Parity Is Not Related to Autoimmune Thyroid Disease in a Population-Based Study of Japanese-Brazilians

Abstract

Background:

It has been suggested that the female preponderance for autoimmune thyroid disease might be associated with hormonal differences, abortion, and fetal microchimerism. Findings emerging from the few epidemiological studies on this matter, however, are controversial. In this study, we investigated the hypothesis whether parity, abortion, and the use of estrogens are associated with a higher risk for thyroid autoimmunity.

Methods:

This cross-sectional population-based study examined 675 women from a Japanese-Brazilian population aged above 30 years. Thyroid peroxidase antibodies (TPOAbs), thyroglobulin antibodies (TgAbs), thyrotropin, and free T₄ were measured by immunofluorimetric assays. Urinary iodine concentration was measured using a colorimetric method. Data were analyzed in logistical regression models to obtain the odds ratio (OR) and 95% confidence intervals.

Results:

TPOAbs and TgAbs were present in 11.6% and 13.6% of women, respectively. After adjustment for age, smoking, and urinary iodine concentration, the OR for positive TPOAb (OR, 1.22 [95% confidence interval, 0.73–2.02]) and for positive TgAb (OR, 1.01 [0.63–1.62]) among women who had one or more parities did not differ from those who had never given birth. In addition, we found no association between the presence of thyroid antibodies and previous abortions or the use of estrogens.

Conclusions:

Parity, abortion, and the use of estrogens are not associated with thyroid autoimmunity in this population. These findings reinforce previous reports that advocated against a key role of fetal microchimerism in the pathogenesis of autoimmune thyroid disease.

Introduction

A utoimmune thyroid disease (AITD) is one of the most common autoimmune disorders, affecting around 5% of iodine-sufficient populations (1,2). Interaction of genetic susceptibility and environmental factors appears to be of fundamental importance in initiating the process of thyroid autoimmunity (3). It has been postulated that 79% of the AITD susceptibility can be attributed to genetic factors, whereas 21% can be attributed to environmental factors (4). However, there is no clear evidence of causality, and the mechanisms by which environmental factors trigger thyroid autoimmunity in genetically predisposed individuals remain unclear (3,4).

Similar to other human autoimmune disorders, AITD preferentially develops in women of childbearing age and appears to be modulated by pregnancy (5). Possible reasons for this and for the female preponderance for AITD might include X-chromosomal factors, the potential effect of exogenous estrogens use on the immune system, and abortions (3,4,6). In addition, a phenomenon known as microchimerism, defined as a bi-directional trafficking of maternal and fetal cells during pregnancy, has emerged to explain the female preponderance in AITD (7–8). The persistence of microchimeric fetal cells in maternal tissues, such as thyroid, skin, and pancreas (9,10), is believed to play a role in the pathogenesis of a number of autoimmune diseases (e.g., AITD, systemic sclerosis, and type I diabetes mellitus) (8 –10).

Whether the last factor is natural or pathogenic and whether it may be involved in the pathogenesis of AITD is still not clear. Studies of fetal-maternal microchimerism in the thyroid have documented a higher prevalence of fetal cells in association with Hashimoto's thyroiditis and Graves' disease when compared with nonautoimmune thyroid disorders (11 –13). However, recent data emerging from epidemiological studies show conflicting results (14 –16). Two previous population-based studies (14,15) found no association of pregnancy and parity with thyroid antibodies or thyroid dysfunction, suggesting no role of fetal microchimerism in AITD, whereas another study demonstrated that parity was a potential risk factor for AITD (16).

In the present study, we estimated the influence of previous parities, abortions, and the use of estrogens on the risk of thyroid autoimmunity in an entire community.

Materials and Methods

The present population (n = 675) represents the female portion of a nonmixed Japanese-Brazilian population living in Bauru (Human Development Index, HDI 0.825; www.ipeadata.com.br Accessed May 20, 2010), State of São Paulo, Brazil. Detailed descriptions of this survey have been previously reported (17,18). Briefly, the entire population ≥30 years of age (n = 1751) was invited to participate, and 1330 (76%) agreed. We excluded individuals who self-reported previous thyroid disease or were taking thyroid drugs (n = 47), those using drugs that could affect thyroid function such as amiodarone, lithium, or corticosteroids (n = 6) and individuals for whom serum was not available for testing for thyroid antibodies (n = 5). Of the remaining 1272 (95.6%) individuals, 675 women took part in the present analysis.

The participants answered a standardized questionnaire including information concerning family and personal history of thyroid disease, concomitant medications, smoking habits, pregnancy, parity, and previous or present use of estrogens.

Fasting blood and urine samples were collected and stored at −80°C for analysis of thyrotropin (TSH), free T₄ (FT₄), thyroid antibodies, and urinary iodine concentration (UIC). Serum TSH was measured in duplicate by a sensitive immunofluorimetric assay (Wallac-Delphia). The functional sensitivity of the assay was 0.05 mU/L, and the reference range was 0.4–4.5 mU/L. Serum FT₄ was measured using a competitive immunoassay (Wallac–Delphia, Finland), and the normal reference range was 0.7–1.5 ng/dL.

Serum antithyroid peroxidase antibodies (TPOAbs) were determined by a time-resolved immunofluorimetric assay (AutoDelfia^® TPOAb; PerkinElmer Life and Analytical Sciences, Wallac Oy). Tests were considered positive when values were above 35 U/mL, which corresponded to the functional sensitivity given by the manufacturer. Serum antithyroglobulin antibody (TgAb) concentrations were measured using an in-house immunofluorimetric assay (19) with a sensitivity of 40 UI/mL and an interassay error less than 5% along the standard curve.

UIC was measured in early morning urine samples using a modified colorimetric semiautomated method that was previously described (20). The detection limit for the method was 10 μg/L with a normal range between 100 and 299 μg/L.

Euthyroidism was defined as both serum TSH and FT₄ within the normal reference range, subclinical hyperthyroidism as a TSH level less than 0.45 mU/L with normal FT₄ level, overt hyperthyroidism as a TSH level less than 0.1 mU/L in the presence of an FT₄ level above the upper limit of the normal reference range, subclinical hypothyroidism as a TSH level above 4.5 mU/L with a normal FT₄ level, and overt hypothyroidism as a TSH level above 4.5 mU/L with an FT₄ level below the inferior limit of the normal reference range or a TSH concentration above 20 mU/L.

This study was approved by the Ethics Committee of Escola Paulista de Medicina, Federal University of São Paulo, and written informed consent was obtained from all participants.

Statistical analysis

All statistical analyses were performed using SAS statistical software version 9.1 (SAS Institute Inc.). The assumed level of significance was p < 0.05 (2-tailed). Parity was analyzed both as a binary variable (0, ≥1) and as a continuous variable (0, 1, 2, and ≥3). Continuous variables were described with the mean and standard deviation and nominal variables with the absolute (n) or relative (%) frequencies according to the number of parities, abortions, and use of estrogen. The Kruskal-Wallis test was used for continuous data followed by the Mann-Whitney U test when it was significant. Variables without a normal distribution as determined by the Kolmogorov-Smirnov test underwent logarithmic transformations before statistical analysis. Frequencies were compared by the chi-square test or Fisher test when more than 20% of expected values were lower than 5 or any expected value was less than 1. Logistic regression models, adjusted for age, UIC and smoking, were used to determine the association of parities, abortions, and the use of estrogens with the presence of thyroid antibodies; regression models were also used to determine the odds ratio and 95% confidence interval.

Results

Among the 675 women from this population, 367 (54.4%) had never given birth and 308 (45.6%) mentioned previous childbirths (range, 1–12). One or more abortions were disclosed by 35 (5.2%) participants, and 67 (9.9%) had previously or were presently using estrogens (oral contraceptives or hormone replacement therapy). The characteristics of these women are shown in Table 1. Women who had never given birth tended to be older (p < 0.0001) and were more likely to be smokers (p < 0.04) and to have a lower UIC (p = 0.005) than those who had at least one parity. Conversely, there were no apparent differences in the prevalence of goiter or the mean TSH and FT₄ levels among women according to their obstetrical history.

Table 1.

Sociodemographic and Thyroid-Related Characteristics, According the Number of Parities, Abortions, and the Use of Estrogens

	No. of parities			No. of abortions			Use of estrogens
Parameter	0 (n = 367)	≥1 (n = 308)	p-Value	0 (n = 640)	≥1 (n = 35)	p-Value	No (n = 608)	Yes (n = 67)	p-Value
Age, years	60.5 ± 12.7	52.9 ± 10.3	<0.0001	57.2 ± 12.4	53.7 ± 10.3	0.07	57.4 ± 12.6	54.0 ± 8.2	0.01
Smokers, n (%)	52 (14.2)	28 (9.1)	0.04	74 (11.6)	6 (17.1)	0.32	75 (12.3)	5 (7.5)	0.24
Goiter, n (%)	48 (13.1)	43 (14)	0.73	88 (13.8)	3 (8.6)	0.6	80 (13.2)	11 (16.4)	0.45
Urinary iodine concentration, μg/L	190 ± 187	220 ± 201	0.005	210 ± 198	185 ± 149	0.23	200 ± 182	210 ± 196	0.96
Thyrotropin, mU/L	2.9 ± 13.7	2.6 ± 4.7	0.7	2.7 ± 10.7	3.6 ± 7.4	0.54	2.1 ± 1.9	2.9 ± 11.2	0.6
Free T₄, ng/dL	1.1 ± 0.7	1.08 ± 0.5	0.054	1.09 ± 0.6	1.0 ± 0.2	0.17	1.05 ± 0.37	1.09 ± 0.6	0.2
TPOAb, U/mL	27.7 ± 127.9	24.4 ± 90.5	0.23	15.5 ± 52.6	26.7 ± 114.8	0.007	25.6 ± 109.3	31.9 ± 137.3	0.8
TgAb, UI/mL	42.8 ± 194.2	56.3 ± 274.9	0.9	28.9 ± 91.5	49.9 ± 240.2	0.018	47.9 ± 222.8	58.8 ± 324.6	0.95
Positive TPOAb, n (%)	38 (10.4)	38 (12.3)	0.42	72 (11.3)	3 (8.6)	0.78	71 (11.7)	5 (7.5)	0.3
Positive TgAb, n (%)	51 (14.0)	41 (13.3)	0.8	87 (13.6)	4 (11.4)	1.0	84 (13.8)	8 (11.9)	0.9
Positive TPOAb and TgAb, n (%)	26 (7.2)	21 (6.9)	0.87	43 (6.7)	3 (8.6)	0.72	43 (7.1)	4 (6.1)	1.0
Positive TPOAb and/or TgAb, n (%)	63 (17.2)	58 (18.8)	0.6	116 (18.1)	4 (11.4)	0.29	112 (18.4)	9 (13.4)	0.31

Data presented as mean ± SD, unless noted otherwise.

TPOAb, thyroid peroxidase antibody; TgAb, thyroglobulin antibody.

The median UIC for all women was 210 μg/dL, and no statistical difference was found among thyroid disease categories. The greatest proportion of the participants had euthyroidism (80.1%), and the most prevalent thyroid dysfunctions were subclinical hypothyroidism (10.4%) and subclinical hyperthyroidism (6.9%). Overt hyperthyroidism (1.6%) was more prevalent than overt hypothyroidism (1%). We found no association of thyroid status with parity, abortion, or use of estrogens (Table 2).

Table 2.

Thyroid Dysfunctions According to the Number of Parities, Abortions, and the Use of Estrogens

Parameter	Euthyroidism (n = 541)	Overt hyperthyroidism (n = 11)	Subclinical hyperthyroidism (n = 46)	Overt hypothyroidism (n = 7)	Hypothyroidism (n = 70)	p-Value
No. of parities
0	302 (55.8)	3 (27.3)	24 (52.2)	4 (57.1)	34 (48.6)	0.31
≥1	239 (44.2)	8 (72.7)	22 (47.8)	3 (42.9)	36 (51.4)
No. of abortions
0	531 (94.8)	11 (100.0)	43 (93.5)	6 (85.7)	65 (92.9)	0.45
≥1	26 (4.8)	—	3 (6.5)	1 (14.3)	5 (7.1)
Use of estrogens
No	487 (90.0)	10 (90.9)	42 (91.3)	7 (100.0)	62 (88.6)	0.97
Yes	54 (10.0)	1 (9.1)	4 (8.7)	—	8 (11.4)

Data presented as no. of subjects (%).

p indicates χ ² test or Fisher test.

Thyroid antibodies (TPOAb and/or TgAb) were positive in 17.9% of the participants (Table 1). Both tests were positive in 47 (6.9%) individuals, TPOAb was positive in 76 (11.3%) individuals, and TgAb was positive in 92 (13.6%) individuals. There was no statistical difference in the prevalence of positive thyroid antibodies between women who had never been given birth and those with one or more previous parity. We repeated the analysis using number of parity as a continuous variable (0, 1, 2, ≥3), but again no relationship was found between parity and thyroid autoimmunity (data not shown).

In addition, no difference was found in the prevalence of positive thyroid antibodies for participants who reported a previous abortion or for those exposed to estrogens (present or previous) through oral contraceptives or hormone replacement therapy (Table 1). However, when we analyzed these data by quantitative levels of thyroid antibodies rather than by dichotomous classification, we found significantly increased mean serum levels of both TPOAb and TgAb in women who suffered one or more abortions.

As shown in Table 3, the risk (odds ratio) for positive TPOAb and/or for positive TgAb was not different between the following groups: (a) women who had never given birth and those who had one or more previous childbirths; (b) women who had a previous abortion and those who never had an abortion; (c) women who reported present or previous use of estrogens and those who had not previously used estrogens. After adjustment for age, UIC, and smoking, the risk for positive thyroid antibodies (TPOAb and/or TgAb) remained unchanged among the groups.

Table 3.

Odds Ratios with 95% Confidence Intervals for the Association Between Positive Thyroid Antibodies and Previous Parity, Abortion, and Use of Estrogens

Parameter	≥1 Parities (reference: no parity)	≥1 Abortion (reference: no abortion)	Use of estrogens (reference: no use)
Positive TPOAb
Crude	1.21 (0.75–1.95)	0.73 (0.22–2.44)	0.61 (0.24–1.88)
Adjusted^a	1.22 (0.73–2.02)	0.71 (0.21–2.45)	0.59 (0.23–1.54)
Positive TgAb
Crude	0.95 (0.61–1.47)	0.81 (0.28–2.35)	0.85 (0.39–1.85)
Adjusted^a	1.01 (0.63–1.62)	0.82 (0.28–2.39)	0.86 (0.39–1.86)
Positive TPOAb and TgAb
Crude	0.95 (0.52–1.73)	1.28 (0.38–4.36)	0.84 (0.29–2.42)
Adjusted^a	0.94 (0.51–1.76)	0.79 (0.23–2.27)	1.2 (0.42–3.48)
Positive TPOAb and/or TgAb
Crude	1.11 (0.75–1.65)	0.63 (0.23–1.73)	1.39 (0.68–2.84)
Adjusted^a	0.84 (0.55–1.28)	0.57 (0.19–1.66)	0.68 (0.33–1.4)

Adjusted for age, smoking, and urinary iodine concentration.

Discussion

In this iodine-sufficient population of Japanese-Brazilians, we found no relationship between previous parity and the presence of thyroid antibodies (TPOAb and/or TgAb), thyroid dysfunction, or goiter. Our results are in close agreement with two previous larger population-based studies (14,15), in which previous pregnancy or parity were not risk factors for thyroid autoimmunity. Walsh et al. (14) examined available serum samples from 1045 of 2142 female participants in the context of the Busselton Health Study in Western Australia, which is considered an iodine-sufficient region. Hypo- and hyperthyroidism (including subclinical and overt dysfunctions) were prevalent in 9% and 4% of the participants, respectively, and thyroid antibodies (TPOAb and/or TgAb) were prevalent in 20%. After adjustment for age, no evidence was found for increased risk of thyroid autoimmunity or abnormal TSH with increasing number of reported pregnancies. Pedersen and colleagues (15) examined 3283 women randomly selected from the general population of Aalborg and Copenhagen as part of the Danish investigation of iodine intake and thyroid diseases (DanThyr). With median UICs of 45 μg/L (Aalborg) and 61 μg/L (Copenhagen), the overall prevalence for TPOAb and/or TgAb was 20.9%. Again, no association was found between the risk of having thyroid antibodies and a previous pregnancy, number of pregnancies, parity, previous abortion, or the use of estrogens.

The current study differs from the Study of Health in Pomerania (SHIP) (16), which identified a significant association between parity and both positive TPOAb and hypoechogenic thyroid ultrasound patterns in a population-based sample of 2156 women.

The reasons for the discrepancies between the SHIP study (16), previous studies (14,15), and this study are not clear but may include differences in population characteristics in terms of age, ethnicity, iodine intake, different assays used to determine thyroid antibodies, and different cut-offs used to define thyroid autoimmunity. The authors of the SHIP (16) study justified their different results by arguing that the Busselton (14) and DanThyr (15) studies were limited by relatively low response proportions of 64% and 51%, respectively, as compared with 68.8% in their study. In addition, they believed that the exclusion of women aged 31–39 years and 46–59 years might have led to an underestimation of the relationship between parity and AITD in the DanThyr study (15). They also commented that the study was performed in an area of mild-to-moderate iodine deficiency, whereas no exact data on iodine supply exist for the Busselton study (14). In fact, all of these arguments could have limited the ability of such studies (14,15) to detect an effect of parity on AITD. However, in the present study, in which no association was found between parity and AITD, the proportion of responders was approximately 76% (greater than that estimated in the SHIP study), and all women more than 30 years of age were invited to participate in the study; additionally, the median UIC found in our population was 210 μg/L, which is more than optimal according to current recommendations (21). The SHIP population was investigated during a time when the iodine supply was shifted from deficient to a sufficient supply (22). This is of particular importance, as previous studies have reported a transient or persistent increase in thyroid autoantibody titers after iodine prophylaxis (23 –26). Thus, it is not possible to exclude the possibility that the parity-related risk of AITD found in such a study (16) could have been influenced by changes in iodine supply during the study recruitment period, especially considering that the study design did not allow the evaluation of changes in thyroid autoimmunity by comparing the situation before and after the iodine supplementation program (22). Further, the thyroid antibody assays and cut-offs chosen to define positive titers of TPOAb differ between the studies. The SHIP study differentiated between elevated (>60 mIU/mL) and positive (>200 mIU/mL) TPOAb titers, whereas the other studies (14,15) and our study used a TPOAb cut-off of approximately 30–35 kIU/L.

The present study also differs from a more recent case-control study (27), in which both female and male twins from opposite-sex pairs had an increased frequency of thyroid antibodies compared with monozygotic pairs, indicating a potential role of microchimerism in developing thyroid autoimmunity. However, these findings should be cautiously interpreted, as reflected by the relatively wide confidence intervals, and because a number of other intrauterine factors such as number of placentas and birth weight may have influenced the results (27).

The female preponderance in AITD might also be attributed to hormonal influences, as there is limited evidence concerning the influence of the X chromosome in the pathogenesis of AITD (3). However, there are relatively few studies exploring potential relationship between the use of estrogens and thyroid autoimmunity, and the results are conflicting (3). The present study is fully in agreement with the DanThyr study (15), in having found no association between thyroid autoantibodies and exogenous estrogen use.

Associations between abortions and thyroid autoimmunity have been reported in some studies and a meta-analysis (6,28). The reasons for these associations might include the copresence of thyroid autoimmunity with other autoimmune diseases, direct actions of thyroid antibodies on placenta, and a greater age of women with positive thyroid antibodies and mild thyroid failure (6,28). In fact, in the current study, women who reported at least one previous abortion had significantly higher mean serum thyroid antibodies levels compared with those who had never experienced abortions (Table 1). Despite this, no increased risk for thyroid autoimmunity was found in women who reported previous abortions, which is also in line with the DanThyr study (15).

In this study, age, smoking, and UIC differed significantly between women with one or more parities and those with no previous parity (Table 1). We do not have a reasonable explanation for age differences among the groups, but search for a causal role of parity on thyroid autoimmunity remained insignificant even after a logistic regression analysis adjusted for these variables.

The major strength of our study lies in the fact that an entire iodine-sufficient population (and not a sample) was invited to participate, with a higher proportion of agreement than the three previous population-based studies in this area. A weakness of this study is the cross-section design, which is common to all previously published studies. Also, we cannot exclude the possibility of interaction effects across the different predictor variables, as it was not possible to do this kind of analysis in this study. Finally, we cannot guarantee that our findings can be generalized due to our selected population of Japanese-Brazilians.

In summary, parity, abortion, and the use of estrogens are not associated with AITD in the Japanese-Brazilian female population. These findings reinforce previous reports that argued against a key role of fetal microchimerism in the pathogenesis of AITD.

Footnotes

Acknowledgments

We are grateful to the Japanese-Brazilian population and to The Japanese-Brazilian Diabetes Study Group. The authors gratefully acknowledge Dr. Heloisa Villar for helping in the collection of clinical and thyroid data. We thank Sirlei Siriane for the statistical assistance. We are also grateful to Gilberto Furuzawa and Patricia Hiroka for technical assistance and to Angela Faria for administrative assistance. This study was supported by a grant from the São Paulo State Research Foundation (Fundação de Amparo à Pesquisa do Estado de São Paulo), grant 06/59737-9. R.M.B.M. is an investigator of the Brazilian Research Council and of the Fleury Group.

Disclosure Statement

The authors have no conflict of interest.

References

Wang

, Crapo

. 1977. The epidemiology of thyroid disease and implications for screening. Endocrinol Metab Clin North Am, 26:189–219.

Tumbridge

, Vanderpump

MPJ

. 2000. Population screening for autoimmune thyroid disease. Endocrinol Metab Clin North Am, 29:239–253.

Prummel

, Strieder

, Wiersinga

. 2004. The environment and autoimmune thyroid disease. Eur J Endocrinol, 150:605–618.

Sgarbi

, Maciel

RMB

. 2009. Pathogenesis of autoimmune thyroid diseases. Arq Bras Endocrinol Metab, 53:5–14.

Philips

, Lazarus

, Butland

. 1990. The influence of pregnancy and reproductive span on the occurrence of autoimmune thyroiditis. Clin Endocrinol (Oxf), 32:301–306.

Prummel

, Wiersinga

. 2004. Thyroid autoimmunity and miscarriage. Eur J Endocrinol, 150:751–755.

Ando

, Davies

. 2003. Postpartum autoimmune thyroid disease: the potential role of fetal microchimerism. J Clin Endocrinol Metab, 88:2965–2971.

Waldorf

KMA

, Nelson

. 2008. Autoimmune disease during pregnancy and the microchimerism legacy of pregnancy. Immunol Invest, 37:631–644.

Koopmans

, Hovinga

ICLK

, Baelde

, Harvey

, Heer

, Bruijn

, Bajema

. 2008. Chimerism occurs in thyroid, lung, skin and lymph nodes of women with sons. J Reprod Immunol, 78:68–75.

10.

Rust

, Bianchi

. 2009. Microchimerism in endocrine pathology. Endocr Pathol, 20:11–16.

11.

Ando

, Imaizumi

, Graves

, Unger

, Davies

. 2002. Intrathyroidal fetal microchimerism in Graves' disease. J Clin Endocrinol Metab, 87:3315–3320.

12.

Renné

, Ramos Lopez

, Steimle-Grauer

, Ziolkowski

, Pani

, Luther

, Holzer

, Encke

, Wahl

, Bechstein

, Usadel

, Hansmann

, Badenhoop

. 2004. Thyroid fetal male microchimerism in mothers with thyroid disorders: presence of y-choromosomal immunofluorescence in thyroid-infiltrating lymphocytes is more prevalent in Hashimoto's thyroiditis and Graves' disease than in follicular adenomas. J Clin Endocrinol Metab, 89:5810–5814.

13.

Klintschar

, Immel

, Kehlen

, Schwaiger

, Mustafa

, Mannweiler

, Regauer

, Kleiber

, Hoang-Vu

. 2006. Fetal microchimerism in Hashimoto's thyroiditis: a quantitative approach. Eur J Endocrinol, 154:237–241.

14.

Walsh

, Bremner

, Bulsara

, O'Leary

, Leedman

, Feddema

, Michelangeli

. 2005. Parity and the risk of autoimmune thyroid disease: a community-based study. J Clin Endocrinol Metab, 90:5309–5312.

15.

Bülow Pedersen

, Laurberg

, Knudsen

, Jørgensen

, Perrild

, Ovesen

, Rasmussen

. 2006. Lack of association between thyroid autoantibodies and parity in a population study argues against microchimerism as a trigger of thyroid autoimmunity. Eur J Endocrinol, 154:39–45.

16.

Friedrich

, Schwarz

, Thonack

, Jonh

, Wallaschofski

, Volzke

. 2008. Association between parity and autoimmune thyroiditis in a general female population. Autoimmunity, 41:174–180.

17.

Gimeno

, Ferreira

, Franco

, Hirai

, Matsumura

, Moisés

. 2002. Prevalence and 7-year incidence of Type II diabetes mellitus in a Japanese-Brazilian population: an alarming public health problem. Diabetologia, 45:1635–1638.

18.

Sgarbi

, Matsumura

, Kasamatsu

, Ferreira

, Maciel

RMB

. 2010. Subclinical thyroid dysfunctions are independent risk factors for mortality in a 7.5-year follow-up: the Japanese-Brazilian thyroid study. Eur J Endocrinol, 162:569–577.

19.

Vieira

JGH

, Tachibana

, Fonseca

RMG

, Nishida

, Maciel

RMB

. 1996. Development of a immunofluorimetric assay for the measurement of anti-thyroglobulin antibodies. Arq Bras Endocrinol Metab, 40:232–237.

20.

Esteves

, Kasamatsu

, Kunii

, Furuzawa

, Vieira

JGH

, Maciel

RMB

. 2007. Development of a semi-automated method for measuring urinary iodine and its application in epidemiological studies in Brazilian schoolchildren. Arq Bras Endocrinol Metab, 51:1477–1484.

21.

Delange

, de Benoist

, Burgi

. the ICCIDD Working Group. 2002. Determining median urinary iodine concentration that indicates adequate iodine intake at population level. Bull World Health Organ, 80:633–636.

22.

Völzke

, Lüdemann

, Robinson

, Spieker

, Schwahn

, Kramer

, John

, Meng

. 2003. Prevalence of undiagnosed thyroid disorders in a previously iodine-deficient area. Thyroid, 13:803–810.

23.

Weaver

, Nishiyami

, Burton

, Batsakis

. 1966. Surgical thyroid disease: a survey before and after iodine prophylaxis. Arch Surg, 92:796–801.

24.

Laurberg

, Pedersen

, Knudsen

, Ovesen

, Andersen

. 2001. Environmental iodine intake affects the type of nonmalignant thyroid disease. Thyroid, 11:457–469.

25.

Zois

, Stavrou

, Svarna

, Seferiadis

, Tsatsoulis

. 2006. Natural course of autoimmune thyroiditis after elimination of iodine deficiency in northwestern Greece. Thyroid, 16:289–293.

26.

Teng

, Shan

, Teng

, Guan

, Li

, Teng

, Jin

, Yu

, Fan

, Chong

, Yang

, Dai

, Yu

, Li

, Chen

, Zhao

, Shi

, Hu

, Mao

, Gu

, Yang

, Tong

, Wang

, Gao

, Li

. 2006. Effect of iodine intake on thyroid diseases in China. N Engl J Med, 354:2783–2793.

27.

Brix

, Hansen

, Kyvik

, Hagedus

. 2009. Aggregation of thyroid autoantibodies in twins from opposite-sex pairs suggests that microchimerism may play a role in the early stages of thyroid autoimmunity. J Clin Endocrinol Metab, 94:4439–4443.

28.

Toulis

, Goulis

, Venetis

, Kolibianakis

, Negro

, Tarlatzis

, Papadimas

. 2010. Risk of spontaneous miscarriage in euthyroid women with thyroid autoimmunity undergoing IVF: a meta-analysis. Eur J Endocrinol, 162:643–652.