Relationship Between Thyroid Hormone Levels and the Pathology of Alzheimer's Disease in Euthyroid Subjects

Abstract

Background:

Although several studies have reported an association between thyroid dysfunction and Alzheimer's disease (AD), the effect of mild thyroid dysfunction within the normal range of thyrotropin (TSH) on the development of AD remains unclear. The aim of this study was to evaluate the association between thyroid hormones and the pathology of AD in euthyroid subjects.

Methods:

Sixty-nine subjects with a TSH level between 0.5 and 4.5 μIU/L who underwent ¹⁸F-florbetaben positron emission tomography were included in this prospective cross-sectional study. The levels of serum free thyroxine (fT4) and TSH were quantified using radioimmunoassay. Neuropsychological tests were performed to assess cognitive function. Differences in cerebral amyloid-β (Aβ) burden were compared between high-normal TSH (TSH ≥2.5 μIU/mL) and low-normal TSH (TSH <2.5 μIU/mL) groups. Multiple linear regression analyses, adjusted for age, sex, education level, and Neuropsychiatric Inventory scores, were performed to evaluate relationships between thyroid hormone levels and both cerebral Aβ burden and cognitive function.

Results:

The cerebral Aβ burden in the high-normal TSH group was significantly higher than in the low-normal TSH group (1.53 ± 0.31 vs. 1.35 ± 0.22, p = 0.009). The fT4 levels were negatively correlated with cerebral Aβ burden (β = −0.240, p = 0.035), and TSH levels were positively correlated with cerebral Aβ burden (β = 0.267, p = 0.020). The fT4 level was also positively associated with cognitive function, as inferred from neuropsychological test scores.

Conclusions:

Thyroid hormone concentrations were associated with AD pathology in euthyroid subjects. Our findings suggest that AD is likely to occur even in individuals with high-normal TSH levels.

Introduction

Thyroid hormones are well-known key neuroregulators of functional development and maturation of the central nervous system and play significant roles in normal neurocognitive function in adulthood. Recently, studies have reported an association between thyroid hormone levels and Alzheimer's disease (AD) (1), and have revealed that dysregulation of thyroid hormone expression and function is significantly associated with an increased risk of dementia and AD (2 –5). Preclinical studies and human postmortem analyses have suggested that thyroid hormone dysfunction contributes to increased production of amyloid-β (Aβ) in the brain, and serum free thyroxine (fT4) levels are negatively correlated with cerebral Aβ deposition (6,7).

In regard to the possible association between subclinical hypothyroidism (SCH) and AD, considerable controversies remain in the literature. Although some studies have reported an association between SCH and increased risk of dementia (8,9), others have failed to find any association (10,11), and another study reported conflicting findings, suggesting that SCH was associated with dementia (12). Furthermore, the appropriate upper limit of the normal range of serum thyroid thyrotropin (TSH) levels remains controversial (13). It has been suggested that the upper limit of the normal TSH level should be set to 2.5 mU/L, based on the distribution of TSH levels in individuals without thyroid disease (14,15). However, only a few studies have addressed the relationship between thyroid function and AD (16 –18) in euthyroid states. Moreover, the level of TSH within the normal range at which elderly patients should be prescribed levothyroxine (LT4) to improve cognitive function is unclear. Thus, the relationship between serum TSH levels and the in vivo neuropathological changes associated with AD requires investigation, to determine the level of TSH that may be associated with cognitive dysfunction in euthyroid individuals.

Although the etiology of AD is not fully understood, deposition and accumulation of certain Aβ peptides play a significant role in the pathogenesis of AD (19). Positron emission tomography (PET) is a reliable imaging modality for tracking the progression of AD in vivo. PET imaging with ¹⁸F-florbetaben (¹⁸F-FBB) has been validated by postmortem histology as a reliable means of inferring cerebral Aβ burden in vivo (20), and evaluating the correlation between thyroid hormone levels and cerebral Aβ burden measured by ¹⁸F-FBB PET could help to elucidate the impact of thyroid dysfunction on AD development. Thus, the aim of this study was to evaluate the relationships between thyroid hormone levels and cerebral Aβ burden in euthyroid subjects.

Materials and Methods

Subjects

This was a prospective study of a consecutive series of patients (50–90 years of age) who visited the memory clinic at Keimyung University Dongsan Medical Center for the evaluation of cognitive function between June 2015 and January 2017. Standard clinical and neuropsychological assessments were conducted, and the evaluating clinicians categorized syndromal cognitive staging based on the 2018 National Institute on Aging-Alzheimer's Association Research Framework. Patients were categorized as being cognitively unimpaired (CU), having mild cognitive impairment (MCI), or meeting criteria for a diagnosis of dementia (21). Neuropsychological tests, including the Mini-Mental State Examination (MMSE), the Digit Span Memory Test, Korean-Boston Naming Test (KBNT), and the Rey-Osterrieth Complex Figure Test and Recognition Trial (RCFT) were used to assess cognitive function (22). The Neuropsychiatric Inventory (NPI) was used to evaluate neuropsychiatric symptoms, including delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, disinhibition, irritability, aberrant motor behavior, nighttime behavior, and food intake and appetite changes (23). Total NPI scores were calculated as the sum of the subscores of all symptoms. All participants underwent brain magnetic resonance imaging (MRI) and ¹⁸F-FBB PET within four weeks of visiting the clinic. Patients meeting the following criteria were excluded: an MMSE score <10; having a condition known to affect cognition; having a previous diagnosis of dementia; having thyroid disease or receiving LT4 or antithyroid drug therapy; exhibiting abnormal results on a thyroid function test (TFT) (TSH levels <0.4 or >4.5 μIU/mL); having a history of psychiatric episodes, substance abuse, cerebrovascular disease, brain injury, cardiovascular disease, malignancy, or infection. The Institutional Review Board of our hospital approved this study. Subjects were informed about the study methodology and risks associated with participation in the study, and written informed consent was obtained from all subjects or from their legally authorized representatives.

Thyroid function test

Venous blood sampling was performed around the same time of the day (9:00–10:00 AM) on all patients after overnight fasting to evaluate serum levels of total triiodothyronine (T3), fT4, and TSH. Serum TSH levels were measured by an immunoradiometric assay (TSH-CTK-3; DiaSorin, Saluggia, Italy) with a functional sensitivity of 0.07 mIU/L. Serum fT4 levels were measured with an fT4 radioimmunoassay kit (Immunotech, Prague, Czech Republic) according to the manufacturer's instructions. Serum total T3 levels were measured by radioimmunoassays (T3-CTK; DiaSorin). The normal range for total T3 was 98–180 ng/dL; for fT4, it was 0.70–1.94 ng/dL; and for TSH, it was 0.40–4.50 μIU/mL. Thyroid status was determined based on serum TSH level; subjects with a serum TSH level of ≥2.5 μIU/mL were classified as the high-normal serum TSH (HN_TSH) group, while subjects with a serum TSH level of <2.5 μIU/mL were classified as the low-normal serum TSH (LN_TSH) group (24).

¹¹⁸F-florbetaben positron emission tomography

All imaging was performed within one month following the TFT. Subjects underwent 3D T1-weighted MRI (Signa VHi 3.0T scanner; GE Healthcare, Milwaukee, WI, USA), and a PET/computed tomography (CT) system (Biograph mCT-64; Siemens Healthcare, Knoxville, TN, USA) was used to acquire ¹⁸F-FBB PET images. ¹⁸F-FBB PET images were acquired from 90 to 100 minutes after a single-dose intravenous injection of 300 MBq of ¹⁸F-FBB. Nonenhanced low-dose CT was performed for attenuation correction and localization in the spiral mode at 120 kVp and 150 mAs, using the True X algorithm. The PET images were subjected to iterative reconstruction using ordered subset expectation maximization. Attenuation correction of the PET images was performed using attenuation data from the nonenhanced CT images.

Quantitative analyses of ¹⁸F-FBB PET imaging were conducted for certain volumes of interest (VOI) using the software program PMOD (version 3.9; PMOD Technologies Ltd., Zurich, Switzerland) by an experienced nuclear medicine physician blinded to each subject's clinical status, including thyroid hormone levels and AD pathology, as previously described (25). Image processing was performed using Statistical Parametric Mapping 12 (SPM12; Wellcome Department of Imaging Neuroscience, Institute of Neurology, University College London, United Kingdom) within MATLAB 2013a (MathWorks, Inc., MA, USA) and MRIcro software, version 1.37 (Chris Rorden, Columbia, SC, USA). Each MRI and PET image was coregistered with a standard mutual information algorithm and spatially normalized. An automated anatomical labeling template was subsequently applied for standardized, regional brain VOI sampling of count densities (26).

For quantification of cerebral Aβ burden, VOIs were individually defined in the frontal, temporal, and parietal cortices, cingulate cortex, and cerebellar cortex on amyloid PET images. Standardized ¹⁸F-FBB uptake values were obtained from the defined regional VOIs, and a regional standardized ¹⁸F-FBB uptake value ratio (SUVR) was calculated by dividing the standardized ¹⁸F-FBB uptake value for each individual target region by the uptake value for the cerebellar cortex. A composite SUVR was calculated by dividing the mean of the standardized ¹⁸F-FBB uptake values for the frontal, temporal, and parietal cortices, and the cingulate cortex by the uptake value for the cerebellar cortex as a reference region, as previously described (27).

Statistical analyses

Numerical data (age, education, body mass index [BMI], thyroid hormone levels, NPI score, scores on the MMSE, KBNT, RCFT, and Digit-span tests, and SUVR from PET scans) are expressed as means ± standard deviations, and were compared between the HN_TSH and LN_TSH groups using two-sample t-tests. Fisher's exact tests were performed to evaluate differences in the frequency of type 2 diabetes mellitus (DM), hypertension (HTN), hyperlipidemia, and cognitive staging between the HN_TSH and LN_TSH groups. Pearson correlation analysis was performed to evaluate associations between serum levels of thyroid hormones and SUVRs. Based on the results from correlation analyses, thyroid hormones with p-values <0.05 were selected for further regression analysis. We performed the multivariate linear regressions with the selected thyroid hormones as independent variables and the SUVRs as dependent variables, controlling for age and sex as covariates. Also, we assessed the relationship between the thyroid hormone levels and their scores on the cognitive function tests using Pearson's correlation analyses and multiple linear regression analyses, adjusted for age, sex, education level, and NPI score. All statistical analyses were performed using the Statistical Package for the Social Sciences for Windows version 25.0 (SPSS, Inc., Chicago, IL, USA). The p-values were corrected for multiple comparisons using a false discovery rate correction. A p-value <0.05 was considered to be statistically significant.

Results

Subject demographic and clinical characteristics

A total of 69 subjects were enrolled in this study; their demographic and clinical characteristics are summarized in Table 1. There were no significant differences in age, sex, education level, type 2 DM, HTN, hyperlipidemia, BMI, NPI score, thyroid hormone levels, or neuropsychological test scores between the HN_TSH (n = 19) and LN_TSH (n = 50) groups. In the HN_TSH group, four subjects were categorized as CU, eight exhibited MCI, and seven met the criteria for dementia. In the LN_TSH group, 14 subjects were categorized as CU, 14 exhibited MCI, and 22 met the criteria for dementia. The mean time interval between TFT and ¹⁸F-FBB PET scans was 9.2 ± 7.2 days. The composite SUVRs of subjects with dementia were significantly higher than those of patients with CU (1.48 ± 0.30 vs. 1.27 ± 0.13, p = 0.020). There were no significant differences in composite SUVRs between patients with CU and MCI (1.27 ± 0.13 vs. 1.40 ± 0.24, p = 0.358), or between patients with MCI and dementia (1.40 ± 0.24 vs. 1.48 ± 0.30, p = 0.703).

Table 1.

Clinical Characteristics of Patients

	Total (n = 69)	High-normal TSH group ^a (n = 19)	Low-normal TSH group ^b (n = 50)	p
Age, years (SD)	66.0 (9.4)	66.9 (9.1)	65.6 (9.5)	0.623
Sex, female (%)	42 (60.9)	10 (52.6)	32 (64.0)	0.419
Education^c, years (SD)	11.2 (5.5)	10.0 (5.6)	11.7 (5.5)	0.218
Type 2 diabetes mellitus, n (%)	13 (18.8)	2 (10.5)	11 (22.0)	0.491
Hypertension, n (%)	25 (36.2)	6 (31.6)	19 (38.0)	0.620
Hyperlipidemia, n (%)	14 (20.3)	4 (21.1)	10 (20.0)	1.000
Body mass index (SD)	23.3 (3.5)	23.9 (3.2)	23.1 (3.6)	0.435
Thyroid hormone levels
TSH, μIU/mL (SD)	1.97 (0.93)	3.19 (0.44)	1.51 (0.57)	<0.001
T3, mg/dL (SD)	137.63 (29.76)	140.70 (28.13)	136.5 (30.56)	0.617
fT4, ng/dL (SD)	1.31 (0.23)	1.28 (0.20)	1.31 (0.25)	0.636
NPI score (SD)	7.4 (11.6)	5.3 (6.5)	8.4 (13.1)	0.333
Cognitive function tests
MMSE (SD)	24.3 (5.5)	24.1 (4.2)	24.3 (5.9)	0.369
KBNT (SD)	42.7 (13.2)	43.2 (11.8)	42.4 (13.8)	0.833
RCFT copy (SD)	28.9 (7.1)	29.1 (5.1)	28.9 (7.9)	0.936
RCFT immediate recall (SD)	10.8 (6.9)	10.6 (7.0)	11.2 (6.8)	0.771
RCFT delayed recall (SD)	9.9 (7.0)	10.0 (7.1)	9.9 (6.9)	0.964
Digit-span forward (SD)	5.9 (1.6)	5.8 (1.7)	6.0 (1.5)	0.698
Digit-span backward (SD)	3.7 (1.5)	3.8 (1.7)	3.5 (0.8)	0.378
Cognitive staging
Cognitively unimpaired, n (%)	18 (26.1)	4 (21.1)	14 (28.0)	0.527
Mild cognitive impairment, n (%)	22 (31.9)	8 (42.1)	14 (28.0)
Dementia, n (%)	29 (42.0)	7 (36.8)	22 (44.0)

High-normal TSH group was defined as subjects with a serum TSH level of ≥2.5 μIU/mL.

Low-normal TSH group was defined as subjects with a serum TSH level of <2.5 μIU/mL.

The education level was defined as the period of education from elementary school to university and graduate school.

free T4, free thyroxine; KBNT, Korean Boston Naming Test; NPI, Neuropsychiatric Inventory; MMSE, Mini-Mental State Examination; RCFT, Rey Complex Figure Test; T3, triiodothyronine; TSH, thyrotropin.

Cerebral Aβ burden and thyroid hormone levels

The HN_TSH group exhibited a significantly higher composite SUVR than the LN_TSH group (1.53 ± 0.31 vs. 1.35 ± 0.22, p = 0.009). In addition, the regional SUVRs in the bilateral frontal, temporal, and parietal cortices, and the cingulate cortex in the HN_TSH group were significantly higher than those in the LN_TSH group (Table 2).

Table 2.

Comparison of Cerebral Amyloid-β Burden Between High-Normal and Low-Normal Thyrotropin Groups

Regions	High-normal TSH group ^a (n = 19)	Low-normal TSH group ^b (n = 50)	p
Composite SUVR	1.53 (0.31)	1.35 (0.22)	0.009
Regional SUVR
Frontal cortex
Left	1.45 (0.32)	1.28 (0.22)	0.019
Right	1.42 (0.30)	1.26 (0.22)	0.022
Temporal cortex
Left	1.57 (0.32)	1.38 (0.19)	0.024
Right	1.45 (0.30)	1.30 (0.20)	0.022
Parietal cortex
Left	1.57 (0.39)	1.36 (0.26)	0.020
Right	1.47 (0.34)	1.28 (0.25)	0.021
Cingulate cortex
Left	1.78 (0.39)	1.56 (0.27)	0.024
Right	1.84 (0.32)	1.64 (0.22)	0.016

High-normal TSH group was defined as subjects with a serum TSH level of ≥2.5 μIU/mL.

Low-normal TSH group was defined as subjects with a serum TSH level of <2.5 μIU/mL.

Aβ, amyloid-β; SUVR, standardized ¹⁸F-FBB uptake value ratio.

Pearson's correlation analyses revealed that TSH levels were positively correlated with composite SUVRs (r = 0.294, p = 0.014) and regional SUVRs in the bilateral frontal, temporal, and parietal cortices, and the left cingulate cortex (Table 3). The fT4 levels were negatively correlated with composite SUVRs (r = −0.265, p = 0.028; Fig. 1) and regional SUVRs in the right temporal cortex (p = 0.048). Furthermore, multivariate linear regression analysis, adjusted for age and sex, showed that both TSH and fT4 levels were associated with the composite SUVRs (β = 0.267, p = 0.020 and β = −0.240, p = 0.035, respectively), with a higher level of TSH and a lower level of fT4 associated with a higher cerebral Aβ burden. Linear regression analyses also showed that TSH levels were associated with regional SUVRs in the bilateral frontal, temporal, and parietal cortices, and the left cingulate cortex (p = 0.037, p = 0.041, p = 0.032, p = 0.043, p = 0.043, p = 0.037, and p = 0.040, respectively). The fT4 levels were associated with regional SUVRs in the right temporal cortex (r = −0.289, p = 0.048).

FIG. 1.

Pearson correlation analysis between thyroid hormone level and composite SUVR. There was a significant positive correlation between serum TSH levels and composite SUVR (a), and a significant negative correlation between serum fT4 levels and composite SUVR (c). However, there was no significant correlation between serum T3 levels and composite SUVR (b). fT4, free thyroxine; SUVR, standardized ¹⁸F-FBB uptake value ratio; T3, triiodothyronine; TSH, thyrotropin.

Table 3.

Univariate and Multivariate Linear Regression Analysis Between Global Cerebral Amyloid-β Burden and Thyroid Hormone Levels

Dependent variables	Independent variables	Univariable model		Multivariable model
Dependent variables	Independent variables	r	p	Adjusted R²	Standardized β	p
Composite SUVR	TSH	0.294	0.014	0.162	0.267	0.020
	fT4	−0.265	0.028		−0.240	0.035
	T3	0.097	0.441
Regional SUVR
Left frontal cortex	TSH	0.262	0.030	0.111	0.250	0.037
	fT4	−0.273	0.061
	T3	0.121	0.999
Right frontal cortex	TSH	0.263	0.039	0.130	0.250	0.041
	fT4	−0.306	0.088
	T3	0.078	0.856
Left temporal cortex	TSH	0.340	0.032	0.103	0.340	0.032
	fT4	−0.226	0.099
	T3	0.071	0.764
Right temporal cortex	TSH	0.290	0.043	0.142	0.277	0.043
	fT4	−0.301	0.048		−0.289	0.048
	T3	0.028	0.942
Left parietal cortex	TSH	0.277	0.034	0.118	0.257	0.043
	fT4	−0.231	0.112
	T3	0.079	0.999
Right parietal cortex	TSH	0.263	0.033	0.100	0.246	0.037
	fT4	−0.206	0.090
	T3	0.019	0.882
Left cingulate cortex	TSH	0.291	0.030	0.122	0.272	0.040
	fT4	−0.221	0.091
	T3	0.131	0.999
Right cingulate cortex	TSH	0.293	0.056
	fT4	−0.219	0.081
	T3	0.097	0.999

Values represent standardized linear regression coefficients (β) of the correlation between thyroid hormone levels and cerebral Aβ burden, after adjusting for age and sex.

Cognitive function and thyroid hormone levels

In univariate analyses, there were significant positive correlations between subjects' scores on the MMSE, KBNT, and RCFT (immediate and delayed recall) and their fT4 levels (p = 0.006, p = 0.024, p = 0.021, and p = 0.027, respectively), but no significant correlations were observed between subjects' scores on any of the cognitive function tests and their TSH or T3 levels. Multiple linear regression analyses, adjusted for age, sex, education level, and NPI score, also showed that fT4 levels were positively associated with subjects' scores on the MMSE, KBNT, and RCFT (copy and delayed recall) (p = 0.003, p = 0.013, p = 0.001, and p = 0.040, respectively; Table 4). However, TSH, and T3 levels were not significantly correlated with scores on the cognitive function tests, even after adjusting for age, sex, education level, and NPI score. There were no significant differences in scores on the cognitive function tests between the HN_TSH and LN_TSH groups.

Table 4.

Univariate and Multivariate Linear Regression Analysis Between Cognitive Function and Thyroid Hormone Levels

Dependent variables	Independent variables	Univariable model		Multivariable model
Dependent variables	Independent variables	r	p	Adjusted R ²	Standardized β	p
MMSE	TSH	0.009	0.946
	fT4	0.384	0.006	0.705	0.228	0.003
	T3	0.086	0.776
KBNT	TSH	0.086	0.493
	fT4	0.323	0.024	0.634	0.213	0.013
	T3	0.143	0.395
RCFT (copy)	TSH	0.171	0.294
	fT4	0.290	0.078	0.527	0.358	0.001
	T3	0.168	0.215
RCFT (immediate recall)	TSH	0.147	0.270
	fT4	0.303	0.021
	T3	0.157	0.371
RCFT (delayed recall)	TSH	0.125	0.349
	fT4	0.338	0.027	0.376	0.238	0.040
	T3	0.150	0.405
Digit-span test (forward)	TSH	0.057	0.650
	fT4	0.198	0.333
	T3	0.113	0.566
Digit-span test (backward)	TSH	−0.020	0.876
	fT4	0.205	0.294
	T3	0.100	0.656

Values represent standardized linear regression coefficients (β) of the correlation between thyroid hormone levels and cognitive function, after adjusting for age, sex, education level, and NPI score.

Discussion

The present study discovered significant associations between thyroid hormone levels and the in vivo neuropathological changes of AD in euthyroid subjects. Several epidemiological studies have reported that thyroid dysfunction can be a risk factor for dementia (4,28). There is objective evidence to suggest that brain structure and function are altered in patients with overt hypothyroidism (1,29). However, the impact of SCH on cognitive function in older people remains controversial (9,11,12). Furthermore, there is considerable controversy regarding the appropriate upper limit of normal TSH levels (13). Although the upper limit of the normal TSH reference range is generally suggested to be 4–5 mU/L (24), some studies have reported that the upper limit should be 2.5 or 3 mU/L in healthy individuals without thyroid disease (15). In contrast, a recent study suggested that the upper limit of the TSH range should be raised to 7 mIU/L in individuals older than 80 years, because of the age-related shift toward higher TSH levels in older individuals and the lack of benefits of LT4 treatment for the improvement of hypothyroid symptoms in older individuals with SCH (30). The present study found a positive correlation between TSH levels and the cerebral Aβ burden, as well as a negative correlation between fT4 levels and the cerebral Aβ burden in subjects with TSH levels between 0.4 and 4.5 mIU/L. Moreover, the subjects with HN_TSH exhibited a higher cerebral Aβ burden than those with LN_TSH. These findings raise the question of whether the upper limit of the normal TSH range should be lowered, because mild thyroid dysfunction within the normal range of TSH may affect the pathology of AD.

The deposition of Aβ in the brain is a hallmark pathophysiology of patients with AD, and PET imaging with ¹⁸F-FBB or ¹¹C-Pittsburgh compound B (PIB) has proven validity for the in vivo identification of cerebral Aβ plaque deposition in patients with AD (20,21). In addition to the deposition and accumulation of specific Aβ peptides, the pathological tau proteins play an important role in AD pathology through the formation of intracellular neurofibrillary tangles (31). A few studies have evaluated the association between thyroid hormone levels and AD pathology, including the Aβ peptide and phosphorylated tau protein. One postmortem histological study reported that serum T3 levels are lower in patients with a high burden of neurofibrillary tangles compared with controls without any neurological disease (6). Recently, a PET study using ¹¹C-PIB reported a negative association between serum fT4 levels and cerebral Aβ burdens in 148 CU subjects (18). In accordance, the present study also observed that the fT4 level was negatively correlated with cerebral Aβ burden, although the present study also showed a positive correlation between serum TSH levels and cerebral Aβ burdens, which was not found in that previous PET study (18). This discrepancy could be attributed to differences in the study population, as the present study included not only CU subjects but also those with MCI and dementia, while the former included only CU subjects. Solely within the CU group, the difference in cerebral Aβ burden between subjects may have been too small to observe a statistical association between TSH levels and AD pathology (19).

The mechanism through which thyroid hormones contribute to the pathophysiology of AD is not yet clear, although there may be a relationship between the hypothalamic/pituitary/thyroid axis and AD, as T3 plays a regulatory role in Aβ synthesis (6,32). Serum thyroid hormones, predominately T4, cross the blood/brain barrier (BBB) of the choroid plexus via the monocarboxylate transporter 8 or organic anion transporting polypeptide 1C1 (33). After crossing the BBB, T4 can be converted into T3, mainly via type II iodothyronine deiodinase in the cerebral cortex; converted T3 could repress the cerebral gene expression of Aβ precursor protein (APP), the proteolysis of which generates Aβ (34,35). Furthermore, T3 treatment modulates alternative APP gene splicing and secretion of APP isoforms, leading to alterations in cerebral Aβ deposition (32). In addition, another mechanism related to transthyretin may explain the association between thyroid hormone levels and AD pathology. Transthyretin, a transport protein that carries T4 in the cerebrospinal fluid, is reduced in AD patients compared with age-matched controls (36). Furthermore, transthyretin may play a preventive role in cerebral Aβ deposition in AD patients by inhibiting its aggregation through the formation of soluble Aβ complexes (37). Stabilized transthyretin exerts a positive effect on Aβ clearance, and its stability is decreased in AD, suggesting a protective role of transthyretin in AD (38).

Hypothyroidism is a well-known risk factor for the development of neuropsychiatric symptoms, including cognitive impairment, especially in elderly subjects (12,39). However, the effect of SCH on cognitive function in older people remains controversial (40). One prospective, open-label, interventional study showed a specific deficit in hippocampal memory processes in patients with SCH (41). Furthermore, that study reported that thyroid hormone replacement therapy improved verbal and spatial memories in patients with SCH. A meta-analysis of 13 studies reported a significant risk of cognitive impairment in SCH patients younger than 75 years (8); however, another meta-analysis of seven studies revealed no significant association between SCH and dementia risk. The latter study also reported that the decline in MMSE scores over the years did not significantly differ between SCH and euthyroid groups (42). On the contrary, 1 prospective cohort study with 15,792 participants reported that SCH was associated with a lower risk of dementia, while hyperthyroidism was associated with a higher risk of dementia compared with euthyroid subjects (43). Moreover, a population-based prospective study conducted in Rotterdam also revealed that SCH in the elderly increases the risk of dementia and AD (3). The Framingham longitudinal, community-based, observational study assessed 1864 euthyroid subjects. It found that not only SCH but also subclinical hyperthyroidism increases the risk of dementia and AD after adjusting for covariates, including age and education level (4). Therefore, it seems that mild thyroid dysfunction, including SCH and subclinical hyperthyroidism, can affect the development of AD, with a U-shaped relationship between TSH levels and AD risk, although the results for the correlation between TSH level and AD have been discordant based on the study design and population.

In the present study, although there was a significant relationship between TSH level and cerebral Aβ burden, there was no correlation between TSH levels and subjects' cognitive functions. These conflicting results could be a result of the study design, as the relationship was assessed between thyroid dysfunction and a broad spectrum of clinical dementia diagnoses based on cognitive function tests. The scores on cognitive function tests can be substantially affected by many factors, ranging from several forms of neurodegenerative disease, as well as genetics, level of physical activity, education level, alcohol use, diabetes, and cardiovascular disease (44), whereas small changes in thyroid hormone levels will have a relatively low impact on their cognitive function. In addition, even if cerebral Aβ burden is abnormal in CU subjects, it typically takes 10–20 years for those affected to begin to exhibit cognitive decline, so there may be a discrepancy in the association between TSH level and cerebral Aβ burden or cognitive function (19). To clarify the effects of mild thyroid dysfunction on the development of AD, longitudinal studies that evaluate pathological changes, such as changes in tau protein or Aβ plaque deposition, will be needed, rather than studies based on a broad spectrum of clinical dementia diagnoses. Also, the present study showed that serum fT4 levels were positively correlated with subjects' scores on the MMSE, KBNT, and RCFT, but no correlation was observed between TSH levels and scores on any of the neuropsychological tests; this could indicate that cognitive function may be more affected by fT4 levels, which can directly impact test scores in the short term, while TSH may be more likely to indirectly affect scores on cognitive function tests in the long term (33).

This study has some limitations. First, this was a cross-sectional study, so it is difficult to establish a causal relationship between thyroid hormone levels and global cerebral Aβ burden based on our findings. Thyroid hormones may also contribute to AD development in an indirect manner through other mechanisms involving systemic changes associated with cardiovascular disease, cerebral small vessel disease, or metabolic syndrome (45,46). Second, participants in the present study had impaired cognition of greater severity than those in previous studies that evaluated the association between thyroid dysfunction and AD (18,40). Thus, the difference in cognitive function of the study population should be considered, to apply the results of the present study to subjects with normal cognition or MCI. Third, because serum TSH levels may vary according to the time or day of measurement, there might have been some variability in TSH levels. In an attempt to minimize this variability, however, venous blood sampling was performed in the mornings, at the same time of day in all patients. Despite these efforts, there may still have been temporary fluctuations in subjects' TSH levels. Future prospective study with repeated blood sampling could reduce this type of potential error (47) and more reliably reveal the relationship between thyroid hormone levels and AD pathology.

In conclusion, thyroid hormone levels are associated with the pathology of AD. We observed a positive relationship between serum TSH levels and the cerebral Aβ burden, and a negative relationship between serum fT4 levels and the cerebral Aβ burden in euthyroid subjects. Our findings suggest that AD is likely to occur even in individuals with high-normal TSH levels. Further longitudinal studies with a larger population will be needed to clarify the effect of mild thyroid dysfunction within the normal range of TSH on AD development.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korea Government (MSIP) (Nos. 2018R1C1B5047075, 2014R1A5A2010008, and 2017R1C1B5017721).

References

Cooke

, Mullally

, Correia

, O'Mara

, Gibney

. 2014. Hippocampal volume is decreased in adults with hypothyroidism. Thyroid, 24:433–440.

Ganguli

, Burmeister

, Seaberg

, Belle

, DeKosky

. 1996. Association between dementia and elevated TSH: a community-based study. Biol Psychiatry, 40:714–725.

Kalmijn

, Mehta

, Pols

, Hofman

, Drexhage

, Breteler

. 2000. Subclinical hyperthyroidism and the risk of dementia. The Rotterdam study. Clin Endocrinol, 53:733–737.

Tan

, Beiser

, Vasan

, Au

, Auerbach

, Kiel

, Wolf

, Seshadri

. 2008. Thyroid function and the risk of Alzheimer disease: the Framingham Study. Arch Intern Med, 168:1514–1520.

de Jong

, Masaki

, Chen

, Remaley

, Breteler

, Petrovitch

, White

, Launer

. 2009. Thyroid function, the risk of dementia and neuropathologic changes: the Honolulu-Asia aging study. Neurobiol Aging, 30:600–606.

Davis

, Podolanczuk

, Donahue

, Stopa

, Hennessey

, Luo

, Lim

, Stern

. 2008. Thyroid hormone levels in the prefrontal cortex of post-mortem brains of Alzheimer's disease patients. Curr Aging Sci, 1:175–181.

Contreras-Jurado

, Pascual

. 2012. Thyroid hormone regulation of APP (beta-amyloid precursor protein) gene expression in brain and brain cultured cells. Neurochem Int, 60:484–487.

Pasqualetti

, Pagano

, Rengo

, Ferrara

, Monzani

. 2015. Subclinical hypothyroidism and cognitive impairment: systematic review and meta-analysis. J Clin Endocrinol Metab, 100:4240–4248.

Resta

, Triggiani

, Barile

, Benigno

, Suppressa P

A G

, iagulli

, Guastamacchia

, Sabbà

. 2012. Subclinical hypothyroidism and cognitive dysfunction in the elderly. Endocr Metab Immune Disord Drug Targets, 12:260–267.

10.

de Jong

, den Heijer

, Visser

, de Rijke

, Drexhage

, Hofman

, Breteler

. 2006. Thyroid hormones, dementia, and atrophy of the medial temporal lobe. J Clin Endocrinol Metab, 91:2569–2573.

11.

Formiga

, Ferrer

, Padros

, Contra

, Corbella

, Pujol R; Octabaix Study

Group

. 2014. Thyroid status and functional and cognitive status at baseline and survival after 3 years of follow-up: the OCTABAIX study. Eur J Endocrinol, 170:69–75.

12.

Moon

, Park

, Kim

, Han

, Choi

, Lim

, Park

, Kim

, Jang

. 2014. Lower-but-normal serum TSH level is associated with the development or progression of cognitive impairment in elderly: Korean Longitudinal Study on Health and Aging (KLoSHA). J Clin Endocrinol Metab, 99:424–432.

13.

Cappola

. 2019. The thyrotropin reference range should be changed in older patients. JAMA, 322:1961–1962.

14.

Baloch

, Carayon

, Conte-Devolx

, Demers

, Feldt-Rasmussen

, Henry

, LiVosli

, Niccoli-Sire

, John

, Ruf

, Smyth

, Spencer

, Stockigt JR; Guidelines Committee

NAoCB

. 2003. Laboratory medicine practice guidelines. Laboratory support for the diagnosis and monitoring of thyroid disease. Thyroid, 13:3–126.

15.

Wartofsky

, Dickey

. 2005. The evidence for a narrower thyrotropin reference range is compelling. J Clin Endocrinol Metab, 90:5483–5488.

16.

van Osch

, Hogervorst

, Combrinck

, Smith

. 2004. Low thyroid-stimulating hormone as an independent risk factor for Alzheimer disease. Neurology, 62:1967–1971.

17.

Chiaravalloti

, Ursini

, Fiorentini

, Barbagallo

, Martorana

, Koch

, Tavolozza

, Schillaci

. 2017. Functional correlates of TSH, fT3 and fT4 in Alzheimer disease: a F-18 FDG PET/CT study. Sci Rep, 7:6220.

18.

Choi

, Byun

, Yi

, Sohn

, Lee

, Kim

, Lee

. 2017. Associations of thyroid hormone serum levels with in-vivo Alzheimer's disease pathologies. Alzheimers Res Ther, 9:64.

19.

Villemagne

, Burnham

, Bourgeat

, Brown

, Ellis

, Salvado

, Szoeke

, Macaulay

, Martins

, Maruff

, Ames

, Rowe

, Masters CL; Australian Imaging Biomarkers and Lifestyle (AIBL) Research

Group

. 2013. Amyloid beta deposition, neurodegeneration, and cognitive decline in sporadic Alzheimer's disease: a prospective cohort study. Lancet Neurol, 12:357–367.

20.

Sabri

, Sabbagh

, Seibyl

, Barthel

, Akatsu

, Ouchi

, Senda

, Murayama

, Ishii

, Takao

, Beach

, Rowe

, Leverenz

, Ghetti

, Ironside

, Catafau

, Stephens

, Mueller

, Koglin

, Hoffmann

, Roth

, Reininger

, Schulz-Schaeffer

; Florbetaben Phase 3 Study Group. 2015. Florbetaben PET imaging to detect amyloid beta plaques in Alzheimer's disease: phase 3 study. Alzheimers Dement, 11:964–974.

21.

Jack

Jr. , Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

, Contributors 2018. NIA-AA Research Framework: toward a biological definition of Alzheimer's disease. Alzheimers Dement, 14:535–562.

22.

Folstein

, Folstein

, McHugh

. 1975. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res, 12:189–198.

23.

Cummings

. 1997. The Neuropsychiatric Inventory: assessing psychopathology in dementia patients. Neurology, 48:S10–S16.

24.

Biondi

, Cappola

, Cooper

. 2019. Subclinical hypothyroidism: a review. JAMA, 322:153–160.

25.

Barthel

, Gertz

, Dresel

, Peters

, Bartenstein

, Buerger

, Hiemeyer

, Wittemer-Rump

, Seibyl

, Reininger

, Sabri O; Florbetaben Study

Group

. 2011. Cerebral amyloid-beta PET with florbetaben (18F) in patients with Alzheimer's disease and healthy controls: a multicentre phase 2 diagnostic study. Lancet Neurol, 10:424–435.

26.

Tzourio-Mazoyer

, Landeau

, Papathanassiou

, Crivello

, Etard

, Delcroix

, Mazoyer

, Joliot

. 2002. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage, 15:273–289.

27.

Landau

, Fero

, Baker

, Koeppe

, Mintun

, Chen

, Reiman

, Jagust

. 2015. Measurement of longitudinal beta-amyloid change with 18F-florbetapir PET and standardized uptake value ratios. J Nucl Med, 56:567–574.

28.

Parsaik

, Singh

, Roberts

, Pankratz

, Edwards

, Geda

, Gharib

, Boeve

, Knopman

, Petersen

. 2014. Hypothyroidism and risk of mild cognitive impairment in elderly persons: a population-based study. JAMA Neurol, 71:201–207.

29.

Constant

, de Volder

, Ivanoiu

, Bol

, Labar

, Seghers

, Cosnard

, Melin

, Daumerie

. 2001. Cerebral blood flow and glucose metabolism in hypothyroidism: a positron emission tomography study. J Clin Endocrinol Metab, 86:3864–3870.

30.

Mooijaart

, Du Puy

, Stott

, Kearney

, Rodondi

, Westendorp

RGJ

, den Elzen

WPJ

, Postmus

, Poortvliet

RKE

, van Heemst

, van Munster

, Peeters

, Ford

, Kean

, Messow

, Blum

, Collet

, Watt

, Dekkers

, Jukema

, Smit

JWA

, Langhorne

, Gussekloo

. 2019. Association between levothyroxine treatment and thyroid-related symptoms among adults aged 80 years and older with subclinical hypothyroidism. JAMA, 322:1–11.

31.

Iqbal

, Alonso Adel

, Chen

, Chohan

, El-Akkad

, Gong

, Khatoon

, Li

, Liu

, Rahman

, Tanimukai

, Grundke-Iqbal

. 2005. Tau pathology in Alzheimer disease and other tauopathies. Biochim Biophys Acta, 1739:198–210.

32.

Latasa

, Belandia

, Pascual

. 1998. Thyroid hormones regulate beta-amyloid gene splicing and protein secretion in neuroblastoma cells. Endocrinology, 139:2692–2698.

33.

Schroeder

, Privalsky

. 2014. Thyroid hormones, t3 and t4, in the brain. Front Endocrinol, 5:40.

34.

Bárez-López

, Obregon

, Bernal

, Guadaño-Ferraz

. 2018. Thyroid hormone economy in the perinatal mouse brain: implications for cerebral cortex development. Cereb Cortex, 28:1783–1793.

35.

Belakavadi

, Dell

, Grover

, Fondell

. 2011. Thyroid hormone suppression of beta-amyloid precursor protein gene expression in the brain involves multiple epigenetic regulatory events. Mol Cell Endocrinol, 339:72–80.

36.

Serot

, Christmann

, Dubost

, Couturier

. 1997. Cerebrospinal fluid transthyretin: aging and late onset Alzheimer's disease. J Neurol Neurosurg Psychiatry, 63:506–508.

37.

, Song

, Sanders

, Buxbaum

. 2016. Transthyretin suppresses amyloid-beta secretion by interfering with processing of the amyloid-beta protein precursor. J Alzheimers Dis, 52:1263–1275.

38.

Alemi

, Silva

, Santana

, Cardoso

. 2017. Transthyretin stability is critical in assisting beta amyloid clearance—relevance of transthyretin stabilization in Alzheimer's disease. CNS Neurosci Ther, 23:605–619.

39.

Samuels

. 2014. Psychiatric and cognitive manifestations of hypothyroidism. Curr Opin Endocrinol Diabetes Obes, 21:377–383.

40.

Yamamoto

, Ishizawa

, Ishikawa

, Yamanaka

, Murakami

, Hiraiwa

, Okumiya

, Ishine

, Matsubayashi

, Otsuka

. 2012. Cognitive function with subclinical hypothyroidism in elderly people without dementia: one year follow up. Geriatr Gerontol Int, 12:164–165.

41.

Correia

, Mullally

, Cooke

, Tun

, Phelan

, Feeney

, Fitzgibbon

, Boran

, O'Mara

, Gibney

. 2009. Evidence for a specific defect in hippocampal memory in overt and subclinical hypothyroidism. J Clin Endocrinol Metab, 94:3789–3797.

42.

Rieben

, Segna

, da Costa

, Collet

, Chaker

, Aubert

, Baumgartner

, Almeida

, Hogervorst

, Trompet

, Masaki

, Mooijaart

, Gussekloo

, Peeters

, Bauer

, Aujesky

, Rodondi

. 2016. Subclinical thyroid dysfunction and the risk of cognitive decline: a meta-analysis of prospective cohort studies. J Clin Endocrinol Metab, 101:4945–4954.

43.

George

, Lutsey

, Selvin

, Palta

, Windham

, Folsom

. 2019. Association between thyroid dysfunction and incident dementia in the atherosclerosis risk in communities neurocognitive study. J Endocrinol Metab, 9:82–89.

44.

, Prina

, Brayne

. 2015. The association between community environment and cognitive function: a systematic review. Soc Psychiatry Psychiatr Epidemiol, 50:351–362.

45.

Yang

, Lv

, Yue

, Wei

, Liu

, Zhang

. 2016. Subclinical hypothyroidism and the risk of metabolic syndrome: a meta-analysis of observational studies. Endocr Res, 41:158–165.

46.

Zhang

, Xie

, Ding

, Xiao

, Tang

, Jiang

, Shan

, Lin

, Zhu

, Li

, Hu

, Ling

, Xu

, Sheng

. 2017. Subclinical hypothyroidism and risk of cerebral small vessel disease: a hospital-based observational study. Clin Endocrinol, 87:581–586.

47.

Polak

, Le Gac

, Vuillard

, Guibourdenche

, Leger

, Toubert

, Madec

, Oury

, Czernichow

, Luton

. 2004. Fetal and neonatal thyroid function in relation to maternal Graves' disease. Best Pract Res Clin Endocrinol Metab, 18:289–302.

Relationship Between Thyroid Hormone Levels and the Pathology of Alzheimer's Disease in Euthyroid Subjects

Abstract

Background:

Methods:

Results:

Conclusions:

Introduction

Materials and Methods

Subjects

Thyroid function test

118F-florbetaben positron emission tomography

Statistical analyses

Results

Subject demographic and clinical characteristics

Cerebral Aβ burden and thyroid hormone levels

Cognitive function and thyroid hormone levels

Discussion

Footnotes

Author Disclosure Statement

Funding Information

References

¹¹⁸F-florbetaben positron emission tomography