| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Endocrinology M (P.S.H., T.H.B., L.H.), Odense University Hospital, and The Danish Twin Registry (P.S.H., K.O.K.), Epidemiology, Institute of Public Health, University of Southern Denmark, DK-5000 Odense C, Denmark; and Danish Epidemiology Science Centre (T.I.A.S.), Institute of Preventive Medicine, Copenhagen University Hospital, DK-1399 Copenhagen K, Denmark
Address all correspondence and requests for reprints to: Pia Skov Hansen, The Danish Twin Registry, Epidemiology, Institute of Public Health, University of Southern Denmark, Odense, Sdr. Boulevard 23A, DK-5000 Odense C, Denmark. E-mail: piaskovhansen{at}dadlnet.dk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A representative sample of self-reported healthy twin pairs was identified through the Danish Twin Registry. A total of 284 monozygotic (MZ), 286 dizygotic same-sex (DZ), and 120 opposite-sex (OS) twin pairs were investigated. A classical twin study was performed. After adjustment for age, sex, and other covariates, the intraclass correlations of serum TSH, free T4, and free T3 were calculated. To elucidate the relative importance of hereditary and environmental factors on the variation of these hormone levels, quantitative genetic modeling was used.
The intraclass correlations were consistently higher for MZ twin pairs than for DZ twin pairs. Regression analysis suggested that iodine intake played a small but significant role for the concentration of serum TSH and free T4, whereas cigarette smoking was without influence. In quantitative genetic modeling, the heritability (with 95% confidence intervals) accounted for 64% (57–70%) of the variation in serum TSH concentration and 65% (58–71%) and 64% (57–70%), for the concentrations of free T4 and free T3, respectively.
Genetic factors play a substantial role in controlling the pituitary-thyroid axis, indicating that each individual has a genetically determined thyroid function set-point. Whether this is of importance when treating individuals in whom pituitary-thyroid function has been disrupted by, e.g. hypo- or hyperthyroidism, remains to be clarified.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
On the other hand, it is well accepted that factors such as iodine intake (7, 8), sex (9, 10), and time of blood sampling (1, 11, 12) influence the serum TSH concentrations. Furthermore, factors such as age (13, 14, 15), cigarette smoking (16, 17), and phase of menstrual cycle (18) seem to influence concentrations of serum TSH, free T3, and free T4 although the latter findings are less unanimous. The importance of these factors compared with hereditary factors is largely unknown. To elucidate the relative impact of genetic and environmental influence on serum TSH, free T4, and free T3 concentrations, we studied a large twin sample.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The present study is part of a nationwide project (GEMINAKAR) investigating the relative influence of genetic and environmental factors on a variety of different traits among Danish twins.
A representative sample of complete twin pairs was recruited from the population-based Danish Twin Registry (19, 20). The majority of these twins participated in a questionnaire survey regarding physical health and health-related behavior. The twins included in the GEMINAKAR study were healthy as reported by themselves. To get an equal distribution of twin pairs, sampling was stratified according to age, sex, and zygosity.
The examinations were running in parallel, throughout the year, at The Danish Twin Registry in Odense and at The Institute of Preventive Medicine in Copenhagen from August 1997 to November 2000. The twins in a pair were examined on the same day. All twins with at least one partner living in the western part of Denmark were examined in Odense, whereas all those where both partners were living in the eastern part of Denmark were examined in Copenhagen. With the exception of 39 twin pairs, both twins in a pair lived in the same geographic region of Denmark. Denmark is regarded as an area with mild to moderate iodine deficiency (21). Blood samples were drawn between 0800 and 0900 h after a 12-h fast and followed by a clinical examination. During the day the twins filled in additional questionnaires regarding their general health and lifestyle including questions regarding thyroid disease, smoking habits, and medicine intake.
In all, 1512 individuals (756 twin pairs) were examined. However, due to a missing blood sample (39 persons in 20 twin pairs) and self-reported thyroid disease (32 persons in 28 twin pairs), 71 subjects distributed in 48 pairs were excluded. Moreover, three subjects had overt biochemical hyperthyroidism defined as decreased serum TSH (serum TSH < 0.3 mU/liter) and increased serum free T4 (serum free T4 > 17.7 pmol/liter) and/or increased serum free T3 (serum free T3 > 7.4 pmol/liter). Additionally, 16 individuals had overt hypothyroidism defined as serum TSH greater than 4.0 mU/liter and serum free T4 less than 9.9 pmol/liter. These individuals and their co-twins were also excluded, leaving a total of 1380 healthy individuals or 690 twin pairs distributed in 284 monozygotic (MZ), 286 dizygotic same-sex (DZ), and 120 opposite-sex (OS) twin pairs. One individual in a DZ twin pair had a missing serum free T3 value. Therefore, only 285 complete twin pairs were available in the regression and twin analyses. In an attempt to identify additional cases of thyroid disease, information was sought from The National Discharge Register as a part of a record linkage between The Danish Twin Registry and The National Discharge Register. No additional cases were identified.
Individuals with chronic diseases such as back pain, asthma, migraine, etc. were included in the study population, but no twins were taking medicine that could affect the pituitary-thyroid axis. Mean age of the MZ twins were 37.1 yr (range 17–59 yr) and 36.9 yr (range 18–66 yr) and 37.2 yr (range 17–57 yr) for the DZ same-sex twins and DZ OS twins, respectively.
Written informed consent was obtained from all participants, and the study was approved by all regional Danish Scientific-Ethical Committees (case file 97/25 PMC).
Assays
Serum TSH was measured using a time-resolved fluoroimmunometric assay (AutoDELFIA hTSH Ultra Kit, Perkin-Elmer/Wallac, Turku, Finland). Reference range is 0.30–4.00 mU/liter. The intra- and interassay coefficient of variation (CV) at serum TSH concentrations between 0.046 and17.6 mU/liter is from 1.3–4.7% and 1.7–3.7%, respectively.
Serum free T4 and serum free T3 were determined using the AutoDELFIA FT4 and FT3 (Perkin-Elmer/Wallac), respectively. For free T4, the reference range is 9.9–17.7 pmol/liter, and for free T3 it is 4.3–7.4 pmol/liter. The intra- and interassay CV for free T4 at serum free T4 concentrations between 9.2 and 19.2 pmol/liter are from 1.3–2.0% and 3.9–5.4%, respectively. For free T3, the intra- and interassay CV at serum free T3 concentrations between 4.7 and 9.7 pmol/liter are from 3.9–5.0% and 2.9–4.2%, respectively. Thyroid peroxidase antibodies (TPOAbs) and thyroglobulin antibodies (TgAbs) were measured by solid-phase, two-step, time-resolved fluoroimmunoassays (AutoDELFIA TPOAb kit and hTgAb kit, respectively, Perkin-Elmer/ Wallac). Intra- and interassay CV for TPOAb and TgAb were from 3.2–8.4% and 3.8–10.1%, respectively, in the range of 50–155 U/ml. Values above 60 U/ml were regarded as positive for TPOAb as well as for TgAb. Subjects were considered as antibody positive if either of the tests was positive. Twin pairs were analyzed within the same run. All the serum samples were analyzed at the same laboratory in Odense. Zygosity was established by analysis of nine highly polymorphic restriction fragment length polymorphisms and microsatellite markers widely scattered through the genome with a PE Applied Biosystems (Foster City, CA) AmpFISTR Profiler Plus kit (22).
Statistical analyses
Before analysis, the frequency distribution of serum TSH, free T4, and free T3 was examined to detect any deviation from a Gaussian distribution. Only the serum TSH concentration was skewed, and after descriptive analyses but before twin analysis, serum TSH transformed using the natural logarithm (lnTSH). In the descriptive analyses, a modified Wilcoxon test was used testing the differences between the groups (23).
The potential effects of serum free T4, serum free T3, age, sex, examination place, thyroid antibody status, cigarette smoking (smokers were defined as former or current smokers, whereas nonsmokers were subjects who never had been smoking), and body mass index (BMI, defined as weight in kilograms divided by the square of height in meters) on the serum TSH concentration were analyzed using backward stepwise multiple regression analysis (with a limit for entry into the model of 0.05) and with cluster option (taking the dependence of the twin data into account). MZ, DZ, and OS twin pairs were used in the descriptive and the regression analyses.
The intraclass correlation coefficients and heritability estimates for lnTSH, free T4, and free T3 were calculated using the adjusted residuals resulting from the regression. Due to the gender-related difference in mean values of serum TSH, free T4, and free T3, only MZ and DZ same-sex pairs were used in these calculations. Including OS twin pairs would have violated some of the assumptions in the twin model used.
The analyses of serum free T4 and free T3 were performed in the same way. The adjustment for serum free T4 included lnTSH and free T3, whereas the adjustment for serum free T3 included lnTSH and free T4.
Quantitative genetic model fitting of twin data
The classical twin study compares phenotypic resemblances of MZ and DZ twins and is based on the assumption that MZ twins are genetically identical, whereas DZ twins share, on average, 50% of their genes (24, 25, 26, 27).
Structural equation modeling was used to estimate the magnitude of the genetic and environmental effects. This technique quantifies sources of individual differences by decomposing the observed phenotypic variance into genetic and environmental contributions (25, 27). The genetic contribution is divided into an additive (A) component (represents the influence of alleles at several gene loci acting in an additive manner) and a dominant (D) genetic component (represents intralocus interaction). The environmental contribution is divided into a shared/common environmental component (C) (refers to environmental factors that are affecting both twins in a pair in the same way and are a source of their similarity) and a unique (E) environmental component (the environmental factors that are not shared by twins in a pair and are a source of their dissimilarity). The latter component (E) also includes measurement error, which in this case is the analytical and intraindividual variation. The heritability can be defined as the proportion of the total variance attributable to additive genetic variation.
The significance of A, C, D, and E were tested by removing them sequentially in specific nested submodels and comparing them with the full model as described elsewhere (25). The following etiological models were fitted to the data: ACE, AE, CE, E, ADE, and DE. The selection of the best fitting model was based on a balance between goodness of fit and parsimony (25). The fit of the models were assessed by 
2 tests (a small 2 value and a high P value indicates a good agreement between the model and the observed data). The Akaike’s information criterion (AIC) (28), which corresponds to 2 - 2 x degrees of freedom, was used to assess parsimony (explain the observed data as well as possible with as few parameters as possible). Models with the lowest AIC were preferred.
Statistical software
The statistical analyses were carried out using STATA (version 7) (29). Level of significance was set to 0.05. Univariate quantitative genetic modeling was carried out using Mx software (30).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Table 1
shows the basic descriptive statistics for serum concentrations of TSH, free T4, and free T3 with respect to sex, examination place, antibody status, smoking, and zygosity. The serum TSH concentration was significantly higher in women than in men, whereas the serum free T4 and free T3 concentrations were significantly lower in women than in men. Individuals examined in Copenhagen (and therefore living in the eastern part of Denmark) had a significantly higher serum TSH concentration than the twins examined and living in the western part of Denmark. Individuals that were TgAb positive and/or TPOAb positive had a higher TSH level compared with antibody-negative persons. The TSH concentration was slightly higher in nonsmokers than in smokers. There was no difference in serum TSH, serum free T4, or free T3 between the MZ and DZ groups.
|
Table 2 presents the results of the regression analysis on the serum lnTSH, the serum free T4, and serum free T3 concentration, respectively. In general, the regression coefficients were small, and the SDs and variances were reduced only minimally (data not shown).
|
Figure 1 is a scatterplot of the transformed serum TSH levels for the 284 MZ twin pairs and 285 DZ same-sex twin pairs. In MZ twins, a clear clustering is seen at the line of identity, whereas DZ twins show a more scattered pattern.
|
. The correlation coefficients for the transformed TSH concentrations were significantly higher for MZ pairs than for DZ pairs (P < 0.00005). Because of the feedback loop between serum TSH and the free thyroid hormone concentrations, we also performed the analysis without adjusting for serum free T3 and free T4 concentrations in the regression analyses. The intraclass correlations were largely unchanged (data not shown). As evident from Fig. 1, the MZ correlations for serum free T4 and free T3 were also significantly higher than the corresponding DZ correlations, although the difference in correlations between the twin groups was less pronounced looking at serum free T3 (free T4, P < 0.00005; free T3, P = 0.0004). Subdividing the analysis according to sex did not change the overall results (data not shown). Biometric analyses
The results of the model-fitting analyses are presented in Table 3. Comparing the full models (ACE and ADE), the univariate model-fitting procedure revealed that the ADE model provided a slightly better fit to the data for lnTSH, free T4, and free T3. However, in the ADE model, the components were estimated with large confidence intervals in all three groups. Reduction to nested submodels indicated that the AE model was the best fitting model, because of small 2 statistics and high P values. This simpler model was favored because it provides a more simple and parsimonious explanation of the observed data. This is reflected by the low AIC values. Thus, additive genetic factors and unique environmental factors were the most important in explaining the interindividual differences in concentrations of serum TSH, free T3, and free T4. As evident from Table 3 and Fig. 2, additive genetic effects (i.e. the heritability) explained 64% (confidence interval, 57–70%) of the total variance in serum TSH concentration, whereas the estimate for the unique environmental (E) effect was 36% (confidence interval, 31–43%). Lack of adjusting for serum free T4 and free T3 concentrations did not influence this estimate (data not shown). Additive genetic effects explained 65% (58–71%) and 64% (57–70%) of the variance for serum free T4 and free T3 concentrations, respectively.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Given the standardized conditions and the large number of subjects in our study we are confident that our results are reliable. However, it is worth emphasizing that the heritability estimate is population specific. Nevertheless, within the range of sampling error, the estimates tend to be rather similar across populations (31), and a heritability estimate of serum lnTSH, T4, and T3 in other populations will most likely be within the 95% confidence intervals reported here. Based on just 15 MZ and 15 DZ male twin pairs and less sophisticated statistical methods, Meikle et al. (32) suggested that heritability accounted for 44% of the variation in plasma free T4. Their study size did not allow for estimating heritability of plasma T3 and TSH or the evaluation of possible environmental influences.
Whether the demonstrated genetic influence is related to the intrinsic capacity to synthesize and secrete TSH, the level of pituitary sensitivity to T4 and T3, the level of secretion of thyroid releasing hormone, the peripheral conversion of T4 to T3, or a combination of these cannot be explained by our study. A recent study by Peeters et al. (33) identifies specific polymorphisms located in the coding sequences of the iodothyronine deiodinases and the TSH receptor, which is associated with plasma thyroid hormone levels and TSH levels, respectively. These possible loci might be involved in the regulation of thyroid function, and these genetic contributions are a part of the genetic effect estimated in our study.
The feedback loop interconnecting the pituitary and the thyroid gland is well characterized. In stable phases, the serum TSH level is the most reliable indicator of thyroid status (34), and thus we have focused on the serum TSH level as the primary marker of thyroid function. The serum TSH concentration displays pulsatile (12, 35) and diurnal variation (1, 11, 12), the levels being highest at night and lowest in the daytime. These factors together with the short half-life of TSH (
60 min) (11) contribute to a substantial intraindividual variability. In our study, all blood samples were drawn in the morning between 0800 and 0900 h, and the twins in a pair were examined on the same day and at the same time, which means that the diurnal and the intraindividual variations are negligible.
We regard this sample of twins representative for the general population. We have no reason to believe that twins differ in thyroid function compared with singletons in the general population, because basic thyroid-related characteristics such as sex and regional differences in serum TSH level were similar to those in comparable studies of thyroid function in Denmark (7, 9), an area with mild to moderate iodine deficiency. The difference in TSH levels, with higher values in the eastern part of Denmark (i.e. subjects examined in Copenhagen), is best explained by a higher iodine intake in the eastern part as demonstrated in previous studies (36, 37). As found by others (7), we could demonstrate an age-related decline in serum TSH level in both regions of Denmark. This decline was more pronounced in the region with moderate iodine deficiency (data not shown) and is probably due to autonomous functioning thyroid tissue among the elderly in the most iodine-deficient region.
The potential effects of age, sex, examination place, antibody status, cigarette smoking, and BMI on the serum TSH level were examined by multiple linear regression analysis. However, these variables explained only a small fraction of the variation as reflected by a minimal reduction in SDs and variances. Similarly, we found small effects of the same parameters looking at the serum free T4 and serum free T3 concentrations. One explanation may be that we are trying to explain the variation with linear relations, whereas these relations may be nonlinear. As an example, we excluded persons with self-reported thyroid disease and biochemical thyroid disease. Thus, in our final study population we were looking only at a narrow range of thyroid function. The relationship between serum TSH and serum free T4 concentration was now almost nonexistent, reflected in a circular scatter when plotting the serum TSH concentration vs. free T4 concentration (data not shown). However, the regression analysis revealed a significant influence. Therefore, it seems wrong not adjusting for serum free T4 concentration as well as for the other parameters.
This study is a classical twin study and is built upon certain premises. The most important might be the assumption that MZ twin pairs are not treated more similarly than DZ twin pairs (25). The consequence is an overestimation of the genetic influence, because it may not be clear whether the effect on the phenotype is elicited by the genotype or is simply a matter of identical individuals being exposed to similar environmental factors. In addition, we assume that no gene-environment interaction exists (25). This assumption implies that an environmental factor has the same effect on the phenotype regardless of the genotype of the individual (25). This may not be the case.
In conclusion, our results demonstrate that interindividual differences in the pituitary-thyroid axis are predominantly determined by genetic factors in healthy subjects, and offer a possible explanation for the relatively narrow limits of thyroid hormones in the individual, as opposed to considerable interindividual variation. This is compatible with a unique thyroid function set-point predominantly under genetic influence. Because we do not know the exact set-point in the healthy individual, our findings imply that it may be difficult to restore normal thyroid function in the individual when for instance treating hypo- or hyperthyroidism (38). However, the possible clinical importance of our findings still remains unclarified.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: A, Additive component; AIC, Akaike’s information criterion; BMI, body mass index; C, shared/common environmental component; CV, coefficient(s) of variation; D, dominant genetic component; DZ, dizygotic same-sex; E, unique environmental component; lnTSH, the natural logarithm of the serum TSH level; MZ, monozygotic; OS, opposite-sex; TgAb, thyroglobulin antibody; TPOAb, thyroid peroxidase antibody.
Received September 22, 2003.
Accepted November 26, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Biondi and D. S. Cooper The Clinical Significance of Subclinical Thyroid Dysfunction Endocr. Rev., February 1, 2008; 29(1): 76 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. S. Hansen, T. H. Brix, I. Iachine, T. I. A. Sorensen, K. O. Kyvik, and L. Hegedus Genetic and environmental interrelations between measurements of thyroid function in a healthy Danish twin population Am J Physiol Endocrinol Metab, March 1, 2007; 292(3): E765 - E770. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. J. de Jong, R. P. Peeters, T. den Heijer, W. M. van der Deure, A. Hofman, A. G. Uitterlinden, T. J. Visser, and M. M. B. Breteler The Association of Polymorphisms in the Type 1 and 2 Deiodinase Genes with Circulating Thyroid Hormone Parameters and Atrophy of the Medial Temporal Lobe J. Clin. Endocrinol. Metab., February 1, 2007; 92(2): 636 - 640. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. P Peeters, W. M van der Deure, and T. J Visser Genetic variation in thyroid hormone pathway genes; polymorphisms in the TSH receptor and the iodothyronine deiodinases. Eur. J. Endocrinol., November 1, 2006; 155(5): 655 - 662. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R THOMAS ZOELLER Collision of Basic and Applied Approaches to Risk Assessment of Thyroid Toxicants. Ann. N.Y. Acad. Sci., September 1, 2006; 1076: 168 - 190. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G Brabant, P Beck-Peccoz, B Jarzab, P Laurberg, J Orgiazzi, I Szabolcs, A P Weetman, and W M Wiersinga Is there a need to redefine the upper normal limit of TSH? Eur. J. Endocrinol., May 1, 2006; 154(5): 633 - 637. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Gumieniak, S. Hurwitz, T. S. Perlstein, U. C. Ngumezi, P. N. Hopkins, X. Jeunemaitre, and G. H. Williams Aggregation of High-Normal Thyroid-Stimulating Hormone in Hypertensive Families J. Clin. Endocrinol. Metab., November 1, 2005; 90(11): 5985 - 5990. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Knudsen, P. Laurberg, L. B. Rasmussen, I. Bulow, H. Perrild, L. Ovesen, and T. Jorgensen Small Differences in Thyroid Function May Be Important for Body Mass Index and the Occurrence of Obesity in the Population J. Clin. Endocrinol. Metab., July 1, 2005; 90(7): 4019 - 4024. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. S. Hansen, T. H. Brix, F. N. Bennedbaek, S. J. Bonnema, K. O. Kyvik, and L. Hegedus Genetic and Environmental Causes of Individual Differences in Thyroid Size: A Study of Healthy Danish Twins J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2071 - 2077. [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 |
