Effects of Hyperthyroidism on Venous Thromboembolism: A Mendelian Randomization Study

Objective Observational studies show the correlation between thyroid dysfunction and risk of venous thromboembolism. However, the causal effects remain uncertain. Our study was conducted to evaluate whether thyroid function and dysfunction were causally linked to the risk of venous thromboembolism. Methods Publicly available summary data of thyrotropin (TSH) and free thyroxine (FT4), hypothyroidism, and hyperthyroidism were obtained from the ThyroidOmics Consortium and the UK Biobank. With single nucleotide polymorphisms (SNPs) as instrumental variables, the casual effects of genetically predicted TSH and FT4 and hypo- and hyperthyroidism on venous thromboembolism outcome were estimated through Mendelian randomization analysis methods (inverse variance weighted (IVW), MR-Egger, weighted median, simple mode, and weighted mode). Cochran's Q test was performed to evaluate the heterogeneity and horizontal pleiotropy. Results Our study selected 15 FT4-, 36 TSH-, 3 hyperthyroidism-, and 79 hypothyroidism-associated SNPs as instrumental variables. The IVW analysis results showed that the odds ratio of venous thromboembolism for hyperthyroidism was 1.124 (95% confidence interval: 1.019-1.240; p = 0.019), demonstrating the casual effect of hyperthyroidism not FT4, TSH, and hypothyroidism on venous thromboembolism. No heterogeneity or horizontal pleiotropy was observed according to Cochran's Q test. Conclusion Our Mendelian randomization analysis supports the causal effect of hypothyroidism on risk of venous thromboembolism. There is no evidence that genetically predicted TSH, FT4, and hypothyroidism have casual effects on venous thromboembolism. Future studies should be conducted to elucidate the underlying pathophysiological mechanisms.


Introduction
Venous thromboembolism, including deep vein thrombosis and pulmonary embolism, is a chronic disease that affects approximately 10 million individuals per year globally [1,2]. The annual incidence of acute venous thromboembolism is 1-2 cases per 1000 people and is four times higher in highincome countries compared with low-income countries [3]. The diagnostic methods of venous thromboembolism consist of sequential examinations combining clinical probabilities, D-dimer test, and imaging assessments [4]. Venous thromboembolism therapy is aimed at preventing thrombus extension and embolism, cardiopulmonary failure, deaths, recurrence, and long-term complications [5]. Direct oral anticoagulants and thrombin inhibitor are the major treatment strategies in the management of venous thromboembolism [6]. Venous thromboembolism is a multifactorial illness induced by the interaction of multiple predisposing factors [4]. Observational studies have showed the correlation between thyroid dysfunction and risk of venous thromboembolism. A prospective cohort study showed that hyperthyroidism not hypothyroidism was associated with lower risk of recurrent venous thromboembolism among elderly patients [7]. A meta-analysis of cohort studies also showed that hyperthyroidism was a risk factor of venous thromboembolism [8]. However, in a retrospective study, increased risk of venous thromboembolism was observed among patients with hypothyroidism not hyperthyroidism [9]. Nevertheless, it remains unclear whether the above observed associations are causal, because observational studies could have selection bias, residual confounding, and reverse causality.
Mendelian randomization is a research method that uses genetic variation as an instrumental variable to establish a model to test the causal relationship between variable exposure factors and diseases [10,11]. Single nucleotide polymorphism (SNP) loci discovered by genome-wide association study (GWAS) based on large samples provide a large number of genetic variation tool variables for Mendelian randomization analysis, which can effectively reduce the bias of causality estimates [12][13][14]. Mendelian randomization has been widely used in venous thromboembolism studies, revealing the intrinsic link between various phenotypes and venous thromboembolism [15][16][17]. For instance, a Mendelian randomization study showed that obesity is casually correlated with venous thromboembolism [18]. Another study found that taller height acts as a risk factor of venous thromboembolism [19]. Based on previous research, this study collected currently available GWAS data on thyroid function and dysfunction. The causal relationships of free thyroxine (FT4), thyrotropin (TSH), hyperthyroidism, and hypothyroidism with venous thromboembolism were assessed using two-sample Mendelian randomization, which might provide a basis for prevention and treatment of venous thromboembolism.

Summary Data and Study Population.
Our study acquired publicly available genetic summary data from two large GWAS cohorts (the ThyroidOmics Consortium [20] and the UK Biobank [21]) ( Table 1). GWAS summary data on FT4 and TSH were accessed from the ThyroidOmics Consortium, composed of 72,167 venous thromboembolism samples. Meanwhile, GWAS summary data on hyperthyroidism and hypothyroidism were obtained from the UK Biobank, comprising 337,159 venous thromboembolism samples.

Two-Sample Mendelian Randomization Analysis for
Evaluating Causal Effects of FT4, TSH, Hyperthyroidism, and Hypothyroidism on Venous Thromboembolism. Five different two-sample Mendelian randomization analysis approaches (IVW, MR-Egger, weighted median, sample mode, and weighted mode) were conducted to assess the casual effects of FT4, TSH, hyperthyroidism, and hypothyroidism on risk of venous thromboembolism, as shown in Figures 2(a)-2(d). For the IVW analysis results, the OR of venous thromboembolism for hyperthyroidism was 1.124 (95% CI: 1.019-1.240; p = 0:019), without statistical significance for the MR-Egger, weighted median, sample mode, and weighted mode analysis results (Table 6). Meanwhile, no statistical significance from the five Mendelian randomization analysis approaches was found for FT4, TSH, and hypothyroidism on venous thromboembolism. The above data demonstrated that hyperthyroidism was a risk factor of venous thromboembolism.

Discussion
Randomized controlled trials are the gold standard for causal inference in clinical research but are expensive and time-consuming and often difficult to implement due to ethical factors and subject constraints [23,24]. Observational studies are relatively easy to implement and are widely applied for initial determination of etiology [25,26]. Nonetheless, when there are reverse causal effects and potential confounding factors, the results of observational studies are often difficult to be recognized [27]. Mendelian randomization method provides an effective way to solve the above problems, using genetic variation as an instrumental variable effectively to avoid confounding factors caused by the environment [28]. Furthermore, compared with the immediate results obtained from randomized controlled trials, the exposure factors obtained from a genetic point of view are often accompanied for life and can even be passed on to the next generation [29]. Based on above evidence, we carried out the two-sample Mendelian randomization study to analyze the casual effects of FT4, TSH, hyperthyroidism, and hypothyroidism on venous thromboembolism.
The IVW method applies the ratio method to calculate causality estimates for individual instrumental variables and aggregates each estimate to perform a weighted linear regression to obtain an overall estimate [30]. However, this method cannot calculate the bias caused by pleiotropy, so an additional Cochran's Q test is needed to evaluate pleiotropy [31]. Our IVW analysis results showed that the ORR of venous thromboembolism for hyperthyroidism was 1.124 (95% CI: 1.019-1.240; p = 0:019), demonstrating the casual effect of hyperthyroidism not FT4, TSH, and hypothyroidism on venous thromboembolism. Additionally, there was no heterogeneity or horizontal pleiotropy in accordance with Cochran's Q test. Genome-wide association analysis uncovers that most variations in thyroid function are genetically determined [20]. In a nationwide populationbased cohort study, hypothyroidism was in relation to higher risk of venous thromboembolism [32]. Decreased venous thromboembolism risk of hypothyroid patients treated with thyroxine replacement therapy was not investigated. However, our Mendelian randomization analysis showed that there was no causal effect of thromboembolism on venous thromboembolism. Another cohort study showed that TSH levels within the normal range were not associated with venous thromboembolism risk, but abnormal TSH levels were associated with moderate risk of venous thromboembolism [33]. Case-control studies have shown the significant association between serum TSH levels and increased risk of venous thromboembolism [34]. Consistently, TSH concentration was an independent risk factor of venous thromboembolism regardless of thyroid function [35]. Our causal analysis demonstrated that TSH was not casually associated with venous thromboembolism.
However, there are several limitations in our study. Firstly, the sample sets included in this study are all from the European region. Whether the research conclusions can be extended to other populations still needs further research to be proven. Secondly, there may be population sharing of sample sets from Europe, leading to overuse of genetic information.

Conclusion
Altogether, in the two-sample Mendelian randomization study with genome-wide association variants, our evidence demonstrated that hypothyroidism was casually associated with venous thromboembolism outcome. However, there was no evidence that TSH, FT4, and hypothyroidism had casual effects on venous thromboembolism. Despite this, further studies are required to validate the biological mechanisms underlying the casual association.

Data Availability
The datasets analyzed during the current study are available from the corresponding author on reasonable request.