Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Sep 11.
Published in final edited form as: Arch Gerontol Geriatr Plus. 2025 Aug 6;2(3):100195. doi: 10.1016/j.aggp.2025.100195

Thyroid-stimulating hormone and cognitive impairment in non-depressed non-demented multiethnic middle-aged and older US adults: Assessing sex-specific risk prediction

Asma Hallab, for the Alzheimer’s Disease Neuroimaging Initiativea,b,c,d
PMCID: PMC12422142  NIHMSID: NIHMS2105998  PMID: 40937123

Abstract

Background:

Understanding the particularities of thyroid-cognition interactions in older adults is crucial in assessing the risks and evaluating therapeutic options.

Methods:

Cross-sectional analyses where participants from Alzheimer’s Disease Neuroimaging Initiative (ADNI) with mild cognitive impairment (MCI) and healthy controls (HC), with complete neurocognitive tests, thyroid stimulating hormone (TSH) <10 μIU/mL, and geriatric depression scale (GDS) <5, were eligible. Linear and logistic regression models, including testing for non-linearity, were performed. Sex strata were explored.

Results:

Of the total 1845 multiethnic US-participants, with a median age of 73 (IQR: 68, 78); 887 (48 %) were females, and 1056 (57 %) had MCI. The median TSH level was 1.70 μIU/mL (IQR: 1.15, 2.40); significantly lower in MCI than HC (1.66 vs. 1.74 μIU/mL, p-value=0.02). There was a significant association between TSH and overall cognition only in males (adj. ßMales=−0.40[−0.74, −0.07], p-value=0.019). The odds of being diagnosed with MCI at baseline decreased with higher TSH levels in the total study population (adj. ORTotal=0.87[0.79 0.95], p-value=0.002) and in males (adj. ORMales=0.80[0.70, 0.92], p-value=0.001).

Conclusions:

There was a sex-specific, statistically significant association between TSH levels and cognition in multiethnic middle-aged and older ADNI adults. Lower TSH levels and worse global cognition were statistically associated only in males. To precisely delineate the chronological onset of these disorders, longitudinal clinical studies are needed.

Keywords: Thyroid, Aging, Cognition, Sex difference, Hormone, Neuroendocrinology

1. Introduction

Cognitive impairment is a serious public health concern, affecting millions worldwide and severely impacting their quality of life and independence. It is estimated that 152.8 million persons worldwide will experience some form of dementia by 2050. (Nichols et al., 2022) Understanding related risk factors is crucial for preventive and interventional strategies. Various physiological systems are involved in cerebral functioning and cognition, and the thyroid plays a pivotal role in this body-brain association. (Liu & Brent, 2021) Thyroid-brain interactions begin very early during the neurodevelopmental stages and persist over the lifetime. (Lain et al., 2016; Korevaar et al., 2016) During adulthood and the aging process, thyroid hormones maintain a central role in modulating the risk of mental health, being significantly associated, in a sex-dependent manner, (Baksi & Pradhan, 2021) with the onset and complications of affective and psychotic disorders. (Freuer & Meisinger, 2023; Chen et al., 2022; Barbero et al., 2015; Labad et al., 2016; Saglam et al., 2024; Lai et al., 2021) Thus, the impact of thyroid hormones on cognitive functions, particularly the association between dysthyroidism, whether hypothyroidism (low functioning) or hyperthyroidism (high functioning), on one side, and cognitive decline during the aging process, on the other side, is an emerging research field with multiple edges and of high complexity. Studies in middle-aged and older adults reported controversial outcomes, depending on the underlying demographic particularities and the assessed physiological risk factors. (Pasqualetti et al., 2015; Ge, Xu, Tan & Tan, 2020; Gan & Pearce, 2012; Akintola et al., 2015; van Vliet et al., 2021) Owing to a physiologically increased risk of subclinical hypothyroidism in older adults, (Biondi, Cappola & Cooper, 2019; Biondi & Cappola, 2022) concomitantly with an increased risk of neurodegeneration and cognitive decline, understanding the interaction between both conditions is crucial to minimize the risks. Furthermore, women are more prone to thyroid disorders (Vanderpump, Luster, Duntas & Wartofsky, 2019) and exhibit different patterns of cognitive decline than men. (Jockwitz, Wiersch, Stumme & Caspers, 2021; Sohn et al., 2018) It is therefore unclear how sex-related biological factors might interfere with the thyroid-cognition association.

The general aim of this study was to explore the associations between thyroid-stimulating hormone (TSH) and cognition in non-depressed, non-demented middle-aged and older US-adults and identify eventual sex-related effects. The first aim was to explore whether there is a significant association between TSH and overall cognition in middle-aged and older adults, and whether this association is sex-dependent. The second aim was to evaluate whether TSH and the odds of being diagnosed with mild cognitive impairment (MCI) are concomitantly associated.

2. Methods

2.1. Study population

Alzheimer’s Disease Neuroimaging Initiative (ADNI) study participants with complete baseline data on demographic characteristics, TSH, depression assessment, and overall cognition were included. ADNI, an NIH-funded project (ADNI; National Institutes of Health Grant U19 AG024904 - more in acknowledgments and funding), is a multicenter observational cohort aiming to study the cognitive decline in older adults and better understand Alzheimer’s disease risks. Dr. Michael Weiner is ADNI’s principal investigator. The first cohort (ADNI 1) started in 2004, and several medical centers around the United States and Canada were involved in recruiting study participants. Following ADNI-1, ADNI-Go, −2, −3, and −4 (ongoing) included more study participants and applied more sophisticated neuropsychiatric and neuroimaging methods. Participants completed written consents for taking part in ADNI. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki De claration of 1975, as revised in 2013. All procedures involving human subjects/patients were approved by local IRBs corresponding to each of the over 60 ADNI study centers (a map summarizing the > 60 different ADNI sites can be found at https://adni.loni.usc.edu/about/governance/#core-details). Study protocols and information on ethical approvals, consent, detailed methods, funding, and data can be found at https://adni.loni.usc.edu. The current study is based on a secondary analysis of de-identified data.

The study is reported according to STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines. Details can be found at https://www.strobe-statement.org.

2.2. Thyroid-stimulating hormone

Fasting blood samples were collected at baseline. The time between blood collection to freezing did not exceed 120 min. Central thyroid function was explored, and TSH levels, reported in μIU/mL (= mIU/L), were used for this analysis. In the case of repeated measurements, only the first value was retained since the possibility that the study participant received a hormonal supplementation or intervention between the two assessments cannot be precisely verified or excluded. Very low TSH levels (reported as “<0.01″) were converted to 0.01 for the analysis (in three cases). Only participants with TSH < 10 μIU/mL were included, excluding cases with eventual overt hypothyroidism.

Central thyroid function was explored in the main analysis based on a continuous variable of TSH (μIU/mL). TSH values of 0.4 and 4.5 μIU/mL are clinically relevant cutoff values and are largely considered in literature and guidelines. (Garber et al., 2012) Clinical cutoff values of “euthyroidism”, “hypothyroidism”, and “hyperthyroidism” depend on various factors and might vary across populations, laboratories, sex groups, during pregnancy, and across age spans. (Xing et al., 2021) For the sensitivity analysis, central thyroid function was classified into three TSH ranges limited by 0.4 and 4.5 μIU/mL, which were considered relevant cutoff values for this study (“TSH < 0.4″, “0.4 ≤ TSH ≤ 4.5″, and “TSH > 4.5”), without further dysthyroidism-related labeling.

2.3. Neuropsychological testing

Alzheimer’s Disease Assessment Scale – Cognitive 13-item (ADAS13) score was explored in all ADNI phases and was included as a valid tool to assess the overall cognitive status of the ADNI population. (Rosen, Mohs & Davis, 1984) Moreover, the Clinical Dementia Rating Scale – Sum of Boxes (CDR-SB), (Hughes et al., 1982) the Functional Activity Questionnaire total score (FAQ), (Pfeffer et al., 1982) and the Geriatric Depression Score (GDS) – 15 items total score (Sheikh & Yesavage, 1986) were added to the descriptive and comparative analyses.

2.4. Inclusion criteria

In this study, “middle-aged and older adults” refers to individuals aged 50 years and above. HC and MCI ADNI participants with complete TSH assessment < 10 μIU/mL (to exclude cases with overt hypothyroidism, a risk factor of cognitive impairment), complete biometric/demographic data (age, sex, educational background, height, and weight), complete GDS baseline assessment with GDS total score < 5 (to exclude cases with depression symptoms, a further risk factor of cognitive impairment), (Pocklington, Gilbody, Manea & McMillan, 2016) and complete ADAS13 scores at baseline were eligible. A detailed flow chart is presented in Fig. 1A.

Fig. 1.

Fig. 1.

Open in a new tab

Included cases, TSH levels (μIU/mL) in cognition-related groups, and its association with age.

A. Flow chart of included cases.

B. Linear regression between TSH levels (μIU/mL) and age (years).

C. Comparison of TSH levels (μIU/mL) between healthy controls and participants with mild cognitive impairment.

Footnote: ADAS13: Alzheimer’s Dementia Assessment Scale – Cognition subscale – 13 items, GDS: Geriatric Depression Scale, HC: Healthy Controls, MCI: Mild Cognitive Impairment, TSH: Thyroid Stimulating Hormone.

2.5. Statistical analysis

2.5.1. Descriptive analysis

The study was based on a cross-sectional analysis of ADNI data at baseline. Median (IQR) and number (%) were reported according to the variable type (continuous and count, respectively). Wilcoxon rank sum test, Pearson’s Chi-squared test, Kruskal-Wallis rank sum test, and Fisher’s exact test were performed depending on the assessed variable and the number of compared groups (cognition-related groups and TSH ranges-related groups). Test statistic results and corresponding p-value were reported for each analysis. For further analyses, data were stratified by sex (male and female).

2.5.2. ADAS13 total score

Main analyses

Linear regression was performed to study the association between TSH levels (μIU/mL) and ADAS13 total scores (points) in the total study population, first with a univariable model, then with a multivariable one; adjusted for age (years), sex (male/female), educational background (years of education), ethnic profile (“White”, “Black”, “Other”), Apolipoprotein (APOE) ε4 status (no alleles, one allele, and two alleles), cognition-related main diagnosis (healthy controls (HC) or MCI), and body-mass index (BMI=Weight(kg)/Height(m) (Liu & Brent, 2021)). ADAS13 score was included as a dependent variable, and continuous TSH levels (μIU/mL) as an independent variable.

After stratification, univariable linear regression was performed in each stratum. Then, the models were adjusted accordingly for confounding factors (after removing sex as a confounder).

Results of linear regressions were reported as regression coefficient (ß), 95 % confidence interval (CI), and p-value.

Sensitivity analyses

A sensitivity analysis was performed based on TSH ranges as a categorical variable. TSH ranges were included in the linear regression models as independent variables, first in the main study population, and then in the different strata. Both crude and adjusted models, as previously described, were evaluated. The middle TSH range was set as a reference, and the regression coefficient ß, 95 % CI, and p-value in both groups of “TSH < 0.4″ and “TSH > 4.5” were reported in relation to the reference group.

2.5.3. Mild cognitive impairment

Main analyses

In a second step, a logistic regression analysis was performed to assess the association between TSH and the odds of being diagnosed with MCI at baseline. In the main analysis, MCI (binary: yes/no) as a dependent variable, and continuous TSH levels (μIU/mL) as an independent variable were included in the model. First, a crude model was assessed. Then, the same confounders (age, sex, educational background, ethnic profile, APOE ε4 status, and BMI) were added to the adjusted model. The analysis was repeated in sex strata, after adjusting accordingly for the relevant confounders (age, educational background, ethnic profile, APOE ε4 status, and BMI). Odds ratios (OR), 95 % CI, and p-values were reported.

Sensitivity analyses

The sensitivity analysis followed the previously described steps, and TSH ranges were included as independent variables in the models.

2.5.4. Further methodologies

The variance inflation factor (VIF) was calculated for all models, and the risk of multicollinearity was excluded (all values around 1). Restricted cubic splines were introduced to the models to assess non-linearity in the association. The significance level for the null hypothesis was fixed at 0.05 (two-tailed). Statistical analyses and data visualization were performed with RStudio (version 2024.12.1, Posit®, USA).

3. Results

3.1. Characteristics of study participants

A total of 1845 participants with a median age of 73 years (IQR: 68, 78) met the inclusion criteria (age range: 50–90 years); 958 (52 %) of whom were males, 887 were females (48 %), 789 (42.76 %) were HC, while 1056 (57.24 %) had MCI. The median TSH level in the main population at baseline was 1.70 μIU/mL (IQR: 1.15, 2.40), ranged from < 0.01 (three cases that were transformed to 0.01 for the analysis) to 9.58 μIU/mL, and 58 (3.14 %) participants had TSH levels < 0.4, 48 (2.60 %) TSH > 4.5, and 1739 (94.25 %) had TSH levels between 0.4 and 4.5 μIU/mL. The population characteristics and differences between groups are summarized in Table 1. TSH levels were significantly associated with age in the total study population (ß=0.01; 95 % CI: 0.00, 0.02; p-value=0.006) (Fig. 1B).

Table 1.

Characteristics of the total study population and group comparison.

Characteristics Total study population
 Main diagnosis
 Thyroid function
N Overall N =18451 Healthy controls N = 7891 MCI N = 1,0561 Chi 2 /W 2 p-value (Liu & Brent, 2021) TSH < 0.4 N = 581 0.4 ≤ TSH ≤ 4.5 N = 1,7391 TSH > 4.5 N = 481 Chi 2 /K 3 p-value (Lain et al., 2016)
Age (years) 1845 73 (68, 78) 72 (68, 77) 73 (68, 78) W = 396,806 0.081 71 (67, 78) 73 (68, 78) 76 (70, 80) K = 5.2491 0.072
Sex 1845 Chi2=35.411 <0.001 Chi2=12.522 0.002
Female 887 (48 %) 443 (56 %) 444 (42 %) 41 (71 %) 825 (47 %) 21 (44 %)
Male 958 (52 %) 346 (44 %) 612 (58 %) 17 (29 %) 914 (53 %) 27 (56 %)
Marital status 1845 Chi2=10.975 <0.001 Chi2=1.7124 0.4
Currently married 1362 (74 %) 551 (70 %) 811 (77 %) 46 (79 %) 1278 (73 %) 38 (79 %)
Currently not married or unknown 483 (26 %) 238 (30 %) 245 (23 %) 12 (21 %) 461 (27 %) 10 (21 %)
Ethnic profile 1845 Chi2=26.697 <0.001 Chi2=12.203 0.016
White 1613 (87 %) 655 (83 %) 958 (91 %) 57 (98 %) 1512 (87 %) 44 (92 %)
Black 148 (8.0 %) 91 (12 %) 57 (5.4 %) 1 (1.7 %) 147 (8.5 %) 0 (0 %)
Other 84 (4.6 %) 43 (5.4 %) 41 (3.9 %) 0 (0 %) 80 (4.6 %) 4 (8.3 %)
Educational level (years) 1845 16.00 (14.00, 18.00) 16.00 (15.00, 18.00) 16.00 (14.00, 18.00) W = 457,440 <0.001 16.00 (13.25, 18.00) 16.00 (14.00, 18.00) 16.50 (15.75, 18.00) K = 5.8013 0.055
APOE ε4 status 1672 Chi2=74.578 <0.001 Chi2=0.32148 >0.9
0 allele 977 (58 %) 487 (70 %) 490 (50 %) 33 (59 %) 916 (58 %) 28 (62 %)
1 allele 565 (34 %) 190 (27 %) 375 (39 %) 19 (34 %) 532 (34 %) 14 (31 %)
2 alleles 130 (7.8 %) 22 (3.1 %) 108 (11 %) 4 (7.1 %) 123 (7.8 %) 3 (6.7 %)
Missing values 173 90 83 2 168 3
ADAS13 total score (points) 1845 12 (8, 18) 8 (5, 12) 16 (11, 21) W = 148,408 <0.001 12 (7, 16) 12 (8, 18) 11 (7, 17) K = 0.63804 0.7
CDR-SB (points) 1845 0.50 (0.00, 1.50) 0.00 (0.00, 0.00) 1.50 (1.00, 2.00) W = 10,762 <0.001 1.00 (0.00, 1.88) 0.50 (0.00, 1.50) 0.00 (0.00, 1.00) K = 6.9835 0.030
FAQ total score (points) 1831 0.0 (0.0, 2.0) 0.0 (0.0, 0.0) 1.0 (0.0, 5.0) W = 166,812 <0.001 1.0 (0.0, 2.8) 0.0 (0.0, 2.0) 0.0 (0.0, 0.0) K = 10.469 0.005
Missing values 14 5 9 0 14 0
BMI 1845 26.5 (24.1, 29.5) 26.8 (24.1, 30.0) 26.3 (24.1, 29.2) W = 440,422 0.035 25.5 (23.2, 29.3) 26.5 (24.1, 29.5) 26.4 (24.7, 30.8) K = 2.3593 0.3
TSH level (μIU/mL) 1845 1.70 (1.15, 2.40) 1.74 (1.19, 2.49) 1.66 (1.09, 2.36) W = 443,560 0.017 0.15 (0.06, 0.26) 1.71 (1.18, 2.36) 5.24 (4.80, 6.02) K = 299.83 <0.001
Cognition-related diagnosis Chi2=5.8141 0.055
Healthy controls 18 (31 %) 745 (43 %) 26 (54 %)
MCI 40 (69 %) 994 (57 %) 22 (46 %)
Open in a new tab
1

Median (IQR); n ( %)

2

Wilcoxon rank sum test; Pearson’s Chi-squared test

3

Kruskal-Wallis rank sum test; Pearson’s Chi-squared test; Fisher’s exact test

ADAS13: Alzheimer’s Disease Assessment Scale-13 Items, APOE: Apolipoprotein E, BMI: Body-Mass Index, CDR-SB: Clinical Dementia Rating-Sum of Boxes, FAQ: Functional Activities Questionnaire, MCI: Mild Cognitive Impairment, TSH: Thyroid Stimulating Hormone.

The median TSH value was significantly lower in the MCI group than in HC (1.66 vs. 1.74 μIU/mL, p-value=0.017) (statistical analysis detailed in Fig. 1C). No statistically significant differences were observed in the number of MCI cases across the TSH groups (Table 1).

3.2. Thyroid-stimulating hormone and overall cognition

Models of the linear regression are detailed in Table 2. No significant association between TSH levels (μIU/mL) and ADAS13 total score (points) was found in the main population. After sex stratification, significant associations were reported only in males in the crude model (crude ßMales=−0.65; 95 % CI: −1.0, −0.26; p-value=0.001). The significant association remained after adjustment for confounding factors (adj. ßMales=−0.40; 95 % CI: −0.74, −0.07; p-value=0.001).

Table 2.

Crude and adjusted models of linear regression: Overall cognition (ADAS13 total score) as an outcome in the total study population, female, and male strata - Continuous TSH levels (μIU/mL) as an independent variable.

Characteristics Total study population Crude model N = 1845 Adjusted model N = 1672
 Females Crude model N = 887 Adjusted model N = 789
 Males Crude model N = 958 Adjusted model N = 883
ß (95 % CI) p-value ß (95 % CI) p-value ß (95 % CI) p-value
Crude model
TSH levels (μIU/mL) −0.22 (−0.50, 0.07) 0.14 0.15 (−0.25, 0.56) 0.5 −0.65 (−1.0, −0.26) 0.001
Adjusted model
TSH levels (μIU/mL) −0.07 (−0.30, 0.17) 0.6 0.24 (−0.09, 0.57) 0.2 −0.40 (−0.74, −0.07) 0.019
Age (years) 0.19 (0.15, 0.23) <0.001 0.21 (0.15, 0.27) <0.001 0.17 (0.12, 0.23) <0.001
Diagnosis <0.001 <0.001 <0.001
Healthy controls — — —
MCI 6.6 (6.1, 7.2) 7.4 (6.6, 8.2) 5.7 (4.9, 6.5)
Educational level (years) −0.33 (−0.43, −0.23) <0.001 −0.28 (−0.43, −0.13) <0.001 −0.38 (−0.52, −0.25) <0.001
APOE ε4 status 1.5 (1.1, 2.0) <0.001 2.0 (1.4, 2.6) <0.001 1.2 (0.64, 1.8) <0.001
Ethnic profile > 0.9 0.3
White — — —
Black 0.01 (−1.2, 1.2) 0.55 (−0.98, 2.1) −0.71 (−2.7, 1.3)
Other −0.08 (−1.5, 1.3) −1.5 (−3.5, 0.56) 1.2 (−0.78, 3.2)
BMI −0.01 (−0.07, 0.05) 0.7 −0.02 (−0.09, 0.05) 0.7 0.01 (−0.08, 0.10) 0.9
Sex <0.001
Female —
Male 1.6 (1.0, 2.1)
Open in a new tab

DAS13: Alzheimer’s Disease Assessment Scale – 13 items, APOE: Apolipoprotein E, ß: regression coefficient, BMI: Body Mass Index, CI: Confidence Interval, MCI: Mild Cognitive Impairment, OR: Odds Ratio, TSH: Thyroid Stimulating Hormone.

3.3. Thyroid-stimulating hormone and mild cognitive impairment

The increase in TSH levels of one μIU/mL reduced 10 % the odds of being diagnosed with MCI at baseline (crude ORTotal=0.90; 95 % CI: 0.83, 0.98; p-value=0.013) in the total study population. The association remained statistically significant after adjusting for confounders (adj. ORTotal=0.87; 95 % CI: 0.79, 0.95; p-value=0.002). After stratification, only in the male stratum, the increase of TSH levels was significantly associated with lower odds of MCI at baseline (crude ORMales=0.82; 95 % CI: 0.73, 0.93; p-value=0.001), even after adjustment for confounders (adj. ORMales=0.80; 95 % CI: 0.70, 0.92; p-value=0.001) (Table 3).

Table 3.

Crude and adjusted models of logistic regression: Odds of MCI in the total study population, female, and male strata – Continuous TSH levels (μIU/mL) as an independent variable.

Characteristics Total study population
 Females
 Males
N Event OR (95 % CI) p-value N Event OR (95 % CI) p-value N Event OR (95 % CI) p-value
Crude model
TSH level (μIU/mL) 1845 1056 0.90 (0.83, 0.98) 0.013 887 444 0.97 (0.86, 1.08) 0.555 958 612 0.82 (0.73, 0.93) 0.001
Adjusted model
TSH level (μIU/mL) 1672 973 0.87 (0.79, 0.95) 0.002 789 401 0.94 (0.82, 1.06) 0.292 883 572 0.80 (0.70, 0.92) 0.001
Age (years) 1672 973 1.00 (0.98, 1.01) 0.536 789 401 1.00 (0.98, 1.02) 0.707 883 572 0.99 (0.97, 1.02) 0.561
BMI 1672 973 1.0 (0.97, 1.02) 0.629 789 401 1.00 (0.97, 1.03) 0.990 883 572 0.99 (0.95, 1.02) 0.469
APOE ε4 status (number) 1672 973 2.07 (1.74, 2.48) <0.001 789 401 1.98 (1.56, 2.53) <0.001 883 572 2.19 (1.71, 2.85) <0.001
Educational level (years) 1672 973 0.90 (0.87, 0.94) <0.001 789 401 0.91 (0.86, 0.97) 0.001 883 572 0.89 (0.85, 0.95) <0.001
Ethnic profile 1672 973 0.024 789 401 0.073 883 572 0.319
White — — —
Black 0.53 (0.33, 0.84) 0.51 (0.28, 0.91) 0.57 (0.27, 1.27)
Other 1.08 (0.62, 1.88) 0.90 (0.41, 1.97) 1.25 (0.58, 2.90)
Sex 1672 973 <0.001
Female —
Male 2.00 (1.62, 2.47)
Open in a new tab

APOE: Apolipoprotein E, BMI: Body Mass Index, CI: Confidence Interval, OR: Odds Ratio, TSH: Thyroid Stimulating Hormone.

The introduction of a non-linear factor to the model using the restricted cubic splines showed non-linear patterns. The Q3 TSH value (2.40 μIU/mL) was the highest relevant cutoff value for the association.

Sensitivity analyses

In linear regression models based on TSH ranges as independent variables, TSH did not show any significant association with the ADAS13 total score (Supplementary Table 1).

In the logistic regression and particularly in the male stratum, sensitivity analysis based on TSH ranges as independent variables showed similar results with TSH 〈 0.4 being associated with higher odds of MCI, while TSH 〉 4.5 with lower odds, compared to the intermediate range group (Supplementary Table 2).

Relevant statistical models and their summarized results are visualized in Fig. 2.

Fig. 2.

Fig. 2.

Open in a new tab

Summary plots of relevant statistical models

A. Linear regression analysis of ADAS13 total score as an outcome variable and TSH levels as an independent variable in the total study population and sex strata

B. Splines of predicted odds ratios of MCI in the total study population and sex strata

C. Summary plots of predicted odds ratios of MCI in the total study population and sex strata

Footnote: ADAS13: Alzheimer’s Dementia Assessment Scale – Cognition – 13 items, MCI: Mild Cognitive Impairment, TSH: Thyroid Stimulating Hormone.

4. Discussion

The exclusion of dementia, overt hypothyroidism, and depression as relevant confounding risk factors for cognitive impairment, the high number of study participants, and the sex stratification were major strengths of this study. The first outcome was the significant association between TSH levels and ADAS13 scores. Although no statistically significant results have been found in the main study population, significant associations between low TSH levels and worse overall cognition were revealed in male participants after stratifying by sex. The second outcome was the statistical association between concomitant lower TSH levels and MCI diagnosis at baseline in middle-aged and older males. The association between TSH and MCI diagnosis was particularly significant in those with TSH levels lower than the Q3 TSH value.

4.1. Thyroid function and cognition

The association between TSH levels and cognition in older adults was controversial in the literature. While some studies on older euthyroid adults did not identify significant associations between TSH, free T4 (FT4, thyroxine) levels, and cognition, (Booth, Deary & Starr, 2013) other studies reported significant associations between lower thyroid hormones, mainly free T3 (FT3, tri-iodothyronine), and cognitive impairment. (Chen et al., 2023)

In studies where both euthyroid patients as well as patients with subclinical hypo- and hyperthyroidism were included, no significant association was found between thyroid hormones and mini-mental status examination (MMSE) scores, despite the higher age of included patients (85 years). (Formiga et al., 2014) TSH did not show a significant association with CERAD subdomains in patients of 75 years and older, but higher TSH levels were significantly correlated with better CDR scores than lower levels. (Ojala et al., 2016)

While those studies focused on associations between thyroid function and cognitive tests, a further prospective cohort study did not show a significant value of TSH in predicting the progression from MCI to dementia within three years in euthyroid and hypothyroid older patients. (Pyun, Park & Kim, 2022) In contrast, lower TSH levels at baseline predicted progression to MCI or dementia over five years in Korean euthyroid and at baseline non-demented older patients. (Moon et al., 2014) In an Italian population of older adults, high TSH levels at baseline were associated with an increased risk of vascular dementia over a follow-up period of four years. (Forti et al., 2012) Furthermore, meta-analytical approaches showed higher risks of incident dementia in older adults with overt and subclinical hyperthyroidism. (Ye et al., 2024) It is, therefore, important to consider the potential higher risk of developing any type of dementia in those who showed a TSH-associated MCI diagnosis. Also, there is a lack of larger longitudinal population-based studies to assess the sex-specific evolution in high-risk groups.

The sample size and inclusion criteria seem to affect the reported results and explain the controversies in published data. The current study was based on estimating the overall cognition using ADAS13 scores, but the outcome of published data was sometimes based on continuous scores (MMSE, CERAD), or binary outcomes such as conversion to MCI or dementia, and particular interest was given to exploring the associations in higher TSH levels or hypothyroidism.

ADAS is a sensitive biomarker of global cognition, largely explored and validated in ADNI, and applied in all ADNI phases, motivating its inclusion in the current analysis. ADAS13 covers: Memory (immediate word recall, word recognition, orientation), Language (naming objects/fingers, following commands, spoken language ability, word-finding difficulty), Praxis (constructional praxis, ideational praxis), Attention and processing (number cancellation – specific to ADAS13), Executive function (Maze task – specific to ADAS13), Working memory (delayed word recall – specific to ADAS13), and Comprehension (spoken language comprehension) (Rosen, Mohs & Davis, 1984).

The variation in the ADAS13-specific subdomains outperforms the MMSE or ADAS11. This highlights the originality of the current study compared to published data.

Patients with dementia were not eligible for the current study, and other dementia-specific tests (CDR, FAQ…) are generally less sensitive to explore the initial and subtle stages of cognitive impairment, particularly MCI, where participants might have some sort of impairment but are still independent in their daily activities. The difference in their values between TSH subgroups was, however, statistically significant, as participants with lower TSH ranges had significantly worse CDR-SB and FAQ total scores.

Despite the statistical significance of the association between an increase of one unit in TSH levels and a decrease of 0.4 in the ADAS13 total score, this value does not represent a clinically relevant outcome. Studies consider modifications of ± 3 points predictive of clinically relevant cognitive changes. (Rockwood et al., 2017) Therefore, the second part of the current study was dedicated to evaluating the odds of MCI, a dichotomous outcome, which is more conventional and solid in estimating the current overall degree of cognitive impairment.

TSH levels tend to increase with age, and physiological ranges have to be adapted accordingly. Subclinical hypothyroidism is a more common condition in older adults and does not necessarily imply a pathological status or necessitate hormone therapy. (Biondi, Cappola & Cooper, 2019; Biondi & Cappola, 2022) In the current study, higher age was associated with higher TSH levels in the total study population. In a placebo-controlled randomized clinical trial, hormonal supplementation in older adults with subclinical hypothyroidism did not show a significant impact on cognitive function. (Parle et al., 2010)

In older populations (85 years), a longitudinal decrease in TSH levels was associated with a concomitant reduction in MMSE scores over a three- and five-year follow-up period. (Gan et al., 2021) Thyrotoxicosis is a rarer condition in which patients present very high thyroid hormones, either of endogenous or exogenous etiology (corresponding to very low TSH levels). In a study including 65,931 patients aged 65 years and older, thyrotoxicosis was associated with a significantly higher risk of cognitive decline. (Adams et al., 2023) Among the included cases from the ADNI cohort, three participants had TSH levels of < 0.01 and were converted to 0.01 for the analysis. Despite that, the present study showed a significant association between lower TSH levels (independently of thyrotoxicosis), worse cognitive scores, and higher odds of being diagnosed with MCI. These results were particularly significant under a specific TSH value in males, but not females. The low number of participants with TSH levels beyond clinically relevant TSH ranges explains the large confidence intervals in the sensitivity analysis.

4.2. Effect of sex on the relationship between thyroid function and cognition

Samuels and colleagues studied, in a cross-sectional and longitudinal design, the eventual associations between the variation of thyroid function within age-adapted normal ranges and cognition in older men (≥ 65 years). Neither TSH nor FT4 showed statistically significant results. (Samuels et al., 2016) This contradicts another study in older men aged between 70 and 89 years, where higher FT4 (but not TSH) levels were associated with higher dementia incidence. (Yeap et al., 2012) In women between 30 and 64 years old, higher TSH levels were significantly associated with a faster decline in clock-command scores. (Beydoun et al., 2015) Moreover, in a cross-sectional analysis of 2563 euthyroid persons aged between 50 and 80 years, higher TSH levels were significantly associated with the diagnosis of MCI only in women. (Winkler et al., 2016)

The present findings, particularly the male-specific association with MCI risk, confirm previous studies in favor of a sex-dependent association between thyroid and cognition in older adults. (Beydoun et al., 2013; Hallab, 2025) Lower TSH levels and higher odds of informant-perceived anxiety were significantly associated in older males but not females in the same source cohort. (Hallab, 2025) Animal studies demonstrated a sex-specific association between thyroid hormones and neurogenesis in the hippocampal subfields of male, but not female, mice. (Kapri, Pradhan, Vuruputuri & Vaidya, 2024) Furthermore, thyroid hormone receptors are expressed in a sex-specific manner, also in the brain. (Minakhina et al., 2020) Exposing mice to environmental endocrine disruptors, like pesticides, shows a sex-specific down-regulation of hippocampal TSH-specific receptors, anxiety, and cognitive impairment only in females. (Yang et al., 2020) In addition to the TSH-associated genetic expression and regulation of thyroid hormone receptors, (Baksi & Pradhan, 2021) it has been hypothesized that female-specific gonadal hormones might play a protective role in thyroid-related cognitive decline by regulating hippocampal TSH-specific receptors. (Dumas et al., 2006) Yet, eventual sex-specific confounding factors in human subjects need to be better understood. Randomized controlled trials and neurobiological studies are needed to address this gap.

4.3. Effect of age on the relationship between thyroid function and cognition

Thyroid function and cognition are differently associated across age groups. While studies in older adults reported controversial results, a large cross-sectional analysis including 10,362 middle-aged adults (mean: 49.5 ± 7.4 years) showed that lower TSH levels are associated with impaired executive function. (Szlejf et al., 2018) Moreover, higher TSH levels were associated with poorer cognition in adults under 59 years and better performance in those older than 60 years. (Beydoun et al., 2012)

4.4. Effect of neurodegeneration on the thyroid-cognition relationship

Although neither older adults with dementia nor TSH measurements in cerebrospinal fluid were included in the current analysis, studies have reported a biological association between thyroid hormones and Alzheimer’s pathology, even within euthyroid older adults. (Ge et al., 2022; Choi et al., 2020; Johansson et al., 2013) The association between TSH and Alzheimer’s disease biological biomarkers was out of the scope of this study, which was primarily dedicated to the sex difference in the association between TSH and overall cognition in non-demented persons. However, the study accounted for this risk, and models were adjusted for the APOE ε4 status. Further investigations are needed, particularly to assess whether sex similarly impacts the association in cases with Alzheimer’s disease.

4.5. Limitations

The first limitation of this study is related to the lower number of participants with thyroid levels beyond clinical cutoff values compared to participants with relatively “normal” TSH levels at baseline. This might be related to the healthcare awareness and compliance in persons with health-seeking behavior, who are more likely to consent and participate in a longitudinal cohort like ADNI. Despite the multiethnicity of the cohort, ADNI participants might be of a higher educational level and healthcare awareness compared to the general population. Therefore, larger population- and register-based studies are needed to replicate the findings. Moreover, people with severe comorbidities (including active thyroid cancer) were not eligible for the study. Furthermore, the study was only based on TSH levels without evaluating the potential effect of peripheral thyroid hormones (FT3 and FT4). This is mainly due to the lack of these assessments in the ADNI database. While FT3 and FT4 might have a direct effect on the brain and show associations with cognitive function, the lack of information on their levels limits the interpretability of the results. A way to reduce this bias was to include only cases with TSH<10 μIU/mL. FT3 and FT4 variations could, however, be predicted by TSH levels, especially when TSH is within normal ranges.

The second limitation is the non-inclusion of the medical history related to thyroid pathologies, particularly whether there was an underlying supplemented hypothyroidism or an eventual history of thyroid (surgery/radioiodine) ablation. This information was missing from the database of the included population. Moreover, information on potential concomitant autoimmune diseases such as Graves’ or Hashimoto thyroiditis was out of the scope of the study. This is primarily due to the coding system’s reliability, based on self-reporting of medical history, and the lack of this specific information in the included ADNI stages. The main exposure of the study was the central thyroid function based on TSH levels independently from etiologies and endocrine comorbidities. Using TSH as a biomarker of central thyroid function highlighted the aim of the current study, since it presented it independently from any underlying pathology or treatment, and focused only on TSH levels, as secreted by the anterior pituitary gland, without labelling it based on any ranges of function, pathology, medication, or intervention. The possibility of ongoing treatment cannot be ruled out even if TSH was within the normal range. However, the probability of whether any treatment or pathology caused TSH to fall to very low levels and be consequently associated with cognitive impairment needs to be explored in clinical longitudinal interventional studies.

The cross-sectional design is a further limitation of the study, and no causal inference can be concluded from the current analysis. This also implies that reverse causality can not be excluded (a theoretical hypothesis might be that cognitive function / neurodegeneration might impact thyroid function, but can not be explored in a cross-sectional design).

Larger longitudinal population-based and clinical studies are needed to infer a potential causal effect and assess the associations in risk groups, particularly to better understand the role of sex in modulating the observed associations.

4.6. Perspectives and significance

This study highlighted the importance of accounting for the role of sex in the association between TSH and cognition. Low TSH levels and concomitant MCI were significantly associated, and this association might be revealed only after stratifying by sex and exploring separately males and females. Most published studies focused on the effect of hypothyroidism (high TSH levels) on cognition. By excluding cases with TSH equal to or higher than 10 μIU/mL, this current study highlights, however, the association between lower TSH levels and worse cognition in advanced ages. This finding might have been masked in studies where hypothyroidism was much more prevalent.

Hypothyroidism-related cognitive impairment is largely believed to be a reversible condition. However, further longitudinal studies need to test whether lower TSH levels have a causal effect on cognitive decline (exp., iatrogenic TSH suppression in thyroid carcinoma-treated patients, or an over-supplementation of thyroid hormones in older patients, and their effect on cognition) and whether potential associations might be reversible.

Several studies have pointed out that the shifting of TSH toward higher levels in the older population has no pathological significance. Recommendations tended to be against thyroid hormone supplementation if only the TSH levels were slightly high in older adults. This study highlights a different cutoff value (higher than the commonly known 0.4 μIU/mL), under which the odds of concomitant cognitive impairment were higher in the ADNI population. Regular cognitive screenings may be appropriate in older patients under thyroid hormone supplementation, with a downregulation of TSH to lower levels. Regular cognitive screenings might be considered in TSH-suppressed thyroid-carcinoma-treated patients, as well. As previously mentioned, this cross-sectional study neither suggests causality nor excludes or rules out the hypothesis that middle-aged and older adults with cognitive impairment might express lower TSH levels. The sex-dependent neurobiology of cognition in the context of thyroid function needs to be further explored to decipher the differences between males and females. It is also important to replicate the findings in different populations and test for non-linearity. The predictive value and the biological significance of TSH-specific cutoff values in older males need to be further studied and better understood. It is also important to screen for thyroid dysfunction in middle-aged and older adults at higher risk of cognitive decline and those who already suffer from cognitive impairment.

5. Conclusions

This study provides additional arguments supporting the theory of a multifaceted brain-thyroid interaction in advanced age. Sex significantly influences the association between TSH and cognition in older populations, and this association might be subject to a cutoff value different from the commonly used clinical references. Lower TSH values and cognitive impairment were associated in older males but not females, but the underlying mechanisms remained unexplained and need to motivate further research.

Supplementary Material

MMC1

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.aggp.2025.100195.

Acknowledgments

“Data collection and sharing for the Alzheimer’s Disease Neuroimaging Initiative (ADNI) is funded by the National Institute on Aging (National Institutes of Health Grant U19 AG024904). The grantee organization is the Northern California Institute for Research and Education. In the past, ADNI has also received funding from the National Institute of Biomedical Imaging and Bioengineering, the Canadian Institutes of Health Research, and private sector contributions through the Foundation for the National Institutes of Health (FNIH) including generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; Bio-Clinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research &Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics.”

Funding

AH did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sector for this work. Data collection and sharing for the ADNI project were funded by the National Institutes of Health (Grant U19 AG024904). ADNI is made possible with funding from the NIH and private sector support detailed at https://adni.loni.usc.edu/about/.

Footnotes

Research material availability

The information on research material supporting the conclusions of this article is available at http://adni.loni.usc.edu.

Clinical trial number Not applicable.

Ethics statement

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki De claration of 1975, as revised in 2013. All procedures involving human subjects/patients were approved by local IRBs in each of the over 60 ADNI recruitment sites (a map summarizing the over 60 ADNI study sites can be found at https://adni.loni.usc.edu/about/governance/#core-details). The current study is based on a secondary analysis of anonymized data and complies with ADNI’s data use agreement (Hallab’s DUA with ADNI). An additional ethical approval was therefore not required.

Consent statement

Written informed consent was obtained from every ADNI study participant. More information is available at http://adni.loni.usc.edu.

CRediT authorship contribution statement

Asma Hallab: Visualization, Formal analysis, Writing – review & editing, Supervision, Data curation, Writing – original draft, Methodology, Conceptualization.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The datasets supporting the conclusions of this article are available at http://adni.loni.usc.edu.

Analytic code availability

The analytical code can be obtained upon written request to the corresponding author (asma.hallab@charite.de).

References

  1. Adams R, Oh ES, Yasar S, Lyketsos CG, & Mammen JS (2023). Endogenous and exogenous thyrotoxicosis and risk of incident cognitive disorders in older adults. JAMA Internal Medicine, 183(12), 1324–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Akintola AA, Jansen SW, van Bodegom D, van der Grond J, Westendorp RG, de Craen AJ, et al. (2015). Subclinical hypothyroidism and cognitive function in people over 60 years: A systematic review and meta-analysis. Frontiers in Aging Neuroscience, 7, 150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baksi S, & Pradhan A (2021). Thyroid hormone: Sex-dependent role in nervous system regulation and disease. Biology of Sex Differences, 12(1), 25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Barbero JD, Gutiérrez-Zotes A, Montalvo I, Creus M, Cabezas Á, Solé M, et al. (2015). Free thyroxine levels are associated with cognitive abilities in subjects with early psychosis. Schizophrenia Research, 166(1–3), 37–42. [DOI] [PubMed] [Google Scholar]
  5. Beydoun MA, Beydoun HA, Shroff MR, Kitner-Triolo MH, & Zonderman AB (2012). Serum leptin, thyroxine and thyroid-stimulating hormone levels interact to affect cognitive function among US adults: Evidence from a large representative survey. Neurobiology of Aging, 33(8), 1730–1743. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Beydoun MA, Beydoun HA, Kitner-Triolo MH, Kaufman JS, Evans MK, & Zonderman AB (2013). Thyroid hormones are associated with cognitive function: Moderation by sex, race, and depressive symptoms. The Journal of Clinical Endocrinology and Metabolism, 98(8), 3470–3481. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Beydoun MA, Beydoun HA, Rostant OS, Dore GA, Fanelli-Kuczmarski MT, Evans MK, et al. (2015). Thyroid hormones are associated with longitudinal cognitive change in an urban adult population. Neurobiology of Aging, 36(11), 3056–3066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Biondi B, & Cappola AR (2022). Subclinical hypothyroidism in older individuals. The Lancet. Diabetes & Endocrinology, 10(2), 129–141. [DOI] [PubMed] [Google Scholar]
  9. Biondi B, Cappola AR, & Cooper DS (2019). Subclinical hypothyroidism: A review. JAMA, 322(2), 153–160. [DOI] [PubMed] [Google Scholar]
  10. Booth T, Deary IJ, & Starr JM (2013). Thyroid stimulating hormone, free thyroxine and cognitive ability in old age: The Lothian birth cohort study 1936. Psychoneuroendocrinology, 38(4), 597–601. [DOI] [PubMed] [Google Scholar]
  11. Chen G, Lv H, Zhang X, Gao Y, Liu X, Gu C, et al. (2022). Assessment of the relationships between genetic determinants of thyroid functions and bipolar disorder: A mendelian randomization study. Journal of Affective Disorders, 298(Pt A), 373–380. [DOI] [PubMed] [Google Scholar]
  12. Chen H, Hu J, Yang X, Zhou Q, Hu Y, Tang X, et al. (2023). Low levels of free triiodothyronine are associated with risk of cognitive impairment in older euthyroid adults. Scientific Reports, 13(1), Article 22133. [Google Scholar]
  13. Choi BW, Kim S, Kang S, Won KS, Yi HA, & Kim HW (2020). Relationship between thyroid hormone levels and the pathology of Alzheimer’s disease in euthyroid subjects. Thyroid, 30(11), 1547–1555. [DOI] [PubMed] [Google Scholar]
  14. Dumas J, Hancur-Bucci C, Naylor M, Sites C, & Newhouse P (2006). Estrogen treatment effects on anticholinergic-induced cognitive dysfunction in normal postmenopausal women. Neuropsychopharmacology : Official Publication of the American College of Neuropsychopharmacology, 31(9), 2065–2078. [DOI] [PubMed] [Google Scholar]
  15. Formiga F, Ferrer A, Padros G, Contra A, Corbella X, & Pujol R (2014). Thyroid status and functional and cognitive status at baseline and survival after 3 years of follow-up: The OCTABAIX study. European Journal of Endocrinology, 170(1), 69–75. [DOI] [PubMed] [Google Scholar]
  16. Forti P, Olivelli V, Rietti E, Maltoni B, Pirazzoli G, Gatti R, et al. (2012). Serum thyroid-stimulating hormone as a predictor of cognitive impairment in an elderly cohort. Gerontology, 58(1), 41–49. [DOI] [PubMed] [Google Scholar]
  17. Freuer D, & Meisinger C (2023). Causal link between thyroid function and schizophrenia: A two-sample Mendelian randomization study. European Journal of Epidemiology, 38(10), 1081–1088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gan EH, & Pearce SH (2012). Clinical review: The thyroid in mind: Cognitive function and low thyrotropin in older people. The Journal of Clinical Endocrinology and Metabolism, 97(10), 3438–3449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gan EH, Jagger C, Yadegarfar ME, Duncan R, & Pearce SH (2021). Changes in serum thyroid function predict cognitive decline in the very old: Longitudinal findings from the Newcastle 85+ study. Thyroid, 31(8), 1182–1191. [DOI] [PubMed] [Google Scholar]
  20. Garber JR, Cobin RH, Gharib H, Hennessey JV, Klein I, Mechanick JI, et al. (2012). Clinical practice guidelines for hypothyroidism in adults: Cosponsored by the American Association of clinical endocrinologists and the American Thyroid Association. Thyroid, 22(12), 1200–1235. [DOI] [PubMed] [Google Scholar]
  21. Ge F, Dong L, Zhu D, Lin X, Shi J, & Xiao M (2022). Comparison of serum triiodothyronine with biomarkers for Alzheimer’s Disease continuum in euthyroid subjects. Journal of Alzheimer’s Disease : JAD, 85(2), 605–614. [DOI] [PubMed] [Google Scholar]
  22. Ge YJ, Xu W, Tan CC, & Tan L (2020). Blood-based biomarkers in hypothalamic-pituitary axes for the risk of dementia or cognitive decline: A systematic review and meta-analysis. Aging, 12(20), 20350–20365. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hallab A (2025). Sex-modulated association between thyroid stimulating hormone and informant-perceived anxiety in non-depressed older adults: Prediction models and relevant cutoff value. Scientific Reports, 15(1), 2526. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hughes CP, Berg L, Danziger WL, Coben LA, & Martin RL (1982). A new clinical scale for the staging of dementia. The British Journal of Psychiatry, 140, 566–572. [DOI] [PubMed] [Google Scholar]
  25. Jockwitz C, Wiersch L, Stumme J, & Caspers S (2021). Cognitive profiles in older males and females. Scientific Reports, 11(1), 6524. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Johansson P, Almqvist EG, Johansson JO, Mattsson N, Hansson O, Wallin A, et al. (2013). Reduced cerebrospinal fluid level of thyroxine in patients with Alzheimer’s disease. Psychoneuroendocrinology, 38(7), 1058–1066. [DOI] [PubMed] [Google Scholar]
  27. Kapri D, Pradhan A, Vuruputuri RM, & Vaidya VA (2024). Sex differences in the influence of adult-onset hypothyroidism on hippocampal progenitor survival and neuronal differentiation in mice. Journal of Neuroendocrinology, 36(12), Article E13453. [Google Scholar]
  28. Korevaar TI, Muetzel R, Medici M, Chaker L, Jaddoe VW, de Rijke YB, et al. (2016). Association of maternal thyroid function during early pregnancy with offspring IQ and brain morphology in childhood: A population-based prospective cohort study. The Lancet. Diabetes & Endocrinology, 4(1), 35–43. [DOI] [PubMed] [Google Scholar]
  29. Labad J, Barbero JD, Gutiérrez-Zotes A, Montalvo I, Creus M, Cabezas Á, et al. (2016). Free thyroxine levels are associated with cognitive changes in individuals with a first episode of psychosis: A prospective 1-year follow-up study. Schizophrenia Research, 171(1–3), 182–186. [DOI] [PubMed] [Google Scholar]
  30. Lai S, Zhong S, Zhang Y, Wang Y, Zhao H, Chen G, et al. (2021). Association of altered thyroid hormones and neurometabolism to cognitive dysfunction in unmedicated bipolar II depression. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 105, Article 110027. [Google Scholar]
  31. Lain SJ, Bentley JP, Wiley V, Roberts CL, Jack M, Wilcken B, et al. (2016). Association between borderline neonatal thyroid-stimulating hormone concentrations and educational and developmental outcomes: A population-based record-linkage study. The Lancet. Diabetes & Endocrinology, 4(9), 756–765. [DOI] [PubMed] [Google Scholar]
  32. Liu YY, & Brent GA (2021). The role of thyroid hormone in neuronal protection. Comprehensive Physiology, 11(3), 2075–2095. [DOI] [PubMed] [Google Scholar]
  33. Minakhina S, Bansal S, Zhang A, Brotherton M, Janodia R, De Oliveira V, et al. (2020). A direct comparison of thyroid hormone receptor protein levels in mice provides unexpected insights into thyroid hormone action. Thyroid, 30(8), 1193–1204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Moon JH, Park YJ, Kim TH, Han JW, Choi SH, Lim S, et al. (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). The Journal of Clinical Endocrinology and Metabolism, 99(2), 424–432. [DOI] [PubMed] [Google Scholar]
  35. Nichols E, Steinmetz JD, Vollset SE, Fukutaki K, Chalek J, Abd-Allah F, et al. (2022). Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: An analysis for the global burden of disease study 2019. The Lancet Public Health, 7(2), E105–Ee25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ojala AK, Schalin-Jäntti C, Pitkälä KH, Tilvis RS, & Strandberg TE (2016). Serum thyroid-stimulating hormone and cognition in older people. Age and Ageing, 45(1), 155–157. [DOI] [PubMed] [Google Scholar]
  37. Parle J, Roberts L, Wilson S, Pattison H, Roalfe A, Haque MS, et al. (2010). A randomized controlled trial of the effect of thyroxine replacement on cognitive function in community-living elderly subjects with subclinical hypothyroidism: The Birmingham Elderly Thyroid study. The Journal of Clinical Endocrinology and Metabolism, 95(8), 3623–3632. [DOI] [PubMed] [Google Scholar]
  38. Pasqualetti G, Pagano G, Rengo G, Ferrara N, & Monzani F (2015). Subclinical hypothyroidism and cognitive impairment: Systematic review and meta-analysis. The Journal of Clinical Endocrinology and Metabolism, 100(11), 4240–4248. [DOI] [PubMed] [Google Scholar]
  39. Pfeffer RI, Kurosaki TT, Harrah CH Jr, Chance JM, & Filos S (1982). Measurement of functional activities in older adults in the community. Journal of Gerontology, 37(3), 323–329. [DOI] [PubMed] [Google Scholar]
  40. Pocklington C, Gilbody S, Manea L, & McMillan D (2016). The diagnostic accuracy of brief versions of the geriatric depression scale: A systematic review and meta-analysis. International Journal of Geriatric Psychiatry, 31(8), 837–857. [DOI] [PubMed] [Google Scholar]
  41. Pyun JM, Park YH, & Kim S (2022). Subclinical hypothyroidism and cognitive impairment. Journal of Alzheimer’s Disease : JAD, 88(2), 757–762. [DOI] [PubMed] [Google Scholar]
  42. Rockwood K, Howlett SE, Hoffman D, Schindler R, & Mitnitski A (2017). Clinical meaningfulness of Alzheimer’s disease assessment scale–cognitive subscale change in relation to goal attainment in patients on cholinesterase inhibitors. Alzheimer’s & Dementia, 13(10), 1098–1106. [Google Scholar]
  43. Rosen WG, Mohs RC, & Davis KL (1984). A new rating scale for Alzheimer’s disease. The American Journal of Psychiatry, 141(11), 1356–1364. [DOI] [PubMed] [Google Scholar]
  44. Saglam Y, Ermis C, Tanyolac D, Oz A, Turan S, Korkmaz HA, et al. (2024). The association between plasma thyroxine levels and neurocognitive impairment in early-onset schizophrenia and other psychosis spectrum disorders. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 131, Article 110940. [Google Scholar]
  45. Samuels MH, Kaimal R, Waring A, Fink HA, Yaffe K, Hoffman AR, et al. (2016). Thyroid function variations within the reference range do not affect quality of life, mood, or cognitive function in community-dwelling older men. Thyroid : Official Journal of the American Thyroid Association, 26(9), 1185–1194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Sheikh JI, & Yesavage JA (1986). Geriatric Depression scale (GDS): Recent evidence and development of a shorter version. Clinical Gerontologist: The Journal of Aging and Mental Health, 5(1–2), 165–173. [Google Scholar]
  47. Sohn D, Shpanskaya K, Lucas JE, Petrella JR, Saykin AJ, Tanzi RE, et al. (2018). Sex differences in cognitive decline in subjects with high likelihood of mild cognitive impairment due to Alzheimer’s disease. Scientific Reports, 8(1), 7490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Szlejf C, Suemoto CK, Santos IS, Lotufo PA, Haueisen Sander Diniz MF, Barreto SM, et al. (2018). Thyrotropin level and cognitive performance: Baseline results from the ELSA-Brasil study. Psychoneuroendocrinology, 87, 152–158. [DOI] [PubMed] [Google Scholar]
  49. van Vliet NA, van Heemst D, Almeida OP, Åsvold BO, Aubert CE, Bae JB, et al. (2021). Association of thyroid dysfunction with cognitive function: An individual participant data analysis. JAMA Internal Medicine, 181(11), 1440–1450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Vanderpump MPJ, Luster M, Duntas LH, & Wartofsky L (2019). Epidemiology of thyroid disorders. The thyroid and its diseases: A comprehensive guide for the clinician (pp. 75–85). Cham: Springer International Publishing. [Google Scholar]
  51. Winkler A, Weimar C, Jockel KH, Erbel R, Dragano N, Broecker-Preuss M, et al. ¨ (2016). Thyroid-stimulating hormone and mild cognitive impairment: Results of the Heinz Nixdorf Recall study. Journal of Alzheimer’s Disease : JAD, 49(3), 797–807. [Google Scholar]
  52. Xing D, Liu D, Li R, Zhou Q, & Xu J (2021). Factors influencing the reference interval of thyroid-stimulating hormone in healthy adults: A systematic review and meta-analysis. Clinical Endocrinology, 95(3), 378–389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Yang Y, Zhan F, Wang YC, Wang B, Shi MX, Guo C, et al. (2020). Pubertal fenvalerate exposure impairs cognitive and behavioral development partially through down-regulating hippocampal thyroid hormone receptor signaling. Toxicology Letters, 332, 192–201. [DOI] [PubMed] [Google Scholar]
  54. Ye J, Huang Z, Liang C, Yun Z, Huang L, Liu Y, et al. (2024). Thyroid dysfunction and risk of different types of dementia: A systematic review and meta-analysis. Medicine, 103(34), Article E39394. [Google Scholar]
  55. Yeap BB, Alfonso H, Chubb SA, Puri G, Hankey GJ, Flicker L, et al. (2012). Higher free thyroxine levels predict increased incidence of dementia in older men: The Health in Men Study. The Journal of Clinical Endocrinology and Metabolism, 97 (12), E2230–E2237. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

MMC1
NIHMS2105998-supplement-MMC1.docx (57.1KB, docx)

Data Availability Statement

The datasets supporting the conclusions of this article are available at http://adni.loni.usc.edu.

RESOURCES