Sleep traits and thyroid gland: results from National Health and Nutrition Examination Survey 2007-2012 and Mendelian randomization analyses

doi:10.21203/rs.3.rs-4840632/v1

Download PDF

Research Article

Sleep traits and thyroid gland: results from National Health and Nutrition Examination Survey 2007-2012 and Mendelian randomization analyses

https://doi.org/10.21203/rs.3.rs-4840632/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Common sleep problems reduce quality of life and increase chronic disease risk. The relationship between sleep traits and thyroid function is unclear. This study explores the association between sleep traits and thyroid using NHANES data and Mendelian randomization (MR) analysis.

Materials and Methods: Data from NHANES 2007-2012 were used to assess the relationship between sleep traits and thyroid function using weighted multivariable-adjusted logistic regression. A two-sample MR study was conducted using GWAS summary statistics, and methods like Inverse Variance Weighted (IVW) were used to explore the causal relationship between sleep traits and thyroid disease. Sensitivity analysis ensured robustness.

Results: The study included 6919 NHANES participants. Logistic regression showed higher TSH levels in the long sleep group (P < 0.0001, β= 0.85, 95% CI: 0.54, 1.15). Lower FT3 levels were found in the normal sleep group (P = 0.0030, β= -0.06, 95% CI: -0.06, -0.00). TT4 levels were lower in those with sleep disorders (P = 0.0157, β= -0.11, 95% CI: -0.20, -0.02). Long sleep was positively associated with TGAb positivity (P = 0.0288, OR = 1.81, 95% CI: 1.06, 3.07), while sleep disorders were negatively associated with TGAb positivity (P = 0.0176, OR = 0.72, 95% CI: 0.56, 0.95). MR analysis indicated a positive association between long sleep and Graves' disease (GD) risk (P = 0.0240, OR = 99.98, 95% CI: 1.83, 5453.63), and a negative association between sleep duration and Hashimoto's thyroiditis (HT) risk (P = 0.0294, OR = 0.72, 95% CI: 0.54, 0.97).

Conclusion: Sleep duration is associated with thyroid hormone levels and autoimmunity. Long sleep is linked to higher TSH levels, lower FT3 levels, and increased TGAb positivity risk, while sleep disorders are linked to lower TT4 levels and decreased TGAb positivity risk. MR studies suggest long sleep may increase GD risk, while shorter sleep may decrease HT risk.

Epigenetics & Genomics

Nutrition & Dietetics

Endocrinology & Metabolism

Head & Neck Surgery

Mendelian randomization

National Health and Nutrition Examination Survey

sleep traits

cross-sectional study

causality

The thyroid gland is the largest endocrine gland in the body and its main function is to secrete thyroid hormones. The regulation of the availability of thyroid hormones, including triiodothyronine (T3) and thyroxine (T4), is a complex and diverse process. This regulation is not only at the central level, i.e., the hypothalamic-pituitary-thyroid (HPT) axis but also involves fine-tuning at the local level. At the local level, thyroid hormone availability is influenced by the differential expression of a variety of factors such as thyroid hormone transporter proteins, deiodinases, and thyroid nuclear hormone receptors¹. In addition, transcriptional co-repressors and co-activators are involved in this regulatory process, which together ensure the homeostasis and efficient utilization of thyroid hormone in the body^1,2. Active free thyroid hormones in the body include free triiodothyronine (FT3) and free thyroxine (FT4), which are converted to total T3 (TT3) and total T4 (TT4) when FT3 and FT4 are depleted³. Under physiologic conditions, circulating concentrations of thyroid-stimulating hormone (TSH) and FT4 are regulated by negative feedback through the HPT axis. In thyroid dysfunction, central regulation may fail, leading to abnormal changes in thyroid hormone levels and causing the corresponding symptoms of hypothyroidism or hyperthyroidism. Hashimoto’s thyroiditis (HT) is characterized by the presence of thyroid peroxidase antibodies (TPOAb) and thyroglobulin antibodies (TGAb), as well as diffuse lymphocytic infiltration of the thyroid tissue ⁴. In contrast, Graves’disease (GD) is associated with autoantibodies against the TSH receptor⁵. HT and GD are the result of an attack by the immune system on the thyroid tissue, which ultimately manifests as changes in thyroid hormone levels, namely hyperthyroidism and hypothyroidism. Various lifestyle habits have been shown to cause changes in thyroid hormone levels, such as smoking, alcohol consumption, diet, and exercise⁶. In addition, almost all hormones are produced in a cyclical rhythm over 24 hours, and sleep has different effects on the regulation of this rhythm⁷.

Sleep plays a vital role in maintaining overall physical and mental health. In the United States, approximately one-third of adults do not get the recommended 7 to 9 hours of sleep per night⁸. More worryingly, millions of Americans self-report sleep disorders each year⁹. It is worth noting that, based on observational data from the past few decades, some Western countries are also gradually showing a trend of sleep-related disorders and shortened sleep time¹⁰. Common sleep problems such as insufficient sleep, insomnia, and difficulty falling asleep not only reduce the quality of life, but also cause economic burdens and are closely related to physical and mental health and an increased risk of various chronic diseases (such as inflammation, metabolic syndrome, obesity, stroke, diabetes, and cancer)^11–16. Hyperthyroidism and hypothyroidism are known to be common causes of sleep disorders. One study found that short sleepers have a higher risk of subclinical hyperthyroidism (elevated TSH)¹⁷. Another study showed that people with low thyroid hormone levels or even subclinical hypothyroidism (reduced TSH) typically have longer sleep latency and shorter sleep duration compared to people with normal thyroid function ¹⁸. In contrast, the study by Lynn Kessler et al. evaluated the effects of prolonged, moderate sleep deprivation on the circadian rhythm of TSH secretion. While acute, extreme deprivation increases TSH secretion and release, chronic, moderate sleep deprivation suppresses TSH secretion¹⁹. However, the aforementioned studies were observational and had small sample sizes. Therefore, the present study was designed to explore the correlation between sleep traits (sleep duration and sleep disorder) and thyroid using cross-sectional data from the National Health and Nutrition Examination Survey (NHANES). In addition, quantifying the causal effects of traditional observational studies is challenging as observational studies do not completely eliminate potential bias against confounders. Therefore, this study will also incorporate Mendelian randomization (MR) analyses to validate the findings of cross-sectional studies from the perspective of genetic variation and further assess the causal relationship between sleep characteristics (sleep duration, long sleep, short sleep, and insomnia) and thyroid disorders (hyperthyroidism, hypothyroidism, HT, and GD).

2.1 Study design

This study combined an observational epidemiologic study with MR analysis. Clinical associations between sleep traits (sleep duration and sleep disorder) and thyroid function were first explored by epidemiologic analysis of NHANES. Next, a two-sample MR analysis using genome-wide association study (GWAS) statistical pooled data was performed to validate the causal association between sleep traits (sleep duration, long sleep, short sleep, and insomnia) and thyroid disorders (hyperthyroidism, hypothyroidism, HT, and GD). MR relies on three basic assumptions: (1) the association hypothesis: genetic variation is associated with exposure; (2) the independence hypothesis: genetic variation is independent of confounding factors between exposure and outcome; and (3) the exclusivity hypothesis: genetic variation affects the outcome only through exposure. The flowchart of the MR study design is shown in Fig. 1.

2.2 Epidemiological observation study

2.2.1 Data source

The NHANES project is part of the National Center for Health Statistics (NCHS) and is a cross-sectional survey that uses a complex multistage probability sampling design to generate a nationally representative sample of the noninstitutionalized civilian population of the United States. The NCHS Research Ethics Review Board approved the NHANES study, and all NHANES participants signed informed consent. This study used data from the 2007 to 2012 NHANES cycles as a secondary analysis from a public data source, so no additional ethics approval or informed consent was required.

Data on 30,442 participants from 2007–2012 were publicly available from the NHANES database (https://www.cdc.gov/nchs/nhanes/index.htm). Among all subjects, we excluded the following individuals: (1) individuals with missing data on thyroid function. (2) Individuals with missing information on sleep duration and sleep disorder (for participants who answered "don't know" or "refused", their data were considered as missing values). (3) Previous thyroid disease. (4) Pregnancy women. (4) Individuals with incomplete data on education, body mass index (BMI), poverty-to-income ratio (PIR). Finally, 6919 individuals were included. The flowchart of patient screening is shown in Fig. 2.

2.2.2 Measurements and definitions

The sleep traits addressed in this study included sleep duration and sleep disorder. Sleep duration was assessed based on a self-report questionnaire (SLD010H) in which participants were asked to answer the question "How much sleep do you usually get at night on weekdays or workdays?". If they reported 12 hours or more, it was recorded as 12 hours. According to the National Sleep Foundation's recommendations for sleep duration²⁰, sleep duration was categorized into three groups in this study: short sleep (< 7 hours/day), normal (7–9 hours/day), and long sleep (> 9 hours/day). The assessment of sleep disorder was based on a self-report questionnaire (SLD050H) and participants were asked to answer the question "Have you ever told a doctor or other health professional that you have trouble sleeping? ". Sleep disorder were categorized into two groups based on the responses: non-sleep disorder and sleep disorder.

The serum samples involved in the outcome of this study for thyroid function indices included TSH, FT3, FT4, TT3, TT4, thyroglobulin (TG), TPOAb, and TGAb.The NHANES Laboratory/Healthcare Technician Procedures Manual for specimen collection and processing provides a comprehensive and detailed description of the methodology used in the collection and processing of serum specimens. TPOAb > 9.0 IU/mL and TGAb > 4.0 IU/mL were classified as positive, respectively.

2.2.3 Covariates

Covariates in this study included race (Mexican American, other Hispanic, non-Hispanic white, non-Hispanic black, and other races), age, sex (male and female), education (less than high school, high school or equivalent, and high school above), marital status (married/living with partner, widowed/divorced/separated, never married), BMI (less than normal, normal, overweight, obesity), PIR (low, medium, high), smoking status (no and yes), diabetes (no and yes), hypertension (no and yes) data were analyzed. Socioeconomic status was assessed using PIR values, and PIR was divided into three groups: low (< 1.5), medium (1.5–3.5), and high (> 3.5). Weight was divided into four groups: below normal (≤ 18.5 kg/m2), normal (> 18.5 and ≤ 25.0 kg/m2), overweight (> 25.0 and ≤ 30.0 kg/m2), and obese (> 30 kg/m2).

2.3 Mendelian randomization analysis

2.3.1 Data source

The exposures included in this study were sleep traits, where sleep traits were categorized as long sleep, short sleep, sleep duration, and insomnia; and the outcomes included were thyroid-related disorders, which included hyperthyroidism, hypothyroidism, HT, and GD. Ethical approval or informed consent was not required as the data were obtained from secondary analyses of publicly available data sources. Specific information on the GWAS data in this study is detailed in the supplementary material: Table 1.

Table 1

Thyroid function and thyroid autoimmunity of NHANES (2007–2012) study population in sleep duration groups and sleep disorder groups.
Characteristics	Sleep duration				P-value	Sleep disorder			P-value
Characteristics	Total	Short sleep	Normal	Long sleep	P-value	Total	Non-sleep disorder	Sleep disorder	P-value
N^b	6919(100)	2816(40.7)	3934(56.86)	169(2.44)		6919(100)	5346(77.27)	1573(22.73)
Gender^b					< 0.0001				< 0.0001
Male	2989(43.20)	770(27.35)	2150(54.66)	69(40.79)		3117(45.05)	2747(51.39)	370(23.53)
Female	3930(56.80)	2046(72.65)	1784(45.34)	100(59.21)		3802(54.95)	2599(48.61)	1203(76.47)
Age (year)^a		60.07 ± 18.89	49.66 ± 17.22	51.02 ± 21.47	< 0.0001		52.18 ± 19.36	55.69 ± 13.06	< 0.0001
Race^b					< 0.0001				< 0.0001
Mexican American	232(3.35)	100(3.53)	119(3.03)	13(7.92)		219(3.16)	184(3.44)	35(2.22)
Other hispanic	162(2.35)	83(2.96)	63(1.60)	16(9.79)		143(2.07)	111(2.08)	32(2.02)
Non-Hispanic White	5345(77.25)	2324(82.53)	2912(74.01)	109(64.43)		5340(77.18)	3942(73.74)	1398(88.86)
Non-Hispanic Black	986(14.25)	209(7.42)	756(19.23)	21(12.15)		1040(15.03)	961(17.98)	79(5.03)
Other-race	194(2.80)	100(3.56)	84(2.13)	10(5.71)		177(2.56)	148(2.76)	29(1.87)
BMI (kg/m2)^b					< 0.0001				< 0.0001
Less than normal	47(0.68)	16(0.55)	26(0.65)	5(3.18)		43(0.62)	36(0.67)	7(0.47)
Normal	2567(37.10)	939(33.34)	1570(39.92)	58(33.96)		2713(39.21)	1720(32.19)	993(63.13)
Overweight	2515(36.35)	1343(47.70)	1116(28.37)	56(33.22)		2306(33.33)	2050(38.34)	256(16.27)
Obesity	1790(25.87)	518(18.40)	1222(31.06)	50(29.64)		1857(26.84)	1540(28.80)	317(20.12)
Smoke^b					< 0.0001				< 0.0001
No	4031(58.26)	1253(44.50)	2688(68.33)	90(53.00)		4260(61.57)	3125(58.46)	1135(72.14)
Yes	2888(41.74)	1563(55.50)	1246(31.67)	79(47.00)		2659(38.43)	2221(41.54)	438(27.86)
Poverty-to-income ratio^b					< 0.0001				< 0.0001
Low	1355(19.59)	970(34.45)	311(7.91)	74(44.09)		1114(16.10)	896(16.77)	218(13.89)
Medium	4362(63.04)	1284(45.60)	3015(76.63)	63(37.19)		4618(66.74)	3618(67.67)	1000(63.58)
High	1202(17.37)	562(19.94)	608(15.46)	32(18.72)		1187(17.16)	832(15.56)	355(22.54)
Education level^b					< 0.0001				< 0.0001
Less than high school	1775(25.65)	822(29.19)	898(22.83)	55(32.36)		1661(24.01)	1515(28.34)	146(9.29)
High school or equivalent	702(10.15)	355(12.61)	290(7.37)	57(33.79)		642(9.28)	460(8.61)	182(11.57)
High school above	4442(64.20)	1639(58.20)	2746(69.79)	57(33.85)		4616(66.71)	3371(63.05)	1245(79.14)
Marital status^b					< 0.0001				< 0.0001
Married, or living with partner	4476(64.69)	1710(60.72)	2679(68.09)	87(51.51)		4598(66.45)	3352(62.69)	1246(79.21)
Widowed, divorced, or separated	1078(15.58)	826(29.32)	209(5.31)	43(25.63)		881(12.74)	684(12.80)	197(12.54)
Never married	1365(19.73)	280(9.96)	1046(26.60)	39(22.86)		1440(20.81)	1310(24.51)	130(8.26)
Hypertension^b					< 0.0001				< 0.0001
No	4887(70.63)	1829(64.94)	2955(75.12)	103(61.18)		5005(72.34)	3754(70.22)	1251(79.52)
Yes	2032(29.37)	987(35.06)	979(24.88)	66(38.82)		1914(27.66)	1592(29.78)	322(20.48)
Diabetes^b					< 0.0001				< 0.0001
No	6036(87.24)	2701(95.91)	3185(80.95)	150(88.49)		5951(86.01)	4469(83.60)	1482(94.19)
Yes	883(12.76)	115(4.09)	749(19.05)	19(11.51)		968(13.99)	877(16.40)	91(5.81)
TSH (mIU/L)^a		1.60 ± 1.20	1.91 ± 1.42	2.32 ± 4.11	< 0.0001		1.69 ± 1.44	2.40 ± 1.06	< 0.0001
FT3 (pg/mL)^a		3.12 ± 0.43	3.16 ± 0.29	3.11 ± 0.41	0.0002		3.15 ± 0.35	3.12 ± 0.26	0.0167
FT4 (ng/dL)^a		0.86 ± 0.23	0.74 ± 0.10	0.81 ± 0.18	< 0.0001		0.78 ± 0.17	0.79 ± 0.10	0.0037
TT3 (ng/dL)^a		107.33 ± 18.54	113.73 ± 16.53	111.61 ± 22.99	< 0.0001		109.51 ± 16.26	121.80 ± 19.03	< 0.0001
TT4 (ug/dL)^a		8.07 ± 1.45	6.79 ± 1.66	8.03 ± 1.64	< 0.0001		6.89 ± 1.64	8.51 ± 1.35	< 0.0001
TG (ng/mL)^a		17.62 ± 19.82	11.09 ± 18.73	18.54 ± 25.06	< 0.0001		12.12 ± 19.42	17.61 ± 18.42	< 0.0001
TPOAb^b					0.0008				0.1501
Negative	6654(96.17)	2687(95.41)	3813(96.93)	154(90.85)		6669(96.39)	5163(96.58)	1506(95.75)
Positive	265(3.83)	129(4.59)	121(3.07)	15(9.15)		250(3.61)	183(3.42)	67(4.25)
TGAb^b					0.0029				0.3568
Negative	6750(97.56)	2739(97.26)	3856(98.01)	155(91.87)		6761(97.72)	5229(97.82)	1532(97.39)
Positive	169(2.44)	77(2.74)	78(1.99)	14(8.13)		158(2.28)	117(2.18)	41(2.61)
^a :Continuous Variables: weighted mean (SD).
^b :Categorical variable: actual frequency (weighted percentage).

Abbreviations: TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxine; TT3, total T3; TT4, total T4; TG, thyroglobulin; TPOAb, thyroid peroxidase antibody; TGAb: thyroglobulin antibody; BMI, body mass index; SD, Standard deviation.

Hassan S Dashti et al.²¹ obtained genetic association data on sleep duration using GWAS data published by the UK Biobank for 446,118 adults of European ancestry. In this study, the mean self-reported habitual sleep duration was 7.2 hours with a standard deviation of 1.1 hours. Subsequent analysis converted the continuous variable of sleep duration into a categorical variable. There were 106,192 cases of short sleep (< 7 hours) and 34,184 cases of long sleep (≥ 9 hours).

Jacqueline M Lane et al.²² analyzed self-reported insomnia symptoms in 453,379 European participants in the UK Biobank. Participants were asked to answer the question “Do you have difficulty falling asleep or wakeing up in the middle of the night?”. The analysis showed that 29% reported frequent insomnia symptoms (“usually”), and the prevalence was higher in women (32% vs. 24%) and older participants, shift workers, and those who self-reported shorter sleep duration.

The GWAS data of hyperthyroidism and hypothyroidism were obtained from IEU OPEN GWAS. The genetic association data of hyperthyroidism included 3545 cases and 459,388 controls. The genetic association data of hypothyroidism included 22,687 cases and 440,246 controls.

The GWAS data for HT and GD were compiled and published in 2021 by Saori Sakaue et al.²³. The pooled data for GD included 395,640 European participants (15,654 cases and 379,986 controls). The pooled GWAS data for HT included 458,620 European participants (1678 cases and 456,942 controls).

2.3.2 Selection of instrument variables

The selection criteria for instrumental variables (IVs) were as follows:(1) single nucleotide polymorphism (SNP) loci with genome-wide significance were screened from the exposure database. (2) Perform the PLINK clustering algorithm with a 10,000 kb linkage disequilibrium (LD) window to ensure independence (R² < 0.0001). (3) Excluded SNP with F statistic < 10, as F > 10 indicates strongly associated IVs.(4) Harmonized the exposure and outcome datasets and excluded palindromic sequences to ensure that the effects of SNP on exposure and outcome were from the same allele. SNP used as IVs in this study are listed in Supplementary material: Tables 2, 3, 4, and 5.

Table 2

Effects of sleep duration and sleep disorders on thyroid function.
	Model 1^a	P-value	Model 2^a	P-value
	β (95% CI)	P-value	β (95% CI)	P-value
TSH (mIU/L)
Sleep duration
Short sleep	Reference		Reference
Normal	0.15(0.06,0.25)	0.0018	0.09(−0.01,0.18)	0.0767
Long sleep	0.97(0.66,1.28)	< 0.0001	0.85(0.54,1.15)	< 0.0001
Sleep disorder
Non-sleep disorder	Reference		Reference
Sleep disorder	0.01(−0.10,0.12)	0.8354	−0.06(−0.17,0.05)	0.3043
FT3 (pg/mL)
Sleep duration
Short sleep	Reference		Reference
Normal	−0.04(−0.07,−0.01)	0.0036	−0.03(−0.06,−0.00)	0.0030
Long sleep	−0.12(−0.21,−0.04)	0.0053	−0.06(−0.14,0.02)	0.1517
Sleep disorder
Non-sleep disorder	Reference		Reference
Sleep disorder	−0.08(−0.11,−0.05)	< 0.0001	−0.02(−0.06,,0.01)	0.1149
FT4 (ng/dL)
Sleep duration
Short sleep	Reference		Reference
Normal	0.01(−0.00,0.01)	0.0875	0.01(−0.00,0.01)	0.1088
Long sleep	0.02(−0.00,0.04)	0.1099	0.01(−0.01,0.04)	0.2219
Sleep disorder
Non-sleep disorder	Reference		Reference
Sleep disorder	−0.01(−0.02,0.00)	0.0670	−0.01(−0.02,0.00)	0.0614
TT3 (ng/dL)
Sleep duration
Short sleep	Reference		Reference
Normal	−1.08(−2.26,0.10)	0.0731	−0.47(−1.62,0.67)	0.4180
Long sleep	−5.06(−8.85,−1.28)	0.0088	−2.83(−6.46,0.81)	0.1275
Sleep disorder
Non-sleep disorder	Reference		Reference
Sleep disorder	−1.83(−3.20,−0.46)	0.0089	−0.11(−1.47,1.24)	0.8690
TT4 (ug/dL)
Sleep duration
Short sleep	Reference		Reference
Normal	-−0.00(−0.08,0.08)	0.9727	0.04(−0.04,0.11)	0.3589
Long sleep	0.25(−0.01,0.50)	0.0455	0.21(−0.04,0.45)	0.0961
Sleep disorder
Non-sleep disorder	Reference		Reference
Sleep disorder	−0.04(−0.13,0.05)	0.3646	−0.11(−0.20,−0.02)	0.0157
Model 1^a: no covariates were adjusted.
Model 2^b: age, gender, race, poverty-to-income ratio, marital, body mass index, smoke, education level, marital status, hypertension, diabetes were adjusted.
Abbreviations: TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxine; TT3, total T3; TT4, total T4; TG, thyroglobulin.

Table 3

Effects of sleep duration and sleep disorders on thyroid autoimmunity.
	Model 1^a	P-value	Model 2^a	P-value
	OR (95% CI)	P-value	OR(95% CI)	P-value
TPOAb^b
Sleep duration
Short sleep	Reference		Reference
Normal	1.14(0.96,1.36)	0.1309	1.03(0.86,1.23)	0.7563
Long sleep	1.64(1.02,2.65)	0.0416	1.39(0.85,2.27)	0.1857
Sleep disorder
Non-sleep disorder	Reference		Reference
Sleep disorder	0.90(0.73,1.10)	0.2993	0.83(0.67,1.02)	0.0803
TGAb^b
Sleep duration
Short sleep	Reference		Reference
Normal	1.21(0.98,1.50)	0.0737	1.10(0.88,1.36)	0.4096
Long sleep	2.23(1.33,3.74)	0.0024	1.81(1.06,3.07)	0.0288
Sleep disorder
Non-sleep disorder	Reference		Reference
Sleep disorder	0.77(0.60,1.00)	0.0478	0.72(0.56,0.95)	0.0176
Model 1^a: no covariates were adjusted.
Model 2^b: age, gender, race, poverty-to-income ratio, marital, body mass index, smoke, education level, marital status, hypertension, diabetes were adjusted.
Abbreviations: TSH, thyroid-stimulating hormone; FT3, free triiodothyronine; FT4, free thyroxine; TT3, total T3; TT4, total T4; TG, thyroglobulin.

2.4 Statistical Analyses

2.4.1 Cross-Sectional study

All statistical analyses were performed according to the recommendations of the Centers for Disease Control and Prevention using appropriate NHANES sampling weights. Continuous variables are expressed as standard deviation (x ± s), whereas categorical variables are expressed as n (%). Between-group differences were assessed by weighted linear regression for continuous variables or weighted chi-square test for categorical variables. The association between sleep behavior and thyroid was studied by multiple linear regression using two different models. The first model did not adjust for covariates. The second model adjusted for age, gender, race, education, marital status, PIR, BMI, smoking, diabetes, and hypertension.

All analyses were performed by EmpowerStats (4.0) (http://www.empowerstats.com) and R software (http://www.r-project.org) using MEC weights. p values < 0.05 were considered statistically significant.

2.4.2 Two-Sample mendelian randomization

All MR analyses were performed using the TwoSampleMR package in R (4.2.1) software. In this study, the inverse variance weighted (IVW) method was used to pool the Wald estimates of individual SNP for weighted linear regression to summarize the total effect value. IVW, as the main analysis method, can provide unbiased causal effect estimates in the absence of horizontal pleiotropy. At the same time, to improve the robustness of the results, MR-Egger regression and weighted median estimator (WME) were used as supplementary methods of IVW because these methods can provide more reliable effect estimates under more relaxed conditions. In addition, we also used Cochran's Q test calculated by the IVW method to estimate heterogeneity. At the same time, the MR-Egger intercept test was used to evaluate and correct for potential pleiotropy. Finally, the robustness of the results was further verified by observing the funnel plot and leave-one-out plot.

3.1 Baseline characteristics

The baseline characteristics of the participants are summarized in Table 1. After strict screening, a total of 6919 subjects were included in this study. In the sleep duration group, 3930 (56.80%) were female, 5345 (77.25%) were non-Hispanic white, 2567 (37.10%) had normal BMI, 4362 (63.04%) had a medium PIR level, 4442 (64.20%) had an education level higher than high school, and 4476 (64.69%) were married or living with a partner. In addition, there were 2032 cases of hypertension, 883 cases of diabetes, and 2888 cases of smoking, of which 265 were TPOAb positive and 169 were TGAb positive. In the sleep disorder group, 3802 (54.95%) were female, 5340 (77.18%) were non-Hispanic white, 2713 (39.21%) had normal BMI, 4618 (66.74%) had a medium family income, 4616 (66.71%) had an education level higher than high school, and 4598 (66.45%) were married or living with a partner. In addition, 1914 had hypertension, 968 had diabetes, and 2659 smoked, of which 183 were TPOAb positive and 158 were TGAb positive.

3.2 Relationship between sleep traits and thyroid

The results of the multivariate regression analyses have been presented in Tables 2 and 3. In Model 2 adjusted for covariates, we found significantly higher TSH levels in the long-sleep group compared to those in the short-sleep group (P < 0.0001 β = 0.85, 95% CI: 0.54, 1.15). FT3 levels were lower in the normal group compared to those with short sleep (P = 0.0030 β = -0.06, 95% CI: -0.06, -0.00). Lower TT4 levels were also found in people with sleep disorder compared to those with non-sleep disorder (P = 0.0157 β = -0.11, 95% CI: -0.20, -0.02).

In terms of thyroid autoimmunity, from model 2, it was found that long sleep was positively associated with the risk of TGAb positivity (P = 0.0288 OR = 1.81, 95% CI 1.06, 3.07), whereas sleep disorder were negatively associated with TGAb positivity (P = 0.0176 OR = 0.72, 95% CI: 0.56, 0.95).

3.3 Mendelian randomization analysis

All MR analysis results have been presented in Figs. 3, 4, and 5. IVW results indicate that there may be a positive association between long sleep and the risk of GD (P = 0.0240 OR = 99.98, 95% CI 1.83, 5453.63), while there may be a negative association between sleep duration and the risk of HT (P = 0.0294 OR = 0.72, 95% CI 0.54, 0.97). In this study, there was heterogeneity between sleep traits and hypothyroidism, heterogeneity between sleep duration and hyperthyroidism, heterogeneity between short sleep and insomnia and GD, and the remaining results showed no heterogeneity. or pleiotropic effects (P > 0.05) (Supplementary material: Table 6). We also observed the funnel plot, which was roughly symmetrical, indicating that the risk of bias was relatively low and the results were highly reliable (Supplementary material: Fig. 1). The leave-one-out method was used to eliminate SNP one by one, and no specific SNP was found to cause significant changes in the results (Supplementary material: Fig. 2).

This study is the first to combine large-scale observational study data and MR analysis to explore the association between sleep traits and thyroid. In terms of thyroid function, cross-sectional studies have shown that compared with short sleep, long sleep TSH levels are significantly higher, and normal sleep FT3 levels are lower. It was also found that people with sleep disorder had lower TT4 levels than those with non-sleep disorder. In terms of thyroid autoimmunity, the study found that long sleep was positively correlated with the risk of TGAb positivity, while sleep disorder were negatively correlated with the risk of TGAb positivity. In addition, further two-sample MR analysis verified this conclusion, indicating that there may be a positive correlation between long sleep and GD risk, and there may be a negative correlation between sleep duration and HT risk.

The HPT axis is controlled by the circadian rhythm system, which plays a key role in the sleep-wake cycle, and disruption of one axis leads to dysregulation of the other²⁴. The results of previous studies have assessed, to varying degrees, the effects of sleep on thyroid function. There is evidence that sleep affects thyroid hormone secretion, leading to a decrease in the circadian amplitude of TSH, which in turn feedback suppresses thyroid hormones through the HPT axis²⁵. Previous observational studies have observed circadian rhythm changes in TSH secretion to varying degrees^26–29. Epidemiologic findings suggest that there are long-term effects of chronic circadian rhythm disruption on human health, and indeed chronic sleep deprivation disrupts rhythmic TSH secretion^24,30. Although acute, extreme deprivation increases TSH secretion and increases the surge in TSH release, chronic, moderate sleep deprivation suppresses the effects of circadian TSH secretion¹⁹. The results of a recent study showed an increase in TSH levels and a decrease in FT3 levels with increasing sleep duration, which is broadly similar to our study³¹. A similar circadian rhythmicity of FT3 was found in another study by observing that FT3 peaked 90 minutes after TSH secretion in 86–100% of participants²⁹. In contrast, FT4 did not show a clear circadian rhythm²⁹. Regardless of the effects of sleep, TSH is regulated by circadian rhythms. Starting in the evening, concentrations increase dramatically, peak at 2–3 am, and do not decrease until the afternoon. Thus, although NAHANES self-reported the same sleep duration, the effect of sleep duration on TSH secretion may not be attributable only to duration, but may also vary depending on the individual's sleep onset and termination times. For example, Participant A slept from 9 p.m. to 3 a.m. (when TSH secretion was in the ascending phase), whereas Participant B slept from 2 a.m. to 9 a.m. (when TSH secretion was in the descending phase), and both individuals slept for the same duration of 7 hours; however, the effect of sleep on TSH secretion would differ depending on the period.

Previous observational studies have shown a correlation between hypothyroidism and sleep. A study based on a Chinese population showed that subclinical hypothyroidism was a risk factor for poor sleep quality¹⁸. In addition, the results of another single-center retrospective study showed that patients with hypothyroidism were more likely to have early chronotype³². A large Korean cohort study showed that either too much or too little sleep was associated with an increased risk of subclinical thyroid dysfunction¹⁷. However, some studies have taken the opposite view, suggesting that there is no significant difference in subjective sleep measures between individuals with hyperthyroidism and normal thyroid function³³. Similarly, a study by Akatsu et al. did not find a relationship between subclinical hypothyroidism and sleep quality³⁴. The current evidence on the association between sleep and thyroid disorders is controversial, and the results of our MR study found no correlation between sleep traits (sleep duration, long sleep, short sleep, and insomnia) and thyroid disorders (hyperthyroidism and hypothyroidism).

The association between the HPT axis and sleep can affect thyroid function, and the mutual interference between autoimmunity and sleep duration may further exacerbate the potential impact of sleep on thyroid function^35,36. Although previously understudied in the context of thyroid autoimmune diseases, sleep disorder have been reported to increase the risk of multiple autoimmune diseases such as systemic lupus erythematosus, rheumatoid arthritis, and ankylosing spondylitis^37,38. Some clinical studies have also observed that autoimmune disease activity is associated with sleep^39,40. In addition, sleep is bi-directionally related to the immune system, and sleep can also regulate the immune system³⁵. Therefore, sleep may play a role in the immune mechanism of autoimmune thyroid disease. Our cross-sectional study found that long sleep was positively associated with the risk of TGAb positivity, while sleep disorder were negatively associated with TGAb positivity. This was verified in further MR study results, indicating that long sleep may be positively associated with the risk of GD, while sleep duration may be negatively associated with the risk of HT.

The main strength of this study is that it is the first to combine the NHANES observational study with the MR method to explore the association between sleep traits and thyroid. The strong sample size and thorough adjustment for multiple confounders allowed us to increase the confidence of the study results in the multivariate regression model. Although the cross-sectional study design limits causal inference, the MR analysis essentially produces a natural randomized controlled trial and provides high-power (p < 5×10^− 8) and strongly associated (F statistic > 10) SNP to evaluate the causal association between sleep traits and thyroid disease. However, we still need to consider the limitations of the study. First, although we tried to control various confounders, the data on sleep traits were self-reported, which may be subject to recall bias. Second, the sleep duration measurement only included sleep time on weekdays or workdays and did not include additional information related to sleep (such as shift work, working conditions, and sleep onset and end times), which limited the further analysis of the study results. Finally, our NHANES study focused on the U.S. population, while the MR study focused on people of European ancestry, which limits the generalizability of the study results. It is necessary to validate the current findings in the same ethnic or other ethnic populations in the future.

Our cross-sectional study found that sleep duration was associated with TSH and FT3 levels, and sleep disorder were associated with TT4 levels. In terms of thyroid autoimmunity, long sleep was positively correlated with the risk of TGAb positivity, while sleep disorder were negatively correlated with TGAb positivity. Further through MR research, we speculate that long sleep may be positively correlated with the risk of GD, and sleep duration may be negatively correlated with the risk of HT. In the future, it is necessary to conduct more in-depth basic and clinical research to verify the accuracy of the research results and explore its potential mechanism of action.

Acknowledgments

We sincerely thank all the projects (NHANES, UK Biobank and IEU OPEN GWAS) who participated in this study.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Rongliang Qiu and Jinbo Fu. The first draft of the manuscript was written by Rongliang Qiu and Jinbo Fu, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

No funding was received for this research.

Ethical Approval

Not applicable.

Conflict of interest

All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers’ bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent-licensing arrangements), or non-financial interest (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript.

Data availability statement

The dataset analyzed in this study is publicly available and can be accessed as described below. No new data were generated. Data for observational studies were taken from the NHANES website: https://www.cdc.gov/nchs/nhanes/index.html. GWAS statistical summary data were obtained from UK Biobank and IEU OPEN GWAS, and the corresponding studies can be found in the Supplementary Material: tables 2, 3, 4 and 5.

Brent GA. Mechanisms of thyroid hormone action. J Clin Invest. Sep 2012;122(9):3035-43. doi:10.1172/JCI60047
Jing L, Zhang Q. Intrathyroidal feedforward and feedback network regulating thyroid hormone synthesis and secretion. Front Endocrinol (Lausanne). 2022;13:992883. doi:10.3389/fendo.2022.992883
Brdar D, Gunjaca I, Pleic N, et al. The effect of food groups and nutrients on thyroid hormone levels in healthy individuals. Nutrition. Nov-Dec 2021;91-92:111394. doi:10.1016/j.nut.2021.111394
Ragusa F, Fallahi P, Elia G, et al. Hashimotos' thyroiditis: Epidemiology, pathogenesis, clinic and therapy. Best Pract Res Clin Endocrinol Metab. Dec 2019;33(6):101367. doi:10.1016/j.beem.2019.101367
Smith TJ, Hegedus L. Graves' Disease. N Engl J Med. Oct 20 2016;375(16):1552-1565. doi:10.1056/NEJMra1510030
Babic Leko M, Gunjaca I, Pleic N, Zemunik T. Environmental Factors Affecting Thyroid-Stimulating Hormone and Thyroid Hormone Levels. Int J Mol Sci. Jun 17 2021;22(12)doi:10.3390/ijms22126521
Richards J, Gumz ML. Advances in understanding the peripheral circadian clocks. FASEB J. Sep 2012;26(9):3602-13. doi:10.1096/fj.12-203554
Liu Y, Wheaton AG, Chapman DP, et al. Prevalence of Healthy Sleep Duration among Adults--United States, 2014. MMWR Morb Mortal Wkly Rep. Feb 19 2016;65(6):137-41. doi:10.15585/mmwr.mm6506a1
Covassin N, Singh P. Sleep Duration and Cardiovascular Disease Risk: Epidemiologic and Experimental Evidence. Sleep Med Clin. Mar 2016;11(1):81-9. doi:10.1016/j.jsmc.2015.10.007
Bin YS, Marshall NS, Glozier N. Secular trends in adult sleep duration: a systematic review. Sleep Med Rev. Jun 2012;16(3):223-30. doi:10.1016/j.smrv.2011.07.003
Smiley A, King D, Bidulescu A. The Association between Sleep Duration and Metabolic Syndrome: The NHANES 2013/2014. Nutrients. Oct 26 2019;11(11)doi:10.3390/nu11112582
Chen Y, Tan F, Wei L, et al. Sleep duration and the risk of cancer: a systematic review and meta-analysis including dose-response relationship. BMC Cancer. Nov 21 2018;18(1):1149. doi:10.1186/s12885-018-5025-y
He Q, Sun H, Wu X, et al. Sleep duration and risk of stroke: a dose-response meta-analysis of prospective cohort studies. Sleep Med. Apr 2017;32:66-74. doi:10.1016/j.sleep.2016.12.012
Anothaisintawee T, Reutrakul S, Van Cauter E, Thakkinstian A. Sleep disturbances compared to traditional risk factors for diabetes development: Systematic review and meta-analysis. Sleep Med Rev. Dec 2016;30:11-24. doi:10.1016/j.smrv.2015.10.002
Hui SK, Grandner MA. Trouble Sleeping Associated With Lower Work Performance and Greater Health Care Costs: Longitudinal Data From Kansas State Employee Wellness Program. J Occup Environ Med. Oct 2015;57(10):1031-8. doi:10.1097/JOM.0000000000000534
Dowd JB, Goldman N, Weinstein M. Sleep duration, sleep quality, and biomarkers of inflammation in a Taiwanese population. Ann Epidemiol. Nov 2011;21(11):799-806. doi:10.1016/j.annepidem.2011.07.004
Kim W, Lee J, Ha J, et al. Association between Sleep Duration and Subclinical Thyroid Dysfunction Based on Nationally Representative Data. J Clin Med. Nov 18 2019;8(11)doi:10.3390/jcm8112010
Song L, Lei J, Jiang K, et al. The Association Between Subclinical Hypothyroidism and Sleep Quality: A Population-Based Study. Risk Manag Healthc Policy. 2019;12:369-374. doi:10.2147/RMHP.S234552
Kessler L, Nedeltcheva A, Imperial J, et al. Changes in serum TSH and free T4 during human sleep restriction. Sleep. Aug 2010;33(8):1115-8. doi:10.1093/sleep/33.8.1115
Hirshkowitz M, Whiton K, Albert SM, et al. National Sleep Foundation's sleep time duration recommendations: methodology and results summary. Sleep Health. Mar 2015;1(1):40-43. doi:10.1016/j.sleh.2014.12.010
Dashti HS, Jones SE, Wood AR, et al. Genome-wide association study identifies genetic loci for self-reported habitual sleep duration supported by accelerometer-derived estimates. Nat Commun. Mar 7 2019;10(1):1100. doi:10.1038/s41467-019-08917-4
Lane JM, Jones SE, Dashti HS, et al. Biological and clinical insights from genetics of insomnia symptoms. Nat Genet. Mar 2019;51(3):387-393. doi:10.1038/s41588-019-0361-7
Sakaue S, Kanai M, Tanigawa Y, et al. A cross-population atlas of genetic associations for 220 human phenotypes. Nat Genet. Oct 2021;53(10):1415-1424. doi:10.1038/s41588-021-00931-x
Ikegami K, Refetoff S, Van Cauter E, Yoshimura T. Interconnection between circadian clocks and thyroid function. Nat Rev Endocrinol. Oct 2019;15(10):590-600. doi:10.1038/s41574-019-0237-z
Romijn JA, Adriaanse R, Brabant G, et al. Pulsatile secretion of thyrotropin during fasting: a decrease of thyrotropin pulse amplitude. J Clin Endocrinol Metab. Jun 1990;70(6):1631-6. doi:10.1210/jcem-70-6-1631
Ehrenkranz J, Bach PR, Snow GL, et al. Circadian and Circannual Rhythms in Thyroid Hormones: Determining the TSH and Free T4 Reference Intervals Based Upon Time of Day, Age, and Sex. Thyroid. Aug 2015;25(8):954-61. doi:10.1089/thy.2014.0589
Roelfsema F, Veldhuis JD. Thyrotropin secretion patterns in health and disease. Endocr Rev. Oct 2013;34(5):619-57. doi:10.1210/er.2012-1076
Roelfsema F, Pereira AM, Veldhuis JD, et al. Thyrotropin secretion profiles are not different in men and women. J Clin Endocrinol Metab. Oct 2009;94(10):3964-7. doi:10.1210/jc.2009-1155
Russell W, Harrison RF, Smith N, et al. Free triiodothyronine has a distinct circadian rhythm that is delayed but parallels thyrotropin levels. J Clin Endocrinol Metab. Jun 2008;93(6):2300-6. doi:10.1210/jc.2007-2674
Kettner NM, Katchy CA, Fu L. Circadian gene variants in cancer. Ann Med. Jun 2014;46(4):208-20. doi:10.3109/07853890.2014.914808
Wang M, Lu X, Zheng X, et al. The relationship between sleep duration and thyroid function in the adult US population: NHANES 2007-2012. PLoS One. 2023;18(9):e0291799. doi:10.1371/journal.pone.0291799
Arosemena MA, Ramos AR, Marcus EN, et al. Primary hypothyroidism and chronotypes in adult women. BMC Res Notes. Feb 14 2022;15(1):52. doi:10.1186/s13104-022-05934-3
Grabe HJ, Volzke H, Ludemann J, et al. Mental and physical complaints in thyroid disorders in the general population. Acta Psychiatr Scand. Oct 2005;112(4):286-93. doi:10.1111/j.1600-0447.2005.00586.x
Akatsu H, Ewing SK, Stefanick ML, et al. Association Between Thyroid Function and Objective and Subjective Sleep Quality in Older Men: The Osteoporotic Fractures in Men (MrOS) Study. Endocr Pract. Jun 2014;20(6):576-86. doi:10.4158/EP13282.OR
Besedovsky L, Lange T, Haack M. The Sleep-Immune Crosstalk in Health and Disease. Physiol Rev. Jul 1 2019;99(3):1325-1380. doi:10.1152/physrev.00010.2018
Irwin MR, Olmstead R, Carroll JE. Sleep Disturbance, Sleep Duration, and Inflammation: A Systematic Review and Meta-Analysis of Cohort Studies and Experimental Sleep Deprivation. Biol Psychiatry. Jul 1 2016;80(1):40-52. doi:10.1016/j.biopsych.2015.05.014
Cervilla O, Miro E, Martinez MP, et al. Sleep quality and clinical and psychological manifestations in women with mild systemic lupus erythematosus activity compared to women with fibromyalgia: A preliminary study. Mod Rheumatol. Nov 2020;30(6):1016-1024. doi:10.1080/14397595.2019.1679973
Hsiao YH, Chen YT, Tseng CM, et al. Sleep disorders and increased risk of autoimmune diseases in individuals without sleep apnea. Sleep. Apr 1 2015;38(4):581-6. doi:10.5665/sleep.4574
Young KA, Munroe ME, Harley JB, et al. Less than 7 hours of sleep per night is associated with transitioning to systemic lupus erythematosus. Lupus. Aug 2018;27(9):1524-1531. doi:10.1177/0961203318778368
Son CN, Choi G, Lee SY, et al. Sleep quality in rheumatoid arthritis, and its association with disease activity in a Korean population. Korean J Intern Med. May 2015;30(3):384-90. doi:10.3904/kjim.2015.30.3.384

The authors declare no competing interests.

SupplementaryMaterial.xlsx

Download PDF

Version 1

posted

You are reading this latest preprint version

Sleep traits and thyroid gland: results from National Health and Nutrition Examination Survey 2007-2012 and Mendelian randomization analyses

Status:

Version 1

Abstract

Figures

1. INTRODUCTION

2. MATERIAL AND METHODS

2.1 Study design

2.2 Epidemiological observation study

2.2.1 Data source

2.2.2 Measurements and definitions

2.2.3 Covariates

2.3 Mendelian randomization analysis

2.3.1 Data source

2.3.2 Selection of instrument variables

2.4 Statistical Analyses

2.4.1 Cross-Sectional study

2.4.2 Two-Sample mendelian randomization

3. RESULTS

3.1 Baseline characteristics

3.2 Relationship between sleep traits and thyroid

3.3 Mendelian randomization analysis

4. DISCUSSION

5. CONCLUSION

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1