Coronavirus disease 2019 (COVID-19) is an epidemic disease caused by SARS-CoV-2, which is a sizable enveloped RNA virus, exhibiting approximately 80% sequence similarity with SARS-CoV, and predominantly impacts the respiratory system owing to its main attachment to angiotensin converting enzyme-2 (ACE-2) receptors (1, 2). SARS-CoV-2 can also commonly affect the cardiovascular, brain, kidney, liver, and immune systems, leading to their compromise (3). Myocarditis, pericarditis, acute coronary syndrome, arrhythmias, cardiomyopathy; heart failure, and thromboembolic disease are common cardiovascular damage manifestations of COVID-19 (2, 4).

Myocarditis is an inflammatory disease of the cardiac muscle, mainly diagnosed by histological, immunological, and immunohistochemical criteria (5). Pericarditis the most common form of pericardial disease refers to the inflammation of the pericardial layers (6). The etiology of myocarditis and pericarditis includes infectious and non-infectious forms. Viral infection and postviral immune-mediated responses are the most common causes of myocarditis (5). At the same time, viral infection is also a common cause of pericarditis. Sometimes myocarditis and pericarditis caused by viral infection can coexist (6, 7). According to some reports, myocarditis or pericarditis is a common cardiac complication caused by COVID-19 (8–10). During the COVID-19 pandemic, a study conducted by the Centers for Disease Control and Prevention (CDC) observed that the incidence of myocarditis infected with COVID-19 is 15 times that of non-infected COVID-19 (11). In European and American countries, it is estimated that the annual morbidity of acute pericarditis in the general population is about 27.7 cases per 100,000 people (12, 13). However, in the context of COVID-19 infection, real-world data demonstrate that new-onset pericarditis developed in approximately 1.5% of cases (14). More attention has been paid to myocarditis or pericarditis during the prevalence of COVID-19. Although there is growing evidence suggesting a potential association between COVID-19 and myocarditis and pericarditis, the precise causal relationship has not been definitively established. To date, limited reports with pathological evidence have demonstrated direct myocardial or pericardium invasion by the SARS-CoV-2 virus (15, 16). It is notable that even three years after the COVID-19 epidemic, our understanding of the pathogenesis of myocardial or pericardium injury in relation to COVID-19 remains incomplete. Several experts suggest that systemic inflammation may be the primary cause of myocardial injury, rather than direct viral infection of the heart (4, 17). The potential for long-term evolution into forms of inflammatory cardiomyopathy remains also unclear (18). Therefore, the relationship between COVID-19 and myocarditis and pericarditis still deserves further study, especially new research protocols, such as Mendelian randomization (MR).

Clinical research is susceptible to confounding factors and reverse causation. MR as a new and promising approach uses genetic variants as instrumental variables (IVs) to assess the genetic relationship between two phenotypes while avoiding potential confounding factors and reverse causation (19), which can be used to reveal the genetic sensitivity between COVID-19 and myocarditis or pericarditis. The reverse direction MR analysis was also performed (i.e., evaluating the causal association between myocarditis or pericarditis on COVID-19).

Using GWAS summary statistics, the 2-sample MR analyses did not indicate any causal effect of genetically predicted COVID-19 on the risk of myocarditis or pericarditis in individuals of European descent. The F-values of individual SNP and total SNPs are greater than 10. IVW models indicated that genetically predicted severe COVID-19 was not associated with an increased risk of myocarditis (OR = 1.00, 95% CI, 0.89–1.12; P = 0.99), and pericarditis (OR = 0.90, 95% CI, 0.78–1.04, P = 0.17). Similar results can be observed in hospitalized COVID-19, and COVID-19 infection (Figure 2). At the same time, null associations were observed for myocarditis or pericarditis with COVID-19 traits (Figure 3). IVW models showed thatmyocarditis was not associated with an increased risk of severe COVID-19 (OR = 0.96, 95% CI, 0.91–1.00; P = 0.07), hospitalized COVID-19 (OR = 0.97, 95% CI, 0.94–1.00, P = 1.00), and COVID-infection (OR = 0.99, 95% CI, 0.97–1.00; P = 0.09). Similar results can be observed in pericarditis. No directional pleiotropic effect across the genetic variants was detected in the MR-Egger regression (Tables 1, 2). No horizontal pleiotropic effect across the genetic variants was detected in the MR-PRESSO (Tables 1, 2). The results of the IVW analysis did not suggest the existence of heterogeneity (Tables 1, 2). Associations of each variant with exposure and the risk of outcome are shown in Supplementary Tables S1–S12. The scatter plot for exposure on outcome was shown in Supplementary Figures S1–S12A. A forest plot of exposure on outcome was shown in Supplementary Figures S1–S12B. The leave-one-out sensitivity analysis showed that the association between exposure and outcome was not substantially driven by any individual SNP (Supplementary Figures S1–S12C). The funnel plot was symmetrical around the IVW estimates, revealing that the causal association was less susceptible to potential bias (Supplementary Figures S1–S12D).

This study conducted the first-ever 2-sample MR analysis examining the association between COVID-19 and myocarditis or pericarditis utilizing large-scale GWAS data. There is no evidence that COVID-19 is independently and causally associated with myocarditis or pericarditis.

Some studies indicated that during the COVID-19 pandemic, there is a substantial increase in the incidence of myocarditis and pericarditis. In a study from March 2020 to January 2021, the incidence of myocarditis was 0.146% among patients diagnosed with COVID-19 and 0.009% among patients without COVID-19, with a 15-fold difference compared to each other (11). In another study, among 718,365 patients diagnosed with COVID-19, 35,820 (5.0%) occurred myocarditis, and 10,706 (1.5%) occurred pericarditis (14). However, in a study, the incidence of myocarditis and pericarditis among 181,656 hospitalized COVID-19 patients was 0.08% (21). Another large study reported that among 277 postmortem examinations, the true incidence of myocarditis was likely less than 2% (22). A published meta-analysis about the rate of myocarditis in COVID-19 fatalities was also less than 2% (23). Our careful analysis of the above articles can conclude that the large variation in the incidence of COVID-19-associated myocarditis is due to diagnostic heterogeneity. The definitive diagnosis of myocarditis should be based on established histological and immunohistochemical (ICH) criteria with endomyocardial biopsy or autopsy (24). Due to the relatively lower rates of endomyocardial biopsies (EMBs) performed in patients with COVID-19, the diagnosis of myocarditis relies more extensively on evidence of myocardial injury or the use of cardiac magnetic resonance imaging (CMR) in conjunction with clinical symptoms (25, 26). These methods are employed to meet the criteria outlined in the CDC working definition of myocarditis. For example, in an Italy study, researchers conducted an autopsy series on 40 patients who had died from SARS-CoV-2 infection. The authors performed a thorough histological analysis of the heart tissues. Interestingly, only two out of the 40 patients exhibited indications of myocarditis, and only one patient had clear evidence of lymphocytic myocarditis. Notably, despite observing histological signs of lymphocytic infiltrate in three cases, the researchers did not detect the presence of the virus using reverse transcription polymerase chain reaction (RT-PCR), in situ hybridization (ISH), and IHC techniques (27). At the same time, the diagnosis of most patients with COVID-19 associated pericarditis is still based on symptoms and imaging examinations, but not pathological examinations (28, 29). Differences in diagnostic criteria have a greater impact on the incidence of myocarditis and pericarditis during the COVID-19 epidemic (20). We analyzed that this reason may significantly affect the result of this Mendelian analysis.

The mechanisms of COVID-19-associated myocarditis or pericarditis are complex and mainly include direct viral damage, cytokine storm, immune system dysregulation, and endothelial dysfunction (9, 30). Firstly, direct viral damage includes endocytosis of the virus and degradation of cellular structures, and overall myocyte injury (31, 32). Secondly, cytokine storm, as a well-known physiological mechanism associated with SARS-CoV-2 infection, leads to the excessive production of pro-inflammatory cytokines. This overproduction of pro-inflammatory cytokines has been linked to myocardial injury, as evidenced by elevated troponin levels observed during these conditions. Following the initial infection of the lungs by SARS-CoV-2, a cytokine storm is generated, characterized by the release of various cytokines such as tumor necrosis factor-alpha (TNFα), interleukin-1β (IL-1β), and IL-6. Alongside this, extracellular vesicles (EVs) carrying the virus or virus particles are also released. These EVs have the potential to travel through the bloodstream or lymphatic system to the heart, where they can infect cardiac cells expressing specific receptors such as ACE2, transmembrane serine protease-2 (TMPRSS2), and neuropilin-1 receptor (NRP1) (33). Thirdly, immune-mediated myocarditis is another important mechanism wherein both the innate and acquired immune responses lead to myocardial injury, ultimately leading to dilated cardiomyopathy. Autoimmune-mediated myocarditis can occur when cryptic antigens in cardiomyocytes, are released following virus-mediated injury. In cases where viral antigens successfully evade the innate immune system, they replicate and produce virus-associated proteins that directly cause myocardial injury by inducing apoptosis and necrosis. Similar to other viral pathogens, SARS-CoV-2 may induce myocarditis in humans through a similar pathway. Noteworthily, cytokine storm promotes the activation of T-cells and releases cytokines to maintain the exaggerated immune response (34). The fourth mechanism involves the direct entry of the virus into endothelial cells within the heart, without necessarily infecting myocytes. This direct infection of endothelial cells has been observed in autopsy hearts and glomerular endothelial cells, through the use of electron microscopy. However, in some cases, the appearance and location of these virus particles within cells were not typical of those observed in coronavirus-infected cells (30, 35).

Although myocarditis or pericarditis caused by COVID-19 may be related to the above mechanisms, there is still some controversy. First, some studies believed that the evidence for direct damage to myocardial cells caused by viruses is not sufficient. In a study involving 39 consecutive autopsy cases from Germany, the viral load of SARS-CoV-2 was quantified. The results showed that SARS-CoV-2 was detected in the hearts of 24 out of 39 cases (61.5%), with 16 out of 39 cases (41%) exhibiting viral copy numbers higher than 1,000 copies per microgram of RNA. Virus replication was observed in five patients with the highest viral load. However, ISH confirmed the presence of the virus in interstitial cells within the cardiac tissue, rather than in the myocytes themselves. Notably, the presence of the virus was not associated with increased infiltration of mononuclear cells into the myocardium, and no cases of myocarditis were identified according to the Dallas criteria (15). At the same time, other researchers also have not found a SARS-CoV-2 genome detected within the myocardium (16, 36). Second, drug-induced myocarditis should also be considered because it has been confirmed in autopsy reports of COVID-19 deaths receiving antivirals (37, 38). Third, Takotsubo cardiomyopathy and myocarditis have similar CMR manifestations, mainly including injury to myocytes, and infiltration of lymphocytes and macrophages in the autopsied specimens. Some so-called COVID-19-related myocarditis cannot be ruled out as Takotsubo cardiomyopathy, especially in patients with a lack of evident myocardial SARS-CoV-2 viral genome (39).

Another concerning issue is confounding factors, such as immunodeficiency. A study showed an increased risk of more severe presentations and higher mortality caused by COVID-19 for patients with solid-organ transplants or cancer (40). According to a recent report, patients with cancer treated with checkpoint inhibitors were at particularly high risk of severe COVID-19 (41). Another study including 60 patients with primary immunodeficiency (PID) and 40 patients with secondary immunodeficiency (SID) included that, compared with the general population, adult patients with PID and symptomatic SID display greater morbidity and mortality from COVID-19 (42). A study concluded that myocarditis was found in 37 cases (52%) according to a retrospective evaluation of 71 consecutive necropsy patients who died from acquired immunodeficiency syndrome (AIDS) (43). Another study included that among 181 consecutive patients at all stages of HIV infection, 109 patients (60%) occurred different degrees of pericarditis (44). We could find that immunodeficiency as a confounding factor is not only associated with COVID-19 infection but also increases the incidence of myocarditis and pericarditis. Therefore, we cannot ignore that confounding factors have a great effect on the results of MR analysis.

Although there is already a large number of literature on the relationship between COVID-19 and myocarditis or pericarditis, our article still has outstanding novelty. Firstly, this is the first time that we have used a new research method—MR to clarify the causal relationship between them. Compared to traditional clinical studies, MR is less prone to potential confounding and avoids reverse causation because genetic variation is allocated at conception, and thus, it can strengthen the evidence for causal inference (45, 46). At the same time, in order to verify whether there is a reverse causal relationship, our article uses bidirectional MR. Secondly, our data comes from the latest GWAS database. The huge sample size and the latest data further improve the stability and credibility of the results. Finally, we conducted an assessment of pleiotropy, and heterogeneity and invited a professional from the data center to assist and verify our statistical and computational processes to ensure the stability and reliability of the results.

Our study has several limitations. Firstly, most of the included patients with myocarditis or pericarditis were diagnosed by clinical symptoms and signs, auxiliary examinations (electrocardiogram, myocardial enzymes, cardiac color Doppler ultrasound, CMR, etc.), and no EMB or autopsy was performed. This may have a significant impact on our analysis. Secondly, myocarditis or pericarditis is equivalent to the number of COVID-19 patients being too small, and the screened SNPs are relatively insufficient. Finally, the majority of participants were of European ancestry. Hence, the generalizability of the results cannot be derived for other races.

Our 2-sample MR analysis did not yield any supportive genetic evidence for a causal association between COVID-19 and an elevated risk of myocarditis or pericarditis. Although growing evidence indicates a causal relationship between COVID-19 and myocarditis or pericarditis, we need to consider the diagnostic heterogeneity of myocarditis or pericarditis and the effect of confounders. According to our conclusions and other evidence, the relationship between COVID-19 and myocarditis or pericarditis still deserves further research and observation.