SPECIAL FEATURE Clinical Review Current State and Future Perspectives in the Diagnosis of Diabetes Insipidus: A Clinical Review Wiebke Fenske and Bruno Allolio Department of Internal Medicine |, Endocrine and Diabetes Unit, University Hospital Wrzburg, 97080 \Wareburg, Getmany Context: The differential diagnosis of diabetes insipidus (01) is often challenging but essential, because treatment may vary substantially. This article analyzes the database and performance of currently used differential diagnostic tests for Dl and discusses future perspectives for diagnostic improvement. Evidence Acquisition: A review of electronic and print data comprising original and review articles retrieved from the PubMed or Cochrane Library database up to January 2012 was conducted. The search term “polyuria polydipsla syndrome" was cross-referenced with underlying formsof disease and associated clinical, diagnostic, and therapeutic MeSH terms. In addition, references from re- View articles and textbook chapters were screened for papers containing original data. Search results were narrowed to articles containing primary data with a description of criteria for the differential diagnosis of Dl Evidence Synthes ifteen articles on differential diagnosis of DI were identified, mainly con- sisting of small series of patients, and mostly covering only part of the differential diagnostic spectrum of DI. Test protocols differed, and prospective validation of diagnostic criteria was con- sistently missing, Inconsistent data were reported on the diagnostic superiority of direct plasma arginine vasopressin determination over the indirect water deprivation test. Both test methods revealed limitations, especialy in the differentiation of disorders with a milder phenotype. Conclusion: The available data demonstrate limitations of current biochemical tests for the di ferential diagnosis of Dl, potentially leading to incorrect diagnosis and treatment. The newly available assay for copeptin, the Cterminusof the vasopressin precursor, holds promise fora higher diagnostic specificity and simplification of the differential diagnostic protocol in DI. ( lin Endo- crinol Metab 97: 3426-3437, 2012) Dresree bb clon soshespeccun of poly uric and polydipsic diseases, a group of hereditary ‘or acquired disorders mainly associated with an inade- quate arginine vasopressin (AVP) secretion or renal re- sponse to AVP, which clinically results in hypotonic poly- uuria and a compensatory or underlying polydipsia. Under physiological conditions, body water and os- motic homeostasis are balanced by a sensitively regulated AVP system essentially depending on ewo determinants, the serum osmolality and arterial blood volume (1). AVP activates the renal vasopressin-2 receptor (V2R) at the Copy 62012 by Me Endocrine Sacey 3425 —_jcemendojournalsorg basolateral membrane of the principal cells and inereases| the tubular fluid permeability via cAMP-induced phos- phorylation andinsertion of the water channelsaquaporin 2(AQP2) into the apical membrane (2). Thirst servesasan important backup mechanism, becoming indispensable in the moment when pituitary and renal mechanisms fail to ‘maintain body fluid homeostasis. Under these circum- stances, intact thirst sensation is essential to restore nor- ‘mal osmolality, but often at the expense of a clinically pronounced polyuria and polydipsia. Differentiating pri- ‘mary polydipsia (PP) from secondary polydipsia is impor- “atria: ROPE Rquparn HAV aginnesapesan:CD etallorneuogen) rary pole VR, prs 1 Clin Endocrinol Metab, October 2012, 97(10)3426-34371 Clin Endocrinol Metab, October 2012, 97(10):3426-3437 tant because incorrect treatment may lead t0 serious complications. Although simple dehydeation should the- oretically enable the differentiation of both disorders, this approach often fails, especially in mild forms of DI (3) ‘The purpose of this review is to provide a comprehen- sive discussion ofthe available database on the differential diagnosis of DI to better appreciate the diagnostic evi dence and strengths, but also the imitations and pitfalls of the currently available test concepts. Central or neurogenic DI (CDI) is the most common form of polyuric and polydipsic disorder, which occurs mainly due to lesions of the neurohypophysis or the hy- pothalamic median eminence resulting in deficient syn- thesis and/or release of AVP. A number of acquired and congenital disorders may cause CDI (Table 1) (4~6) with variable clinical manifestation, depending on the extent of neuronal destruction, Usually 80 to 90% of the magno- cellularneuronesin the hypothalamus need to be damaged before symptoms of DI arise (7). Acquired CDI mostly results from transsphenoidal sur- gery orhead trauma. Twenty to 30% of pituitary surgeries lead to transient CDI, and 2~10% lead to permanent dis- ease (8, 9), with a risk of postoperative neuronal damage depending on the height of hypothalamic-hypophyseal tract destruction; sections above the median eminence cause a permanent form of disease, whereas sections be- Iow the median eminence produce only transient CDI (10). CDI due to pituitary adenomas is extremely rare (11, 12), even when associated with complete anterior pitu- itary insufficiency (13). Similarly rare, but important to consider, is an autoimmune destruction of the neurohy- pophysis caused by lymphocytic infundibuloneurohy- pophysitis, an entity presenting with a thickened pituitary stalk and enlargement of the neurohypophysis in cranial magnetic resonance imaging (CMRI) (14-16), which in individual cases may spontaneously go into remission (6, 14, 17). This autoimmune process has also been described ina subset of patients initially classified as idiopathic CDI (14). But idiopathic CDI, accounting for about 25% of disease in adulthood, provides no clinical evidence of in- jury oF diseases that could be linked to Dl, and CMRI generally revealsno abnormality other than the absence of the posterior pituitary bright spot and sometimes varying, degrees of thickening of the pituitary stalk (6) ‘Much more rare are the congenital forms of CDI (Table 1), mostly presenting with an autosomal dominant mode emendojoumalsorg 3427 TABLE 1. Eliologies to consider in the differential diagnosis of central diabetes insipidus Cental diabetes insipidus equred “Trauma (surgery, deceleration injury) ‘Vascular (cerebral nemorthage,tnfarctionanteror communicating artery aneurysm or gation, intrahypothalamic hemorrhage) Neoplastic (crantopharyngioma, meningioma, germinoma, pituitary tumor or metastases) Granulomatous (histocytou, sarcoidosis) Infectous (meningits, encephalitis) Inflammatoryfautoimmune tymphocjtic infundibuloneuronypophysit) Drughotnindiced tena, phenytoin, rake ‘ther disorders hydrocephalus, ventrcuarsupaseliar st trauma, degenerative diseases) liopatie ° Congenital ‘Congental malformations “Autozomal dominant. AVP-neurophysin gene mutations ‘Autosomal cessive (21,22) Walram Synorome (DIDMOAD) 23), inked recessive (24) idopathe Nephrogenic diabetes insipidus Recuited Drug-induced (demeclocyclne, lithium, cisplatin, tmethoo/lurane, ete Hypercalcemia, hypokalemia Inftrating lesions (sarcodosis, amides, multiple iryelama, Soergen's disease) vascular sickle cel tsease) congenital Selinked recessive (OMIM 304800): AVP V, receptor gene mutations Aurozomal recessive: AQP2 water channel gene mutations Primary polyps Pyrogen Dipsogente (downward reseting ofthis threshold) Increased AVP metabolism Pregnancy of inheritance (18-20), and only very infrequently due to aan autosomal recessive (21-23) or X-linked recessive in- heritance (24). Underlying mutations, located in the cod- ing region of the AVP-neurophysin precursor gene medi- ate the generation of an abnormal precursor protein, ‘which accumulates within the neuron and causes cell ap- ptosis (25). Therefore, congenital CDI typically develops ‘months to years after birth and then gradually progresses (26,27). Only rare cases of congenital CDI were reported, which later went into remission (28, 29) without evidence for recovered AVP release (30). Osmoreceptor dysfunction Different lesions in the anterior hypothalamus near the thied ventricle may entail a dysfunction of the osmorecep- tors resulting in an impaired AVP responsivity and thirst sensation (Table 1) (31, 32). Patients manifest with dehy- dration and hypernatremia, sometimes resulting in serious3428 Ferske and lal Differential Diagnosis of Diabetes nsipidus cardiovascular complications like hypotension, renal fail ure, and hypovolemic shock (33). The severity of the neu- ological signs is closely related to the degeee of hyperto- nicity, ranging from irritability to disorientation, seizuces, and coma (34). Importantly, patents can stil demonstrate a physiological antidiuretic response to baroreceptor-me- diated (35) and other non-osmotic stimuli of AVP release (36, 37) Nephrogenic diabetes insipidus ‘Nephrogenic DI (NDI) results from renal resistance to the antidiuretic action of AVP, which may also be due to acquired or inherited conditions (Table 1). The acquired form of NDI is much more common than the congenital form, but it is rarely severe. The ability to elaborate a hypertonic urine is usually pre- served, despite the impairment of the maximal concen- trating ability of the nephron. Therefore, the symptoms of disease are usually more moderate with a polyuria and polydipsia seldom above 3 to 4 liters/d. Lithium therapy poses the most frequent cause of acquired NDI (38), involving about 10-20% of patients treated long, term (39). Even at therapeutic concentrations, lithium ‘may interfere with the cAMP system (40, 41), leading to decreased renal AQP2 expression (42) with subsequent polyuria and tubular acidosis (40, 43, 44). This defect in AQP? is slow to correct (45, 46), and in some cases may even persist when associated with glomerular or tubulointerstitial nephropathy (47). Also, electrolyte disorders like hypokalemia or hyper- calcemia may involve NDI, probably due to a temporary down-regulation of AQP? expression (45, 48, 49). But the manifestation is reversible once electrolytes are corrected. Almost 90% of the more infrequent congenital NDI is inherited in an X-linked recessive manner due to muta- tionsinthe V2R gene. Only 10% of congenital diseases are due to autosomal mutations in the AQP2 gene (50). More than 200 putative disease-causing mutations within the \V2R gene have been described (51), and functional char- acterization of some of these mutant receptors revealed reduced binding affinity for AVP (type 1), defective intra- cellular trafficking (type 2), or reduced receptor transcrip- tion (type 3) (50). The genetic form normally presents from birth with basal serum sodium levels being either normal or elevated if fluid administration is also impaired. Primary polydipsia P differs from the other causes of Dl in that it does not arise from a deficient AVP secretion or activity, but rather results from an excessive fluid intake practiced over an extended period of time (52, 53). PP comprises subjects ‘with a defective thirst mechanism or increased thirst sen- 41Clin Endocrinol Metab, October 2012, 97(10):3426~3437 sation (dipsogenic DI) and patients with a more unknown ‘motive for polydipsia frequently associated with psychi- atric disorders (psychogenic polydipsia) (3, 54, 55). The starting point of disease is always the same: an excessive consumption of fluids, resulting in excessive body fluid, decreased serum osmolality, and suppressed AVP release. ‘The subsequent increase in water excretion compensates forthe high fluid intake, and the tonicity stabilizesata new set point around the osmotic threshold for AVP secretion. ‘Therefore, despite intact pituitary and renal function, pa- tients with PP share essential characteristics of DI Because dipsogenic DI can be caused by the same hypo- thalamic lesions as found in CDI (Table 1) iti essential to perform MRI scansin all patients with PP before idiopathic ‘or psychiatric illness can be diagnosed (56). PP may alsohave aniatrogenic origin, either due to drug intake associated with, oral dryness (57} or duetothe medical advice to increase fluid intake for certain health benefits (58). These patients have generally nosigns of mentalillnessor increased thirst, andthe term “habitual polydipsia” may be the more accurate de- scription of this disorder. Gestational diabetes insipidus For the sake of completeness, the gestational form of DI should also be mentioned. The manifestation is triggered by the degradation of AVP by the placental enzyme cys- teine aminopeptidase (59-61), although the hormone lev- els are often normal (62), suggesting that pregnancy un- masks a subtle underlying deficiency of AVP (59, 63). Although spontaneous remission of gestational DI nor- ‘mally occurs within? to 3 wk postpartum, diagnostic eval- uation for possible underlying disorders should be con- sidered (Table 1), Differenti \gno: ‘Once hypotonic polyuria has been confirmed, patient his- tory and clinical presentation can provide useful diagnos- tic clues. But very often, clinical data are of limited help (64) because patients with a preserved thirst mechanism do biochemically notdiffer significantly. And although the assessment of the integrity of the neurohypophyseal sys- tem may be conceptually simple and should theoretically allow a straightforward diagnostic differentiation of Dl, interpretational difficulties often arise (65) andnot seldom lead to misclassification (64, 66, 67) with a high risk of inappropriate therapy. Therefore, a reliable and generally applicable diagnostic approach to suspected DI is ex- tremely important (3, 68, 69). To better understand and characterize the limitations and pitfalls of current diagnostic test standards, we per-1 Clin Endocrinol Metab, October 2012, 97(10):3426-3437 formed an extensive literature review, incorporating elec- tronic and manual components with the search term “polyuria polydipsia syndrome” and its underlying dis- eases. The electronic search was conducted using the PubMed and the Cochrane Library database (up to Jan- uuary 2012) with the following MeSH terms: “diagnosis of polyuria/polydipsia/diabetes. insipidus/(posterior) _pitu- itary disease/hypopituitarism”; “psychogenic. polydip- -ompulsive water drinking”; “water deprivation test”; “hypertonic saline infusion test”; “impaired urinary concentration”; “desmopressin”; and “vasopressin.” In addition, the scientific work of individual authors with exceptional activity inthis field (including G. L. Robert- son, J. G. Verbalis, D. G. Bichet, P. H. Baylis, and C. J Thompson) was analyzed separately. Data and references from major textbooks (70-74) were also included in the analysis. Fifteen original articles dealing with the evalua- tion of differential diagnostic test concepts in DI were re- trieved. Design, endpoints, and main findings of these studies are chronologically summarized in Table 2, and the diagnostic key issues in differentiation of DI are dis- cussed in the following paragraphs against this database. The urinary concentration test as an indirect measure of AVP activity Our current diagnostic test concepts of DI date back to animal studies from Gilman and Goodman (75, 76) in the 1930s, which first demonstrated thar osmotic stimulation by dehydration (75, 76) oF hypertonic saline infusion (77) induces an antidiuretic urinary response that is not ob- served in hypophysectomized dogs (78). Hickey and Hare (79) first applied these findings to ‘man, demonstrating that patients with CDI (n = 2) have ahigher tubular chloride reabsorption and a higher water diuresis during osmotic stimulation than patients with PP. (n = 1) (Table 2). Carter and Robbins (80) equally re- ported a much lower reduction in urine flow in subjects ‘with CDI (n = 8) when compared with PP (n = 3) (Table 2). Both studies, therefore, suggest that the urinary con- centration capacity may provide useful information in the diagnostic differentiation of DI. But the urinary concen- tration mechanism may be disturbed under certain cir- cumstances (e.g. in states of anterior pituitary insuffi- ciency, overhydration, or reduced glomerular filtration ate), essentially limiting the test validity (64) (Table 2). ‘An important modification of the urine concentration test was proposed in the 1950s by Barlow and de War- dener (3). The authors aimed to improve the diagnostic value of urine concentration by adding the evaluation of the urine response to a subsequent injection of pitressin (3). They hypothesized that patients with CDI should re- spond with a further concentration of theie urine, whereas emendojoumalsorg 3429 patients with PP should not. In fact, except for two pa- tients all subjects in their study with PP (n = 8) revealed greater urine concentration after dehydration than after pitressin, despite significant variations in the absolute urine concentration during dehydration (Table 2).Incon- trast, all patients with CDI (n = 12) consistently showed a higher urine concentration after pitressin than after de- hydration. Similar findings were reported by Dies et al. (64), who found a decrease in urine osmolality after va- sopressin injection in patients with PP (n = 2), but an increased urine concentration in CDI (n = 2) (Table 2). Thus, this test concept seemed much more promising than the evaluation of the absolute urine concentration alone. But, its validation in a larger series of patients showing the entire differential diagnostic spectrum of DI has long been missing, A first standardized test protocol with detailed diag- nostic test criteria was proposed by Miller e¢ al. (67) in 1970 after the evaluation of 36 patients with different disorders of polyuria and polydipsia syndrome (Table 2). Patients with severe CDI (n = 18) were reported to fail in concentrating the urine above the level of plasma osmo- lality during dehydration but to feature a more than 50% urine osmolality increase (range, 51-792%) after exoge- nous vasopressin injection; patients with NDI (n = 2) showed a further rise of only 39 and 45%. In contrast, subjects with mild forms of CDI (n = 11) were able to concentrate their urine above their plasma osmolality but still demonstrated a further rise in urine concentration after vasopressin by more than 9% (range, 9~67%). A ‘much higher urine concentrating ability was found in pa- tients with PP (n = 5), who showed only a minor further increase in urine concentration alter vasopressin, canging, between 2 and 13%. Unfortunately, the diagnostic criteria derived from this study (67) (see Table 2 for details) have never been val dated in an independent, prospective approach, and sen- sitivity and specificity for identification of differential di- agnoses of DI have not been defined. Furthermore, it was shown that some patients with partial CDI may demon- strate quite normal urine concentrations, in particular in case of a decreased glomerular filtration rate (3, 64, 81). [And the urine concentration ability may be variably re- duced in different forms of chronic polyuria (82), leading to overlapping degrees of pathologically diminished uri- nary responses to osmotic stimulation (67). This diagnos- tic problems not surprising, given some principles of kid- ney physiology. First, chronic polyuria itself can decrease the urinary concentration capacity (83-85), probably through a washout mechanism of the renal medullary con- centration gradient as well as by down-regulation of kid- ney AQP2 water channels. Both may explain the some-330 Fenske and Allolo Differential Diagnoss of Diabetes nsipidus 41Clin Endocrinol Metab, October 2012, 97(10):3426~3437 ‘TABLE 2._Collecion of the original data on the differential dagnosis of central diabetes insipidus in chronological order oo pain ee (compat Dp Method of tring ohn ep and ‘av veocivy ay ent ‘lute Major fing patents with CO trie) 2 naclten tow eenkpoaee nce Ufo eng on: amet 2s acl ans repreta Nolin Sujet psa ape, sate: 2 change Oem ne nUomighr compet ‘Sipdaton sean eragenin eater, 59 ‘Uoimaueeapein 3 ser nope wae ong oF Tee nvstenne tencatre, prea, sale on nbc th "hasan to cine bu 355s repre toss he tr ere tascam 191316) cation retosamamiy ato cea thpnayo38 ——unemunmomaaly toler nF ond sas mien tone? Vom 3 nero ee 1Op=3.¥=38 Ne cangem low, Ueno UR cna 1870167 Owhtaton atowed by 1. Maximum Cam ding cen te tn ‘nse CD are ND U-Csm rem below 0am oorean npn ‘hater 2 denge non pee to ‘erg Syson a er ey 25250 ‘iat narra by 29% frac Daren 9 =a ohety ue 1981 (86) S% Naina ‘ret plana AVP response to eaScnSpet Unctecable AVP ee compete CDI = 5) toate stenectnense constneynderal ms «AWE =7.€= 8.823, Saha ceomeysertdcampee CD = ‘sation at ‘awe cate por Two pater pal Ca reo 12 rescrmenoaretedhyMkt WOT =7.c=6.n=0, patent, scary theyre |e) stenactnsense Sane enipant it 82 BPs he-9.0-2, — dahtemcamcty earaeatein oh Moshe ‘esaton at poe ‘abject wth nda OP ete worce=%.¢~2.0=a tanned an crite lye dayne Po? Senet dad prc on 1. Deering Soya dB pater th by te ation tone aera agree ee ee Insteeaten er th eloing Sam at by hyaon nce (havea nsep sale or crest ner cra. ra aton toms) Thaghadmayete agate pete he pope go he potty i recnarceimagig ‘eachpope bhp for ‘sein Din = pease oes paencueeke ee Bone) oatydatn et Dieceanpaanat 1 here eemenaann Le oncaeid ‘at scp Band dmc aspra cgretfn $s tpaern pe opp rie tear FIVC, Free water clearance; S-Osm, serum osmolality, flow, urine flow: UN‘, uin sodium excretion, chloride UP rat, relation of chloe concentration in the reabsorbate to that in plasma; U-Osm, urine osmolality; WOT, water deprivation test.1 Clin Endocrinol Metab, October 2012, 97(10):3426-3437 times lower than expected urinary response to injected vasopressin in CDI (86). Second, an enhanced antidiuretic response to low levels of plasma AVP has been reported (81, 87), possibly due to an up-regulated V2R expression inchronic states of CDI (88). On the other hand, patients with acquired NDIare often only incompletely resistant to AVP, and therefore may resemble patients with partial CDL-Thus, the consequence of these mechanisms is con- vergence of different underlying disorders to a similar uri- nary phenotype. Accordingly, when we recently evaluated in a prospective study the diagnostic accuracy of the pre- viously described urine concentration test (67) in 50 pa- tients presenting the entie differential diagnostic spec- trum of DI, we only found a total diagnostic accuracy of 41% (68) (Table 2) Direct measurement of plasma AVP activity With the availability ofa firstsensitive and specific AVP RIA in the 1970s, it was hoped that direct plasma AVP. ‘measurement would overcome the limitations of the urine concentration tests and allow a less cumbersome differ- ential diagnosis of DI (87). The new test method, first implemented by Baylis and Robertson in the early 1980s (89, 90), exploited the direct assessment of AVP release in response to a 2-h hypertonic saline infusion test, and the data were interpreted as a function of the corresponding, serum osmolality (Table 2 and Fig. 1). The conceptual superiority of this method was assumed to consist of an intact AVP responsivity to osmotic stimulation even after a long period of overhydration or dehydration (54). If compared with the normal physiological relationship be- ‘owen plasma AVP and serum osmolality (Fig. 1, gray area), patients with CDI were hypothesized to demon- strate data pairs below this area of normality, whereas patients with NDI or PP show data pairs above or within this normative area (89, 90) (Fig. 1 and Table 2) Pasa realy nos) FIG. 1. Plasma AVP and osmolality response to hyperton saline infusion, The shaded area defines the normal response. Red rangle symbolizes P, and green cice symbolizes CD. LD the limit of detection of plasma AVE (modified from Ref. 90. eemendojoumalsorg 3431 Two peer-reviewed studies investigating the diagnostic performance of direct AVP measurements in differential diagnosis of DI are available in the literature. The first study, from Zetbe and Robertson (82) (Table 2}, co pared the diagnostic performance of plasma AVP mea- surement with that of the urine concentration test (67) in 23 patients. Patients with complete CDI (n = 7) could be easily identified by both test methods. Discrepant results, however, were found in patients with partial CDI and PP: ‘wo of six subjects diagnosed as partial CDI based on their urine levels revealed an entirely normal AVP response;one was later diagnosed as PP, the other as NDI. Vice versa, three of 10 patients diagnosed as PP by urine concentra- tion had suppressed plasma AVP levels, clearly suggestive of AVP deficiency (82) Similar diagnostic discrepancies between both test methods were reported by Milles t al. (69), who repeated. the study in another 19 patients (Table 2). Both tests came to the diagnosis of PP in the same six patients, and also eight of nine patients with subnormal AVP levels could be confirmed as partial CDI by the urine test results; only one patient was diagnosed as NDI. But, two of four patients with undetectable plasma AVP levels were categorized as partial CDI (n = 1) or NDI{(n = 1) according to their urine results ‘An important difficulty in the interpretation of both studies is the lack of a diagnostic “gold standard” with which the test results could be compared. Thus, the final diagnosis and, therefore, the diagnostic accuracy of both test methods remain undetermined. Another eritical point refersto the area of normality describing the physiological behavior of plasma AVP release in relation to the corre- sponding serum osmolality. The exact definition of this area is the fundamental basis for the identification of ab- errant AVP secretion in patients tested for disease. On this crucial point, however, both studies (89, 90) relied on a previously described slope model of linear regression cal- culated on the basis of a very small number of healthy volunteers and without the definition of the correspond- ing reference area or stability limit (69, 87, 89). But more recent evidence, froma higher number of healthy controls, suggests that the relation between AVP and serum osmo- Lality is less linear and close than has previously been de- scribed (68). Consequently, a total diagnostic accuracy of only 46% could be found when direct AVP measurement ‘was prospectively tested in 50 patients with polyuria poly- dipsia disease (68) (Table 2). Finally, the well-character- ized technical difficulties of the AVP assay with its high preanalytical instability (87, 91, 92) certainly also account for the still extremely low acceptance of the AVP assay in the differential diagnosis of DI2432 Ferske and Aloo ifrentalDagnoss of Diabetes sis 2 z Zw & 2 Bo FG. 2. Th four rian memsred at 16 Getycaton test and ilsatd standard box and whaker pts or he cntl group, nd fr patents Paral CO ar compete COL 2.005 compared wth contol, bP 005 compared wth Parka COI) sct measurement of plasma copeptin (C-terminal proAVP) A possibly novel approach to the diagnosis of DI offers. the measurement of plasma copeptin in response to an, appropriate osmotic stimulus. Within the last few years, copeptin, the C-terminal glycoprotein of the AVP prohor ‘mone, has been established as an easy to measure and stable surrogate for endogenous plasma AVP (91, 93, 94). Recent data have also promoted copeptin as a promising diagnostic marker for identification of patients with severe neurohypophyseal damage after transsphenoidal surgery (n = 38) (95). Based on these findings, another prospective study evaluated the diagnostic potential of copeptin across the FIG. 3. Sagittal TT -neighted magnetic resonan (aronhead and pituitary sta (i (thc i the posterior ptutary gland larouheach imaging cans ofthe hypothalan region ina patient with FP (A) and a patent with CDI). A, Arora anterior ptuitary nite arrow) and hyperintensty ofthe posterior pituitary row 8, Sogital TI-veighted scan obtained patent with CD ue to ymahoma shows amacivel thickened pitutary talk (thin white arrav) and an absent hyperintense signal in 41Clin Endocrinol Metab, October 2012, 97(10):3426~3437 full differential diagnostic specteum of DI (n = $0) (68) (Table 2}. In accordance with the direct AVP test method (see The Urinary Concentration Test as an Int direct Measure of AVP Activity and Fig. 1), osmotically stimulated copeptin release was interpreted by plotting the hormone levels together with the corresponding se rum osmolality into a nomogeam and estimating the data paits as to theie position to the area of normality In this way, copeptin revealed an overall diagnostic ac curacy of 83%, whereas patients with complete CDI (n = 7) and NDI (n = 2) could already be identified at plasma baseline values. With respect to the particularly challenging differentiation of patients with PP and partial CDI, the ratio of copeptin increase duringan 8-hdehydeation period to the serum sodium concentration measured after, 16 h of water deprivation (ACopeptinys 16 uYS-Na* x6 »)) proved to be a highly useful parameter, providing a di agnostic yield of 94%, and a specificity and sensitivity of 100 and 86%, respectively, to separate PP (n = 22), from partial CDI (n = 17) (Fig. 2). The main criticism, to this approach, again, is the missing “gold standard” test with which the copeptin results could have been compared. To cope with this problem, the reference diagnosis in this study was made retrospectively, with consideration of patient history, clinical presentation, and treatment response. Moreover, the clinical use of ccopeptin levels as a surrogate of AVP secretion during os ‘motic stimulation still requires a larger database of normal and abnormal responses, and the reported diagnostic cutoff values that were determined in a post hoc analysis stil need. to be validated prospectively. Also important to note is the factthat the sandwich immunoluminomettic copeptin as- say (LUMItest CT-proAVP) is not yet commercially available in countries outside Europe, currently limiting its clinical implementation as a new diagnostic standard. Further proposed test procedures [Ag might be expected, most patients with DL also exhibit a subnormal antidi: uretic response to nonosmotic stimuli such as nicotine (96), hypotension, nau: sea, and hypoglycemia (90). But for di agnostic purposes, tests of neurohypophyseal function do not provide advantages over the osmotic ee difficult to con- these nonosmotic stimuli because they twol or quantitate and generally cause a markedly variable AVP response (90, 96). Moreover, they can lead to false-pos- pituitary itive or false-negative test results because there are patients who exhibit little or no4.Clin Endocrinol Meta, 0 "2012, 97(10)3426. rise in AVP after hypotension or emesis, yet lack polyuria and have a normal response to osmot versely, patients with osmoreceptor dysfunction exhibit, litele or no antidiuretic response to hypertonicity but c stimuli emendojoumalsorg 3433 show a normal response to induced hypotension (35) or nicotine infusion (96) (Table 2), An interesting diagnostic test method has been pursued by Thompson et al. (97) (Table 2), who examined the Diagnostic Approach to Differential Water Deprivation Test (after hypotonic polyuria has been confirmed) < 200mOsmikg HO ‘increase in Urine Osmolality following desmopressin, Compl Cental DI > 800mOsmikg HO ! ws Primary Plypsia 300 to 300 to <800 mOsm/kg) are diagnostically indeterminate. In these cases, measurement of plasma AVP may be useful, provided that itis performed with a sensitive and validated assay for AVP, reliable preanalyti- cal handling of samples is guaranteed, and corresponding, serum osmolality is in the hypertonic range, preferably above 300 mOsm/kg. Otherwise, the differentiation be- tween PP, partial CDI, and partial NDI should currently rely on a comprehensive assessment of patient and family history, clinical data, cMRI imaging of the hypothalamic- pituitary region, anda therapeutic trial with desmopressin (Fig. 4) There are emerging data on copeptin measurement in the differential diagnosis of DI to suggest that this surto- gate provides a novel and more robust test concept of neurohypophyseal function than current test standards. But importantly, before its implementation in differential diagnostic routine procedures, previous findings on co- peptin in DI still need to be confirmed in larger prospective trials, The ultimate aim should be to use the diagnostic potential of copeptin for a simplification of the currently still cumbersome diagnostic workup of DI Acknowledgments [Address all correspondence and requests for reprints to: Dr. Wiebke Fenske, M.D., and Prof. Bruno Allolio, M.D., Univesity Hospital of Witreburg, Department of Medicine, Division of En- dlocrinology and Diabetology, Oberdkirebacher St. 6,97080 Wir- hung, Germany. E-mail: Fenske w@medizin.uni- waereburg.de and Allolio b@medirin.sni-wuerzburg de. \W.P. was supported by the German Research Society (DFG). Disclosure Summary: The authors have nothing to declare. References 1. Robertson GL. 1995 Posterior pituitary. In: FeligP, Baxter J, Fro- hhman L, eds. Endocrinology and metabolism. New York McGraw-Hill, 385-432 2. Knepper MA 1997 Molecular physiology of urinary concentrating1 Clin Endocrinol Metab, October 2012, 97(10):3426-3437 20 24 mechanism: regulation of aquaporin water channel by vasopees- sin, Am [Physiol 2723-12 Barlow ED, De Wardener HE 1959 Compulsive water drinking, Q J Med 28:235-258 (Gaernichow PPR, Brauner R, Czernichow P, PomaredeR, Brauner 1985 Neurogenic diabetes insipidus in children. I: Caenichow , Robiason AG, es. Frontier of hormone eesearch, Vl 13.Di- abbetes insipidus in man. Basel, Switzerland: S Karger; 190 GGeeger NG, Kikland RT, Clayton GW, Kiekland JL 1986 Cental diabetes insipidus: 22 yeas’ experience Am JDis Child 140-551 5st -Maghnie M,CosiG,GenoveseE, Mancs-Biti ML. Cohen A, Zecea 5, TinliC Gallucci M, Bernasconi 8, BoscheriiB, SeveiE, Arico 1M 2000 Cental diabetes insipidus in chldren and young adults. 1 Engl J Med 343:998 1007 Heinbecker P, White HL 1939 The role ofthe pituitary gland in water balance. Ana Surg 110:1037-1049 Heasen J, Henig A, Fahlbusch R, Meyer M, Bochnett M, Buch felder M 1999 Prevalence, predictors and pater of postoperative polyuria and hyponatraemia in the immediate course after teans- sphenoidal surgery foe pituitary adenomas. Clin Endocsnol (Oxf) 50:431-439 Nemergut EC, Zuo Z, Jane Je JA, Laws Jr ER 2005 Predicioes of diabetes insipidas ater eanssphenoidal surgery: a review of 881 patients. J Neurosurg 103-448-454 Lipset MB, Maclean JP, West CD, Li MC, Pearson OH 1956 Aa analysisof the polyuria induced by ypophysectomy aman. Jin Endocrinol Mecab 16:183-195 Hoshimaru M, Hashimoto N, Kikuchi H 1992 Cental diabetes insipidus eesuling from a nonneoplastic tiny mas esio localized inthe neurohypophyseal system, Surg Neurol 38:1~6 Koshiyama H, Sato H, Yorta 8, Koh T, Kanasuna T, Nishimura K, Hayakawa K, Takahashi J, Hashimoro N 1994 Lymphocytic hypophysitis presenting with diabetes insipidus: case report and. literature review. Endoce 4193-97 Vesbalis JG 1984 Postoperative and posttraumatic diabetes insip- ids. In; Cernichow P, Robinson AG, eds. Diabetes insipidus in ‘man, Basel, Switzerland: S Karger Tea H, Nakao K, Shimatsu A, Ogawa Y,Sando"T, Fujisawa I, Yamabe H 1993 Lymphocytic infundibuloneurobypophysiisas a Val inthe signal peptide ofthe arginine vasopeesin/neurophysin Mopeptin pee- utsor. J Clin Endoctinol Metab 82:51-56 Robertson GL. 1988 Riddles posed by clinical studies of vasopees- sin. ln: Cowley Je AW, Liaed JE, Ausiello DA, eds. Vasopressin cellular and integrative functions. New York: Raven Press; 169— 176 Kaplowitz PB, D'Ercole AJ, Robertson GL 1982 Radioimmuso- sey of vasopressin in familia central diabetes insipidus. Pediat 100:76-81 Slillee PG, Corcoan AC, Schebel AL. 1956 Renal mechanismsin| coincident Addison's disease and diabetes insipidus: effect of va- sopeesin and hydrocortisone J Cin Endocrinol Metab 16:171— 182 Buggy J,Jonhson AK 1977 Peeoptic-hypothalamic periventriculae lesions: thst deficits and hypernatremia, Am J Physiol 233:R44— RS? Thrasher TN, Keil LC, Ramsay DJ 1982 Lesions ofthe organu sasculosum ofthe lamina tesmialis(OVIT) ateauate osmotical- Iy-indaced deiaking and vasopeesin secretion inthe dog. Endo- erinology 110:1837-1839 AAdrogu Hi, Madias NE 2000 Fypernatremia.N Engl] Med 342: 1493-1499 Riggs JE 1989 Neurologic manifestations of uid and elacteoyte disturbances. Neucol Cin 7:509-523 Smith D, MeKenna K, Moore K, Tormey W, Finucane J Philips J, BaylisP, Thompson C}2002 Baroregulation of vasopeesin lease in adipic diabetes insipidus. J Clin Endoceinol Metab 87:4564— 4568 Halter JB, Goldberg AP, Robertson GL, Porte JD 1977 Selective csmoreceptor dysfunction in the syndrome of chronic hyperna- twemia. J Clin Endocrinol Metab 44:609-616 DeRtubertis FR, Michelis ME, Beck N, Feld JB, Davis BB 1971 “Essential” hyperatemia due © ineffective osmotic and intact, volume regulation of vasopressin secetion. J Clin Invest 0:97— a Boton , Gavia M, Batlle DC 1987 Prevalence, pathogenesis, and treatment of renal dysfunction associated with chronic lithium therapy. Am J Kidney Dis 10:329-345 Bendz H, Aurel M 1999 Drug-induced diabetes insipidus: nci- ‘dence, prevention and management. Drug Saf 2149-456 Christensen 5, Kusano E, Yusufi AN, Murayama N, Dousa TP 1985 Pathogenesis of nephrogenic diabetes insipidus due co heoaicadinistation of lihium ia rats. J Clin lnvest 75:1869— 1879,2936 4 4 45, 46 7 48 so. 1 8. ss. ss. se. s. ss. 8. oo 61 Fenske and Allolo Differential Diagnoss of Diabetes nsipidus Kortenocven ML, Schwser H, Cox R, Wetzls JF, Deen PM 2012 Lithium reduces aquaporia-2transcripcion independent of pros taglandins. Am J Physiol Cell Physiol 302:C131-C140 Marples D, Chestensen S, Christensen El, Ottosen PD, Nilsen S 1995 Lithium-iaduced downregulation ‘of aguaporin-2 water chanael expession in at kidney medul. J Clin Invest 95:1838— 1845 Cogan E, Svoboda M, Abramow M 1987 Mechaniss of lihium- ‘asopressin interaction in rabbit corial collecting rubule. Am J Physiol 252:F1080-F1087 Goldberg H, Clayman P, SkoreckiK 1988 Mechanism of Liinhi- bition of vasopressin-sensitve adenylate cyclase in cultured renal epithelial cells. Am J Physiol 255:F995-F1002 Marples D, Frokiaer J, Dorup J, Knepper MA, Nielsen $ 1996 Hypokalemiainduced doseneegulation of aguaporin2 water chanel expression in rat kidney medulla and cortex. J Clin Invest 97:1960-1968 Bend2H, Sjédin |, Aull M1996 Renalfunctiononandofflichina in paints tated wih ithium for 15 years ormoce- A coatelled, prospective lthium-wichdeawal study. Nephrol Dial Tansplant 11:457~460 Markowitz GS, Radhaksshnan J,Kambham N, Valeri AM, Hines WH, D’Agati VD 2000 Lithium nepheoroxcitysa progressive com- bined glomerular and tubuloinertial nephropathy. J Am Soe Nephol 11:1439—1448 Amal H, Krane CM, Chea Q Solcimani M 2000 Ealy polyuria and urinary concentrating defect in potassium deprivation. Am J Physiol Renal Physiol 279F658-Fo63 Sands JM, lores FX, Kato A, Baum MA, Brown EM, Waed DT, Hebert SC, Haris HW 1998 Vasopressin aliited water and urea permeabilities ate altered in IMCD in hypercalcemic rats. Atm J Physiol 274:F975-F98S Fujiwara TM, Bichet DG 2005 Molecular biology of hereditary Aiabetes insipidus. J Am Soc Nepheol 16:2836-2846 Spanakis E, Mlord E, Gragnoli C2008 AVPR2 variants and mu tations in nephrogenic diabetes insipidus: eview and missense mi tation significance. J Cell Physiol 217:605-817 Robertson GL. 1987 Dipsogenic diabetes insipidus a newly cog nized syndrome caused by a selective defect inthe osmoregulation of thes. Trans Assoe Am Physicians 100-241-249 Robertson GL. 1991 Disouder of thirst in aan, I: Ramsey D, ed “Thies physiological and psychological aspets. London: Springer Verlag, 453 Goldiman MB, Luchias DJ, Robertson GL. 1988 Mechasisas of altered water metabolism in psychotic patients with poydipsin and hnyponateemia. N Engl J Med 318:397—403 ‘decom J, VergheseC, Tracy I, Jsiassen RC, Simpson GM 1994 Polyipsia and water intoxication in psyehiati patients: a review ofthe epidemiological lterature, Biol Poyhiatry35:408—419 VerbalisJG2003 Disorders of body water homeostasis. Best Pract, Res Clin Endocrinol Metab 17:471-503 Rogees PW, Kurtman NA 1973 Renal failure, uncontrollable thirst, and hypeereninemia. Cessation ofthiest with bilateral ne- phrectomy. JAMA 225:1236-1238 NaltiaH 2002 “Drink atleast ight glasses of watera day." Realy? Isthere scientific evidence for “8 X 872 Ams Physiol Regul Inege Comp Physiol 283:R093-R 1004 Iwasaki Y,Ois0 Y, Kondo K, Takagi, Takatsuli K, Hasegawa H, shikawa K, Fujimura Y,KazetoS, Tomita A 1991 Aggravation of subclinical diabetes insipidus during pregnancy. N Engl Med 324: 522-526 Braak H, Braak E 1992 The human entorhinal cortex: normal ‘morphology and lamina-specifie pathology in various diseases Neurosci Res 1536-31 Davison JM, Shill EA, Barron WM, Robinson AG, Lindheimer MD 1989 Changes inthe metabolic dearance of vasopressin and in plasma vasopressinase throughout human pregnancy. J Chin la sese SEISTS-13I8 6 6s. 66 67. 68 6 7. 7. a 7 7. % 7. 7” ” 0, a1 45 41Clin Endocrinol Metab, October 2012, 97(10):3426~3437 Braak H,JelligerK, Beaak E, Bohl J 1992 Allocorical newofi- bvillary changes in progvessive supranuclese palsy. Acta Neueo- pathol 84478483, ‘Zhang YH, Lu J, Elmquist JK, Saper CB 2003 Specific roles of «yclooxygenase-t and cyclooxygenase-2 in lipopoysaccharie-in- Alced fever and Fos expression in eat bain. J Comp Neurol 463: ary Dies F, Rangel 8, Rivera A 1961 Differential diagnosis hetween diabetes insipidus and compulsive plydipsa. Ana Inern Med 54 ras Carter AC, Robbins J1947 The use of hypertonic sine infusions in the diferetal diagnosis of dabetes insipidus and psychogenic polyipsia. J Clin Endoctinol Metab 7:753-756 Dashe AM, Cramm RE, Crist CA, Habener JF, Solomon DH 1963 [A water deprivation test forthe differential deagnoss of polyuria JAMA 185:699-703 Miller M, Dalakos T, Moses AM, FllermanH, Sreten DH 1970 ‘Recognition of partial defects in ancidiueetic hoemone secretion, Ann Intern Med 73:721-729 Fenske W, Quiakler M, Lorenz D, Zopf K, Haagen U, Papas- sotiiow J, Peiffer AF, Fassnacht M, Stk 8, Allao B 2011 Co- peptin in the differential diagnosis ofthe pol dipsia- polyuria sya Arome—revisiting the direct and inivect water deprivation ess J Cla Endocrinol Metab 96:1506-1515 Milles J, Spruce B, Baylis PH 1983 A comparison of diagnostic ‘methods to iferentate diabetes insipidus from primary polyuria ‘review of 21 patients. Acta Endocrinol (Copenh) 104:410—816 Schrier RW, Bichet DG 2007 Diseases ofthe kidney and urnasy teact Vo. II. 8th ed. Chap 87. Philadelphia: Lippincott Wiliams, Wilkins; 2250-2269 Robertson GL. 1991 Disorders of thts in man, In: Ramsay DI, Booth DA, eds. Thiet: physiological and psychological aspects London: Springer-Verlag: 453, Robertson GL. 1985 Diagnosis of diabetes insipidus. Ine Czer- Aichow AP, Robinson Ay eds. Diabetes insipidus in man, Basel, Switzerland: Karger 176 ValtiaH, North WG, Edwards BR, Gell M 1985 Animal models of diabetes insipidus. Ins Czernichow P, Robinson AG, eds. Dia- betes insipidus in man, ase, Switzerland: Karger;105-126 Robiasoa AG, Vesbalis JG 2012 Posterior pituitary In: Mlmed, Poloasky KS, Larsen PR, Kronenberg HM eds. Williams extbook of endocrinology. 12th ed. Chap” 10. Philadelphia: Saunders Elsevier Gilman A, Goodman LS 1936 The secretion of an antidiuretic hypophyseal hormone in response tothe ace for enal water con- servation. Science #4:24-28 Gilman A, Goodman L 1937 The secretory response ofthe poste: ‘ior pituitary tothe ned for water conservation. J Physiol 90:11 naa Hare K, Hickey RC, Hare RS 1941 The renal excction ofan an tidiretic substance bythe dog. Am J Physiol 134:240-244 Hare RS, HareK, Phillips DM 1943 The ena excretion ofchloide by the normal and by the diabetes insipidus dog. Am J Physiol 140334338 Hickey RC, Hare K 1944 The renal excretion of chlorideand water in diabetes insipidus. J Cia Invest 23:768-775, Carter AC, Robbins 1947 The us of hypertonic sine infusions in the diferetal diagnosis of dabetes insipidus and psychogenic polydipsia. J Clin Endocsinol Metab 7-464 Berliner RW, Davidson DG 1957 Production of hypertonic urine in the absence of pituitary antidiuretic hormone. J Clin Iaves 36: 1416-1427 Zesbe RL, Robertson GL 1981 A comparison of plasma vasopres- fin measurements with standard indirect cet in the differentia diagnosis of polyuria. N Engl J Med 305:1539-1546 De Wardener HE, Heexheimer A 1957 The eect of high water intake on the kidney’ ability to concenteate the uine in man, Physiol 139-42-521 Clin Endocrinol Metab, October 2012, 97(10):3426-3437 84. Epstcin FH, Kleeman CR, Hendrix A 1957 The influence of bodily hydration on the renal concentrating process. J lin Invest 36:629-634 85. Hareington AR, Vala H_1968 Inpaited urinary concentration after vasopressin and its gradual correction in hypothalamie di betes insipid. J Cin lnvest 47:502-510 6, Cadoapaphornchai MA, Summer SN, Fall, Thurman JM, Knep- per MA, Schrier RW 2003 Eifect of primary polyipsia on agua porin and sodium tansporter abundance. Am J Physiol Renal Physiol 285:F965-F971 87. Robertson GL, Mabr EA, Atha, Sinha'T 1973 Development and clinical aplication ofa new method forthe radioimmunoassay of arginine vasopressin ia human plasina.JChinlavest 52:2340-2382, 88. Block LH, Fuster J, Locher RA, Sicenthaler W, Vetter W 1981 Changes in issue sensitiv to vasopressin in hereditary bypotha- lam diabetes insipidus. Kin Wochenschr 59:831-836 89, Baylis PH, Roberton GL. 1980 Plasma vasopressin response 0 hypertonic saline infusion o assess posterior ptitay function. J R Soc Med 73.255-260 90, Baylis PH, Gaskill MB, Robertson GL.1981 Vasopeessin secretion in primary polyps and cranial diabetes insipidus. Q J Med 50: 3s-358 91. Morgenthaler NG, Struck}, Alonso C Bergman A 2006 Assay for the measurement of copeptin, a stable peptide derived feom the precursor of vasopressin. Clin Chem 2112-119, 92. Czacakes JW, Kleeman CR, Koenig M 1964 Physiologic studies of antidiuretic hormone by ts direct measurement in human plasma, J lin Invest 4316251640 93. Morgenthaler NG, Stuck J, Jochbergr S, Dusee MW 2008 Co- peptin: clinical us ofa new biomarker. Trends Endocrinol Metab 19343-49 94. BalanescuS, Kopp P, Gaskill MB, Morgenthaler NG, Schindler, RutishauserJ 2011 Corzltion of plasma copeptin and vasopees” sin concentrations in hypo, iso, and hyperosmolar states. J Chia Endocrinol Metab 96:1046-1052 95. Katan M, MorgenthalerNG, Di Miller B, Cheis-Crain M2007 Anterior and posterior pituitary function testing with simultaneous insulin tolerance test and a novel copeptia assay. J Clin Endocrinol Metab 92:2640-2643, 96. Dingman JF, Beieschke K, Thorn GW 1957 Sudies of neuroby- pophyseal function in man. Am J Med 23:226-238 97. Thompson CJ, Edwards CR, Baylis PH 1991 Osmosic and non- KC, Rutishause J, Brabant GE, 98 9 100, 101 102, 103. 104, 10s. 10 108. emendojoumalsorg 3437 osmotic regulation of dist and vasopeesin secretion in patients with compulsive wate drinking, Clit Endocrinol (Oxf) 35:221— Ds Fujisawa I, AsatoR, Nishimura K, TogashiK, lth K, Nakano Y, toh H, Hashimoto N, Takeuchi], Torizuka K 1987 Anterior and posterior lobes of the pituitary gland: assessment by 1.5 T MR. "nang. J Compue Assist Tomoge 11:214-220 Moses AM, Clayton B, Hochhauser L 1992 Use of T1-weighted MR imaging to diferensiateberween primary polydipsa and cen- teal diabetes insipidus. AJNR Am J Neutoradiol 13:1273-1277 ‘Maghnie M, Genovese E, Bemascoai 8, Binda 8, Ard M1997 Persistenchigh MR signal ofthe posterior pituitary glandincentral labets insipidus. AJNR Am J Neuroradiol 18:1749-1752 Hannon MJ, Orr C, Moran C, Behan LA, Agha A, Ball SG, ‘Thompion CJ 2012 Anterior hypopituitarism is ace and auton mune disease is common in adults with idiopathic central diabetes insipidus. Clin Endocrinol (Ox 76:725-728 Brooks BS, el Gammal T, Allison JD, Hoffman WH 1989 Fee- ‘quency and variation ofthe posterior pituitary bright signalon MR. images. AJR Am J Roentgenol 153:1033-1038 ‘Maghnic M, Villa A, Arico M, Lariza D, Pezoua S, Belufi G, Genovese E, Seveti F 1992 Correlation berween magnetic reso” nace imaging of posterior pimitary and neurobypophysel func tion in chldeen with diabetes insipidus. Cin Endoceinol Metab 74:795-800 ‘Verbais JG 2011 Disorders of water balance. I: Taal MW, Cher- tow GM, Marsden PA, Skorecki K, Yu ASL, Brenner BM, eds Brenner and Rectr’s The Kidney, th ed. Chap 15. Philadephis Saunders, 552-569 Bonneville JF 1986 The pituitary stalk In Bonaeille JF, CatinF, Distemann JL, eds. Computed tomography ofthe pituitary gland New York Spriagee Verlag: 106-114 Leger J, Velasquez A, Gare C, Hassan M, C2emnichow P 1999 “Thickened pituitary stalk on magnetic resonance imaging in chile cea with central dabetes insipidus. J Clin Endocrinol Metab 84: 1954-1960 Robertson GL, Bed T 1987 Water metabolism In: Brenner BM, Rector FC, eds. The kidney. Chap 11 3 ed Philadelphia Saun- ders; 85-432 ‘Moses AM, Stseten DH 1967 Differentiation of polyurie states by measurement of responsesto changes in plasma osmolality induced by hypertonic saline infusions, Am J Med 42:368-377,