435 Water disturbances at different stages of primary thyroid failure

M Sahún, C Villabona, P Rosel1, M A Navarro1, J M Ramón2, J M Gómez and J Soler Department of Endocrinology, 1Hormone Unit and 2Department of Preventive Medicine, Hospital Princeps d’Espanya, Ciutat Sanitària i Universitària de Bellvitge, Barcelona, Spain (Requests for offprints should be addressed to M Sahún, Bonsoms 29–33 4  1a esc A, 08028 Barcelona, Spain)

Abstract The aim of the present study was to study salt and water the AVP and thirst thresholds did not differ between the metabolism in thyroid deficiency. groups, the lag between them was lower in the PRE We performed an oral water loading test (OWL) and a (4·13·2 mOsm/kg) and SUB group (2·62·1 mOsm/ hypertonic 5% saline infusion test (HSI) in 16 patients with kg) than in the CG group (13·39·2 mOsm/kg) overt primary hypothyroidism before replacement treat- (P<0·05). There were no differences in atrial natriuretic ment (PRE group) and after, in eight patients with hormone (ANH), plasma renin activity (PRA) and plasma subclinical hypothyroidism (SUB group) and in 16 normal aldosterone among the groups. individuals (CG group). These results indicate that plasma hypo-osmolality and In the PRE group, a lower free water clearance was low levels of AVP are present in primary hypothyroidism, detected in the OWL (P<0·022), with lower plasma and indeed are already present in the subclinical phase of osmolality (OWL: P<0·005; HSI: P<0·001) and arginine the disease. An overlap between the thresholds of thirst (AVP) (OWL: P<0·001; HSI: P<0·001) than and AVP seem to play a role in these abnormalities, but the CG group, across both tests; they normalized with the ANH, PRA and plasma aldosterone do not appear to replacement treatment. The same plasma abnormalities contribute. were detected in the SUB group with the HSI. Although Journal of Endocrinology (2001) 168, 435–445

Introduction 1990, Barna et al. 1994, Ota et al. 1994). This decrease could influence the sodium and water regulation of Disturbances in salt and water metabolism, mainly im- patients with hypothyroidism, although its exact contri- paired water excretion, have been described in primary bution remains undefined. hypothyroidism. Severe can appear, es- In patients with subclinical hypothyroidism, disturb- pecially in the presence of myxedema coma (Moses & ances in various organs have been described, even in the Scheinman 1996). However, controversy persists concern- absence of thyroid hormones abnormalities (Surks & ing the physiopathological causes and the involvement Ocampo 1996). However, in-depth studies on their water of hormonal abnormalities, hemodynamic changes, or metabolism are lacking. alterations of the itself. The aim of this study was to assess water and salt One hormonal mechanism that has been suggested as a metabolism and its regulation in patients with overt possible explanation of these abnormalities is an altered primary and subclinical hypothyroidism, as well as the of arginine vasopressin (AVP) secretion. changes observed after treatment with thyroid hormones. Levels of this hormone have been reported to be high in some hypothyroid patients, mimicking the state of Patients, Materials and Methods inappropriate antidiuretic hormone secretion syndrome (SIADH) (Skowsky & Kikuchi 1978, Laczi et al. 1987). Patients However, more recently, plasma AVP has been found to be normal or low in hypothyroid patients (Iwasaki et al. The study was approved by the ethical committee of the 1990, Ota et al. 1994). In addition, low plasma levels of hospital and informed consent was obtained from all atrial natriuretic hormone (ANH) have been found in patients before the beginning of the study. The clinical some studies (Kohno et al. 1987, Zimmerman et al. 1987, characteristics of the individuals are detailed in Table 1. Woolf & Moult 1988, Vesely et al. 1989, Widecka et al. The studies were carried out in the following groups.

Journal of Endocrinology (2001) 168, 435–445 Online version via http://www.endocrinology.org 0022–0795/01/0168–435  2001 Society for Endocrinology Printed in Great Britain

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access 436 M SAHUuN and others · Water metabolism at thyroid failure

Table 1 Characteristics of the patients that were studied. Data are meansS.D.

PRE POST SUB CG (n=16) (n=16) (n=8) (n=16)

Male/female 4/12 — 0/8 4/12 Age (years) 43·714·3 — 38·48·1 41·916·4 (range 17–72) (range 29–54) (range 18–78) BMI (kg/m2)28·75·4 — 25·02·829·88·0 TSH (mU/l) 129·090·63·85·313·53·51·70·7     T4 (nmol/l) 24·6 16·9 100·1 22·587·1 15·899·6 12·1     FT4I (nmol/l) 21·5 14·895·0 24·272·1 14·092·6 14·6     T3 (nmol/l) 0·84 0·51 1·84 0·31·89 0·33 2·08 0·4

The clinical data of those in the POST group were the same as those in the PRE group.

PRE group These were patients with overt primary were no significant differences between the groups either hypothyroidism without treatment. Individuals with in age or in body mass index. serum thyrotropin (TSH) levels >10 mU/l, low serum Plasma volume (PV), effective renal plasma flow thyroxine (T4) levels and with clinical symptoms were (ERPF) and glomerular filtration rate (GFR) were studied selected. All the patients presented spontaneous non- in all patients. Compared with the POST group, all three iatrogenic hypothyroidism. They had not been treated parameters were lower in the PRE group but not in the with thyroid hormones before the study. The etiologic SUB group (data published in Villabona et al. 1999). diagnosis was idiopathic thyroid atrophy in eleven and Hashimoto’s thyroiditis in five patients. Methods SUB group These were patients with subclinical On two different non-consecutive days, an oral water hypothyroidism. Individuals with serum TSH >10 mU/l, loading test and a hypertonic saline infusion test were normal serum T4 and tri-iodothyronine (T3) levels and performed in all subjects. The trials were conducted at without recognized symptoms of hypothyroidism were outpatient centers, in the morning after an overnight fast included. In all cases in this group the disease was and a ban on smoking and alcohol intake of at least 10 h, spontaneous. No patients in this group were taking thyroid with a free salt and water diet on the days before the start hormones. Etiology was Hashimoto’s thyroiditis in four, of all studies. There were no tolerance problems in either idiopathic thyroid atrophy in three and postpartum test. thyroiditis in one patient. Oral water loading test The patient was maintained in CG group These were normal individuals who were age a supine position for at least 30 min before time 0, with the (5 years)- and sex-matched healthy volunteers recruited head at approximately 30, throughout the test, standing for each hypothyroid patient. upright only for micturition. A single arm was cannulated at 15 min, and the oral administration of 20 ml/kg tap POST group These were patients with primary hy- water began at time 0 and continued for 30 min. Blood pothyroidism, after treatment. The PRE group individuals samples were obtained at 30-min intervals for 240 min, were studied again after replacement treatment, when and samples were obtained hourly. Weight, blood serum TSH levels had returned to normal. The time- pressure (BP) and heart rate were determined at the interval between the first and second study was beginning and end of the test; in plasma, Na, K, glucose, 14·34·0 months. The mean dose of levothyroxine AVP and osmolality were measured at all times, ANH administered at that moment was 2·20·5 µg/kg weight. hourly and renin activity (PRA) and aldosterone at 0 and Individuals with congestive heart failure or arrhythmia, 240 min. In the urine, Na, K, urea and osmolality were severe arterial hypertension (systolic>180 and diastolic determined every hour. >100 mmHg), renal failure, diabetes mellitus, diabetes insipidus, SIADH, adrenal insufficiency and hepatic dis- Hypertonic saline infusion test This test was per- ease were excluded from the study. Patients who at the formed under the same conditions as the oral water loading time of the study were taking thyroid hormones, diuretics, test. The patient was maintained in a supine position with hypotensors, lithium, demeclocycline, glucocorticoids the head at approximately 30 throughout the infusion or any other drug that could affect water and sodium period. Both arms were cannulated at 15 min, and the homeostasis were also excluded. In menstruating women, 5% hypertonic saline infusion (3 ml/kg per h) using an studies were preferably made in the follicular phase. There infusion pump began at time 0 and continued for 120 min.

Journal of Endocrinology (2001) 168, 435–445 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access Water metabolism at thyroid failure · M SAHUuN and others 437

Blood samples were extracted in the contralateral arm at and with one-way ANOVA when more than two groups 30-min intervals and urine samples were collected hourly. were studied, with Scheffé’s and Bonferroni’s a posteriori Weight was determined at the beginning and end of the contrasts to determine the differences group by group. In test and BP and heart rate each 30 min. In plasma, Na, K, the variables with non-normal distribution, the Kruskal– glucose, AVP and osmolality were measured at all times, Wallis test was applied. In the variables in which evolution ANH every hour, and renin activity and aldosterone at at different points of the tests was of interest, an ANOVA 0 and 120 min. In the urine, Na, K, urea and urinary of repeated measures was applied, directly if the distribu- osmolarity (UOsm) were measured hourly. Thirst was tion was normal, or after neperian logarithm or squared evaluated every 30 min by asking how many glasses of root transformation if it was not. In variables in which a water the patient would drink at that moment, showing a significant difference was found, each point was studied by diagram with drawings of several glasses (maximum eight a one-way ANOVA test. The differences between groups glasses) (Goldman et al. 1988). The patients were allowed at each point were studied with Scheffé and Bonferroni to indicate fractions (e.g. a glass and a half). tests. The comparisons of continuous quantitative variables were performed by correlation test, with Pearson’sor Analytical measurements Plasma and urine sodium, Spearman’scoefficient as appropriate. In the case of AVP potassium, glucose, urea and were determined and thirst, correlation with plasma osmolality was deter- with a Hitachi 911 Automatic Autoanalyzer (Boehringer mined in each person, using the individual thresholds and Mannheim Immunodiagnostics, Mannheim, Germany). slopes to determine the mean and standard deviation of the Plasma and urinary osmolalities were determined by the group. For the correlation test with thyroid hormones, the freezing point depression method (Advanced Cryomatic base values of plasma osmolality, plasma AVP, plasma Osmometer 3C2; Advanced Instruments, Needham ANH, PRA and plasma aldosterone were calculated as the Heights, MA, USA). These determinations were used in pool of the water loading and saline infusion test values. the calculation of the free water clearance using standard The area under the curve (AUC) was calculated with the formula. Plasma effective osmolality was also calculated by trapezoidal method. The increase of one variable was the formula: 2(plasma Na+plasma K)+plasma glucose calculated as the maximumbase/base, and the decrease  ff (all in mmol/l). The TSH, as well as T4,T3 and as base minimum/base. For all the tests, a di erence of thyroxine-binding capacity (TBK) were determined with P<0·05 was considered significant. All results are ex- autoanalyzer ES-700 using an enzyme-linked immuno- pressed in the text as the mean.. unless otherwise sorbent assay method (Boehringer Mannheim Immuno- indicated. For the statistical analysis, undetectable values of diagnostics). With the T4 and TBK values, the free AVP were considered at 0·1 pmol/l. thyroxine index (FT4I) was calculated. PRA was deter- mined using competitive radioimmunoassay (RIA) (Incstar, Stillwater, MN, USA). Plasma aldosterone was Results determined by competitive RIA (Sorin, Saluggia, Italy). For determination of AVP, the samples were deposited Oral water loading test in chilled heparinized tubes and centrifuged at 4 C, The percentage of water excreted at the end of the test, 2500 g for 15 min. The supernatant was maintained in ice compared with that ingested at the beginning, did until extraction, which was completed within 6 h. The not differ significantly between groups (PRE: 108·01 extraction was done with ethanol 98%. The determination 26·8%; SUB: 139·3733·4%; CG: 122·6924·4%; was performed using RIA (Bühlmann, Schönenbuch, POST: 118·5527·0%). None of the individuals pre- Switzerland). The between-assay coefficient of variation of sented an excretion at the end of the test of less than 80% the method was 14·1%. The minimal detectable levels of the ingested volume. were 0·15 pmol/l. For ANH, the extracted blood was The response of the free water clearance (CH2O)tothe deposited in a chilled tube containing trasylol (400 µl) and water loading differed between the four groups (P<0·022, EDTA (15 mg). The sample was placed in ice until ANOVA) (Fig. 1A). The CH2O at 120 min was signifi- its centrifugation at 4 C, 2500 g for 15 min. The resulting cantly lower in the PRE (329·44201·77 ml/h) than in supernatant was frozen to c20 C until analysis was the POST group (510·24165·12 ml/h) (P<0·017, performed. The sample was extracted in C18 columns. paired t-test). Conversely, at 240 min it was significantly The determination was performed by RIA (Nichols Insti- lower in the POST group (6·6367·11 ml/h) than in tute, San Juan Capistrano, CA, USA). The between-assay the PRE group (96·28111·37 ml/h) (P<0·017, paired coefficient of variation of the method was 16·6%, and the t-test). The AUC of CH2O throughout the 240 min of the minimal detectable levels were 15 pg/ml. test showed significant intergroup differences. In the PRE group (39·411·7 liters/h per min) and the POST group Statistical analysis The means in the variables with (46·911·3 liters/h per min) the AUC was significantly normal distribution were compared by a Student’s t-test of lower than in the CG group (64·411·0 liters/h per min) paired data, in the case of the PRE and the POST groups, (P<0·05, one-way ANOVA). The AUC of the SUB www.endocrinology.org Journal of Endocrinology (2001) 168, 435–445

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access 438 M SAHUuN and others · Water metabolism at thyroid failure

   Figure 1 (A) CH2O and (B) UOsm throughout the oral water loading test. PRE ( ), SUB ( ), CG ( ) and POST ( ). Values are meansS.E.M. *P<0·05 compared with CG group.

group (44·916·3 liters/h per min) was also lower than significant (P<0·05) at most points. The plasma osmolality that in the CG group, but without reaching significance. in the SUB and POST groups showed levels between The base UOsm and its decrease was similar in all groups those of the PRE and the CG groups, and presented no (Fig. 1B). significant differences, except between the POST and the There were significant differences in the plasma osmo- CG groups at 60 min. lality response to the oral water loading in the four groups The evolution of plasma AVP showed significant dif- (P<0·005, ANOVA) (Fig. 2A). Beforehand and through- ferences between the four groups (P<0·001, ANOVA) out the test the PRE group showed a lower plasma (Fig. 2B). All levels in the PRE group were lower than osmolality than the CG group, the difference being those of the CG group, the difference being significant at

Journal of Endocrinology (2001) 168, 435–445 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access Water metabolism at thyroid failure · M SAHUuN and others 439

Figure 2 (A) Plasma levels of osmolality (POsm), (B) AVP and (C) ANH throughout the oral water loading test. PRE (), SUB ( ), CG () and POST (). Values are meansS.E.M. *P<0·05 compared with CG group; §P<0·05 compared with POST group.

www.endocrinology.org Journal of Endocrinology (2001) 168, 435–445

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access 440 M SAHUuN and others · Water metabolism at thyroid failure

Table 2 PRA and aldosterone in water and salt loading tests. Values are meansS.D.

PRE SUB CG POST (n=16) (n=8) (n=16) (n=16) Oral water loading Base PRA (kat/l per h) 31·4526·27 27·2220·83 33·2153·09 25·8621·37 Final PRA (kat/l per h) 26·425·72 20·9317·94 27·9943·99 19·321·19 PRA decrement (%) 12·745·430·9321·419·7433·334·8625·5 Base plasma aldosterone (nmol/l) 0·330·11 0·430·43 0·60·60·300·17 Final plasma aldosterone (nmol/l) 0·160·09 0·150·12 0·170·17 0·190·14 Plasma aldosterone decrement (%) 52·6915·058·537·156·8731·637·3119·9 Hypertonic saline infusion Base PRA (kat/l) 16·5513·928·3717·725·0512·620·219·2 Final PRA (kat/l) 5·913·613·485·510·119·67·256·7 PRA decrement (%) 40·840·957·618·453·031·964·222·9 Base plasma aldosterone (nmol/l) 0·450·50 0·480·30 0·470·33 0·340·17 Final plasma aldosterone (nmol/l) 0·120·20 0·160·20 0·130·09 0·180·17 Plasma aldosterone decrement (%) 76·129·067·419·066·411·351·124·4

Table 3 Urinary parameters during hypertonic saline infusion test. Values are meansS.D.

PRE SUB CG POST (n=16) (n=8) (n=16) (n=16)

Urinary volume (ml) 350·0192·3 445·0219·9 284·7111·9 289·2187·7 Basal urinary osmolality (mOsm/kg) 713·2254·5 747·0392·8 850·1208·2 824·0117·3 Final urinary osmolality (mOsm/kg) 662·3120·5 652·1241·1 760·8269·5 613·0200·0 Urinary osmolality increase (%) 8·6466·51 47·24179·78 3·5640·04 23·5324·99 Total sodium excretion (mmol) 85·550·276·653·160·422·862·844·1 Osmolar clearance (ml/h) 348·4167·3 442·9219·5 340·0132·6 328·2172·2 Free water clearance (ml/h) 190·097·1 220·4132·5 199·1108·7 168·284·4

most points (P<0·05). In the POST group, plasma AVP plasma osmolality was significantly lower in the PRE was higher, beforehand and throughout the test, than in group than in the CG group throughout the test (P<0·05). the other three groups; the differences between the PRE The POST group maintained levels of plasma osmolality and the POST groups in this variable were significant at similar to those of the CG and higher than the PRE each point (P<0·05). In the SUB group, plasma AVP groups, significantly so in the base times and at 30 and levels were similar to those of the CG group although 60 min (P<0·05). In the SUB group, evolution of the slightly lower, without significant differences either with plasma osmolality was similar to that of the PRE group, respect to this group or to the PRE group. Plasma ANH and significantly lower than the CG group at most points during the oral water loading test was similar in the four (P<0·05). The peaks of plasma osmolality also differed groups (Fig. 2C). significantly in the four groups (P<0·002, one-way In this test, the PRA and plasma aldosterone was found ANOVA); it was lower in the PRE (305·125·0 mOsm/ to be similar in the four groups, without significant kg) and SUB groups (306·464·66 mOsm/kg), than in differences in the base, end or decline (Table 2). the CG group (313·466·69 mOsm/kg) (P<0·05). There were no significant differences between the POST group (310·375·61 mOsm/kg) and the others. The base Hypertonic saline infusion plasma osmolality showed a positive correlation with T4 < < With the hypertonic saline infusion there were no signifi- (r=0·49, P 0·001), FT4I(r=0·54, P 0·001) and T3 cant differences between the four groups in the response of (r=0·45, P<0·001). The peak of plasma osmolality showed < the urinary volume, base UOsm, final UOsm, UOsm a positive correlation with T4 (r=0·49, P 0·001), FT4I < < increase, total urinary sodium excretion, osmolar clearance (r=0·52, P 0·001) and T3 (r=0·42, P 0·003). The ff or in CH2O (Table 3). plasma osmolality increase showed no significant di er- Hypertonic saline infusion produced the expected ences in the four groups (PRE: 5·11·5; SUB: 5·81·2; increase in plasma osmolality in the four groups, although CG: 5·91·3; POST: 5·21·3%). there were significant differences between them The saline infusion stimulated plasma AVP in all (P<0·001, ANOVA) (Fig. 3A). As with the water loading, individuals, but there were significant differences between

Journal of Endocrinology (2001) 168, 435–445 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access Water metabolism at thyroid failure · M SAHUuN and others 441

Figure 3 (A) Plasma levels of osmolality (POsm), (B) AVP and (C) ANH throughout the hypertonic saline infusion. PRE (), SUB ( ), CG () and POST (). Values are meansS.E.M. *P<0·05 compared with CG group; §P<0·05 compared with POST group.

www.endocrinology.org Journal of Endocrinology (2001) 168, 435–445

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access 442 M SAHUuN and others · Water metabolism at thyroid failure

Table 4 AVP and thirst correlation with plasma osmolality (POsm). The osmolar threshold and sensitivity of each group is shown as the meanS.D. of the threshold and sensitivity of the individuals of each group. Threshold difference is the difference between thirst and AVP threshold

AVP Thirst Threshold Male/ Age Base POsm Base AVP Threshold Threshold difference female (years) (mOsm/kg) (pmol/l) (mOsm/kg) Sensitivity (mOsm/kg) Sensitivity (mOsm/kg) Group PRE (n=6) 0/6 40·67 286·93 0·37 286·29 0·115 290·42 0·125 4·13 16·74 2·70 0·27 5·77 0·045 5·22 0·051 3·2 SUB (n=5) 0/5 40·80 288·92 0·64 287·24 0·159 289·85 0·195 2·60 10·32 5·19 0·29 5·12 0·060 4·45 0·045 2·08* GC (n=8) 2/6 32·38 293·29 1·31 279·42 0·130 292·73 0·258 13·31 16·15 6·18 0·52 11·02 0·079 4·49 0·108 9·24 POST (n=9) 1/8 45·89 295·85 1·44 284·72 0·149 295·17 0·155 10·45 15·96 4·32 0·48 7·38 0·060 5·81 0·100 6·52

*P<0·05 compared with CG group.

the four groups (P<0·001, ANOVA) (Fig. 3B). The PRE A significant difference was found between the PRE group showed significantly lower plasma AVP levels than group and the CG group (P<0·017, t-test). the CG group throughout the test. In the POST group, Comparison of the thirst and AVP thresholds showed plasma AVP levels were similar to those of the CG and that in all four groups the secretion of plasma AVP began significantly higher than in the PRE group (P<0·05). at lower plasma osmolality levels than the thirst sensation. Plasma AVP levels in the SUB group were similar to those The separation between the thresholds (Table 4) showed of the PRE group, and lower than the CG group, differences between the four groups (P<0·0186, one-way significantly so at several points (P<0·05). Significant ANOVA). The only significant difference was between differences were also detected in the peak of plasma AVP the SUB group and the CG group. When the PRE and reached, which was lower in the PRE (2·030·93 pmol/ CG groups were studied by t-test, the difference between l) than in the CG group (3·771·46 pmol/l) (P<0·014, thresholds was also significantly lower in the PRE group one-way ANOVA). The peak of plasma AVP in the SUB (P<0·04). group (3·311·82 pmol/l) and in the POST group (3·591·17 pmol/l) did not show significant differences with regard to the others. Base plasma AVP showed a Discussion < positive correlation with T4 (r=0·62, P 0·001), FT4I < < ff (r=0·49, P 0·001) and T3 (r=0·55, P 0·001). The peak In our study, there were no significant di erences in the of plasma AVP showed a positive correlation with T4 four groups in terms of percentage of excreted water, and < < (r=0·50, P 0·001), FT4I(r=0·47, P 0·001) and T3 none of the patients fulfilled the criteria for inappropriate (r=0·38, P<0·02). secretion of ADH after administration of oral water load- The saline infusion stimulated plasma ANH similarly in ing. However, CH2O presented alterations in the PRE and the four groups (Fig. 3C). There were no significant SUB groups: quantitative (lower AUC) and qualitative differences between the four groups in the base, final and (slower increase). This behavior was not due to an decrease values of PRA and plasma aldosterone (Table 2). impairment in gastrointestinal water absorption, since Base plasma aldosterone showed a significant correlation plasma osmolality was low at the beginning of the test, and < < with T4 (r=0·44, P 0·001) and T3 (r=0·49, P 0·001). PRA and aldosterone were inhibited. These findings In the individuals whose thirst was also tested, a coincided with a lower plasma osmolality in these patients, regression study of plasma osmolality and plasma AVP which tended to normalize after replacement treatment. showed that the osmotic threshold and the sensitivity of There was no complete loss of osmoregulation of AVP, plasma AVP did not differ significantly between the since it was stimulated with the saline infusion and groups (Table 4). suppressed with the water loading. However, throughout In all patients, the saline infusion stimulated the sen- both tests, plasma AVP was always significantly lower in sation of thirst, with a similar threshold in the four groups the hypothyroid patients than in the control group. Base (Table 4). The sensitivity to thirst was different in each of and stimulated plasma ANH concentrations showed no the four groups (P<0·041, one-way ANOVA), although differences between the groups. In addition, our results no group was found to be clearly different from the others. suggested the integrity of the PRA–aldosterone axis.

Journal of Endocrinology (2001) 168, 435–445 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access Water metabolism at thyroid failure · M SAHUuN and others 443

Most studies detect a decrease in the elimination of a but not in others (Weissel et al. 1986, Ladenson et al. water loading in hypothyroid patients, (DeRubertis et al. 1987). One of the possible explanations for these discrep- 1971, Skowsky & Kikuchi 1978, Laczi et al. 1987, Iwasaki ancies may lie in the comparability of the groups, since in et al. 1990, Ota et al. 1994) and a lower plasma osmolality most studies the mean age of the hypothyroid patients is than euthyroid individuals (Laczi et al. 1987, Iwasaki et al. higher than that of the control group. In our study the age 1990, Barna et al. 1994, Ota et al. 1994). of the hypothyroid patients and that of the controls was The studies of plasma AVP in hypothyroid patients totally comparable. The principal cause of the discrepan- show a wide variety of results. Some authors have found cies may be differences in the severity of the hypothy- high plasma AVP and postulate the role of this hormone in roidism, with significant differences only in the patients the development of hyponatremia of hypothyroidism with serious deficit, as suggested by Kohno et al. (1987). (Skowsky & Kikuchi 1978, Laczi et al. 1987). Others have The duration of the hypothyroidism could also influence found that plasma AVP is normal or even suppressed the plasma ANH levels of these patients, although (Iwasaki et al. 1990, Vargas et al. 1991, Arnaout et al. 1992, Widecka et al. (1990) found differences after only a short Ota et al. 1994). Probably the discrepancies are due to the period of levothyroxine withdrawal. In our study, as in all different characteristics of the individuals studied, princi- those in which the hypothyroid patients have been diag- pally the duration and severity of the hypothyroidism, nosed recently and have not received treatment, the real which may cause an increase in plasma AVP by raising duration of the disease is unknown. Given the variability non-osmotic stimuli. These stimuli may be due to a fall in of the results in the different studies, as well as the overlap cardiac output and/or the blood volume, caused by the of the plasma levels in the groups, it seems unlikely that disease itself or by additional alterations (Skowsky & ANH plays an important role in water disorders in Kikuchi 1978). Our series was homogeneous in that we hypothyroidism. included only patients with recently diagnosed and un- We detected no abnormalities in the urinary excretion treated hypothyroidism, of mild or moderate severity, and parameters in the patients with subclinical hypothy- in whom all drugs and clinical entities that might have roidism, either in the oral water loading or in the saline influenced salt and water metabolism were ruled out. infusion tests. However, we found lower plasma osmolality Our results are in accordance with those of the authors and AVP than in the control group, similar to that seen in who report plasma AVP to be normal or suppressed. The patients with clinical primary hypothyroidism, especially possibility that the alterations in water excretion and in the saline infusion test, in which the levels of plasma decreased plasma osmolality found in hypothyroid patients osmolality and those of plasma AVP of the SUB group are due to an inappropriately high secretion of AVP seems were very similar to those of PRE group, and significantly improbable. Other possible explanations could be an lower than the CG group. No abnormalities were found in increase in the sensitivity to AVP, although some exper- plasma ANH, PRA or plasma aldosterone. The PV, ERPF imental studies in rats suggest, in fact, that there is a and GFR in this group were normal (Villabona et al. resistance to its action (Seif et al. 1979, Kim et al. 1987). 1999). To our knowledge, no other studies have been Another possibility is the action of other AVP- undertaken specifically to assess this aspect in subclinical independent mechanisms at the level of the renal tubule, hypothyroidism. In the study of Cooper et al. (1984) in 17 such as the decrease of urine flow at the distal tubule women with subclinical hypothyroidism, although the (DeRubertis et al. 1971, Emmanouel et al. 1974, Laczi levels of water excretion after water loading were relatively et al. 1987) or both morphological (Davis et al. 1983) or low, there was no variation pre- and post-normalization of functional (Michael et al. 1976, Garg & Tisher 1985, Kim TSH with levothyroxine. Our results suggest that, in spite et al. 1987) abnormalities of the distal tubule in hypo- of the theoretical normality of the thyroid hormones, there thyroidism. It has been suggested that in some hypo- are abnormalities in plasma sodium and AVP in subclinical natremia cases with characteristics of inappropriate hypothyroidism, similar to those detected in patients with secretion of ADH and with decreased AVP, there may be primary hypothyroidism, as described in other systems an increase in a non-filiated substance with an action (Surks & Ocampo 1996). The real clinical implication similar to AVP (Robertson et al. 1982, Kern et al. 1986). remains unclear, and more accurate studies in these The levels of AVP are lower in hypothyroid patients in subjects are required, analysing their basal situation as well spite of a lower plasma volume. as their response to the replacement treatment. In patients with hypothyroidism, plasma ANH concen- As regards thirst, in our study no differences were trations have been reported to be low (Kohno et al. 1987, detected in the thirst thresholds of the groups (Table 4). Zimmerman et al. 1987, Woolf & Moult 1988, Widecka The PRE slope indicated a lower sensitivity than the CG. et al. 1990, Barna et al. 1994, Ota et al. 1994) or normal One interesting point is the difference between the (Ladenson et al. 1987, Rolandi et al. 1992). After thyroid thresholds of AVP secretion and of thirst. In our control replacement these concentrations have increased in some group, the threshold of plasma AVP (279·411·0 reports (Kohno et al. 1987, Zimmerman et al. 1987, Woolf mOsm/kg) was lower than the thirst threshold & Moult 1988, Rolandi et al. 1992, Bernstein et al. 1997) (292·74·5 mOsm/kg), with an average difference of www.endocrinology.org Journal of Endocrinology (2001) 168, 435–445

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access 444 M SAHUuN and others · Water metabolism at thyroid failure

13·39·2 mOsm/kg. This coincides with some authors, Barna I, Földes J, Tóth M, Büki B, Lang RE & de Châtel R 1994 like Robertson et al. (1982), who also find a thirst Atrial natriuretic peptide (ANP) responsiveness in patients with 50 threshold higher than the AVP threshold and with a hypothyroidism. Acta Medica Hungarica 23–32. ff Bernstein R, Midtbo K, Urdal P, Morkrid L, Smith G, Muller C, similar di erence. However, other authors (Thompson BjoroT&HaugE1997Serum N-terminal pro-atrial factor 1–98 et al. 1986) report that the thresholds are similar. The before and during thyroxine replacement therapy in severe difference between the osmotic thresholds of AVP and hypothyroidism. Thyroid 7 415–419. thirst varied significantly in all four groups. It was sig- Cooper DS, Halpern R, Wood LC, Levin AA & Ridgway EC  nificantly lower in the PRE and the SUB groups than in 1984 -thyroxine therapy in subclinical hypothyroidism. A double blind, placebo-controlled trial. Annals of International Medicine 101 the CG. Thus, the thresholds seem to be closer than 18–24. normal in the PRE group. This smaller difference be- Davis RG, Madsen KM, Fregly MJ & Tisher CC 1983 Kidney tween thresholds already appears in subclinical hypothy- structure in hypothyroidism. American Journal of Pathology 113 roidism and disappears after replacement treatment. Only 41–49. a few studies have evaluated thirst in hypothyroidism. In a DeRubertis FR, Michelis MF, Bloom ME, Mintz DH, Field JB & Davis BB 1971 Impaired water excretion in myxedema. American review, Fitzsimons (1972) describes the increase in water Journal of Medicine 51 41–53. intake in rats with antithyroid treatment, although the Emmanouel D, Lindheimer MD & Katz AI 1974 Mechanism of author doubts whether it can be considered a direct effect impaired water excretion in the hypothyroid rat. Journal of Clinical of the thirst mechanism or an effect of the increase in Investigation 54 926–934. natriuresis. The presence in patients with primary and Fitzsimons JT 1972 Thirst. Physiological Reviews 52 468–561. subclinical hypothyroidism of a thirst threshold that prac- Garg LC & Tisher CC 1985 Effects of thyroid hormone on Na-K-adenosine triphosphatase activity along the rat nephron. tically overlaps with the AVP threshold could explain, in Journal of Laboratory and Clinical Medicine 106 568–572. part, the decrease in plasma AVP, inhibited by the Goldman MB, Luchins DJ & Robertson GL 1988 Mechanism of beginning of the water intake at lower osmolalities. This altered water metabolism in psychotic patients with polydipsia relatively premature water intake could also contribute and hyponatremia. New England Journal of Medicine 318 to the decrease of the plasma osmolality, suggesting a 397–403. pathological thirst. Iwasaki Y, Oiso Y, Yamauchi K, Takatsuki K, Kondo K, Hasegawa H & Tomita A 1990 Osmoregulation of plasma vasopressin in In summary, primary thyroid hypofunction is ac- myxedema. Journal of Clinical Endocrinology and Metabolism 70 companied by impairment in water excretion and a 534–539. tendency towards plasma hypo-osmolality, even in early Kern PA, Robbins RJ, Bichet D, BerlT&Verbalis JG 1986 phases such as subclinical hypothyroidism. These distur- Syndrome of inappropriate antidiuresis in the absence of arginine bances could be due to a direct action of the decrease of vasopressin. Journal of Clinical Endocrinology and Metabolism 62 148–152. the thyroid hormones on the renal dilution mechanisms, Kim JK, Summer SN & Schrier RW 1987 Cellular action of arginine since hemodynamic factors do not seem to be determi- vasopressin in the isolated renal tubules of hypothyroid rats. nant, nor is there an increase in plasma AVP. The American Journal of Physiololgy 253 F104–F110. narrowing of the gap between plasma AVP and thirst Kohno M, Murakawa K, Yasunari K, Nishizawa Y, Morii H & osmotic thresholds could contribute to these disturbances. Takeda T 1987 Circulating atrial natriuretic peptides in hyperthyroidism and hypothyroidism. American Journal of Medicine 83 With this trend towards plasma hypo-osmolality, it is 648–652. conceivable that hyponatremia may appear, due to exces- Laczi F, Janáky T, IványiT,JuleszJ&László FA 1987 sive water administration, or to an increase in plasma AVP Osmoregulation of arginine-8-vasopressin secretion in primary secondary to a substantial decrease in the effective volemia hypothyroidism and in Addison’s disease. Acta Endocrinologica 114 or of the cardiac output. Plasma ANH and the renin– 389–395. aldosterone axis do not appear to play a determinant role Ladenson PW, LangevinH&Michener M 1987 Plasma atriopeptin concentrations in hyperthyroidism, euthyroidism, and in the water and sodium disturbances in this disease. hypothyroidism: studies in man and rat. Journal of Clinical Endocrinology and Metabolism 65 1172–1176. Michael UF, Kelley J, AlpertH&VaamondeC1976Roleofdistal Acknowledgements delivery of filtrate in impaired renal dilution of the hypothyroid rat. American Journal of Physiology 230 699–705. This work was supported by a FIS grant 91–1123. We Moses AM & Scheinman SJ 1996 The kidneys and thank Dr Pere Ginés for his critical suggestions and Mr metabolism in hypothyroidism. In Werner and Ingbar’s The Thyroid, Michael Maudsley for his help in the preparation of the edn 7, pp 812–815. Eds LE Braverman & RD Utiger. Philadelphia: Lippincott-Raven. manuscript. Ota K, Kimura T, Sakurada T, Shoji M, Inoue M, Sato K, Ohta M, Yamamoto T, Funyu T, Yoshida K et al. 1994 Effects of an acute water load on plasma ANP and AVP, and renal water handling in References hypothyroidism: comparison of before and after -thyroxine treatment. Endocrine Journal 41 99–105. Arnaout MA, Awidi AS, El-Najdawi AM, Khateeb MS & Ajlouni Robertson GL, AycinenaP&Zerbe RL 1982 Neurogenic KM 1992 Arginine-vasopressin and endothelium-associated proteins disorders of osmoregulation. American Journal of Medicine 72 in thyroid disease. Acta Endocrinologica 126 399–403. 339–353.

Journal of Endocrinology (2001) 168, 435–445 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access Water metabolism at thyroid failure · M SAHUuN and others 445

Rolandi E, Santaniello B, Bagnasco M, Cataldi A, Garibaldi C, Villabona C, Sahun M, Roca M, Mora J, Gomez N, Gomez JM, FranceschiniR&Barreca T 1992 Thyroid hormones and atrial PuchalR&SolerJ1999 Blood volumes and renal function in overt natriuretic hormone secretion: study in hyper- and hypothyroid and subclinical primary hypothyroidism. American Journal of Medical patients. Acta Endocrinologica 127 23–26. Science 318 277–280. Seif SM, Robinson AG, Zenser TV, Davis BB, Huellmantel AB & Weissel M, Punzengruber C, Hartter, LudvikB&Woloszczuk W Haluszczak C 1979 Neurohypophyseal peptides in hypothyroid rats: 1986 Thyroid hormones and pericardial effusion may influence plasma levels and kidney response. Metabolism 28 137–143. plasma levels of atrial natriuretic peptide (ANP) in humans. Skowsky WR & Kikuchi TA 1978 The role of vasopressin in the Klinische Wochenschrift 64 93–96. impaired water excretion of myxedema. American Journal of Medicine Widecka K, Gozdzik J, Dutkiewicz T, MamosE&Czekalski S 1990 64 613–621. Low plasma concentrations of atrial natriuretic peptide in untreated Surks MI & Ocampo E 1996 Subclinical thyroid disease. American hypothyroid patients. Journal of International Medicine 228 39–42. Journal of Medicine 100 217–223. Woolf AS & Moult PJA 1988 Plasma concentrations of atrial ThompsonCJ,BlandJ,BurdJ&Baylis PH 1986 The osmotic natriuretic peptide in hypothyroidism. British Medical Journal 296 threshold for thirst and vasopressin release are similar in healthy 531. man. Clinical Science 71 651–656. Zimmerman RS, Gharib H, Zimmerman D, HeubleinD&Burnett Vargas F, Baz MJ, Luna JD, Andrade J, JodarE&HaroJM1991 JC 1987 Atrial natriuretic peptide in hypothyroidism. Journal of Urinary excretion of digoxin-like immunoreactive factor and Clinical Endocrinology and Metabolism 64 arginine-vasopressin in hyper- and hypothyroid rats. Clinical Science 353–355. 81 471–476. Vesely DL, Winters CJ & Sallman AL 1989 Prohormone atrial natriuretic peptides 1–30 and 31–67 increase in hyperthyroidism Received 4 August 2000 and decrease in hypothyroidism. American Journal of Medical Science Revised manuscript received 1 November 2000 297 209–215. Accepted 10 November 2000

www.endocrinology.org Journal of Endocrinology (2001) 168, 435–445

Downloaded from Bioscientifica.com at 10/02/2021 01:05:07AM via free access