Skip to main content
Download PDF

Causal associations between epigenetic age and thromboembolism: a bi-directional two-sample Mendelian randomization study

Clinical Epigenetics volume 17, Article number: 75 (2025) Cite this article

Abstract

Background

Thromboembolism is one of the most prevalent cardiovascular conditions affecting the elder population. The associations between epigenetic aging and thromboembolism risks remain incompletely elucidated. Through Mendelian randomization (MR), this research seeks to assess the causal links between genetically determined epigenetic aging factors and thromboembolism.

Results

Genetic variants were extracted from genome-wide association studies (GWAS) under stringent threshold as instrumental variables (IVs). Bi-directional two-sample MR analyses were conducted to determine the direction of causal associations. We employed the inverse variance weighted (IVW), weighted median, weighted mode and MR Egger to estimate the causal effect, with sensitivity analyses such as Cochran’s Q tests, MR-PRESSO and leave-one-out performed to avoid potential heterogeneity and pleiotropy. Our MR analysis revealed a causal association between intrinsic epigenetic age acceleration and deep vein thrombosis of lower extremities (IVW: OR 0.963, 95% CI 0.934–0.992, P = 0.014), and between the genetically determined levels of plasminogen activator inhibitor-1 and other arterial embolism and thrombosis (IVW: OR 1.000, 95% CI 1.000–1.0005, P = 0.029). Causality was also identified between the genetically predicted levels of FGF23 and other arterial embolism and thrombosis (IVW: OR: 1.661, 95% CI 1.051–2.624, P = 0.029) and arterial embolism and thrombosis of lower extremity artery (IVW: OR 1.68, 95% CI 1.031–2.725, P = 0.037). Moreover, bi-directional MR showed reverse effects between portal vein thrombosis and PhenoAge (IVW: OR 0.871, 95% CI 0.765–0.992, P = 0.037) and between venous thromboembolism and GrimAge (IVW: OR 1.186, 95% CI 1.048–1.341, P = 0.007). Sensitivity analysis using Cochran’s Q tests, MR-PRESSO and leave-one-out excluded the influence of heterogeneity, horizontal pleiotropy, and outliers.

Conclusion

Our results identified a causal association between genetically predicted epigenetic aging factors and thromboembolism. The findings highlight the necessity for further exploration into the underlying etiology of thromboembolism.

Background

Thromboembolism, a critical cardiovascular condition that is frequently underdiagnosed and undertreated, significantly impacts a large portion of the global population, accounting for approximately one in four deaths worldwide [1]. Venous thromboembolism (VTE), which includes deep vein thrombosis (DVT), pulmonary embolism (PE), thrombophlebitis and portal vein thrombosis (PVT), is a significant health issue affecting approximately 10 million people globally each year [2]. Characterized by abnormal blood clotting in the deep veins, VTE disrupts normal venous return, which can result in severe complications and long-term consequences. Arterial thromboembolism, on the other hand, resulting from clot formation within arteries or embolization from the heart or vessel walls, obstructs blood flow and is a leading cause of ischemic heart disease and stroke [3]. Despite advances in diagnosis and treatment, venous and arterial thromboembolism remain a challenge for clinicians, highlighting the need for primary prevention strategies to reduce the risk of thrombosis. The underlying mechanisms of venous and arterial thromboembolism are complex and multi-factorial, with risk factors including dietary habits, tobacco use, and sedentary behavior and accelerated epigenetic aging [4, 5].

Epigenetic aging, which refers to the biological aging process as indicated by specific epigenetic markers, such as DNA methylation and histone modifications, is a significant area of research in the field of gerontology and epidemiology [6]. Epigenetic age, an indicator of biological age that can differ from chronological age, is a reflection of health status and aging rate. This measure is linked to various health outcomes and can predict longevity [7]. Epigenetic clocks, such as HannumAge, PhenoAge and GrimAge, are tools developed to predict biological age and assess the risk of age-related diseases based on these methylation patterns [8,9,10]. The relationship between epigenetic aging and thrombosis is an emerging area of research [11]. Recent studies have shown that epigenetic age acceleration is associated with a prothrombotic hemostatic profile, characterized by increased levels of clotting factors such as fibrinogen and PAI1, as well as decreased clotting time, which may contribute to the increased risk of thrombotic events [5]. However, prothrombotic profiles may contribute to accelerated epigenetic aging, as chronic inflammation and oxidative stress commonly seen in diabetes or obesity, both of which are risk factors for thrombosis, can contribute to epigenetic aging acceleration [12]. Therefore, the precise causal links between epigenetic aging and thrombosis are not yet fully understood. This gap in knowledge underscores the need for further research to uncover the underlying host factors that could inform more effective prevention and treatment strategies for thrombotic events.

MR offers an alternative to conduct causality assumptions that cannot be readily obtained from conventional observational studies [13]. Using random genetic variants as IVs, MR explores the causal relationships between two variables, helping to control for confounding factors and address issues of reverse causality [14, 15]. Here, an MR analysis was executed to test whether these SNP alleles affecting epigenetic aging are causally associated with the risk of thromboembolism at the population level.

Methods

Study design

A two-sample MR analysis was conducted to assess the casual connections between genetically predicted epigenetic aging factors and thromboembolism. SNPs were selected as IVs based on three criteria: 1) directly linked to the exposure, 2) not correlated with potential confounders, and 3) an exclusive influence on the outcome through the exposure. [14, 16]. Figure 1 summarizes the overall study design. Our study was reported in line with the “STrengthening the Reporting of OBservational studies in Epidemiology using Mendelian Randomization (STROBE-MR)” checklist [16].

Fig. 1
figure 1

Overall study design of the MR analysis. A flowchart depicts how the bi-directional MR analysis was conducted in this study

Full size image

Data sources

The GWAS summary data of venous and arterial thromboembolism was collected from the FinnGen database, a large public–private partnership in Finland aimed at collecting and analyzing genomic and health data to better understand the genetic basis of diseases, which were categorized into VTE (21,021 cases and 391,160 controls), DVT (6501 cases and 357,111 controls), PE (10,046 cases and 401,128 controls), thrombophlebitis (7683 cases and 357,111 controls), PVT (394 cases and 357,111 controls), and arterial embolism and thrombosis of lower extremity artery (1076 cases and 381,977 controls) and other arterial embolism and thrombosis (869 cases and 381,977 controls) of European ancestry.

The GWAS summary data of PhenoAge, GrimAge, HannumAge and intrinsic epigenetic age acceleration (IEAA) derived from the HorvathAge were obtained from a previous study [17]. In addition, genetically predicted levels of five factors that have been associated with epigenetic aging, including granulocyte proportions, plasminogen activator inhibitor-1 (PAI1), telomere length, circulating plasma α-Klotho and fibroblast growth factor 23 (FGF23) were considered for exposure as well, and the genetic data were acquired from public studies [18,19,20,21,22]. The detailed information for each phenotypic exposure and outcome data is provided in Supplemental Table 1.

Selection of instrumental variables

SNPs for each exposure phenotype were selected as candidate IVs at a threshold of p < 5e-8 [23], with the exception of analyses involving granulocyte proportions, PVT, arterial embolism, and thrombosis of the lower extremity artery and other arterial sites, where a less stringent threshold of p < 5e-6 was applied due to limited number of SNPs at p < 5e-8 [24]. Subsequently, SNPs with MAF (minor allele frequency) no more than 0.01 were discarded. Linked disequilibrium (LD)-clumping was conducted to ensure the independence of each SNPs under a stringent criteria (r2 < 0.001, window size = 10,000 kb). For SNPs missing in the outcome, we identified the most highly correlated proxy SNPs within a ± 500 kb window around the original SNP locus, ensuring a high degree of correlation (r2 > 0.8) to maintain their validity as IVs. Furthermore, F-statistics for these IVs were calculated using the formula F = R2*(N-2)/(1-R2), of which R2 represents the proportion of phenotypic variance explained by a single SNP and N refers to sample sizes [25]. SNPs with an F-statistic below 10 were considered as weak IVs and subsequently discarded [26]. Finally, during the harmonization process of aligning exposure and outcome data, alleles that were non-concordant and those with palindromic sequences were eliminated. These stringently selected SNPs were utilized as eligible IVs for MR analyses.

Mendelian randomization analysis

Bi-directional two-sample MR analysis was executed between genetically determined epigenetic aging factors and thromboembolism by calculating odds ratio (OR) and 95% confidence interval (CI). We applied inverse variance weighted (IVW) [27], weighted median [28], weighted mode [29] and MR Egger [30], with IVW as the main approach and other methods utilized for the robustness of the results.

MR Egger considers the presence of intercept terms and can provide accurate causal effect estimates in the presence of pleiotropic bias. The weighted median analyzed the causality based on the assumption that half of the IVs were valid. We initially computed the causal estimates using the fixed effects IVW methods for each IV. The random effects IVW was utilized in case of significant heterogeneity (p < 0.05). Forest plots, scatter plots and funnel plots were also produced to depict the causality and the influence of each SNP on the outcome. All MR analyses were run in R 4.0.5 along with the package “Two-sample MR.”

Sensitivity analysis

To exclude pleiotropy, MR Egger was utilized. The intercept term in MR Egger regression is crucial as it indicates the presence of horizontal pleiotropy. Specifically, a significant intercept term (p < 0.05) suggests that there is unbalanced horizontal pleiotropy [31]. Conversely, a non-significant intercept term (p ≥ 0.05) suggests that there is no strong evidence of horizontal pleiotropy, thereby supporting the validity of the causal estimates.

Cochran’s Q was utilized for heterogeneity identification among IVs of each phenotypic exposure [32]. A significant Q statistic (P < 0.05) indicates heterogeneity, suggesting inconsistent effect sizes across IVs. Conversely, a non-significant Q statistic (P ≥ 0.05) suggests no significant heterogeneity and indicates that effect sizes are consistent and can be reliably pooled. Furthermore, Mendelian Randomization Pleiotropy RESidual Sum and Outlier (MR-PRESSO) was applied for outlier elimination (p < 0.05) and correcting for horizontal pleiotropy [31]. Leave-one-out analysis, which discarded each SNP prior to MR analysis of remaining SNPs on the risk of thromboembolism, was conducted to obtain the effect of each SNP on the overall causality [33].

Results

Screening of IVs associated with epigenetic aging factors

We selected SNPs from GWAS summary data for each phenotypic exposure, applying a stringent threshold. During the harmonization process, we eliminated alleles that were non-concordant and palindromic. Therefore, we acquired a total of 224 IVs associated with the exposure, with 11 for PhenoAge, 4 for GrimAge, 9 for HannumAge, 24 for IEAA, 19 for granulocyte proportions, 5 for PAI1, 138 for telomere length, 8 for FGF23 and 6 for α-Klotho, which all showed an F-statistics > 10 (Supplemental Table 2). Certain SNPs were not found in the outcome and therefore replaced with appropriate proxy SNPs (Supplemental Table 3).

Causal effects of epigenetic aging factors on thromboembolism

Based on the GWAS summary data for thromboembolism, MR analysis revealed significant causality between IEAA and DVT of lower extremities (IVW: OR 0.963, 95% CI 0.934–0.992, P = 0.014), suggesting that IEAA may serve as a risk factor for this condition (Table 1). In addition, genetically predicted levels of FGF23 and other arterial embolism and thrombosis (IVW: OR 1.661, 95% CI 1.051–2.624, P = 0.029) and arterial embolism and thrombosis of lower extremity artery (IVW: OR 1.68, 95% CI 1.031–2.725, P = 0.037) were found to be causally associated, consistent with the weighted median method (Table 1). This suggested the potential of FGF23 in exacerbating thrombotic events. A weak association was observed between PAI1 and other arterial embolism and thrombosis (IVW: OR 1.000, 95% CI 1.000–1.0005, P = 0.029), suggesting a minimal impact on the risk of these conditions (Table 1). No heterogeneity and pleiotropy were identified among these associations (Table 2). MR-PRESSO revealed outliers among associations between α-Klotho and DVT of lower extremities, thrombophlebitis, VTE and PE, between FGF23 and DVT of lower extremities, thrombophlebitis, PE and VTE, between PhenoAge and arterial embolism and thrombosis of lower extremity artery, between telomere length and thrombophlebitis, PE and VTE, the causal associations of which remained negative after outliers were removed (Table 3). We additionally produced scatter plots and forest plots, visually examining for any potential influence of outliers (Figs. 2A–D, 3A–D). Leave-one-out analysis plots and funnel plots further consolidated the stability of the results (Supplemental Figs. 1A–D, 2A–D).

Table 1 Causal estimates of epigenetic aging factors on thromboembolism
Full size table
Table 2 Heterogeneity and pleiotropy for each exposure factor on thromboembolism
Full size table
Table 3 MR-PRESSO results of the MR analysis
Full size table
Fig. 2
figure 2

MR analyses of the causal relationships between epigenetic aging factors and thromboembolism. Scatter plots of causal association between A PAI1 and other arterial embolism and thrombosis; B IEAA and DVT of lower extremities; C FGF23 and arterial embolism and thrombosis of lower extremity artery; D FGF23 and other arterial embolism and thrombosis. The slopes depicted the causal associations, with each line representing a different methodology

Full size image
Fig. 3
figure 3

Forest plots for MR analyses of the causal relationships between epigenetic aging factors and thromboembolism. The causal effect of A PAI1 on other arterial embolism and thrombosis; B IEAA on DVT of lower extremities; C FGF23 on arterial embolism and thrombosis of lower extremity artery; D FGF23 on other arterial embolism and thrombosis. The MR estimate derived from the MR Egger and IVW methods was shown for comparison

Full size image

Screening of IVs associated with thromboembolism

We extracted SNPs from GWAS summary statistics of each thromboembolism subtype. In total, we acquired 92 IVs associated with the exposure, with 33 for VTE, 16 for PE, 4 for PVT, 11 for thrombophlebitis, 13 for DVT of lower extremities, nine for arterial embolism and thrombosis of lower extremity artery and six for other arterial embolism and thrombosis, which all showed an F-statistics > 10 (Supplemental Table 2). Some SNPs missing in the outcome data were replaced with suitable proxy SNPs, as detailed in Supplemental Table 3.

Causal effects of thromboembolism on epigenetic aging factors

To assess any reverse causality between genetically determined levels of epigenetic aging factors and thromboembolism, we considered each thromboembolism subtype as the exposure and epigenetic aging factors as the outcome. In the context of PVT, the causal association with PhenoAge (IVW: OR 0.871, 95% CI 0.765–0.992, P = 0.037) was identified, suggesting that PVT might decelerate epigenetic aging (Table 4). In addition, VTE was found to be causally associated with GrimAge (IVW: OR 1.186, 95% CI 1.048–1.341, P = 0.007), indicating VTE as a potential contributor to epigenetic aging (Table 4). Sensitivity analysis identified no heterogeneity and pleiotropy among these associations (Table 5). Additionally, MR-PRESSO revealed outliers among associations between DVT of lower extremities and FGF23, HannumAge, IEAA and telomere length, and between other arterial embolism and thrombosis and telomere length, between thrombophlebitis and FGF23 and HannumAge, between PE and FGF23, and between VTE and FGF23 and telomere length (Table 6). Finally, the robustness of our findings was confirmed and visualized via leave-one-out analysis plots and funnel plots (Supplemental Figs. 3A, B, 4A, B).

Table 4 Causal estimates of thromboembolism on epigenetic aging factors
Full size table
Table 5 Heterogeneity and pleiotropy for each exposure factor on epigenetic aging factors
Full size table
Table 6 MR-PRESSO results of the MR analysis
Full size table

Discussion

We conducted a bi-directional two-sample MR analysis using GWAS summary statistics from European ancestry to explore the causal relationships between genetically predicted epigenetic aging factors and thromboembolism. Our results indicated that IEAA may be a potential protective factor for the occurrence of DVT of lower extremities. In addition, the levels of PAI1 and FGF23 were demonstrated as potential risk factors for other arterial embolism and thrombosis and arterial embolism and thrombosis of lower extremity artery. Furthermore, we found evidence for causal associations between VTE and GrimAge, as well as between PVT and PhenoAge. This indicates that VTE is associated with an increase in GrimAge, while PVT is associated with a decrease in PhenoAge.

The relationship between epigenetic aging and thromboembolism is increasingly recognized, particularly concerning how alterations in epigenetic markers can influence hemostatic factors. Elevated epigenetic age has been associated with a pro-coagulation profile, characterized by increased levels of fibrinogen and PAI1, which are critical components in the coagulation cascade [5]. This can lead to a higher risk of clot formation and reduced clotting time, potentially causing thromboembolic events [5]. Moreover, studies indicate that epigenetic modifications, such as DNA methylation changes, play a significant role in the pathogenesis of thromboembolic diseases. For instance, specific epigenetic alterations have been observed in conditions like coronary artery disease and cerebrovascular disease, which are closely linked to thromboembolic events [34]. These findings suggest that understanding the epigenetic mechanisms underlying aging could provide insights into the prevention and management of thromboembolism. Our study provided evidence supporting causal links between epigenetic aging factors and thromboembolism and its subtypes, achieved by inferring causality through genetic prediction using MR, which also effectively addresses confounding variables.

Our results showed that genetically determined levels of FGF23 and PAI1 acted as detrimental factors in arterial thromboembolism. FGF23, a hormone mainly produced by osteocytes, regulates phosphate balance and vitamin D metabolism by enhancing renal phosphate excretion and curbing active vitamin D synthesis [35, 36]. In addition, FGF23-deficient mice exhibited aging phenotypes, such as reduced life span, atherosclerosis, osteoporosis, skin atrophy, infertility, or emphysema [37]. Moreover, clinical studies increasingly link elevated FGF23 levels to cardiovascular risks like arterial stiffness and vascular calcification, which contribute to thromboembolism [36]. Mechanistically, it has been proposed that the effect of FGF23 on mineral metabolism can lead to systemic inflammation and vascular changes that predispose individuals to thromboembolic conditions [35, 38]. On the other hand, PAI1 is a serine protease inhibitor that plays a crucial role in the regulation of fibrinolysis and thrombosis. It primarily inhibits tissue plasminogen activator (t-PA) and urokinase plasminogen activator (u-PA), which are essential for the breakdown of blood clots. Elevated levels of PAI1 can lead to impaired fibrinolysis, contributing to an increased risk of thrombotic events such as VTE and arterial thrombosis [39, 40]. More specifically, genetic polymorphisms, particularly the 4G/5G polymorphism, have been linked to varying levels of PAI1 expression and an individual’s susceptibility to VTE [41]. Furthermore, elevated plasma levels of PAI1 are associated with longer clot lysis times, indicating a higher risk for both venous and arterial thrombosis [40, 41]. In line with existing evidence, our study identified FGF23 and PAI1 as potential detrimental factors for thrombotic events. Given the associations between elevated FGF23 and PAI1 levels and thrombosis risks, targeting them may reduce thromboembolic events. Potential interventions include lowering FGF23 or PAI1 levels or modulating their effects on mineral metabolism and fibrinolysis through pharmacological agents, lifestyle changes, or other approaches.

Our results indicated that IEAA was inversely associated with the incidence of DVT of lower extremities. HorvathAge, developed by Steve Horvath in 2013, is an epigenetic clock that estimates biological age based on DNA methylation patterns [6]. This method utilizes a specific set of CpG sites across the genome to predict an individual’s biological age. Research indicates that accelerated biological aging, as reflected by higher IEAA, may correlate with increased risk factors for thromboembolism. This relationship suggests that individuals who exhibit signs of accelerated aging could be at a heightened risk for developing VTE due to underlying physiological changes associated with aging [42]. However, our results demonstrated the opposite. This discrepancy may arise from the fact that the biological aging processes reflected by epigenetic clocks do not uniformly translate to clinical outcomes across different populations. Additionally, variations in how VTE is diagnosed or reported, as well as the specific epigenetic markers selected for analysis, could lead to conflicting results.

In terms of reverse causation, our results revealed that VTE and PVT were causally linked to increased and reduced GrimAge and PhenoAge, respectively. Unlike HorvathAge, GrimAge is specifically designed to predict mortality risk, lifespan, and healthspan. It incorporates not only more DNA methylation data, but also considers lifestyle factors such as smoking history and levels of specific plasma proteins associated with health outcomes, allowing GrimAge to be a more robust predictor of age-related diseases and overall mortality [43]. Individuals with VTE could experience an increased GrimAge, suggesting that thrombosis and related conditions, such as chronic inflammation or endothelial dysfunction, might exacerbate the biological aging markers and increase the risk of developing age-related diseases [44]. Nonetheless, our results revealed that individuals with PVT may exhibit decelerated epigenetic aging as measured by PhenoAge. PhenoAge is an epigenetic clock that incorporates not only DNA methylation information but also the levels of clinical biomarkers such as creatinine and C-reactive protein. It focuses on integrating clinical biomarkers that reflect physiological aging and disease risk, making it particularly useful for assessing healthspan and age-related diseases. The varying connections between epigenetic clocks and thromboembolism may stem from different focuses of these clocks on biological aging, highlighting the multifaceted nature of the aging process. PhenoAge may reflect a different aspect of health status that is less sensitive to the effects of thromboembolic events. Therefore, for individuals with VTE, the increased GrimAge suggests a heightened risk of age-related diseases and mortality, emphasizing the need for more aggressive preventive measures and monitoring of comorbidities. Conversely, the decelerated epigenetic aging observed in PVT patients, as indicated by PhenoAge, may suggest a unique biological profile that warrants further investigation. Understanding these distinct associations could help in the development of targeted interventions to mitigate the adverse effects of thromboembolism on biological aging and overall health.

This study represents the first evaluation of the associations between epigenetic aging factors and thromboembolism using MR. The main strength of this study lies in its utilization of MR, enabling the assessment of the associations between genetically determined levels of epigenetic aging factors and thromboembolism in independent European populations simultaneously. Genetic associations aid in elucidating the diverse relationships between epigenetic aging factors and thromboembolism by considering shared genetic risk factors. The results were validated through alternative MR methods and sensitivity analyses for pleiotropy. Nonetheless, some limitations have to be addressed for this study. Firstly, the lack of individual-level data restricts our ability to categorize patients into finer subgroups based on disease progression. Future studies can address this limitation by using large-scale datasets or electronic health records to obtain detailed individual-level data, enabling more nuanced analyses of disease subtypes and progression. Secondly, we could not sufficiently account for unmeasured confounders like smoking and alcohol consumption, which are known to influence thromboembolism risk. Further efforts could address this issue by incorporating comprehensive lifestyle assessments. Thirdly, the applicability of this study to populations outside of European ancestry is limited due to its focus on this specific demographic. Thus, caution is advised when extrapolating our findings to populations with different ancestral backgrounds. Future studies should aim to include more diverse samples from various ancestral backgrounds to improve the generalizability of the findings.

Conclusions

Our findings revealed that genetically determined levels of epigenetic aging factors affect thromboembolism risk in European ancestry. Our findings suggested one protective causal association between IEAA and DVT of lower extremities; two detrimental causal associations between PAI1 and FGF23 and other arterial embolism and thrombosis and arterial embolism and thrombosis of lower extremity artery; one protective causal associations between PVT and PhenoAge; one detrimental causal association between VTE and GrimAge. These results suggest that these epigenetic aging factors could serve as potential predictors for thromboembolism, offering avenues for future research to develop targeted prevention strategies. However, further studies are needed to confirm these associations and explore their applicability in diverse populations, particularly those of non-European ancestry, to enhance the external validity of our findings.

Availability of data and materials

The GWAS summary statistics for thromboembolism can be accessed via the FinnGen database. The genetic data regarding epigenetic clocks can be accessed from previous studies.

Abbreviations

VTE:

Venous thromboembolism

DVT:

Deep vein thrombosis

PE:

Pulmonary embolism

PVT:

Portal vein thrombosis

CVD:

Cardiovascular diseases

IEAA:

Intrinsic epigenetic age acceleration

FGF23:

Fibroblast growth factor 23

CI:

Confidence interval

References

  1. Wendelboe AM, Raskob GE. Global Burden of Thrombosis: Epidemiologic Aspects. Circ Res. 2016;118:1340–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/circresaha.115.306841.

    Article  CAS  PubMed  Google Scholar 

  2. Raskob GE, et al. Thrombosis: a major contributor to global disease burden. Arterioscler Thromb Vasc Biol. 2014;34:2363–71. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/atvbaha.114.304488.

    Article  CAS  PubMed  Google Scholar 

  3. Mackman N. Triggers, targets and treatments for thrombosis. Nature. 2008;451:914–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/nature06797.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Meah MN, Dweck MR, Newby DE. Cardiovascular imaging to guide primary prevention. Heart (British Cardiac Society). 2020;106:1267–75. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/heartjnl-2019-316217.

    Article  PubMed  Google Scholar 

  5. Ward-Caviness CK, et al. DNA methylation age is associated with an altered hemostatic profile in a multiethnic meta-analysis. Blood. 2018;132:1842–50. https://doiorg.publicaciones.saludcastillayleon.es/10.1182/blood-2018-02-831347.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Horvath S. DNA methylation age of human tissues and cell types. Genome Biol. 2013;14:R115. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/gb-2013-14-10-r115.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Chen BH, et al. DNA methylation-based measures of biological age: meta-analysis predicting time to death. Aging. 2016;8:1844–65. https://doiorg.publicaciones.saludcastillayleon.es/10.18632/aging.101020.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Hannum G, et al. Genome-wide methylation profiles reveal quantitative views of human aging rates. Mol Cell. 2013;49:359–67. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.molcel.2012.10.016.

    Article  CAS  PubMed  Google Scholar 

  9. Levine ME, et al. An epigenetic biomarker of aging for lifespan and healthspan. Aging. 2018;10:573–91. https://doiorg.publicaciones.saludcastillayleon.es/10.18632/aging.101414.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Lu AT, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging. 2019;11:303–27. https://doiorg.publicaciones.saludcastillayleon.es/10.18632/aging.101684.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Liu D, Aziz NA, Pehlivan G, Breteler MMB. Cardiovascular correlates of epigenetic aging across the adult lifespan: a population-based study. GeroScience. 2023;45:1605–18. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s11357-022-00714-0.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Franzago M, Pilenzi L, Di Rado S, Vitacolonna E, Stuppia L. The epigenetic aging, obesity, and lifestyle. Front Cell Develop Biol. 2022;10: 985274. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fcell.2022.985274.

    Article  Google Scholar 

  13. Sanderson E, et al. Mendelian randomization. Nat Rev Methods Primers. 2022. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s43586-021-00092-5.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Davies NM, Holmes MV, Smith GD. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ. 2018. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bmj.k601.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Verduijn M, Siegerink B, Jager KJ, Zoccali C, Dekker FW. Mendelian randomization: use of genetics to enable causal inference in observational studies. Nephrology, Dialysis, Transplantation: Official Publication Eur Dialysis Transplant Association - Eur Renal Assoc. 2010;25:1394–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ndt/gfq098.

    Article  Google Scholar 

  16. Skrivankova VW, et al. Strengthening the reporting of observational studies in epidemiology using mendelian randomization: the STROBE-MR statement. JAMA. 2021;326:1614–21. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jama.2021.18236.

    Article  PubMed  Google Scholar 

  17. Roberts JD, et al. Epigenetic age and the risk of incident atrial fibrillation. Circulation. 2021;144:1899–911. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/circulationaha.121.056456.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Portales-Castillo I, Simic P. PTH, FGF-23, Klotho and Vitamin D as regulators of calcium and phosphorus: genetics, epigenetics and beyond. Front Endocrinol. 2022;13: 992666. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fendo.2022.992666.

    Article  Google Scholar 

  19. McCartney DL, et al. Genome-wide association studies identify 137 genetic loci for DNA methylation biomarkers of aging. Genome Biol. 2021;22:194. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13059-021-02398-9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Codd V, et al. Polygenic basis and biomedical consequences of telomere length variation. Nat Genet. 2021;53:1425–33. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41588-021-00944-6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Gergei I, et al. GWAS meta-analysis followed by Mendelian randomization revealed potential control mechanisms for circulating α-Klotho levels. Hum Mol Genet. 2022;31:792–802. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/hmg/ddab263.

    Article  CAS  PubMed  Google Scholar 

  22. Folkersen L, et al. Genomic and drug target evaluation of 90 cardiovascular proteins in 30,931 individuals. Nat Metab. 2020;2:1135–48. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s42255-020-00287-2.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Jannot AS, Ehret G, Perneger T. P < 5 × 10(-8) has emerged as a standard of statistical significance for genome-wide association studies. J Clin Epidemiol. 2015;68:460–5. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jclinepi.2015.01.001.

    Article  PubMed  Google Scholar 

  24. Yu M, Li Y, Li B, Ge Q. Inflammatory biomarkers and delirium: a mendelian randomization study. Front Aging Neurosci. 2023;15:1221272. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fnagi.2023.1221272.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Staiger D, Stock J. Instrumental variables regression with weak instruments. Econometrica. 1997;65:557–86.

    Article  Google Scholar 

  26. Li B, Martin EB. An approximation to the F distribution using the chi-square distribution. Comput Stat Data Anal. 2002;40:21–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/S0167-9473(01)00097-4.

    Article  Google Scholar 

  27. Burgess S, Small DS, Thompson SG. A review of instrumental variable estimators for Mendelian randomization. Stat Methods Med Res. 2017;26:2333–55. https://doiorg.publicaciones.saludcastillayleon.es/10.1177/0962280215597579.

    Article  PubMed  Google Scholar 

  28. Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in mendelian randomization with some invalid instruments using a weighted median estimator. Genetic Epidemiol. 2016;40:304–14.

    Article  Google Scholar 

  29. Hartwig FP, Smith GD, Bowden J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol. 2017;46(6):1985–98. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ije/dyx102.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44:512–25. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ije/dyv080.

    Article  PubMed  PubMed Central  Google Scholar 

  31. 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. Nat Genet. 2018;50:693–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41588-018-0099-7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Greco MF, Minelli C, Sheehan NA, Thompson JR. Detecting pleiotropy in Mendelian randomisation studies with summary data and a continuous outcome. Stat Med. 2015;34:2926–40. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/sim.6522.

    Article  Google Scholar 

  33. Hemani G, et al. The MR-Base platform supports systematic causal inference across the human phenome. eLife. 2018. https://doiorg.publicaciones.saludcastillayleon.es/10.7554/eLife.34408.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Patsouras MD, Vlachoyiannopoulos PG. Evidence of epigenetic alterations in thrombosis and coagulation: a systematic review. J Autoimmun. 2019;104: 102347. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jaut.2019.102347.

    Article  CAS  PubMed  Google Scholar 

  35. Ewendt F, Feger M, Föller M. Role of fibroblast growth Factor 23 (FGF23) and αKlotho in cancer. Front Cell Develop Biol. 2020;8: 601006. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fcell.2020.601006.

    Article  Google Scholar 

  36. Buchanan S, Combet E, Stenvinkel P, Shiels PG. Klotho, aging, and the failing kidney. Front Endocrinol. 2020;11:560. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fendo.2020.00560.

    Article  Google Scholar 

  37. Kuro-o M, et al. Mutation of the mouse klotho gene leads to a syndrome resembling ageing. Nature. 1997;390:45–51. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/36285.

    Article  CAS  PubMed  Google Scholar 

  38. Clemens Z, et al. The biphasic and age-dependent impact of klotho on hallmarks of aging and skeletal muscle function. eLife. 2021. https://doiorg.publicaciones.saludcastillayleon.es/10.7554/eLife.61138.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Huang G, Wang P, Li T, Deng X. Genetic association between plasminogen activator inhibitor-1 rs1799889 polymorphism and venous thromboembolism: evidence from a comprehensive meta-analysis. Clin Cardiol. 2019;42:1232–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/clc.23282.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Meltzer ME, et al. Venous thrombosis risk associated with plasma hypofibrinolysis is explained by elevated plasma levels of TAFI and PAI-1. Blood. 2010;116:113–21. https://doiorg.publicaciones.saludcastillayleon.es/10.1182/blood-2010-02-267740.

    Article  CAS  PubMed  Google Scholar 

  41. Wang J, et al. Association between the plasminogen activator inhibitor-1 4G/5G polymorphism and risk of venous thromboembolism: a meta-analysis. Thromb Res. 2014;134:1241–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.thromres.2014.09.035.

    Article  CAS  PubMed  Google Scholar 

  42. Matías-García PR, et al. DNAm-based signatures of accelerated aging and mortality in blood are associated with low renal function. Clin Epigenetics. 2021;13:121. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13148-021-01082-w.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Zhang Y, et al. DNA methylation signatures in peripheral blood strongly predict all-cause mortality. Nat Commun. 2017;8:14617. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/ncomms14617.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Meier HCS, Mitchell C, Karadimas T, Faul JD. Systemic inflammation and biological aging in the Health and Retirement Study. GeroScience. 2023;45:3257–65. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s11357-023-00880-9.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

The study was supported by Wuhan Municipal Health Commission guided the project (WX2Z29). The funders had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript.

Author information

Authors and Affiliations

  1. Department of Cardiovascular Surgery, Wuhan Asia Heart Hospital, Wuhan City, 430000, Hubei Province, China

    Bowen Jin, Yunyan Li, Dingyang Li, Chi Jing & Qunshan Sheng

Authors
  1. Bowen Jin
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Yunyan Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Dingyang Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Chi Jing
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Qunshan Sheng
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Bowen Jin and Yunyan Li carried out the studies, participated in collecting data, and drafted the manuscript. Chi Jing and Qunshan Shen performed the statistical analysis and participated in its design. Dingyang Li and Bowen Jin participated in acquisition, analysis, or interpretation of data and draft the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Bowen Jin.

Ethics declarations

Ethics approval and consent to participate

This study utilized summary statistics from public GWAS studies, for which ethical approvement has been obtained. Consequently, no further ethical approval was necessary.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

13148_2025_1875_MOESM1_ESM.jpg

Supplementary file 1: Figure 1. Funnel plots for MR analyses of the causal associations between epigenetic aging factors and thromboembolism in forward analysis.

13148_2025_1875_MOESM2_ESM.jpg

Supplementary file 2: Figure 2. Leave-one-out analysis for the causal associations between epigenetic agingfactors and thromboembolism in forward analysis.

13148_2025_1875_MOESM3_ESM.jpg

Supplementary file 3: Figure 3. Funnel plots for MR analyses of the causal associations between thromboembolism and epigenetic aging factors in reverse analysis.

13148_2025_1875_MOESM4_ESM.jpg

Supplementary file 4: Figure 4. Leave-one-out analysis for the causal associations between thromboembolism and epigenetic aging factors in reverse analysis.

Supplementary file 5.

Supplementary file 6: Table 1. Basic characteristics of the study populations.

Supplementary file 7: Table 2. IVs used in this study.

Supplementary file 8: Table 3. Proxy SNPs used in this study.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Jin, B., Li, Y., Li, D. et al. Causal associations between epigenetic age and thromboembolism: a bi-directional two-sample Mendelian randomization study. Clin Epigenet 17, 75 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13148-025-01875-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13148-025-01875-3

Keywords

Clinical Epigenetics

ISSN: 1868-7083

Contact us