Cardiovascular diseases (CVDs) are the leading cause contributing to mortality and morbidity worldwide. In China, national economic losses caused by CVDs are estimated to approach $8.8 trillion between 2012 and 2030 (Bloom, 2013). Although multifactorial contributors such as genetic factors, environmental factors, and immune and inflammatory factors are identified and under exploration in the pathogenesis of CVDs, studies have increased in recent indicating the role of microbiome (Finlay and Humans, 2020) in non-communicable diseases (NCDs) including CVDs, with the Human Microbiome Project initiated over 2007 and 2016. Recent studies suggest that the human microbiome, including oral and gut, contributes to the pathogenesis of inflammation (Alam et al., 2017; Bruce-Keller et al., 2018) and immune dysfunction (Kong et al., 2019), which may contribute to CVD pathophysiology.

In addition to the bacterial microbiome and their metabolites already analyzed as a pathogenic factor, biomarker, and potential intervention targets in NCDs, viruses (Zou et al., 2016) (virome) constituting a portion of the microbiome present with limited investigations. Virus infections are long-term issues for associations with CVDs (Kawana, 1985; Pothineni et al., 2017; Du et al., 2018) via multiple epidemiological, serological, molecular, and animal studies. Of note, the high cardiovascular manifestations under the COVID-19 pandemic (Chung et al., 2021) further underscore the importance of more explorations of viruses on CVDs. A recent study has also revealed that circulating cell-free DNA of microbiome in CVDs are enriched with bacteriophages and eukaryotic viruses (Dinakaran et al., 2014) in circulation compared with healthy controls. Moreover, viral metagenomic profiling of fecal samples in patients with coronary heart disease revealed reduced gut virome of the family Microviridae ( Guo et al., 2017 ), a viral taxon dominant in the healthy population, but the direct correlation between virome and coronary heart disease remains unclear. The above evidence indicates that viruses (virome) are involved in the development of CVDs.

Cytomegalovirus (CMV) is one of the double-stranded DNA (dsDNA) viruses that belong to the herpes virus family with relatively common prevalence and is under investigation for association with CVDs. Many observational studies have revealed the association between CMV and CVDs. A meta-analysis of community-based prospective studies covering 34,564 subjects and 4,789 patients with CVDs including coronary artery disease, ischemic heart disease, stroke, and heart failure reveals a 22% increase in CVD (Wang et al., 2017) incidence risk after CMV infection. Seropositivity of CMV, particularly CMV immunoglobulin (Ig) G levels (Roberts et al., 2010), are also observed as predictors of all-cause and/or CVD-related mortality in elderly Swedes (Wikby et al., 2005), Latinos (Hadrup et al., 2006), and Americans with higher C-reactive protein (CRP) levels (Simanek et al., 2011). In addition, although not in total accordance, histopathological studies (Zhu and Liu, 2020) detected CMV DNA in serum of coronary artery disease patients (Westphal et al., 2006) and atherosclerotic plaque or vascular wall specimens (Shi and Tokunaga, 2002; Gredmark-Russ et al., 2009). A rare propensity of vascular thrombosis after CMV infections has also been observed since the 1980s (Boers and Haak, 1984). The potential contribution of CMV to multiple CVDs may thus be indicated. However, it is worth noting that we cannot provide evidence for a causal association between CMV and CVDs. First, all the aforementioned findings are based on observational studies. Therefore, no conclusion can yet be reached about whether individuals infected with CMV would be at higher risk of CVDs or vice versa. Moreover, findings based on observational studies may have been affected by some confounders, even some unknown or unmeasured risk factors.

Mendelian randomization (MR) is a novel computational method to assess the causal associations between risk factors and particular diseases (Smith and Ebrahim, 2003; Davey Smith and Hemani, 2014; Davies et al., 2018) due to the following reasons: 1) the allocation of genetic variants is entirely random; 2) the inherited genetic variants have been determined at conception and will not change due to non-differential measurement error or confounding. In this study, we first investigate the role of CMVs in all available CVDs in the databases, including atrial fibrillation, coronary artery disease, hypertension, heart failure, peripheral artery disease, pulmonary embolism (PE), deep vein thrombosis (DVT) of the lower extremities, rheumatic valve diseases, and non-rheumatic valve diseases, by using 2-sample summary MR based on the genetic variants of anti-cytomegalovirus IgG levels as instruments.

Herein, we explored the causal association between genetically predicted anti-cytomegalovirus IgG levels and the risk of 9 CVDs. In the present MR, we first found direct evidence that genetically predicted higher anti-cytomegalovirus IgG levels were causally associated with a higher risk of coronary artery disease, peripheral artery disease, and DVT.

The human microbiome includes the oral and gut microbiome. Many studies have revealed that both gut and oral microbiome may be modified by systemic diseases, including CVDs (Graves et al., 2019). The collection of microbes living in the human intestinal tract and oral cavity has been known to influence more than digestion. Indeed, the microbiota, which collectively is widely considered the body’s largest endocrine organ, can generate biologically active metabolites and impact many aspects of host physiology (Witkowski et al., 2020). Human CMV is a prevalent beta-herpes virus with approximately 50% European and American population under infection (Bate et al., 2010) and is difficult to be cleared from the host. With such common prevalence and higher cardiovascular complications after CMV infection in observational studies, whether the observational association is causal remains unclear. CMV infection enables a lifetime latent phase with capacities of infecting cells, generally epithelial cells and peripheral blood mononuclear cells, and reactivation with higher titers of IgG. Interestingly, CMV also affects vessel wall cells, with its infection linked to coronary heart disease and atherosclerosis (Wang et al., 2017; Zhu and Liu, 2020), which was also consistent with the conclusions of our study, revealing a causal association between genetic predisposition of anti-cytomegalovirus IgG levels and higher risk of coronary artery disease. Notably, our study first establishes the causative correlation for CMV IgG level in peripheral artery disease, a disease that mainly shares similar mechanisms of atherosclerosis with coronary artery disease but involves peripheral vessel beds, in line with a population-based case–control study (Bloemenkamp et al., 2002) previously observing a positive correlation in women between CMV IgG titer and peripheral artery disease (OR = 1.6, 95% CI, 1.1–2.3). The presence of CMV antigen and DNA is observed in both human cardiovascular samples and animal models (Melnick et al., 1983; Melnick et al., 1994). CMV as an intracellular pathogen could involve diverse phases of atherogenesis (Pothineni et al., 2017) including endothelium activation, leucocyte migration, lipid core formation, smooth muscle proliferation, plaque stability, and thrombus formation (Strååt et al., 2009). Other systemic mechanisms of inflammation and host immune response interplaying with CMV (Forte et al., 2020), and molecular mimicry between serum antibodies of CMV and host proteins such as heat shock protein 60 (Mayr et al., 2000), are also in a recent investigation. The CMV and its IgG antibody as biomarkers for atherosclerotic diseases such as peripheral artery disease and coronary artery disease, and their pathogenic roles need further exploration. Additionally, CMV is a novel target for CVD prevention and treatment via therapies such as antiviral drugs, chimeric antigen receptor T cells, immunotherapy, or CMV vaccines (Vasilieva et al., 2020) (i.e., Triplex, a Modified Vaccinia Ankara (MVA) vaccine encoding CMV antigens (La Rosa et al., 2017; Aldoss et al., 2020), Poxvirus Vectored Cytomegalovirus Vaccine) may be developed.

Moreover, we first confirmed the causal association between genetic predisposition of anti-CMV IgG level and DVT, a thrombotic disorder of the veins with a higher risk for life-threatening PE. Although with the limited investigation, several case reports or case series have revealed a trend for correlation between CMV infection and venous thrombosis-related diseases such as venous thromboembolism (VTE; the combination of DVT and PE) (Ceccarelli et al., 2018), portal vein thrombosis (Squizzato et al., 2007), and superior mesenteric vein thrombosis (Walter et al., 2021). Additionally, case–control studies observed a higher incidence rate of thrombosis in the CMV infection group (6.4% vs. 0%) (Atzmony et al., 2010) and a higher rate of anti-CMV IgG positivity in the VTE group compared with healthy controls (Schimanski et al., 2012). A meta-analysis analyzed 97 reports of thrombosis associated with CMV infection, presenting DVT and PE as the most prevalent diseases, particularly in immunocompromised populations (Justo et al., 2011). A recent retrospective study analyzed 1,007 VTE patients and observed 0.1% co-occurrence with acute CMV infection, predominantly in women and the younger population (Yildiz et al., 2016). In addition, a prospective study covering approximately 90,000 VTE patients demonstrated anti-CMV IgM seropositivity as an over 2-fold independent risk factor for VTE onset (Paran et al., 2013). Plausible mechanisms for thrombosis after CMV infection include the following: 1) infection of endothelial cells promoting their procoagulant activity, such as Virchow’s triad activation (Sherman et al., 2014); 2) transient hypercoagulable state after anti-phospholipid antibody formation caused by CMV-derived peptide, which mimics human beta-2 glycoprotein 1 (Uthman and Gharavi, 2002; Gharavi et al., 2002; Nakayama et al., 2014); 3) host response to phospholipids from CMV envelopes with procoagulant propensities, for instance, phosphatidylserine (Pryzdial and Wright, 1994); 4) platelet aggregation via upregulated production of von Willebrand factor and expression of ICAM-1 and VCAM-1 (Rahbar and Söderberg-Nauclér, 2005). Interestingly, the host procoagulant reaction to CMV infection appears to be transient with a window period after infection and thrombosis (Justo et al., 2011). Anticoagulant and antiviral therapy choices and duration in those populations may be further explored with close monitoring for virus and host statuses.

In the present MR, we have implemented some key measures to satisfy the assumptions of MR analysis. First, only statistically significant SNPs of anti-cytomegalovirus IgG levels based on a genome-wide significant level valid association were used in our MR analysis to satisfy the IV assumption 1 as described in our methods. To minimize the bias and confounders in anti-cytomegalovirus IgG levels and CVDs, we only chose the GWAS of anti-cytomegalovirus IgG levels and CVDs, which were finished just in European ancestry populations. Therefore, the potential bias and confounders in the present study are small. To ensure that these SNPs only affect CVDs through anti-cytomegalovirus IgG levels (no pleiotropic effects), MR-Egger regression was performed, and no evidence of directional pleiotropic effects was observed in our MR.

Despite these strengths, some limitations are still worth noting. First, anti-cytomegalovirus IgG levels can only represent a state after infection, and it is not sufficient for us to assess the viral infection activity or acute infection. Second, even though the effect size of anti-cytomegalovirus IgG levels is quite modest, it is estimated that a large number of patients are at an increased risk of coronary artery disease, peripheral artery disease, and DVT because of the large population with anti-cytomegalovirus IgG. Most importantly, it has been widely accepted that the risk factors are various in populations with different races and ethnicities (Tan et al., 2021). However, all of the summary-level statistics in the present MR were based on European ancestry populations. Therefore, further studies are needed to clarify whether the conclusions apply to other populations.

In essence, our MR analysis first revealed the casual association between genetic predisposition of anti-cytomegalovirus IgG levels and the risk of coronary artery disease, peripheral artery disease, and DVT. Future studies on CMV infection in the pathogenesis of atherosclerosis and thrombosis, as well as anti-CMV therapy development for disease prevention, are of significance. In addition, preventive anticoagulation treatment may be investigated for individuals infected by CMV.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material.

Ethical approval was not provided for this study on human participants because the summary data were previously published and the ethics approvals have been obtained in their institutions. Therefore, no additional ethics approvals were required in this study. The patients/participants provided their written informed consent to participate in this study.

JC contributed to the conception and design of the study. J-ST organized the database. J-ST and LF performed the statistical analysis and wrote the first draft of the manuscript. YY contributed to interpret the results and review the revised manuscript. J-MR and S-SZ helped with the revision of the manuscript. All authors contributed to manuscript revision and read and approved the submitted version.

The work was supported by grants from Capital‘s Funds for Research and Application of Clinical Diagnosis and Technology(Z191100006619121), Treatment National Center for Clinical Research in Cardiovascular Diseases (No. NCRC2020015), the CAMS Innovation Fund for Medical (CIFMS, Sciences 2021-I2M-1-007 and 2022-GSP-GG-26), Natural National Science Foundation of China (Project ID. 81825002), Beijing Outstanding Young Scientist Program (Project ID. BJJWZYJH01201910023029), and the project for the distinguishing academic discipline of Fuwai hospital (Project ID. 2022-FWQN02).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent 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.