Introduction

Polycyclic aromatic hydrocarbons (PAHs) are hydrocarbons composed of two or more benzene rings. They are highly soluble solvents and possess lipophilic properties. PAHs are ubiquitous in the environment and widely distributed.

Numerous studies have demonstrated that certain PAHs are hazardous, persistent chemicals linked to cancer, teratogenicity, cardiovascular disease,1 endocrine disruption,2 immune disorders3 and nervous system diseases,4 posing serious risks to human health. Environmental PAHs primarily stem from the inadequate combustion of organic matter. Major sources of PAH pollution include anthropogenic activities, such as smoking, high-temperature cooking, vehicle exhaust emissions, mining and metal processing. Exposure occurs mainly through breathing, ingestion and dermal contact. Once absorbed, PAHs are rapidly metabolised into hydroxyl metabolites, which are primarily excreted into urine. Consequently, urinary hydroxyl metabolites of PAHs (OH-PAHs) serve as reliable indicators of PAH exposure.5

In humans, the thyroid gland is an essential endocrine organ responsible for synthesising triiodothyronine (T3) and thyroxine (T4). It plays a significant role in endocrine metabolism and the functions of the nervous, cancer and reproductive system.6–8 Previous studies as early as 2008 suggested that PAHs influence the transcriptional regulation of thyroid hormone (TH) release and may impair thyroid function.9 Numerous studies have concentrated on the relationships between PAHs and THs across various populations,10 as well as the implications of combined PAH exposure on TH levels.11

PAHs function as endocrine disruptors by imitating hormones essential for homeostasis.12 Their connections with aromatic hydrocarbon receptors (AHR) and hormone signalling pathways, such as THs,13 emphasise their role in endocrine disruption. AHR signalling pathway plays a crucial role in regulating immune function, cell cycle development and reproductive system processes.14

Substantial evidence indicates that even subtle changes in thyroid function, even within the normal range, can significantly impact the body. Impaired TH sensitivity has been associated with metabolic diseases, cardiovascular ailments and increased all-cause mortality risk.15 16 TH sensitivity is classified into two types: peripheral sensitivity, which regulates the metabolism in body tissues, and central sensitivity, which impacts pituitary gland feedback set points.

To date, no existing studies have examined the association between PAHs and TH sensitivity. Therefore, the current study aimed to analyse seven urinary PAH metabolites in 2356 adults with normal thyroid function and explore their relationship with central TH sensitivity indicators. These indicators include the thyroid feedback quantile-based index (TFQI), which assesses TH sensitivity. Thyroid feedback quantile index of FT4 (TFQIFT4) is derived from the empirical joint distribution of free T4 (FT4) and thyroid-stimulating hormone (TSH), offering the advantage of minimising extreme values in cases of thyroid dysfunction. Similarly, thyroid feedback quantile index of FT3 (TFQIFT3) is based on the empirical joint distribution of free T3 (FT3) and TSH levels. Other indicators include the thyroid-stimulating hormone index (TSHI), the thyrotropin T4 resistance index (TT4RI) and the thyrotropin T3 resistance index (TT3RI). Peripheral TH sensitivity was assessed using the FT3/FT4 ratio.17 Analyses were stratified by sex and age. The findings of the current study enhanced our understanding of PAH’s exposure impact on TH sensitivity.

MethodsData

The National Health and Nutrition Examination Survey (NHANES) is a research programme that assesses the nutritional and health conditions of both children and adults in the USA. Its results help determine the prevalence of significant illnesses and identify the contributing factors. NHANES uses a nationally representative sample and obtains informed consent from all participants. The current study used publicly available data from the 2007–2012 NHANES. Participants were included based on the following criteria: aged≥20 years, no history of thyroid disorders, not pregnant, complete PAH metabolite data and TH levels within the normal range (FT3: 2.5–3.9 pg/mL, FT4: 7.74–20.64 pmol/L and TSH: 0.34–5.6 µIU/mL). Metabolites of PAHs in urine were detected employing capillary gas chromatography combined with high-resolution mass spectrometry during the 2007–2008 period and isotope dilution capillary gas chromatography–tandem mass spectrometry during the 2009–2012 period. The urine specimens encompassed numerous PAH metabolites, including 1-hydroxynaphthalene (1-NAP), 2-hydroxynaphthalene (2-NAP), 2-hydroxyfluorene (2-FLU), 3-hydroxyfluorene (3-FLU), 9-hydroxyfluorene (9-FLU), 1-hydroxyphenanthrene (1-PHE) and 1-hydroxypyrene (1-PYR). The seven urinary PAH metabolites were normalised to urine creatinine (Cr) levels and represented as ng/g Cr as independent variables. The current study used the publicly available NHANES dataset,18 accessible at https://www.cdc.gov/nchs/nhanes/. As this research involved a secondary analysis of anonymised data, it was exempted from ethical review by our institution’s ethics committee.

TH sensitivity index

TSH was measured using the access HYPER sensitive hTSH test method, which employs a two-site immunoenzymatic (‘sandwich’) approach with low cross reactivity. The specimens were combined with precise reagents and incubated, and the quantity of TSH was determined by measuring light intensity. FT4 levels were assessed using the access FT4 test method, in which a biotin-conjugated antibody binds to the target, followed by the binding and separation of T3-alkaline phosphatase. The analyte level was determined by measuring the light intensity and using a calibration curve. FT3 levels were determined using the access FT3 assay, in which T3 in the sample binds to anti-T3 antibodies and is connected with particles coated with a biotinylated T3 mimic. Following separation, a chemiluminescence substrate (Lumi-Phos530) was introduced to quantify the brightness, which was inversely proportional to the T3 concentration. A calibration curve was then used to estimate the analyte level.

Based on the test results and previously reported methods for calculating TH sensitivity, we calculated six different TH sensitivity indices, including TFQIFT4, TFQIFT3, TT4RI, TT3RI, TSHI and FT3/FT4 ratio.19 20

TFQIFT4=cdf FT4 (pmol/L) − (1 − cdf TSH (mIU/L)).

TFQIFT3=cdf FT3 (pmol/L) − (1 − cdf TSH (mIU/L)).

TT4RI=FT4 (pmol/L) × TSH (mIU/L).

TT3RI=FT3 (pmol/L) × TSH (mIU/L).

TSHI=ln TSH (mIU/L) + 0.1345 ∗ FT4 (pmol/L).

FT3/FT4=FT3 (pmol/L)/FT4 (pmol/L).

TFQI values span from –1 to 1. Positive TFQI values indicate impaired responsiveness to THs, while negative TFQI values denote enhanced sensitivity. A TFQI value of 0 indicates normal TH sensitivity. Elevated TT4RI, TT3RI and TSHI values suggest reduced responsiveness to THs, whereas an increased FT3/FT4 ratio indicates greater peripheral TH activity.

Covariates

The demographic information comprised the following: age (≤ 60 or > 60 years), gender (male or female), ethnic background (Hispanic, White, Black and others), family income-to-poverty ratio (PIR), educational attainment (none or less than high school, high school or its equivalent) and marital status (married or cohabiting, widowed, divorced, separated or unmarried). Additionally, the body mass index (BMI) was recorded. Medical history encompassed smoking status (non-smoker, former smoker and current smoker), alcohol use (yes or no) and the presence of hyperlipidaemia, diabetes mellitus (DM), hypertension (HBP), stroke, coronary heart disease (CHD), angina pectoris, congestive heart failure (CHF) and myocardial infarction (MI). Laboratory tests measured aspartate aminotransferase (AST), alanine aminotransferase (ALT), albumin, serum creatinine (Scr), total protein, total bilirubin and serum glucose levels. Smoking status was determined based on the participants’ responses to whether they had ever smoked at least 100 cigarettes in their lifetime. Those who answered ‘no’ were classified as never smokers. Participants who responded ‘yes’ were further queried about their current smoking status. Those who smoked over 100 cigarettes but no longer smoked were classified as former smokers, whereas those who affirmed both questions were classified as current smokers. Alcohol consumption was defined as drinking alcohol at least once per month. HBP was characterised by Low-Density Lipoprotein Cholesterol(LDL-C) levels≥3.4 mmol/L, thyroglobulin (TG) levels≥1.7 mmol/L,Total Cholesterol (TC) levels≥5.2 mmol/L or High-Density Lipoprotein Cholesterol(HDL-C) levels<1.0 mmol/L. Diabetes was defined by plasma glucose levels≥11.1 mmol/L after 2 hours, Hemoglobin A1c(HbA1c) levels≥6.5% or active use of medicines or insulin for diabetes treatment. HBP was diagnosed when systolic blood pressure≥140 mm Hg or diastolic blood pressure was 90 mm Hg.

Statistical analysis

Participants were grouped by age or sex. Continuous variables with a skewed distribution are summarised using the median and IQR, and group comparisons were conducted using the Rao–Scott χ2 test. Categorical variables are presented as unweighted frequencies with weighted percentages, and group comparisons were performed using the Wilcoxon rank-sum test, which is appropriate for complex survey samples.

To assess the relationship between PAH metabolites (1-NAP, 2-NAP, 2-FLU, 3-FLU, 9-FLU, 1-PHE and 1-PYR) and TH sensitivity indices (TFQIFT4, TFQIFT3, TT4RI, TSHI, FT3/FT4 and TT3RI), we conducted weighted multiple linear regression analyses. The svydesign function in R accounted for sampling weights, stratification and clustering, ensuring that our findings accurately represent the US population.

Before performing multiple linear regression, we calculated the skewness of each TH sensitivity index. Indices with a non-normal distribution (skewness>0.5), such as TT4RI and TT3RI, were log-transformed. To prevent instability in statistical parameter estimates due to small values, all TH sensitivity indices were multiplied by 1000. When interpreting regression coefficients, this unit transformation should be considered carefully. Notably, rescaling variables (multiplying by a constant) does not affect significance, only the magnitude of the coefficients.

The regression models were adjusted for multiple covariates, including age, sex, race or ethnicity, marital status, PIR, education level, BMI, smoking status, alcohol consumption, hyperlipidaemia, diabetes, HBP, stroke, CHF, CHD, angina, MI, ALT, AST, albumin, Scr, total protein, total bilirubin and serum glucose.

We also conducted stratified analyses by sex and age, as well as an interaction-stratified analysis combining both factors. In these analyses, covariate adjustments remained consistent with the overall model, except for the stratifying variable. To further explore the dose–response relationship between PAH metabolites and TH sensitivity indices, we employed weighted restricted cubic spline models, adjusting for the same covariates as in the weighted multiple linear regression.

For missing data, we applied multiple imputations using the fivefold imputation method.

We employed weighted quantile sum (WQS) regression analysis to assess the impact of PAH mixtures on TH sensitivity indices. The current study sample was randomly divided into a training dataset (40%) and a validation dataset (60%). This 40%–60% allocation prioritises validation robustness, as the larger validation set enhances reliability in assessing model generalisability. Additionally, the training set, comprising over 900 participants, provided a sample size sufficient for stable weight estimation in WQS regression (WQSR).

In the training dataset, each PAH metabolite was categorised into quartiles. A total quantile score was then created for each individual by summing the quartile ranks of all PAH metabolites. Bootstrapping within the training dataset was used to estimate empirical weights for each PAH metabolite in the mixture. These weights were then used to construct the WQSR score, representing the overall PAH mixture, and its statistical significance was tested in the validation dataset. PAH metabolites with estimated weights≥0.142 (1/7) were considered major contributors to the WQSR score.

Because the WQSR method assumes that all PAH metabolites in the mixture have effects in the same direction, we separately evaluated positive and negative associations of the WQSR score. Weights were estimated using 1000 bootstrap samples from the training dataset (40%) and tested for statistical significance in the validation dataset (60%). WQSR was implemented using the gWQS package in R.

All statistical analyses were conducted using R (V.4.3.2), with a two-sided P value<0.05 considered statistically significant.

Patient and public involvement

No patients or members of the public were involved in the study.

ResultsCharacteristics of participants

The current study used publicly NHANES data from 2007 to 2012, collected through a sophisticated multistage stratified cluster survey of 30 442 participants. Exclusion included 12 729 individuals under 20 years of age, 1720 with thyroid disorders and 127 pregnant participants. Additionally, those without PAH metabolism data or with levels exceeding the normal range were removed (figure 1).

Figure 1

Flowchart of study participants. FT3, free triiodothyronine; FT4, free thyroxine; NHANES, National Health and Nutrition Examination Survey; PAHs, polycyclic aromatic hydrocarbons; TSH, thyroid-stimulating hormone.

A total of 2536 participants were included in the current study, comprising 1343 men and 1193 women. The proportion of individuals aged≥60 years was similar between men and women (31.87% vs 27.41%), with no statistically significant difference (p>0.05). Furthermore, no significant differences were observed between men and women in race, education level, stroke history or total protein levels (p>0.05).

However, several other clinical characteristics differed significantly between the two groups (p<0.05), including marital status, PIR, BMI, smoking status, alcohol consumption, hyperlipidaemia, DM, HBP, CHF, CHD, angina pectoris, MI, ALT, AST, albumin, Scr, total bilirubin, serum glucose and the thyroid secretion marker FT3 (table 1).

Table 1

Characteristics of included participants (n (%)/median (Q1, Q3))

Statistical description of TH sensitivity index and OH-PAHs concentration in urine in different populations

A gender-based analysis revealed that urinary levels of 2-NAP, 1-PHE and 1-PYR were significantly higher in women than in men. However, women received lower TFQIFT4, TT4RI, TT3RI, TSHI and FT3/FT4 ratios than men (online supplemental table 1). An age-based analysis indicated that individuals aged≥60 years received significantly higher urinary 1-NAP concentrations than those under 60, while levels of 2-NAP, 2-FLU, 3-FLU and 1-PYR were significantly lower. Additionally, older individuals exhibited higher TFQIFT4, TFQIFT3, TT4RI, TT3RI and TSHI levels but a lower FT3/FT4 ratio than younger individuals (online supplemental table 2).

Connection between OH-PAHs and TH sensitivity index

To explore the relationship between PAH metabolites and TH sensitivity indices, we performed multiple linear regression analyses. After adjusting for confounders, no significant associations were observed in the overall population (online supplemental table 3). However, subgroup analyses stratified by sex and age revealed notable associations following adjustment for confounders.

In women, 1-NAP exhibited positive correlations with TFQIFT4 (weighted β: 0.0072, 95% CI 0.0018 to 0.0127, p=0.022), TT4RI (weighted β: 0.0099, 95% CI 0.0024 to 0.0173, p=0.021) and TSHI (weighted β: 0.0107, 95% CI 0.0038 to 0.0176, p=0.012) (online supplemental table 4). Additionally, among individuals aged 60 years or older, 1-PHE was negatively correlated with the FT3/FT4 ratio (weighted β: –0.9328, 95% CI –1.5055 to –0.3602, p=0.014) (online supplemental table 5).

Further subgroup analyses incorporating both age and sex demonstrated more nuanced findings. In women under 60 years of age, 1-NAP was positively associated with TFQIFT4 (weighted β: 0.0099, 95% CI 0.0025 to 0.0173, p=0.018), TFQIFT3 (weighted β: 0.0095, 95% CI 0.0042 to 0.0148, p=0.006), TT4RI (weighted β: 0.0157, 95% CI 0.0067 to 0.0247, p=0.006), TSHI (weighted β: 0.0152, 95% CI 0.0064 to 0.024, p=0.007) and TT3RI (weighted β: 0.018, 95% CI 0.0077 to 0.0283, p=0.006). Conversely, in men under 60 years of age, 1-NAP was negatively correlated with the FT3/FT4 ratio (weighted β: –0.0063, 95% CI –0.0123 to –0.0002, p=0.044).

Among men aged 60 years or older, 1-NAP was positively correlated with TT4RI (weighted β: 0.025, 95% CI 0.0048 to 0.0452, p=0.026), TSHI (weighted β: 0.0212, 95% CI 0.0007 to 0.0416, p=0.045), FT3/FT4 (weighted β: 0.0078, 95% CI 0.0027 to 0.0129, p=0.013), TFQIFT3 (weighted β: 0.0164, 95% CI 0.0075 to 0.0254, p=0.007) and TT3RI (weighted β: 0.0398, 95% CI 0.0164 to 0.0632, p=0.009). Moreover, in this subgroup, 1-PHE was negatively correlated with FT3/FT4 (weighted β: –0.8165, 95% CI –1.5288 to –0.1043, p=0.033), and 1-PYR also indicated a negative correlation with FT3/FT4 (weighted β: –2.165, 95% CI –4.2395 to −0.0905, p=0.044) (online supplemental table 6).

The dose–response link between the TH sensitivity index and the concentrations of OH-PAHs

In females, no nonlinear relationships were observed between 1-NAP and central TH sensitivity indices, including TFQIFT4 (figure 2A), TT4RI (figure 2B) and TSHI (figure 2C). Similarly, no nonlinear correlation was found between 1-PHE and the peripheral TH sensitivity index FT3/FT4 in individuals aged 60 years and older (figure 3).

Figure 2

Dose–response connections between TH sensitivity indices and urinary PAH metabolites in females. (A) TFQIFT4 and 1-NAP. (B) TT4RI and 1-NAP. (C) Thyrotropin index (TSHI) and 1-NAP. 1-NAP, 1-hydroxynaphthalene; PAH, polycyclic aromatic hydrocarbon; TFQIFT4, thyroid feedback quantile index of FT4; TH, thyroid hormone; TSHI, thyroid-stimulating hormone index.

Figure 3

Dose–response connections between TH sensitivity indices and urinary PAH metabolites in adults aged 60 or older. FT3/FT4, free triiodothyronine/free thyroxine; PAH, polycyclic aromatic hydrocarbon; 1-PHE, 1-hydroxyphenanthrene; TH, thyroid hormone.

Further subgroup analyses revealed consistent findings across different demographic groups. In females under 60 years of age, no nonlinear relationships were detected between 1-NAP and central TH sensitivity indices, including TFQIFT4 (online supplemental figure 1A), TT4RI (online supplemental figure 1B), TSHI (online supplemental figure 1C), TFQIFT3 (online supplemental figure 1D) and TT3RI (online supplemental figure 1E). Similarly, in males under 60 years of age, no nonlinear relationship was observed between 1-NAP and the peripheral TH sensitivity index FT3/FT4 (online supplemental figure 2).

Among males aged 60 years and older, no nonlinear relationships were identified between 1-NAP and central TH sensitivity indices, including TT4RI (online supplemental figure 3A), TSHI (online supplemental figure 3B), FT3/FT4 (online supplemental figure 3C), TFQIFT3 (online supplemental figure 3D) and TT3RI (online supplemental figure 3E). Furthermore, no nonlinear relationships were found between 1-PHE and FT3/FT4 (online supplemental figure 3F) or between 1-PYR and FT3/FT4 (online supplemental figure 3G).

Association between OH-PAHs and TH sensitivity indices in WQS analysis

To analyse the impact of mixed exposure on TH sensitivity indices, we employed a WQS model to evaluate the effects of seven OH-PAHs on TH sensitivity indices (table 2). The results revealed that OH-PAHs exerted statistically significant effects on TH sensitivity indices, including TFQIFT4, TFQIFT3, TT4RI, TT3RI and TSHI. However, different components or concentrations of OH-PAHs may exert varying effects on these indices. Further analysis indicated that 9-FLU might contribute to an increase in TFQIFT4 (figure 4A), while 1-PHE, 1-PYR and 9-FLU could lead to an increase in TFQIFT3 (figure 4B). Additionally, 1-PYR was associated with an increase in TT4RI (figure 4D), and 2-PYR, 9-FLU and 1-PHE were linked to an increase in TT3RI (figure 4E). Besides, 1-PYR was found to potentially elevate TSHI (figure 4F). Conversely, 2-FLU, 2-NAP and 9-FLU were associated with a decrease in TSHI (figure 4G), and 2-NAP was linked to a reduction in TFQIFT3 (figure 4C).

Figure 4

WQSR weights in the WQSR model for the association between TH sensitivity indices and WQSR index of seven PAH metabolites mixtures. (A) WQSR index weights for models of TFQIFT4 negative. (B) WQSR index weights for models of TFQIFT3 positive. (C) WQSR index weights for models of TFQIFT3 negative. (D) WQSR index weights for models of TT4RI positive. (E) WQSR index weights for models of TT3RI positive. (F) WQSR index weights for models of TSHI positive. (G) WQSR index weights for models of TSHI negative. 2-FLU, 2-hydroxyfluorene; 3-FLU, 3-hydroxyfluorene; 9-FLU, 9-hydroxyfluorene; 1-NAP, 1-hydroxynaphthalene; 2-NAP, 2-hydroxynaphthalene; PAH, polycyclic aromatic hydrocarbon; 1-PHE, 1-hydroxyphenanthrene; 1-PYR, 1-hydroxypyrene; TFQIFT3, thyroid feedback quantile index of FT3; TFQIFT4, thyroid feedback quantile index of FT4; TH, thyroid hormone; TSHI, thyroid-stimulating hormone index; TT3RI; thyrotropin triiodothyronine resistance index; TT4RI, thyrotropin thyroxine resistance index; WQSR, weighted quantile sum regression.

Table 2

Association between PAH metabolites and TH sensitivity indices based on WQSR analysis

Discussion

To the best of our knowledge, the current study is the first to investigate the connection and dosage–response correlation between PAH biomarkers and TH sensitivity in individuals with normal thyroid function. Our analysis revealed that PAH metabolites, such as 2-NAP, 1-PHE and 1-PYR, were significantly higher in women compared with men. However, TFQIFT4, TT4RI, TT3RI, TSHI and FT3/FT4 ratios were lower in women than in men. Regarding age groups, urinary 1-NAP concentrations were significantly higher in individuals aged≥60 compared with those aged<60. In contrast, levels of 2-NAP, 2-FLU, 3-FLU and 1-PYR were lower in the≥60 age group compared with the<60 age group. Furthermore, TFQIFT4, TFQIFT3, TT4RI, TT3RI and TSHI were higher in individuals aged≥60, whereas FT3/FT4 ratio was lower. These findings suggest that women face a higher risk of PAH exposure, while men and older adults exhibit reduced TH sensitivity, likely influenced by factors, such as blood glucose levels, blood pressure and uric acid levels.16 21 The analysis revealed distinct differences between men and women across various factors, including marital status, PIR, BMI, smoking, alcohol use, hyperlipidaemia, diabetes, HBP, CHF, CHD, angina pectoris, MI and levels of ALT, AST, albumin, Scr, total bilirubin and serum glucose. Compared with women, men were more likely to be married or unmarried, received higher PIR, increased body weight, higher smoking rates, lower alcohol consumption and higher incidence of HBP, diabetes, hyperlipidaemia, CHF, CHD, angina pectoris and MI. Furthermore, men received elevated levels of FT3, ALT, AST, albumin, Scr, total bilirubin and serum glucose.

In the current study, no correlation was observed between PAHs and markers of TH sensitivity in the general population. After adjusting for confounders, 1-NAP was positively associated with TFQIFT4, TT4RI and TSHI in women, but no such association was evident in men. Gender-specific PAH exposures have been investigated for their impact on TH levels. Manthar Ali Mallah et al. discovered positive correlations between 1-NAP and FT4, as well as between 2-FLU and total triiodothyronine (TT3) in men. In women, 2-FLU, 2-hydroxybenzopyrene (2-OHP), 1-HP and 9-FLU were associated with elevated TH levels, with 2-NAP, 2-FLU and 9-FLU specifically linked to increased FT4.1 Another study revealed that, in females, heightened levels of 2-hydroxynaphthalene, 2-hydroxyphenanthrene and 1-HP were linked to increased TT3 levels. Conversely, in males, elevated levels of 1-PHE, 2-hydroxyphenanthrene and 9-hydroxypyrene were linked to lower FT4 levels.22 This variation may be influenced by the hypothalamus–pituitary–thyroid (HPT) axis and its interaction with the hypothalamus–pituitary–gonad axis, as the association between TH and sex hormones is intertwined and inseparable.23 Women exhibited increased susceptibility to chromosomal damage and oxidative stress from PAH exposure compared with men.24 In the 2013–2016 NHANES, 1-NAP demonstrated a negative correlation with estradiol, an association that was particularly pronounced in postmenopausal women.25

Among individuals aged 60 and above, we found that 1-PHE was negatively linked to the FT3/FT4 ratio. Notably, the duration and mode of PAH exposure to PAHs are key determinants of health outcomes. Older adults, compared with younger individuals under similar exposure conditions, may be more vulnerable to PAHs' impacts.26

Given that both gender and age influence the results, we conducted a joint stratified analysis by gender and age. Among females under 60 years of age, central TH sensitivity is more susceptible to PAH exposure, whereas in males under 60, peripheral TH sensitivity is more likely to be influenced. In males aged 60 and above, both central and peripheral TH sensitivity may be affected by PAH exposure. The WQS analysis further confirmed the association between OH-PAHs and TH sensitivity. The seven metabolites generated during PAH metabolism—1-NAP, 2-NAP, 2-FLU, 3-FLU, 9-FLU, 1-PHE and 1-PYR—originate from different parent PAHs: naphthalene, fluorene, phenanthrene and pyrene. Among these, 1-NAP is a common biomarker for naphthalene exposure, which primarily originates from combustion pollutants, industrial emissions, transportation, tobacco and certain household chemicals, making it a significant environmental contaminant. 2-NAP, another biomarker for naphthalene exposure, is produced through a slightly different metabolic pathway than 1-NAP, depending on cytochrome P450 enzyme activity. 2-FLU is a biomarker for fluorene exposure, widely present in environmental and industrial sources. Similarly, 3-FLU also indicates a fluorene exposure but may follow a slightly different metabolic pathway due to variations in cytochrome P450 enzyme specificity. 9-FLU is typically used to track long-term or cumulative fluorene exposure. 1-PHE acts as a biomarker for phenanthrene exposure, which is prevalent in automobile emissions and cigarette smoke. 1-PYR, the most reliable biomarker for pyrene exposure, is widely used in environmental and occupational health studies. The findings of the current study indicate that naphthalene exposure likely impairs TH sensitivity in both females and males under the age of 60. In males aged 60 and above, exposure to naphthalene, phenanthrene and pyrene is associated with a higher risk of impaired TH sensitivity.

Although existing studies suggest an association between PAHs and TH sensitivity, the mechanism underlying this link remains unclear. Several factors contribute to this complexity. Evidence suggests a connection between PAH exposure and thyroid function, with elevated 2-NAP concentrations aligning with the negative feedback mechanism of the HPT axis.10 Initially, PAH exposure appears to regulate pituitary TSH secretion, though findings remain inconsistent. Notably, in urine samples, 2-FLU was linked to elevated TSH levels in men.27 A cross-sectional study of children and adolescents in Iran found strong positive associations between PAHs (1-, 2-NAP, 1-PYR and 9-PHE) and serum TSH levels, suggesting that prolonged PAH exposure may impair thyroid function.28 Conversely, a recent study on pregnant women in Iran revealed a negative correlation between ethanolic phenol concentrations in maternal blood and neonatal TSH levels.29 Yang et al. also discovered a negative correlation between adult exposure to 2-NAP and TSH levels. Moreover, in adults, increased 1-PYR levels were linked to higher FT4 levels. In females, elevated 1-PYR levels correlated with increased FT3, FT4 and TT3, suggesting an overall elevation in TH levels. Conversely, in adolescents, higher levels of FT4 were significantly linked to increased 9-FLU concentrations, implying that PAH exposure may augment deiodinase (Deio) activity, resulting in reduced T4 and raised T3 levels.10 Previous PROTECT longitudinal-birth cohort investigations have revealed significant positive correlations between PAH exposure and maternal T3 levels.30 This aligns with zebrafish experiments, where larvae exposed to benzopyrene exhibited increased TT3 and decreased TT4 concentrations.31 Furthermore, PAH exposure increased the biliary excretion of FT4.32 Variations in findings may stem from differences among study participants, species’ variations and varying PAH exposure levels.

PAHs may regulate TH synthesis by influencing the thyroid peroxidase (TPO) activity. TPO plays a vital role in TH synthesis by facilitating iodine transfer to TG.33 Previous studies have established that PAHs disrupt the TH axis by altering TPO activity. Specifically, dibenzo(a,h)anthracene enhances TPO activity, whereas 3-methylchloroanthracene, benzo(k)fluoroanthracene, pyrene and benzo(e)anthracene hinder it.34 Additionally, elevated 2-OHP levels have been linked to TPO antibodies in men.11 Experimental studies have explored the mechanisms by which PAH exposure affects thyroid development and function. In rockfish embryos, PAH exposure altered genes associated with thyroid regulation, including fibroblast growth factor receptor 2, Hoxa3a, Deio1, transthyretin and TG.35 Thyroid hormone receptors (TRs) are nuclear receptor proteins that regulate TH activity. 1-NAP acts as a TR antagonist, disrupting the TR signalling pathway.9 However, organic pollutants linked to diesel emissions, including PAHs and their derivatives, have been demonstrated to activate thyroid receptor alpha (TRα) and significantly amplify the effectiveness of the natural TRα ligand T3 in vitro.36 More rigorous experimental studies are required to further explore the regulatory mechanisms of PAHs on thyroid axis hormones.

The current study examined the connection between PAH exposure and TH sensitivity in euthyroid individuals—a topic not extensively addressed in the existing research. Strengths of the current study include a large sample size and the use of various analytical techniques, such as multiple linear regression and dose–response analysis. We observed that women under the age of 60 are more susceptible to naphthalene exposure affecting central TH sensitivity, while men are more prone to peripheral sensitivity changes. In men aged 60 and above, exposure to naphthalene, phenanthrene, and anthracene may simultaneously affect both central and peripheral sensitivity. However, several limitations must be acknowledged. The cross-sectional methodology of the current study limits our ability to establish causal relationships. Variations in measurement methods and limits of detection for OH-PAHs across different study periods may introduce errors despite our adjustments using urinary Cr values. Besides, assessing OH-PAHs from a single urine sample may inaccurately reflect long-term PAH exposure.