Epidemiology of metabolic syndrome and its components in Chinese patients with a range of thyroid-stimulating hormone concentrations

Abstract

Objective

To investigate the relationship between thyroid stimulating hormone (TSH) concentration and the risks of developing metabolic syndrome and its components.

Methods

A total of 10,140 residents of the Yunyan district of Guiyang (Guizhou, China) who were ≥40 years old were selected by cluster random sampling between May and August 2011, of whom 5692 were eligible. TSH concentration and indices of metabolic syndrome were documented at baseline and 3 years later. Participants were allocated to a euthyroid (TSH 0.55–4.78 mIU/L) or high TSH concentration (TSH >4.78 mIU/L) group. Patients with overt hypothyroidism or were undergoing treatment for hypothyroidism were excluded.

Results

The crude and adjusted prevalences of metabolic syndrome were 39.9% and 33.9% in the euthyroid group and 44.3% and 37.5% in the high TSH group, respectively. Binary logistic regression analysis revealed a positive correlation between a high TSH concentration at baseline and the cumulative incidence of metabolic syndrome during follow up.

Conclusions

High TSH is associated with a higher risk of developing metabolic syndrome or one of its components; therefore, people with a high TSH concentration should be screened regularly to permit the early identification of metabolic syndrome and followed up thoroughly.

Introduction

Metabolic syndrome consists of a combination of central obesity, hyperglycemia (diabetes and/or impaired glucose tolerance), hypertension, high serum triglyceride concentration, and low serum high-density lipoprotein (HDL) concentration, which are associated with higher risks of heart disease, stroke, and type 2 diabetes mellitus. It has also been shown that metabolic syndrome is a risk factor for aberrant cardiovascular metabolism and early arteriosclerosis.^1–8 Thus, metabolic syndrome represents a major emerging public health challenge globally.^9,10 The easy availability of energy-dense foods, coupled with a sedentary lifestyle, has significantly increased the prevalence of this condition.^11–13 According to data published by the International Diabetes Federation, about one-quarter of the world’s population has metabolic syndrome.¹⁴ In addition, numerous clinical studies have demonstrated that the presence of metabolic syndrome is associated with three-fold higher incidence of cardiovascular disease and a three-fold higher mortality rate from this disease.^14–17 Furthermore, metabolic syndrome has recently been recognized as a diabetogenic state;¹⁸ the incidence of type 2 diabetes mellitus has been found to be five times higher in patients with metabolic syndrome.¹⁴ Therefore, both clinicians and the wider public have stakes in establishing deeper understanding of the pathophysiology and improving the treatment of metabolic syndrome.

Thyroid disorders are some of the most common endocrine diseases in the world, with more than 300 million people experiencing thyroid dysfunction.¹⁹ Forty-two million people are estimated to have thyroid disorders in India, and the prevalence of thyroid disorders in the Korean population has also significantly increased, according to a national population-based study.^20,21 In addition, some previous studies have suggested that thyroid disorders are positively related to atherosclerotic heart disease, because these disorders can lead to alterations in cardiovascular hemodynamics.^22,23 Thyroid stimulating hormone (TSH) has become a subject of much attention because it is a critical endocrine regulator of thyroid function, and a reduction in thyroid function can directly increase the risk of cardiac dysfunction and atherosclerosis.^24–26

Recent studies have suggested a significant overlap between thyroid disorders and metabolic syndrome.^27–29 It has been reported that high TSH is associated with a higher prevalence of metabolic syndrome, not only in patients with subclinical hypothyroidism, but also in euthyroid individuals with TSH concentrations within the normal range.^30,31 A cross-sectional study of participants with normal thyroid function indicated that high serum TSH is associated with high serum triglyceride concentrations in people with metabolic syndrome in a specific sex and age range.³² Therefore, it is crucial to elucidate the relationship between thyroid function, TSH, and metabolic syndrome. A high TSH concentration may induce changes in the hypothalamus–pituitary–leptin axis, which would cause resistance to thyroid hormones. This would lead to increases in body weight, waist circumference, and body mass index, resulting in central obesity.³³ To investigate this possibility, we recruited patients with a range of TSH concentrations to investigate the relationships between TSH and the prevalence and cumulative incidence of metabolic syndrome and its components (including central obesity, low HDL cholesterol, and hypertriglyceridemia) over a 3-year period. The findings of this study should provide useful information regarding the prevention and treatment of metabolic syndrome.

Materials and methods

Research participants

Recruitment

Residents of the Yunyan district of Guiyang (Guizhou, China) were selected by cluster random sampling between May and August 2011. The present study was part of the baseline survey for the Risk Evaluation of cAncers in Chinese diabeTic Individuals: a lONgitudinal (REACTION) study, which was conducted in 259,657 adults aged ≥40 years in 25 communities across mainland China from 2011 to 2012.^34–37 The history, anthropometric parameters, and biochemical test results were recorded for each participant. The study protocol was approved by the ethics committee of Guizhou Medical University Affiliated Hospital (2011 (51)) and each participant provided their written informed consent prior to the start of the study.

Follow-up protocol

After evaluation, participants with metabolic syndrome at baseline were excluded from the follow-up component of the study. After the study period (38.55 ± 1.59 months from May to August of 2011 to July to December of 2014), all the anthropometric and biochemical parameters that were evaluated at baseline were re-evaluated.

Classification of the participants

According to the reference range for TSH at Guizhou Medical University Affiliated Hospital, all the eligible participants were allocated to one of two groups: a euthyroid group with TSH concentrations within the normal range for the laboratory (0.55–4.78 mIU/L), and a high TSH group with concentrations >4.78 mIU/L.

Definitions used

Metabolic syndrome was defined as per the International Diabetes Federation criteria that were published in 2005.⁹ The definition of metabolic syndrome was as follows: 1. Central obesity (required condition), waist circumference: Chinese: male ≥90 cm, female ≥80 cm; European descent: male ≥94 cm, female ≥80 cm; South Asian: male ≥90 cm, female ≥80 cm; 2. Two or more of the following components: (1) TG ≥1.7 mmol/L, or treatment for hypertriglyceridemia; (2) Low HDL-C: male <1.04 mmol/L, female <1.3 mmol/L, or treatment for dyslipidemia; (3) blood pressure ≥130/85 mmHg, or a diagnosis of hypertension and antihypertensive treatment; and (4) fasting plasma glucose (FPG) ≥5.6 mmol/L or type 2 diabetes. A history of smoking was defined as the smoking of at least one cigarette per day for at least 6 months,³⁸ and a history of excess alcohol intake was defined as the drinking of at least 30 g of ethanol weekly for at least a 1-year period. An exercise status of “active” was defined as the participation in exercise for at least 30 mintes on at least 3 days a week.³⁹

Exclusion criteria

Individuals with TSH concentrations below the normal range; patients with thyroid disease who were already under treatment; people who were taking medications that could affect thyroid function; patients with malignant tumors, critical illnesses, extreme exhaustion, cachexia, or liver/renal insufficiency; women who were pregnant or within 1 year of childbirth; individuals with psychiatric illness; and people unwilling to complete the questionnaire, or undergo physical examination or blood tests, were excluded from the study.

Investigation methods

General information

A questionnaire was administered to all the participants by professionally trained endocrinologists during a face-to-face interview. The basic information collected in the questionnaire included the participant’s name, sex, age, smoking history, history of alcohol intake, and exercise status. In addition, their past medical history, including of hypertension, diabetes, thyroid disorders, and dyslipidemia, was collected through the questionnaire during the interview.

Physical examination

All the participants were examined by endocrinologists and their height, body mass, waist circumference, and blood pressure were recorded. During the blood pressure measurement, the participants were asked to sit in a quiet room for 5 minutes, then the mean of three readings obtained using an electronic blood pressure monitor (Omron Model HEM-725FUZZY, Omron Company, Dalian, China) was recorded. The cuff was placed around an upper arm and there were 3-minute intervals between each measurement. For the measurement of waist circumference, the participants were required to stand upright with their feet 30 to 40 cm apart, while wearing underwear and with an unfastened belt, and breathing normally. A measuring tape was used to measure the circumference at the midpoint between the line connecting the anterior superior iliac crest and the inferior edge of the twelfth rib. The measuring tape was held against the skin, but without pressure being applied, and the measurements were made to the nearest 0.1 cm. Body mass index (BMI) was calculated using the equation: BMI = body mass (kg)/height (cm)².

Biochemical tests

Serum lipid profile analysis was performed on venous samples that were collected from each participant after fasting for ≥10 hours using an Abbott automatic biochemical immunoassay analyzer (Architect-ci16200; Abbott Laboratories, Abbott Park, IL, USA). Oral glucose tolerance testing (OGTT) was performed on all the participants with no history of diabetes. Plasma glucose concentrations were measured using the hexokinase method. For participants with a history of diabetes, the FPG and 2-hour post-meal glucose concentrations were determined after the consumption an 80-g instant noodle cake. TSH concentrations were measured using an Advia Centaur XP automated immunoassay analyzer (Siemens Healthineers Global, Malvern, PA, USA). This assay had an intra-assay coefficient of variation of 1.70% to 4.69% and an inter-assay coefficient of variation of 0.88% to 4.20%.

Statistical analysis

All the data were double-entry verified using EpiData software (www.epidata.dk). Normally distributed continuous data are expressed as means ± standard deviations, and categorical data are expressed as frequencies and percentages. Comparisons of the differences in rates and proportions between the high TSH and euthyroid groups were performed using the chi-square test, and Student’s t test or analysis of variance (ANOVA) were used to identify significant differences in mean values between two or among multiple groups, respectively (SPSS Statistics for Windows, Version 17.0, SPSS Inc., Chicago, IL, USA). The prevalences and cumulative incidences of metabolic syndrome were normalized by age and sex using the 2010 National Census data for China and the risk factors for metabolic syndrome were identified using binary logistic regression analysis. Differences were considered statistically significant when P<0.05.

Results

Participants

A total of 10,140 residents were selected, of whom 9618 qualified for the present study. After evaluation, 3926 participants with metabolic syndrome at baseline were excluded from the follow-up component of the study. Therefore, 5692 participants were followed up for 3 years. A total of 4292 participants completed the full 3-year period of the study, which yielded a completion rate of 75.40%.

Comparison of the baseline characteristics of the euthyroid and high TSH groups

Table 1 shows a comparison of the baseline characteristics of the two groups. The serum triglyceride, low-density lipoprotein (LDL) cholesterol, and HDL cholesterol concentrations were all significantly higher in the high TSH group than in the euthyroid group. The high TSH group contained a relatively larger number of women. The 2-hour post-meal plasma glucose concentration of the high TSH concentration group was lower than that of the euthyroid group. In addition, the high TSH group contained significantly fewer participants with histories of excess alcohol consumption and/or smoking. However, there were no significant differences in the waist circumference, BMI, systolic blood pressure, diastolic blood pressure, FPG, homeostatic model assessment of insulin resistance, or exercise history between the two groups at baseline. Finally, individuals >65 years old accounted for 20.7% of the enrolled participants.

Table 1.

Comparison of the characteristics of participants in the euthyroid and high TSH groups at baseline.

Variable	Euthyroid	High TSH	P-value
	Group (n = 7706)	Group (n = 1912)
Thyroid stimulating hormone concentration (mIU/L)/Range (mIU/L)	2.47 (1.75–3.34)	6.22 (5.41–8.11)	<0.001
Age (years)	58.30 ± 8.14	58.37 ± 8.30	0.71
Sex (female/male)	5437/2269	1565/347	<0.001
Waist circumference (cm)	84.30 ± 8.88	84.26 ± 8.51	0.864
Body mass index (kg/m2)	24.03 ± 3.22	24.09 ± 3.22	0.502
Systolic blood pressure (mmHg)	122.33 ± 18.92	122.24 ± 18.96	0.848
Diastolic blood pressure (mmHg)	77.24 ± 11.01	77.09 ± 11.08	0.573
Fasting plasma glucose (mmol/L)	6.21 ± 1.62	6.15 ± 1.62	0.117
2-hour post-meal blood glucose (mmol/L)	8.78 ± 3.65	8.56 ± 3.59	0.021
Triglycerides (mmol/L)	1.73 ± 1.34	1.90 ± 1.41	<0.001
Low-density lipoprotein cholesterol (mmol/L)	2.62 ± 0.88	2.73 ± 0.92	<0.001
High-density lipoprotein cholesterol (mmol/L)	1.23 ± 0.37	1.25 ± 0.38	0.049
Homeostatic model assessment of insulin resistance	2.44 ± 2.29	2.47 ± 2.67	0.597
History of smoking, n (%)	1636/21.23%	244 (12.76%)	<0.001
History of excessive alcohol consumption, n (%)	822/10.67%	134 (7.01%)	<0.001
History of exercise, n (%)	1649/21.40%	405 (21.18%)	0.836

Data are mean ± SD or n (%). Student’s t test.

TSH, thyroid stimulating hormone.

Relationships between indices related to metabolic syndrome and its components and serum TSH concentration

Of the total of 9618 randomly selected participants, 3926 were found to have metabolic syndrome. The total crude and standardized prevalences of metabolic syndrome were 40.8% and 34.5%, respectively, in the complete sample at baseline (Table 2). The crude prevalences of metabolic syndrome, central obesity, low HDL cholesterol concentration, and high triglyceride concentration were higher in the high TSH group than in the euthyroid group at baseline, before the exclusion of the individuals who had metabolic syndrome (P < 0.05). The crude prevalences of hypertension and high FPG were similar in the two groups. However, the standardized prevalence of metabolic syndrome (P < 0.01) and its components (central obesity, P < 0.01; low HDL cholesterol concentration, P = 0.045; high triglyceride concentration, P <0.01) tended to be higher in the high TSH group.

Table 2.

Comparison of the crude and standard prevalences of metabolic syndrome and its components in the euthyroid and high TSH groups at baseline.

Parameter	Euthyroid group (n=7,706)			High TSH group (n=1,912)			P-value
	Number of affected individuals	Crude prevalence (%)	Standardized prevalence (%)	Number of affected individuals	Crude prevalence (%)	Standardized prevalence (%)
Metabolic syndrome	3079	39.96	33.90	847	44.30	37.56	<0.01
Central obesity	4393	57.01	49.28	1193	62.40	52.53	<0.01
Low high-density lipoprotein cholesterol	2447	31.75	26.84	653	34.15	27.75	0.045
High triglycerides	1831	23.76	20.93	598	31.28	27.31	<0.01
Hypertension	2118	27.49	23.20	545	28.50	24.02	0.373
High fasting plasma glucose	2981	38.68	32.37	771	40.32	34.36	0.188

Chi-square test.

TSH, thyroid stimulating hormone.

Relationship between the cumulative incidence of metabolic syndrome and its components and serum TSH concentration

Over the course of the 3-year follow-up period, 966 participants out of the 4292 who completed the study were diagnosed with metabolic syndrome, which yielded cumulative crude and standardized 3-year incidences of metabolic syndrome of 22.5% and 20.6%, respectively (Table 3). The cumulative crude incidence of low HDL cholesterol was significantly higher in the high TSH group than in the euthyroid group (P = 0.014), whereas those of the other metabolic syndrome components were similar in the two groups. The incidence of metabolic syndrome was also found to increase with increasing serum TSH concentration when the participants were placed into four groups according to their TSH values (0 to 2.49, 2.5 to 4.99, 5.0 to 7.49, and 7.5 to 10.0 mIU/L) (Figure 1).

Table 3.

Comparison of the crude and standard cumulative incidences of metabolic syndrome and its components in the euthyroid and high TSH groups.

Parameter	Euthyroid group (n = 7,706)			High TSH group (n = 1,912)			P-value
	Number ofaffected individuals	Crude incidence (%)	Standardized incidence (%)	Number of affected individuals	Crude incidence (%)	Standardized incidence (%)
Metabolic syndrome	768	22.01	20.22	198	24.69	23.20	0.101
Central obesity	986	28.25	25.11	252	31.42	26.50	0.074
Low high-density lipoprotein cholesterol	562	16.10	15.86	158	19.70	27.91	0.014
High triglycerides	1042	29.86	33.05	253	31.55	30.31	0.347
Hypertension	1313	37.62	37.56	291	36.28	38.24	0.480
High fasting plasma glucose	1177	33.72	35.25	248	30.92	35.28	0.129

P values refer to comparisons of the crude incidence rates between the groups. Chi-square test.

TSH, thyroid stimulating hormone.

Figure 1.

Incidence of metabolic syndrome increases alongside increases in serum TSH concentration.

TSH, thyroid stimulating hormone.

To identify factors that influence the cumulative incidence of metabolic syndrome in the high TSH group, binary logistic regression analysis was performed, with metabolic syndrome incidence as the dependent variable and all the other characteristics as covariates (Table 4). We found a significant positive correlation between the cumulative incidence of metabolic syndrome and TSH concentration.

Table 4.

Identification of risk factors for metabolic syndrome using binary logistic regression analysis.

Parameter	β coefficient	Standard deviation	Wald chi-square statistic	Odds ratio	95% confidence interval	P-value
Female sex	3.135	0.416	56.714	22.997	10.169–52.007	<0.001
Age (years)	0.013	0.016	0.630	1.013	0.982–1.045	0.427
TSH (mIU/L)	0.047	0.023	4.222	1.048	1.002–1.097	0.04
Body mass index (kg/m²)	0.082	0.060	1.846	1.098	0.976–1.234	0.174
Waist circumference (cm)	0.101	0.022	20.484	1.106	1.059–1.156	<0.001
Fasting plasma glucose (mmol/L)	0.088	0.137	0.412	1.092	0.835–1.429	0.521
2-hour post-meal blood glucose (mmol/L)	0.081	0.053	2.297	1.084	0.977–1.203	0.13
Systolic blood pressure (mmHg)	0.029	0.009	10.367	1.030	1.012–1.049	0.001
Diastolic blood pressure (mmHg)	0.020	0.015	1.715	1.020	0.990–1.052	0.19
High-density lipoprotein cholesterol (mmol/L)	−3.646	0.554	43.368	0.026	0.009–0.077	<0.001
Triglycerides (mmol/L)	0.415	0.114	13.341	1.514	1.212–1.892	<0.001
Low-density lipoprotein cholesterol (mmol/L)	0.300	0.141	4.525	1.350	1.024–1.781	0.033
Homeostatic model assessment of insulin resistance	0.207	0.079	6.913	1.230	1.054–1.436	0.009
History of smoking	−0.129	0.376	0.118	0.879	0.421–1.837	0.731
History of excessive alcohol consumption	1.221	0.512	5.688	3.391	1.243–9.250	0.017
History of exercise	0.044	0.235	0.035	1.045	0.659–1.657	0.851

Discussion

Relationships between the prevalence of metabolic syndrome and its components and TSH

A study performed in the United States between 2003 and 2012 showed the total standardized prevalence of metabolic syndrome to be 33%, having increasing from 32.9% (95% confidence interval: 31.6, 34.2) in 2003 to 2004 to 34.7% (95% confidence interval: 33.5, 36.0) in 2011 to 2012.⁴⁰ In the present study, the total crude and standardized prevalences of metabolic syndrome at baseline were 40.8% and 34.5%, respectively, which were similar to the prevalence estimates in the United States.

The results of a long-term, large-scale follow-up study conducted in Taiwan suggested that the prevalence of metabolic syndrome is positively related to serum TSH concentration: a significant difference in the prevalence of metabolic syndrome was found in participants with TSH concentrations in the highest quartile, compared with those with concentrations in the lowest quartile.⁴¹ In addition, a retrospective study conducted in the United States found that patients with hypothyroidism had a 2.6-fold higher prevalence of metabolic syndrome than euthyroid individuals.⁴² Similarly, a study of the relationship of metabolic syndrome and its components with thyroid dysfunction in Algerian patients found that high TSH was associated with a higher incidence of metabolic syndrome, and that TSH concentration positively correlated with the prevalence of metabolic syndrome.⁴³ Furthermore, a study of 949 euthyroid postmenopausal women found that the prevalence of metabolic syndrome in women in the highest TSH concentration quintile was about 1.95 times higher than that for women in the lowest quintile.⁴⁴ A Korean study also showed a significant increase in the prevalence of metabolic syndrome alongside an increase in TSH concentration.⁴⁵ Another study of 2760 young euthyroid women with normal TSH concentrations suggested that women with TSH concentrations at the higher end of the normal range had a two-fold higher risk of metabolic syndrome than those with values at the lower end, even after adjustment for age and BMI.⁴⁶ Finally, a 3-year follow-up study of 5998 Korean people >18 years of age concluded that a high TSH concentration could be considered a predictor of the future development of metabolic syndrome.⁴⁷ Thus, a high TSH concentration, even when within the normal range, may represent a risk factor for metabolic syndrome.

In the present study, the high TSH group had a higher crude prevalence of metabolic syndrome, central obesity, low HDL cholesterol, and high triglycerides than the euthyroid group (P < 0.05). This implies that people with TSH concentrations above the normal range may be at a higher risk of developing metabolic syndrome and its components. Thyroid hormones are significant regulators of lipid biosynthesis and metabolism, and thyroid dysfunction can therefore result in lipid abnormalities.⁴⁸ Previous studies have shown that a high circulating TSH concentration can cause alterations to the hypothalamic–pituitary–leptin axis, and such alterations, along with the partial inactivation of TSH proteins, could further cause resistance to thyroid hormones. Resistance to thyroid hormone action would likely retard nutrient metabolism, leading to increases in body weight, waist circumference, and BMI, and ultimately central obesity.³³ In addition, it has been reported that TSH concentration is positively correlated with plasma cholesterol ester transfer protein concentrations. High TSH might increase the accumulation of cholesterol ester transfer protein, which could in turn accelerate the inter-conversion of HDL cholesterol, LDL cholesterol, and extremely low LDL cholesterol. This might result in greater transport of cholesterol from peripheral tissues to hepatocytes in HDL cholesterol, which would reduce the HDL cholesterol concentration.⁴⁹ In addition, TSH could increase triglyceride accumulation in hepatocytes by activating the TSH receptor/peroxisome proliferator-activated receptor/adenylate-activated protein kinase/sterol regulatory element-binding protein 1c signaling pathway.⁵⁰ Previous studies have shown a positive correlation between TSH and vascular resistance in the small renal arteries and a negative correlation with renal blood flow and glomerular filtration rate. Therefore, a high TSH concentration might induce pathophysiological changes in the renal circulation, and thereby increase the risk of high blood pressure.⁵¹ Furthermore, a high TSH concentration might have an impact on glucose metabolism by affecting insulin sensitivity, or by altering the release of thyroxine (T4), which has an important influence on glucose and energy homeostasis.⁵²

Relationships between the incidence of metabolic syndrome and its components and TSH

In the present study, the cumulative crude and standardized incidences of metabolic syndrome in all the participants, over a 3-year period, were 22.5% and 20.6%, respectively. The cumulative crude and standardized incidences of the euthyroid group were 22.0% and 20.2%, respectively, but they were 24.7% and 23.2% in the high TSH group. The cumulative crude incidence of low HDL cholesterol was significantly higher in the high TSH group than in the euthyroid group (27.9% vs. 15.9%, P = 0.014). However, both groups exhibited similar crude incidences of metabolic syndrome, central obesity, high triglyceride concentration, hypertension, and high FPG. Because there were more women than men in the high TSH group, the crude incidence was corrected for the participants’ sex and age. After this adjustment, the standardized incidences of metabolic syndrome and its components were higher in the high TSH group, with the exception of the serum triglyceride concentration, which did not differ between the groups.

A 6-year prospective cohort study²⁷ of an elderly population, with a mean age of 73.6 years, showed an independent positive correlation between TSH and triglyceride, and a negative correlation with HDL cholesterol. These relationships were similar to those found in the present study. However, the same study did not show relationships between TSH and metabolic syndrome and its components (including central obesity, hypertension, and high FPG). This might have been because of the small sample size or the advanced age of the participants, because many previous studies of elderly populations have shown that older people have higher circulating TSH concentrations⁵³.

A 3-year population-based cohort study showed that both overt and subclinical hypothyroidism, especially in older people, are associated with metabolic syndrome.⁵⁴ This finding was mirrored by that of a hospital-based cross-sectional study conducted in Nepal,⁵⁵ and another cross-sectional study conducted in India showed that overt hypothyroidism is associated with a higher risk of metabolic syndrome.⁵⁶ Finally, a population-based survey conducted in Italy showed a positive correlation between circulating TSH and triglyceride concentrations, and positive relationships between free thyroxine (FT4) and the components of metabolic syndrome.⁵⁷

Binary logistic regression analysis showed that the odds ratio (OR) of developing metabolic syndrome over a 3-year period in the high TSH group was 1.047 (P = 0.043, 95% confidence interval: 1.001, 1.095). Insulin resistance is considered to be a primary underlying cause of metabolic syndrome. To eliminate confounding of the results by the presence of insulin resistance, logistic regression analysis was performed using homeostasis model assessment of insulin resistance as a covariate, but there was a similar OR for metabolic syndrome in the high TSH group (OR = 1.048, P = 0.040, 95% confidence interval 1.002, 1.097), which indicates that the TSH concentration is associated with metabolic syndrome independently of insulin resistance. There were significantly larger numbers of men than women in the high TSH group, and because most of the female participants were not smokers or drinkers, the incidences of metabolic syndrome in participants with a history of smoking and excessive alcohol intake were below the expected levels, given that most of the smokers and drinkers were in the euthyroid group. However, because the comparison of the incidences between the two groups was normalized using the sex of the participants, these variations did not affect the overall results of this study.

The study had a few limitations. First, we did not have the serum total T3 and T4 concentrations of the participants. Second, we did not study the relationships of markers of thyroid immunity, such as anti-thyroid peroxidase antibody titer, with the incidences of metabolic syndrome and its components. Third, there were significantly fewer male than female participants, which could also have had a bearing on the results. Finally, further studies are needed to establish whether high TSH is a cause or a consequence of metabolic syndrome and the associated inflammatory milieu.

Conclusions

A 3-year prospective cohort study of 5692 participants, with a completion rate of 75.40%, revealed a higher crude prevalence of metabolic syndrome in participants with high serum TSH concentrations than those who were euthyroid. Similarly, participants with high TSH concentrations had a higher cumulative standardized incidence of metabolic syndrome than euthyroid participants. Furthermore, binary logistic regression analysis showed that TSH concentration was positively related to the incidences of metabolic syndrome and its components. Our findings indicate that there is an association between high TSH and the development of metabolic syndrome. However, further studies are required to determine whether treatment of high TSH concentration is capable of preventing the development or reducing the severity of incident metabolic syndrome and its components.