aycinena P, Zerbe RL. Neurogenic disor- tion achieved during the WDT may provide a ders of osmoregulation. Am J Med 1982;72:339–53. hypovolaemic stimulus to the osmoregulatory 22 Ball SG, Vaidja B, Baylis PH. Hypothalamic adipsic 4 syndrome: diagnosis and management. Clin Endocrinol system suYcient to stimulate ADH release. (Oxf) 1997;47:405–7. The diagnosis of selective osmoreceptor de- 23 Dunger DB, Lightman S, Williams M, Preece MA, Grant DB. Lack of thirst, osmoreceptor dysfunction, early fects is therefore impossible to achieve with the puberty and abnormal behaviour in two boys. Clin Endocri- standard WDT, but is easily characterised with nol (Oxf) 1985;22:469–78. 24 Halter JB, Goldberg AP, Robertson GL, Porte D. Selective the hypertonic saline test as this does not osmoreceptor dysfunction in the syndrome of chronic reduce blood volume. hypernatremia. J Clin Endocrinol Metab 1977;44:609–16.