Cardiac adverse drug reactions to COVID-19 vaccines. A cross-sectional study based on the Europe-wide data

Abstract

Aims

We aimed to analyse serious cardiac adverse drug reactions to COVID-19 vaccines from the Europe-wide EudraVigilance database.

Methods and results

In this retrospective, cross-sectional study, the EudraVigilance database was searched to identify suspected serious cardiac post-vaccination adverse drug reactions to COVID-19 vaccines. This data was coupled with the number of total vaccine doses administered in the European Economic Area for Comirnaty (Pfizer BioNTech), Spikevax (Moderna), Vaxzevria (AstraZeneca), Jcovden (Janssen), Nuvaxovid (Novavax), products, available from the European Centre for Disease Prevention and Control ‘Vaccine Tracker’ database. The analysis included 772 228 309 administered doses of eligible vaccines from the ‘Vaccine Tracker’ database and 86 051 eligible records of cardiac adverse drug reactions from the EudraVigilance database. The frequency of most of the investigated adverse drug reactions was very rare (<1/10 000 i.e. <100/1 000 000 doses). The lowest risk of any serious cardiac adverse drug reactions was noticed for vaccination with Comirnaty (135.5 per million doses), while Spikevax, Jcovden, Vaxzevria, and Nuvaxovid were characterized by higher risk (respectively, 140.9, 194.8, 313.6, and 1065.2 per million doses). The most common complications of vaccinations included syncope, arrhythmia, tachycardia, palpitations, angina pectoris, hypertension, myocarditis, thrombosis, and pulmonary embolism.

Conclusion

The risk of serious cardiac adverse drug reactions to COVID-19 vaccines is low and the benefit of active immunization against that disease seems to outweigh the potential risk of serious post-vaccination cardiac adverse drug reactions.

Graphical Abstract

Open in new tabDownload slide

Introduction

Vaccination against Coronavirus disease 2019 (COVID-19) has significantly reduced the likelihood of hospitalization and deaths resulting from SARS-CoV-2 infection.^1–4 The benefit of active immunization against COVID-19 substantially outweighs the potential risk of post-vaccination severe adverse drug reactions (ADRs).^1–4 On the contrary, COVID-19 vaccinations are linked to various suspected ADRs from many human organs, including skin, respiratory, allergic, neurological as well as cardiac manifestations.^1,5,6 According to the European Medicines Agency (EMA), an ADR is a noxious and unintended response to a medicine.⁷ EMA also defines a serious ADR as ‘a reaction that corresponds to any untoward medical occurrence that at any dose results in death, is life-threatening, requires inpatient hospitalization or prolongation of existing hospitalization or results in significant disability’.⁸ Serious ADRs to COVID-19 vaccinations requiring hospitalization, organ failure, and/or resulting in death are reported rarely.^1,2 To date, the most common suspected serious cardiac ADRs reported for COVID-19 vaccinations include thrombosis, stroke, myocarditis, myocardial infarction, pulmonary embolism, and arrhythmia.^6,9–11 Many of them are life-threatening events and have a substantial impact on quality of life, reduce life expectancy, and may result in death.^9–11 However, there is limited information about the frequency of serious cardiac ADRs to the most commonly administered COVID-19 vaccines. Some studies analyse selected ADRs (e.g. acute myocardial infarction, thrombosis, or myocarditis) or are based on case reports.^9–15 Thus, sometimes it is impossible to compare the results of different studies due to methodological differences. Therefore, there is a need for large-scale studies that will estimate the frequencies of many ADRs in a single analysis. This will provide detailed and comparable insights into a range of potential cardiac ADRs after immunization against COVID-19.

Suspected ADRs to COVID-19 vaccines are reported in the EudraVigilance (EV) database.⁸ EMA groups the frequency of ADRs using the following nomenclature: very common (≥1/10); common (≥1/100 to <1/10); uncommon (≥1/1000 to <1/100); rare (≥1/10 000 to <1/1000); very rare (<1/10 000).^1,16–20 The incidence rates are optimally studied during cohort studies and registration clinical trials with long-term follow-up.^1,16–20 However, given the complexity and expensiveness of such studies as well as the rarity of observed cardiovascular complications, it is hard to find the exact frequencies of the cardiac ADRs.^11,33

Thus, there is a need for studies monitoring a large population to estimate the frequency of potential cardiac ADRs observed after vaccination against COVID-19. This would uncover new information and deepen the understanding of the frequency of many suspected cardiac ADRs to COVID-19 vaccines.

Aim

This study aims to analyse the burden of suspected serious cardiac ADRs following vaccination against COVID-19 from the Europe-wide EudraVigilance database.

Materials and methods

This is a retrospective, cross-sectional study that was performed according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies.²¹

Database search

On 21 January 2024, a database search of the EudraVigilance database was performed to identify any suspected post-vaccination ADRs to COVID-19 vaccines. No filters were used. The preliminary search returned a total of anonymized 2 302 495 ADRs to 14 COVID-19 vaccine products (Figure 1).²²

Moreover, on the same day, the ECDC ‘Vaccine Tracker’ (VT) database was searched to collect data on the total number of vaccine doses administered to the EEA for the most extended period available, from 7 December 2020 to 5 October 2023. For all vaccines approved in the EU/EEA, over 1.4 billion doses were distributed, and over 980 million doses were administered.²³

EudraVigilance database filtering

To match the VT database, any potential ADRs reported outside the EEA region were excluded (n = 714 578). In addition to that, ADRs reported as non-serious events were also removed (n = 1 174 295). Due to a low number of suspected ADRs for Novavax XBB.1.5 (NVX-COV2373), Valneva, Vidprevtyn Beta, and Spikevax XBB1.5 (n = 1, n = 3, n = 4, and n = 9, respectively), these vaccines were excluded from further analyses.

Further on, to match the period from the VT database, any reports added after the 5th of October 2023 were removed from the EV database (n = 5051). Moreover, suspected ADRs with unspecified sex (n = 5597) were also excluded. As no records for Spikevax Omicron BA.4-5 (Moderna) met the previously defined criteria, this vaccine was excluded from further analysis. All evaluated vaccine products were categorised into five groups, according to their manufacturing company and commonly used names: Pfizer-BioNtech (Comirnaty, n = 265 737 records), AstraZeneca (Vaxzevria, n = 71 386), Moderna (Spikevax, n = 47 940), Janssen (Jcovden, n = 17 546), and Novavax (Nuvaxovid, n = 348, Supplementary material online, S1). In total, there were 402 957 records available (Figure 1). The VT database extracted the vaccine doses administered in the EEA countries for five analysed vaccines.²³ In total, 772 228 309 doses of eligible vaccines were administered.

Identification of cardiac adverse drug reactions from EudraVigilance

Next, to find ADRs related to the cardiac system, a list of all available pathologies that were classified as ‘cardiac disorders’ and ‘vascular disorders’ was downloaded from the EV database (n = 527, Supplementary material online, S2). For 59 ADRs, no records were found. Moreover, 227 ADRs were excluded due to a very small number of reported cases (n <10 for each ADR).

Next, the remaining ADR pathologies were reviewed by experts in cardiac medicine. ADR pathologies that were (1) classified as valvular heart disease with highly improbable ‘post-vaccination’ aetiology; (2) whose descriptions were too general; (3) were classified as ‘other ADRs’ (not directly related to heart disease and/or chronic conditions with highly improbable ‘post-vaccination’ aetiology) were removed from further analysis (n = 90). Finally, the eligible specific ADR names (n = 151) were classified into 10 broader ‘clinical’ categories: arrhythmia, inflammation, ischaemia, heart failure, life-threatening event, conduction disorder, cardiomyopathy, blood pressure abnormality, thrombosis, syncope (Supplementary material online, S2). The classification into categories was performed by an expert in cardiology (LDS).

Data analysis

Overall, 402 957 records were available, of which 86 051 were classified as suspected cardiac ADRs using the eligible cardiac 151 ADR names (Figure 1, Supplementary material online, S2).

The records were analysed according to three general characteristics: primary source qualification (healthcare/non-healthcare professional), sex (male/female), and age category (under 18 years old, 18–65 years old, 65–85 years old, over 85 years old).

To calculate the frequency of a given suspected ADR per one million administered vaccine doses, the data from the EV database on possible serious cardiac ADRs was coupled with the total number of doses administered for each vaccine product from the ECDC database (Tables 1, 2, Supplementary material online, S3). An interactive online web application was created to visualize the data presented in Tables 1 and 2. The application can be downloaded as a Supplementary material online, S4. The application should be opened with a web browser (e.g. Google Chrome or Apple Safari).

Machine learning

Machine learning algorithms were developed following the Proposed Requirements for Cardiac Imaging-Related Machine Learning Evaluation (PRIME) framework.²⁴

The aim of machine learning was (1) to predict a given category of suspected ADRs; (2) to explore clinical characteristics that could be identified as potential risk factors for a given ADR category. This approach may provide a foundation for personalized vaccinations based on accurate vaccine-patient matching using the patient's clinical characteristics. It can ultimately improve patient safety by avoiding the administration of vaccines that may carry a higher risk of triggering a serious ADR in certain groups of patients.

Logistic regression, Random Forest and XGBoost models were trained for each ADR category using their default hyperparameters. These algorithms were chosen as they can be trained using the balanced class weights (the algorithm adjusts for the unequal frequencies of classes in classification tasks), and they fit in the scope of explainable and interpretable artificial intelligence.

Categorical input variables included primary source qualification, sex, age, and vaccine. The primary endpoint was a binary outcome defined as the presence or absence of the ADR category. Mean accuracy, balanced accuracy, sensitivity, specificity, and concordance index (c-index) were calculated from five-fold cross-validation to evaluate the models.

Statistical analysis

Categorical data were summarized with the use of frequencies and proportions. The Chi-square test with Yates’ correction was used to test for statistical significance.

The threshold of statistical significance was set at P < 0.05. All analyses were performed using Python 3.10 as well as Pandas 2.1.3, Scikit-learn 1.2.1, and XGBoost 2.0.1 libraries.

Results

General characteristics

In total, the analysis included 772 228 309 administered doses of eligible vaccines from the ‘Vaccine Tracker’ database and 86 051 eligible records of suspected cardiac ADRs from the EudraVigilance database (Figure 1, Table 1, Supplementary material online, S1 and S2).

The greatest absolute number of suspected cardiac ADRs was reported for Comirnaty (n = 56 325) followed by Spikevax (n = 13 816), Vaxzevria (n = 13 424), Jcovden (n = 2328), and Nuvaxovid vaccines (n = 158, Supplementary material online, S6). Most of the potential ADRs were reported by non-healthcare professionals and ranged from 52.8% for Vaxzevria to 73.4% for Nuvaxovid (P < 0.001). Moreover, the vast majority of suspected ADRs were reported for patients aged 18–64 years old (ranging from 65.1% for Vaxzevria to 88.6% for Nuvaxovid, P < 0.001). In addition to that, only for Jcovden, there were more reports for male patients (58.1%), while for all other vaccines, the women constituted 55.0% (Spikevax) to 67.1% (Nuvaxovid) of the investigated cohort (P < 0.001).

Frequency of suspected cardiac ADRs

All frequencies of potential ADRs described in this section are considered as the number of reported ADRs per million administered vaccine doses.

Overall, vaccination with Comirnaty was associated with the lowest risk of any serious cardiac suspected ADR (135.5) and was followed by Spikevax (140.9), Jcovden (194.8), Vaxzevria (313.6), and Nuvaxovid (1065.2, Figure 2).

For Comirnaty, arrhythmias (12.1), tachycardia (10.8), pulmonary embolism (10.1), palpitations (9.7), and syncope/myocarditis (8.2 each) were most frequently reported (Table 1). For Spikevax, these were arrhythmia (12.7), syncope (12.4), tachycardia (11.1), palpitations (10.7), and myocarditis (10.3). For Jcovden, the most prevalent potential ADRs were syncope (30.8), pulmonary embolism (15.9), arrhythmia (14.8), tachycardia (14.1), and palpitations (10.7). For patients vaccinated with Vaxzevria pulmonary embolism (39.0), deep vein thrombosis (28.2), syncope (26.1), tachycardia (22.1), and thrombosis (19.7) were most frequently noted. Subjects vaccinated with Nuvaxovid reported most commonly tachycardia (168.7), arrhythmias (159.8), palpitations (93.2), hypertension (75.5), and angina pectoris (66.6). Compared to other vaccines, vaccination with Nuvaxovid was more frequently associated with many sporadic suspected ADRs, for example, blood pressure fluctuation (17.8 vs. maximally 1.8 for Vaxzevria), pericardial effusion (13.3 vs. maximally 1.7 for Comirnaty), myopericarditis (17.8 vs. maximally 3.2 for Spikevax), circulatory collapse (26.6 vs. maximally 4.4 for Jcovden), or angina pectoris (66.6 vs. maximally 6.1 for Jcovden).

All investigated suspected ADRs (n = 151) are available in Supplementary material online, S2.

Frequency of categories of suspected cardiac ADRs

All frequencies of potential ADRs described in this section are considered as the number of reported ADRs per million administered vaccine doses.

For Comirnaty, the most widely administered vaccine, suspected ADRs from arrhythmia (35.7), thrombosis (21.8), and inflammation (17.4) categories were most commonly observed, whereas cardiomyopathies (0.8), conduction disorders (1.8), and heart failure (3.1) were the least frequently reported categories of ADRs (Table 2). For Spikevax, the most frequently reported potential ADRs were arrhythmias (37.1), inflammations (21.2), and thromboses (20.0). Inoculation with Jcovden was most frequently linked to events of arrhythmias (41.3), thromboses (36.2), and syncopes (35.0). The use of Vaxzevria was reportedly most commonly associated with thrombosis (99.0), arrhythmias (68.0), and syncopes (30.0). For Nuvaxovid, the most frequently reported events included arrhythmia (368.4), blood pressure abnormalities (106.5), and thrombosis (97.6). On the contrary, the least common were ADRs from cardiomyopathy (4.4), heart failure (13.3), and conduction disorder (26.6) categories. The same types of ADRs were the rarest ones for Spikevax (0.8 for cardiomyopathies, 1.6 for conduction disorders, and 2.6 for heart failure), Jcovden (1.2, 2.2, and 2.7, respectively) as well as for Vaxzevria (1.2, 3.5, and 4.8, respectively).

The frequency of the reported life-threatening events was low for Comirnaty (4.6) and Spikevax (4.6), moderate for Jcovden (8.6) and Vazxevria (10.1) and high for Nuvaxovid (31.1). Moreover, for Nuvaxovid, higher reporting rates of suspected ADR categories per million doses are observed than for other investigated vaccines.

Overall, there were statistically significant differences for all 30 most commonly reported ADRs and all 10 broader categories of ADRs (P < 0.05 for each comparison).

Machine learning

The prediction capability of a given ADR category using machine learning algorithms was low to moderate (Supplementary material online, S5). The highest achieved mean balanced accuracies were 0.70 (95% CI 0.70–0.70) for heart failure using logistic regression (c-index 0.75), 0.68 (95% CI 0.68–0.68) for arrhythmia using Random Forest (c-index 0.71) and 0.66 (95% CI 0.66–0.66) for inflammation using XGBoost (c-index 0.72). The lowest prediction accuracy was calculated for blood pressure abnormalities (0.58 using XGBoost, 95% CI 0.58–0.58; c-index 0.62), conduction disorders (0.59 using logistic regression, 95% CI 0.59–0.59; c-index 0.61) and cardiomyopathies (0.61 using XGBoost, 95% CI 0.61–0.61, c-index 0.64).

There was no one clinical characteristic that would be a strong risk factor for a given suspected ADR category. Most commonly, algorithms identified primary source qualification (healthcare professional/non-healthcare professional) as the most predictive factor. Interestingly, vaccination with Vaxzevria and female sex were identified as the most predictive factors for potential ADRs for the ‘Inflammation’ category by all algorithms.

Discussion

Based on our cross-sectional analysis of over 86 000 suspected serious cardiac ADRs following over 772 000 000 COVID-19 vaccinations, it seems that (1) cardiac ADRs to COVID-19 vaccines are very rare; (2) vaccination with Comirnaty (Pfizer-BioNTech) is associated with the lowest risk of cardiac complications followed by Spikevax (Moderna), Jcovden (Janssen), Vaxzevria (AstraZeneca), and Nuvaxovid (Novavax); (3) mRNA vaccines provide enhanced safety in comparison to adenovirus-based vector vaccines and protein-based vaccines.

Suspected cardiac ADRs

According to the official product characteristics authorized by EMA, during the clinical trials and post-authorization experience in individuals 12 years of age and older, after vaccination with Spikevax, Comirnaty, or Nuvaxovid, the only cardiac ADRs were very rare cases of myocarditis and pericarditis (<1/10 000 i.e. <100/1 000 000 administered doses).^18–20 For Jcovden, myocarditis and pericarditis (frequency unknown), as well as thrombosis with thrombocytopenia (very rare) and venous thromboembolism (rare, ≥1/10 000 to <1/1000), were documented.¹⁷ Following the product characteristics, vaccination with Vaxzevria was very rarely associated with ‘blood clots often in unusual locations’ (thrombosis) as well as brain thrombosis with thrombocytopenia or venous thromboembolism (frequency unknown).¹⁶ Moreover, EMA officially warned about a ‘possible link between the Vaxzevria vaccine and very rare cases of unusual blood clots with low blood platelets’ and ‘confirms overall benefit-risk remains positive’.²⁵

In our study, vaccination with Comirnaty was associated with the lowest risk of any suspected serious cardiac ADR (135.5 per million doses). It was followed by Spikevax (140.9 per million doses), Jcovden (194.8 per million doses), Vaxzevria (313.6 per million doses), and Nuvaxovid (1065.2 per million doses, Figure 2). Therefore, mRNA vaccines provide enhanced safety in comparison to adenovirus-based vector vaccines and protein-based vaccines. Moreover, we also observed similar results to the official product characteristics from EMA, as the frequency of the majority of the investigated suspected ADRs per number of administered vaccine doses was very rare, except for rare ADRs for Nuvaxovid from the arrhythmia and blood pressure abnormalities categories (368.4 and 106.5 per million doses, respectively). Moreover, another discrepancy is the frequency of venous thromboembolism for Jcovden that EMA documented as rare. However, in our study, any suspected event from the thrombosis category for Jcovden was observed only for 36.2 per million administered doses, which suggests a very rare frequency.¹⁷

Further, official documentation mentions cardiac ADRs from only two major categories identified in our study: inflammation and thrombosis.^16–20 According to our study, the risk of suspected ADRs from the inflammation category (myocarditis, pericarditis) was the highest for Nuvaxovid (71.0 per million doses) and was followed by Spikevax, Comirnaty, Jcovden, and Vaxzevria (21.2, 17.4, 10.9, and 9.1 per million doses, respectively; Table 2). Additionally, potential ADRs from the thrombosis category occurred most frequently after vaccination with Vaxzevria (99.0 per million doses). They were followed by Nuvaxovid, Jcovden, Comirnaty, and Spikevax (97.6, 36.2, 21.8, and 20.0 per million doses, respectively). Suspected ADRs from both categories were observed very rarely for any vaccine.

In addition to that, in our study, the diversity of analysed suspected serious cardiac ADRs seems to be much broader and, except for inflammation and thrombosis, also includes arrhythmias, ischaemia, heart failure, life-threatening events, conduction disorders, cardiomyopathies, blood pressure abnormalities, and syncope. Moreover, the frequency of directly life-threatening events (e.g. ventricular fibrillation, cardiac arrest, or circulatory collapse) that were potentially linked to COVID-19 vaccinations was very rare for all studied vaccines (<1/10 000 i.e. <100/1 000 000 administered doses, Tables 1, 2). Of note, the estimated frequency of life-threatening events was lower for mRNA vaccines (Comirnaty and Spikevax–4.6 per million administered vaccine doses) than for adenovirus-based vector vaccines (8.9 for Jcovden and 10.1 for Vaxzevria) and protein-based vaccine (Nuvaxovid, 31.1 per million administered vaccine doses, Table 2).

Based on our analysis, the most common serious specific suspected ADRs associated with COVID-19 vaccines are very rare and include syncope, arrhythmia, tachycardia, palpitations, angina pectoris, hypertension, myocarditis, deep vein thrombosis, thrombosis, and pulmonary embolism. According to other studies, most cardiac ADRs are mild to moderate and serious events are rarely reported.^26,27 Thus, the results of our study are in line with the literature. In the general population, Kaur et al. most commonly observed tachycardia, hypertension, and hypotension, whereas acute myocardial infarction, cardiac arrest, and circulatory collapse were linked to the vaccines in the age group >75 years.²⁷ Other studies mention that the main cardiac complication was myocarditis.^28,29 Hana et al. report that according to the Vaccine Adverse Event System, a safety signal detection system that reports adverse events to the United States Food and Drug Administration during post-licensure vaccine safety monitoring, the frequency of suspected cardiovascular events is very rare.²⁶ After the administration of over 379 million doses, the most commonly reported potential cardiac post-vaccination events were palpitations (11 252 cases), hypertension (5272 cases), tachycardia (4908 cases), thrombosis (3066), pulmonary embolism (2666), deep vein thrombosis (1960 cases), atrial fibrillation (1830 cases), myocarditis (1670 cases), pericarditis (1115 cases), and myocardial infarction (1125 cases). Thus, the frequency of ADRs, as well as their relative proportions observed in our study, are in line with the already published literature.^26–32

Overall, suspected cardiac ADRs after COVID-19 vaccine administration are rarely observed, and the benefit of active immunization against COVID-19 with any of the investigated vaccines outweighs the potential risk of serious post-vaccination ADRs.

Machine learning for the prediction of the category of suspected ADRs and identification of risk factors

The prediction accuracy of a given category of suspected ADRs using machine learning algorithms was low to moderate, from 0.58 to 0.70 (c-index from 0.61 to 0.75; Supplementary material online, S5). Therefore, the models should be used with caution to predict the possible post-vaccination occurrence of a suspected ADR from a given category. Low prediction capabilities are likely caused by the low number of input data types (just four) and the low prevalence of categories of potential cardiac ADRs compared to all reported suspected ADRs. Moreover, except for identifying vaccination with Vaxzevria and female sex as the most predictive factor for suspected ADRs for the ‘Inflammation’ category, vaccination with any investigated vaccine was not predictive for any other categories of potential cardiac ADRs.

Limitations and future directions

Determining a causative link between the event (vaccination) and suspected ADR remains a challenge due to the ageing population and increasing frequency of comorbidities.³³ As a consequence of no randomization, these unmeasured variables may act as confounders, which is a methodological limitation of our retrospective cross-sectional study. The study is also sensitive to the recall bias, as some participants may have needed to recall the association between the vaccination and an ADR. Further on, the dataset has a limited granularity due to a low number of clinical characteristics available for each record in the EV database. Moreover, based on this cross-sectional study, we cannot establish a cause-and-effect relationship or analyse changes over time.³⁴ Thus, it is impossible to determine that inoculation with one vaccine is of higher risk than with the other, and future longitudinal studies are needed to validate these correlations. However, as this analysis encompasses a large sample and the results are consistent with the previously published studies, it seems reasonable to consider that the insights from this study are reliable and provide a solid foundation for future research.^26–32

Furthermore, reporting potential ADRs to COVID-19 vaccines is based on a spontaneous surveillance system.^1,3 Therefore, the suspected ADRs may be over- or under-reported.^1,35 Spontaneous reporting might also be influenced by the information shared by legislation bodies. For example, the frequency of reporting of actual ADRs might have been larger for the Vaxzevria vaccine because EMA ‘raised awareness of clinical care recommendations to manage suspected thrombosis with thrombocytopaenia syndrome’.²⁵ This was followed by an increase in the amount of news regarding this information on X (Twitter).³⁶ Thus, the frequency of suspected ADRs documented for Vaxzevria may be over-reported compared to other vaccines. Moreover, there are large discrepancies between the number of administered vaccines, from 225 thousand for Nuvaxovid to over 567 million doses for Comirnaty (about 2000 times more, Figure 1).²³ Thus, the documented frequency of the most common suspected ADRs might be lower for the most widely products (Comirnaty and Spikevax). To reduce this bias, we analysed serious ADRs only, which are usually more consistently reported in comparison to mild ADRs.^35,37

Additionally, in the EV database, it is possible to report suspected ADRs using general descriptions, for example, ‘arrhythmia’ instead of ‘ventricular fibrillation’.²² To overcome this bias, we classified specific ADRs into broader categories to find the most frequent groups of pathologies that might occur after inoculation against COVID-19.

Conclusions

Based on this study and available literature data, the benefit of active immunization against COVID-19 seems to outweigh the potential risk of serious post-vaccination cardiac ADRs. The frequency of the majority of the suspected serious cardiac ADRs appears to be very rare and shows the high safety of COVID-19 vaccines. Inoculation with Comirnaty and Spikevax (mRNA vaccines) seems to be associated with the lowest risk of potential serious cardiac ADRs, followed by Jcovden, Vaxzevria (adenovirus-based vector vaccines), and Nuvaxovid (protein-based vaccine).

Funding

This research received no external funding.

Conflict of interest: The authors declare no conflict of interest.

Data availability

The data underlying this article are available in the article, its online supplementary material, and the EudraVigilance and ECDC ‘Vaccine Tracker’ databases.

Author contributions

Conceptualization: W.N., L.D.S.; methodology: W.N., J.R., L.D.S.; investigation: W.N.; statistical analysis: W.N.; writing—original draft preparation: W.N.; writing—review and editing: J.R., M.N., L.D.S.; supervision: M.N., L.D.S. All authors have read and agreed to the published version of the manuscript.

References