Osteoarthritis and autoimmune thyroid disease: A Mendelian randomization study

doi:10.21203/rs.3.rs-4366584/v1

Download PDF

Research Article

Osteoarthritis and autoimmune thyroid disease: A Mendelian randomization study

https://doi.org/10.21203/rs.3.rs-4366584/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Objective:

Previous epidemiological observational researches have suggested a potential connection between autoimmune thyroid disease (AITD) and osteoarthritis (OA); however, the results are inconclusive. It remains unclear whether the association between AITD and OA is causal. This Mendelian randomization (MR) study was conducted to specifically explore whether there is a genetic causal link between OA and AITD disease.

Methods:

Derived from a comprehensive genome-wide association study involving individuals of European descent, our study carefully chose specific single nucleotide polymorphisms (SNPs) as instrumental variables for two distinct forms of KOA and HOA. We evaluated two types of AITD: autoimmune hypothyroidism and autoimmune hyperthyroidism. MR Egger, Weighted median, Simple mode, Weighted mode and Inverse Variance Weighting (IVW) were used in our study to estimate whether there was a genetic level of causality between AITD and OA, and the IVW method prevailed in the assessment. Additionally, we conducted tests for heterogeneity and sensitivity to ensure the robustness of our findings.

Results:

IVW analysis revealed a significant impact of autoimmune hyperthyroidism on the incidence of KOA. (OR: 1.0512; 95% CI: 1.0197–1.0836, P = 0.001285396). Other results had a P-value > 0.05, showing all negative.

Conclusions

Our findings indicated a genetic causal association between autoimmune hyperthyroidism and the risk of KOA, despite observational studies reporting an association between autoimmune hypothyroidism and OA and provided new insight into the treatment strategies for OA and AITD.

Mendelian randomization

Knee osteoarthritis

Hip osteoarthritis

Autoimmune thyroid disease

OA is a prevalent orthopedic disease that occurs most often in older women and is a major reason of joint impairment and functional limitations [1], and healthcare spending [2]. As the global population ages, obesity, and other osteoarthritis-causing factors on the rise, the incidence of OA is also increasing, and research indicates that a staggering 250 million people worldwide are grappling with this condition [3].However, osteoarthritis of the knee is also very different from osteoarthritis of the hip. From an epidemiological perspective, Symptomatic osteoarthritis confirmed through radiography is more prevalent in the knee compared to the hip, it impacts around 10% of individuals aged 45 years and above, compared to 16% for the knee[4, 5]. However, estimates generated by the Global Burden of Disease initiative suggest a notably greater contrast in the prevalence of hip and knee OA worldwide [6]. According to a meta-analysis, the occurrence of HOA is consistent across genders, while KOA affects a higher proportion of women than men [7]. In Chinese women, HOA is less prevalent compared to their US White counterparts, whereas KOA is more prevalent in Chinese women compared to women of US White ethnicity [8]. Unlike KOA, there doesn't seem to be a higher prevalence of HOA among African Americans compared to White individuals. However, there may be racial differences in specific features such as osteophyte formation [4, 5, 8].Furthermore, KOA and HOA differ significantly in genes[9], gene expression[10], molecular pathophysiology[11, 12] and cytokine expression[13, 14].

AITD, including Graves' disease and Hashimoto's thyroiditis, is a disease in which autoantibodies attack one's own tissues. Autoantibodies include antibodies such as TPOAb, Tgab, TRAb, TSAb and TSBAb. TRAb is associated with Graves' disease, and TPOAb and TgAb are associated with Hashimoto's thyroiditis. The clinical manifestations of these disorders often exhibit significant overlap. Graves’ disease is characterized by the production of antibodies targeting the thyrotropin receptor, activation of thyroid follicular cells, modest lymphocytic infiltration of the thyroid gland, and resultant hyperthyroidism. In contrast, individuals with Hashimoto's thyroiditis may manifest either hyperthyroidism or hypothyroidism, with hypothyroidism being the predominant characteristic[15].

While previous several studies have identified a connection between AITD and OA [16], it remains unconfirmed whether a causal relationship exists between these two conditions. Several previous studies have established links between various orthopedic disorders and thyroid dysfunction. For example, Tagoe et al[17]. reported a heightened likelihood of chondrocalcification in individuals with elevated antithyroid peroxidase antibody (TPOAb) levels. Cellular studies support the idea that abnormal thyroid hormone signaling increases susceptibility to some degenerative orthopedic diseases [18, 19]. Additional research has shown that thyroid hormones can influence the functionality of both osteoblasts and osteoclasts [20]. But some studies have also concluded that the functional status of the thyroid gland is not associated with osteoarthritis[17, 21].

MR analysis is a technique that uses genetic variation linked to a particular exposure to investigate the causal impacts of that exposure on health, social, and economic outcomes [22]. This approach helps mitigate confounding factors and addresses issues of reverse causality often encountered in traditional observational epidemiological studies. MR is adopted to enhance result reliability by minimizing bias when evaluating the association between risk factors and prognosis of a disease [23]. In our study, the causal effect of exposure on outcomes was estimated using single nucleotide polymorphisms (SNPs) associated with exposure-related genes, subject to specific assumptions [24]. Two-sample MR analysis necessitates summary-level data from two independent genome-wide association studies (GWASs) to identify potential exposures and outcomes. By eliminating confounders, the reliability of the results obtained through MR analysis can be enhanced [25]. By searching the available literatures, we found that this was the first two-sample MR study to determine the causal relationship between AITD and OA.

Study overview

On the basis of our MR analysis, we chose KOA and HOA for our research, which is the outcome in MR analysis. To represent AITD, we specifically chose autoimmune hyperthyroidism and autoimmune hypothyroidism, which are the exposure in MR analysis. According to our MR method analysis, SNPs closely associated with autoimmune hyperthyroidism and autoimmune hypothyroidism were chosen as instrumental variables. To mitigate potential bias stemming from the intersection of samples between exposure and outcome, we meticulously sourced exposure and outcome data from separate databases. In the selection of exposure and outcome GWAS datasets, we considered several factors, including the population of European descent, a more extensive range of disease types, larger population sample sizes and SNPS, and the timing of data publication. Figure 1 illustrates the key assumptions of the MR study.

Key assumptions of the Mendelian randomization study: Assumption 1: Instrumental variables should demonstrate a sturdy and consistent connection with the exposure. Assumption 2: Instrumental variables should show no association with any confounders Assumption 3: Instrumental variables should have no association with HOA and KOA except through their connection with the exposure.

GWAS data for KOA and HOA

The data for KOA and HOA were obtained from the Integrative Epidemiology Unit (IEU) GWAS database (https://gwas.mrcieu.ac.uk). The data for KOA has a sample sizes of 393873 and the data for HOA has a sample sizes of 403124.The original dataset is available from the research conducted by Tachmazidou et al [26]. The data exclusively comprised person of European ancestry.

GWAS data for autoimmune hyperthyroidism and autoimmune hypothyroidism

The GWAS datasets related to AITD were constructed by the FinnGen Biobank. The autoimmune hyperthyroidism GWAS dataset included 173938 samples and 16380189 SNPs with the ID number of “finn-b-AUTOIMMUNE_HYPERTHYROIDISM”, The autoimmune hypothyroidism GWAS dataset was comprised of 198472 samples and 16380353 SNPs with the ID numbers of “finn-b-E4_HYTHY_AI_STRICT”(https://gwas.mrcieu.ac.uk/), and the data collection exclusively involved individuals of European descent.

Ethics statement

The GWAS datasets used in our investigation have been publicly released, and the necessary ethical approval has been obtained. This study did not require separate ethical approval.

Mendelian randomization analysis

In our investigation, the MR analysis for each step involved the following procedures: First, we identified independent SNPs strongly associated with the exposure. We applied stringent filtration criteria (P < 5x10^− 8) within a 10000 kb window around a lead SNP, ensuring low connection with other SNPs in the region (linkage disequilibrium, r2 < 0.001). Subsequently, we cross-referenced these SNPs in PhenoScanner (www.phenoscanner.medschl.cam.ac.uk) and excluded SNPs associated with confounding factors and outcomes. Confounders primarily denote other potential risk factors in addition to the specific exposure of interest that may contribute to the observed outcome. Common risk factors associated with OA include but are not restricted to factors such as smoking, alcohol consumption, and obesity [27]. Second, extract SNPs effects from the outcome GWAS dataset. Next, we acquired valid IVs by harmonizing the effects of exposure and outcome and eliminating SNPs with F-statistics < 10 or those failing to harmonize, which encompassed SNPs with incompatible alleles and displayed palindromic patterns with intermediate allele frequencies. Then, the subsequent MR evaluation of valid IVs involved the application of multiplicative random-effects-inverse variance weighted (IVW), weighted mode (WM), weighted median (WME), and MR-Egger regression methods. Lastly, we performed sensitivity analysis and visually presented the MR results. The study flowchart is depicted in Fig. 2.

Flowchart of MR analysis in this study.

In the MR analysis, AITD were considered as the exposures, while OA was designated as the outcome. We identified 7 strongly correlated SNPs from the GWAS dataset of autoimmune hyperthyroidism and 58 SNPs from the GWAS dataset of autoimmune hypothyroidism. (Supplementary Table 1 and Supplementary Table 2)

Among 7 SNPs effects extracted from the OA GWAS for the association of autoimmune hyperthyroidism on OA, rs9265890 was excluded due to its inability to harmonize, and rs9275576 was removed for associating with confounders. (Supplementary Table 3). Rs7310615 exhibited associations with confounding variables and excluded from the 58 SNPs. 16 SNPs failed the harmonization process and were consequently excluded from the 58 SNPs extracted from the OA GWAS concerning the relationship between Autoimmune hypothyroidism and OA, including rs10223666, rs12897126, rs13293763, rs17786733, rs1964690, rs1993945, rs2983511, rs3132487, rs3775291, rs4549506, rs707937, rs7090504, rs7310615, rs7754251, rs9296422, rs9842232 and rs61938963 was removed for associating with confounders. (Supplementary Table 4). Autoimmune hypothyroidism was considered as the exposures, while HOA was designated as the outcome, rs9277542 was excluded from the MR-presso process. (Supplementary Table 5)

Statistical analysis

The statistical analyses for our research were conducted using R version 4.3.1 with RStudio, and the TwoSampleMR package was employed within the RStudio environment. In the coordination process, the F statistic plays a key role in estimating the valid IV. SNPs with an F statistic > 10 were selected as valid instrumental variables. Various formulas are commonly employed to compute the F-statistic, such as \(F=（N-k-1)÷k\times R2÷(1-R2)\)or \(F={b}^{2}÷{se}^{2}\).In our study, we used the formula \(F={b}^{2}÷{se}^{2}\) ( where b represents the effect size of the exposure and se represents the standard error of the exposure), and we used the formula with reference to the univariate and multivariate studies done by Chen et al [28]. Only valid IVs with F-statistics greater than 10 were considered highly correlated and were used in subsequent MR estimation, sensitivity analysis, and visualization. Sensitivity analyses used heterogeneity tests, multiple validity tests, and leave-one-out tests to ensure stability and reliability. Cochran's Q statistic and I2 statistic were used to assess the Heterogeneity [29]. Causality was assessed based on the IVW results if there was heterogeneity in the results (p < 0.05), which was not the case for the positive results from our study [30]. During the pleiotropy test, intercepts computed through MR-Egger regression were employed to assess the horizontal pleiotropy of valid IVs. A greater deviation of intercepts from zero indicated a higher likelihood of horizontal pleiotropy (P < 0.05) [31]. In the leave-one-out test, each valid IV was systematically removed, and IVW was used to assess whether the exclusion significantly altered the MR results. Consistency between the distribution of the "ALL" line and the MR estimates indicated stable and reliable results [32].Presentation of MR results employed scatter plots, forest plots, and funnel plots in our study.

3.1. Mendelian randomization analysis

IVW analysis and WME analysis demonstrated such a result, when autoimmune hyperthyroidism is the exposure and KOA is the outcome, there is a positive causal relationship between the two (P < 0.05, OR > 1) (Table 1).

Overall, the results of our statistical analyses are mainly referenced in the IVW analysis: There was a positive causal relationship between Autoimmune hyperthyroidism and KOA. Figure 2 displays the outcomes of the different statistical approaches.

3.2. Sensitivity analysis

We performed sensitivity analysis to confirm the authenticity of our study’s results. Both the IVW and MR-Egger tests for heterogeneity indicated no observed heterogeneity between autoimmune hyperthyroidism and KOA (p > 0.05) (Table 1). Results from MR-Egger regression Suggested the lack of pleiotropy in the Mendelian randomization analysis results (p > 0.05) (Table 1). To assess the influence of individual SNPs, we implemented a leave-one-out method as part of the sensitivity analysis. The leave-one-out method revealed that no SNPs were present in the instrumental variables for the particular connection of autoimmune hyperthyroidism with KOA. (Supplementary Fig. 13).

Table 1

Mendelian randomization estimates of autoimmune hyperthyroidism and autoimmune hypothyroidism on KOA and HOA.
Exposure	Outcomes	IVW					MR-Egger				F-statistic
		OR	95%CI	P-value	Q-value	Heterogeneity P-value	OR	95%CI	P-value	Pleiotropy P-value
Autoimmune hyperthyroidism	Knee OA	1.051	1.020–1.084	0.001	1.292	0.863	1.034	0.890–1.201	0.693	0.838	41.9
Autoimmune hyperthyroidism	Hip OA	0.972	0.935–1.010	0.140	2.173	0.704	0.999	0.828–1.205	0.989	0.790	41.9
Autoimmune hypothyroidism	Knee OA	1.033	0.994–1.072	0.096	76.249	0.001	1.037	0.952–1.130	0.408	0.912	84
Autoimmune hypothyroidism	Hip OA	0.978	0.939–1.018	0.274	51.388	0.088	0.964	0.881–1.054	0.423	0.721	84.1

3.3. Visualization of MR

The scatter plots illustrating the estimated effect sizes of exposure (autoimmune hyperthyroidism and autoimmune hypothyroidism) on the outcome (HOA) can be found in Supplementary Fig. 1–4. Supplementary Fig. 5–8 depict the funnel diagram, while Supplementary Fig. 9–12 present the forest plot.

Our MR analysis concluded a positive causal correlation between autoimmune hyperthyroidism and KOA, while no significant correlation was found with HOA. Conversely, no genetic causality was found between autoimmune hypothyroidism and knee osteoarthritis or hip osteoarthritis.

This study marks the pioneering endeavor to investigate the genetic causality between autoimmune hyperthyroidism and OA. Interestingly, our findings align with a recent comprehensive review indicating that elevated levels of FT3 and FT4 are linked to an increased risk of OA, consistent with our MR results [33]. And previously there is a genetic relationship between thyroid function status and osteoarthritis, it has been shown that DIO2 is a gene that can make inactive T4 to active T3, and upregulation of DIO2 can promote the increase of T3([34] [35]), and in the cartilage, the intracellular bioavailability of T3 stimulates the differentiation of chondrocytes, inhibits the proliferation of chondrocytes and induces chondrocytes to hypertrophy, and initiates terminal differentiation and calcification. Chondrocyte hypertrophy impairs cartilage viability by altering the expression of bone-specific collagen, which triggers calcification of the cartilage matrix and upregulation of cartilage-specific protein hydrolases. It is well known that most of the cartilage matrix consists of collagen and proteoglycans, and that chondrocyte-specific protein hydrolases hydrolyze collagen and proteoglycans, which in turn damage articular cartilage, leading to osteoarthritis༈[34] [36] [37]༉, which is also elevated by autoimmune hyperthyroidism T3, which likewise leads to osteoarthritis by the same mechanism. And further research is needed as to why it leads to KOA but not HOA. Although many epidemiologic studies had found an association between arthritis incidence and hypothyroidism[15], our study's genetic perspective did not establish a causality between autoimmune hypothyroidism and OA. However, this does not exclude the possibility of other non-genetic factors potentially influencing the association between autoimmune hypothyroidism and OA. Several studies have proposed that the reduced TSH in patients with autoimmune hypothyroidism may be a contributing factor to knee osteoarthritis [15].

Our study has several strengths. First, in contrast to observational research, we employed MR for causal inference, utilizing genetic variations as IVs. This approach mitigates the influence of confounding factors and reverse causality, enhancing the robustness of our findings. Second, the exclusive inclusion of samples from European populations bolsters the internal consistency of our results by minimizing the impact of demographic variations. Most notably, we unveil, for the first time, a causality between AITD and OA in genetical, potentially opening new avenues for OA treatment.

On the other hand, we should also recognize certain limitations. First, we did not use SNP data for various thyroid hormone subtypes (e.g., T3, T4) and thyroid autoantibodies (e.g., TPOAb and TgAb) due to the study conditions. Therefore, we were unable to determine whether thyroid hormone subtypes or thyroid autoantibodies caused osteoarthritis of the knee. Secondly our study concluded that AITD is not genetically correlated with hip osteoarthritis, and due to conditioning we did not use thyroid autoantibodies and thyroid hormone subtypes as exposures to determine if there was a correlation between them and hip osteoarthritis. Third, this MR study only examined the relationship between AITD and OA; the relationship between other thyroid diseases and OA will be discussed in further studies. Fourth, because all participants in our study were of European descent, it remains uncertain whether the conclusions can be generalized to other population groups.

Moreover, although some of the methods used in our study yielded results that deviated from the IVW method, we dominated the results derived from IVW. Nevertheless, all SNPs used in our MR analysis followed the assumption of valid instrumental variables, and the MR-presso findings supported the IVW outcomes., which affirms the reliability of our findings.

Our MR study demonstrated a correlation between autoimmune hyperthyroidism and the development of KOA from a genetic point of view, providing a basis for the treatment and prevention of knee osteoarthritis.

Data availability statement

Publicly available datasets can be found here: the Integrative Epidemiology Unit (IEU) GWAS database (https://gwas.mrcieu.ac.uk) and FinnGen Biobank.

Conflict of interest

The authors declare that they have no competing interests.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Funding

This research was supported by the Sponsored by Tianjin Health Research Project (TJSQNYXXR-D2-132), the Exceptional Young Talents Fostering Foundation 2021 of the Tianjin Fourth Central Hospital(tjdszxyy20210011)and Natural Science Foundation of Tianjin (21JCYBJC00280).

Author Contribution

ZZ and YZ designed the study. ZZ, YZ and QY performed data analysis and drafted the manuscript. YZ, HX, JD and HC revised the manuscript. YZ and HX collected the data. JD and HC provided the resources. ZZ leads the study. All authors contributed to the article and approved the submitted version.

Data Availability

Data availability statementPublicly available datasets can be found here: the Integrative Epidemiology Unit (IEU) GWAS database (https://gwas.mrcieu.ac.uk) and FinnGen Biobank. Furthermore, our selected data is provided within the manuscript or supplementary information files

James RJE, Walsh DA. Trajectories of pain predict disabilities affecting daily living in arthritis. Br J Health Psychol. 2019. https://doi.org/10.1111/bjhp.12364.
Dieleman JL, Cao J. JAMA. 2020. https://doi.org/10.1001/jama.2020.0734. US Health Care Spending by Payer and Health Condition, 1996–2016.
Hunter DJ, Bierma-Zeinstra S. Osteoarthritis. Lancet (London, England). 2019; https://doi.org/10.1016/s0140-6736(19)30417-9.
Jordan JM, Helmick CG. Prevalence of knee symptoms and radiographic and symptomatic knee osteoarthritis in African Americans and Caucasians: the Johnston County Osteoarthritis Project. J Rhuematol. 2007.
Jordan JM, Helmick CG. Prevalence of hip symptoms and radiographic and symptomatic hip osteoarthritis in African Americans and Caucasians: the Johnston County Osteoarthritis Project. J Rhuematol. 2009. https://doi.org/10.3899/jrheum.080677.
Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet (London, England). 2020; https://doi.org/10.1016/s0140-6736(20)30925-9.
Srikanth VK, Fryer JL. A meta-analysis of sex differences prevalence, incidence and severity of osteoarthritis. Osteoarthr Cartil. 2005. https://doi.org/10.1016/j.joca.2005.04.014.
Allen KD. Racial and ethnic disparities in osteoarthritis phenotypes. Curr Opin Rheumatol. 2010. https://doi.org/10.1097/BOR.0b013e32833b1b6f.
den Hollander W, Ramos YF. Knee and hip articular cartilage have distinct epigenomic landscapes: implications for future cartilage regeneration approaches. Ann Rheum Dis. 2014. https://doi.org/10.1136/annrheumdis-2014-205980.
Coutinho de Almeida R, Ramos YFM. RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage. Ann Rheum Dis. 2019. https://doi.org/10.1136/annrheumdis-2018-213882.
Alexopoulos LG, Youn I. Developmental and osteoarthritic changes in Col6a1-knockout mice: biomechanics of type VI collagen in the cartilage pericellular matrix. Arthritis Rheum. 2009. https://doi.org/10.1002/art.24293.
Christensen SE, Coles JM. Altered trabecular bone structure and delayed cartilage degeneration in the knees of collagen VI null mice. PLoS ONE. 2012. https://doi.org/10.1371/journal.pone.0033397.
Ren G, Lutz I. Serum and synovial fluid cytokine profiling in hip osteoarthritis: distinct from knee osteoarthritis and correlated with pain. BMC Musculoskelet Disord. 2018. https://doi.org/10.1186/s12891-018-1955-4.
Grieshaber-Bouyer R, Kämmerer T. Divergent Mononuclear Cell Participation and Cytokine Release Profiles Define Hip and Knee Osteoarthritis. J Clin Med. 2019. https://doi.org/10.3390/jcm8101631.
Tagoe CE, Zezon A. Rheumatic manifestations of autoimmune thyroid disease: the other autoimmune disease. J Rhuematol. 2012. https://doi.org/10.3899/jrheum.120022.
Gillan MM, Scofield RH. Hashimoto's thyroiditis presenting as bilateral knee arthropathy. J Okla State Med Assoc. 2002.
Tagoe CE, Wang W. Association of anti-thyroid antibodies with radiographic knee osteoarthritis and chondrocalcinosis: a NHANES III study. Therapeutic Adv Musculoskelet disease. 2021. https://doi.org/10.1177/1759720x211035199.
Bassett JH, Williams GR. Role of Thyroid Hormones in Skeletal Development and Bone Maintenance. Endocr Rev. 2016. https://doi.org/10.1210/er.2015-1106.
Nehls V. [Osteoarthropathies and Myopathies associated with Disorders of the Thyroid Endocrine System]. Deutsche medizinische Wochenschrift (1946). 2018; https://doi.org/10.1055/s-0043-121381.
Williams GR. Thyroid hormone actions in cartilage and bone. Eur thyroid J. 2013. https://doi.org/10.1159/000345548.
Chaisson CE, McAlindon TE. Lack of association between thyroid status and chondrocalcinosis or osteoarthritis: the Framingham Osteoarthritis Study. J Rhuematol. 1996.
Richmond RC, Davey Smith G. Mendelian Randomization: Concepts and Scope. Cold Spring Harbor perspectives in medicine. 2022; https://doi.org/10.1101/cshperspect.a040501.
Emdin CA, Khera AV, Mendelian Randomization. JAMA. 2017. https://doi.org/10.1001/jama.2017.17219.
Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014. https://doi.org/10.1093/hmg/ddu328.
Hartwig FP, Davies NM. Two-sample Mendelian randomization: avoiding the downsides of a powerful, widely applicable but potentially fallible technique. Int J Epidemiol. 2016. https://doi.org/10.1093/ije/dyx028.
Tachmazidou I, Hatzikotoulas K. Identification of new therapeutic targets for osteoarthritis through genome-wide analyses of UK Biobank data. Nat Genet. 2019. https://doi.org/10.1038/s41588-018-0327-1.
Felson DT, Lawrence RC. Osteoarthritis: new insights. Part 1: the disease and its risk factors. Ann Intern Med. 2000. https://doi.org/10.7326/0003-4819-133-8-200010170-00016.
Chen L, Yang H. Insights into modifiable risk factors of cholelithiasis: A Mendelian randomization study. Hepatology (Baltimore MD). 2022. https://doi.org/10.1002/hep.32183.
Huedo-Medina TB, Sánchez-Meca J. Assessing heterogeneity in meta-analysis: Q statistic or I2 index? Psychological methods. 2006; https://doi.org/10.1037/1082-989x.11.2.193.
Gill D. Heterogeneity Between Genetic Variants as a Proxy for Pleiotropy in Mendelian Randomization. JAMA Cardiol. 2020. https://doi.org/10.1001/jamacardio.2019.4281.
Verbanck M, Chen CY. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018. https://doi.org/10.1038/s41588-018-0099-7.
Burgess S, Bowden J, Epidemiology. (Cambridge, Mass.). 2017; https://doi.org/10.1097/ede.0000000000000559.
Chen S, Sun X. Association between sensitivity to thyroid hormone indices and the risk of osteoarthritis: an NHANES study. Eur J Med Res. 2022. https://doi.org/10.1186/s40001-022-00749-1.
Meulenbelt I, Min JL. Identification of DIO2 as a new susceptibility locus for symptomatic osteoarthritis. Hum Mol Genet. 2008. https://doi.org/10.1093/hmg/ddn082.
Valdes AM, Spector TD. The clinical relevance of genetic susceptibility to osteoarthritis. Best practice & research. Clinical rheumatology. 2010; https://doi.org/10.1016/j.berh.2009.08.005.
Bay-Jensen AC, Slagboom E. Role of hormones in cartilage and joint metabolism: understanding an unhealthy metabolic phenotype in osteoarthritis. Menopause (New York, N.Y.). 2013; https://doi.org/10.1097/GME.0b013e3182745993.
Bassett JH, Williams GR. The molecular actions of thyroid hormone in bone. Trends in endocrinology and metabolism: TEM. 2003; https://doi.org/10.1016/s1043-2760(03)00144-9.

No competing interests reported.

Download PDF

Editor assigned by journal
06 May, 2024
Submission checks completed at journal
06 May, 2024
First submitted to journal
03 May, 2024

You are reading this latest preprint version

Osteoarthritis and autoimmune thyroid disease: A Mendelian randomization study

Status:

Version 1

Abstract

Objective:

Methods:

Results:

Conclusions

Figures

Introduction

Methods

Study overview

GWAS data for KOA and HOA

GWAS data for autoimmune hyperthyroidism and autoimmune hypothyroidism

Ethics statement

Mendelian randomization analysis

Statistical analysis

Results

3.1. Mendelian randomization analysis

3.2. Sensitivity analysis

3.3. Visualization of MR

Discussion

Conclusion

Data availability statement

Declarations

Conflict of interest

Publisher’s note

Funding

Author Contribution

Data Availability

References

Additional Declarations

Supplementary Files

Status:

Version 1