Short-term administration of a small thyroxine dose to euthyroid type 2 diabetic patients improves the fasting lipoprotein profile^#

^#These authors contributed equally to this work.

ABSTRACT

Background: Although several studies have assessed the association between thyroid hormones and dyslipidaemia, whether influencing thyroid function improves the lipid profile in euthyroid diabetic patients has not been studied.

Methods: Fasting lipids were assessed in 11 euthyroid, treatment naive patients with type 2 diabetes (T2DM) and a micronodular texture of the thyroid gland (age: 43 ± 3.8 years, body mass index (BMI) 27.5 ± 1.4 kg/m², triiodothyronine (T3) 119 ± 5.7 ng/dl, thyroxine (T4) 8.13 ± 0.46 μg/dl, thyroid- stimulating hormone (TSH) 1.51 ± 0.14 μIU/ml, free thyroxine (FT4) 1.272 ± 0.047 ng/dl) before and after administration of 50 μg of T4 once daily for 2 months. A placebo was given to 11 age, sex and BMI-matched euthyroid, treatment naive patients with T2DM. Care was taken to avoid even subclinical hyperthyroidism.

Results: TSH fell significantly post-treatment (1.51 ± 0.11 vs. 0.79 ± 0.11 μIU/ml, p < 0.0001), but remained within the reference range. Total cholesterol (212 ± 21 vs. 158 ± 10 mg/dl, p = 0.003), low-density lipoprotein cholesterol (146 ± 17 vs. 112 ± 9 mg/dl, p = 0.007), high density lipoprotein cholesterol (51 ± 4 vs. 40 ± 3 mg/dl p = 0.001), triglycerides (93 ± 13 vs. 72 ± 8 mg/dl, p = 0.015), apolipoprotein A1 (167 ± 15 vs. 127 ± 8 mg/dl, p = 0.004), apolipoprotein B (101 ± 13 vs. 72 ± 7 mg/dl, p = 0.009) and lipoprotein (a) (60 ± 15 vs. 41 ± 11 mg/dl p = 0.009) all fell significantly after T4 administration for 2 months. No changes were observed in the placebo group.

Conclusions: Small doses of T4 administered to euthyroid patients with T2DM significantly improved lipid levels. This could contribute to a reduced risk of macrovascular complications.

KEYWORDS:

• Dyslipidaemia
• type 2 diabetes
• euthyroid
• thyroxine

Introduction

Diabetic dyslipidaemia has been used to describe the pathophysiology surrounding the effects of insulin resistance on abnormal lipid levels [1–3]. This concept supports the hypothesis that defects in insulin action and glucose dysregulation can lead to a deranged lipoprotein composition and concentration in the bloodstream. The subsequent changes in lipids, secondary to a state of glucose intolerance, add to the risk of progression of atherosclerosis and macrovascular complications in insulin resistant individuals [1–3]. Even slight changes in lipid levels in diabetic patients are associated with a substantial increase in cardiovascular disease, more than in the general population [1–3].

Thyroid hormones influence all major metabolic pathways. Their most important action is to increase in basal energy expenditure [4,5]. Regarding lipid metabolism, thyroid hormones affect the synthesis, mobilisation and degradation of lipids, although degradation is influenced more than synthesis [6]. In particular, they induce the 3-hydroxy-3-methylglutaryl-coenzyme A reductase, which is the first step in cholesterol biosynthesis, they upregulate the low-density lipoprotein (LDL) receptors by controlling the sterol regulatory element-binding protein-2 (SREBP-2) and LDL receptor gene activation. They may also influence high density lipoprotein (HDL) and triglyceride (TG) metabolism by increasing cholesteryl ester transfer protein (CETP), lipoprotein lipase and hepatic lipase activity [7].

Hypothyroidism is more often associated with abnormalities of lipid metabolism which are one of the major risk factors of coronary artery disease (CAD) in these patients [4].

The association between thyroid function and dyslipidaemia is well documented not only in overt hypothyroidism, but also in subclinical disorders with low-normal thyroid function and a thyroid-stimulating hormone (TSH) level within the normal range [8]. Replacement therapy with L-thyroxine (T4) is beneficial, improving lipid metabolism abnormalities [9].

Each person seems to have a unique set point for hypothalamo-pituitary-thyroid axis function [10]. This specific set point variation within the euthyroid range has recently been recognised as a contributing factor for cardiometabolic effects, and this notion has raised a question about the actual definition of “normal” thyroid function in terms of TSH levels and the interaction with metabolism [10–12]. Low-normal thyroid function may contribute to the pathogenesis of cardiovascular disease as it relates to several traditional or non-traditional risk factors, like the components of the metabolic syndrome, hepatic steatosis/steatohepatitis and lipid abnormalities [13,14].

Although several studies evaluated the association between thyroid hormones and lipid abnormalities, few have actually assessed this association in patients with type 2 diabetes (T2DM) [15–17]. In a recent study, a low thyroid function was positively associated with a lipid dysregulation in a diabetic population [18]. Whether the latter could be restored in humans by altering thyroid function has not been studied.

The present study was undertaken in euthyroid patients with T2DM with a micronodular goitre, to assess the effect of administration of small subthyrotoxic doses of T4 on fasting lipid parameters.

Patients and methods

Patients

Eleven euthyroid, treatment-naive, type 2 diabetic patients with a micronodular texture of the thyroid gland, [aged 42 ± 3.8 years, body mass index (BMI) 27.5 ± 1.4 kg/m², triiodothyronine (T3) 119 ± 5.7 ng/dl (1.832 ± 0.087 nmol/l), T4 8.13 ± 0.46 μg/dl (104.6 ± 5.92 nmol/l) thyroid-stimulating hormone (TSH) 1.51 ± 0.14 μIU/ml, free thyroxine (FT4) 1.272 ± 0.047 ng/dl], were studied before and after administration of 50 μg of T4 once daily for 2 months.

In parallel, a placebo group was also studied. Eleven euthyroid treatment-naïve patients with T2DM and a micronodular texture of the thyroid gland, matched for age, sex, body mass index (BMI) and basal thyroid function, [aged 43 ± 3.74 years, BMI 27.8 ± 1.28 kg/m², T3 84.44 ± 3.013 ng/dl (1.3 ± 0.046 nmol/l), FT4 1.3 ± 0.046 ng/dl, TSH 1.554 ± 0.181 μIU/ml] were studied before and after administration of a placebo (placebo for 50 μg thyroxine, Unipharma, Greece) once daily for 2 months.

We used a small dose of T4 for two months under strict surveillance so as to maintain thyroid function within the euthyroid range. Special care was taken so that not even subclinical hyperthyroidism developed, as the latter could influence metabolic regulation. The study was approved by the hospital ethics committee and the patients gave their informed consent.

Experimental protocol

This is an open-label, randomised and placebo-controlled intervention. All subjects were admitted to the hospital at 07.00 h after an overnight fast and had a forearm vein catheterised. Blood samples were drawn at −30 and 0 min for measurements of thyroid hormones, total cholesterol (TC), LDL-C (low-density lipoprotein cholesterol), HDL-C (high density lipoprotein cholesterol), triglycerides (TG), apolipoprotein (Apo) A1, Apo B and lipoprotein (a) [Lp(a)], fasting plasma glucose, fasting plasma insulin and glycated haemoglobin. The values obtained from the two samples (at −30 and 0 min) were averaged to give a 0′ time value. After the first visit, treatment with 50 μg of T4 once daily or placebo was initiated for two months. Then a second identical visit was repeated. Special care was taken in order to avoid the induction of subclinical hyperthyroidism (i.e. TSH below 0.27 μIU/ml) as it has recently been shown that the latter is also an insulin-resistant condition [19].

Laboratory assessment

Thyroid hormones were measured by chemiluminescence immunometric assays (Roche Diagnostics, GmbH Manheim, Germany).

Plasma TC, HDL-C, Apo A, Apo B, Lp(a) and TG concentrations were assayed enzymatically on a Roche Modular Analytics (Modular Analytics EE, Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer’s specifications. LDL-C was calculated using the Friedewald equation. Fasting insulin (Linco Research, St. Charles, MO) and glucose (Yellow Springs Instruments, Yellow Springs, OH) were also assessed.

Statistical analysis

Grouped data are expressed as mean ± SEM. Differences between pre- and post-treatment values within group were tested using a two-sided paired t-test. The comparison between the two groups of diabetic subjects was carried out using unpaired t-tests. The statistical analysis was performed using Graph Pad In Stat (San Diego, CA, USA) software.

Results

TSH levels fell significantly post-treatment (1.51 ± 0.11 vs. 0.79 ± 0.11 μIU/ml, p < 0.0001) but remained within the normal range. FT4 values also changed significantly (1.272 ± 0.046 vs. 1.454 ± 0.05 ng/dl, p = 0.0016). No significant changes occurred in the placebo group.

The treatment protocol was adjusted so that the dose of T4 was small enough and could not cause subclinical hyperthyroidism. The dose was fixed to a level of 50 μg for all subjects, since they had comparable body weight and basal TSH levels.

All lipid parameters fell significantly post-treatment in the group that received thyroxine (Table 1). No changes were detected in the placebo group (Table 2). The comparison of the changes (deltas) in each lipid parameter between the treatment and the placebo group revealed significant differences in all values (Table 3). Improvements in glucose and insulin values as well as in glycated haemoglobin (HbA1c) were also observed in the group that received thyroxine (Tables 3 and 4). These data were recently published in detail [20].

Table 1. Fasting lipid values before and after administration of 50 μg of thyroxine (treatment group). All values are expressed as mean and standard error of mean (mean ± SEM).

Lipid Values	Pre-treatment	Post-treatment	p value
TC (mg/dl)	212 ± 21	158 ± 10	0.0028
LDL-C (mg/dl)	147 ± 17	112 ± 9	0.0071
HDL C (mg/dl)	50 ± 4	40 ± 3	0.0013
TG (mg/dl)	93 ± 13	72 ± 8	0.0154
Apo AI (mg/dl)	167 ± 15	128 ± 9	0.0041
Apo B (mg/dl)	101 ± 13	72 ± 7	0.0092
Lp(a) (mg/dl)	60 ± 16	42 ± 11	0.0091

Notes: TC: total cholesterol, LDL-C: low-density lipoprotein cholesterol, HDL-C: high density lipoprotein cholesterol, TG: triglycerides, Apo AI: apolipoprotein AI, Apo B: apolipoprotein B, Lp(a): lipoprotein (a).

Table 2. Fasting lipid values before and after administration of placebo. All values are expressed as mean and standard error of mean (mean ± SEM).

Lipid Values	Pre-treatment	Post-treatment	p value
TC (mg/dl)	215 ± 6	212 ± 10	0.383
LDL-C (mg/dl)	141 ± 6	139 ± 8	0.654
HDL C (mg/dl)	57 ± 3	55 ± 2	0.282
TG (mg/dl)	84 ± 6	85 ± 8	0.8
Apo AI (mg/dl)	140 ± 5	136 ± 4	0.351
Apo B (mg/dl)	93 ± 4	92 ± 4	0.757
Lp(a) (mg/dl)	45 ± 8	46 ± 8	0.344

Notes: TC: total cholesterol, LDL-C: low-density lipoprotein cholesterol, HDL-C: high density lipoprotein cholesterol, TG: triglycerides, Apo AI: apolipoprotein AI, Apo B: apolipoprotein B, Lp(a): lipoprotein (a).

Table 3. Metabolic variables before and after administration of 50 μg of thyroxin (treatment group). All values are expressed as mean and standard error of mean (mean ± SEM).

Variables	Pre-treatment	Post-treatment	p value
BMI kg/m^2	27.5 ± 1.4	27.4 ± 1.4	0.6497
FPG mmol/l	7.7 ± 0.2	6.6 ± 0.3	0.0001
FPI μU/ml	16.26 ± 2	8.35 ± 0.9	0.004
HOMA-IR	5.55 ± 0.68	2.38 ± 0.2	0.0013
HbA1c %	8.3 ± 0.2	7.8 ± 0.2	0.0003

Notes: BMI: body mass index, FPG: fasting plasma glucose, FPI: fasting plasma insulin, HOMA-IR: homeostatic model assessment of insulin resistance, HbA1c: glycated haemoglobin.

Table 4. Metabolic variables before and after administration of placebo. All values are expressed as mean and standard error of mean (mean ± SEM).

Variables	Pre-treatment	Post-treatment	p value	p basal values treatment group vs placebo
BMI kg/m^2	27.8 ± 1.3	27.7 ± 1.2	0.4561	0.8683
FPG mmol/l	7.6 ± 0.2	7.5 ± 0.17	0.361	0.8931
FPI μU/ml	15.63 ± 1.7	14.9 ± 1.6	0.099	0.8166
HOMA-IR	5.27 ± 0.54	4.96 ± 0.49	0.061	0.7623
HbA1c %	8.2 ± 0.1	8.1 ± 0.1	0.221	0.8359

Notes: BMI: body mass index, FPG: fasting plasma glucose, FPI: fasting plasma insulin, HOMA-IR: homeostatic model assessment of insulin resistance, HbA1c: glycated haemoglobin.

For this specific time period, our patients did not demonstrate any weight change, sinus rhythm alteration, blood pressure changes or any other symptom that could imply a clinically relevant effect of the intervention.

Discussion

Thyroid hormones affect a wide range of metabolic functions including regulation of lipid, carbohydrate, protein, electrolyte and mineral metabolisms. In patients with overt hypothyroidism, there is an increase in serum TC, LDL-C, Apo B, Lp(a) levels and possibly TG levels [6,8]. The increase in lipid levels can be reversed by thyroid hormone supplementation [21]. In subjects with subclinical hypothyroidism, significant increases in the levels of TC, LDL-C and TC/HDL-C ratio compared with euthyroid subjects have been observed [22]. Levothyroxine treatment in subjects with subclinical hypothyroidism mainly reduces TC and LDL-C levels, but also TGs, Apo A, Apo B and Lp(a), as shown in several studies [23]. On the other hand, hyperthyroid patients exhibit lower levels of TC, HDL-C and LDL-C [24]. The association between thyroid function and dyslipidaemia is well demonstrated not only in hypo- and hyperthyroidism, but also in low-normal thyroid function with TSH level within the reference range [8,12,14].

However, the relationship between thyroid hormones and lipid profile in euthyroid patients with T2DM is less clarified. A recent study addressed this issue, and found that TSH is positively associated with serum TC and LDL-C levels in euthyroid diabetic women, suggesting that further investigations are needed to clarify the complex mechanisms of lipid metabolism in T2DM with respect to thyroid function [18]. Another study in both a diabetic and a non-diabetic cohort has demonstrated that low-normal thyroid function may augment plasma TG levels and large very low-density lipoprotein (VLDL) particles and also increase VLDL particle size [25]. The same group has also recently shown that the above traits of dyslipidaemia in low-normal thyroid function are mediated via impairment in Apo E regulation that affects the metabolism of TG-rich lipoproteins [26].

In addition to the laboratory findings, there is evidence linking low-normal thyroid function to actual clinical and subclinical atherosclerosis. In a large cohort of healthy euthyroid men, low-normal T4 levels were found associated with the severity of coronary artery calcification [27]. Carotid atherosclerosis as expressed by Intima Media Thickness has repeatedly been found to be negatively correlated with FT4 values within normal range, even after adjustment for lipid parameters as a confounder [28,29]. In another study examining the effect of low-normal thyroid function in patients with T2DM, low-normal FT4 was associated with reduced levels of serum bilirubin which is an endogenous antioxidant, thus impairing endothelial protection [30].

In the current study, we used an intervention with a small subthyrotoxic dose of T4 for two months and found that TC, LDL-C, HDL-C, TG, Apo A1, Apo B and Lp(a) levels were significantly reduced after T4 treatment.

These findings can be explained by the role of thyroid hormones on lipid metabolism. Thyroid hormones induce 3-hydroxy-3-ethylglutaryl coenzyme A (HMG-CoA) reductase, which catalyses the first step in cholesterol biosynthesis [7]. Moreover, T3 upregulates LDL receptors by controlling gene activation. This T3-mediated gene activation is mediated by the direct binding of T3 to specific thyroid hormone responsive elements (TREs) [7,31]. Furthermore, T3 controls sterol regulatory element-binding protein, which in turn regulates LDL receptor gene expression [32]. The upregulation of LDL receptors by thyroid hormones can explain the decrease of TC and LDL-C after treatment with T4.

HDL-C levels were found to be diminished after treatment with T4. The cause of reduced levels of HDL-C is due to increased CETP and hepatic lipase activity stimulated by thyroid hormones. The latter results in an increased transport of cholesteryl esters from HDL-2 to VLDL and intermediate density lipoprotein (IDL) [33]. On the other hand, low-normal thyroid function has been associated to increased plasma CETP activity in patients with T2DM. This seems to be more pronounced in subjects with exaggerated hypertriglyceridaemia, and may add to the increased cardiovascular risk in such individuals [12]. Moreover, in people with dysglycaemia, low-normal FT4 correlates with reduced HDL antioxidant capacity, thus impairing HDL function [14,34].

Thyroid hormones have dual effects on apolipoprotein gene activity [35–37]. They exert a major effect on transcription and also affect translation of apo A-I mRNA [35]. T3 has been shown to increase apo A-I mRNA levels in rat liver but not in the intestine [38,39]. Strobl et al. found that T3 increased both APOA1 gene expression and posttranscriptional stability of RNA [40]. Romney et al. have shown that the regulation of APOA1 gene expression by T3 is mediated in part by a thyroid hormone response element (THRE) located at the 5′ end of APOA1. APOA1 promoter activity is increased in the presence of the THRE site and decreased in its absence [41]. Effects of thyroid hormones on apo AI have been demonstrated in numerous studies [37]. In our study, we found reduced levels of Apo AI and Apo B after T4 treatment for two months.

In the present study, there was a fall in serum TG levels. A single fasting value of TG is not always a reliable indicator of overall lipid metabolism, since it depends on the physical activity of the previous day and the amount of carbohydrates and fats that the patient had consumed the day before [42]. However, this decrease could represent an improvement in insulin sensitivity.

Our study has limitations. It has an open-label design and thus we cannot exclude the possibility of that interfering with the results. However, we specified that all patients should not change their diet or exercise patterns throughout the study period. Their weight remained the same, and antidiabetic or hypolipidaemic treatment did not start until after the end of the study. Also, the number of participants was small.

To our knowledge, this is the first intervention study to address the effect of reducing TSH levels within the normal range using T4 in euthyroid patients with T2DM and micronodular goitre. Since a debate is emerging about the definition of “optimal” TSH level and its relationship to atherosclerotic risk in insulin-resistant subjects, our findings could be of interest.

In conclusion, we showed that administering a small subthyrotoxic dose of T4 to euthyroid patients with T2DM, improved the lipid profile. Whether, this effect is clinically relevant remains to be established.

Author contributions

VL and FS wrote the manuscript and researched data, EV researched data, EH researched data, PM researched data, EM researched data, NP researched data, VL and GD reviewed the manuscript. All authors approved the final version of the text.

Disclosure statement

FS is currently an employee of MSD Greece; has attended conferences sponsored by Novo Nordisk, Lilly, Sanofi, Unipharma. EV and EM have nothing to declare. PM has been an advisory board member, participated in sponsored studies and has received fees for lectures from Novartis, Lilly, MSD, Sanofi, Menarini, Angelini, Elpen. EH has received fees for advisory work and lectures as well as research funding from following pharmaceutical companies that manufacture diabetes medicines: Eli Lilly, MSD, Bristol/AstraZeneca, Novo Nordisk, Menarini, WinMedica, Boehringer, Ingelheim, Sanofi, Elpen. NP has been an advisory board member of Astra-Zeneca, Boehringer Ingelheim, MSD, Novo Nordisk, Pfizer, Takeda and TrigoCare International; has participated in sponsored studies by Astra-Zeneca, Eli-Lilly, GSK, MSD, Novo Nordisk, Novartis and Sanofi-Aventis; has received honoraria as a speaker for Astra-Zeneca, Boehringer Ingelheim, Eli-Lilly, ELPEN, Galenica, MSD, Mylan, Novo Nordisk, Pfizer, Sanofi-Aventis, Takeda and Vianex; and attended conferences sponsored by TrigoCare International, Eli-Lilly, Galenica, Novo Nordisk, Pfizer and Sanofi-Aventis.GD has been an advisory board member of Astra-Zeneca, Boehringer Ingelheim, MSD, Novo Nordisk, Sanofi, Amgen, Eli-Lilly, Novartis, Abbott, Vianex, Elpen, Menarini; has participated in sponsored studies by Eli-Lilly, GSK, Sanofi, Novo Nordisk; has received honoraria as a speaker for Astra-Zeneca, Boehringer Ingelheim, Eli-Lilly, ELPEN, MSD, Novo Nordisk, Sanofi, Vianex, Novartis, Abbott, Menarini and attended conferences sponsored by Eli-Lilly, Novo Nordisk, Sanofi, Novartis, MSD. VL has been an advisory board member of Astra-Zeneca, Boehringer Ingelheim, MSD, Novo Nordisk, Sanofi, Amgen, Eli-Lilly, Novartis; has participated in sponsored studies by Eli-Lilly, GSK, Sanofi, Novo Nordisk; has received honoraria as a speaker for Astra-Zeneca, Boehringer Ingelheim, Eli-Lilly, ELPEN, MSD, Novo Nordisk, Sanofi, Vianex, Novartis, and attended conferences sponsored by Eli-Lilly, Novo Nordisk, Sanofi, Novartis, MSD.

Funding

This work was supported by the Research Grant Authority of Research Grant Authority of Athens University [Trial registration number: NCT02509858].