Revealing the causal relationship between thyroidectomy, thyroid function status, and eczema: a bidirectional two-sample Mendelian Randomization study

Abstract

This Mendelian Randomization (MR) study aims to explore the potential bidirectional causal relationship between thyroidectomy, thyroid function status, and eczema. We utilized summary statistics from genome-wide association studies of European ancestry for thyroidectomy (4360 cases and 458,573 controls), hyperthyroidism (2547 cases and 334,612 controls), and hypothyroidism (337,199 samples), employing genetic instruments to investigate their association with eczema (461,199 samples), and vice versa. Primary causal estimates were obtained using the inverse variance weighting method (IVW), with a series of sensitivity analyses conducted to assess the robustness of the results. The study found that genetically predicted thyroidectomy has a negative effect on eczema (OR = 0.645; 95% CI 0.498–0.837; P < 0.001), while neither hyperthyroidism nor hypothyroidism showed significant causal associations with eczema. Reverse Mendelian Randomization analysis indicated that genetically predicted eczema does not significantly affect the incidence of thyroidectomy, hyperthyroidism, or hypothyroidism. To correct for the impact of post-thyroidectomy hypothyroidism on eczema, a multivariable Mendelian Randomization analysis was performed, treating thyroidectomy and hypothyroidism as exposures and eczema as the outcome. After adjusting for hypothyroidism, thyroidectomy still reduced the risk of developing eczema (OR = 0.632; 95% CI 0.476–0.838; P = 0.001). This study elucidates the bidirectional causal relationship between thyroidectomy, thyroid function status, and eczema, providing ancillary scientific evidence on the influence of thyroidectomy on the mechanism of eczema development.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00403-024-03591-z.

Introduction

Thyroidectomy [1–5] is a widely implemented surgical procedure used to treat various thyroid conditions, including malignant tumors (such as papillary carcinoma, follicular carcinoma, and anaplastic carcinoma), benign tumors (which may carry a risk of malignant transformation or compress adjacent organs), hyperthyroidism (when medical and radioactive iodine treatments fail, such as in Graves’ disease), large goiters (causing compressive symptoms or cosmetic concerns), and recurrent thyroid diseases. Depending on the extent of removal, thyroidectomy can be categorized into total thyroidectomy, subtotal thyroidectomy, lobectomy, and isthmectomy.

Eczema [6–10] is a chronic, relapsing inflammatory skin disease with a multifactorial etiology, including genetic predisposition, immune system abnormalities, and environmental factors. The clinical manifestations of eczema are diverse, typically presenting as dry skin, pruritus, erythema, papules, vesicles, erosion, and crusting. In severe cases, it can lead to skin thickening and lichenification. The pathogenesis of eczema involves skin barrier dysfunction and immune dysregulation, characterized by abnormalities in the composition and function of the stratum corneum lipids and Th2 cell-mediated immune responses. Due to its significant impact on quality of life, eczema has received considerable attention from both the public and researchers.

Current research on the mechanisms underlying eczema primarily focuses on immune system and hormonal levels [11]. Studies have suggested a higher incidence of eczema in patients with primary thyrotoxicosis [12]. However, the specific mechanisms linking these conditions remain unclear. M.V. Kachuk et al. demonstrated that inducing hyperthyroidism in guinea pigs via the pituitary-thyroid axis significantly increased the incidence of eczema and allergic dermatitis [13]. Conversely, other studies have reported that while hypothyroidism does not promote the onset of eczema, it can exacerbate the severity of existing eczema [14]. Additionally, some research indicates that eczema occurrence may be related to the hypothalamic-pituitary-adrenal (HPA) axis [15].

In summary, the pathogenesis of eczema is not yet fully understood, and the role of the thyroid in eczema development remains to be elucidated. It is uncertain whether thyroid hormone levels influence eczema onset, whether the thyroid impacts the pituitary and subsequently the HPA axis via neuroendocrine pathways, or whether thyroid peroxidase antibodies play a role in eczema development. Investigating the causal relationship between thyroidectomy and eczema could provide scientific evidence for understanding the specific mechanisms of eczema development and offer comprehensive guidance on thyroid hormone supplementation post-thyroidectomy, aiding in clinical decision-making.

Mendelian Randomization (MR) leverages single nucleotide polymorphisms (SNPs) to infer causal relationships between exposures and disease outcomes based on genetic variation. In MR studies, genetic variants associated with phenotypes are used as instrumental variables to make causal inferences about exposure-outcome associations. Genetic variants are randomly assorted from parents to offspring according to Mendel’s laws and are determined at conception, thus minimizing confounding factors typically present in observational studies.

This study employs a bidirectional, multivariable two-sample Mendelian Randomization approach to investigate the causal relationships between thyroidectomy and eczema.

Materials and methods

To conduct a Mendelian Randomization (MR) study and ensure the validity and reliability of the conclusions, the selection of instrumental variables must satisfy three assumptions: (1) the genetic variants must be strongly associated with the exposure; (2) they must influence the outcome only through the exposure and not via other pathways; (3) they must not be associated with any confounding factors that relate to both the risk factor and the outcome.

For thyroidectomy, this study used summary statistics from genome-wide association studies (GWAS) conducted by The Neale Lab for Genomic Medicine, including 4360 cases and 458,573 controls of European ancestry. For hyperthyroidism, we included 2547 cases and 334,612 controls of European ancestry, and for hypothyroidism, we used 337,199 samples of European ancestry. The GWAS summary statistics for eczema, provided by Loh PR, included data on eczema from 461,199 samples of European ancestry.

To minimize the effects of linkage disequilibrium (LD) [16, 17], we selected single nucleotide polymorphisms (SNPs) as instrumental variables (IVs) based on established genome-wide significance thresholds (P < 5 × 10–8, r2 ≤ 0.001, Hardy–Weinberg equilibrium (H-W) met, genetic distance < 10,000 kb). IVs with F-values greater than 10 were included to avoid bias from weak instruments. For forward MR analysis with thyroidectomy, hyperthyroidism, and hypothyroidism as exposures and eczema as the outcome, we selected 130, 53, and 21 SNPs as IVs, respectively. For reverse MR analysis, we selected SNPs associated with eczema as the exposure.

Data availability

Exposure and outcome data were obtained from the GWAS OPEN database (https://gwas.mrcieu.ac.uk/), which is publicly accessible and does not require any permissions.

Statistical analysis

The inverse variance weighting (IVW) MR analysis was employed as the primary analysis method. Cochran’s Q and Rücker’s Q tests were conducted using IVW and MR-Egger methods for heterogeneity, and pleiotropy was assessed using the Egger-intercept method and leave-one-out sensitivity analysis.

To strengthen the evidence, we conducted an independent MR study using the PhenoScanner database and excluded pleiotropic SNPs related to confounding factors in the exposure-outcome association. Effect estimates were expressed as odds ratios (ORs) with 95% confidence intervals (CIs).

Due to multiple testing, associations with P-values below the Bonferroni correction threshold (α = 0.05/6 = 0.008) were considered statistically significant, while those with P-values ≥ 0.008 and < 0.05 were deemed suggestively significant. All analyses were conducted using open-source statistical software R (version 4.3.2). Reporting adhered to the STROBE-MR [18, 19] (Strengthening the Reporting of Observational Studies in Epidemiology—Mendelian Randomization) guidelines.

Ethics statement

This study was based on publicly available GWAS data. All original studies had obtained ethical approvals.

Results

The SNP selection process is shown in Fig. 1. Using the IVW method as the primary analysis, the estimates were reported as follows. Forward MR analysis indicated that genetically predicted thyroidectomy reduced the risk of eczema (OR = 0.645; 95% CI 0.498–0.837; P < 0.001). Neither hyperthyroidism nor hypothyroidism showed significant causal associations with eczema, with ORs and 95% CIs of (OR = 2.133; 95% CI 0.656–0.930; P = 0.208) and (OR = 1.126; 95% CI 0.082–15.449; P = 0.929), respectively (Fig. 2). Reverse MR analysis showed that genetically predicted eczema did not significantly affect the incidence of thyroidectomy, hyperthyroidism, or hypothyroidism (Fig. 3). Heterogeneity tests using Cochran’s Q and Rücker’s Q methods with IVW and MR-Egger indicated no significant heterogeneity. All MR analyses assessed pleiotropy using the Egger-intercept method, with no pleiotropy detected (P > 0.05). MR PRESSO confirmed the significance of all positive results from the IVW method (P < 0.05).

Fig. 1.

Flowchart (Left column shows the number of remaining SNPs after each step of forward MR analysis, middle column indicates the selection criteria for each step, and right column shows the number of remaining SNPs after each step of reverse MR analysis)

Fig. 2.

Forest Plot of Forward MR Analysis Results(The forest plot from top to bottom shows the MR analysis results with thyroidectomy, hyperthyroidism, and hypothyroidism as exposures and eczema as the outcome. The left column displays the MR analysis methods along with the corresponding OR values and 95% confidence intervals, with an asterisk indicating P < 0.05; the arrows on the right indicate that the actual upper limit of the OR value exceeds the range of the X-axis of the plot)

Fig. 3.

Forest Plot of Reverse MR Analysis Results (The forest plot from top to bottom shows the MR analysis results with thyroidectomy, hyperthyroidism, and hypothyroidism as outcomes and eczema as the exposure. The left column displays the MR analysis methods along with the corresponding OR values and 95% confidence intervals, with an asterisk indicating P < 0.05; the arrows on the right indicate that the actual upper limit of the OR value exceeds the range of the X-axis of the plot)

Validation

To account for the impact of hypothyroidism following thyroidectomy on the results, this study also employed multivariable Mendelian Randomization (MR) analysis. By simultaneously treating thyroidectomy and hypothyroidism as exposure factors and eczema as the outcome, the results indicated that, after adjusting for the influence of hypothyroidism, genetically predicted thyroidectomy still reduced the risk of developing eczema (OR = 0.632; 95% CI 0.476–0.838; P = 0.001).

Discussion

This study utilized bidirectional, multivariable, two-sample Mendelian Randomization analysis to elucidate the bidirectional causal relationship between thyroidectomy, thyroid function status, and the incidence of eczema. Notably, this study discovered a protective effect of thyroidectomy on the occurrence of eczema, which contrasts with previous reports that did not find an influence of thyroid function status on eczema incidence. This finding may offer new insights into the mechanisms underlying eczema development. Below, I will discuss our findings in the context of current research on the pathogenesis of eczema.

Eczema, also known as atopic dermatitis (AD), has seen a dramatic increase in prevalence in industrialized countries over the past 30 years, affecting 15–30% of children and 2–10% of adults [20]. Several potential AD-associated loci have been identified on chromosomes 3q21, 1q21, 16q, 17q25, 20p, and 3p26. Cytokines on these chromosomes regulate the synthesis of interleukin-4 (IL-4), IL-5, IL-12, IL-13, and granulocyte–macrophage colony-stimulating factor (GM-CSF), influencing the development of symptoms. Research has shown that skin inflammation is central to the pathogenesis of AD [21]. Lesions in AD patients’ skin predominantly feature T-cell infiltration, characterized by CD4 expression. This complex inflammation involves the activation of resident inflammatory dendritic cells, innate lymphoid cells, and Langerhans cells. The release of alarmins due to epidermal barrier disruption activates inflammatory dendritic cells and type 2-mediated responses. Activated Th2 cells release IL-4 and IL-13, promoting B-cell IgE class switching through the STAT pathway and producing antigen-specific IgE. The disrupted skin barrier with dissolved tight junctions in the granular layer allows dendrites to extend beyond this second barrier to sense and present antigens.

There have been numerous studies exploring the association between eczema and thyroid diseases [22]. An exploratory study by Brandon Smith et al. reported a significant association between AD and thyroid dysfunction in American adults, with a higher risk of thyroid diseases in younger individuals and males with AD [23]. This study’s findings, however, contradict the negative results of our research, possibly due to differences in the study populations—our study involved a Mendelian Randomization analysis of an all-age European ancestry cohort, while Smith’s study focused on an American population. Additionally, Smith’s study performed subgroup analyses of eczema concurrent with thyroid dysfunction, which our study could not replicate due to the limitations of the GWAS OPEN database. Nevertheless, our multivariable MR analysis revealed that after adjusting for hypothyroidism, the protective effect of thyroidectomy on eczema persisted, suggesting that hypothyroidism may be a risk factor for eczema, aligning with Smith’s conclusions. Smith’s research also suggested that the interrelationship between AD and thyroid diseases might result from immune dysregulation and overlapping cytokine pathways common to AD and autoimmune diseases. Tejas P. Joshi et al. conducted a case–control study indicating that AD patients were more likely to develop Graves’ disease (0.2% vs 0.04%, P < 0.001) and Hashimoto’s thyroiditis (2% vs 0.5%, P < 0.001), conditions representing hyperthyroidism and hypothyroidism, respectively [24]. These findings are inconsistent with our study, possibly due to the inclusion of other thyroid diseases beyond Graves’ disease and Hashimoto’s thyroiditis in our analysis or population differences. Given that skin inflammation is core to AD pathogenesis, a study by Xinyu Hu et al. based on the NHANES database demonstrated a negative correlation between systemic immune-inflammation index and thyroid function, potentially explaining the decreased systemic immune-inflammation following thyroidectomy-induced hypothyroidism and its effect on reducing eczema incidence [25].

In summary, while there seems to be a link between thyroid hormones and eczema, our study did not observe this relationship, but a protective effect of thyroidectomy on eczema was noted. This protective effect theoretically should relate to thyroid dysfunction, but whether and to what extent thyroid dysfunction explains this effect remains to be further studied. Future research should focus on the intrinsic connection between the thyroid gland and eczema, exploring pathways beyond thyroid hormones, such as neural or other endocrine axes (e.g., hypothalamic-pituitary-adrenal axis or hypothalamic-pituitary-gonadal axis), to better interpret these findings.

This study has several strengths: (1) employing bidirectional two-sample Mendelian Randomization, making the results more robust; (2) multivariable Mendelian Randomization analysis correcting for the impact of hypothyroidism, making the conclusions more scientific; (3) novel elucidation of the thyroid gland’s effect on eczema beyond thyroid hormones; (4) using public databases, avoiding resource wastage, and leveraging the advantages of Mendelian Randomization to minimize confounding factors, providing evidence comparable to randomized controlled trials; (5) exposure and outcome data from different original databases, avoiding the possibility of false positives due to sample overlap [26].

However, there are limitations: (1) the inability to analyze the impact of specific thyroid diseases on eczema incidence separately; (2) data limited to European ancestry, restricting the generalizability of the findings; (3) difficulty performing subgroup analysis on the specific extent of thyroidectomy due to database constraints; (4) lack of data from other original databases for cross-validation of thyroidectomy results.

Conclusion

Thyroidectomy is a protective factor against the occurrence of eczema, whereas thyroid function status did not show a significant causal relationship with eczema incidence. Additionally, eczema did not exhibit a clear impact on thyroid function status or thyroidectomy.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Y.Y. wrote the main manuscript text and prepared Figs. 1, 2 and 3. X.Z. prepared supplementary documents. All authors reviewed the manuscript.

Funding

No funding sources.

Data availability

No/Not applicable (this manuscript does not report data generation or analysis).

Declarations

Conflict of interest

The author(s) declare that they have no conflict of interest.

Ethical approval

The data for this study were obtained from the GWAS public database and the patients’ personal information was anonymised, so it does not require ethical committee approval as it does not involve personal privacy or informed consent.

Consent for publication

Not applicable.