Sex-specific effect of urinary metabolites of polycyclic aromatic hydrocarbons on thyroid profiles: results from NHANES 2011–2012

The current study aims to evaluate the associations between 10 urinary polycyclic aromatic hydrocarbon (PAH) metabolites and thyroid profiles. The levels of 10 PAH metabolites and thyroid profiles were obtained from National Health and Nutrition Examination Survey (NHANES) 2011–2012. Spearman analysis was utilized to evaluate the correlation coefficients among these 10 PAH metabolites. Multivariate linear and logistic regression models assessed the relationship between urinary PAH metabolite levels, thyroid hormones, and thyroid autoantibodies after adjusting potential confounders. Stratified analysis by gender was performed to evaluate sex-specific effect of urinary metabolites of PAH on thyroid profiles. One thousand six hundred forty-five eligible adult participants with complete research data were enrolled. Of note, the concentrations of the majority of urinary PAH metabolites were remarkedly higher in females compared with males. 2-hydroxyfluorene (2-FLU) was associated with higher total triiodothyronine (T3) levels in whole population (β = 2.113, 95% CI 0.339–3.888). In males, positive associations were observed in 1-hydroxynaphthalene (1-NAP) and free thyroxine (T4) (β = 0.0002, 95% CI 0.0000–0.0004). 2-FLU was also found positively associated with total T3 (β = 2.528, 95% CI 0.115–4.940) in male subjects. While in female participants, 2-hydroxynaphthalene (2-NAP) was associated with free T3 (β = 0.002, 95% CI 0.000–0.005). 2-FLU was associated with total T3 (β = 2.683, 95% CI 0.038–5.328), free T3 (β = 0.050, 95% CI 0.012–0.087), and total T4 (β = 0.195, 95% CI 0.008–0.382). 2-Hydroxyphenanthrene (2-OHP), 1-hydroxypyrene (1-HP), and 9-hydroxyfluorene (9-FLU) were all positively related to total T3 levels, and the corresponding coefficients were 16.504, 6.587, and 3.010. 9-FLU was also associated with free T3 (β = 0.049, 95% CI 0.008–0.090). No statistical significances were found between PAH metabolite levels and increased prevalence of increased thyroglobulin antibody (TgAb)/thyroid peroxidase antibody (TPOAb) when PAH metabolites were treated as continuous variables. Meanwhile, in the quartile analyses, increased prevalence of elevated TgAb was observed in participants with quartile 2 2-NAP compared with lowest quartile (OR = 1.753, 95% CI 1.021–3.008). Male subgroup analyses indicated that increased prevalence of elevated TgAb was observed in higher quartile of 1-NAP, 2-NAP, and 3-hydroxyfluorene (3-FLU). Increased prevalence of elevated TPOAb was associated with higher 2-NAP quartile. However, in subgroup analysis of females, no statistical significances were found between PAH quartiles and increased TgAb/TPOAb. Significant correlations were found among these 10 PAH metabolites. In conclusion, the cross-sectional study indicated that exposure to PAH might disturb the concentrations of thyroid hormones and thyroid autoantibodies. It is noteworthy that significant differences existed in males and females. Further prospective research is warranted to explore the causal relationship and underlying mechanism of PAH exposure on thyroid dysfunction.


Introduction
Accumulated evidence revealed that large numbers of pollutants have contributed to detrimental effect on human health (Ambade 2018;Vithanage et al. 2022). Among them, polycyclic aromatic hydrocarbons (PAHs) are a group of hazardous and persistent organic pollutants with at least 2 fused benzene rings (Kim et al. 2013). PAHs are ubiquitously encountered in ambient environmental compartments, including soil, air, water, and food (Kumar et al. 2020a, b). Incomplete combustion of carbonaceous materials including petroleum, oil, biomass, and coal, as well as biological transformation of organic substances, is a predominant anthropogenic and natural source of PAHs (Ambade et al. 2022a, b;Ambade et al. 2022a, b;Kumar et al. 2020a, b). China is a country with large amount of PAH emissions; over one-fifth of the global emission of PAHs were originated from China (Han et al. 2019). Absorbed PAHs could rapidly metabolize to mono-hydroxylated PAH (OH-PAHs) forms and mainly excreted through urine (Li et al. 2010). Therefore, urinary OH-PAHs have been recognized as common internal biomarkers for PAH exposure (Shi et al. 2021). Environmental endocrine-disturbing chemicals (EDCs) are pervasively present in environment, and large number of exogenous agents has been identified as potential EDCs (Diamanti-Kandarakis et al. 2009). Thyroid gland is susceptible to disturbing compounds. Considering the wellestablished essential role of thyroid hormones in physical development and metabolism regulation (Gilbert et al. 2020), adverse effect of thyroid dysfunction on nervous and cardiovascular system (Goodman andGilbert 2007, Rodondi et al. 2006), thyroid-disturbing chemicals have attracted increasing attention these years (Calsolaro et al. 2017). Growing number of thyroid-disturbing substances has been found over the past decades (Bai et al. 2018;Zhang et al. 2022). Mechanistically, EDCs interfere with thyroid function through hypothalamic and anterior pituitary levels (Yilmaz et al. 2020).
Effects of PAHs on human health risk have been extensively documented (Ambade et al. 2021), including teratogenicity, reproductive-disturbing effects, endocrinedisturbing effects, neurotoxicity, and immunotoxicity . However, few previous researches have reported the association between different PAH metabolites and thyroid hormones. The research findings remained conflicting. Study investigating the relationship between exposure to PAH and thyroid dysfunction in non-occupational population was firstly undertaken in 480 Chinese adult males. Only 2-hydroxyfluorene (2-FLU) concentration was found positively associated with thyroid-stimulating hormone (TSH) (Zhu et al. 2009). Data from 2007-2008 NHANES indicated that 2-hydroxynapthalene (2-NAP), 2-hydroxyphenanthrene (2-OHP), and 1-hydroxypyrene (1-HP) were positively associated with total triiodothyronine (T3) in females. However, 1-hydroxyphenanthrene (1-OHP), 2-OHP, and 9-hydroxypyrene (9-HP) were negatively associated with the levels of free thyroxine (T4) (Jain 2016). Inconsistently, a nationwide survey conducted in Korean suggested no significant association of exposure to PAH and thyroid hormones (Kim et al. 2021). The association between PAH metabolites and thyroid autoantibodies has been poorly explored. Limited studies investigated in the involvement of PAH exposure in thyroid autoantibodies. In Korean National Environmental Health Survey 2015-2017, 4 urinary PAH metabolites 1-HP, 2-NAP, 2-FLU, and 1-OHP were measured. However, no association was found between PAH metabolites and the presence of thyroglobulin antibody (TgAb) or thyroid peroxidase antibody (TPOAb) (Kim et al. 2021).
The sex-specific effect of 10 urinary PAH metabolites on both thyroid hormones and thyroid autoantibodies remains uninvestigated. Hence, the objective of the present study was to provide a detailed profile of 10 urinary PAH metabolites in 1645 adult participants. Additionally, the association between these 10 PAH metabolites and thyroid profiles (total T3, free T3, total T4, free T4, TSH, Tg, TgAb, and TPOAb) was evaluated. Additionally, stratified analysis by gender was performed for further exploration of PAH exposure on thyroid dysfunction in males and females. NHANES 2011NHANES -2012 cycle was selected because it is the latest NHANES data with obtained complete thyroid profiles. Figure 1 shows the diagram of sample selection. Among the 9756 study participants, subjects aged below 18, with thyroid cancer/diseases, and pregnant females were all excluded. In the remaining 5864 participants, 4081 were absent for the data of urinary PAH metabolites or creatinine, 118 were absent for thyroid hormones or antibody parameters. As a result, 1645 participants (836 males and 809 females) were eventually enrolled in the present study. NHANES program was approved by the National Center for Health Statistics Ethics Review Board. All participants provided written informed consent at recruitment.

PAH metabolites
Urine specimens were collected, processed, and stored under − 20 °C condition until analysis for PAH metabolites. The specific OH-PAHs were measured in this method. This procedure involves enzymatic hydrolysis of glucuronidated/ sulfated OH-PAH metabolites in urine, extraction, derivatization, and subsequent analysis using isotope dilution capillary gas chromatography/tandem mass spectrometry (GC-MS/MS) (Li et al. 2006). A total of 10 PAH metabolites were detected in urine samples; the abbreviations and lower limit of detection (LLOD) for the 10 urinary PAH metabolites are listed in Table 1. If the concentrations of PAH metabolites were below LLOD, LLOD divided by the square root of 2 was used to substitute original data. PAH metabolite levels were adjusted by urinary creatinine (CR) to correct for variable urine dilution in the spot urine samples (Wahlberg et al. 2019), and CR was measured via enzymatic assay using Roche/Hitachi Modulator P Chemistry Analyzer.

Thyroid hormones
Thyroid blood specimens were collected, processed, stored, and shipped to Collaborative Laboratory Services, Ottumwa, Iowa. Thyroid-stimulating hormone (TSH) was quantified with a third-generation two-site immunoenzymatic ("sandwich") HYPERsensitive human TSH assay. Total T3, free T3, and total T4 were quantified with competitive binding immunoenzymatic assay. Free T4 was quantified with a two-step enzyme immunoassay. Tg was assessed by a simultaneous one-step sandwich assay. TgAb and TPOAb were quantified with a sequential two-step immunoenzymatic sandwich assay. The normal ranges for anti-TgAb and TPOAb were < 4.0 IU/mL and < 9.0 IU/ mL, respectively . The detailed description of specimen collection, processing, and laboratory method for thyroid profile could be obtained in NHANES 2011-2012.

Covariates
The covariates included demographic and questionnaire data in the current research. Demographic information included age, gender, body mass index (BMI), education level, and marital status. The educational level was defined as below high school, high school, and above high school. Marital status was classified as married/living with partner, widowed/ divorced/separated, and never married. Questionnaire data consisted smoking and drinking status. Smoking status was divided into current smoker (smoked at least 100 cigarettes in life and currently smoked), former smoker (smoked at least 100 cigarettes in life but now quitted smoking), and non-smoker (smoked less than 100 cigarettes in life). Drinking status was classified as drinker (had at least 12 alcohol drinks per year) and non-drinker (had less than 12 alcohol drinks per year).

Statistical analysis
Normally and non-normally distributed continuous data were presented as mean ± standard deviation, and median (interquartile range), respectively. While categorical data was presented as numbers and percentages. Student's t tests, Kruskal-Wallis tests, and chi-squared tests were used for comparisons of continuous and categorical variables, respectively. Spearman analysis was performed to assess the correlation between these 10 PAH metabolites. The concentrations of 10 PAH metabolites were treated as both continuous and categorical variables in the statistical analyses. Each PAH metabolite was equally divided into 4 quartiles when recognized as categorical variable. The weighed multivariate linear regression and logistic regression models adjusting for potential confounders were performed to evaluate the regression coefficients of thyroid profiles and odds ratios for thyroid autoantibodies of PAH metabolites. Furthermore, subgroup analysis stratified by gender was performed to assess the gender-specific effect of PAH exposure on thyroid profiles. Sample weights were considered accounting for complex survey design, unequal probabilities of individual sampling, and non-response of eligible participants. Statistical analyses were conducted using R software and SPSS software. Two-sided p < 0.05 was recognized as statistically significant.

Demographic characteristics
A total of 1645 eligible participants were enrolled in the present study. The average age was 46.48 ± 18.52, and 836 was male. Table 2 shows the characteristics of study population by gender. Male participants showed higher percentage of married/living with partner and never married, higher smoking, and higher drinking rates compared with female individuals. Meanwhile, enrolled males were less likely to receive an education above high school. The distributions of age, race, and BMI were similar in males and females.

Discussion
To our best knowledge, this is the first research addressing the effect of 10 urinary PAH metabolites on thyroid hormones and thyroid autoantibodies in a gender-specific analysis. The present study provided 10 kinds of CR-adjusted urinary PAH metabolite distribution. Additionally, mild to Models adjusted for: age, race, BMI, education level, marital status, smoking, and drinking status T3 triiodothyronine, T4 thyroxine, TSH thyroid-stimulating hormone, TG thyroglobulin The bold values represents p value < 0.05   after adjusting potential confounders. However, no statistical significances were observed between higher quartile of PAH metabolites and the prevalence of TPOAb in the whole population. Gender-specific effect of PAH exposure on thyroid profiles was evaluated, and significant different results were found between males and females. Collectively, the present research revealed the positive association between PAH and thyroid dysfunction. Therefore, enhanced surveillance of PAH exposure in environment media is warranted. PAH is one kind of widely distributed environmental persistent organic pollutants (Famiyeh et al. 2021). Acute and chronic exposure to PAH could contribute to skin irritation, liver dysfunction, kidney dysfunction, DNA damage. Seriously, PAH exposure even lead to gastrointestinal carcinoma and increased cardiopulmonary mortality (Cheng et al. 2021;Lawal 2017;Lee et al. 2011;Zhu et al. 2018). However, the relationship between PAH exposure and thyroid effect remains largely uninvestigated. OH-PAHs in urine samples were most frequently used for evaluation of internal exposure to PAH due to its high excretion rates in urine, convenient, and invasive sampling method (Chetiyanukornkul et al. 2006). Therefore, the current study used urinary PAH metabolites to evaluate the exposure of PAH on thyroid hormones and thyroid autoantibodies.
There existed some evidence showing the association between exposure to PAH and thyroid effect. However, the research findings of previous reports remained remarkably controversial. 2-FLU in urine samples was found related to increased male TSH levels (Zhu et al. 2009), whereas no significant relationship was observed between 4 PAH metabolites and thyroid hormones in the Korean National Environmental Health Survey (Kim et al. 2021). Additionally, recent research assessing PAH exposure in Iran pregnant women indicated that acenaphthene concentrations in maternal blood samples were negatively associated with neonatal TSH levels (β = − 0.99) (Dehghani et al. 2022). The research population with different demographic background and distinct exposure to PAH might explain the discrepancies between these research results. Moreover, different studies focused on and detected distinctive kinds of PAH metabolites; no uniform regulation was formed. However, limited research explored gender-specific effect of a total of 10 PAH metabolites on both thyroid hormones and thyroid autoantibodies in adults.
The gender-specific effect of PAH exposure on adverse health outcomes has been investigated, including depression symptoms ) and skeletal muscle atrophy . Moreover, the gender-specific effect was also observed in thyroid nodular goiter and papillary thyroid carcinoma (Liu et al. 2021). However, gender-dependent association between PAH and thyroid hormones was limited. In the present research, positive associations were observed between 1-NAP and free T4, and 2-FLU and total T3 in males. While in females, 2-FLU, 2-OHP, 1-HP, and 9-FLU were positively related to total T3. 2-NAP, 2-FLU, and 9-FLU were associated with increased free T3 levels. Higher 2-FLU levels were also associated with increased concentrations of total T4. Consistently, previous findings showed different effects of PAH exposure on thyroid hormones in men and women. Furthermore, they also revealed that elevated total T3 levels were associated with 1-HP in adult females (Jain 2016). The reasons for discrepancies might be as follows: Thyroid hormones and sex hormones are inseparable due to the interaction between the hypothalamus-pituitary-thyroid axis and the hypothalamic-pituitary-gonadal axis. (Kjaergaard et al. 2021). Compared with males, females are more susceptible to chromosome damage and oxidative stress induced by PAH exposure (Guo et al. 2014). According to previous reports, 2-FLU and 2-OHP were both 3-ring OH-PAHs presenting some estrogenic activity (Hayakawa et al. 2007). The structural characteristic and estrogenic activities might explain the significant association between 2-FLU, 2-OHP, and thyroid profiles in females. Furthermore, 4-ring OHPAH 1-HP also presented some estrogenic activity (Hayakawa et al. 2007). Statistical significant relationship between 1-HP and total T3 was observed in adult females. Considering the distinct distribution of PAH metabolites and dramatically distinctive gender-specific associations between PAH and thyroid profiles, more work is needed to further elucidate the mechanisms of PAH exposure on thyroid dysfunction in participants with different gender. Experimental researches have conducted preliminary exploration on the mechanisms of the adverse development and function of thyroid gland induced by PAH exposure. Gentes and his colleagues firstly reported naturally heavy PAH exposure contributed to altered thyroid content in nestling in Athabasca Oil Sands Region (AOSR) (Gentes et al. 2006). In rockfish embryos, pyrene exposure decreased the expression of thyroid primordium markers and thyroid receptor genes. In addition, thyroid development and function-related genes, including Fgfr2, Hoxa3a, Deio1, Ttr, and Tg, were all altered after PAH exposure (He et al. 2012). TPO acts as a crucial factor in thyroid hormone (TH) synthesis through catalyzation iodine transfer to thyroglobulin (Taurog et al. 1996). Previous research determined that PAH disturbed TH axis via alteration of TPO activity. Specifically, dibenzo(a,h)anthracene increased TPO activity, whereas 3-methylchloranthracene, benzo(k) fluoranthene, pyrene, and benzo(e)pyrene suppressed TPO activity (Song et al. 2012). The current research suggested that higher 2-OHP levels were related to the prevalence of TPOAb in males. The association between PAH, thyroid autoantibodies, and thyroid autoimmune disorders remains largely unknown. Future research is merited to investigate the underlying mechanisms of the involvement of thyroid autoantibodies in the effect of PAH exposure on thyroid dysfunction.
The present research has several advantages. Firstly, this is a nationwide, representative research with relatively large sample. Secondly, a total of 10 PAH metabolites were measured in urine samples for each participant, which provided abundant information for PAH exposure. Lastly, gender stratification analysis broadened our understanding of PAH exposure on thyroid dysfunction. However, the study has some limitations. Firstly, the association between PAH metabolites in urine samples and thyroid dysfunction was assessed in a crosssectional study. Further prospective research is warranted for exploration of the causal relationship between PAH exposure and thyroid dysfunction. In addition, data of urinary PAH metabolites and/or thyroid profiles in many NHANES cycles were absent. Therefore, the latest dataset in NHANES 2011-2012 was selected for analysis. Future research with larger sample and complete research data is need to be conducted to expand our understanding of PAH and abnormal thyroid hormones/autoantibodies. Lastly, the research population was non-occupational-exposed general population with relatively low exposure to PAH. Different research findings might be obtained in participants with heavy exposure to PAH.

Conclusion
In conclusion, the current research provided the diverse distribution of a total of 10 PAH metabolites in males and females. The findings also suggested a positive association between 2-FLU and total T3. Higher quartile of 2-NAP was associated with the prevalence of elevated TgAb. Furthermore, significant gender differences were observed on the relationship between PAH metabolite, thyroid hormones, and thyroid autoantibodies. Further studies were needed to confirm the association between these urinary PAH metabolites and thyroid profiles, and explore the underlying mechanisms.