| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Department of General Internal Medicine (H.P., P.H.E.M.d.M., P.K.C.S., H.B., A.E.M.), Dietetics (J.L., C.I.G.M.C.), and Centre for Human Drug Research (A.H.M.v.d.B., R.C.S., A.F.C., J.B.), Leiden University Medical Center, Leiden 2300 RC, The Netherlands
Abstract
We explored energy and macronutrient intake in patients with
Graves’ hyperthyroidism. We specifically hypothesized that
hyperthyroidism is associated with increased appetite for
carbohydrates, because of enhanced sympathetic tone and diminished
serotonin mediated neurotransmission in the brain. To test this
hypothesis, we measured food intake by dietary history and food
selected for lunch in the laboratory in 14 patients with Graves’
hyperthyroidism. Twenty-four-hour catecholamine excretion was used as a
measure of activity of the sympathetic nervous system (SNS) and the
plasma [Trp]/[NAA] ratio was measured to estimate (rate limiting)
precursor availability for brain 5-hydroxytryptamine synthesis. All
measurements were repeated after the subjects had been treated to
establish euthyroidism. In addition, the effects of nonselective
ß-adrenoceptor blockade upon these parameters were studied to
evaluate the influence of ß-adrenergic hyperactivity on food intake.
Hyperthyroidism was marked by increased SNS activity and reduced plasma
[Trp]/[NAA] ratio. Accordingly, energy intake was considerably and
significantly increased in hyper vs. euthyroidism, which
was fully attributable to enhanced carbohydrate consumption, as protein
and fat intake were not affected. These results suggest that
hyperthyroidism alters the neurophysiology of food intake regulation.
Increased SNS activity and reduced Trp precursor availability for
5-hydroxytryptamine synthesis in the brain may drive the marked
hyperphagia and craving for carbohydrates that appears to characterize
hyperthyroid patients. Because propranolol did not affect food intake
in hyperthyroidism, the potential effect of catecholamines on food
intake might be mediated by 
-adrenoceptors.
MOST MEDICAL textbooks consider a voracious appetite to be one of the clinical features of hyperthyroidism. However, as far as we are aware, food intake has never been measured in hyperthyroid humans. Moreover, the mechanism underlying this presumed need for food is unclear.
Appetite is governed by extremely complex interactions between neurotransmitter systems in specific brain nuclei (1). Serotonin (5-hydroxytryptamine, 5-HT) mediated neurotransmission appears to play an important role. Stimulation of postsynaptic 5-HT receptors specifically inhibits carbohydrate consumption in rats (2, 3). More vigorous stimulation reduces total calorie intake as well. In humans, serotoninergic drugs exert a similar influence upon food intake (4, 5). Conversely, low 5-HT levels in the brain are associated with enhanced appetite (3). Brain 5-HT is produced locally in serotoninergic neurons. 5-HT synthesis is critically dependent on the availability of its precursor amino acid tryptophan (Trp). Because transport of Trp across the blood brain barrier is competitive with transport of other neutral amino acids (NAA), e.g. isoleucine, leucine, valine, phenylalanine and tyrosine, its brain concentration is largely determined by the plasma ratio of Trp to other NAA (6, 7). A reduced plasma [Trp]/[NAA] ratio hampers transport of Trp into brain tissue and thereby diminishes 5-HT synthesis in the brain. This potentially stimulates appetite and energy (carbohydrate) intake.
One of the other neurotransmitters involved in the complex regulation of food intake is norepinephrine (NE). Injection of NE in the paraventricular nucleus specifically stimulates carbohydrate intake (8, 9) and a spontaneous increase of brain NE levels precedes carbohydrate consumption in rats (10).
The activity of the sympathetic component of the autonomic nervous system (SNS) is enhanced in hyperthyroidism (11). Moreover, the plasma [Trp]/[NAA] ratio appears to be reduced in experimental hyperthyroidism in rodents. Several mechanisms may underlie this phenomenon. Firstly, thyroid hormones are proteolytic and promote NAA release from muscle tissue (12), and secondly, tissue uptake of NAA is hampered, because their transport across cell membranes competes with thyroid hormones (13, 14). Finally, hyperthyroidism is associated with insulin resistance (15, 16), which may contribute to the diminution of the plasma [Trp]/[NAA] ratio (7, 17, 18).
Both neuroendocrine changes potentially stimulate energy (carbohydrate) consumption. A reduction of the plasma [Trp]/[NAA] ratio can diminish brain 5-HT synthesis and thereby enhance appetite for carbohydrates. Increased sympathetic tone in hyperthyroidism may stimulate carbohydrate intake directly by activation of postsynaptic adrenergic receptors in the brain.
We hypothesized that food choice in hyperthyroidism is characterized by high energy intake and a preference for carbohydrate-rich food items because of increased SNS tone and reduced precursor availability for brain 5-HT synthesis. To test this hypothesis, we measured food intake and neuroendocrine parameters in patients with Graves’ hyperthyroidism and compared the results with those of the same patients in the euthyroid state. We reasoned that this design allows optimal evaluation of the effects of hyperthyroidism on neuroendocrine systems and food intake within an individual patient. 24-h catecholamine excretion in urine was used as a measure of activity of the SNS and the plasma [Trp]/[NAA] ratio was measured to estimate precursor availability for brain 5-HT synthesis. In addition, the effects of nonselective ß-adrenergic receptor blockade upon these parameters were studied to evaluate the influence of ß-adrenergic hyperactivity on food intake.
Subjects and Methods
Subjects
Newly diagnosed, untreated hyperthyroid patients (HT) were recruited from the outpatient clinic of the department of Internal Medicine of Leiden University Medical Center. The diagnosis of Graves’ disease was established on the basis of clinical, biochemical and immunological data in all patients. Severe Graves’ opthalmopathy, any serious concomitant disease, the use of medication (except oral contraceptives) or diet and pregnancy were exclusion criteria. The Ethics Committee of Leiden University Medical Center approved the study protocol and all subjects gave written informed consent. The study was conducted according to the principles of the Helsinki declaration.
Study design
Hyperthyroid patients were studied on three occasions in this open study. The first occasion took place at the time of diagnosis. Subsequently, treatment with propranolol (10 mg orally qid) was initiated. The second occasion took place after 1 wk of propranolol treatment. Thereafter, the subjects started using thiamazol (Strumazol; 10 mg orally tid) to completely suppress thyroid function. L-thyroxine (Thyrax; 100 µg starting dose, rising up to 2 µg/kg body weight) was added to establish clinical and biochemical euthyroidism. Propranolol treatment was continued for several weeks until most clinical symptoms had vanished. The third occasion, which was required to be at least one month after the last dose of propranolol, took place in a stable euthyroid state, which was between 3–10 months after treatment initialization.
Study days
After an overnight fast, the subjects were admitted to the clinical research unit, where they handed in the urine collected over the previous 24 h. Blood samples for determination of plasma glucose, insulin, valine, isoleucine, leucine, phenylalanine, tyrosine and tryptophan, T3, T4, and TSH concentrations were collected. Subsequently, the subjects received a pure carbohydrate solution (75 g of glucose in 200 ml water). Blood samples for measurement of serum glucose and insulin concentrations were collected 30, 60, 90, and 120 min after glucose ingestion. Blood sampling for determination of amino acid levels was repeated at 120 min after the glucose load. The subjects were not allowed to eat or drink anything until lunch at 180 min after glucose ingestion. Spontaneous food choice was determined by offering the subjects a lunch buffet comprising 29 different food items of known macronutrient content and weight (19). All items were commercially available and known to the subjects. The subjects were asked to select their lunch from these items. During the meal, the subjects were allowed to take additional proportions of any product. They were instructed to eat to satiety. Remaining food was weighed and subtracted from the records.
Daily food consumption in free living conditions was determined using the dietary history with a food frequency list as cross-check (20). Patients were interviewed by a dietician about their dietary history at occasion 1 and 3. Macronutrient composition and daily caloric intake were calculated with the use of food composition tables (21).
Assays
All measurements were performed using standardized routine methodology. Free T4 (fT4) was measured on an IMx [Abbott, Abbott Park, IL; interassay coefficient of variation (CV), 3.8–7.1% at different levels]. T3 was measured by RIABEAD of the same company (interassay CVs of 2.0–4.4%). TSH was determined with an immunofluorometric assay (Wallac, Inc., Turku, Finland, interassay CVs: 2.4–5.9%). Serum insulin was measured by RIA (Medgenix, Fleurus, Belgium) and glucose was measured by a fully automated Hitachi system. NE, dopamine, and vanillylmandelic acid in urine were determined by HPLC.
Blood for amino acid measurements was collected in EDTA containing tubes. Plasma was deproteinized using an equal volume of 5% (wt/vol) sulfosalicylic acid in water and analyzed for amino acid concentrations by ion-exchange chromatography and ninhydrin derivatization on an LKB 4151 ALPHA PLUS automated amino acid analyzer (LKB Biochrom, Cambridge, UK) using standard conditions. For tryptophan the sum of albumin-bound and free plasma concentrations was determined using a standard addition calibration curve. All assays were performed at the clinical chemistry laboratories of Leiden University Medical Center.
Calculations and statistics
Insulin resistance was calculated by the HOMA index described by Matthews et al. (22): IRI = I/22.5·e-InG, where IRI = insulin resistance index, I = fasting serum insulin concentration, and G = fasting serum glucose concentration. The time-integrated glucose and insulin response to the breakfast was calculated as AUC.
Paired t tests were used to detect differences within groups. Correlation between variables was established by Pearson’s statistic. All calculations were carried out using SPSS for Windows (SPSS, Inc., Chicago, IL). Results are reported as mean (±SD).
Results
Subjects characteristics
Fifteen patients were included. The data from 1 subject were not used in the analysis. This patient completed only the first occasion. Another patient withdrew from the study before the third occasion. The available data from this subject were used in the analysis. Thus, the data represent 14 patients (13F/1M). At inclusion they were 38.9 ± 9.7 (mean ± SD; range 21–56) yr; body mass index, 23.1 ± 4.4 (16.2–30.0) kg/m2.
Thyroid hormone concentrations
Thyroid hormone levels are summarized in Table 1
. TSH and
fT4 concentrations were frequently lower,
respectively higher than the limit of detection in untreated patients
and propranolol-treated patients. In these cases, the limit of
detection for TSH (0.06 mU/liter) or the valid upper range value for
fT4 (77.2 pmol/liter) was used.
fT4 concentrations exceeded the valid upper range
value in so many cases for the untreated (8 cases) and
propranolol-treated (10 cases) patients that statistical analysis on
these parameters was not performed. Propranolol had no major influence
on TSH and fT4 levels but reduced
T3 levels by 0.74 nmol/liter. The euthyroid state
was adequately established by the use of thiamazol and
L-thyroxine in all patients.
|
Total daily energy intake in free living conditions was
significantly increased in hyperthyroidism, whereas energy intake for
lunch was not affected (Table 2). The
increase of daily energy intake was fully attributable to an increase
of carbohydrate consumption because protein and fat intake were not
affected (Fig. 1). Thus, the percentage
of carbohydrate consumption was significantly increased, whereas the
percentage protein intake was significantly reduced and percentage fat
intake not affected in hyperthyroid compared with euthyroid subjects,
both in free living conditions and for lunch (Table 2). The percentage
carbohydrate, fat, and protein that was chosen at lunch correlated with
dietary history data in hyper as well as in euthyroid conditions
(Pearson’s r = 0.46–0.87 (except for protein intake in euthyroid
conditions: r = 0.04), data not shown), which underscores the
reliability of our food intake data.
|
|
). Also, IRI was increased in
hyperthroidism (Fig. 3).
|
|
).
The percentage change of this ratio in response to glucose ingestion
was not affected by hyperthyroidism (Table 3).
|
).
|
Propranolol did not significantly affect food selection at lunch
(Table 2
), the plasma glucose or insulin concentrations in fasting
conditions or in response to glucose ingestion (Fig. 2), the plasma
[Trp]/[NAA] ratio (Table 3), or urinary catecholamine excretion
(Table 4).
Discussion
We explored energy and macronutrient intake in patients with Graves’ hyperthyroidism. Many medical textbooks (and our clinical impressions) consider hyperphagia to be an important feature of hyperthyroidism. We specifically hypothesized that hyperthyroidism is associated with increased appetite for carbohydrates because enhanced sympathetic tone and a decreased plasma [Trp]/[NAA] ratio potentially promote carbohydrate consumption.
The results of our work suggest that hyperthyroidism profoundly changes the neuroendocrine milieu that governs food intake. Specifically, the plasma [Trp]/[NAA] ratio was significantly reduced and SNS activity increased in hyperthyroid vs. euthyroid conditions. These neuroendocrine changes potentially hamper 5-HT synthesis and activate adrenoceptors in the brain respectively. Accordingly, energy intake was considerably and significantly higher in the hyperthyroid state, which was fully attributable to an increase of carbohydrate consumption. Protein and fat intake were not affected by hyperthyroidism.
Hyperthyroidism was marked by a reduction of the plasma [Trp]/[NAA] ratio compared with the euthyroid state, probably because thyroid hormones promote the release of NAA from protein depots (12) and hamper uptake of NAA by peripheral tissues (13), whereas Trp metabolism remains relatively unaffected (14). Moreover, hyperthyroidism was associated with insulin resistance, as was consistently demonstrated by the fasting glucose and insulin levels, the response to glucose ingestion and by the IRI. These findings are in keeping with previous work by other authors (15, 16). Insulin increases the plasma [Trp]/[NAA] ratio because it promotes uptake of NAA in muscle tissue, whereas Trp uptake is hampered by its chemical bond to albumin in plasma (7, 17, 18). This mechanism was sustained by our observations of the plasma [Trp]/[NAA] ratio in response to glucose ingestion. Therefore, insulin resistance may reduce the plasma [Trp]/[NAA] ratio. However, the change of the plasma [Trp]/[NAA] ratio in response to glucose ingestion was not different in the hyper vs. the euthyroid state, suggesting that insulin resistance did not significantly affect the ratio in hyperthyroidism. This probably reflects the fact that the protein sparing effects of insulin are preserved in the hyperthyroid state, which may serve to counteract the proteolytic effects of thyroid hormones (12). Thus, compensatory hyperinsulinemia presumably mitigates the thyroid hormone-induced increase of plasma NAA levels in hyperthyroidism.
Catecholamine excretion in urine was considerably increased in the hyperthyroid state. This finding suggests that sympathetic tone is enhanced in hyperthyroidism, which is underscored by spectral analysis of heart rate fluctuations in hyperthyroid patients (11). However, it is in apparent contrast to earlier papers reporting normal or even reduced sympathetic activity in hyperthyroidism (23, 24, 25, 26). Importantly, these reports are generally based on plasma norepinephrine concentrations in single samples of venous forearm blood (25, 26) or whole body catecholamine turnover rates, calculated from specific activities of radiolabeled catecholamines in venous plasma (23, 24). It is now known that catecholamine levels are more appropriately determined in arterial(ized) blood, inasmuch as extraction from venous circulation occurs across various organs (27, 28). Urinary excretion of catecholamines and their metabolites also appears to be a reliable tool to estimate their plasma concentration and whole body turnover (28, 29).
Nonselective ß-adrenergic receptor blockade did not affect food
intake in hyperthyroid patients. This finding corroborates observations
in rats, which show that the -adrenoceptor antagonist phentolamine,
but not propranolol, blocks NE-induced feeding (30). More
recent data indicate that the stimulatory action of NE on eating is
specifically linked to activation of
2-adrenoceptors in the paraventricular nucleus
(31, 32). Thus, the fact that propranolol did not affect
food intake in hyperthyroid patients does not discard the possibility
that hyperactivity of adrenoceptors stimulates appetite for
carbohydrates in hyperthyroidism. It seems important to emphasize that
our data do not prove a causal relationship between enhanced
sympathetic tone or reduced brain 5-HT synthesis and food intake in
hyperthyroid patients. Our inferences in this context are based on data
from animal experiments (10, 33). In fact, the positive
association between catecholamine excretion in urine and carbohydrate
intake did not attain statistical significance (P =
0.1) in our study.
It seems important to emphasize that other neuroendocrine systems may affect food intake in hyperthyroidism in addition to the autonomic nervous system and 5-HT mediated pathways. Complex interactive neural routes in various brain nuclei orchestrate food intake and macronutrient selection. Numerous neuropeptides and neuramines are involved (1). Several investigators have hypothesized that alterations of leptin secretion might disrupt energy balance in hyperthyroidism. However, most experiments have refuted this thesis (34, 35, 36, 37). Interestingly, it was recently reported that T3 down-regulates expression of the tub gene, encoding the tubby protein, in the ventral and dorsomedial hypothalamus in rodents (38). Tubby and tubby-like proteins have been implicated as transcription factors with potentially broad biological function (39). Targeted deletion of the tub gene induces hyperphagia and maturity onset obesity in mice (40). Thus, although the (neurobehavioral) actions of tubby proteins in humans are not known, it is conceivable that hyperthyroidism modulates energy balance via down-regulation of tub gene expression in man. The effects of thyroid hormones on other neurotransmitter systems involved in the regulation of energy balance remain to be established.
What picture of the neurophysiology of food intake regulation in
hyperthyroidism emerges from our data? It appears that hyperthyroid
patients are hyperphagic and specifically crave for carbohydrate rich
foods. Increased activity of the sympathetic nervous system and reduced
availability of Trp for brain 5-HT synthesis may be involved in the
pathogenesis of aberrant eating behavior in these patients. Because
propranolol did not affect food intake in hyperthyroidism, the
potential effect of catecholamines on food intake might be mediated by
-adrenoceptors.
Acknowledgments
Received May 16, 2001. Accepted September 6, 2001.
Address all correspondence and requests for reprints to: Dr. H. Pijl, M.D., Leiden University Medical Center, Department of Internal Medicine, C1-R39, P.O. Box 9600, 2300 RC Leiden, The Netherlands. E-mail: h.pijl@lumc.nl.
Footnotes
Abbreviations: AUC, Areas under curve; CV, coefficient of variation; fT4, free T4; 5-HT, 5-hydroxytryptamine; NAA, neutral amino acids; NE, norepinephrine; SNS, sympathetic nervous system; Trp, tryptophan.
References
2-noradrenergic receptors mediate
feeding induced by paraventricular nucleus injection of norepinephrine
and clonidine. Eur J Pharmacol 115:11–19[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  O. Gimenez-Palop, G. Gimenez-Perez, D. Mauricio, E. Berlanga, N. Potau, C. Vilardell, J. Arroyo, J.-M. Gonzalez-Clemente, and A. Caixas Circulating ghrelin in thyroid dysfunction is related to insulin resistance and not to hunger, food intake or anthropometric changes Eur. J. Endocrinol., July 1, 2005; 153(1): 73 - 79. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. M. Kong, N. M. Martin, K. L. Smith, J. V. Gardiner, I. P. Connoley, D. A. Stephens, W. S. Dhillo, M. A. Ghatei, C. J. Small, and S. R. Bloom Triiodothyronine Stimulates Food Intake via the Hypothalamic Ventromedial Nucleus Independent of Changes in Energy Expenditure Endocrinology, November 1, 2004; 145(11): 5252 - 5258. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Silva The Thermogenic Effect of Thyroid Hormone and Its Clinical Implications Ann Intern Med, August 5, 2003; 139(3): 205 - 213. [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. D. Riis, T. K. Hansen, N. Moller, J. Weeke, and J. O. L. Jorgensen Hyperthyroidism Is Associated with Suppressed Circulating Ghrelin Levels J. Clin. Endocrinol. Metab., February 1, 2003; 88(2): 853 - 857. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
,
P < 0.01 vs. untreated.