Low selenium status in the elderly influences thyroid hormones

Transcription

1 Clinical Science (1995) 89, (Printed in Great Britain) 637 Low selenium status in the elderly influences thyroid hormones Oliviero OLIVIERI, Domenico GIRELLI, Margherita AZZINI, Anna Maria STANZIAL, Carla RUSSO, Massimiliano FERRONI* and Roberto CORROCHER Institute of Medical Pathology, Chair of Internal Medicine, University of Verona, Italy, and *Public Health Service, Verona, Verona, Italy (Received 10 May/31 July 1995; accepted 10 August 1995) 1. lodothyronine 5 -deiodinase, which is mainly responsible for peripheral triiodothyronine ( T3) production, has recently been demonstrated to be a selenium-containing enzyme. In the elderly, reduced peripheral conversion of thyroxine (T,) to T3 and overt hypothyroidism are frequently observed. 2. We measured serum selenium and erythrocyte glutathione peroxidase (as indices of selenium status), thyroid hormones and thyroid-stimulating hormone in 109 healthy euthyroid subjects (52 women, 57 men), carefully selected to exclude abnormally low thyroid hormone levels induced by acute or chronic diseases or calorie restriction. The subjects were subdivided into three age groups. To avoid conditions of undernutrition or malnutrition, dietary records were obtained for a sample of 24 subjects, randomly selected and representative of the whole population for age and sex. 3. In order to properly assess the influence of selenium status on iodothyronine 5 -deiodinase type I activity, a double;blind placebo-controlled trial was also carried out on 36 elderly subjects, resident at a privately owned nursing home. 4. In the free-living population, a progressive reduction of the T3/T, ratio (due to increased T, levels) and of selenium and erythrocyte glutathione peroxidase activity was observed with advancing age. A highly significant linear correlation between T4, T3/T, and selenium was observed in the population as a whole (for T,, R= , P<O.O02; for T3/T, ratio, R=0.32, P<O.Ol) and in older subjects (for T,, R= -0.40, P<0.05; for T3/T4 ratio, R=0.54, P < 0.002) * 5. The main result of the double-blind placebocontrolled trial was a significant improvement of selenium indices and a decrease in the T, level in selenium-treated subjects; serum selenium, erythrocyte glutathione peroxidase activity and thyroid hormones did not change in placebo-treated subjects. 6. We concluded that selenium status influences thyroid hormones in the elderly, mainly modulating T, levels. INTRODUCTION Full activity of the thyroid hormones requires the deiodination of thyroxine (T,) to triiodothyronine (T,). T, is produced by the thyroid gland but results mainly from the peripheral deiodination of T, catalysed by iodothyronine 5 -deiodinase (ID) [ 11. Three isoenzymes of iodothyronine 5 -deiodinase have been identified [l]: type I (ID-I) is found in the liver, kidney and thyroid; ID-I1 is abundant in the brain, brown adipose tissue and pituitary; and ID-I11 is present in the brain and placenta. ID-I, which is mainly responsible for peripheral T3 production, has recently been demonstrated to be a selenium (Se)-containing enzyme [2, 31; this observation agrees with previous reports that a Sedeficient diet inhibits ID-I in rats [4, 51. Advanced age is a well-recognized condition characterized by important changes in thyroid function [6]. In addition to a high frequency of overt hypothyroidism [6], reduced iodine uptake, weight, colloid content, follicular volume of the gland and reduced peripheral conversion of T, to T, with a lower T3/T4 ratio are usually observed in elderly euthyroid people [7]. Numerous features of the aging process associated with impaired or slowed cellular functions to some extent mimic those of hypothyroidism, so it is sometimes very difficult to clinically distinguish the manifestations of the illness from those related to aging per se. As a result of these similarities, the concept of aging as a sort of tissue hypothyroidism has also been proposed [7]. No studies relating Se status, thyroid function and aging in humans have been published to date. However, since Se status tends to decline in people over 60-65years of age [S, 91, reduced ID-I activity due to impaired Se availability may explain the Key words: aging, glutathione peroxidase, nutrition, selenium, thyroid hormones, thyroid-stimulating hormone. Abbreviations: FT, free thyroxine; C-SH-Px, glutathione peroxidase; ID, iodothyronine S deiodinase; T,, triiodothyronine; T, thyroxine; TBG, thyroxinebinding globulin; TSH, thyroid-stimulating hormone. Correspondence: Professor Roberto (:orrocher, lstituto di Patologia Medica, Cattedra di Medicina Interna, Universita di Verona, Policlinico Borgo Roma, Verona, Italy.

2 Olivieri et al. reduction in the serum T,/T, ratio observed in elderly people. To test this hypothesis, we investigated the relationships between age, Se status and thyroid hormones in three groups of free-living healthy subjects of different age, matched for sex distribution. In addition, in a smaller group of institutionalized elderly subjects, we investigated the effects of Se supplementation on thyroid hormones. MATERIAL AND METHODS Subjects An initial age-based, sex-balanced selection of 500 subjects was obtained by applying tables of random numbers to the population of Nove (a village near Vicenza, northern Italy) appearing on the electoral register. A further selection was performed in order to study the consequences of aging without the confounding effects of chronic or acute diseases that might affect levels of both Se and circulating thyroid hormones. Thus, very strict criteria were adopted to define the healthy population. We therefore excluded all subjects known to be suffering from hypertension, diabetes, hyperlipidaemia, liver disease, neoplastic disease, renal disease, endocrinological or immunological diseases, coagulative disorders or acute intercurrent illness. Three seemingly euthyroid subjects with thyroid-stimulating hormone (TSH) values higher than 5 m-units/l were also excluded. One hundred and nine subjects (52 women, 57 men) were finally admitted to the study. None was institutionalized or on a special diet, and none was taking preparations containing multivitamins and/or trace elements. The population was divided into three sex-matched groups, according to age: group I, 36 subjects (2W4years old); group 11, 36 subjects (45-65years old); group 111, 37 subjects (over 65years old). In order to exclude conditions of undernutrition or malnutrition, a 7-day food record was obtained for a representative sample of the subjects (n=24, eight per group), randomly selected from the age groups. Informed consent was obtained from all participants according to the ethical guidelines of the Helsinki Declaration. Supplementation trial In order to properly assess the influence of Se status on ID-I activity, a double-blind placebocontrolled trial was also carried out. As changes in Se and thyroid hormones were observed to a larger extent in older subjects, the trial was conducted with 40 subjects aged 85f7years, who were resident at a privately owned nursing home in Verona, Italy. People with intercurrent nutritional disturbances, thyroid or gastrointestinal diseases that could negatively influence thyroid hormones or Se status were excluded. Several subjects were taking digoxin, theophylline and/or non-steroidal antiinflammatory drug, but none were taking preparations containing multivitamins and/or other trace elements. Four subjects were excluded from the study early on because of intercurrent acute illness interfering with oral absorption of the tablets. All the other subjects consumed their habitual diet, prepared in the kitchen of the nursing home, and had no known nutritional problems during the study. The subjects were randomly allocated to the Se (Se+) or placebo (Se-) group, according to the number of their bedroom. Every day for 3months, one tablet of sodium selenite (loopg) or placebo (starch) was taken during breakfast, under the supervision of a nurse. Neither the patient nor the investigator knew which treatment was allocated and only at the end of the study was the randomization code broken. Thirty six subjects (Se- group, 14 women and 3 men; Se+ group, 14 women and 5 men) concluded the study and their results were suitable for statistical evaluation. No side effects were recorded during the study. Blood samples for measuring Se parameters and thyroid hormones were collected before and after the supplementation period (3 months). Informed consent was obtained from all subjects and the study was approved by the local ethics committee. Biochemical analysis Blood samples were collected after overnight fasting and processed within 1 h. Haemogram was automatically measured using the Technicon H-2 System (Technicon Instruments, Tarrytown, NY, U.S.A.). Plasma cholesterol and serum albumin were determined using a Technicon DAX 96 automated analyser (Technicon Instruments). Serum T,, T,, free thyroxine (FT,) and TSH were assayed by a chemiluminescence immunoassay system (LIA-mat, Byk-Sangtec Diagnostica, Germany) [ 10, 111. For the quantitative determination of thyroxine-binding globulin (TBG), a radioimmunoassay kit (Gamma Dab 25 TBG Radioimmunoassay, Sorin Biomedica, Vercelli, Italy) was employed. To prepare haemolysates for enzyme assays, the erythrocytes were washed and then filtered in cellulose. For serum Se assay, blood was collected in trace-element-free vacutainer tubes, containing no additives, for metal and metalloid determinations (Becton Dickinson, Rutherford, NJ, U.S.A.). Erythrocyte glutathione peroxidase assay. Activity of erythrocyte glutathione peroxidase (GSH-Px, EC ), which accurately reflects the Se intake over 2-3months [8], was measured in 50pl of cell lysate, according to the Gunzler method, using t- butylhydroperoxide as acceptor substrate, as previously described [ 123. GSH-Px activity was expressed as IU, defined aspmol of NADPH oxidized per min per g of Hb (IU/g Hb). Se assay. Serum Se was determined as previously

3 Selenium and thyroid hormone in aging 639 Table I. Main features of free-living subjects, subdivided according to age groups. Statistical significance: *P<0.05 for group I compared with group II; tp<0.05 for group I compared with group 111. Comparisons between age groups were performed using analysis of variance and differences at the 5% Significance level were assessed by Tukey's test. Means not sharing a common superscript symbol are not significantly different. Age (Yean) Body mass index (kg/m') Cholesterol (mmol/l) HDL-cholesterol (mmol/l) Triacylglycerols (mmol/l) Albumin (g/l) Haemoglobin (g/dl) Packed cell volume (%) Dietary intake (n = 24) Total calories (cal kg-lday-') Lipids (gkg-'day-') Carbohydrates (g kg-lday-0 Proteins (g kg- I day- I) Group I Group II Group f 3.2* 4.85 f I.w*t I 52 f f f f 2. I 41 f f 4. I 0.99 f & I.05 I.2 f f 2.5* 5.90 f 0.87* I.45 f f k k f f7.4 I.a9 f f I.7 1.3f0.3 > f f 0.19t 1.53 f fo f4.6 13k f f I.04 f f I.8 I.4 f 0.6 described [ 121, using hydride generation atomic absorption spectrometry on serum stored at - 20 C until analysed. Briefly, 1 ml of serum was treated by wet digestion in nitric acid/hydrochloric acid ( 1.55, v/v). Se concentrations were determined using a G.B.C. 902 atomic absorption spectrophotometer equipped with a hydride generation system and a selenium hollow cathode lamp. Sodium borohydride was used as reductant. A batch of Seronorm trace element serum (Nycomed Pharma As, Oslo, Norway) with a defined Se concentration (mean recommended value 1.19 pmol/l, range ; mean value obtained in our laboratory 1.20pmol/l) was used as a control reference. The within- and between-run coefficients of variation were 1.7% and 4%, respectively. Statistical methods Statistical analysis was carried out with an Apple Macintosh SE/30 computer using the Systat 5.0 statistical package. The data are given as means & SD. Comparison of the various parameters among age groups was performed using one-way analysis of variance; differences at the 5% significance level (P < 0.05) were assessed by Tukey's 'post hoc' probability test for multiple comparisons. Paired Student's t-test was used to compare the same parameters before and after supplementation in Se- or placebo-treated subjects. Simple correlations were determined using Pearson's correlation coefficient. RESULTS Population study Table 1 gives the main features of the subjects, subdivided according to the various age groups. All our subjects had normal nutritional indices such as Table 2. Thyroid parameters and indices of Se in free-living subjects. Statistical significance: *P<0.05 for group I compared with group 111; tp<0.05 for group I1 compared with group 111. Comparisons between age groups were performed using analysis of variance and differences at the 5% significance level were assessed by Tukey's test. Means not sharing a common superscript symbol are not significantly different. Age (years) m, (pmol/l) I 1.5 f 2.7 T, (nmol/l) I.87 k 0.34 T, (nmol/l) 85.1 f 16* TSH (m-units/l) I.39 f 0.55 TIP, ratio f 0.007* Se (pmol/l) I. 19 f 0.19* Erythrocyte GSH-Px (IU/g Hb) 7.24f 1.3' 0.5 t I ~~ Group I Group II Group I.% f 6t I.03 k f fo.l9t 6.46 f I.48 > 65 I I.w f 2.3 I.95 f f25*t l.43f * 0.94 f O.l8*t 5.86 f I.68* r= I00 I T, (nmol/l) Fig. 1. Correlation between serum Se and T4 in a free-living population as a whole body mass index, Hb, packed cell volume and serum albumin; dietary records showed no significant differences between the groups. Levels of thyroid hormones, TSH, T,/T, and serum Se and erythrocyte GSH-Px activity of the subjects, subdivided according to age groups, are listed in Table 2. Erythrocyte GSH-Px and the T,/T, ratio were lower in older (group 111) than in younger subjects (group I). Serum Se was lower and T, was higher in group 111 compared with the values in the other two groups. T,, FT, and TSH levels were similar in all groups (Table 2). In the population as a whole, serum Se was positively correlated with the T,/T, ratio (R =0.32, P < 0.01) and negatively correlated with T, (R=-0.31, P<O.Ol; Fig. 1). When the analysis was performed separately for the different age groups, a significant correlation was found only for subjects in group I11 (for T,/T,, R=0.54, P<0.002; for T,, R= -0.40, P<O.O5; Fig. 2), who also had lower Se values. The T,/T, ratio and erythrocyte GSH-Px activity were also correlated (R=0.50; P<0.002) in the group I11 subjects.

4 Olivieri et al. :::. :::I, I00 I ,, T, (nrnol/l) Fig. 2. Correlation between serum Se and T4 in free-living subjects of group 111 (age r65years) Supplementation trial Table 3 gives the main features of institutionalized subjects who underwent thc trial. Serum Se, erythrocyte GSH-Px, T,, T,. FT, and T,/T, were generally very low in these institutionalized subjects, while TSH levels were still normal. After supplementation, parameters of Se status strongly improved in Se-treated (Se= +61%, GSH- Px = '2,) patients, but did not change in the placebo-treated subgroup (Table 3). The only noteworthy modification in thyroid hormones induced by Sc supplementation was a significant decrease in T, levels (-7.5%); this change was not due to a parallel decrease in TBG, the major T,-binding protein in serum (Table 3). No changes in hormonal parameters were observed in the placebo-treated subgroup (Table 3). DISCUSSION After the recent demonstrations that ID-I is a selenocysteine-containing enzyme [2, 31, increasing attention has been focused on the relationships between Se status and thyroid hormones. Pathological conditions affecting thyroid function such as Graves' disease [ 131 and hypothyroidism [ have been studied, but there has been little investigation of the topic in healthy humans, particularly in relation to aging. Our study aimed to investigate the topic by two different approaches: firstly, to evaluate age-related changes in Se status and thyroid hormones and their possible correlation in a free-living healthy population; secondly, to verify the consequences on thyroid hormones of a Se supplementation trial in a smaller sample of institutionalized elderly people with a poor Se status. Since serum levels of thyroid hormones could be lowered by a great variety of acute or chronic diseases or calorie restriction in seemingly euthyroid individuals [7], we devoted particular care to the selection of free-living subjects. With the strict criteria adopted, our aim was to avoid possible changes in thyroid hormone concentrations which might reflect diseases associated with age, rather than being an effect of aging per se. Nutritional state was also checked: all our subjects had normal nutritional indices such as body mass index, Hb, packed cell volume and serum albumin; moreover, a 7-day food record randomly obtained for a representative sample of subjects showed similar calorie intake without any significant differences between the age groups (Table 1). We found a low T,/T, ratio, high T, and reduced Se and erythrocyte GSH-Px activity in the older subjects; moreover, T, and T,/T, were significantly correlated with Se. Although these results supported the hypothesis that an impaired Se status may explain the hormonal changes observed in elderly people, conceivably via impaired activity of the Secontaining enzyme ID-I, direct evidence of a causative link was still lacking. For this purpose, a double-blind placebocontrolled supplementation trial was carried out, whose main result was a significant decrease in T, levels in Se-treated subjects (Table 3). In this case, we adopted three criteria for the selection of subjects: no evidence for gastrointestinal or other diseases interfering with absorption of tablets by the oral route, a poor Se status and a constant diet during the trial (for the latter reasons we chose very old and institutionalized people). All subjects included in the study met all these requirements. As reported in Table 3, the subjects presented a very poor Se status. Compliance of the subjects to the Se treatment was proved by the remarkable improvement of Se indices in Se-treated subjects, whereas in placebo-treated subjects the same parameters remained unchanged (Table 3). Moreover, the diet provided by the nursing home was constant over a period of 3months. However, in these very old subjects (85 k 7 years; twenty eight subjects were more than 80years of age) a generalized reduction of thyroid hormone synthesis and an impaired nutritional state was evident (Table 3). In these respects, the institutionalized subjects were very different from the freeliving older subjects of group 111 (mean age 76+6years; only five were more than 80years of age). In particular, Se deficiency was more severe and T, levels were lower than in the subjects of group 111. Clearly, these two groups of elderly people were representative of two different steps in the aging process in that a progressive impairment of several functions, i.e. thyroid function, occurs. In institutionalized subjects of a more advanced age and poorer nutritional status, the functional decline of thyroid gland (the only source of T,) determined a significant reduction of circulating T, levels, whereas in free-living elderly people thyroid hormone synthesis was still effective and T, levels were high.

5 Selenium and thyroid hormones in aging 641 Table 3. Nutritional and thyroidal parameters of the institutionalized subjects before and after Se (Se+) or placebo (Se-) supplementation. *Paired Student f-test: significant difference (P <0.05) between values at baseline and after supplementation period. Baseline Se - Baseline Se + Age (85 f 7yean) Body mass index (kg/m') Cholesterol (mmol/l) Albumin (g/l) Hb (g/dl) Packed cell volume (%) FT, (pmol/l) T, (nmol/l) T, (nmol/l) TSH (m-units/l) T,/T, ratio Se (pmol/l) Erythrocyte GSH-Px (IU/g Hb) TBG (mg/l) 84.5 f f f I 30. I f I.55 I1.9f * f * I3 I. 14 f f ko f I f f 4 5.2f I 29.3 f I f f 4.8 lof f f * 0.7 I 0.015fO f f f f f f I 31.3f f f f I.5 I f9 1.2 fo f k f f f 5 5. I f f f f f f * I.I8 * & I.34 f 0.3* 7.78 f 2' 24.2 f 6 In our institutionalized elderly people, the recoverv of an adequate Se status Droduced a 7.5X decrease in T, leiels in spite of impaired thyroid hormone synthesis. This percentage could probably be higher in subjects (as for example in our freeliving elderly people) presenting normal or increased availability of T,, the substrate for ID-I. Supposing that the increase in T, levels observed in the free-living subjects of group 111 (mean T,= 104nmol/l), compared with that of groups I (mean T, = 85.1 nmol/l) or I1 (mean T, = 92.9 nmol/l), was only due to an impaired ID-I activity, the percentage increase in T, values dependent on impaired ID-I function could be estimated to be between 10.6 and 18%. Therefore, a 7.5% reduction in T,levels obtained after Se supplementation is a noteworthy result, provided that institutionalized elderly people had a T, reduction of 19.&26.3% in comparison with free-living subjects of group I or 11, respectively. Results of the Se supplementation trial suggest that T, is the most sensitive parameter to Se repletion and in conditions where thyroid T, synthesis is blunted, such as in very advanced age. To the best of our knowledge, this is the first study demonstrating that Se supplementation modifies the parameters of thyroid function in euthyroid elderly people. Previous observations by other authors are in agreement with the present results. In animals [ and in children [20], low Se status was demonstrated to lower the T,/T, ratio, mainly through an increase in T4 rather than by affecting T, levels. In Se-deficient children, affected by phenylketonuria, after 46 weeks of Se supplementation the only modification was a significant decrease in T,, which increased back to the initial values within 3 weeks after stopping supplementation [2 13. The only apparent discrepancy between our results and previous findings concerns the results of the fraction free of T,. FT, levels were unchanged after 3 months of Se supplementation in our elderly subjects; conversely, in rats [22] and in children [20], changes in Se status give rise to changes in both total T, and FT,. We do not have any logical explanation for this discrepancy, except the obvious difference of species and age of our subjects. In this and in other studies, TSH levels were not affected by impaired Se status [4, 15, 20, 221, suggesting normal T, and deiodinase activity within the pituitary, since inadequate T, concentrations act as a signal for TSH secretion. In rats, aging was reported to be associated with reduced ID-I activity in liver and thyroid, but with normal circulating TSH levels, normal T, and increased deiodinase activity within adenohypophysis [23]. In conclusion, a definite percentage of circulating T, is sensitive to the variations of Se, the trace element cofactor of ID-I. Due to the high frequency of Se deficiency in elderly subjects, abnormalities in T, levels as a result of impaired peripheral deiodinase activity are probably frequent in older people, but are clinically underscored because T, and TSH values are normal. The long-term consequences of Se depletion on thyroid metabolism during the aging process remain to be established and to be evaluated in terms of preventive action and health advice. In fact, a Se deficiency-induced increase in T, may have inhibitory effects on ID-I1 [24], whose activity is the major determinant of T, levels in the neuronal and pituitary cells. Further investigations, specifically addressed to the effects of long-term Se supplementation, will be necessary to clarify this potentially very important point. ACKNOWLEDGMENT This work was supported by grants from the National Research Council, No , from the Ministry of the University and of Scientific and

6 Olivieri et al. Technological Research, and from the Veneto activity and serum selenium concentration in patients with chronic renal failure: relations to dialysis treatments, diet and cardiovascular complications. Region Department of Health. Clin Sci 1993; W Reglinski 1. Smith WE, Wilson R, Halls DJ, McKillop JH, Thomson ]A. Selenium in Graves' disease (Letter). Clin Chim Acta : Contempre B. Dumont JE. Ngo B, et al. Effects of selenium supplementation REFERENCES in hypothyroid subjects of an iodine and selenium deficient area: the possible danger of indiscriminate supplementation of iodinedeficient subjects with selenium. J Clin Endocrinol Metab 1991; IS. Contempre B, Duale NL. Dumont JE, Ngo B, Diplock AT, Vanderpas J. Effect of selenium supplementation on thyroid hormone metabolism in an iodine and selenium deficient population. Clin Endocrinol 1992; Roti E, Minelli R. Gardini E, et al. Selenium administration does not cause thyroid insufficiency in subjects with mild iodine deficiency and sufficient selenium intake. J Endocrinol Invest 1993; 16: I. Kohrle J. Hesch RD, Leonard JL. lntracellular pathways of iodothyronine metabolism. In: Braverman LE, Utiger RD. eds. The thyroid. Philadelphia: Lippincott, 1991: Behne D, Kyriakopoulos A, Meinhold H, Kohrle J. Identification of typd iodothyronine S'deiodinase as a selenoenzyme. Biochem Biophys Res Commun 1990; Berry MJ, Banu L, Larsen PR. Type1 iodothyronine deiodinase is a selenocysteine-containing enzyme. Nature (London) 1991; Beckett GJ, Beddows SE, Morrice PC, Nicol F. Arthur JR. Inhibition of hepatic deiodination of thyroxine is caused by selenium deficiency in rats. Biochem J 1987; 248: Beckett GI. MacDougall DA, Nicol F. Arthur JR. Inhibition of type I and type II deiodinase activity in rat liver, kidney and brain produced by selenium deficiency. Biochem J 1989; Rae P, Farrar J, Beckett G, Toft A. Assessment of thyroid status in elderly people. Br Med J 1993; 307: lngbar SH. The thyroid gland. In: Wilson ID, Foster DW, eds. Textbook of endocrinology. Philadelphia: Saunden. 1985: Lockitch G. Selenium: clinical significance and analytical concepts. Crit Rev Clin Lab Sci 1989; 27: Olivieri 0, Stanzial AM, Girelli D. et al. Selenium status, fatty acids, vitamins A and E and aging: the Nove study. Am j Clin Nutr k Zucchelli GC, Pilo A, Masini S. Chiesa MR. Prontera C. A new chemiluminescence immunoassay for triiodothyronine and thyroxine: evaluation using quality control sera assayed in an interlaboratory survey. J Clin Chem Clin Biochem 1990; 28: I I. Pilo A, Zucchelli GC, Chiesa MR. Evaluation of precision and relative accuracy of a new automated luminescence immunoassay for TSH and ft4. J Clin lmmunoassay 1994; 17: Girelli D, Olivieri 0, Stanzial AM, et al. Low platelet glutathione peroxidase 17. Beckett GI. Nicol F. Rae PWH, Beech S, Guo Y. Arthur JR. Effects of combined iodine and selenium deficiency on thyroid hormone metabolism in rats. Am J Clin Nutr 1993; 57 (Suppl.): Vadhanavikit S, Ganther HE. Selenium requirements of rats for normal hepatic and thyroidal S'deiodinase (type I) activities. J Nutr 1993; Chanoine 1, Braverman LE, Farwell AP, et al. The thyroid gland is a major source of circulating T, in the rat. J Clin Invest 1993; 91: Terwolbeck W, Behne D. Meinhold H. Menzel H, Lombeck I. Increased plasma T,-levels in children with low selenium state due to reduced type I iodothyronine S'deiodinase activity? J Trace Elem Electrolytes Health Dis 1993; Vanderpas J, Calomme M. Francois B, et al. Thyroid function parameters during selenium repletion/depletion in phenylketonuric children [Abstract]. J Endocrinol Invest (Suppl.): Arthur JR, Nicol F. Beckett GJ. Selenium deficiency, thyroid hormone metabolism, and thyroid hormone deiodinases. Am J Clin Nutr 1993; 57 (Suppl.): 236-9s. 23. Donda A, Lemarchand Beraud T. Aging alters the activity of S'deiodinase in the adenohypophysis, thyroid gland and liver of the male rat. Endocrinology (Baltimore) 1989; 124: Arthur JR. Nicol F, Hutchinson AR, Beckett GJ. The effects of selenium depletion and repletion on the metabolism of thyroid hormones in the rat. J lnorg Biochem 1990;