Cardiovascular outcomes between COVID-19 and non-COVID-19 pneumonia: a nationwide cohort study

This large-scale, population-based study of 154,302 Korean patients did not reveal a higher incidence of post-acute cardiovascular outcomes in patients with COVID-19 hospitalization compared to those hospitalized with non-COVID-19 pneumonia. The findings were consistent after adjustment for preexisting cardiovascular disease. However, the risks were higher in patients with COVID-19 aged 18–39 years. COVID-19 patients had a higher risk of acute myocarditis or deep vein thrombosis, regardless of vaccination status. The risks were more pronounced among younger patients with less comorbidities and illness severity. Finally, the current results suggest that COVID-19 vaccination may prevent cardiovascular outcomes.

Several studies from the large US, UK, or global database have provided substantial evidence that the risk of cardiovascular outcomes was significantly higher in COVID-19 survivors compared to non-COVID-19 controls [6, 16, 17]. COVID-19 is associated with endothelial dysfunction and a prothrombic state [18], which can result in micro- and macrovascular thrombosis with subsequent injury to various organs [19]. Furthermore, putative mechanisms include fibrosis and scarring of cardiac tissue due to increased proinflammatory cytokine secretion and activation of transforming growth factor-β signaling and metabolic changes that favor viral survival, resulting in persistent inflammation and tissue injury [20, 21].

Despite these mechanisms, COVID-19 hospitalization was not associated with the increased risk of cardiovascular outcomes in the present study. There are several explanations for these observations. First, a weighted control group was selected from hospitalized patients with non-COVID-19 pneumonia. This study design was based on the mechanisms that the risk of cardiovascular disease may be increased by proinflammatory changes of the atherosclerotic lesions [22], persistent inflammation [23], and persistent procoagulant state [24], which can be observed in non-COVID-19 pneumonia. Comparable rates of organ dysfunction and intensity of organ support were found between the COVID-19 and non-COVID-19 pneumonia groups. Alternatively, previous studies compared outcomes after COVID-19 and general hospital admission [6, 16, 17], which lacks adequate consideration of differential risk by pneumonia severity, and thereby does not necessarily represent an appropriate counterfactual situation to COVID-19 hospitalization. Second, there may be geographic variations in the prevalence of COVID-19 and cardiovascular disease; most previous studies evaluated cardiovascular outcomes in the Western countries. The current study in an Asian population may be clinically relevant, given that angiotensin-converting enzyme 2 expression in tissues differs in the East Asian populations compared to other non-Asian populations under similar conditions [25].

The present study excluded patients who died or had any cardiovascular outcomes within the first 30 days of hospital admission, because cardiac injury is often associated with multiorgan dysfunction during critical illness rather than the disease itself [26]. This method may lead to an underestimation of the true risk of cardiovascular outcomes among patients with COVID-19. For instance, recent studies have shown that cardiovascular complications often preceded death in non-survivors, compared to patients who survived, and most events occurred within 30 days of infection [16, 27]. However, the incidence of outcomes were generally comparable between non-survivors with COVID-19 and non-COVID-19 pneumonia, although the COVID-19 group were more likely to have cardiovascular risk factors and greater illness severity. Moreover, the median time to MACE was longer than 30 days in both groups.

The finding that the increased risk of cardiovascular outcomes is confined to younger patients (aged 18–39 years), who are less likely to have comorbidities and classic cardiovascular risk factors, implies that COVID-19 may be an independent risk factor in this subgroup. Furthermore, the risks were attenuated after adjustment for preexisting cardiovascular disease. However, these should be interpreted with caution due to relatively small number of events producing wide CIs. Acute myocarditis associated with COVID-19 can present as a direct cardiac injury mediated by SARS-CoV-2, which could trigger a nonspecific inflammation using molecular mimicry between viral proteins and cardiomyocytes [28, 29]. The current finding that the risk of acute myocarditis was significantly higher in the COVID-19 group suggest that myocarditis may be triggered by SARS-CoV-2 independently of the occurrence of pneumonia. Furthermore, the increased risk of acute myocarditis, regardless of vaccination status, among younger COVID-19 patients with no cardiovascular comorbidities and less illness severity supports the notion that SARS-CoV-2 may act as an independent risk factor. Notably, the risk of acute myocarditis was more pronounced among COVID-19 patients without preexisting cardiovascular disease. A higher risk of acute myocarditis was observed in male COVID-19 patients, which contrasts with the Western studies showing that female COVID-19 survivors had a higher risk of myocarditis [17]. The risk of deep vein thrombosis was significantly higher in the COVID-19 group without preexisting cardiovascular disease, and this observation is consistent with the results of previous studies [5, 6, 16, 30]. Finally, the finding that vaccination against COVID-19 was associated with a decreased risk of cardiovascular outcomes is consistent with previous studies supporting vaccination, especially for patients with preexisting cardiovascular disease [12].

To the best of our knowledge, this is the first study to assess whether COVID-19 hospitalization was associated with cardiovascular outcomes compared to hospitalization for pneumonia due to other respiratory pathogens in a nationwide cohort of Asian patients. The study derived its cohorts from an entire population in Korea to increase generalizability and had sufficient power to quantify multiple and rare outcomes overall and across various subgroups. In addition, all available baseline covariates were included in the PS used for weighting, and the primary results were generally consistent with multiple sensitivity analyses. Previous studies may have misclassified individuals due to the inherent limitations of using claims data to define variables. For instance, some COVID-19 patients might have been enrolled in the control group if they did not have enough symptoms to be tested. To overcome this limitation, controls were recruited from patients hospitalized before the pandemic (January 1–December 31, 2019) in the present study. Cerebrovascular outcomes and myocardial infarction were rigorously defined using the ICD-10 codes and related imaging or procedures rather than just the codes to obtain more reliable results.

The current study has several limitations. First, the retrospective observational design precludes any causal inference regarding the associations between COVID-19, non-COVID-19 pneumonia, and cardiovascular outcomes. Moreover, there were differences in baseline characteristics between the groups. A robust model was used to reduce the effect of such differences, but the possibility of unmeasured confounders exists. Second, the accuracy of the ICD-10 codes may be limited. It is possible that all ICD-10 codes for each outcome might not have been captured, although the most common ones were included. In addition, patients who had asymptomatic infections or resided in nursing homes for palliative care who were not hospitalized may also develop post-acute complications but were not included. Meanwhile, this approach might have described a more specific population of severe pneumonia. Third, the database did not include information regarding vital signs and laboratory data, but diagnoses, prescriptions, and procedures were used as surrogates of pneumonia severity. Fourth, the diagnosis of acute myocarditis was not supported by histological evidence or cardiac magnetic resonance imaging [31, 32]. Furthermore, a very small number of events (< 100) may be a limiting factor in analysis. Fifth, the increased risk of acute myocarditis in COVID-19 patients might be a result of increased medical care after SARS-CoV-2 infection. However, this ascertainment bias is unlikely to fully explain the present results because the risk of other cardiovascular outcomes was not higher in the COVID-19 group. Sixth, the cardiovascular manifestations of COVID-19 may change over time with the emergence of new variants that induce varying degrees of severity [33]. However, a subgroup analysis assessing the risk of cardiovascular outcomes according to different SARS-CoV-2 variants was not feasible due to lack of data. Seventh, unequal follow-up in the study groups, probably due to differing distribution of the index dates, may bias the analysis. In addition, the follow-up time was relatively short, although the median time to MACE was less than 100 days for both the COVID-19 and non-COVID-19 pneumonia groups.