Age-related change in thyroid-stimulating hormone: a cross-sectional study in healthy euthyroid population

Abstract

This cross-sectional study aimed to examine changes in thyroid-stimulating hormone (TSH) concentration over age in China and investigate relationship between TSH and risk factors for cardiovascular disease (CVD) among euthyroid subjects. TSH, free triiodothyronine (FT3), free thyroxine (FT4), blood lipid, and glucose were measured. 7,693 individuals were subdivided into different age groups. Associations between TSH and CVD risk factors [age, body mass index (BMI), systolic and diastolic blood pressure, total cholesterol (TC), triglycerides, low density lipoprotein-cholesterol (LDL-C) and high density lipoprotein-cholesterol (HDL-C) and fasting plasma glucose (FPG)] were evaluated with Pearson correlation analysis. Results showed that 2.5th percentile for TSH was consistent across age groups, whereas 97.5th percentile increased in subjects older than 40 years with upper limit being 6.83 mIU/L in subjects aged 60–69 years and 8.07 mIU/L in those older than 70 years. The age-specific upper limits reclassification rate was higher in all age bands as compared to the common cut-off value. TSH was positively associated with age, SBP, DBP, TC and LDL-C and negatively with FT3 and FT4. Serum TSH within new reference range had a linear correlation with SBP, TC and LDL-C in subjects aged <60 years. There were no significant differences in BMI, blood pressure, lipid profile or FPG among subjects 60–69 and older than or equal to 70 years. Elevated TSH within new reference range is associated with risk factors for CVD in subjects aged <60 years. Thus, there might be age-related difference in the relationship between CVD risk factors and elevated serum TSH.

THE SERUM thyroid stimulating hormone (TSH) is the most sensitive indicator in the evaluation of thyroid function. Subclinical hypothyroidism and subclinical hyperthyroidism are defined as abnormally increased or decreased serum TSH level with a normal serum free thyroxine (FT4) level. Serum TSH level is affected by age, genetics, gender, thyroid autoantibodies, iodine intake and thyroid disease [1-5]. Therefore, the National Academy of Clinical Biochemistry recommends that the serum TSH reference range should be established after rigorously screening euthyroid volunteers without evidence of thyroid disease [6]. The National Health and Nutrition Examinations Survey III (NHANES III) in the United States suggests that, in the absence of thyroid disease, serum TSH concentration in adults increases with age [7]. In 14,376 participants with no history of thyroid disease, free of overt hyperthyroidism or hypothyroidism and with negative tests for thyroid peroxidase and thyroglobulin antibodies in the NHANES III study, there was a progressive shift in the TSH curve towards higher values with age [8]. This has been confirmed by other studies [6, 9]. It has been suggested that the age-specific TSH reference range may minimize the under-diagnosis of subclinical hypothyroidism in young adults and prevent over-diagnosis in the elderly [5, 8]. However, questions remain about whether TSH level increases with age in Chinese.

Thyroid dysfunction is associated with the pathogenesis of cardiovascular disease (CVD) and coronary heart disease related death [10]. Dyslipidemia in hypothyroidism, also in some cases of subclinical hypothyroidism, is associated with an increased risk for CVD [10-12]. Available studies have shown that the elevated serum levels of total cholesterol (TC), low density lipoprotein-cholesterol (LDL-C) and triglycerides (TG) in individuals with overt hypothyroidism [13], is reversible after levothyroxine therapy [14]. Recently, the association between TSH and CVD has been investigated among euthyroid individuals. A prospective study shows that higher TSH and lower FT4 within the reference range are positively and linearly associated with CVD mortality in women [15]. Ruhla et al. [16] found that euthyroid subjects with TSH in the upper normal range have higher TG level and an increased incidence of metabolic syndrome. However, other studies fail to show the relationship between TSH and CVD risk factors in euthyroid individuals [17, 18]. Thus, it is necessary to investigate the age-specific reference intervals [19].

In this study, we investigated whether TSH level increased with age in Chinese population and examined the relationship between TSH and CVD risk factors.

Materials and Methods

Subjects

A total of 7,900 Chinese who received routine physical examination in the Nanjing Drum Tower Hospital, were recruited between January 1, 2016 and December 31, 2016. To exclude the influence of thyroid disease and other confounders, following exclusion criteria were used in this study: self-reported thyroid disease or goiter; overt hypothyroidism (TSH >20 mIU/L or TSH >4.2 mIU/L and FT4 <12 pmol/L); overt hyperthyroidism (TSH <0.27 mIU/L and FT4 >22 pmol/L); use of T4 or antithyroid drugs; abnormal thyroid ultrasound or positive thyroid antibodies in those subjects with enlarged thyroid at palpation; use of amiodarone, lithium carbonate, carbamazepine or phenytoin (because of the effects of these drugs on thyroid function or thyroid function tests); and discordant results suggestive of pituitary dysfunction or assay interference (e.g. increased TSH with increased FT4). In addition, pregnant or breast-feeding women were excluded. Finally, 7,693 individuals were recruited as the reference population.

Anthropometric and biochemical measurements

Height and weight were measured, and body mass index (BMI) was calculated as weight (kg) divided by square meters of height (m²). Blood pressure (BP) was measured using an automated device (A & D Co., Japan) in a sitting position, after 10-min rest.

Venous blood samples were taken in the morning after overnight fasting and processed for the detection of TC, TG, LDL-C, high density lipoprotein cholesterol (HDL-C), fasting plasma glucose (FPG), serum TSH, FT4 and free triiodothyronine (FT3). Serum levels of TC, TG, LDL-C, HDL-C and FPG were measured using enzymatic methods with Olympus reagents by automated spectrophotometry on the Olympus AU5400 system (Olympus Corporation, Tokyo, Japan). Serum TSH, FT4 and FT3 were measured by chemiluminescence immunoassay (ADVLA Centaure, Siemens, USA). The reference ranges were as follows: TSH, 0.27–4.2 mIU/L; FT4, 12–22 pmol/L; FT3, 3.1–6.8 pmol/L.

Statistical analysis

Normal distribution was assessed with the Kolmogorov-Smirnov test. Continuous variables with normal distribution are expressed as means ± standard deviation (SD). TSH concentration was log-transformed due to its abnormal distribution. Statistical methodologies from published guidelines were applied to define TSH reference interval between 2.5th percentile and 97.5th percentile.

Comparisons of the median of TSH for different age groups were calculated by Kruskal-Wallis one-way ANOVA in a single gender. The correlation between TSH and clinical variables was evaluated using Pearson’s correlation analysis. The association between TSH and CVD risk factors was assessed in different age groups (<60 years, 60–69 years and ≥70 years). Subjects were divided into five groups according to serum TSH concentration in the new reference range, and then ANOVA was used to compare means among groups. Statistical analysis was performed using SPSS version 21.0 for Windows (USA). A value of p < 0.05 was considered statistically significant.

Results

Baseline characteristics of study population

Of 7,693 subjects, there were 5,397 men and 2,296 women. The mean age was 51.8 ± 14.7 years (range, 21–97 years) and the mean BMI was 24.5 ± 3.0 kg/m² (range, 14.5–43.0 kg/m²). The mean lipid profile was as follows: TC, 4.9 ± 0.9 mmol/L (range, 2.1–8.3 mmol/L), TG, 1.6 ± 1.0 mmol/L (range, 0.3–13.8 mmol/L), LDL-C, 2.6 ± 0.7 mmol/L (range, 0.01–5.2 mmol/L), and HDL-C, 1.2 ± 0.3 mmol/L (range, 0.5–2.9 mmol/L). The mean serum TSH, FT3 and FT4 were 2.6 ± 1.5 mIU/L (range, 0.28–18.1 mIU/L), 4.8 ± 0.7 pmol/L (range, 2.32–6.78 pmol/L) and 17.2 ± 2.3 pmol/L (range, 8.62–22.0 pmol/L), respectively (Table 1).

Table 1 Baseline characteristics of 7,693 euthyroid subjects

Characteristic	Range	Mean ± SD
Male/Female		5,397/2,296
Age (years)	21–97	51.8 ± 14.7
BMI (kg/m2)	14.5–43.0	24.5 ± 3.0
SBP (mmHg)	75.0–183.0	127.1 ± 16.4
DBP (mmHg)	39–121.0	80.5 ± 11.2
FPG (mmol/L)	3.55–25.9	5.3 ± 1.1
TC (mmol/L)	2.1–8.3	4.9 ± 0.9
TG (mmol/L)	0.3–13.8	1.6 ± 1.0
LDL-C (mmol/L)	0.01–5.2	2.6 ± 0.7
HDL-C (mmol/L)	0.5–2.9	1.2 ± 0.3
TSH (mIU/L)	0.28–18.1	2.6 ± 1.5
FT3 (pmol/L)	2.32–6.78	4.8 ± 0.7
FT4 (pmol/L)	8.62–22.0	17.2 ± 2.3

Note: SD, standard deviation; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TC, total cholesterol; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; TSH, thyroid stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxine. Data are presented as mean ± SD.

The possible association between serum TSH level and clinical variables was assessed using Pearson’s correlation analysis. As shown in Table 2, there was positive correlation of TSH with age (r = 0.096, p = 0.000), SBP (r = 0.095, p = 0.000), DBP (r = 0.039, p = 0.001), TC (r = 0.035, p = 0.002) and LDL (r = 0.027, p = 0.017). In addition, serum TSH was negatively associated with FT3 and FT4 (p < 0.001). However, TSH had no relationships with BMI, FPG, TG and HDL-C (p > 0.05).

Table 2 Correlations between serum TSH, FT4 and clinical variables

	TSH		FT4
	r	p value	r	p value
Age	0.096	0.000	–0.115	0.000
BMI	0.017	0.135	–0.033	0.005
SBP	0.095	0.000	–0.022	0.054
DBP	0.039	0.001	0.049	0.000
FBG	0.020	0.084	0.018	0.118
TC	0.035	0.002	–0.033	0.004
TG	0.013	0.263	–0.051	0.000
LDL-C	0.027	0.017	–0.004	0.709
HDL-C	0.021	0.061	–0.002	0.829

Note: BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TC, total cholesterol; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; TSH, thyroid stimulating hormone; FT4, free thyroxine.

There was negative correlation of FT4 with age (r = –0.115, p = 0.000), BMI (r = –0.033, p = 0.005), TC (r = –0.033, p = 0.004) and TG(r = –0.051, p = 0.000). In addition, serum FT4 was positively associated with DBP (r = 0.049, p = 0.000). However, FT4 had no relationship with FPG, SBP, LDL-C or HDL-C (p > 0.05) (Table 2).

Change in serum TSH and FT4 level with age

To determine the age-related serum TSH level, 7,693 participants were divided into six age groups with an interval of 10 years (20–29, 30–39, 40–49, 50–59, 60–69 and ≥70 years). Table 3 shows the median, 2.5th and 97.5th percentiles for TSH in different age groups. The median TSH concentration was 2.25, 2.10, 2.15, 2.26, 2.35 and 2.39 mIU/L in 20–29, 30–39, 40–49, 50–59, 60–69 and ≥70 years groups, respectively. In addition, the serum TSH level increased with age and the highest median TSH was found in ≥70 years group. The 2.5th percentile for TSH was 0.77, 0.74, 0.82, 0.88, 0.81 and 0.78 mIU/L in 20–29, 30–39, 40–49, 50–59, 60–69 and ≥70 years groups, respectively, but 60–69 and >70 years groups showed no change with an increasing age. The 97.5 percentile for TSH increased steadily with age, being 5.07 mIU/L in 20–29 years group, 5.22 mIU/L in 30–39 years group, 6.03 mIU/L in 40–49 years group, 6.26 mIU/L in 50–59 years group, 6.83 mIU/L in 60–69 years group and 8.07 mIU/L in ≥70 years group.

Table 3 The 2.5th and 97.5th percentiles for TSH analyzed by age bands and by gender (n = 7,693)

Age (years)	n	MedianTSH (mIU/L)	2.5th percentile	97.5th percentile
Females
20–29	388	2.58	0.85	5.19
30–39	357	2.21	0.67	5.80
40–49	470	2.27	0.79	6.09
50–59	449	2.39	0.94	6.64
60–69	437	2.47	0.72	6.36
≥70	195	2.82	0.74	9.96
All	2,296	2.38	0.78	6.29
Males
20–29	250	1.90	0.72	4.39
30–39	587	2.05	0.79	5.01
40–49	1,220	2.11	0.83	6.01
50–59	1,416	2.23	0.84	5.89
60–69	1,291	2.30	0.83	7.00
≥70	633	2.31	0.79	7.85
All	5,397	2.19	0.82	6.23
Total
20–29	638	2.25	0.77	5.07
30–39	944	2.10	0.74	5.22
40–49	1,690	2.15	0.82	6.03
50–59	1,865	2.26	0.88	6.26
60–69	1,728	2.35	0.81	6.83
≥70	828	2.39	0.78	8.07
All	7,693	2.25	0.81	6.25

Furthermore, we analyzed the age-related serum TSH level in different age groups of women and men, respectively. The serum TSH level also increased with age (p < 0.001) and the highest median TSH was also found in ≥70 years group of women and men, respectively. The 2.5th percentile for TSH showed no change in different age groups of women and men, respectively. The 97.5th percentile for TSH in women and men increased steadily with age, which was consistent with the total population (Table 3).

According to the age-related upper TSH level, subjects would be reclassified from abnormal to normal, being 6.9% in 20–29 years group, 3.7% in 30–39 years group, 7.0% in 40–49 years group, 8.0% in 50–59 years group and 9.7% in 60–69 years group. The highest reclassification rate was observed in the older participants, being 13.5% in those aged over 70 years.

The median FT4 concentration was significantly decreased with age (p < 0.001) and the lowest median FT4 was found in ≥70 years group in total or in a single gender. In the total population, the 2.5 percentile and 97.5 percentile for FT4 decreased steadily after 30–39 years group. However, the 2.5th percentile and 97.5 percentile for FT4 in a single gender did not change significantly with age (Supplementary Table 1).

Association between TSH and CVD risk factors in different age groups

To explore the influence of age on the relationship between TSH and CVD risk factors (<60, 60–69 and ≥70 years). The subjects were divided into five subgroups based on the quintiles of TSH within age-related reference range.

In <60 years group, 60–69 years group and ≥70 years group, the reference range was 0.82–3.56 mIU/L, 0.81–6.83 mIU/L and 0.78–8.07 mIU/L, repectively. In <60 years group, subjects were further subdivided into five groups according to the TSH quintiles: Q1, 0.82–1.47 mIU/L; Q2, 1.48–1.92 mIU/L; Q3, 1.93–2.47 mIU/L; Q4, 2.48–3.32 mIU/L; Q5, 3.33–5.62 mIU/L. As shown in Table 4, there were significant increases in SBP, TC and LDL-C with the elevation of serum TSH (all p < 0.05), while serum BMI, DBP, FPG, TG, and HDL-C were comparable among five subgroups (all p > 0.05). Especially, compared with Q1 group (the lowest TSH quintile), subjects in Q5 group (the highest TSH quintile) had significantly higher SBP, TC and LDL-C.

Table 4 CVD risk factors in different TSH quintile groups among subjects aged <60 years

	Q1 (0.82–1.47)	Q2 (1.48–1.92)	Q3 (1.93–2.47)	Q4 (2.48–3.32)	Q5 (3.33–5.62)	p
n	913	1,022	1,030	1,024	898
Age (years)	42.97 ± 9.54	43.97 ± 9.74	44.09 ± 9.51	43.62 ± 9.78	43.91 ± 10.01	0.091
BMI (kg/m2)	23.78 ± 4.58	23.84 ± 4.76	23.87 ± 4.93	23.75 ± 4.73	23.43 ± 5.57	0.323
SBP (mmHg)	122.13 ± 15.31	123.90 ± 15.48	122.87 ± 14.87	123.70 ± 15.88	124.58 ± 15.13	0.009
DBP (mmHg)	78.74 ± 11.35	79.75 ± 11.23	79.18 ± 11.56	79.80 ± 11.78	79.70 ± 10.62	0.199
FPG (mmol/L)	5.14 ± 1.12	5.16 ± 1.07	5.18 ± 1.07	5.15 ± 1.06	5.13 ± 0.77	0.826
TC (mmol/L)	4.80 ± 0.83	4.88 ± 0.84	4.89 ± 0.84	4.90 ± 0.88	4.92 ± 0.85	0.029
TG (mmol/L)	1.52 ± 1.04	1.55 ± 1.10	1.54 ± 1.02	1.52 ± 1.09	1.52 ± 0.99	0.966
LDL-C (mmol/L)	2.57 ± 0.65	2.63 ± 0.64	2.65 ± 0.66	2.64 ± 0.70	2.59 ± 0.65	0.036
HDL-C (mmol/L)	1.27 ± 0.35	1.25 ± 0.32	1.25 ± 0.33	1.26 ± 0.34	1.28 ± 0.36	0.307

Note: BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TC, total cholesterol; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.

Subjects aged 60–69 years were also subdivided into five groups according to serum TSH concentration (Q1, 0.81–1.49 mIU/L; Q2, 1.50–2.04 mIU/L; Q3, 2.05–2.66 mIU/L; Q4, 2.67–3.56 mIU/L; Q5, 3.57–6.83 mIU/L). Serum TSH quintile had no significant relationship with age, BMI, SBP, DBP, FPG and lipid profiles (Table 5). As shown in Table 6, participants aged ≥70 years were also subdivided into 5 groups according to the quintiles of serum TSH (Q1, 0.78–1.50 mIU/L; Q2, 1.51–2.11 mIU/L; Q3, 2.12–2.77 mIU/L; Q4, 2.78–3.90 mIU/L; Q5, 3.91–8.07 mIU/L). Age, BMI, SBP, DBP, FPG and lipid profiles had no relationship with the quintiles of TSH.

Table 5 CVD risk factors in different TSH quintile groups among subjects aged 60–69 years

	Q1 (0.81–1.49)	Q2 (1.50–2.04)	Q3 (2.05–2.66)	Q4 (2.67–3.56)	Q5 (3.57–6.83)	p
n	310	343	348	342	301
Age (years)	62.71 ± 2.97	62.97 ± 2.83	62.64 ± 2.80	62.82 ± 2.86	62.76 ± 2.85	0.634
BMI (kg/m2)	24.48 ± 2.82	24.73 ± 2.46	24.59 ± 2.63	24.83 ± 2.76	24.77 ± 2.71	0.443
SBP (mmHg)	130.19 ± 15.27	130.32 ± 16.04	131.15 ± 15.41	132.57 ± 16.70	132.69 ± 15.06	0.115
DBP (mmHg)	83.46 ± 10.45	83.07 ± 10.72	83.70 ± 10.38	83.59 ± 10.31	83.80 ± 10.28	0.911
FPG (mmol/L)	5.54 ± 0.94	5.42 ± 0.94	5.53 ± 1.27	5.51 ± 1.14	5.61 ± 1.14	0.270
TC (mmol/L)	4.87 ± 0.92	4.95 ± 0.93	4.97 ± 0.88	4.94 ± 0.86	5.05 ± 0.87	0.213
TG (mmol/L)	1.61 ± 0.99	1.61 ± 0.87	1.66 ± 1.00	1.66 ± 0.89	1.72 ± 0.97	0.442
LDL-C (mmol/L)	2.60 ± 0.68	2.68 ± 0.68	2.69 ± 0.66	2.69 ± 0.66	2.76 ± 0.68	0.116
HDL-C (mmol/L)	1.28 ± 0.34	1.24 ± 0.33	1.25 ± 0.34	1.24 ± 0.33	1.26 ± 0.33	0.678

Note: BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TC, total cholesterol; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.

Table 6 CVD risk factors in different TSH quintile groups among subjects older than 70 years

	Q1 (0.78–1.50)	Q2 (1.51–2.11)	Q3 (2.12–2.77)	Q4 (2.78–3.90)	Q5 (3.91–8.07)	p
n	148	168	163	165	145
Age (years)	77.62 ± 6.19	78.40 ± 6.83	78.13 ± 6.14	78.92 ± 6.72	79.42 ± 6.68	0.146
BMI (kg/m2)	24.45 ± 2.85	24.77 ± 3.12	24.84 ± 3.15	24.87 ± 3.20	24.43 ± 3.16	0.472
SBP (mmHg)	137.39 ± 16.41	137.31 ± 16.95	138.87 ± 12.99	139.90 ± 16.08	138.80 ± 14.17	0.767
DBP (mmHg)	81.04 ± 10.39	81.08 ± 10.03	83.11 ± 10.19	80.90 ± 10.48	80.27 ± 9.97	0.139
FPG (mmol/L)	5.54 ± 0.91	5.59 ± 1.07	5.53 ± 0.94	5.56 ± 1.08	5.66 ± 0.99	0.804
TC (mmol/L)	4.57 ± 0.96	4.69 ± 0.97	4.77 ± 1.04	4.77 ± 0.96	4.67 ± 0.91	0.344
TG (mmol/L)	1.55 ± 0.93	1.46 ± 0.79	1.45 ± 0.80	1.55 ± 0.87	1.51 ± 0.81	0.711
LDL-C (mmol/L)	2.41 ± 0.76	2.52 ± 0.75	2.54 ± 0.82	2.55 ± 0.74	2.42 ± 0.69	0.297
HDL-C (mmol/L)	1.23 ± 0.32	1.23 ± 0.29	1.30 ± 0.32	1.27 ± 0.34	1.31 ± 0.32	0.072

Note: BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TC, total cholesterol; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.

Discussion

The present study showed that the upper limit of serum TSH level increased with age in euthyroid subjects. This confirms the routine use of age-specific TSH reference range which may improve the diagnostic accuracy.

The findings on the association between age and TSH in our study were consistent with previously reported [4, 7, 8, 20, 21]. In the US NHANES III study, the distribution of TSH and peak frequency were shifted towards higher TSH level with age, and the 97.5th percentile for TSH increased from 3.56 mIU/L in 20–29 years group to 7.49 mIU/L in ≥80 years group [8]. A more recent study of Sriphrapradang et al. [4] reported that the upper normal limit of TSH tended to increase with age in Thais, particularly in those aged ≥80 years. Vadiveloo et al. [21] also reported that there was a significant increase in median serum TSH level (1.58 mIU/L at 31–40 years to 1.86 mIU/L at >90 years) and 97.5th percentile for TSH (3.98 to 5.94 mIU/L) with age in subjects from the reference population. Another study investigated serum TSH reported that, although the median TSH and 97.5th percentile for TSH increased steadily with age, the use of age-specific reference range for TSH had minor effect on the thyroid status, except in the very old population [20]. A close association between TSH and age was observed in our subjects with the upper limit increasing from 6.03 mIU/L at 40–49 years to 8.07 mIU/L at ≥70 years. Our finding was inconsistent with that in a study conducted in Zhengzhou city of China in which no significant difference was observed in TSH among different age groups [22]. This difference might be ascribed to the sample size because the sample size was 211 in the study conducted in Zhengzhou city and 7,693 in the present study [22]. We also examined whether the trend with age for TSH differed in a single gender and found that the trend was consistent with the total population. However, our results showed that the upper reference range for TSH was higher than in Western countries [7, 21]. One possible explanation may be the ethnic difference. Several changes in thyroid physiology may contribute to the increased serum TSH level with age, such as reduced sensitivity of thyrotropin to negative feedback of thyroid hormone [23], age-related reduction in TSH bioactivity [24, 25], genetic influence [5, 26] and decreased iodine uptake [1, 27]. Further studies are required to clarify the underlying mechanism.

While applying the age-specific upper reference limit, we found some subjects would be reclassified from abnormal to normal, being 7% at 40–49 years, 8% at 50–59 years, 9.5% at 60–69 years and 13.5% at ≥70 years. Although Kalani et al. [20] found that the use of age-specific upper limit reclassified only ≤1% of subjects in most age groups, the reclassification rate was high (2.1–4.7%) in ≥85 years group. Our result suggests that the use of age-specific TSH reference range may improve the diagnostic accuracy.

Thyroid dysfunction has been a significant public health problem. Overt thyroid disorder and subclinical hypothyroidism are progressively associated with increased risk for cardiovascular events and poorer outcomes [10, 12]. It is clear that thyroid dysfunction is a risk for CVD. Meta-analyses of risks of coronary heart disease [28] and stroke [29] among subjects with the reference range of TSH were reported. One cross-sectional study showed that people with serum TSH in upper normal range had higher TG, with 1.7-fold increase in the prevalence of metabolic syndrome [16]. Our study revealed that serum TSH was associated with serum TC and LDL-C. Similarly, positive correlation was also found between TSH and lipid profiles in euthyroid Chinese [30], Spanish [31] and Korean [32]. Our results, in agreement with the findings from the study of Chon et al. [33], showed that serum TSH had no relationship with FPG and BMI. Other studies also investigate the relationship between serum TSH and BP in euthyroid population. One study indicates that people with higher serum TSH level within normal range had elevated BP than those with lower serum TSH level in normal range [32]. However, another study from 4,985 euthyroid individuals reveals that there is no association in between serum TSH and BP [34]. In our study, results showed serum TSH had positive relationship with SBP and DBP. Above findings indicate that TSH has an association with CVD risk factors, such as age, SBP, DBP, TC and LDL-C.

The impact of age on CVD risk in people with subclinical hypothyroidism has been reported by some studies. Cappola AR et al. reported that high prevalence of CVD was not associated with subclinical hypothyroidism in elderly patients [35]. Hyland KA et al. revealed that risk of CVD, heart failure, and cardiovascular death did not increase in older adults with persistent subclinical hypothyroidism [36]. However, some studies showed that the association between subclinical hypothyroidism and CVD was strong [37] or only existed [38] among individuals ≤65 yeas old. Treatment of subclinical hypothyroidism with levothyroxine provided no apparent benefits in older adults [39]. Higher levels of TSH may have different effects for younger and older people [40]. Thus, we speculate that there may be age-related difference in the relationship between CVD risk factors and elevated serum TSH. However, the effect of age on this relationship was neglected and less of them performed an age-stratified analysis. We then examined whether the relationship of TSH with CVD risk factors is influenced by age. Our results showed that higher TSH level within new reference range was associated with SBP, TC and LDL-C only in subjects aged <60 years. There were no significant differences in the BMI, BP, lipid profile and FPG among subjects with different serum TSH levels within new reference range in those aged over 60 years. Thus, though the age-related reference range may affect the association of TSH with cardiovascular risk in younger individuals, the slightly elevated TSH in the elderly may be acceptable, without the need of thyoxine treatment.

A limitation of this study is that thyroid disease was excluded by clinical and laboratory evaluations, and not every subject had thyroid ultrasound and thyroid antibodies tests. Iodine status was not assessed in our study. In addition, whether our results are applicable in other ethnicities with different lifestyles is required to be further studied.

In conclusion, our study reveals that the median TSH and 97.5th percentile for TSH increase over age. Elevated TSH in new reference range is associated with CVD risk factors only in subjects aged <60 years. Thus, in order to avoid misclassification and over-replacement, age-specific serum TSH range is recommended in the clinical practice.

Acknowledgements

The study was partially supported by the Natural Science Foundation of Jiangsu Province, China (BK20141089), and the Foundation of Jiangsu Province Health Bureau (BJ15005). The study was partially supported by Novo Nordisk young scientific talent funding.

Disclosure statement

None of the authors have any potential conflicts of interest associated with this research.

References

• 1   Cho  NH,  Choi  HS,  Kim  KW,  Kim  HL,  Lee  SY, et al. (2010) Interaction between cigarette smoking and iodine intake and their impact on thyroid function. Clin Endocrinol (Oxf) 73: 264–270.
• 2   Jeon  MJ,  Kim  WG,  Kwon  H,  Kim  M,  Park  S, et al. (2017) Excessive iodine intake and thyrotropin reference interval: data from the Korean National Health and Nutrition Examination Survey. Thyroid 27: 967–972.
• 3   Rawal  R,  Teumer  A,  Volzke  H,  Wallaschofski  H,  Ittermann  T, et al. (2012) Meta-analysis of two genome-wide association studies identifies four genetic loci associated with thyroid function. Hum Mol Genet 21: 3275–3282.
• 4   Sriphrapradang  C,  Pavarangkoon  S,  Jongjaroenprasert  W,  Chailurkit  LO,  Ongphiphadhanakul  B, et al. (2014) Reference ranges of serum TSH, FT4 and thyroid autoantibodies in the Thai population: the national health examination survey. Clin Endocrinol (Oxf) 80: 751–756.
• 5   Surks  MI,  Boucai  L (2010) Age- and race-based serum thyrotropin reference limits. J Clin Endocrinol Metab 95: 496–502.
• 6   Bremner  AP,  Feddema  P,  Leedman  PJ,  Brown  SJ,  Beilby  JP, et al. (2012) Age-related changes in thyroid function: a longitudinal study of a community-based cohort. J Clin Endocrinol Metab 97: 1554–1562.
• 7   Hollowell  JG,  Staehling  NW,  Flanders  WD,  Hannon  WH,  Gunter  EW, et al. (2002) Serum TSH, T(4), and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab 87: 489–499.
• 8   Surks  MI,  Hollowell  JG (2007) Age-specific distribution of serum thyrotropin and antithyroid antibodies in the US population: implications for the prevalence of subclinical hypothyroidism. J Clin Endocrinol Metab 92: 4575–4582.
• 9   Yeap  BB,  Manning  L,  Chubb  SA,  Hankey  GJ,  Golledge  J, et al. (2017) Reference ranges for thyroid-stimulating hormone and free thyroxine in older men: results from the Health in Men Study. J Gerontol A Biol Sci Med Sci 72: 444–449.
• 10   Collet  TH,  Gussekloo  J,  Bauer  DC,  den Elzen  WP,  Cappola  AR, et al. (2012) Subclinical hyperthyroidism and the risk of coronary heart disease and mortality. Arch Intern Med 172: 799–809.
• 11   Biondi  B (2007) Cardiovascular effects of mild hypothyroidism. Thyroid 17: 625–630.
• 12   Martin  SS,  Daya  N,  Lutsey  PL,  Matsushita  K,  Fretz  A, et al. (2017) Thyroid function, cardiovascular risk factors, and incident atherosclerotic cardiovascular disease: the Atherosclerosis Risk in Communities (ARIC) Study. J Clin Endocrinol Metab 102: 3306–3315.
• 13   Duntas  LH,  Brenta  G (2012) The effect of thyroid disorders on lipid levels and metabolism. Med Clin North Am 96: 269–281.
• 14   Pearce  EN (2012) Update in lipid alterations in subclinical hypothyroidism. J Clin Endocrinol Metab 97: 326–333.
• 15   Asvold  BO,  Bjoro  T,  Nilsen  TI,  Gunnell  D,  Vatten  LJ (2008) Thyrotropin levels and risk of fatal coronary heart disease: the HUNT study. Arch Intern Med 168: 855–860.
• 16   Ruhla  S,  Weickert  MO,  Arafat  AM,  Osterhoff  M,  Isken  F, et al. (2010) A high normal TSH is associated with the metabolic syndrome. Clin Endocrinol (Oxf) 72: 696–701.
• 17   Jung  CH,  Rhee  EJ,  Shin  HS,  Jo  SK,  Won  JC, et al. (2008) Higher serum free thyroxine levels are associated with coronary artery disease. Endocr J 55: 819–826.
• 18   Waring  AC,  Harrison  S,  Samuels  MH,  Ensrud  KE,  LeBLanc  ES, et al. (2012) Thyroid function and mortality in older men: a prospective study. J Clin Endocrinol Metab 97: 862–870.
• 19   Brabant  G,  Beck-Peccoz  P,  Jarzab  B,  Laurberg  P,  Orgiazzi  J, et al. (2006) Is there a need to redefine the upper normal limit of TSH? Eur J Endocrinol 154: 633–637.
• 20   Kahapola-Arachchige  KM,  Hadlow  N,  Wardrop  R,  Lim  EM,  Walsh  JP (2012) Age-specific TSH reference ranges have minimal impact on the diagnosis of thyroid dysfunction. Clin Endocrinol (Oxf) 77: 773–779.
• 21   Vadiveloo  T,  Donnan  PT,  Murphy  MJ,  Leese  GP (2013) Age- and gender-specific TSH reference intervals in people with no obvious thyroid disease in Tayside, Scotland: the Thyroid Epidemiology, Audit, and Research Study (TEARS). J Clin Endocrinol Metab 98: 1147–1153.
• 22   Wang  P,  Gao  YJ,  Cheng  J,  Kong  GL,  Wang  Y, et al. (2014) Serum thyroid hormone reference intervals in the apparently healthy individuals of Zhengzhou area of China. Genet Mol Res 13: 7275–7281.
• 23   Boelaert  K (2013) Thyroid dysfunction in the elderly. Nat Rev Endocrinol 9: 194–204.
• 24   Persani  L,  Ferretti  E,  Borgato  S,  Faglia  G,  Beck-Peccoz  P (2000) Circulating thyrotropin bioactivity in sporadic central hypothyroidism. J Clin Endocrinol Metab 85: 3631–3635.
• 25   Szkudlinski  MW,  Fremont  V,  Ronin  C,  Weintraub  BD (2002) Thyroid-stimulating hormone and thyroid-stimulating hormone receptor structure-function relationships. Physiol Rev 82: 473–502.
• 26   Teumer  A,  Rawal  R,  Homuth  G,  Ernst  F,  Heier  M, et al. (2011) Genome-wide association study identifies four genetic loci associated with thyroid volume and goiter risk. Am J Hum Genet 88: 664–673.
• 27   Fuse  Y (2017) Iodine and thyroid function: a historical review of goiter and the current iodine status in Japan. Pediatr Endocrinol Rev 14 Suppl 1: 260–270.
• 28   Asvold  BO,  Vatten  LJ,  Bjoro  T,  Bauer  DC,  Bremner  A, et al. (2015) Thyroid function within the normal range and risk of coronary heart disease: an individual participant data analysis of 14 cohorts. JAMA Intern Med 175: 1037–1047.
• 29   Chaker  L,  Baumgartner  C,  den Elzen  WP,  Collet  TH,  Ikram  MA, et al. (2016) Thyroid function within the reference range and the risk of stroke: an Individual Participant Data analysis. J Clin Endocrinol Metab 101: 4270–4282.
• 30   Wang  F,  Tan  Y,  Wang  C,  Zhang  X,  Zhao  Y, et al. (2012) Thyroid-stimulating hormone levels within the reference range are associated with serum lipid profiles independent of thyroid hormones. J Clin Endocrinol Metab 97: 2724–2731.
• 31   Santos-Palacios  S,  Brugos-Larumbe  A,  Guillen-Grima  F,  Galofre  JC (2013) A cross-sectional study of the association between circulating TSH level and lipid profile in a large Spanish population. Clin Endocrinol (Oxf) 79: 874–881.
• 32   Park  SB,  Choi  HC,  Joo  NS (2011) The relation of thyroid function to components of the metabolic syndrome in Korean men and women. J Korean Med Sci 26: 540–545.
• 33   Chon  SJ,  Heo  JY,  Yun  BH,  Jung  YS,  Seo  SK (2016) Serum thyroid stimulating hormone levels are associated with the presence of coronary atherosclerosis in healthy postmenopausal women. J Menopausal Med 22: 146–153.
• 34   Amouzegar  A,  Heidari  M,  Gharibzadeh  S,  Mehran  L,  Tohidi  M, et al. (2016) The association between blood pressure and normal range thyroid function tests in a population based Tehran thyroid study. Horm Metab Res 48: 151–156.
• 35   Cappola  AR,  Fried  LP,  Arnold  AM,  Danese  MD,  Kuller  LH, et al. (2006) Thyroid status, cardiovascular risk, and mortality in older adults. JAMA 295: 1033–1041.
• 36   Hyland  KA,  Arnold  AM,  Lee  JS,  Cappola  AR (2013) Persistent subclinical hypothyroidism and cardiovascular risk in the elderly: the cardiovascular health study. J Clin Endocrinol Metab 98: 533–540.
• 37   Pasqualetti  G,  Tognini  S,  Polini  A,  Caraccio  N,  Monzani  F (2013) Is subclinical hypothyroidism a cardiovascular risk factor in the elderly? J Clin Endocrinol Metab 98: 2256–2266.
• 38   Yang  L,  Zou  J,  Zhang  M,  Xu  H,  Qi  W, et al. (2013) The relationship between thyroid stimulating hormone within the reference range and coronary artery disease: impact of age. Endocr J 60: 773–779.
• 39   Stott  DJ,  Rodondi  N,  Kearney  PM,  Ford  I,  Westendorp  RGJ, et al. (2017) Thyroid hormone therapy for older adults with subclinical hypothyroidism. N Engl J Med 376: 2534–2544.
• 40   Jansen  SW,  Akintola  AA,  Roelfsema  F,  van der Spoel  E,  Cobbaert  CM, et al. (2015) Human longevity is characterised by high thyroid stimulating hormone secretion without altered energy metabolism. Sci Rep 5: 11525.