Associations of free triiodothyronine and thyroxine with blood pressure in Chinese coronary artery disease adults

Abstract

Background

Previous evidence regarding the associations of free triiodothyronine (FT3) and free thyroxine (FT4) with blood pressure (BP) were scarce. We aimed to explore the associations of FT3 and FT4 with BP (systolic BP [SBP], diastolic BP [DBP] and mean arterial pressure [MAP]) and examine the mediating roles of body mass index (BMI) and high–sensitivity C–reactive protein (hsCRP) in Chinese adults with coronary artery disease (CAD).

Methods

This hospital-based cross-sectional study included 4606 subjects. Serum FT3 and FT4 were measured.

Results

Among total populations, significant differences of 2.09 mmHg (1.73%) and 2.10 mmHg (1.73%) for SBP and 1.16 mmHg (1.32%) and 1.10 mmHg (1.25%) for MAP were observed between quartiles 4 and 1 of FT3 and FT4 in multi–variable analysis. Moreover, an intermediate FT3 concentration was associated with a higher risk of hypertension (HTN) (OR: 1.32, 95%CI: 1.03, 1.69). Similar positive association was also observed between FT4 and HTN prevalence. There was no significant association between FT3/FT4 ratio and BP (SBP, DBP and MAP). The mediating effects of BMI and hsCRP on “FT3 or FT4–BP” associations were not found among total populations in path analyses.

Conclusions

FT3 and FT4 were positively associated with SBP, MAP and HTN prevalence in Chinese adults with CAD. The “FT3 or FT4–BP” associations might not be mediated by BMI and hsCRP.

Introduction

Coronary artery disease (CAD) is a predominant risk factor for mortality in worldwide [1]. In China, the mortality of CAD reached 11.35 and 11.87 per one Hundred thousand urban populations and rural residents in 2016, which increased 4–fold During the past 14 years [2]. It has been reported that elevated blood pressure (BP) might lead to clinical cardiovascular disease and increase the mortality of cardiovascular related disease [3, 4]. Thus, the investigation of novel modifiable risk factors for BP regulation in patients with CAD is an urgent need.

Thyroid hormone, an essential indicate of maintaining human health [5], could exert multiple effects on the cardiovascular system [6, 7] via modulating glycolipid metabolism and blood pressure. Epidemiological data demonstrated that abnormal BP levels were associated with adverse prognosis in CAD patients [8, 9]. Free triiodothyronine (FT3) and free thyroxine (FT4) were considered as the vital frontline parameters in the routine evaluation of thyroid function [10, 11]. Roos et al. found that FT3 and FT4 played important roles in cardiovascular disease rather than TSH [12]. In vitro and animal studies revealed that triiodothyronine (T3) and thyroxine (T4) could modulate BP levels via catecholamines, vasopressin, endothelin and the renin-angiotensin-aldosterone system [13]. Several studies have been performed, however no consistent conclusions have been drawn to clarify the associations between free thyroid hormones and BP. Few data assessed the association of FT3 with BP (systolic BP [SBP] or diastolic BP [DBP]) showing positive [12, 14–18] and null [12, 14, 16, 17] associations. Similarly, other studies reported positive [12, 15, 17, 19–26], negative [18, 26] or null associations [16, 17, 22, 25–28] of the FT4 with SBP or DBP. It remains uncertain whether discrepancies among above studies might be attributable to a limited effect size, different statistical analysis (analysis of covariance, pearson, spearman and linear regression) or different subjects. No previous study has directly examined the associations of circulating FT3 and FT4 with BP in Chinese adults with CAD. Therefore, these associations remain largely unclear.

In this hospital–based cross–sectional study, we examined the potential associations of serum FT3 and FT4 with BP in Chinese adults with CAD.

Patients and methods

Design and participants

This hospital–based cross–sectional study was carried out During November 2016 and December 2017 in Wuhan Asia Heart Hospital. 4606 CAD adults aged 60.9 ± 9.64 years were included in this study (n_men=2925, n_women=1681). CAD was diagnosed as meeting one of the following criteria: a). history of coronary intervention, angina pectoris or myocardial infarction; b). coronary angiography showed vascular stenosis > 50%; c). electrocardiogram displayed obvious myocardial infarction [29]. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Wuhan Asia Heart Hospital (No. 2016-B008). Informed consent has been obtained from all subjects.

Epidemiological data and laboratory measurements

We collected detailed information from all patients, including alcohol drinking status (current drinker/non–drinker), smoking status (current smoker/non–smoker), age, sex, lipid–lowering, antihypertensive and hypoglycemic agents use. Height and weight were measured from all subjects and body mass indexes (BMIs) were calculated.

The electronic sphygmomanometer (Omron Corporation, Japan) was used to measure BP after subjects were seated or supine by medical workers using the standard protocol. All readings were accurate to 1 mmHg. BP was measured twice at Least 20 min apart. A third measurement was performed when the difference > 4 mmHg between the previous two measurements. The mean values of BP (SBP and DBP) were used in the present study. The mean arterial pressure (MAP) was calculated (MAP =[SBP+(DBP×2)]/3). High–sensitivity C–reactive protein (hsCRP) was measured in accordance with the manufacturer’s instructions using a Cardiac CRP (Latex) High Sensitive kit.

Serum FT3 and FT4 concentrations were detected using Access FT3 (chemiluminescence method, A13422, REV5.1, Beckman) and Access FT4 (chemiluminescence method, 33880, REV5.2, Beckman) assay kit, respectively. In briefly, a sample was added to a reaction tube containing alkaline phosphatase conjugate anti–T3 monoclonal antibody. During the incubation process, FT3 in the sample reacted with anti–T3 antibodies. The particles coated with streptavidin and biotinylated T3 analogs were added to the mixture. After incubation in the reaction tube, the substance bound to the solid phase was attracted by a magnetic field, while the unbound substance was washed away. Then, the chemiluminescence substrate was added to the reaction tube, and the light generated during the reaction was measured by an illuminance meter. The amount of light generated was inversely proportional to FT3 concentration in the sample. The amount of analyte in the sample was determined by the stored multi-point calibration curve. As for FT4, the monoclonal anti-T4 antibody combined with biotin, sample, buffer protein solution, and solid-phase coated with streptavidin were added to the reaction tube. During this initial incubation process, biological anti-T4 antibodies were combined with FT4 in the solid phase and sample. After incubation in the reaction tube, the substance bound to the solid phase was attracted by a magnetic field, while the unbound substance was washed away. Then, the buffer protein solution and T3-alkaline phosphatase conjugate were added to the reaction tube. After incubation in the reaction tube, the substance bound to the solid phase was attracted by a magnetic field, while the unbound substance will be washed away. Then, the chemiluminescence substrate Lumi Phos * 530 was added to the reaction tube, and the light generated during the reaction was measured by an illuminance meter. The amount of light generated was inversely proportional to the concentration of FT4 in the sample. The amount of analyte in the sample was determined by the stored multi-point calibration curve.

Statistical analysis

The basic characteristics of the subjects were presented as medians (interquartile ranges) or means (± standard deviations, SDs) for continuous variables, and as frequencies (percentages) for categorical variables in men and women, respectively. Partial correlation analysis was performed to evaluate the correlations of serum FT3 and FT4 with BP (SBP, DBP and MAP). The subjects were categorized into quartiles according to serum FT3 and FT4 concentrations. Multivariate analyses of covariance were used to compare the mean differences in BP (SBP, DBP and MAP) and to test for trends in accordance with the serum FT3 and FT4 quartiles. The Bonferroni test was performed for pair–wise comparisons between quartiles. Sex, age, alcohol consumption, BMI, smoking status, and antihypertensive agents use might either confound or mediate the “FT3 (or FT4) – SBP (or MAP)” associations [20, 30] and abnormal glucose and lipid metabolism were usually were observed in CAD patients, so we adjusted for above variables plus lipid–lowering and hypoglycemic agents use. In model 1, sex and age were adjusted. In model 2, alcohol consumption, BMI, smoking status, lipid–lowering, antihypertensive and hypoglycemic agents use were further adjusted. Logistic regression analysis was used to calculate the associations of FT3, FT4 and FT3/FT4 ratio with hypertension (HTN) prevalence. Stratified analyses were conducted in men and women, respectively. Path analysis was used to assess the mediating effects of BMI and hsCRP on the associations of FT3 and FT4 with BP [31] by SPSS AMOS version 21 (IBM Corporation, Armonk, NY, USA). Two parts of path analyses were examined: one related to the associations of FT3 and FT4 with BMI and hsCRP, the other related to the associations of BMI and hsCRP with SBP and MAP Levels. The correlation was assessed by computing standardized regression coefficients in each path. All analyses were performed by IBM SPSS Statistics 21.0 for Windows (IBM, Inc., Armonk, NY, USA). Two–tailed p value < 0.05 was statistically significant.

Results

Basic characteristics

The basic characteristics of men (n = 2925) and women (n = 1681) were showed in Table 1. The mean ages of men and women were 60.3 and 61.8 years, respectively. Lower SBP, FT3, FT4, hsCRP concentrations and lower proportion of smoke or drink were observed in women than in men (all p < 0.05).

Table 1.

The baseline characteristics of the study participants with coronary artery disease

Variables	Men (n = 2925)	Women (n = 1681)	pb value
Age, years	60.3 ± 10.06	61.8 ± 8.80	< 0.001
BMI, kg/m2	26.03 ± 16.31	25.50 ± 13.72	0.267
Smoker, n (%)	733 (25.1)	25 (1.5)	< 0.001
Alcohol drinker, n (%)	620 (21.2)	15 (0.9)	< 0.001
Antihypertensive agent user, n (%)	1329 (45.4)	779 (46.3)	0.553
Hypoglycaemic agent user, n (%)	373 (12.8)	280 (16.7)	< 0.001
Lipid-lowering agent user, n (%)	86 (2.9)	49 (2.9)	0.961
SBP, mmHg	122.53 ± 11.29	121.87 ± 8.80	0.049
DBP, mmHg	71.49 ± 7.70	71.34 ± 7.40	0.511
MAP, mmHg	88.50 ± 7.68	88.17 ± 7.34	0.149
FT3, pg/mL	3.17 ± 0.32	3.11 ± 0.30	< 0.001
FT4, ng/dL	0.92 ± 0.14	0.91 ± 0.14	0.027
FT3/FT4 ratio	3.53 ± 0.63	3.50 ± 0.62	0.197
hsCRPa, %	1.26 (0.53, 3.59)	1.15 (0.50, 2.91)	0.009

BMI Body mass index, DBP Diastolic blood pressure, FT3 Free triiodothyronine, FT4 Free thyroxine, hsCRP High sensitivity C reactive protein, MAP Mean arterial pressure, SBP Systolic blood pressure 

^aValues were median (interquartile range), p values were calculated using Mann–Whitney U test

^bp values were calculated by Student’s t test for the continuous variables and Chi-square test for categorical variables

Correlation of FT3 and FT4 with BP

As shown in Table 2, the weak correlations of serum FT3 and FT4 with BP (SBP, DBP and MAP) were found using the partial correlation analysis adjusted for age and sex (r: 0.030–0.053). Null correlation was observed between FT3/FT4 ratio and BP.

Table 2.

Relationships between serum FT3, FT4 and FT3/FT4 ratio and blood pressure^a

	FT3		FT4		FT3/FT4 ratio
	r′	p	r′	P	r′	p
SBP	0.049	< 0.001	0.053	< 0.001	−0.014	0.355
DBP	0.030	0.044	0.032	0.029	−0.003	0.837
MAP	0.044	0.003	0.048	< 0.001	−0.009	0.557

DBP Diastolic blood pressure, FT3 Free triiodothyronine, FT4 Free thyroxine, MAP Mean arterial pressure, SBP Systolic blood pressure

^aPartial correlation analysis, adjusted for age and sex

Associations of FT3 and FT4 with BP

Overall, serum FT3 and FT4 tended to be positively associated with SBP and MAP, but not with DBP. FT3/FT4 ratio exhibited a null association with BP (SBP, DBP and MAP). Among total populations (Table 3), the sex and age adjusted mean differences of serum FT3 and FT4 between quartiles 4 and 1 were 1.90 mmHg (1.57%) and 2.28 mmHg (1.88%) for SBP, 1.14 mmHg (1.30%) and 1.14 mmHg (1.29%) for MAP. After further adjusting for the other potential covariates in Model 2, significant differences of 2.09 mmHg (1.73%) and 2.10 mmHg (1.73%) for SBP and 1.16 mmHg (1.32%) and 1.10 mmHg (1.25%) for MAP were observed between quartiles 4 and 1 of FT3 and FT4 (p–trends: <0.001 and 0.002). There was a marginal relationship between FT3 and DBP in models 1 and 2 (p–trends: 0.036 and 0.062).

Table 3.

Mean blood pressure levels according to quartiles of serum FT3 and FT4 in all participants (Mean ± SE)

Variables		Quartiles by FT3 and FT4				Diff	% Diff	p-trend
		Q1	Q2	Q3	Q4
FT3
n		1137	1162	1162	1145
SBP, mmHg	Model 1	120.84 ± 0.33	122.92 ± 0.33b	122.63 ± 0.33b	122.74 ± 0.33b	1.90	1.57	< 0.001
	Model 2	120.76 ± 0.33	122.88 ± 0.33b	122.64 ± 0.33b	122.85 ± 0.33b	2.09	1.73	< 0.001
DBP, mmHg	Model 1	70.92 ± 0.23	71.66 ± 0.22	71.49 ± 0.22	71.68 ± 0.22	0.76	1.07	0.036
	Model 2	70.95 ± 0.23	71.68 ± 0.22	71.48 ± 0.22	71.65 ± 0.23	0.70	0.99	0.062
MAP, mmHg	Model 1	87.56 ± 0.22	88.74 ± 0.22b	88.54 ± 0.22a	88.70 ± 0.22a	1.14	1.30	< 0.001
	Model 2	87.55 ± 0.23	88.73 ± 0.22b	88.53 ± 0.22a	88.71 ± 0.23a	1.16	1.32	< 0.001
FT4
n		1157	1179	1089	1133
SBP, mmHg	Model 1	121.48 ± 0.32	122.90 ± 0.32a	122.70 ± 0.33a	123.76 ± 0.32b	2.28	1.88	< 0.001
	Model 2	121.57 ± 0.32	122.93 ± 0.32a	122.68 ± 0.33	123.67 ± 0.32b	2.10	1.73	< 0.001
DBP, mmHg	Model 1	71.45 ± 0.22	71.68 ± 0.22	71.32 ± 0.23	72.02 ± 0.22	0.57	0.80	0.167
	Model 2	71.42 ± 0.22	71.68 ± 0.22	71.33 ± 0.23	72.05 ± 0.23	0.63	0.88	0.129
MAP, mmHg	Model 1	88.12 ± 0.22	88.75 ± 0.22	88.45 ± 0.23	89.26 ± 0.22a	1.14	1.29	0.002
	Model 2	88.14 ± 0.22	88.76 ± 0.22	88.44 ± 0.23	89.24 ± 0.22a	1.10	1.25	0.002
FT3/FT4
n		1122	1120	1123	1121
SBP, mmHg	Model 1	122.37 ± 0.32	122.74 ± 0.32	122.61 ± 0.32	121.83 ± 0.32	−0.54	−0.44	0.219
	Model 2	122.27 ± 0.32	122.68 ± 0.32	122.67 ± 0.32	121.92 ± 0.32	−0.35	−0.29	0.459
DBP, mmHg	Model 1	71.51 ± 0.22	71.29 ± 0.22	71.59 ± 0.22	71.69 ± 0.22	0.18	0.25	0.409
	Model 2	71.55 ± 0.23	71.31 ± 0.23	71.56 ± 0.23	71.65 ± 0.23	0.10	0.14	0.587
MAP, mmHg	Model 1	88.45 ± 0.22	88.44 ± 0.22	88.59 ± 0.22	88.40 ± 0.22	−0.05	−0.06	0.996
	Model 2	88.45 ± 0.22	88.44 ± 0.22	88.60 ± 0.22	88.41 ± 0.22	−0.04	−0.05	0.966

Abbreviations were shown in Table 1 plus SE Standard error and Q Quartile. Model 1 adjusted for sex and age, Model 2 further adjusted for BMI, smoking status, alcohol consumption, hypoglycemic, antihypertensive agents and lipid lowering agents use

Diff: difference of thyroid hormone levels between Q4 and Q1= (Q4 – Q1), % Diff: Percentage difference of thyroid hormone levels between Q4 and Q1= (Q4 – Q1)/Q1 × 100%

^ap < 0.05 compared with Q1

^bp < 0.001 compared with Q1

In analyses stratified by sex, graded and positive associations of serum FT3 and FT4 with SBP were observed in both men and women according to model 2 (all p–trends < 0.05) (Table 4). No statistically significant interactions between serum FT3, FT4 and sex were observed (all p-interaction > 0.05). In men, there were the significant positive associations of serum FT3 and FT4 with MAP and SBP (p–trends: <0.001 and 0.017). In women, serum FT3 and FT4 were positively associated with SBP. No significant associations were observed between FT3/FT4 ratio and BP (SBP, DBP and MAP) in either group.

Table 4.

Mean blood pressure levels according to quartiles of serum FT3 and FT4 in men and women, respectively (Mean ± SE)

Variables		Quartiles by FT3 and FT4				Diff	% Diff	p-trend	p-interaction
		Q1	Q2	Q3	Q4
FT3
SBP, mmHg	Men	121.20 ± 0.42	122.84 ± 0.42a	122.78 ± 0.41a	123.28 ± 0.42a	2.08	1.72	< 0.001	0.497
	Women	120.04 ± 0.53	122.89 ± 0.52b	122.36 ± 0.54a	122.16 ± 0.53a	2.12	1.77	0.015
DBP, mmHg	Men	71.01 ± 0.29	71.60 ± 0.29	71.56 ± 0.28	71.81 ± 0.29	0.80	1.13	0.074	0.912
	Women	70.84 ± 0.36	71.78 ± 0.36	71.36 ± 0.37	71.38 ± 0.36	0.54	0.76	0.459
MAP, mmHg	Men	87.74 ± 0.29	88.68 ± 0.29	88.63 ± 0.28	88.96 ± 0.29a	1.22	1.39	0.005	0.710
	Women	87.24 ± 0.36	88.79 ± 0.35a	88.36 ± 0.37	88.31 ± 0.36	1.07	1.23	0.087
FT4
SBP, mmHg	Men	121.47 ± 0.40	123.17 ± 0.41a	122.75 ± 0.42a	123.53 ± 0.41a	2.06	1.70	0.002	0.669
	Women	121.80 ± 0.53	122.59 ± 0.49	122.52 ± 0.53	123.86 ± 0.53a	2.06	1.69	0.010
DBP, mmHg	Men	71.43 ± 0.28	71.71 ± 0.29	70.94 ± 0.29	72.21 ± 0.29c	0.78	1.09	0.218	0.146
	Women	71.41 ± 0.37	71.65 ± 0.35	72.00 ± 0.37	71.74 ± 0.37	0.33	0.46	0.414
MAP, mmHg	Men	88.11 ± 0.27	88.86 ± 0.28	88.21 ± 0.28	89.32 ± 0.28ac	1.21	1.37	0.017	0.561
	Women	88.20 ± 0.37	88.63 ± 0.34	88.84 ± 0.37	89.09 ± 0.36	0.89	1.01	0.079
FT3/FT4
SBP, mmHg	Men	122.30 ± 0.41	122.69 ± 0.41	123.07 ± 0.41	122.26 ± 0.41	−0.04	−0.03	0.891	0.448
	Women	122.20 ± 0.51	122.62 ± 0.51	121.98 ± 0.51	121.36 ± 0.51	−0.84	−0.69	0.177
DBP, mmHg	Men	71.61 ± 0.29	71.16 ± 0.29	71.72 ± 0.29	71.73 ± 0.29	0.12	0.17	0.467	0.681
	Women	71.45 ± 0.37	71.55 ± 0.37	71.28 ± 0.37	71.54 ± 0.37	0.09	0.13	0.998
MAP, mmHg	Men	88.51 ± 0.28	88.33 ± 0.28	88.84 ± 0.28	88.58 ± 0.28	0.07	0.08	0.574	0.538
	Women	88.34 ± 0.36	88.57 ± 0.35	88.18 ± 0.35	88.14 ± 0.36	−0.20	−0.23	0.550

Abbreviations were shown in Table 1 plus SE: standard error and Q: quartile. Age, BMI, smoking status, alcohol consumption, hypoglycemic, antihypertensive agents and lipid lowering agents use were adjusted. Diff: difference of thyroid hormone levels between Q4 and Q1= (Q4 – Q1)

% Diff: Percentage difference of thyroid hormone levels between Q4 and Q1= (Q4 – Q1)/Q1 × 100%

^ap < 0.05 compared with Q1 

^bp < 0.001 compared with Q1

^cp < 0.05 compared with Q3

Associations of FT3 and FT4 with HTN prevalence

We further explored the associations of FT3 and FT4 with HTN prevalence (Table 5). After controlling for potential confounders in model 2, an intermediate FT3 concentration (Q3 vs. Q1) was found to be associated with a higher risk of HTN (odds ratio [OR]: 1.32, 95% confidence interval [CI]: 1.03, 1.69). Similar positive association was also observed between FT4 and HTN prevalence (OR: 1.29, 95%CI: 1.01, 1.66). There were no significant associations between FT3/FT4 ratio and BP (SBP, DBP and MAP) (all p > 0.05).

Table 5.

Associations of serum FT3, FT4 and FT3/FT4 ratio with the risk of hypertension prevalence

	Risk of hypertension
	Q1	Q2	Q3	Q4
FT3
Model 1	1	1.28 (1.08, 1.52)a	1.25 (1.06, 1.48)a	1.14 (0.97, 1.35)
Model 2	1	1.42 (1.11, 1.81)a	1.32 (1.03, 1.69)a	1.16 (0.91, 1.49)
FT4
Model 1	1	1.03 (0.87, 1.21)	1.13 (0.95, 1.33)	1.07 (0.91, 1.27)
Model 2	1	1.15 (0.90, 1.47)	1.29 (1.01, 1.66)a	1.21 (0.95, 1.56)
FT3/FT4
Model 1	1	1.15 (0.97, 1.36)	1.09 (0.92, 1.30)	0.99 (0.84, 1.18)
Model 2	1	1.16 (0.90, 1.48)	1.07 (0.83, 1.37)	0.85 (0.66, 1.10)

Abbreviations were shown in Table 1 plus Q: quartile

^ap < 0.05 compared with Q1

Path analysis

Path analysis indicated that BMI and hsCRP did not have direct effect on SBP and MAP. In Fig. 1, the mediating effects of BMI and hsCRP on “FT3 (or FT4) –SBP (or MAP)” associations were not observed. FT3 showed significant associations with BMI and hsCRP (standardized r_BMI: 0.055 and r_hsCRP: − 0.063, all p < 0.001). Additionally, we also found that FT4 was positively related to hsCRP (standardized r_hsCRP: 0.149, p < 0.001). Nonetheless, BMI and hsCRP were not associated with SBP and MAP levels (all p > 0.05) (Fig. 1).

Fig. 1.

Path analyses of associations of FT3 and FT4 levels and mediators (hsCRP and BMI) with BP (SBP and MAP) in patients with CAD. BMI, body mass index; BP, blood pressure; CAD, coronary artery disease; FT3, free triiodothyronine; FT4, free thyroxine; hsCRP, high-sensitivity C-reactive protein; MAP, mean arterial pressure; SBP, systolic blood pressure. ^**p < 0.001

Discussion

In this hospital–based cross–sectional study of adults with CAD, we observed graded and positive associations of serum FT3 and FT4 with SBP and MAP, but not DBP. The positive associations appeared to be not mediated (at least in part) by BMI and hsCRP. Similar to previous study [32], no statistically significant interactions between serum FT3, FT4 and sex were observed, which suggested that there were no sex-related differences in the associations of FT3 and FT4 with BP. To the best of the authors’ knowledge, our study was the first to calculate the associations of serum FT3 and FT4 with BP in adults with CAD. Our results suggested that FT3 and FT4 might influence BP levels, and particularly SBP and MAP.

FT3, the active form of thyroid hormone, exerted an essential role in modulating energy expenditure [33]. Accordingly, interest in the health benefits of FT3 is increasing. One cross–sectional study of 1036 euthyroid community–based persons reported that SBP and DBP increased according to the increase of FT3 tertiles (SBP_{tertile 3 vs. 1}: 125 mmHg vs. 122 mmHg; DBP_{tertile 3 vs. 1}: 77 mmHg vs. 73 mmHg) in women, but not in men [14]. Taneichi et al. observed that FT3 was correlated with SBP and DBP in 301 type 2 diabetic subjects with euthyroidism (r_SBP = 0.192, r_DBP = 0.141) [15]. Chen et al. also found that FT3 was correlated with SBP and DBP in 880 school–aged subjects without overt thyroid disease (total: r_SBP=0.129, r_DBP=0.129; men: r_SBP=0.166, r_DBP=0.165) [16]. However, these associations were decreased to non–significance in women [16]. Roos et al. [12] and Xu et al. [17] found that FT3 was positively correlated with SBP, but not with DBP. Similar results also were observed in euthyroid adolescents [18]. In present study, the positive relationships were observed between serum FT3 and SBP and MAP after controlling for a wide range of potential confounding variables. Small DBP changes might attenuate the correlation between FT3 and DBP because of arterial elasticity was decreased with the increase of age. Wahab et al. found that obesity predisposed to hypertension [34], and higher BMI was associated with higher SBP and DBP [35]. Thus, low BMI and BP levels might attenuate the relationship between FT3 and SBP and DBP in women [14]. Besides, a small sample size might induce null associations of FT3 with SBP and DBP in subjects (n_man=445 [14] and n_women=541 [16]).

In terms of FT4, persons with higher serum FT4 concentration tended to exhibit a higher SBP and MAP in our study. Similarly, a higher serum FT4 was found to be associated with higher SBP and DBP in both men and women (men: SBP_{Q5 vs. Q1} = 121.9 mmHg vs. 120.9 mmHg, DBP_{Q5 vs. Q1} = 76.3 mmHg vs. 74.9 mmHg; women: SBP_{Q5 vs. Q1} = 113.2 mmHg vs. 111.8 mmHg, DBP_{Q5 vs. Q1} = 71.6 mmHg vs. 70.4 mmHg, all p–trend < 0.001) [21]. Other studies also showed the positive correlations of FT4 with SBP or DBP [12, 15, 17, 19–26]. Nevertheless, Mehran et al. found that there were no significant differences in SBP and DBP based on the FT4 tertiles [23]. Null correlations also were observed in other studies [16, 17, 22, 25, 27, 28]. Even Shin et al. pointed out that FT4 was associated with lower SBP and DBP (men: SBP_{Q4 vs. Q1} = 112.2 mmHg vs. 113.9 mmHg, DBP_{Q4 vs. Q1} = 72.7 mmHg vs. 73.6 mmHg) in euthyroid Korean men, but not in women [26]. Inverse association tended to be obtained due to a high iodine intake in Koreans, which might affect thyroid hormone reference ranges. Le et al. also found that FT4 was negatively correlated with SBP in euthyroid adolescents (r = – 0.06, p = 0.0335) [18]. It is possible that thyroid hormones within the euthyroid range exert paradoxical effects on central and peripheral tissues Due to the different expression of type 1, type 2 deiodinases and thyroid hormone receptor isoforms in central and peripheral tissues, which modulates the conversion of T4 to T3 in the thyroid gland, kidney, liver and brain [36, 37]. Additionally, other reasons such as the differences in statistic method (multivariate analyses of covariance [21, 23, 26] & this study, partial correlation [26], spearman correlation [15, 16, 28], pearson correlation [19, 24], linear regression [12, 17, 18, 22, 23, 25, 27] and generalized estimating equations [20]), different iodine status of the population, definitions for metabolic syndrome, inclusion criteria, ethnicity, and adjustments in the analyses might also explain the discrepancies among these studies.

Previous studies indicated that whether high normal T3 concentrations increased the risk of metabolic syndrome remains unclear [38, 39]. Epidemiological data reported the nonlinear relationships between high normal TSH levels and metabolic parameters [40–42]. Thus, it is still debatable whether mild thyroid dysfunction induce abnormal BP in human. High circulating T3 concentrations might stimulate an adaptive protective response against the detrimental influences of endocrine metabolism (e.g., obesity and insulin resistance) [43]. Meanwhile, moderate increases in FT3 and T3 levels might induce an increase in energy expenditure [44] which ultimately might lead to the increased risk of hypertension [45]. Furthermore, possible nonlinear “exposure-outcome variables” relations tend to be observed in calculations using continuous variables compared to calculations using categorized variables (analysis of covariance based on the serum FT3 and FT4 quartiles [this present]) [46]. Although we could not characterize the shape of the FT3 (or FT4)–BP (or hypertension) relations due to formal nonlinearity testing was not performed, the results suggested that an intermediate FT3 (or FT4) was important in BP modulation. However, further prospective study is needed to confirm this hypothesis.

Mechanism

The mechanisms underlying the associations of FT3 and FT4 with BP were not fully understood. The positive associations of FT3 and FT4 with BP might be partly explained by some different mechanisms. First, one review indicated that the cardiac myocyte β–adrenergic system of hyperthyroidism and adrenergic hyperactivity might be modulated via the adenylate cyclase, guanine nucleotide regulatory proteins, and β1–adrenergic receptor [47]. Meanwhile, adrenergic system was considered as the major regulator of cardiovascular vascular function, and it was performed by the specific receptors of endothelial surface via local and systemic release of catecholamines [48, 49]. Thus, it is possible that thyroid hormones elevate BP levels through the regulation of adrenergic system. Moreover, Prisant et al. found that SBP was increased by hyperthyroidism (elevated total triiodothyronine levels) via reducing systemic vascular resistance, enhancing heart contractility and heart rate, and increasing cardiac output [50]. Gumieniak et al. observed that the outliers of FT4 might alter BP salt sensitivity [51], which ultimately led to the onset of hypertension. Second, thyroid hormone could increase BP by stimulating lipid synthesis and modulating inflammation [52, 53]. However, our study did not find the mediating effects of BMI and hsCRP on the “FT3 or FT4–BP” associations. A loss of power due to a small sample size might decay the above mediating effects. And other inflammation markers might be mediated in the modulating role of FT3 or FT4 on BP. Thirdly, FT4 could increase BP by high pituitary–thyroid axis set point [54, 55]. Finally, thyroid hormone was able to elevate BP by regulating adrenomedullin, endothelin, vasopressin, atrial natriuretic factor and catecholamines [13].

Strengths and limitations

There several strengths in the present study. This study firstly assessed the association of FT3 and FT4 and their ratio with BP in Chinese adults with CAD and further calculated the mediating effects of hsCRP and BMI on the “FT3 or FT4–BP” associations. Besides, the results of this study were accuracy and reliability based on the relatively large sample size. Moreover, the probability of falsely significant results was avoided by assessing different BP indices (SBP, DBP and MAP) and hypertension. Furthermore, the important confounders were considered and both FT3 and FT4 were assessed in the analysis. Nonetheless, some Limitations still need to be considered. First, the cause and effect relationships were unable to be discerned Due to a cross-sectional design in this study. Second, although this study illustrated the small differences of 1.73% in SBP and 1.25% in MAP between quartiles, our findings were of great clinical value. Li et al. pointed out that the risk of multivessel CAD in patients with stable chest pain suspected of CAD increased significantly as per 1 mmHg increased in brachial pulse pressure (OR = 1.02, P = 0.0002) [56]. As for cardiovascular disease events, Tsai et al. reported that each 1 mmHg increased in systolic blood pressure variability might Lead to the hazard ratio of the cardiovascular death was 1.06 (95% CI: 1.03, 1.09) [57]. Du et al. also observed that per 20 mm Hg increased in nighttime SBP was significantly related to the risk of all–cause mortality or total nonfatal cardiovascular disease events (hazard ratio = 1.775, 95%CI: 1.256, 2.507) [8]. Thus, the minor changes in BP might have crucial impacts on adverse cardiovascular events. Large randomized clinical trials and longitudinal cohort studies are needed to confirm our findings. Third, the absence of CAD drugs using in the course of treatment might lead to the overestimation of the underlying associations in our study. However, the potential confounders (e.g., lipid–lowering, hypoglycemic and antihypertensive agent use) were considered, which might attenuate the effect of the absence of these data. Fourth, it has been reported that BP was substantially influenced by sex and age [58–60]; meanwhile, day length and seasonal BP variations were the important treatment target for the prevention of cardiovascular disease [61, 62]. Nevertheless, the effect of these potential confounding factors on “FT3 (or FT4)-BP” associations in our study was attenuated because blood samples were collected from both men and women in different seasons without being paired from the same one. Finally, it was difficult to exclude potential confounding factors (e.g., dietary iodine and physical activity) even though we had adjusted for many of the related covariables. Furthermore, no attempt was made to assess thyroid related medicines intake, which might attenuate the associations of FT3 and FT4 with BP in present study.

Our findings were of potentially great clinical significance because thyroid dysfunction was associated with cardiovascular disorders. It has been reported that hypothyroidism was considered as an crucial risk factor for atherosclerosis and myocardial infarction [63]. Luboshitzky et al. also observed the higher hypertension prevalence in the subclinical hypothyroidism group than that in the normal control group [64]. In line with these studies, our results elucidated that the reasonable reduction of FT3 and FT4 levels might be an additional protective factor for HTN prevention in CAD patients. Thus, thyroid hormones modulation might be beneficial for CAD patients at risk of abnormal BP. Nevertheless, it is important to explore the nonlinear “FT3 (or FT4) –BP (or HTN)” relations in future interventional and experimental studies.

Conclusions

In summary, our study found that FT3 and FT4 were positively associated with SBP and hypertension prevalence in Chinese adults with CAD. It is plausible that FT3 and FT4 are more sensitive to SBP. Replication of these findings is warranted in large cohort studies in general populations.

Authors’ contributions

Nan Lu designed the research. Hongli Dong, Ping Hu and Jie Wang conducted research. Hongli Dong analyzed the data and wrote the paper. Nan Lu critically revised the manuscript. Nan Lu had primary responsibility for final content. All authors read and approved the final manuscript.

Funding

This work was partly supported by the Medical Talent Project of Nantong Maternal and Child Health Hospital (No. YQR202304). The funders had no role in the design, analysis or writing of this article.

Data availability

All data generated or analysed during this study are included in this published article.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Wuhan Asia Heart Hospital (No. 2016-B008). Informed consent has been obtained from all subjects.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.