Available via license: CC BY 4.0
Content may be subject to copyright.
Page 1/15
Autoimmune Thyroid Disease and Myasthenia
Gravis: A study bidirectional Mendelian
randomization
suijian Wang
First Aliated Hospital of Shantou University Medical College
Shaoda Lin ( Shaoda.2023@outlook.com )
First Aliated Hospital of Shantou University Medical College
Xiaohong Chen
First Aliated Hospital of Shantou University Medical College
Daiyun Chen
First Aliated Hospital of Shantou University Medical College
Research Article
Keywords: Autoimmune Thyroid Disease, Graves disease, hypothyroidism, Mendelian randomization,
GWAS
Posted Date: October 14th, 2023
DOI: https://doi.org/10.21203/rs.3.rs-3427396/v1
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
Additional Declarations: No competing interests reported.
Page 2/15
Abstract
Background
Previous studies have suggested a potential association between AITD and MG, but the evidence is
limited and controversial, and the exact causal relationship remains uncertain.
Objective
Therefore, we employed a Mendelian randomization (MR) analysis to investigate the causal relationship
between AITD and MG.
Methods
To explore the interplay between AITD and MG, We conducted MR studies utilizing GWAS-based
summary statistics in the European ancestry.Several techniques were used to ensure the stability of the
causal effect, such as random-effect inverse variance weighted, weighted median, MR-Egger regression,
and MR-PRESSO. Heterogeneity was evaluated by calculating Cochran's Q value. Moreover, the presence
of horizontal pleiotropy was investigated through MR-Egger regression and MR-PRESSO
Results
The IVW method indicates a causal relationship between both GD(OR 1.31,95%CI 1.08 to 1.60,P = 0.005)
and autoimmune hypothyroidism (OR: 1.26, 95% CI: 1.08 to 1.47, P = 0.002) with MG. However, there is no
association found between FT4(OR 0.88,95%CI 0.65 to 1.18,P = 0.406), TPOAb(OR: 1.34, 95% CI: 0.86 to
2.07, P = 0.186), TSH(OR: 0.97, 95% CI: 0.77 to 1.23, P = 0.846), and MG. The reverse MR analysis reveals
a causal relationship between MG and GD(OR: 1.50, 95% CI: 1.14 to 1.98, P = 3.57e-3), with stable results.
On the other hand, there is a signicant association with autoimmune hypothyroidism(OR: 1.29, 95% CI:
1.04 to 1.59, P = 0.019), but it is considered unstable due to the inuence of horizontal pleiotropy (MR
PRESSO Distortion Test P < 0.001). MG has a higher prevalence of TPOAb(OR: 1.84, 95% CI: 1.39 to 2.42,
P = 1.47e-5) positivity and may be linked to elevated TSH levels(Beta:0.08,95% CI:0.01 to 0.14,P = 0.011),
while there is no correlation between MG and FT4(Beta:-9.03e-3,95% CI:-0.07 to 0.05,P = 0.796).
Conclusion
AITD patients are more susceptible to developing MG, and MG patients also have a higher incidence of
GD.
1.Introduction
Page 3/15
Autoimmune thyroid disorders (AITD) emerge from an immune system malfunction, giving rise to an
immune onslaught against the thyroid gland(1).AITD stand as the most prevalent autoimmune disorders
and hold the position of being the most frequently observed pathological conditions of the thyroid
gland(2). This category encompasses two major clinical manifestations: Graves' disease (GD) and
Hashimoto's thyroiditis (HT), both of which share a common characteristic—lymphocytic inltration of the
thyroid parenchyma(3). The dening clinical traits of GD and HT involve thyrotoxicosis and
hypothyroidism, correspondingly(4).The inltration of the thyroid by autoreactive lymphocytes and the
generation of antibodies against three primary thyroid antigens, namely thyroid peroxidase (TPO),
thyroglobulin (TG), and thyroid-stimulating hormone receptor (TSHR), are instigated by the activation of
T- and B cell pathways(5).AITD's etiology is presently comprehended as multifactorial, resulting from the
intricate interplay between particular susceptibility genes and environmental exposures,with genetic
differences and susceptibility playing an important role in the etiology of GD and HT(6).
Myasthenia gravis (MG) exemplies a classic autoimmune disorder mediated by antibodies, primarily
affecting the neuromuscular junction(7).Antibody-mediated processes underlie MG, where antibodies are
generated against key components such as the acetylcholine receptor (AchR), the muscle-specic kinase
antibody (MuSK), and the agrin receptor low-density lipoprotein receptor-related protein-4 antibody (LRP4)
(8).The precise triggering of the autoimmune response in MG remains undisclosed, however, it is evident
that deviations within the thymus gland (hyperplasia and neoplasia) have a substantial role, particularly
in patients with anti-AChR antibodies(9, 10) and the development of the disorder is plausibly subject to
genetic predisposition(11).
MG and AITD exhibit certain similarities, such as both being organ-specic, antibody-mediated, and
contributing to ocular myopathy and exophthalmos(12).Some studies have documented the rising
incidence of thyroid disorders in MG, with a higher propensity for MG patients to develop HT and other
autoimmune thyroid disorders(13, 14).Patients diagnosed with HT and GD exhibited a heightened
subsequent risk of developing MG(15).However, there is a lack of consistency among the reported results,
with some studies indicating no clinical association between myasthenia symptomatology and thyroid
dysfunction, as well as no signicant impact on myasthenic symptoms when the endocrine disorders
improve(16).The relationship between AITD and MG is still a topic of ongoing debate, and observational
studies are susceptible to the inuence of reverse causality and confounding effects.To explore the
causal association between AITD and MG, we employed a bidirectional Mendelian randomization (MR)
approach in this study. This method utilized genetic variants obtained from genome-wide association
studies as instrumental variables (IVs) to mitigate biases commonly found in observational
epidemiological studies, such as reverse causation.
2.Materials and methods
2.1 Study design and the assumption of MR
Page 4/15
Employing a two-sample Mendelian randomization (MR) design, we ascertained the overall effects, with
the primary objective of assessing the connection between AITD and MG. In a separate two-step MR
investigation, we explored whether thyroid function characteristics acted as intermediaries in the impact
of AITD on MG. A reverse MR analysis was carried out to assess the reciprocal inuence of MG on
AITD(Figure1).The MR analysis was conducted under the following assumptions: (i) the single nucleotide
polymorphisms (SNPs) used as IVs were obtained from GWAS and displayed associations with the
exposures; (ii) the IVs were not associated with confounding factors; (iii) the IVs had an exclusive
inuence on the risk of outcomes solely through the exposures(17).
2.2 Data sources
FinnGen constitutes a substantial collaboration between the public and private sectors, with the objective
of gathering and scrutinizing genomic and health information from 500,000 individuals enrolled in
FinnGen biobanks(https://www.nngen./en). Within this framework, the FinnGen Biobank of European
descent has furnished the Genome-Wide Association Study (GWAS) data associated with ATID,
encompassing GD with 4,462 cases and 320,703 controls, as well as autoimmune hypothyroidism with
40,926 cases and 274,069 controls(18). We obtained the summary data for thyroid function GWAS from
the ThyroidOmics Consortium, an initiative established to investigate the factors inuencing thyroid
disorders and thyroid function(19). In a meta-analysis, the analysis of thyroid-stimulating hormone (TSH)
included data from 22 distinct cohorts, encompassing a total of 54,288 individuals, while analyses of free
thyroxine (FT4) were based on data from 19 cohorts involving 49,269 individuals(19).The GWAS
information for thyroid peroxidase antibodies (TPOAb) was extracted from a separate meta-analysis
conducted on a general population of 18,297 individuals across 11 different populations. Among these
individuals, there were 1,769 cases with TPOAb positivity(20).Samples of individuals with MG were
gathered from collaborative sources in both the United States and Italy, constituting a total of 1,873 cases
and 36,370 controls(21). The diagnosis of MG relied on established clinical criteria, specically the
presence of characteristic, fatigue-induced muscle weakness, alongside electrophysiological and/or
pharmacological anomalies, and further conrmed by the presence of anti-acetylcholine receptor
antibodies(22).The complete information is in Table1.
Table1
Details of GWAS included in MR analyses.
Page 5/15
Traits Consortia Ethnicity Cases Control Sample
size PMID
Graves FinnGen Biobank European 4462 320703 325165 36653562
Autoimmune
hypothyroidism FinnGen Biobank European 40926 274069 314995 36653562
TPOAb The ThyroidOmics
Consortium European 1769 16528 18297 24586183
FT4 The ThyroidOmics
Consortium European / / 49269 30367059
TSH The ThyroidOmics
Consortium European / / 54288 30367059
Myasthenia
gravis HumanOmniExpress
arrays European 1873 36370 38243 35074870
2.3 Selection of genetic instrumental variables
To obtain IVs while satisfying the assumption of strong correlation between the exposure and SNPs, we
applied a genome-wide signicance threshold of P-value (P<5×10-8). Additionally, the datasets were
harmonized through the removal of variants in potential linkage disequilibrium (r2 =0.001, 10,000
kb).Subsequently,we standardized the effect estimates for both exposure and outcome variants and
eliminated any potential SNPs with incompatible alleles or palindromic SNPs(23).To assess the strength
of genetically determined IVs and avoid any bias towards weak IVs, we used F statistics (beta2/se2)
(24)and ensured that F>10 in line with the rst MR assumption(25, 26).
2.4 Mendelian randomization analyses
The main analysis utilized the inverse-variance weighted (IVW) approach under a random-effects model,
which accounts for heterogeneity across SNPs(27). We conducted several sensitivity analyses to ensure
the robustness of the primary analysis. The weighted median (WM) method, requiring over 50% of the
weight corresponding to valid IVs, was also employed to estimate the causal effects(28). Additionally, we
evaluated possible horizontal pleiotropy using MR-Egger intercepts(28, 29). To detect and correct for any
potential horizontal pleiotropic outliers, we utilized the MR-PRESSO framework, adjusting the IVW
estimate through outlier removal(30). Furthermore, we conducted a leave-one-out analysis to investigate
whether the effect estimates were impacted by any singular outlier variant.The analyses were conducted
using the R software (version 4.2.3) Two-Sample MR package.
Result
Forward MR Analysis between AITD and MG
Page 6/15
After data screening, 24 SNPS were extracted from GD data, 117 SNPS were extracted from the
Autoimmune hypothyroidism data, 8 SNPS were extracted from TPOAb data, 41 SNPS were extracted
from TSH data, and 19 SNPS were extracted from FT4 data ( Supplementary1 Tables 1-5 ).The
assessment of the effects of these 24 valid IVs on MG consistently revealed a causal association
between GD and MG (OR 1.31,95%CI 1.08 to 1.60,P=0.005), and this direction of effect remained
consistent when employing both the MR-Egger and Weighted Median methods.Subsequent testing
revealed the presence of heterogeneity (Q-pval =1.180e-05), leading to the adoption of a random-effects
model to estimate the MR effect size. Neither evidence of horizontal pleiotropy was found through the MR
Egger intercept(egger intercept P=0.996), nor was there any signicant difference in results after
removing two outlier identied by the MR Presso test (MR PRESSO Distortion Test P=0.929).The results
remained stable before and after the correction(Supplementary2 Table1).The IVW method, upon analysis,
indicated a signicant association between autoimmune hypothyroidism and an increased risk of MG
(OR: 1.26, 95% CI: 1.08 to 1.47, P =0.002, Figure 1). This direction of effect was consistent with the
Weighted Median method, and the MR Egger intercept (egger intercept P=0.127) did not reveal any
evidence of pleiotropy. After removing six outliers with the MR Presso approach, the results remained
unchanged (MR PRESSO Distortion Test P = 0.620). Furthermore, the IVW method found no signicant
associations between TSH, FT4, and TPOAb with the risk of MG (refer to Figure 1). These consistent
ndings were replicated using alternative methodologies and through replicative
analyses(Supplementary2 Table1).
Reverse MR Analysis between MG and AITD
During the reverse MR analysis, six SNPs were extracted from the MG dataset and utilized as IVs. The
IVW method demonstrated a signicant association between MG and an increased risk of GD(OR: 1.50,
95% CI: 1.14to1.98, P =3.57e-3, Figure 2). This association was corroborated by the Weighted Median
method (OR: 1.21, 95% CI: 1.10to1.33, P =6.04e-5, Figure 2) and the Weighted Mode method (OR: 1.20,
95% CI: 1.10to1.31, P =9.30e-3, Figure 2). Moreover, the MR Egger intercept (Egger intercept P = 0.310)
did not indicate the presence of pleiotropy. After the removal of four outliers with the MR Presso
technique, the results remained stable (MR PRESSO Distortion Test P=1). Causal associations were also
observed between MG and autoimmune hypothyroidism, supported by the IVW method (OR: 1.29, 95% CI:
1.04to1.59, P =0.019, Figure 2), the Weighted Median method (OR: 1.12, 95% CI: 1.07to1.17, P =7.43e-8,
Figure 2), and the Weighted Mode method (OR: 1.12, 95% CI: 1.08to1.17, P =2.02e-3, Figure 2). However,
further MR Presso testing revealed the presence of pleiotropy (MR PRESSO Distortion Test P < 0.001),
indicating instability in the results.
Following harmonization, a combined total of 2 valid IVs were identied for the association between MG
and TPOAb, while 1 valid IV was found for the relationship between MG and FT4/TSH. The IVW method
revealed a causal relationship between MG and TPOAb (OR: 1.84, 95% CI: 1.39to2.42, P =1.47e-5,
Figure3), and the Wald ratio method indicated an association between MG and elevated TSH
(Beta:0.08,95% CI:0.01to 0.14,P =0.011,Figure4), whereas there was no observed correlation with FT4.
Page 7/15
F-statistics and Visualization of MR
F-statistics were employed to calculate the values for each valid IV, with none of them falling below 10
(Supplementary1 Tables9–14). The arrangement of gures from left to right includes forest plots, scatter
plots, funnel plots, and leave-one-out plots showcasing the MR Effect (Figure 5). The scatter plots exhibit
a positive correlation trend between ATID and MG, which is also evident in the reverse MR analysis. The
symmetrical funnel plots indicate result stability. The forest plots allow for the observation of the effects
of each SNP, while the leave-one-out analysis validates the signicance of the results(Figure 5).
Discussion
This is the rst MR analysis conducted on AITD and MG. Previous case reports have described the co-
occurrence of GD and MG(31-34). However, the occurrence of thyroid-associated
ophthalmopathy(TAO)and MG together is extremely rare. A retrospective study of 1482 MG cases
revealed that only 20 cases (1.3%) were identied with TAO(35). The sequence of onset between AITD
and MG remains unclear. Studies have reported the TNF-α -863 polymorphism is likely to be associated
with MG combined with TAO(33).Both AITD and MG demonstrate a noticeable genetic predisposition(7,
36, 37). Thymoma or thymus hyperplasia is commonly linked to MG(7), and the amelioration of
neuromuscular symptoms following thymectomy suggests the involvement of a dysfunctional thymus in
the development of MG(38). The presence of thymus hyperplasia in GD was initially described in 1912
and is a prevalent nding (approximately 40% in histology) in patients with thyrotoxicosis(39-41).
Multiple lines of evidence indicate that thyroid hormones themselves induce thymus hyperplasia(42-44).
In this context, the promiscuous expression of the TSH receptor in thymocytes may be responsible for the
autoimmune-mediated expansion of the thymus in GD, facilitated by TSH receptor-stimulating
autoantibodies(45). Conversely, it has also been observed that the size of the thymus decreases after
thyroidectomy, reecting the correction of thyrotoxicosis as well as the reduction of the autoimmune
response against the TSH receptor(46).Reduction ofTPOAbfollowing Thymectomy in Patients
withMG(47).
The existence of a correlation between AITD and MG remains a subject of debate. This study utilized MR
analysis to provide evidence supporting a causal relationship between AITD and MG based on genetic
variation. The ndings complement the conclusions drawn from previous observational studies. Our
results indicate a higher susceptibility of AITD patients to MG and a greater likelihood of MG patients
developing GD. However, the reliability of the results for Autoimmune hypothyroidism is considered
questionable due to the inuence of horizontal pleiotropy. Furthermore, MG patients exhibit a higher
prevalence of TPOAb positivity. Additionally, a positive correlation between MG and TSH is observed,
although further validation is required as only one SNP was analyzed.
Our study possessed evident advantages.Firstly, it stood as the inaugural research endeavor to analyze
the causal association between AITD and MG using bidirectional two-sample Mendelian
randomization.Moreover, the exposure and outcome datasets were sourced from different databases,
Page 8/15
thereby mitigating the potential interference caused by sample overlap (33). The instrumental variables
(IVs) employed in our study were SNPs exhibiting strong associations (P5e-8) and high intensity (F-
statistics > 10). Consequently, the exposure and outcome samples in this study were more comparable,
lending greater credibility to our conclusions.Furthermore, our study incorporated a comprehensive
sensitivity analysis.
Nonetheless, it is crucial to acknowledge the limitations of our research. Firstly, the available GWAS data
for MG and TPOAb is currently restricted, comprising a small number of cases and a limited set of
extractable SNPs. To ensure further validation, larger sample sizes of GWAS data are required. Secondly,
we did not stratify the causal effects of GD and MG based on gender and age, which may introduce
potential heterogeneity due to variations in health status, age, or gender. Moreover, it is worth noting that
our study population consisted of Europeans, and therefore, the generalizability of our conclusions to a
global population may be limited.
Conclusion
In summary, our bidirectional two-sample MR analysis explores the relationship between AITD and MG,
elucidating the causal associations that retrospective studies fail to address from a perspective of
genetic variation. It reveals the bidirectional causal relationship between GD and MG, as well as the
causal relationship between hypothyroidism and MG. Furthermore, it indicates a higher prevalence of
TPOAb positivity in MG patients, potentially linked to elevated TSH levels. These ndings supplement the
evidence from previous observational studies.
Declarations
Contributor Information
Suijian Wang,Department of Endocrinology, The First Aliated Hospital, School of Medicine,59
Changping Road,Shantou University,Shantou515041,China,21sjwang@stu.edu.cn
Shaoda Lin,Department of Endocrinology, The First Aliated Hospital, School of Medicine,59 Changping
Road,Shantou University,Shantou515041,China,Shaoda.2023@outlook.com
Xiaohong Chen,Department of Endocrinology, The First Aliated Hospital, School of Medicine,59
Changping Road,Shantou University,Shantou515041,China,Sharon.uol2017@outlook.com
Daiyun Chen,Department of Endocrinology, The First Aliated Hospital, School of Medicine,59
Changping Road,Shantou University,Shantou515041,China,14dychen@stu.edu.cn
Authorcontributions
Research design and conceptualization,S.W.,data management,S.W.,S.L.,Investigation and
analysis,S.W.,D.C.,Verication,S.W.,X.C.,visualization,D.C.,writing and review,S.L.,project integration and
Page 9/15
editor,S.W.,X.C,the manuscript has been reviewed and approved by all authors prior to its publication.
Funding
This work was supported by the Guangdong Provincial Science and Technology Special Funds (Project
No. 20211231071-33) and the Clinical Research Enhancement Program (Project No. 2014110108).
Availability of data and material
For access to the data utilized in this research, interested parties can contact the corresponding author
directly.
Conicts of interest
Not applicable
References
1. Antonelli A, Ferrari SM, Corrado A, Di Domenicantonio A, Fallahi P. Autoimmune thyroid disorders.
Autoimmun Rev. 2015 Feb;14(2):174-80.
2. Jacobson DL, Gange SJ, Rose NR, Graham NMH. Epidemiology and estimated population burden of
selected autoimmune diseases in the United States. Clin Immunol Immunopathol. 1997
Sep;84(3):223-43.
3. Tomer Y. Mechanisms of Autoimmune Thyroid Diseases: From Genetics to Epigenetics. In: Abbas AK,
Galli SJ, Howley PM, editors. Annual Review of Pathology: Mechanisms of Disease, Vol 9. Palo Alto:
Annual Reviews; 2014. p. 147-56.
4. Kyritsi EM, Kanaka-Gantenbein C. Autoimmune Thyroid Disease in Specic Genetic Syndromes in
Childhood and Adolescence. Front Endocrinol. 2020 Aug;11:22.
5. McLachlan SM, Rapoport B. Thyroid peroxidase as an autoantigen. Thyroid. 2007 Oct;17(10):939-48.
. McLeod DSA, Cooper DS, Ladenson PW, Whiteman DC, Jordan SJ. Race/Ethnicity and the
Prevalence of Thyrotoxicosis in Young Americans. Thyroid. 2015 Jun;25(6):621-8.
7. Drachman DB. Myasthenia gravis. The New England journal of medicine. 1994 Jun 23;330(25):1797-
810.
. Berrih-Aknin S. Myasthenia Gravis: Paradox versus paradigm in autoimmunity. J Autoimmun. 2014
Aug;52:1-28.
9. Berrih S, Morel E, Gaud C, Raimond F, Lebrigand H, Bach JF. ANTI-ACHR ANTIBODIES, THYMIC
HISTOLOGY, AND T-CELL SUBSETS IN MYASTHENIA-GRAVIS. Neurology. 1984;34(1):66-71.
10. Roxanis I, Micklem K, Willcox N. True epithelial hyperplasia in the thymus of early-onset myasthenia
gravis patients: implications for immunopathogenesis. J Neuroimmunol. 2001 Jan;112(1-2):163-73.
Page 10/15
11. Giraud M, Beaurain G, Yamamoto AM, et al. Linkage of HLA to myasthenia gravis and genetic
heterogeneity depending on anti-titin antibodies. Neurology. 2001 Nov;57(9):1555-60.
12. Chen YL, Yeh JH, Chiu HC. Clinical features of myasthenia gravis patients with autoimmune thyroid
disease in Taiwan. Acta Neurol Scand. 2013 Mar;127(3):170-4.
13. Meng C, Jing Y, Li R, Zhang X, Wang J. [Clinical features of myasthenia gravis with thyroid disease
with 106 patients]. Zhonghua yi xue za zhi. 2016 Mar 22;96(11):854-8.
14. Chou CC, Huang MH, Lan WC, Kong SS, Kuo CF, Chou IJ. Prevalence and risk of thyroid diseases in
myasthenia gravis. Acta Neurol Scand. 2020 Sep;142(3):239-47.
15. Yeh JH, Kuo HT, Chen HJ, Chen YK, Chiu HC, Kao CH. Higher Risk of Myasthenia Gravis in Patients
With Thyroid and Allergic Diseases A National Population-Based Study. Medicine (Baltimore). 2015
May;94(21):5.
1. De Assis JL, Scaff M, Zambon AA, Marchiori PE. [Thyroid diseases and myasthenia gravis]. Arquivos
de neuro-psiquiatria. 1984 Sep;42(3):226-31.
17. Sun BB, Maranville JC, Peters JE, et al. Genomic atlas of the human plasma proteome. Nature. 2018
Jun;558(7708):73-+.
1. Kurki MI, Karjalainen J, Palta P, et al. FinnGen provides genetic insights from a well-phenotyped
isolated population. Nature. 2023 Jan;613(7944):508-+.
19. Teumer A, Chaker L, Groeneweg S, et al. Genome-wide analyses identify a role for SLC17A4 and
AADAT in thyroid hormone regulation. Nat Commun. 2018 Oct;9:14.
20. Medici M, Porcu E, Pistis G, et al. Identication of Novel Genetic Loci Associated with Thyroid
Peroxidase Antibodies and Clinical Thyroid Disease. PLoS Genet. 2014 Feb;10(2):13.
21. Chia R, Saez-Atienzar S, Murphy N, et al. Identication of genetic risk loci and prioritization of genes
and pathways for myasthenia gravis: a genome-wide association study. Proc Natl Acad Sci U S A.
2022 Feb;119(5):10.
22. Renton AE, Pliner HA, Provenzano C, et al. A Genome-Wide Association Study of Myasthenia Gravis.
JAMA Neurol. 2015 Apr;72(4):396-404.
23. Hemani G, Zhengn J, Elsworth B, et al. The MR-Base platform supports systematic causal inference
across the human phenome. eLife. 2018 May;7:29.
24. Paternoster L, Standl M, Waage J, et al. Multi-ancestry genome-wide association study of 21,000
cases and 95,000 controls identies new risk loci for atopic dermatitis. Nat Genet. 2015
Dec;47(12):1449-56.
25. Pierce BL, Burgess S. Ecient Design for Mendelian Randomization Studies: Subsample and 2-
Sample Instrumental Variable Estimators. Am J Epidemiol. 2013 Oct;178(7):1177-84.
2. Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in
epidemiological studies. Hum Mol Genet. 2014 Sep;23:R89-R98.
27. Bowden J, Smith GD, Burgess S. Mendelian randomization with invalid instruments: effect
estimation and bias detection through Egger regression. Int J Epidemiol. 2015 Apr;44(2):512-25.
Page 11/15
2. Burgess S, Bowden J, Fall T, Ingelsson E, Thompson SG. Sensitivity Analyses for Robust Causal
Inference from Mendelian Randomization Analyses with Multiple Genetic Variants. Epidemiology.
2017 Jan;28(1):30-42.
29. Bowden J, Del Greco MF, Minelli C, Smith GD, Sheehan N, Thompson J. A framework for the
investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat Med. 2017
May;36(11):1783-802.
30. Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal
relationships inferred from Mendelian randomization between complex traits and diseases. Nature
Genet. 2018 May;50(5):693-+.
31. Chhabra S, Pruthvi BC. Ocular myasthenia gravis in a setting of thyrotoxicosis. Indian journal of
endocrinology and metabolism. 2013 Mar;17(2):341-3.
32. lonescu L, Stefănescu C, Dănilă R, et al. Myasthenia gravis associated with thymoma and toxic
multinodular goiter. A case report. Revista medico-chirurgicala a Societatii de Medici si Naturalisti din
Iasi. 2012 Apr-Jun;116(2):540-4.
33. Yang HW, Wang YX, Bao J, Wang SH, Lei P, Sun ZL. Correlation of HLA-DQ and TNF-α gene
polymorphisms with ocular myasthenia gravis combined with thyroid-associated ophthalmopathy.
Bioscience reports. 2017 Apr 28;37(2).
34. Levy G, Meadows WR, Gunnar RM. The association of grave's disease with myasthenia gravis, with a
report of ve cases. Annals of internal medicine. 1951 Jul;35(1):134-47.
35. Chen YL, Yeh JH, Chiu HC. Clinical features of myasthenia gravis patients with autoimmune thyroid
disease in Taiwan. Acta Neurol Scand. 2013 Mar;127(3):170-4.
3. Bello-Sani F, Anumah FE, Bakari AG. Myasthenia gravis associated with autoimmune thyroid disease:
a report of two patients. Annals of African medicine. 2008 Jun;7(2):88-90.
37. Tan JH, Ho KH. Familial autoimmune myasthenia gravis. Singapore medical journal. 2001
Apr;42(4):178-9.
3. Castleman B. The pathology of the thymus gland in myasthenia gravis. Annals of the New York
Academy of Sciences. 1966 Jan 26;135(1):496-505.
39. Scheiff JM, Cordier AC, Haumont S. Epithelial cell proliferation in thymic hyperplasia induced by
triiodothyronine. Clinical and experimental immunology. 1977 Mar;27(3):516-21.
40. Michie W, Beck JS, Mahaffy RG, Honein EF, Fowler GB. Quantitative radiological and histological
studies of the thymus in thyroid disease. Lancet (London, England). 1967 Apr 1;1(7492):691-5.
41. Simpson JG, Gray ES, Michie W, Beck JS. The inuence of preoperative drug treatment on the extent
of hyperplasia of the thymus in primary thyrotoxicosis. Clinical and experimental immunology. 1975
Nov;22(2):249-55.
42. Marine D, Manley OT, Baumann EJ. THE INFLUENCE OF THYROIDECTOMY, GONADECTOMY,
SUPRARENALECTOMY, AND SPLENECTOMY ON THE THYMUS GLAND OF RABBITS. The Journal of
experimental medicine. 1924 Sep 30;40(4):429-43.
Page 12/15
43. Marder SN. The effect of thyroxine on the lymphoid-tissue mass of immature female mice. Journal
of the National Cancer Institute. 1951 Jun;11(6):1153-61.
44. Fabris N, Mocchegiani E, Mariotti S, Pacini F, Pinchera A. Thyroid function modulates thymic
endocrine activity. The Journal of clinical endocrinology and metabolism. 1986 Mar;62(3):474-8.
45. van der Weerd K, van Hagen PM, Schrijver B, et al. Thyrotropin acts as a T-cell developmental factor
in mice and humans. Thyroid. 2014 Jun;24(6):1051-61.
4. Giménez-Barcons M, Colobran R, Gómez-Pau A, et al. Graves' disease TSHR-stimulating antibodies
(TSAbs) induce the activation of immature thymocytes: a clue to the riddle of TSAbs generation?
Journal of immunology (Baltimore, Md : 1950). 2015 May 1;194(9):4199-206.
47. Rotondo Dottore G, Leo M, Ricciardi R, et al. Disappearance of Anti-Thyroid Autoantibodies following
Thymectomy in Patients with Myasthenia Gravis. European thyroid journal. 2021 Jun;10(3):237-47.
Figures
Figure 1
A ow chart outlining the study design and the steps involved in MR analysis.
Page 13/15
Figure 2
Forest plots of causal effect estimates in forward MR. GD, Graves disease; SNP, single-nucleotide
polymorphism;IVW,inverse variance weighted.MG,Myasthenia gravis;TPOAb,thyroid peroxidase
antibody;FT4,free thyroxine4;TSH,thyroid stimulating hormone.
Page 14/15
Figure 3
Forest plots of causal effect estimates in reverse MR,GD, Graves disease; SNP, single-nucleotide
polymorphism;IVW,inverse variance weighted.MG,Myasthenia gravis;TPOAb,thyroid peroxidase antibody.
Page 15/15
Figure 4
Forest plots of causal effect estimates in reverse MR,IVW,inverse variance weighted.MG,Myasthenia
gravis;TPOAb,thyroid peroxidase antibody;FT4,free thyroxine4;TSH,thyroid stimulating hormone.
Figure 5
From left to right includes forest plots, scatter plots, funnel plots, and leave-one-out plots showcasing the
MR Effect.A,the MR Effect of GD on MG;B,the MR Effect of hypothyroidism on MG;C,the MR Effect of MG
on GD;D,the MR Effect of MG on hypothyroidism.
Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.
Supplementary1.pdf
Supplementary2.pdf
