Thyroid hormones consist of thyroxine (T4, approximately 80%) and triiodothyronine (T3, approximately 20%) (1). T4 is primarily released by the thyroid gland, whereas only about 20% of T3 is directly produced by the thyroid gland (2). The remaining 80% of T3 is converted from T4 in peripheral tissues by deiodinases (DIO1, DIO2) (2). Among these hormones, T3 is the active form, while T4 primarily serves as a precursor that can be converted into the active form (3). In the bloodstream, most T3 and T4 molecules are bound to serum binding proteins, with only a small fraction circulating freely as free thyroxine (FT4, approximately 0.03%) and free triiodothyronine (FT3, approximately 0.5%) (4). These free and bound portions together constitute total triiodothyronine (TT3) and total thyroxine (TT4) (5). The production of thyroid hormones is controlled by a complex negative feedback system involving thyroid-stimulating hormone (TSH) released by the pituitary gland (3). TSH regulates the synthesis and secretion of T3 and T4, while T3 and T4, in turn, regulate the production of TSH through negative feedback (3). T3 and T4 play crucial roles in many tissues, maintaining normal growth and metabolism, regulating energy expenditure and lipid balance, and playing key roles in the cardiovascular system (6). Studies have shown that abnormal levels of thyroid hormones, whether too high or too low, can have adverse effects on the body (7–9). Low levels of thyroid hormones can lead to coronary heart disease (7, 8), while high levels of thyroid hormones can lead to heart failure (9).

Some research has explored the association between thyroid hormones and lipid metabolism. For instance, research by Kim et al. on the Chinese population identified a significant positive association between FT4 and the levels of both high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) (10). Conversely, a study involving 2,183 adults found a negative association between FT4 and HDL-C (11). Similarly, Wang et al. reported a negative association between thyroid hormones and hyperlipidemia indicators, such as LDL-C (12). Non-high-density lipoprotein cholesterol (non-HDL-C) represents all cholesterol carried by lipoproteins other than HDL-C, including the cholesterol content in very low-density lipoprotein, low-density lipoproteins, and other lipoprotein particles (13). Furthermore, some research has shown that higher levels of non-HDL-C are linked to mild inflammation, which may facilitate the conversion of TT4 to TT3, thereby elevating TT3 levels (14, 15). However, one study suggested that activating inflammatory mediators and exposure to low temperatures could reduce FT3 levels (16).

Currently, the methods for measuring TT3 and TT4 are primarily immunoassays, but these methods have significant errors and limitations (17). A study questioned the accuracy of immunoassay results for low, normal, and high concentrations of TT3 and validated the results using the gold standard liquid chromatography-tandem mass spectrometry (LC-MS/MS) (18). The results showed that the measurements of TT3 using low-concentration immunoassay methods were not ideal (18). Additionally, a recent study by the IFCC C-STFT used various immunoassay methods to measure TT4 and TT3 in the serum of healthy individuals and compared the results with those reported by isotope dilution tandem mass spectrometry (ID-LC-MS/MS) (19). The findings revealed that most methods did not meet the optimal 5% target set by C-STFT, with only a few TT4 measurements falling within 10% of the reference values (19, 20). Particularly, since the concentration of TT3 is ten times lower than that of TT4, its measurement requires higher sensitivity and precision (21). However, most TT3 measurements exhibited a positive bias, indicating that the detection of TT3 and TT4 needs to be re-standardized (19, 20). Although tandem mass spectrometry provides higher accuracy in measuring TT3 and TT4, its complex operational steps, which require the separation of analytes from binding proteins to measure free thyroid hormones, make it difficult to be widely applied in clinical settings in the short term (22). The assessment of thyroid function primarily relies on thyroid hormones and TSH levels (23). Therefore, it is particularly important to establish new blood lipid indicators that can more accurately reflect fluctuations in these levels for evaluating thyroid function.

The non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR) is increasingly regarded as a comprehensive and innovative parameter for assessing lipid components in atherosclerosis (24). NHHR is associated with atherosclerosis and provides significant predictive value for various diseases. For instance, a study on the Asian population demonstrated an association between NHHR and non-alcoholic fatty liver disease; a one-unit increase in NHHR could lead to a 64.5% increase in the ratio of alanine aminotransferase to aspartate aminotransferase (25, 26). Since the liver is the primary site for converting TT4 to TT3 (27), non-alcoholic fatty liver disease indirectly affects thyroid hormone levels by altering the metabolic rate of liver function (28). Another study indicated that NHHR offers more accurate predictions for metabolic syndrome than traditional lipid markers (29). Metabolic syndrome is characterized by elevated lipid levels, which can disrupt the iodination process of TT4, potentially resulting in increased levels of FT3 and TT3 and a decrease in FT4 (30, 31). Therefore, exploring the association between NHHR and thyroid hormones and functions is scientifically significant and positions NHHR as a novel biomarker, paving new pathways for monitoring and regulating thyroid hormone levels.

Mendelian randomization (MR) analysis uses genetic variation, such as single nucleotide polymorphisms (SNPs), as instrumental variables (32). By exploiting the random allocation of genetic variation, MR excludes the influence of confounding factors and helps identify causal relationships (32). Therefore, the MR method can effectively assess the causal association between NHHR and thyroid hormones.

Based on the background information and prior research, we hypothesize that NHHR, as an innovative biomarker, more accurately reflects fluctuations in thyroid hormone levels. Therefore, our study investigates the association between NHHR and thyroid hormones in American adults using data from the NHANES survey conducted between 2007 and 2012. Additionally, the use of two-sample MR deepens the understanding of the causal relationship between these variables.

The results found that NHHR is negatively association with FT4, the FT4/FT3, and TT4/FT3. The distinct non-linear relationships between NHHR and FT3, TT3, TT4, and the FT4/TT4 ratio suggest that changes in NHHR may have a more complex impact on the regulation of these thyroid hormones. In particular, the presence of non-linear associations may indicate that at certain levels of NHHR, thyroid hormone regulation could exhibit a threshold effect. That is, once NHHR reaches a critical point, the secretion or conversion rate of thyroid hormones may undergo a dramatic shift, increasing the sensitivity of thyroid function or making it more prone to disruption. This phenomenon could be due to the differential effects of NHHR on various metabolic pathways, further revealing the complexity of the interactions between metabolism and hormone balance. NHHR is an extremely valuable tool in clinical practice, crucial for predicting thyroid hormone levels and function (29, 39). For instance, patients with metabolic syndrome and dyslipidemia are at high risk for thyroid disorders (39). As an effective indicator for assessing metabolic syndrome, NHHR may also play a significant role in the early detection of thyroid diseases (29). For patients who exhibit thyroid dysfunction, regular monitoring of NHHR values can help doctors adjust treatment plans and effectively control hormone level fluctuations (29, 40). This is particularly critical for promoting healthy growth in children (41). Additionally, research indicates that NHHR is an effective tool for assessing the risk of CVD (42). Both hypothyroid and hyperthyroid patients are at increased risk for CVD (7, 9), and patients undergoing cardiac surgery may also develop acquired hypothyroidism (43). Thus, utilizing NHHR can more effectively manage the progression and prognosis of thyroid-related diseases in clinical practice.

This study discovered a positive association between NHHR and both TT3 and FT3 while noting a negative association with FT4. Although no studies directly report these relationships, previous investigations into the interaction between thyroid hormones and lipid abnormalities lend support to our findings. For example, a study in Malaysia demonstrated a notable negative association between FT3 levels and HDL-C (44). Roos and colleagues observed that FT4 showed a negative association with LDL-C and TG yet exhibited a positive association with HDL-C (45). Likewise, studies involving the Spanish population revealed a positive association between FT4 levels and HDL-C, alongside a marked negative association with LDL-C (46). A different research study showed that TT3 has a negative association with LDL-C by increasing the gene expression of LDL-C receptors and enhancing the clearance rate of LDL-C (47). Additionally, elevated levels of HDL-C and reduced levels of LDL-C are associated with a heightened risk of CVD (48). Many researchers have noted that the ratio of FT3/FT4, utilized to evaluate the conversion rate from thyroxine TT4 to triiodothyronine TT3, associates with heightened mortality rates from CVD (49). Consequently, there exists a remarkable link between high HDL-C, low LDL-C, and heightened TT3 levels (48, 49). Moreover, research indicates that heightened levels of NHHR are connected to an amplified risk of type 2 diabetes in patients with hyperthyroidism (42). Leptin, present in the adipose tissue of type 2 diabetes patients (50), can elevate TSH levels, thereby boosting deiodinase activity, facilitating the conversion of FT4 to FT3, and augmenting basal metabolic rate (51). Thus, the escalation in NHHR closely correlates with the decrease in FT4 and the increase in FT3. An animal experiment on obese dogs revealed that adipose tissue within obese animals stimulates the conversion of TT4 to TT3 in order to increase energy expenditure; 42% of obese dogs showed a decrease in FT4 concentration (52, 53). During this process, the transcriptional regulation of genes in adipose tissue contributes to elevated TG levels and decreased HDL-C levels (54). These findings reflect the correlation between NHHR and thyroid hormone levels, aligning with our conclusions. This association emphasizes the importance of NHHR in detecting thyroid hormones.

The underlying mechanisms of this relationship may involve several aspects: First, thyroid hormones increase the flow of bile acids, which promotes the liver’s consumption and synthesis of cholesterol, as well as its uptake from the blood (55, 56). This helps maintain a cholesterol balance in the liver. However, elevated NHHR levels disrupt this balance, leading to increased conversion of stored TT4 into the active form TT3 to restore equilibrium (57). Second, leptin, an essential neuroendocrine regulator, modulates the hypothalamic-pituitary-thyroid axis and governs the thyrotropin-releasing hormone production in the paraventricular and arcuate nuclei (58). An increase in lipid content raises leptin levels, which may lower FT4 levels by influencing serum TSH levels (59). Studies indicate that while FT4 is negatively associated with BMI, TSH is positively associated (60, 61). Moreover, the higher TSH and TT3 levels seen in individuals with obesity might reflect a physiological adaptation aimed at enhancing energy expenditure and managing body weight (57). Third, cholesterol is vital for the production of thyroid hormones (62). High lipid levels are associated with thyroid dysfunction and metabolic disturbances, which can affect the synthesis of thyroid hormones, potentially promoting the synthesis of TT3, although the impact on TT4 synthesis and release may vary (63). Lastly, elevated non-HDL-C levels are often associated with a low-grade inflammatory state, which could affect hormone metabolism (64). Research shows that inflammatory factors can enhance the expression of deiodinases, enzymes that specifically remove iodine atoms from TT4, producing active TT3 and inactive diiodothyronine (65, 66). Thus, inflammation may promote the conversion of TT4 to TT3 while reducing the overall level of TT4.

This research possesses multiple advantages. Firstly, it represents the most extensive population-based analysis conducted thus far to examine the associations of NHHR with various thyroid hormones. Secondly, our analysis utilized a sample that represents the entire U.S. adult demographic to depict the general population accurately. Additionally, we carried out a comprehensive analysis of different variables and made adjustments for covariates to confirm the reliability of our results.

Several limitations should be considered when interpreting the findings of this study. First, the cross-sectional design limits our ability to establish a causal relationship between NHHR and thyroid hormones. Although a potential association between these two factors was suggested in the present study, the simultaneous measurement of lipid biomarkers and thyroid hormone levels prevents us from determining the directionality or temporality of this relationship. Specifically, dyslipidemia, as indicated by an elevated NHHR, may impact thyroid function through mechanisms such as altered deiodination or disruption of thyroid hormone transport. Conversely, thyroid hormones have also been reported to play a crucial role in regulating lipid metabolism. Changes in thyroid function can lead to alterations in lipid profiles, including HDL-C and non-HDL-C cholesterol levels. The possibility of bidirectional interactions introduces the potential for reverse causality, where changes in thyroid hormones could influence NHHR rather than the other way around.

Secondly, despite adjusting for various confounding factors, residual confounding remains an issue that warrants attention. For example, during the NHANES 2007-2012 period, data on inflammatory markers were lacking. However, changes in inflammatory markers may concurrently affect both NHHR and thyroid hormone levels, thereby leading to biased results. Our analysis may not have sufficiently accounted for this potential factor.

Additionally, we conducted MR analysis to further investigate the relationship. Unfortunately, there is no available GWAS specifically focused on NHHR, and the GWAS summary data for TT3, FT3, TT4 and FT4 are also restricted (67). Therefore, we utilized GWAS data from the Global Lipids Genetics Consortium for HDL, LDL, and TC as exposures (32), and the UKB-based GWAS for hyperthyroidism and hypothyroidism as binary outcomes. However, no significant results were observed. In addition to the racial differences (mixed American vs. European descent), the discrepancy between the cross-sectional findings and MR results, we think, may be due to the use of related, but not fully equivalent, phenotypes. To further investigate, we analyzed the associations between HDL-C, LDL-C, TC, and these two binary phenotypes using NHANES data, and again, no significant results were found. As highlighted in our Introduction section, NHHR may provide greater value for disease prediction than traditional lipid markers, emphasizing the importance of further research into NHHR, including dedicated GWAS.

Additionally, thyroid hormone levels were measured using immunoassay methods, which are known to have lower accuracy, especially at low concentrations. Although NHANES employed high-quality laboratory techniques, the lack of LC-MS/MS validation might introduce measurement bias. Future studies incorporating LC-MS/MS validation could provide more accurate and reliable results.

Finally, the NHANES population primarily represents the U.S. population, and the findings may not be fully generalizable to other populations. Cultural, dietary, and genetic differences may influence the association between NHHR and thyroid hormones in other regions, necessitating validation in diverse cohorts.

In summary, this study revealed that among U.S. adults, an increase in NHHR was significantly negatively associated with FT4, FT4/FT3, and TT4/TT3. However, no significant association was found between HDL-C, LDL-C, TC, and the risk of hyperthyroidism or hypothyroidism, highlighting the importance of monitoring NHHR and thyroid hormone levels in clinical practice.