Rheumatoid arthritis and hypothyroidism: a bidirectional Mendelian randomization study

Object Though significant correlations between rheumatoid arthritis (RA) and hypothyroidism have been found in earlier observational studies, their underlying causal relationship is still unknown. Mendelian randomization (MR) was used in the current study to assess the bidirectional causation between RA and hypothyroidism. Method We gathered summary data from genome-wide association studies (GWASs) of RA and hypothyroidism in people of European descent. Then, using data from the FinnGen consortium, we replicated our findings. Three approaches were employed to assess the causal link between RA and hypothyroidism: MR-Egger, weighted median (WM), and inverse variance weighted (IVW). The pleiotropy and heterogeneity were examined using a variety of techniques, including the MR-Egger intercept, the MR-PRESSO approach, the leave-one-out method, and the Cochran’s Q test. Results The study looked at a bidirectional incidental relationship between RA and hypothyroidism. The risk of hypothyroidism increased with RA (IVW odds ratio (OR) = 1.28, 95% confidence interval (CI) = 1.18–1.39, P = 8.30E-10), as did the risk of secondary hypothyroidism (IVW OR = 1.12, 95% CI = 1.05–1.21, P = 9.64E-4). The results of reverse MR analysis revealed that hypothyroidism (IVW OR = 1.68, 95% CI = 1.51–1.88, P = 4.87E-21) and secondary hypothyroidism (IVW OR = 1.74, 95% CI = 1.50–2.01, P = 1.91E-13) were linked to an increased risk of RA. Additionally, we obtain the same results in the duplicated datasets as well, which makes our results even more reliable. This study revealed no evidence of horizontal pleiotropy. Conclusion The present study established a bidirectional causal link between RA and hypothyroidism. However, it differs slightly from the findings of prior observational studies, suggesting that future research should concentrate on the interaction mechanisms between RA and hypothyroidism.


Introduction
Chronic and systemic autoimmune illness rheumatoid arthritis (RA) can cause ongoing inflammatory polyarthritis and joint damage, increases the likelihood of deformity or disability. A recent study determined that over the past 40 years, the prevalence of RA was nearly 0.46 percent worldwide and is expected to increase in the future (1). In addition, studies have found that RA may be linked to a number of autoimmune conditions, such as hypothyroidism. A frequent side effect of autoimmune rheumatic disease (ARD) is hypothyroidism. The main manifestations are musculoskeletal discomfort, fatigue, edema, and dryness. Hypothyroidism and rheumatoid arthritis have been linked in previous research.
RA and hypothyroidism are the most prevalent chronic specific autoimmune diseases. The etiology of these two diseases remains to be clarified, but they mainly involve genetic, environmental, and other factors, and the genes that lead to the pathogenesis of these two diseases still overlap. But that's not enough to explain the relationship. Previous research found that RA patients frequently have thyroid dysfunction, with a prevalence of 6% to 34% (2). Autoimmune diseases usually have similar pathological pathways, which means that there may be an association between RA and hypothyroidism (3). A recent study confirmed a significant link between rheumatoid arthritis and hypothyroidism (4). But some studies found no correlation (5). Although the association between RA and hypothyroidism has been actively studied for many years, there is still a lack of clear data to support this view. We speculate that RA and hypothyroidism may have similar pathogenesis and exist a relationship, so it is particularly necessary to determine the causal association between RA and hypothyroidism.
Limitations of traditional observational studies include confounding variables and reverse causality. Due to possible biases in confounding factors, these previous observational data may be limited in correlation inference, and the conclusions may even be disputed. A randomized controlled trial (RCT) may d e t e r m i n e t h e c a u s a l c o n n e c t i o n b e t w e e n R A a n d hypothyroidism, but it consumes a lot of money and time. Therefore, the Mendelian randomization (MR) method is used specifically for demonstrating whether the link between risk variables and outcomes can be explained by causal effects (6). MR study used single nucleotide polymorphisms (SNPs) as instrumental variables (IVs) to examine the causal relationship between exposure and outcome. It avoided confounding and reverse causality because genetic variants were identified at the time of conception. In the current study, we used a two-sample bidirectional MR analysis to determine if RA and hypothyroidism are causally related.

Study design
In the European population, to examine the association between RA and hypothyroidism, we used a two-sample MR study. During the MR analysis process, the effective instrumental variables (IVs) must satisfy three key assumptions in order to gain reliable results: 1) The IVs and exposure are tightly associated; 2) The IVs and any confounding factors that might impact both the exposure and the outcome are unrelated; and 3) The IVs only affect the outcome through the exposure (7). To validate each implicational direction, the study involved the following key steps: Genetic IVs associated with appropriate exposures were selected using multiple MR methods, pleiotropy assessment, and heterogeneity and sensitivity analyses. The most recent guidelines (STROBE-MR) (8) were adhered to in this study. Figure 1 shows a design for a bidirectional MR.

Data sources GWAS data for RA
To prevent the effects of population stratification, each SNP and its accompanying summary data were extracted from studies that covered only European ancestry. Firstly, summary data for RA (9) from a massive GWAS meta-analysis encompassing 58,284 individuals with European descent (14,

GWAS data for hypothyroidism
In the main analysis, from the GWAS ATLAS (https:// atlas.ctglab.nl/traitDB), we downloaded summary statistics for hypothyroidism (phenotype field: 20002_1226) and secondary hypothyroidism (phenotype field: 41204_E03), which contained 289,307 and 244,890 individuals from the UK Biobank, respectively (11). In addition, we used summary-level TSH and FT4 data from the Thyroid Omics Consortium because hypothyroidism and circulating TSH and FT4 levels are tightly related (12).
For replication purposes, we also utilized summaries of data from the FinnGen Consortium (https://r7.finngen.fi/) (13). In the FinnGen database (release R7, 2022), we selected hypothyroidism (congenital or acquired), hypothyroidism (levothyroxine purchases), and hypothyroidism (strict autoimmune). There are a total of 309,154 (173,746 females and 135,408 males) participants and 3,095 diseases in the database for the European population. The FinnGen project gives demographic information about the participants, such as their age and gender. The data set is available whenever it is needed. Details on the GWAS data sources used in our study are provided in Table 1.

Instrumental variable selection
As a first step toward verifying the first hypothesis, genome-wide significant (p <5×10 -8 ) SNPs were employed in this study as IVs to evaluate the causal association between RA and hypothyroidism. Second, to remove SNPs linked to significant linkage disequilibrium (LD), we employed a clumping approach with R2<0.001 and a window size of 10 Mb. We then manually searched the PhenoScanner database (http://www.phenoscanner.medschl.cam.ac.uk/phenoscanner) to exclude SNPs that potentially alter the result features. SNPs that were palindromes or ambiguous were eliminated. Additionally, our results were also adjusted for horizontal pleiotropy by applying the MR-Pleiotropy Residual Sum and Outlier (MR-PRESSO) method. We evaluated the strength of the IVs for the screened SNPs using Fstatistics and variance (R2) in order to avoid weak-tool bias (F>10): F = R 2 *(N-K-1)/K*(1-R 2 ) (14).

Mendelian randomization analysis
The MR-Egger (15), weighted median (WM) (16), and inverse variance weighted (IVW) (17) techniques were employed in this study to determine the cause-and-effect relationship between RA and hypothyroidism, with the IVW technique predominating. The IVW  Study design for our Mendelian randomization. RA, rheumatoid arthritis; IVs, instrumental variables; SNP, single-nucleotide polymorphism; WM, weighted median; IVW, inverse variance weighted.
method is our primary statistical analysis technique, although it could be skewed if IVs display horizontal pleiotropy. Due to this, to examine the reliability of our findings and potential pleiotropy, we conducted sensitivity analyses using the weighted median and MR-Egger regression approaches. The weighted median technique produces unbiased causal estimation if more than 50% of the SNPs are invalid genetic instruments. The MR-Egger regression method is able to provide estimates that have pleiotropy corrections applied to them.

Sensitivity analysis
Firstly, we employed the MR-Egger intercept to determine whether pleiotropy was present. The heterogeneity of IVW estimates was quantified using Cochran's Q test. Also, the MR-PRESSO test was run to check for outliers. If outliers are discovered, we will eliminate them and reevaluate the MR effect. In order to evaluate the robustness of the results, we apply the leave-one-out method to remove certain SNPs that have a negative impact on the research outcomes and then recalculate the results. The whole analysis was carried out in RStudio (version 4.2.1) with the "TwoSampleMR" (version 0.5.6) and "MRPRESSO" (version 1.0) packages.

Effects of rheumatoid arthritis on hypothyroidism
The MR results of RA on hypothyroidism are listed in Figure 2. After using strict criteria to rule out SNPs, we used 18 SNPs for hypothyroidism and 27 SNPs for secondary hypothyroidism. All of these IVs' F-statistics were greater than 10, indicating that there was no weak instrument bias.

Discussion
Consistent with the original study hypothesis, we used MR analysis in this work to find that there is indeed a two-way causal relationship between RA and hypothyroidism. This study used aggregated GWAS data to investigate the causal relationship between RA and hypothyroidism. To our knowledge, this is the first study to estimate the bidirectional causal inference between rheumatoid arthritis and hypothyroidism using magnetic resonance techniques. Compared with previous studies, this study not only found a correlation between RA and hypothyroidism but also provided a reference for the early routine examination of hypothyroidism in future clinical RA patients.
As an autoimmune ailment, RA increases a patient's risk of developing thyroid issues, notably hypothyroidism (18). 76% of people with RA had one or more extra-articular features, according to a comprehensive investigation. In 6% to 33% of RA patients, thyroid hormone abnormalities and/or autoimmune thyroid illness were detected (19). Hypothyroidism is the most common clinical sign of thyroid hormone deficiency. To diagnose hypothyroidism, the blood TSH and FT4 levels are checked in laboratories. It's been disputed for a long time (20) whether thyroid dysfunction and rheumatic disease are related (20). Many authors have suggested that hypothyroidism could make rheumatoid arthritis worse (21,22). The prevalence of thyroid dysfunction was found to be considerably greater in RA patients than controls, according to a case-control research, which is in line with our findings (OR = 2.89, P < 0.001). Additional subgroup analysis found RA and hypothyroidism had a positive connection (OR = 2.28, P = 0.006) (3). Population studies have indicated that autoimmune thyroid illness is more prevalent in RA patients and vice versa (23,24). An increase in the incidence of hypothyroidism was observed in RA patients, but not an increase in FT4, according to a case-control study. Instead, there is a decline, which is in line with our study's negative findings regarding FT4 and RA (25). One recent metaanalysis found that RA patients had a threefold higher risk of developing thyroid autoantibodies than healthy controls  (26). Using MR Analysis, our study also found a positive correlation between RA and hypothyroidism, and the correlation between the two was significant. In people with RA, thyroid dysfunction has unknown underlying causes (27). It is now widely acknowledged that RA and hypothyroidism can result from a variety of causes, including genetic predisposition and environmental variables. One of the key processes affecting RA and hypothyroidism is genetics. Interestingly, our study found a substantial relationship between elevated RA and genetic risk factors for hypothyroidism in the analysis of genetic variables. According to a study, people with RA are more likely than people with other types of thyroid disease to experience hypothyroidism. A population-based retrospective cohort study that looked into the risk of hypothyroidism in RA patients discovered that the incidence of the condition was 1.74 times higher in the RA group than in controls, with an adjusted hazard ratio of 1.67 (95% CI =1.39-2.00) according to the Cox method after adjusting for covariates. In all categories, the incidence of women was 3.6 times higher than that of men, making up over 75% of the study population (28). The above studies indicate that RA is significantly correlated with hypothyroidism, which is consistent with our findings, but the specific influencing factors and pathogenesis still need to be further studied. Patients with hypothyroidism frequently experience various rheumatic symptoms, including the possibility of fatigue, myalgia, and joint pain in RA (21,29). Numerous investigations have discovered that increased disease activity in RA patients and blood tumor necrosis factor-a levels have an effect on hypothyroidism (4,30). In light of the aforementioned research, it is clear that early thyroid function testing and active therapy have significant clinical implications (31). Additionally, studies have shown a correlation between the disease activity score -28 (DAS-28) and hypothyroidism in RA patients. Hypothyroidism may increase disease activity due to an increase in joint pain, so RA patients should be aware of thyroid function screening for early detection and better treatment (32). One of the most prevalent comorbidities of RA is hypothyroidism; however, the shared pathophysiology between the two conditions is still unknown, and genes implicated in the control of T cell responses may be important (33). Surprisingly, our study found that hypothyroidism is closely related to the increased risk of RA genetically, which may provide new ideas and directions for our research on the mechanism. In addition, Autoimmunity is a factor in both RA and hypothyroidism. Hypothyroidism is associated with an inflammatory response and pro-inflammatory cytokines such as IL-1 and IL-6 (34, 35). According to studies, there is a relationship between autoimmune diseases and the pathogenesis of RA, as patients with hypothyroidism in their family have a higher risk of developing RA (36). Thyroglobulin antibodies (TgAb) and thyroid peroxidase antibodies (TPOAb) with positive results have been found in RA patients (37). Additionally, inconsistent results have been found in research on the possibility of thyroid malfunction in people with RA (38). To sum up, this study found that RA and hypothyroidism are related to the pathogenesis of genetics, genes and autoimmunity, especially the significant correlation between genetic variables. Although no relationship was found with environmental variables, this study will continue to further explore and study the relationship between the pathogenesis of the two in the future.
The advantage of the bidirectional approach is that it enables the inference of both causal axes between RA and hypothyroidism. The MR is built using the publicly available GWAS summary statistics. Without incurring additional fees for experiments, it is practical for mining reliable genetic data. Second, we validated our findings using extra datasets to make our conclusions more credible. Another benefit is that, for the first time, we utilized MR to find out Mendelian randomization for the association of hypothyroidism on RA. RA, rheumatoid arthritis; FT4, free T4; TSH, thyroid-stimulating hormone; WM, weighted median; IVW, inverse variance weighted; nSNPs, number of SNPs used in MR; OR, odds ratio; CI, confidence interval.
if RA susceptibility is linked to a higher risk of thyroid disease. The findings of this study indicate that RA is related to an elevated risk of hypothyroidism.
There are also more limitations to our study. Firstly, the samples for RA and hypothyroidism were quite small, which may have caused some false positives and negatives. The causal relationship between RA and hypothyroidism was not stratified by age or gender, which is the second flaw. Thirdly, the exposure and outcome analyses included only Europeans in the study population. It has to be determined if this outcome might be repeated in other populations. Some heterogeneity existed in the study. As a result, the conclusions are less credible because we were unable to reduce the heterogeneity. Other ethnic groups also need to have MR studies conducted, and these studies must be supported by extensive GWAS data.
Finally, RA and hypothyroidism shared a bidirectional causal connection. Our findings suggest that RA patients should routinely be checked for hypothyroidism. At present, there are few studies on RA and hypothyroidism, so it is suggested that future studies focus on the mechanisms between them.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Ethics statement
Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent from the participants' legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.
Author contributions DC, YS: conception, technique, formal analysis, data collection, first draft preparation, and visualization; LD, SY: preparation of the first draft preparation, visualization, and data curation in writing; SD, YF: editing, monitoring, and writing evaluation. The final draft of the manuscript was approved by all writers, who also made contributions to the work.