Age and Gender Substantially Influence the Relationship Between Thyroid Status and the Lipoprotein Profile: Results from a Large Cross-Sectional Study

Abstract

Background:

Conflicting data are reported on the association between mild thyroid failure and lipid profile, primarily regarding serum triglyceride values and patients with slightly elevated thyrotropin (TSH, <10 mIU/L). In this study, we assessed the possible influence of gender and age on this relationship.

Methods:

The study included 2308 consecutive patients who were seen for suspected or diagnosed thyroid disease (1874 women, 434 men, mean age 47.5±14.1 and 46.9±14.0 years, respectively) and on whom studies of thyroid status and lipoprotein profile were conducted after an overnight fast. Patients with uncontrolled diabetes mellitus and those taking lipid-lowering drugs were excluded.

Results:

There were 628 patients receiving L-thyroxine who had a diagnosis of hypothyroidism: 200 were hyperthyroid, and 120 were still hypothyroid. Overall, 648 patients were hypothyroid, and 290 were hyperthyroid. No gender difference in the frequency of TSH values in the ranges studied (i.e., TSH frequency distribution) was observed. Total cholesterol (TC) and low-density lipoprotein cholesterol (LDLc) values (p<0.0003 and p<0.003, respectively) as well as the LDL/high-density lipoprotein cholesterol (HDLc) ratio (p<0.03) were elevated not only in unselected women with TSH values in the 4th TSH group (>10 mIU/L) but also in those of the 3rd group (3.6–10.0 mIU/L) who were older than 50 years (TC and LDLc p=0.01, LDL/HDLc ratio p=0.02 vs. euthyroid women). Among unselected men, only those of the 4th TSH group had elevated triglyceride (p<0.0001) but not cholesterol values. However, men of the 3rd and 4th TSH group who were older than 65 years had significantly higher TC, LDLc, and LDL/HDLc values as well (p=0.03, p=0.02 and p=0.01, respectively vs. euthyroid men). In the final model of stepwise regression for predicting each lipid parameter variation on the basis of age, TSH, free thyroxine (FT4), and body mass index (BMI) analysis, age had the highest standardized coefficient (0.36 and 0.37, respectively), followed by TSH (0.20 and 0.11, respectively) and FT4 (−0.11 and −0.09, respectively) when looking at TC and LDLc; whereas BMI had the highest standardized coefficient (0.28), followed by age (0.15) and TSH (0.11) when looking at triglyceride variation.

Conclusions:

This study confirms a gender differentiation in the relationship between hypothyroidism and the lipid profile, which is substantially influenced by age, especially in patients with mild thyroid impairment (TSH<10 mIU/L).

Introduction

T he role of thyroid hormones in the synthesis, mobilization, and breakdown of lipids is widely recognized. Thyroid hormones stimulate the synthesis of lipids by inducing the enzyme HMG CoA reductase (1) and the expression of low-density lipoprotein cholesterol (LDLc) receptors and the subsequent cellular uptake of LDL as well as very low-density lipoprotein cholesterol (VLDLc). Moreover, thyroid hormones increase serum concentration of cholesterol ester transfer protein, which moves cholesterol from high-density lipoprotein cholesterol (HDLc) to LDLc and VLDLc, as well as the activity of the hepatic lipase, which lowers triglyceride levels (2,3). The relationship between overt hypothyroidism (HYPO) and dyslipidemia is well documented (4,5), and the prevalence of newly diagnosed cases of HYPO in patients referred to a lipid clinic is twice that in the general population (6). In addition, HYPO can induce intrinsic qualitative changes of lipoproteins, thereby enhancing their atherogenicity (7).

On the other hand, the relationship between serum lipids and subclinical hypothyroidism (sHYPO), defined as elevated serum thyrotropin (TSH) levels in the face of normal free thyroid hormone values, is still debated (4,8). Several population-based studies investigated the relationship between sHYPO and dyslipidemia with conflicting results (5,9 –11). In the Rotterdam Study, no difference was found between sHYPO and euthyroid postmenopausal women in terms of mean total cholesterol (TC) levels (12), while in the community-based study from Busselton, both TC and LDLc levels were higher in sHYPO patients as compared with euthyroid subjects (13). More recently, in the recent EPIC-Norfolk prospective population study, significantly higher levels of total and LDL cholesterol and triglycerides (TRIG) concentrations as well as lower concentrations of HDLc were observed in sHYPO women but not in men (14). Accordingly, a cross-sectional study, which evaluated the prevalence of sHYPO for different levels of TC in 1200 patients (15) documented that an increase of 1.0 mIU/L in serum TSH was associated with a significant rise in TC values of an average of 0.09 mM in women; in men with sHYPO, there was no significant relationship between these parameters. However, in the recent survey from Tromsø, which included 5143 subjects, a significant and positive correlation between serum TSH and total and LDL cholesterol values was found in both genders, but in women the relationship was not significant if adjustment was made for the influence of age and body mass index (BMI). Interestingly, a significant positive correlation between serum TSH and TRIG values was obtained in women only (11).

Giving the recognized inter-relationship between ageing, gender, and lipid profile in the general population (16), the conflicting data just described could be partially explained by the influence of age. However, the possible interaction of gender and ageing on the controversial association between sHYPO and lipid profile has not yet been extensively assessed. Therefore, the aim of the current study was to evaluate the interrelationship of gender and ageing on lipoprotein profile in a large cross-sectional study.

Materials and Methods

From January to December 2010, 2696 patients attending the thyroid outpatient clinic of the Department of Internal Medicine of the University of Pisa for suspected or diagnosed thyroid disease were consecutively enrolled. All the patients provided written informed consent to the study protocol, which was approved by the Institutional Review Board of the University of Pisa.

Patients (n=384) receiving lipid-lowering drugs and those with uncontrolled diabetes mellitus (DM) were excluded from the study. Patients with DM in a good glycometabolic state (HbA1c<7.0%; n=86) were otherwise included, assuming that well-controlled diabetes had no significant impact on lipoprotein profile. The remaining 2312 patients were submitted to a complete clinical work-up (including neck ultrasound), and their personal and familiar history for DM and hyperlipidemia was recorded. Height and weight as well as systemic blood pressure were measured using standardized protocols (17). Blood samples were drawn after an overnight fast (on the day of the first visit while possible or within one week, before starting therapy in the other cases) for the measurement of serum TSH, free thyroxine (FT₄), free triiodothyronine (FT₃), anti-thyroglobulin and anti-thyroid peroxidase autoantibody (TgAb and TPOAb, respectively) levels, as well as TC, HDLc, and TRIG. LDLc was calculated from the Friedewald's formula and was considered missing when triglyceride levels exceeded 400 mg/dL (n=11). Study participants were divided into subgroups according to age (1st: 10–29 years; 2nd: 30–49 years; 3rd: 50–64 years; 4th: >65 years) and serum TSH values (1st: TSH <0.36 mIU/L; 2nd: TSH 0.36–3.6 mIU/L; 3rd: TSH 3.6–10.0 mIU/L; 4th: TSH >10.0 mIU/L). Six hundred and twenty-six patients had already undergone a diagnosis of HYPO and were taking L-thyroxine (L-T4) therapy. Among them, 120 were still hypothyroid (median TSH, 6.90 mIU/L; range, 3.71–39.92 mIU/L), while 200 had serum TSH values below the lower limit of the normal range (median TSH 0.14 mIU/L; range, 0.0–0.35 mIU/L).

Analytical measures

Serum FT₃ and FT₄ concentrations were measured by specific radioimmunoassay (RIA; Techno-Genetics Recordati). Serum TSH levels were determined by an ultrasensitive immunoradiometric assay (IRMA) method (Cis Diagnostici). TgAbs were measured by a specific IRMA assay (TG-Ab IRMA; Biocode), and TPOAbs were measured by a specific RIA (AB-TPO; Sorin Biomedica). TC and TRIG were assayed using enzymatic methods (Roche Diagnostics). HDLc was measured enzymatically after the precipitation of LDL. LDLc was calculated by the Friedewald's formula. The normal ranges are as follows: FT₃ 2.1–4.6 pg/mL; FT₄ 8.6–18.6 pg/mL; TSH 0.36–3.6 mIU/L; TgAb <100 U/L; TPOAb <40 U/L.

Statistical analysis

Statistical analysis was performed by using the Stat-View software (SAS institute Inc, 5.0.1. version, 1992–1998). Data with normal distribution were expressed as mean±SD or otherwise as median and range. Student's t-test, analysis of variance (ANOVA; normal distribution), and Mann–Whitney U test (nonparametric data) were used to compare groups. The relationship between the parameters was assessed by simple regression (normal distribution), by Spearman correlation test (non-parametric data), or by χ² test (dichotomous variables). Stepwise regression analysis was performed with each lipid parameter as a dependent variable and TSH, FT4, age, and BMI as independent variables, in the whole population and in relation to gender. Statistical significance was assigned for p<0.05.

Results

Clinical and biochemical features

Clinical and biochemical features of the whole cohort of patients are reported in Table 1. In all, 2312 patients (1874 women [81.1%]; mean [±SD] age: 47.4±4.1 years) were submitted to a complete clinical workup; four of them had central HYPO and were, therefore, excluded from further analysis. Among the remaining 2308 patients, 1260 (54.6%) suffered from thyroid autoimmune disorders ([96%] Hashimoto's thyroiditis, [4%] Graves' disease), 470 (20.3%) from nodular goiter, and 220 (9.6%) from postsurgical HYPO; whereas 358 subjects (15.5%) did not show any current thyroid disease. The prevalence of elevated serum thyroid autoantibody titer was significantly higher among women (44.8%) than men (32.1%, p=0.0002).

Table 1.

Clinical and Biochemical Features of the Whole Cohort According to Gender

	Women (n=1870)	Men (n=438)	p
Age^* (years)	47.5±14.1	46.9±14.0	0.6
BMI^* (kg/m²)	25.8±5.4	26.8±4.4	0.02
Systolic BP^* (mm/Hg)	126.5±17.9	129.9±15.9	0.02
Diastolic BP^* (mm/Hg)	78.5±9.8	80.6±10.4	0.01
Autoimmunity^× (%)	44.8%	32.1%	0.0002
DM or IGT^× (%)	4.1%	7.3%	0.04
TSH^{a #} (mIU/L)	1.92 (0–200)	1.49 (0–100)	0.8
FT₄ ^* (pg/mL)	10.7±3.5	11.2±3.9	0.06
FT₃ ^* (pg/mL)	3.12±1.1	3.24±1.0	0.1
TC^* (mg/dL)	220.1±46.1	205.3±45.0	<0.0001
LDLc^* (mg/dL)	138.1±41.1	128.9±38.7	0.006
HDLc^* (mg/dL)	58.8±14.8	47.1±10.2	<0.0001
TRIG^* (mg/dL)	113.6±66.3	143.1±114.8	<0.0001
LDL/HDL^*	2.49±0.97	2.85±1.01	<0.0001

Data represent mean±SD, and were analyzed by ^*ANOVA, ^#Kruskal–Wallis test, or ^×χ² test.

Expressed as median and range.

ANOVA, analysis of variance; BMI, body mass index; BP, blood pressure; DM, diabetes mellitus; FT₃, free triiodothyronine; FT₄, free thyroxine; IGT, impaired glucose tolerance; TC, total cholesterol; TSH, thyrotropin; LDLc, low-density lipoprotein cholesterol; HDLc, high-density lipoprotein cholesterol; TRIG, triglycerides.

Eight hundred fifty patients (36.8%) were treated with L-T4, while 54 (2.3%) were treated with anti-thyroid drugs (methimazole). Patients receiving L-T4 therapy were mostly women (84.9%) and significantly older than those without therapy (51.6±12.3 vs. 44.9±14.4 years; p<0.0001). The prevalence of patients on L-T4 with euthyroidsm was 62.1%, while 23.8% had serum TSH values below and 14.1% had values above the normal reference range. There were 226 patients (9.8%) taking ACE inhibitors; 174 (7.5%), calcium channel antagonists; 148 (6.4%), β-blockers; and 146 (6.3%), thiazides.

Mean age and thyroid hormone profile were similar in men and women who otherwise showed significantly higher levels of TC (p<0.0001), HDLc (p<0.0001), and LDLc (p=0.006). Conversely, BMI (p=0.02), serum TRIG levels (p<0.0001), and LDL:HDLc ratio (p<0.0001) were significantly higher in men than women. Patients with thyroid autoimmunity showed higher mean values of serum TSH (3.08 [0–200] vs. 1.27 [0–100] mUI/L; p<0.0001], TC (220.9±48.4 vs. 212.2±43.6; p=0.004), and LDLc (140.5±41.6 vs. 131.0±38.2 mg/dL; p=0.001) than those without autoimmunity.

Lipid profile according to age and TSH groups

Serum TSH levels progressively decreased with increasing age (p<0.0001), while BMI increased from the 1st to the 3rd group of age (23.7±5.5, 25.6±5.1 and 27.2±5.1 kg/mq², respectively; p<0.0001), reaching a plateau in the 4th group (26.4±4.4 kg/mq²). A similar trend, with a significant progressive increase up to the 3rd group of age followed by a plateau in the 4th group, was observed for serum TG, TC, and LDLc levels as well as for the LDL/HDLc ratio (p<0.0001 for all parameters); while HDLc values did not change (Table 2). This pattern did not show any gender difference for all the analyzed parameters.

Table 2.

Clinical and Biochemical Features of the Whole Cohort According to Age Groups

	1st group (n=306)	2nd group (n=894)	3rd group (n=838)	4th group (n=270)
Gender (% female)	85.6%	76.7%	84.5%	80.0%
Age (years)	24.7±3.6	40.5±5.8	55.7±4.2	70.4±5.0
BMI (kg/mq)	23.7±5.5	25.6±5.1	27.2±5.1^*	26.4±4.4
Systolic BP (mm/Hg)	113.3±12.5	121.4±15.6	132.1±17.1	136.3±16.1
Diastolic BP (mm/Hg)	71.5±8.5	77.0±9.6	81.9±9.6	79.7±8.9
Autoimmunity (%)	56.1%	52.5%	51.2%	50.0%
TSH^a (mIU/L)	3.09 (0.1–119.87)	1.97 (0–200)	1.46 (0–150)	1.32^# (0–46)
FT4 (pg/mL)	9.8±3.6	10.7±3.5	11.1±3.6	11.1±3.7
FT3 (pg/mL)	3.3±1.4	3.1±1.0	3.1±0.9	3.0±1.2
TC (mg/dL)	188.0±40.3	209±44.0	233.9±42.1^*	225.8±50.1^*
LDLc (mg/dL)	109.6±31.6	131.7±39.1	149.6±39.0^*	145.3±43.0^*
HDLc (mg/dL)	56.4±13.1	56.0±13.4	57.4±16.4	55.9±15.5
TRIG (mg/dL)	91.0±5.3	111.6±83.9	133.5±80.4^*	129.6±63.3^*
LDL/HDL	2.06±0.82	2.48±0.94	2.81±1.00^*	2.78±1.02^*

Age groups were defined as: 1st, 10–29 years; 2nd, 30–49 years; 3rd, 50–64 years; 4th, >65 years.

Data represent mean±SD, and were analyzed by ^*ANOVA (p<0.0001 vs. 1st and 2nd age groups) or ^#Kruskal–Wallis test (p<0.0001).

Expressed as median and range.

Splitting the whole cohort into groups of TSH, 290 patients (238 women, 82.1%) had serum TSH values below the normal range (1st group); 1370 (1098 women, 80.1%) had TSH values within the normal range (2nd group), 474 patients (388 women, 81.9%) had mildly elevated TSH values (3.6–10 mIU/L; 3rd group), and 174 (146 women, 83.9%) had markedly elevated TSH values (>10 mIU/L; 4th group), as shown in Table 3. Patients of the 1st TSH group were significantly older than those of the three other groups (p<0.001 vs. 2nd and 4th and p<0.0001 vs. 3rd); no difference was otherwise observed regarding systemic blood pressure, BMI, and gender among TSH groups.

Table 3.

Clinical and Biochemical Features of the Whole Cohort According to TSH Groups

	1st group (n=290)	2nd group (n=1370)	3rd group (n=474)	4th group (n=174)
Women (%)	82.1	80.1	81.8	83.9
Age (years)	52.8±11.9^*	48.1±13.7	42.2±14.1	46.6±15.1
BMI (kg/mq)	26.3±4.3	26.2±5.5	25.2±4.9	25.8±5.1
Systolic BP (mm/Hg)	125.8±15.6	127.9±17.3	125.9±19.3	125.9±20.1
Autoimmunity (%)	28.2	35.8	59.1	88.5^×
TSH^a (mIU/L)	0.14 (0.0–0.35)	1.43 (0.36–3.59)	5.06 (3.6–9.87)	19.8 (10.1–200)
FT4 (pg/mL)	14.6±5.1	11.0±2.6	9.4±2.3	6.5±3.9^**
FT3 (pg/mL)	3.7±2.0	3.2±0.8	3.0±0.6	2.4±0.9^**
TC (mg/dL)	213.9±47.3	215.7±43.3	215.5±44.0	243.2±65.4^**
LDLc (mg/dL)	131.8±39.8	134.1±36.7	134.7±41.2	154.5±58.9^**
HDLc (mg/dL)	55.1±14.9	57.0±14.7	56.5±15.3	54.7±13.2
TRIG (mg/dL)	123.8±65.9	118.7±78.5	116.1±71.7	140.8±117.6^**
LDL/HDL	2.50±0.98	2.51±0.95	2.53±0.94	2.91±1.25^**

TSH groups were defined as: 1st, <0.36 mIU/L; 2nd, >0.36 and <3.6 mIU/L; 3rd, >3.6 and <10.0 mIU/L; 4th, >10.0 mIU/L.

Data represent mean±SD, and were analyzed by ANOVA (^* p<0.001 1st group vs. 2nd and 4th group, p<0.0001 vs. 3rd group; ^** p<0.0001 4th group vs. each other group) or ^×χ² test (p=0.0002).

Expressed as median and range.

The overall prevalence of patients with low TSH value receiving L-T4 therapy was significantly higher than those without (23.8% vs. 6.2%, p<0.0001). Interestingly, the prevalence of low TSH value progressively increased with increasing age (2.4%, 4.8%, 7.7%, and 14.1%, from the 1st to the 4th age group) in patients without L-T4, while it increased in the 1st, 2nd, and 3rd age groups only (11.5%, 20.7%, and 29.8%, respectively) and clearly declined in the 4th group (16.4%) in those receiving L-T4 therapy.

Eighty patients of the 1st TSH group (27.6%) suffered from overt hyperthyroidism (TSH: 0.13 [0.0–11.3] mIU/L; FT₄: 19.9±6.5 pg/mL; FT₃: 6.4±2.9 pg/mL). Patients of the 3rd TSH group showed significantly lower serum FT₄ levels than those of the 2nd group (p<0.0001), although still within the normal range (sHYPO). Patients of the 4th group had significantly lower serum FT₃ and FT₄ levels than those of either the 2nd or the 3rd TSH group (p=0.0001 vs. both groups), and 94 of them (54%) were affected by overt HYPO: frankly elevated TSH value (48.6 [16.3–200] mIU/L) and both FT₄ and FT₃ levels below the normal range (4.2±2.8 and 1.4±0.5 pg/mL, respectively).

According to TSH level, only patients of the 4th group had significantly elevated serum TC (p<0.0001), LDLc (p<0.001), and TRIG levels (p<0.02) as well as LDL/HDLc ratio (p<0.001) than those of each other group, while serum HDLc value was similar in all TSH groups (Table 2). Splitting by gender, significantly elevated serum cholesterol levels were observed in women only (TC: p<0.0003; LDLc p<0.003; LDL/HDL: p<0.03; 4th vs. each other TSH group), while increased serum TRIG values were observed in men (p=0.02; 4th vs. each other TSH group). Correcting data by postmenopausal replacement therapy, the results did not change.

By analyzing the lipid profile according to age and TSH groups, apart from HDLc, which never significantly changed, each serum lipid parameter progressively increased with increasing age, although with different figures in relation to thyroid status (Table 4). Indeed, in hyperthyroid and euthyroid subjects (1st and 2nd TSH group), serum lipids progressively increased from the 1st to the 3rd age group, reaching a plateau or even reducing in individuals of the 4th group (older than 65 years). Conversely, both serum LDLc and LDL/HDLc values significantly increased further in patients of the 3rd TSH group older than 65 years (p<0.01 vs. 3rd age group), while in patients of the 4th TSH group (>10 mIU/L), such an increase was observed in serum TC levels as well (Table 4). Finally, a significant increase in serum TC, LDLc, and LDL/HDLc values was also observed in men of the 3rd TSH group (<10 mIU/L) older than 65 years (p=0.03, p=0.02, and p=0.01 vs. 1st and 2nd TSH group, respectively).

Table 4.

Lipoprotein Profile of the Whole Cohort According to Age and TSH Groups

	Age group
Lipid parameter TSH group	1st	2nd	3rd	4th
TC (mg/dL)
1st	177.5±43.8	203.5±34.1	229.0±40.1^††	205.0±49.1^**
2nd	187.8±43.3	207.7±42.9	230.5±39.8^†††	222.9±36.9^†††
3rd	182.6±31.3	209.9±43.2	241.9±37.3^†††	236.1±41.9^†††
4th	213.1±48.7	231.1±49.8	257.4±41.5^††	283.7±44.3^‡‡
LDLc (mg/dL)
1st	92.7±61.4	135.1±28.7	149.4±28.5^††	132.7±37.9^**
2nd	106.7±30.0	129.4±37.2	147.1±34.6^†††	138.2±30.2^***
3rd	109.9±29.4	131.2±39.7	154.6±41.2^†††	171.0±47.2^‡
4th	131.5±35.1	144.7±43.7	161.4±42.3^††	196.5±43.1^‡‡
HDLc (mg/dL)
1st	52.4±8.1	51.4±10.7	58.4±16.0	55.0±18.3
2nd	58.6±13.3	56.7±14.5	56.7±14.9	57.3±15.1
3rd	54.3±13.5	55.6±10.9	60.1±23.3	59.0±12.1
4th	52.2±9.6	56.8±13.6	53.4±13.0	54.1±16.6
LDL/HDL ratio
1st	1.65±1.02	2.83±0.98	2.75±0.95^**	2.67±1.00^**
2nd	1.94±0.72	2.42±0.89	2.79±1.00^†††	2.56±0.78^***
3rd	2.15±0.83	2.45±0.87	2.78±0.93^†††	3.53±1.01^‡
4th	2.67±1.01	2.70±1.03	3.09±1.05^††	3.83±1.05^‡
TRIG (mg/dL)
1st	99.8±45.5	112.7±53.3	120.2±55.9^**	120.0±54.1^**
2nd	91.7±42.5	113.6±47.7	129.4±53.4^†††	124.4±54.7^***
3rd	85.3±37.9	110.2±53.2	142.8±54.2^†††	152.0±53.3^†††
4th	110.8±53.9	105.2±54.2	185.0±57.3^††	171.3±51.1^††

TSH groups were defined as: 1st, <0.36 mIU/L; 2nd, 0.36–3.6 mIU/L; 3rd, 3.6–10.0 mIU/L; 4th, >10.0 mIU/L. Age groups were defined as: 1st, 10–29 years; 2nd, 30–49 years; 3rd, 50–64 years; 4th, >65 years.

Data represent mean±SD. Only the statistical significance of groups of patients older than 50 years (3rd and 4th age groups) as compared with each other and with groups of patients younger than 50 years (1st and 2nd age groups) is described: ^** p<0.01, ^*** p<0.001 vs. 1st age group;^†† p<0.01, ^††† p<0.001 vs. 1st and 2nd age groups; ^‡ p<0.05, ^‡‡ p<0.01 vs. 3rd age group.

Regression analysis

In the whole cohort, a significant correlation was obtained between age and BMI (r=0.22; p<0.0001), with the relationship being essentially linked to female gender (r=0.23; p<0.0001); while in men, it only approached statistical significance (r=0.15; p=0.05). A significant positive correlation between BMI and each serum lipid parameter was also found (TC [r=0.16; p<0.0001], LDLc [r=0.14; p<0.0001], HDLc [r=0.16; p<0.0001], TRIG [r=0.32; p<0.0001], LDL/HDL ratio [r=0.23; p<0.0001]); the relationship between BMI and TRIG was not observed in men. A significant positive correlation was observed between age and serum lipid profile apart from HDLc (TC [r=0.32; p<0.0001], LDLc [r=0.33; p<0.0001], TRIG [r=0.19; p<0.0001], LDL/HDL ratio [r=0.25; p<0.0001]).

Serum TSH levels showed a significant positive relationship with TC (r=0.20; p<0.0001), LDLc (r=0.12; p=0.0001), and TRIG (r=0.12; p=0.0001) values as well as LDL/HDLc ratio (r=0.11; p=0.0004). A different figure was obtained according to gender: Women did not exhibit a relationship between serum TSH and TRIG values, which was otherwise maintained in men (r=0.31; p<0.0001). Similar, although negative, correlations were obtained while analyzing the various lipid parameters in relation to free thyroid hormone levels (data not shown).

In the final model of stepwise regression analysis, which explained almost 40% of serum TC and LDLc variations (p<0.0001; r ²=0.43 and 0.39, respectively), age had the highest standardized coefficient (SC; 0.36 and 0.37, respectively), followed by TSH (0.20 and 0.11, respectively) and FT4 (−0.11 and −0.09, respectively), while BMI was excluded. Conversely, while utilizing serum TRIG as the dependent variable, in the final model that explained almost 35% of its variation (p<0.0001, r ²=0.36), BMI had the highest SC (0.28), followed by age (0.15) and TSH (0.11), while FT4 was excluded. Finally, in case of serum HDLc as the dependent variable, only BMI was included in the final model (p<0.0001, r ²=0.17) (Table 5).

Table 5.

Stepwise Regression Analysis with Serum TSH Value, FT₄ Values, BMI, and Age as Independent Variables and TC, LDLc, HDLc, and TRIG as Dependent Variables, Alternatively

	Whole cohort	r²	Women	r²	Men	r²
TC
Step 1	Age	0.340	Age	0.359	Age	0.290
Step 2	TSH age	0.423	TSH age	0.437	TSH age	0.376
Step 3	TSH FT₄ age	0.434	FT₄ TSH age	0.448	TSH BMI age	0.414
LDLc
Step 1	TSH FT₄ BMI	0.355	age	0.379	age	0.262
Step 2	TSH age	0.386	TSH; age	0.412
Step 3	TSH FT₄ age	0.394
HDLc
Step 1	BMI	0.170	BMI	0.179	BMI	0.140
TRIG
Step 1	BMI	0.316	BMI	0.331	BMI	0.335
Step 2	BMI age	0.346	BMI age	0.385	TSH BMI	0.450
Step 3	TSH BMI age	0.361

Among women, the result did not change with regard to cholesterol parameters, while only BMI and age were included in the final model with TRIG as the dependent variable (r ²=0.38), and BMI showing the highest SC (0.28), followed by age (0.20). Among men, the result did not change only for TC (r ²=0.41), with age showing the highest SC (0.25), followed by TSH (0.23) and BMI (0.17). However, using LDLc as the dependent parameter, only age was included in the final model (r ²=0.26), and both TSH and BMI were excluded. Regarding TRIG, both BMI and TSH were included in the final model (r ²=0.45; SC 0.31 and 0.30, respectively) (Table 5).

The results did not change after adjustment for a positive family history for DM and/or dyslipidemia as well as current DM or impaired fasting glucose. Similarly, no modifications were observed while correcting data for patients receiving L-T4 as well as ACE inhibitors, β-blockers, or thiazides.

Discussion

The main finding of the current study is that both gender and age significantly affect the relationship between HYPO and the lipid profile, especially in patients with mild thyroid failure (TSH <10 mIU/L). Indeed, in the final model of stepwise regression analysis, age appeared to be the most important factor influencing serum cholesterol parameters followed by TSH, whose effect was markedly more evident in women than men. In women, both age and TSH values strongly influenced either serum TC or LDLc levels, while, in men, TSH significantly affected TC value only. Conversely, in women, serum TSH did not affect TRIG levels, which substantially depend on BMI and age, while, in men, both TSH and BMI essentially influenced serum TRIG concentrations independently of age.

Altogether, our results show that, although with a certain gender differentiation, (mild) HYPO is able to augment and worsen the effect of ageing on serum lipids, especially cholesterol parameters, which showed a continuous increasing trend up to the last decades of life, at odds with euthyroid subjects in whom a plateau was reached in the 4th age group (>65 years). This latter finding is in line with previous population-based, prospective studies that showed an age-dependent increase in serum lipids up to the sixth decade of life followed by a plateau or even a decline in the last decades with a slight gender difference: LDLc levels reach a plateau in men between the age of 50 and 60 years, and in women, between the age of 60 and 70 years (16).

In the whole cohort, a significant positive correlation between serum TSH and TC, LDLc, and TRIG levels as well as LDL/HDLc ratio was found; while an association with HDLc values was never observed. Interestingly, serum cholesterol levels (TC, LDLc, and LDL/HDL ratio) significantly increased only in women of the 4th TSH group (TSH >10 mIU/L), while serum TRIG increased only in men of the same TSH group, suggesting that a slight increase in serum TSH values only marginally influences the lipid profile. These findings are in agreement with most previous population-based and cross-sectional studies, although not all the parameters reported were assessed in each study (8,9, 11,13,18 –21). However, three placebo-controlled studies demonstrated a significant effect of L-T4 therapy, also in patients with slightly increased serum TSH (>3.6 mIU/L) (22 –24).

Discordant results about the influence of TSH on lipoprotein profile in relation to gender have been previously reported (14,25). Most of the conflicting data could be partially explained by the substantial role of age, as shown in the current study. In this setting, the crucial role of age along with gender in influencing lipid parameters was confirmed while analysing data according to TSH and age groups, which showed a progressive effect of TSH on lipid profile with increasing age. Indeed, a slight TSH increase (3.6–10 mIU/L, 3rd group) was associated with significant hypercholesterolemia only in women older than 50 years, including elevated TRIG values in those older than 65 years. Similarly, among men of the 3rd TSH group, only those older than 65 years, besides increased TRIG value, had significant hypercholesterolemia. Overall, these findings suggest a substantial interaction between ageing and increased TSH in influencing lipid level toward an atherogenic profile. To our knowledge, this is the first report which documents that the impact of increased TSH value on lipid profile is substantially influenced by age in both genders. Indeed, most of the previous studies usually included middle-aged subjects (12,14,25,26) or the relationship between TSH and lipoprotein profile was not corrected by age groups (11,13,19).

In the current study, the influence of ageing in promoting the metabolic effect of even slightly elevated TSH levels, although present in both genders, appears earlier in women than men (>50 vs. >65 years, respectively). In this setting, it is well known that menopause is associated with unfavorable changes in body composition, abdominal fat deposition (27), and adverse metabolic changes. Cross-sectional data from large-scale population studies suggest that around the time of the menopause, LDLc levels increase by ∼15–25%. Since this increase is larger than that observed in men over the same age span, it is likely that reduced circulating estrogen levels associated with menopause play a role in the adverse changes in blood lipid levels (28,29). Moreover, epidemiologic studies have suggested that an association exists between menopause and the metabolic syndrome (30,31). In our cohort BMI, strongly correlates with age, especially in women, and largely influences serum TRIG levels; so, we might hypothesize that a certain degree of insulin resistance along with other metabolic changes occurring during menopause can interplay with TSH in affecting the lipoprotein profile after 50 years in women. Interestingly, Bakker et al. (32) demonstrated that in euthyroid nondiabetic adults, the relationship between serum TSH and cholesterol levels appears modified by insulin resistance. Unfortunately, this hypothesis is somewhat speculative, as we did not obtain any measure of insulin sensitivity in our cohort and, the results did not change while corrected by estrogen replacement therapy. These findings, however, outline the complex, not yet completely resolved relationship between thyroid function and intermediate metabolism (33).

In the current study, the overall prevalence of sHYPO was notably higher than that generally reported in literature, without any gender difference (20.7% vs. 19.6%, women and men, respectively) and did not change with increasing age (19). As expected, women showed a significantly higher prevalence of thyroid autoimmunity than men (55.3% vs. 39.7%). These results are probably the consequence of the study design with patients' enrolment in a third-level thyroid outpatient clinic. Although the fact that the study was performed in a population with highly prevalent endocrine disease can be considered a limitation, it can also be regarded as an ideal situation to investigate lipid changes in relation to thyroid alterations. Another potential limitation of the current study is that almost one third of the patients, significantly older than those without therapy, were receiving L-T4 therapy. Thus, L-T4 replacement might have affected the inter-relationship between age, lipids, and serum TSH values. However, while correcting the data (including stepwise regression analysis) by L-T4 therapy, the results did not change. Therefore, we can assume that L-T4 therapy did not act as a bias in the analysis of the data and in their interpretation. Finally, the lack of information on smoking habitus may be a limitation of the study, while the careful exclusion of patients receiving lipid-lowering drugs and the correction of data by treatments potentially influencing lipid profile (ACE inhibitors, β-blockers, thiazides) certainly improves its power.

In conclusion, the current study confirms a gender-related differentiation in the relationship between HYPO and lipid profile. Moreover, this is the first study which documents the fact that the impact of increased TSH value on lipid profile is substantially influenced by age in both genders: HYPO being able to augment and worsen the effect of ageing on serum lipids. Overall, our data delineate a specific interrelationship between thyroid function, gender, and age and may, at least partially, explain the conflicting results of the literature regarding the impact of slightly elevated TSH value on circulating lipid parameters.

Footnotes

Disclosure Statement

The authors have no commercial associations that might create a conflict of interest in connection with this article.

References

Duntas

. 2002. Thyroid disease and lipids. Thyroid, 12:287–293.

Tan

KCB

, Shiu

SWM

, Kung

AWC

. 1998. Plasma cholesteryl ester transfer protein activity in hyper- and hypothyroidism. J Clin Endocrinol Metab, 83:140–143.

Lam

, Chan

, Yeung

. 1986. High-density lipoprotein cholesterol, hepatic lipase and lipoprotein lipase activities in thyroid dysfunction: effects of treatment. Q J Med, 59:513–521.

Duntas

, Biondi

. 2011. New insights into subclinical hypothyroidism and cardiovascular risk. Semin Thromb Hemost, 37:27–34.

Peppa

, Betsi

. Dimitriadis G 2011 Lipid abnormalities, cardiometabolic risk in patients with overt, subclinical thyroid disease. J Lipids, 2011:575840.

Diekman

, Lansberg

, Kastelein

, Wiersinga

. 1995. Prevalence and correction of hypothyroidism in a large cohort of patients referred for dyslipidemia. Arch Intern Med, 155:1490–1495.

Oge

, Sozmen

, Karaoglu

. 2004. Effect of thyroid function on LDL oxidation in hypothyroidism and hyperthyroidism. Endocr Res, 30:481–489.

Biondi

, Cooper

. 2008. The clinical significance of subclinical thyroid dysfunction. Endocr Rev, 29:76–131.

Bell

, Rivera-Woll

, Davison

, Topliss

, Donath

, Davis

. 2007. Well-being, health-related quality of life and cardiovascular disease risk profile in women with subclinical thyroid disease: a community-based study. Clin Endocrinol, 66:548–556.

10.

Takashima

, Niwa

, Mannami

, Tomoike

, Iwai

. 2007. Characterization of subclinical thyroid dysfunction from cardiovascular and metabolic viewpoints: the suita. Study Circ J, 71:191–195.

11.

Iqbal

, Jorde

, Figenschau

. 2006. Serum lipid levels in relation to serum thyroid-stimulating hormone and the effect of thyroxine treatment on serum lipid levels in subjects with subclinical hypothyroidism: the Tromso Study. J Intern Med, 260:53–61.

12.

Hak

, Pols

, Visser

, Drexhage

, Hofman

, Witteman

. 2000. Subclinical hypothyroidism is an independent risk factor for atherosclerosis and myocardial infarction in elderly women: the Rotterdam Study. Ann Intern Med, 132:270–278.

13.

Walsh

, Bremner

, Bulsara

, O'Leary

, Leedman

, Feddema

, Michelangeli

. 2005. Thyroid dysfunction and serum lipids: a community-based study. Clin Endocrinol, 63:670–675.

14.

Boekholdt

, Titan

, Wiersinga

, Chatterjee

, Basart

, Luben

, Wareham

, Khaw

. 2010. Initial thyroid status and cardiovascular risk factors: the EPIC-Norfolk prospective population study. Clin Endocrinol, 72:404–410.

15.

Bindels

, Westendorp

, Frölich

, Seidell

, Blokstra

, Smelt

. 1999. The prevalence of subclinical hypothyroidism at different total plasma cholesterol levels in middle aged men and women: a need for case-finding? Clin Endocrinol, 50:217–220.

16.

Gobal

, Mehta

. 2010. Management of dyslipidemia in the elderly population. Ther Adv Cardiovasc Dis, 4:375–383.

17.

Lohman

, Roche

, Martorell

. 1988. Anthropometric Standardization Reference Manual. First. Champaign, IL: Human Kinetics Books.

18.

Bauer

, Ettinger

, Browner

. 1998. Thyroid function and serum lipids in older women: a population-based study. Am J Med, 104:546–551.

19.

Canaris

, Manowitz

, Mayor

, Ridgway

. 2000. The colorado thyroid disease prevalence study. Arch Int Med, 160:526–534.

20.

Cappola

, Fried

, Arnold

, Danese

, Kuller

, Burke

, Tracy

, Ladenson

. 2006. Thyroid status, cardiovascular risk, and mortality in older adults. JAMA, 295:1033–1041.

21.

Lindeman

, Schade

, LaRue

, Romero

, Liang

, Baumgartner

, Koehler

, Garry

. 1999. Subclinical hypothyroidism in a biethnic, urban community. J Am Geriatr Soc, 47:703–709.

22.

Monzani

, Caraccio

, Kozàkowà

, Dardano

, Vittone

, Virdis

, Taddei

, Palombo

, Ferrannini

. 2004. Effect of levothyroxine replacement on lipid profile and intima-media thickness in subclinical hypothyroidism: a double-blind, placebo- controlled study. J Clin Endocrinol Metab, 89:2099–2106.

23.

Caraccio

, Ferrannini

, Monzani

. 2002. Lipoprotein profile in subclinical hypothyroidism: response to levothyroxine replacement, a randomized placebo-controlled study. J Clin Endocrinol Metab, 87:1533–1538.

24.

Meier

, Staub

, Roth

, Guglielmetti

, Kunz

, Miserez

, Drewe

, Huber

, Herzog

, Müller

. 2001. TSH-controlled L-thyroxine therapy reduces cholesterol levels and clinical symptoms in subclinical hypothyroidism: a double blind, placebo-controlled trial (Basel Thyroid Study) J Clin Endocrinol Metab, 86:4860–4866.

25.

Asvold

, Vatten

, Nilsen

, Bjøro

. 2007. The association between TSH within the reference range and serum lipid concentrations in a population-based study. The HUNT Study. Eur J Endocrinol, 56:181–186.

26.

Hueston

, Pearson

. 2004. Subclinical hypothyroidism and the risk of hypercholesterolemia. Ann Fam Med, 2:351–355.

27.

Ozbey

, Sencer

, Molvalilar

, Orhan

. 2002. Body fat distribution and cardiovascular disease risk factors in pre- and postmenopausal obese women with similar BMI. Endocr J, 49:503–509.

28.

Davidson

, Maki

, Karp

, Ingram

. 2002. Management of hypercholesterolaemia in postmenopausal women. Drugs Aging, 19:169–178.

29.

Derby

, Crawford

, Pasternak

, Sowers

, Sternfeld

, Matthews

. 2009. Lipid changes during the menopause transition in relation to age and weight: the Study of Women's Health Across the Nation. Am J Epidemiol, 169:1352–1361.

30.

Santoro

, Torrens

, Crawford

, Allsworth

, Finkelstein

, Gold

, Korenman

, Lasley

, Luborsky

, McConnell

, Sowers

, Weiss

. 2005. Correlates of circulating androgens in mid-life women: the Study of Women's Health Across the Nation. J Clin Endocrinol Metab, 90:4836–4845.

31.

Ainy

, Mirmiran

, Zahedi Asl

, Azizi

. 2007. Prevalence of metabolic syndrome during menopausal transition Tehranian women: Tehran Lipid and Glucose Study (TLGS) Maturitas, 58:150–155.

32.

Bakker

, ter Maaten

, Popp-Snijders

, Slaets

, Heine

, Gans

. 2001. The relationship between thyrotropin and low density lipoprotein cholesterol is modified by insulin sensitivity in healthy euthyroid subjects. J Clin Endocrinol Metab, 86:1206–1211.

33.

Solini

, Monzani

. 2010. Hypothyroidism and intermediate metabolism: a complex relationship. Thyroid, 20:837–839.