https://orcid.org/0000-0001-7409-9099
Figures
Abstract
Background
Oxidative stress is a significant contributor to the development of various diseases, and the oxidative balance score (OBS) is a valuable tool for assessing the impact of dietary and lifestyle factors on oxidative stress in humans. Nevertheless, the precise relationship between OBS and thyroid function in adults remains elusive.
Methods
This cross-sectional study comprised 6222 adult participants drawn from the National Health and Nutrition Examination Survey (NHANES) conducted from 2007 to 2012. Employing weighted multivariable linear regression modeling, the study estimated the connection between OBS quartiles and thyroid functions. The causal relationship between OBS components and thyroid function was analyzed by Mendelian randomization (MR).
Citation: Song L, Zhou H, Yang Q, He N, Fu F, Li W, et al. (2024) Association between the oxidative balance score and thyroid function: Results from the NHANES 2007–2012 and Mendelian randomization study. PLoS ONE 19(3): e0298860. https://doi.org/10.1371/journal.pone.0298860
Editor: Mukhtiar Baig, King Abdulaziz University Faculty of Medicine, SAUDI ARABIA
Received: November 7, 2023; Accepted: February 1, 2024; Published: March 18, 2024
Copyright: © 2024 Song et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The datasets generated and/or analyzed during the current study are available in the NHANES (National Health and Nutrition Examination Survey) database. The specific datasets utilized for this study can be accessed at the following links: 1) NHANES 2007-2008: https://wwwn.cdc.gov/Nchs/Nhanes/2007-2008/THYROD_E.htm 2) NHANES 2009-2010: https://wwwn.cdc.gov/Nchs/Nhanes/2009-2010/THYROD_F.htm.
Funding: This study was supported by the “Four Batch” Shanxi Medical Science and Technology Innovation Team Construction Project (Clinical Diagnosis and Treatment and Basic Research Innovation Team of Thyroid Cancer) (2020TD16) and the “Youth Science Research Project”: Targeted modification of hUCMSCs-derived engineered extracellular vesicles induces p53 accumulation and inhibits the growth and metastasis of thyroid cancer (202203021222371).The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
1. Introduction
The physiological functions of the thyroid gland include promoting the metabolism of major nutrients, regulating growth and development, increasing tissue oxygen consumption, promoting energy metabolism, increasing thermogenesis, and improving basal metabolism [1]. Thyroid dysfunction is a common clinical disorder that affects approximately 10% of adults in the general population, and it can affect various body systems and lead to the development of other diseases such as atrial fibrillation, coronary artery disease, and stroke [2–4]. However, even fluctuations in thyroid hormones within the clinically recommended range can also be associated with these pathological conditions. Research has demonstrated that elevations in free Thyroxine (FT4) levels, even within the normal range, can lead to heightened vascular resistance, augmented cardiac contractility, elevated heart rate, and increased left ventricular mass. Such increases can potentially raise the risk of premature atrial contractions, thereby becoming a significant risk factor for atrial fibrillation [5]. Conversely, subclinical hypothyroidism (defined by high thyroid-stimulating hormone (TSH) and FT4 within the reference range) can reduce metabolic rate, increase systemic vascular resistance, promote atherosclerosis, and cause changes in endothelial and coagulation functions [6]. Additionally, among young stroke patients with normal thyroid function, heightened levels of anti-thyroperoxidase antibody (TPOAb) were found to be independently associated with intracranial large artery stenosis [7].
Oxidative stress (OS) results in an excessive accumulation of reactive oxygen species (ROS) that can damage the physiological structure of lipids, proteins, DNA, and RNA, affecting the physiological function of tissues and organs [8]. Studies have shown that oxidative stress is implicated in the pathogenesis of several autoimmune disorders, including thyroid diseases. [9, 10]. A diet high in antioxidants, such as vitamins C, D, and E, can effectively prevent OS while smoking, and alcohol consumption, and a high-fat, high-iron intake diet can promote OS, accelerating ROS accumulation and cellular damage [11].
However, when evaluating the effects of individual antioxidants or pro-oxidants on thyroid function, it is simple to ignore the interactions of multiple antioxidants and pro-oxidants about each other. The oxidation balance score (OBS) is a comprehensive indicator derived from information on diet and other lifestyle factors [12]. An increase in OBS usually indicates an increase in antioxidant exposure [13]. Several studies have investigated the relationship between OBS and various chronic diseases and found a negative association between OBS and a variety of diseases, including type 2 diabetes [14], cardiovascular disease [15], and various types of cancer [16–20].
Although an imbalance between free radical production and antioxidant defense is associated with thyroid dysfunction, there is currently no association detected between OBS and Thyroid Function. Identifying the association between OBS and thyroid function could provide important data for future patient education and therapeutic strategies.This study aims to evaluate the cross-sectional connection between OBS and thyroid function using data from the National Health and Nutrition Examination Survey (NHANES) from 2007 to 2012 MR was used for further study of the relationship between the two.
2. Materials and methods
2.1 Study population in NHANES
The National Health and Nutrition Examination Survey (NHANES) is a comprehensive cross-sectional survey that is managed by the National Center for Health Statistics (NCHS) to assess the health and nutritional status of non-institutionalized American nationals. The NCHS assesses the health and nutritional status of non-institutionalized American nationals by randomly selecting a sample of approximately 5,000 individuals on a two-year cycle through a multistage probability sampling design. The research protocol of NHANES was approved by the National Center for Health Statistics Research Ethics Review Board and written informed consent was obtained. The comprehensive thyroid function data in the NHANES database are available for the years 2007–2012 [21]. In this study, we selected the NHANES survey cycle of 2007 to 2012 and excluded the underage sample (age<18) and the data inclusion exclusion criteria are shown in Fig 1.
OBS, oxidative balance score.
2.2 Data collection in NHANES
In this research, OBS was employed as an exposure factor. The NHANES 2007–2012 was furnished with dietary consumption data through two 24-hour dietary recall interviews (24HR) conducted at the Mobile Examination Center. The first dietary data were collected personally in the mobile inspection center, and the second interview was conducted by telephone consultation after 3 to 10 days.The construct validity and computation procedures of OBS have been previously documented [11]. Comprising 20 items, the OBS includes both dietary and lifestyle components. The dietary items investigated were Dietary fiber, Carotene, Riboflavin, Niacin, Total folate, Calcium, Magnesium, Zinc, Copper, Selenium, Total fat, Iron, Vitamin B6, B12, C, and E. The OBS lifestyle components encompassed Physical activity, Alcohol, Body mass index, and Cotinine. To calculate the OBS score for each individual, the corresponding score is assigned to each component, and then the scores for each component are summed for each participant. The score allocation scheme for each component is listed in S1 Table. Higher OBS scores were indicative of greater antioxidant potential. In this study, OBS has been deemed a continuous variable and was categorized into quartiles, with the highest quartile (Q4) representing the highest antioxidant potential.
The study’s outcome variable was thyroid function, which was obtained from laboratory data in NHANES 2007–2012. Thyroid parameters, including total and free Thyroxine (T4), total and free Triiodothyronine (T3), thyroglobulin (Tg), thyroglobulin antibodies (TgAb), thyroid peroxidase antibodies (TPOAb), and thyroid-stimulating hormone (TSH), were assessed. The 2007–2012 NHANES Laboratory/Medical Technologists Procedures Manual (LPM) outlines the collection and processing of these thyroid function parameters, and previous studies provide detailed descriptions of sample processing [22]. To account for the influence of iodine intake on thyroid function, UIC was categorized into iodine deficient (<100), normal (100–299), and excessive iodine intake (≥300μg/L) for participants over 18 years old, as determined by a previous study [22].
To control for potential confounding factors, several covariates were included in the analysis. These covariates were age (in years), ethnicity, sex, educational level, poverty income ratio (PIR), smoking status, urine iodine concentration (UIC), and dietary energy intake (in kcal). Ethnicity was stratified into non-Hispanic white, non-Hispanic black, Mexican American, and others. PIR was estimated using poverty guidelines for family size and was categorized as ≤1.3, 1.3–3.5, and >3.5 to reflect the socioeconomic status of participants. Education levels were graded into five groups: less than 9th grade, 9th-11th grade, high school grade/general equivalent diploma, some college or associate degree, and college graduate or above. As recommended by NHANES, smoking status was categorized into ‘Never’, ‘Now’, and ‘Former’. These covariates were included in the regression models as potential confounding variables to adjust for their effects on the association between OBS and thyroid function.
2.3 Assumptions and sources of two-sample MR
Most of the GWAS summary data about OBS was obtained from the IEU GWAS database (https://gwas.mrcieu.ac.uk/), while a portion of the GWAS data was sourced from the UK biobank database (http://www.nealelab.is/uk-biobank). We selected only Thyroid Function parameters that exhibited cross-sectional association with OBS as outcome variables and obtained the outcome data from the Thyroidomics Consortium database (https://transfer.sysepi.medizin.uni-greifswald.de/thyroidomics/). Single nucleotide polymorphisms (SNPs) were selected as instrumental variables based on the following criteria: a GWAS-correlated P-value of 5×10−8 or less, an LD r2 of less than 0.001, and location less than 1 megabase (MB) from the index variant. However, if the use of a P-value of 5×10−8 for SNP selection yields a limited number of variants, an alternative strategy may involve using less stringent criteria (P-value of 1×10−5, and F-value > 10) to select SNPs. For IVs that were not present in the outcome, we searched for proxy SNPs in online databases using LD with an r2 greater than 0.8.
2.4 Statistical analysis
The statistical analysis in this study was conducted in accordance with CDC guidelines for analyzing complex NHANES data. To account for masked variance, sample weights were assigned to each participant. Continuous variables were reported as mean (SE), while categorical variables were presented as percentages. Multivariable linear regression was used to examine the relationship between OBS and thyroid function, with Model 1 being unadjusted, Model 2 adjusting for age, ethnicity, and sex, and Model 3 being the fully adjusted model. Interaction and stratified analyses were also conducted to explore effect modification by age, sex, and UIC.
For MR analysis, we used the R package "TwoSampleMR". The causal effect was determined using the Wald ratio if one SNP was present in the sample. When the number of SNPs was greater than two, we mostly utilized inverse variance weighting meta-analysis (IVW) for estimating purposes. IVW estimates are the weighted average of the Wald ratios obtained from each SNP. When using the IVW approach, included SNPs cannot have multiple levels of validity to prevent the creation of biased results. MR-Egger is similar to IVW, but it includes an intercept term, allowing for the presence of pleiotropy. Additionally, MR-Egger requires a lower IV concentration than IVW, allowing for the existence of the InSIDE assumption that pleiotropic effects are independent of the correlation between risk variables. The MR-Egger intercept is the average pleiotropic effect of genetic variation, according to the InSIDE hypothesis. Therefore, when pleiotropy was accounted for, MR-Egger estimates were consistent with IVW. The approach of weighted median estimate (WME) can eliminate the effect of erroneous instrumental variables. WME analysis can yield reliable estimates of causal effects even if 50% of the genetic variance information is erroneous. Statistical power analyses were performed with the online mRnd calculator (https://shiny.cnsgenomics.com/mRnd/). The MR Steiger test is used to test the direction of causality in our hypothesis. To detect any outliers and horizontal pleiotropy, we conducted MR Pleiotropy RESidual Sum and Outlier (MRPRESSO), tested directional pleiotropy with MR-Egger intercept, and detected potential heterogeneity with the Cochrane Q test. Radial MR imaging was used to find outliers. Multivariate Mendelian randomization (MVMR) is an extension of conventional univariable MR, which can remove the influence of confounders on causality to obtain a direct effect of exposure on the outcome. Therefore, we performed MVMR to obtain causal estimates adjusted for these genetic correlations.
We utilized R (version 4.1.3) software for all statistical analyses. A p-value of less than 0.05 was considered statistically significant.
3. Results
3.1 Baseline characteristics
A total of 1,1638 subjects were recruited from the National Health and Nutrition Examination Survey (NHANES) during the year 2007–2012. After applying the inclusion and exclusion criteria, our final sample included 6222 adult participants with an average age and standard error (SE) of 46.58(0.42) and OBS of 21.21(0.28). The sociodemographic weighted characteristics of the study subjects are concisely presented in Table 1 and arranged into quartiles based on their OBS scores. Mean OBS scores ranged from 11.33 (most pro-oxidant) to 29.83 (most anti-oxidant), whereas the four OBS quartile ranges were 4~15, 15~21, 21~26, and 26~36, respectively. Among the four OBS quartiles, for thyroid parameters, differences in TT3 and TT4 were statistically significant. For covariates, except for sex and UIC, the distribution of age, ethnicity, educational level, smoking status, mean dietary energy intake, and PIR were statistically different in 4 groups. Age was similar between the OBS quartile groups. In comparison to OBS Q 1 to 3, individuals in the OBS Q4 group had a greater tendency to be non-Hispanic White (77.98(2.28)). Participants belonging to the OBS Q4 group were observed to exhibit a tendency towards increased energy intake, higher educational attainment, decreased smoking frequency, and higher incomes. According to the analysis, participants belonging to the Q4 group had a statistically significant lower level of TT4 when compared to the other groups. The TT3 mean was recorded as 113.27(0.77) mIU/L, and participants in the Q4 group had higher-than-average TT3 levels. No statistically significant differences were detected among the OBS quartiles in other parameters related to thyroid function (all p-values were greater than 0.05).
3.2 OBS score and thyroid function
We carried out an analysis utilizing multivariable linear regression to explore the correlation between the OBS score and thyroid function, as depicted in Table 2. When the OBS score was considered as a continuous variable, our results showed a statistically significant negative correlation between a higher OBS score and lower FT4 and TT4 levels in crude Model 1. Precisely, each incremental unit increase in the OBS score resulted in a corresponding decrease in the TT4 level by 0.02 ng/dl and the FT4 decreased by 0.01 ng / dl. After grouping the OBS score quartiles, this association was still statistically significant. In model 2, our findings indicated a negative correlation between the quartile range of OBS scores and FT3 levels, which, however, ceased to exist after additional adjustment for confounding factors. Furthermore, there was no statistically significant correlation between OBS and other thyroid function parameters (p > 0.05).
We stratified our analysis by age, sex, and UIC to account for the effects of age, sex, and iodine intake on thyroid function.We found a significant negative association between TT4 and FT4 with OBS scores in both female and male groups. Additionally, a strong negative relationship between OBS and TSH and TPOAb was identified in the male group. Specifically, TPOAb decreased by 0.74 IU/mL for every unit increase in OBS; participants in Q4 exhibited a mean decrease of 0.18 mIU/L in TSH compared with those in Q1 (S2 Table). In the excessive iodine intake group, the negative correlation between FT4 and OBS scores disappeared. Unexpectedly, in the normal iodine intake group, a more apparent negative correlation was observed between TT4 and OBS (p for interaction = 0.003) (S3 Table). We categorized age into two groups: <45 years and ≥45 years, and observed a significant negative correlation between both TT4 and FT4 levels and OBS scores. This negative correlation was notably stronger in the age group <45 (p for interaction = 0.044) (Table 3). Furthermore, in the age group ≥45, we noted a significant positive correlation between the level of TT3 and OBS, with statistically significant between-group heterogeneity (p for interaction < 0.001).Our findings suggest that higher OBS scores (antioxidant) may lead to lower free and total T4 levels. After subgroup analysis, the negative association between OBS and total and free T4 appeared to be stronger in the age group <45. We also observed a significant negative correlation between TSH and TPOAb levels and OBS in the male groups, as well as a significant negative correlation between TT4 in the normal iodine intake group and OBS scores.
3.3 The results of MR analysis
After conducting a correlation analysis of the cross-sectional study, we performed an MR analysis of the parameters of Thyroid Function associated with OBS. The figure illustrates the screening process of instrumental variables (IVs) and the flow chart of MR analysis (Fig 2A). As public GWAS data for TT4, dietary fiber, niacin, and total folate have not been published, we only carried out MR analysis for the remaining OBS items and FT4. Table 4 presents detailed information on included single-nucleotide polymorphisms (SNPs).
(a) The flow chart describing the methods and basic flow of MR analysis applied in this study. (b) Scatter plots for MR analyses of the causal effect of BMI on Free T4. (c) Funnel plot of MR analyses of the causal effect of BMI on Free T4.
According to the IVW analysis results, we unexpectedly found a significant causal relationship of BMI with FT4 (β = 0.13, 95% CI = 0.07 ~ 0.18, P < 0.001) (Fig 2B) and no horizontal pleiotropy (intercept of MR-Egger regression = −0.0017, P = 0.204), but there was significant heterogeneity, and MRPRESSO suggested outliers. After excluding the outlier reported by radial MR and MRPRESSO, this result no longer shows heterogeneity (Q statistics = 310, P = 1) (Fig 2C), and no SNP was observed in the leave-one-out plot (S1 Fig) that could significantly affect the overall results. This result removed rs10747488, rs11496125, rs12299814, rs12328930, rs12593036, rs12981256, rs13021737, rs1421334, rs1477199, rs17056301, rs17405819 rs17724992, rs2065418, rs273504, rs2820311, rs331966, rs4072917, rs4639527, rs4757144, rs6985109, rs7334078, rs769449, and rs784944 (S4 Table). According to the PhenoScanner database these SNPs were considered to be associated with confounding, such as schizophrenia [23], depression [24], and systolic and diastolic blood pressure [25]. MR analysis of the other items did not show a causal relationship with freeT4 (Fig 3). In sensitivity analysis, Cochran’s Q statistic indicated significant heterogeneity among SNPs for Vitamin E, Zinc, and Alcohol. Moreover, the intercept of the MR-Egger regression in Body mass index and Vitamin B6 were significantly different from zero, indicating a horizontal pleiotropy of IV in these items. Therefore, we chose MR-Egger as the method of causal effect evaluation in these items. Unfortunately, the MR-Egger results still failed to reveal any significant causal relationship. It is worth noting that all of these IVs have F-values between 20 and 125 and, thus, are largely unaffected by the bias generated by weak instrumental variables. All results of the MR Steiger directionality test illustrated the accuracy of our estimate of causal direction.
The MVMR results showed that BMI remained significantly causally related to FT4 when we included all 17 OBS items in the analysis. Considering that the large number of included exposures is prone to collinearity, we used LASSO analysis to remove the collinearity exposures, and unfortunately they were all collinearity with each other in different degrees. Therefore, we classified them into pro-oxidant and antioxidant groups for MVMR analysis based on their different oxidation. For the pro-oxidant group, the LASSO analysis left BMI, A, and B. After MVMR analysis, we can observe that BMI remains causally related to FT4 (IVW: β = -0.03, 95% CI = 0.02 ~ 0.16, P = 0.02) and smoking also has a causal relationship with free T4 (IVW: β = -0.10, 95% CI = -0.20 ~ -0.00, P = 0.04). The antioxidant group retained Riboflavin, Copper, and C by LASSO analysis, while the results of MVMR were surprising, showing significant causality of Riboflavin (IVW: β = -0.03, 95% CI = -0.10 ~ -0.01, P = 0.01) and Copper (IVW: β = -0.02, 95% CI = -0.04 ~ -0.00, P = 0.02) with free T4 by adjustment.
4. Discussion
Our study aimed to evaluate the connection between OBS and thyroid function with the aid of a multivariable linear regression model. After dividing the participants into four OBS quartiles, we compared the distribution of the different covariates. The findings revealed that non-Hispanic whites, the wealthy, never smokers, and highly educated people in the quartile group Q1–Q4 received higher OBS scores, indicating a greater exposure advantage for antioxidants in their diets. This may be related to the fact that this segment of the population pays more attention to physical health and maintaining good living habits. In the meantime, we discovered a statistically significant negative correlation between OBS and TT3, and TT4. However, when the model was further adjusted for potential confounders, the correlation between TT3 and OBS was no longer significant but the negative correlations between freeT4, totalT4, and OBS were significantly exposed. Similar conclusions to ours have been reported in previous research, where higher dietary selenium (an antioxidant) intake was associated with a decline in serum T4 concentrations but not with a corresponding change in TSH levels [26]. The potential mechanism is still unclear and needs to be proven by further in vitro and in vivo experiments.
Considering the effects of age, gender, and iodine intake on thyroid function, we performed a subgroup analysis for these three variables. We found that the negative correlation between OBS and TSH and TPOAb levels were more significant in men, which may be related to the protective effect of estrogen against oxidative stress [27, 28]. When subgroup analysis was performed based on UIC, we found that OBS was negatively associated with TT4 and FT4 within normal iodine intake in adults. However, in the groups with high iodine intake, this difference lost statistical significance. Previous studies illustrated that iodine has antioxidant properties and can increase the antioxidant capacity of human serum [29], but the effect of iodine on thyroid function may counteract this effect. In the age subgroup analysis, TT4 and FT4 levels were significantly negatively correlated with OBS in both the young and middle-aged as well as the elderly individuals, with the younger and middle-aged groups (age < 45) showing a greater sensitivity to antioxidant diets. This could be attributed to the heightened glutathione peroxidase activity in older cells, rendering them more resistant to oxidative stress compared to younger cells [30]. In the elderly population, we found a significant positive correlation between TT3 level and OBS score, which may be due to the increase of deiodinase activity and the conversion of T4 to T3 caused by the intake of antioxidants against oxidative stress [31]. At the same time, studies have shown that higher T3 levels are associated with longevity [32]. Several studies have demonstrated that higher serum-FT4 levels are considered a risk factor for the development of cardiovascular disease and death within the range of clinical recommendations [33, 34]. By contrast, relatively low levels of FT4 were conducive to delaying aging by reducing metabolic demand, and normal subjects with lower FT4 levels lived 3.7 years longer on average than normal subjects with higher FT4 levels [35, 36]. Therefore, we recommend augmenting the intake of antioxidants in daily life.
After MR analysis between the 16 entries of OBS and FT4, we further found a significant positive causal relationship between BMI with FT4, and this result remained significant under the adjustment of MVMR, and this is consistent with the results of our cross-sectional study. The results of the cross-sectional study showed that FT4 levels were negatively correlated with OBS, higher antioxidant exposure was associated with lower serum FT4 levels, while BMI as a prooxidant, increased BMI was associated with higher serum FT4 levels. Our results of MR are more clear about their relationship, and the robustness of these results gives us reason for believing that BMI is a key factor influencing FT4 levels. It is worth noting that the relationship between BMI and serum FT4 remains a topic of debate, with some studies reporting a negative correlation between the two variables. This phenomenon can be explained by two factors: Firstly, hypothyroidism is known to decrease the body’s metabolic rate and often coexists with obesity [37]. Secondly, increased leptin levels in obesity can influence the hypothalamus-pituitary-thyroid (HPT) axis and consequently the endocrine system. Studies have shown that leptin levels and TSH levels are positively correlated and that TSH can increase the level of deiodinase, leading to the preferential production of T3 by thyroid cells, resulting in an increase in serum FT3 levels, but not FT4. This may account for the decrease in serum FT4 levels observed in individuals with higher BMI [38–42]. However, another study reported a positive association between BMI and FT4.They showed that morbidly obese subjects had higher levels of TT3, FT3, TT4, and TSH than those of the control group, probably as a result of the reset of their central thermostat at higher levels [43]. In addition,individuals with higher BMI often experience increased oxidative stress in accumulated fat, leading to dysregulation of adipocytokine production and selective increases in ROS, which can result in elevated systemic oxidative stress. Excessive production of ROS can cause hyperthyroidism, often accompanied by low TSH, high FT3, and FT4 [44]. Many articles studying smoking usually use cotinine as a surrogate quantitative measure to illustrate the strong correlation between the two [45, 46]. In the present study, smoking data were used as a proxy because GWAS data for cotinine were not available. After MVMR modulation, smoking, and freeT4 produced a causal effect. Smoking can affect levels of perfluoroalkyl acids (PFAS) in humans, and PFAS is now thought to cause effects on thyroid function, with a positive correlation to FreeT4 levels [47].
Riboflavin and Copper showed significant causal effects after elimination for confounders. Copper has been previously suggested as a microelement for maintaining normal thyroid function, stimulating the synthesis of thyroid hormones, and leading to lower levels of FT4 in the body if it is excessively deficient [48]. Riboflavin, also known as vitamin B2, is a micronutrient that has an important impact on thyroid function, as evidenced by the fact that patients with low vitB2 Patients with lower levels of vitamin B2 usually have lower serum T4 [49]. Our results reveal the causal relationship between them and suggest that they are all key factors influencing freeT4 levels in humans. Therefore, we suggest that normal freeT4 levels can be maintained by improvement of BMI and moderate consumption of Riboflavin and Copper.
This is the first to assess the effects of dietary pro-oxidant and antioxidant protective exposures on thyroid function using a nationally representative U.S. general population. In parallel, we acknowledged that it has some restrictions. (1) Due to poor adherence to the questionnaire among younger individuals, our study only included adults aged ≥ 18. Additionally, pregnant individuals were excluded from the study due to potential dietary abnormalities during pregnancy, so the generalizability of our findings to this population is uncertain (2) Since some of the variables were obtained from the questionnaire, this would lead to an inevitable recall bias in the information we obtained. (3) Due to the lack of clinical research and basic experiments, currently, it is difficult for us to explain the intrinsic mechanism associated with OBS and fluctuations in thyroid hormone levels. (4) Individual-level GWAS data are deficient for further testing of nonlinear causality. (5) The GWAS data used in this study was limited to European populations, and thus, our results may not apply to other populations (6) The study of OBS items with FT4 had a lower p threshold due to a paucity of IVs. While our F-statistic tests revealed no significant risk of weak instrument bias, this negative result should be interpreted with caution.
5. Conclusions
In conclusion, our study illustrates that a diet and lifestyle rich in antioxidants is correlated with decreased levels of free and total T4.Therefore, boosting antioxidant intake may serve as a cost-effective measure in preventing cardiovascular and cerebrovascular diseases and in slowing down the aging process, particularly among young and middle-aged individuals. Based on further investigation by MR, we suggest that BMI, Copper, and Riboflavin are important factors affecting freeT4, which can help us to better regulate freeT4 to healthy levels.
Supporting information
S1 Fig. Forest plots of MR leave-one-out sensitivity results (IVW method) for the causal effect of BMI on Free T4.
https://doi.org/10.1371/journal.pone.0298860.s001
(TIF)
S1 Table. Oxidative balance score assignment scheme.
https://doi.org/10.1371/journal.pone.0298860.s002
(DOCX)
S2 Table. The table of the relationship between OBS and thyroid function within gender subgroups.
https://doi.org/10.1371/journal.pone.0298860.s003
(DOCX)
S3 Table. The table of the relationship between OBS and thyroid function within UIC subgroups.
https://doi.org/10.1371/journal.pone.0298860.s004
(DOCX)
S4 Table. Details of the SNPs that were finally included in the MR analysis.
https://doi.org/10.1371/journal.pone.0298860.s005
(CSV)
References
- 1. van den Beld AW, Kaufman JM, Zillikens MC, Lamberts SWJ, Egan JM, van der Lely AJ. The physiology of endocrine systems with ageing. Lancet Diabetes Endocrinol. 2018;6(8):647–58. Epub 2018/07/19. pmid:30017799; PubMed Central PMCID: PMC6089223.
- 2. Cooper DS, Biondi B. Subclinical thyroid disease. Lancet. 2012;379(9821):1142–54. Epub 2012/01/26. pmid:22273398.
- 3. Chaker L, Heeringa J, Dehghan A, Medici M, Visser WE, Baumgartner C, et al. Normal Thyroid Function and the Risk of Atrial Fibrillation: the Rotterdam Study. J Clin Endocrinol Metab. 2015;100(10):3718–24. Epub 2015/08/12. pmid:26262438.
- 4. Chaker L, Baumgartner C, den Elzen WP, Ikram MA, Blum MR, Collet TH, et al. Subclinical Hypothyroidism and the Risk of Stroke Events and Fatal Stroke: An Individual Participant Data Analysis. J Clin Endocrinol Metab. 2015;100(6):2181–91. Epub 2015/04/10. pmid:25856213; PubMed Central PMCID: PMC4454799.
- 5. Baumgartner C, da Costa BR, Collet TH, Feller M, Floriani C, Bauer DC, et al. Thyroid Function Within the Normal Range, Subclinical Hypothyroidism, and the Risk of Atrial Fibrillation. Circulation. 2017;136(22):2100–16. Epub 2017/10/25. pmid:29061566; PubMed Central PMCID: PMC5705446.
- 6. Rodondi N, den Elzen WP, Bauer DC, Cappola AR, Razvi S, Walsh JP, et al. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA. 2010;304(12):1365–74. Epub 2010/09/23. pmid:20858880; PubMed Central PMCID: PMC3923470.
- 7. Shi Z, Zhang X, Chen Z, Liebeskind DS, Lou M. Elevated thyroid autoantibodies and intracranial stenosis in stroke at an early age. Int J Stroke. 2014;9(6):735–40. Epub 2013/09/13. pmid:24024979; PubMed Central PMCID: PMC4142769.
- 8. Panov AV, Dikalov SI. Cardiolipin, Perhydroxyl Radicals, and Lipid Peroxidation in Mitochondrial Dysfunctions and Aging. Oxid Med Cell Longev. 2020;2020:1323028. Epub 2020/09/24. pmid:32963690; PubMed Central PMCID: PMC7499269 publication of this paper.
- 9. Abalovich M, Llesuy S, Gutierrez S, Repetto M. Peripheral parameters of oxidative stress in Graves’ disease: the effects of methimazole and 131 iodine treatments. Clin Endocrinol (Oxf). 2003;59(3):321–7. Epub 2003/08/16. pmid:12919155.
- 10. Nanda N, Bobby Z, Hamide A, Koner BC, Sridhar MG. Association between oxidative stress and coronary lipid risk factors in hypothyroid women is independent of body mass index. Metabolism. 2007;56(10):1350–5. Epub 2007/09/22. pmid:17884444.
- 11. Zhang W, Peng SF, Chen L, Chen HM, Cheng XE, Tang YH. Association between the Oxidative Balance Score and Telomere Length from the National Health and Nutrition Examination Survey 1999–2002. Oxid Med Cell Longev. 2022;2022:1345071. Epub 2022/02/22. pmid:35186180; PubMed Central PMCID: PMC8850082.
- 12. Van Hoydonck PG, Temme EH, Schouten EG. A dietary oxidative balance score of vitamin C, beta-carotene and iron intakes and mortality risk in male smoking Belgians. J Nutr. 2002;132(4):756–61. Epub 2002/04/02. pmid:11925473.
- 13. Cho AR, Kwon YJ, Lim HJ, Lee HS, Kim S, Shim JY, et al. Oxidative balance score and serum gamma-glutamyltransferase level among Korean adults: a nationwide population-based study. Eur J Nutr. 2018;57(3):1237–44. Epub 2017/03/05. pmid:28258305.
- 14. Golmohammadi M, Ayremlou P, Zarrin R. Higher oxidative balance score is associated with better glycemic control among Iranian adults with type-2 diabetes. Int J Vitam Nutr Res. 2021;91(1–2):31–9. Epub 2019/06/25. pmid:31230534.
- 15. Ilori TO, Wang X, Huang M, Gutierrez OM, Narayan KM, Goodman M, et al. Oxidative Balance Score and the Risk of End-Stage Renal Disease and Cardiovascular Disease. Am J Nephrol. 2017;45(4):338–45. Epub 2017/03/13. pmid:28285313.
- 16. Park YM, Shivappa N, Petimar J, Hodgson ME, Nichols HB, Steck SE, et al. Dietary inflammatory potential, oxidative balance score, and risk of breast cancer: Findings from the Sister Study. Int J Cancer. 2021;149(3):615–26. Epub 2021/03/31. pmid:33783833; PubMed Central PMCID: PMC8256885.
- 17. Lakkur S, Goodman M, Bostick RM, Citronberg J, McClellan W, Flanders WD, et al. Oxidative balance score and risk for incident prostate cancer in a prospective U.S. cohort study. Ann Epidemiol. 2014;24(6):475–8 e4. Epub 2014/04/16. pmid:24731700; PubMed Central PMCID: PMC4352102.
- 18. Kong SY, Goodman M, Judd S, Bostick RM, Flanders WD, McClellan W. Oxidative balance score as predictor of all-cause, cancer, and noncancer mortality in a biracial US cohort. Ann Epidemiol. 2015;25(4):256–62 e1. Epub 2015/02/17. pmid:25682727; PubMed Central PMCID: PMC4369443.
- 19. Wang T, Goodman M, Sun YV, Thyagarajan B, Gross M, Bostick RM. DNA base excision repair genetic risk scores, oxidative balance, and incident, sporadic colorectal adenoma. Mol Carcinog. 2017;56(6):1642–52. Epub 2017/01/26. pmid:28120344.
- 20. Agalliu I, Kirsh VA, Kreiger N, Soskolne CL, Rohan TE. Oxidative balance score and risk of prostate cancer: results from a case-cohort study. Cancer Epidemiol. 2011;35(4):353–61. Epub 2010/12/15. pmid:21145797.
- 21. F L, K W, J N, Q F, X L, Y Y, et al.—Relationship between dietary selenium intake and serum thyroid function measures. Front Nutr. 2022;9(1002489).
- 22. Liu N, Ma F, Feng Y, Ma X. The Association between the Dietary Inflammatory Index and Thyroid Function in U.S. Adult Males. Nutrients. 2021;13(10). Epub 2021/10/24. pmid:34684331; PubMed Central PMCID: PMC8540204.
- 23. Jose J, Nandeesha H, Kattimani S, Meiyappan K, Sarkar S, Sivasankar D. Association between prolactin and thyroid hormones with severity of psychopathology and suicide risk in drug free male schizophrenia. Clin Chim Acta. 2015;444:78–80. Epub 2015/02/14. pmid:25677864.
- 24. Frye MA, Denicoff KD, Bryan AL, Smith-Jackson EE, Ali SO, Luckenbaugh D, et al. Association between lower serum free T4 and greater mood instability and depression in lithium-maintained bipolar patients. Am J Psychiatry. 1999;156(12):1909–14. Epub 1999/12/10. pmid:10588404.
- 25. Amouzegar A, Heidari M, Gharibzadeh S, Mehran L, Tohidi M, Azizi F. The Association Between Blood Pressure and Normal Range Thyroid Function Tests in a Population Based Tehran Thyroid Study. Horm Metab Res. 2016;48(3):151–6. Epub 2015/12/17. pmid:26671752.
- 26. Megna G, Bandiera L, Federici A, Mastronardi V, Cerfeda S. [Perception of the articular angles in neurological patients with motor disability. I. Physiopathological considerations]. Acta Neurol (Napoli). 1977;32(3):323–31. Epub 1977/05/01.
- 27. Topcuoglu A, Uzun H, Aydin S, Kahraman N, Vehid S, Zeybek G, et al. The effect of hormone replacement therapy on oxidized low density lipoprotein levels and paraoxonase activity in postmenopausal women. Tohoku J Exp Med. 2005;205(1):79–86. Epub 2005/01/07. pmid:15635276.
- 28. Sener G, Arbak S, Kurtaran P, Gedik N, Yegen BC. Estrogen protects the liver and intestines against sepsis-induced injury in rats. J Surg Res. 2005;128(1):70–8. Epub 2005/08/24. pmid:16115495.
- 29. Winkler R, Griebenow S, Wonisch W. Effect of iodide on total antioxidant status of human serum. Cell Biochem Funct. 2000;18(2):143–6. Epub 2000/05/18. pmid:10814974.
- 30.
Matsuo M, Ikeda H Fau—Sugihara T, Sugihara T Fau—Horiike S, Horiike S Fau—Okano Y, Okano Y Fau—Masaki H, Masaki H. Resistance of cultured human skin fibroblasts from old and young donors to oxidative stress and their glutathione peroxidase activity. (0304-324X (Print)).
- 31. A M, Id O, C DS, Id O, S R, G O, et al.—Thyroid Hormones, Oxidative Stress, and Inflammation. Mediators Inflamm. 2016;6757154(10):8.
- 32. S M G B, P C, L B, P S, F F, et al.—Complex alteration of thyroid function in healthy centenarians. J Clin Endocrinol Metab. 1993;77(5):1130–4. pmid:8077303
- 33. Hogervorst E, Huppert F, Matthews FE, Brayne C. Thyroid function and cognitive decline in the MRC Cognitive Function and Ageing Study. Psychoneuroendocrinology. 2008;33(7):1013–22. Epub 2008/07/22. pmid:18640783.
- 34. Simonsick EM, Chia CW, Mammen JS, Egan JM, Ferrucci L. Free Thyroxine and Functional Mobility, Fitness, and Fatigue in Euthyroid Older Men and Women in the Baltimore Longitudinal Study of Aging. J Gerontol A Biol Sci Med Sci. 2016;71(7):961–7. Epub 2016/01/23. pmid:26791089; PubMed Central PMCID: PMC4906324.
- 35. Bano A, Chaker L, Mattace-Raso FUS, Terzikhan N, Kavousi M, Ikram MA, et al. Thyroid function and life expectancy with and without noncommunicable diseases: A population-based study. PLoS Med. 2019;16(10):e1002957. Epub 2019/10/28. pmid:31652264; PubMed Central PMCID: PMC6814213.
- 36. Chaker L, Cappola AR, Mooijaart SP, Peeters RP. Clinical aspects of thyroid function during ageing. Lancet Diabetes Endocrinol. 2018;6(9):733–42. Epub 2018/07/19. pmid:30017801.
- 37. Danforth E Jr, Horton ES, O’Connell M, Sims EA, Burger AG, Ingbar SH et al. Dietary-induced alterations in thyroid hormone metabolism during overnutrition. J Clin Invest. 1979;64(5):1336–47. Epub 1979/11/01. PubMed Central PMCID: PMC371281. pmid:500814
- 38. Harvey J. Leptin: a diverse regulator of neuronal function. J Neurochem. 2007;100(2):307–13. Epub 2006/11/02. pmid:17076761; PubMed Central PMCID: PMC1781059.
- 39. Duntas LH, Biondi B. The interconnections between obesity, thyroid function, and autoimmunity: the multifold role of leptin. Thyroid. 2013;23(6):646–53. Epub 2012/09/01. pmid:22934923.
- 40. Ishii H, Inada M, Tanaka K, Mashio Y, Naito K, Nishikawa M, et al. Induction of outer and inner ring monodeiodinases in human thyroid gland by thyrotropin. J Clin Endocrinol Metab. 1983;57(3):500–5. Epub 1983/09/01. pmid:6308028.
- 41. Beech SG, Walker SW, Arthur JR, Lee D, Beckett GJ. Differential control of type-I iodothyronine deiodinase expression by the activation of the cyclic AMP and phosphoinositol signalling pathways in cultured human thyrocytes. J Mol Endocrinol. 1995;14(2):171–7. Epub 1995/04/01. pmid:7619206.
- 42. Karavani G, Strich D, Edri S, Gillis D. Increases in thyrotropin within the near-normal range are associated with increased triiodothyronine but not increased thyroxine in the pediatric age group. J Clin Endocrinol Metab. 2014;99(8):E1471–5. Epub 2014/06/01. pmid:24878053.
- 43. Michalaki MA, Vagenakis AG, Leonardou AS, Argentou MN, Habeos IG, Makri MG, et al. Thyroid function in humans with morbid obesity. Thyroid. 2006;16(1):73–8. Epub 2006/02/21. pmid:16487017.
- 44. Asayama K, Kato K. Oxidative muscular injury and its relevance to hyperthyroidism. Free Radic Biol Med. 1990;8(3):293–303. Epub 1990/01/01. pmid:2187767.
- 45. Oh S, Kim S, Sung E, Kim CH, Kang JH, Shin H, et al. The association between cotinine-measured smoking intensity and sleep quality. Tob Induc Dis. 2022;20:77. Epub 2022/09/20. pmid:36118556; PubMed Central PMCID: PMC9443078.
- 46. Connor Gorber S, Schofield-Hurwitz S, Hardt J, Levasseur G, Tremblay M. The accuracy of self-reported smoking: a systematic review of the relationship between self-reported and cotinine-assessed smoking status. Nicotine Tob Res. 2009;11(1):12–24. Epub 2009/02/28. pmid:19246437.
- 47. Shapiro JI, Haug CE, Weil R 3rd, Chan L. 31P nuclear magnetic resonance study of acute renal dysfunction in rat kidney transplants. Magn Reson Med. 1987;5(4):346–52. Epub 1987/10/01. pmid:3316920.
- 48. Kim MJ, Kim SC, Chung S, Kim S, Yoon JW, Park YJ. Exploring the role of copper and selenium in the maintenance of normal thyroid function among healthy Koreans. J Trace Elem Med Biol. 2020;61:126558. Epub 2020/06/02. pmid:32480050.
- 49. Krishnamurthy HK, Reddy S, Jayaraman V, Krishna K, Song Q, Rajasekaran KE, et al. Effect of Micronutrients on Thyroid Parameters. J Thyroid Res. 2021;2021:1865483. Epub 2022/02/11. pmid:35140907; PubMed Central PMCID: PMC8820928 Vibrant Sciences LLC. Reddy, Song, and Rajasekaran are employees of Vibrant America LLC.