Real-world evidence of cardiovascular complications following COVID-19 vaccines

Abstract

The vaccines were considered essential in lessening the severity of the disease and slowing the transmission of COVID-19 during the pandemic. Due to stringent enrolment requirements and a shorter follow-up, the safety evidence from randomised clinical trials of COVID-19 vaccines may not be sufficient, especially in the expedited programs. Therefore, it is important to conduct post-marketing safety surveillance with real-world data for a comprehensive understanding of vaccine safety profiles, especially for novel vaccines with mRNA platforms. Cardiovascular diseases (CVD) have been identified as intriguing adverse complications due to the higher incidence observed after COVID-19 infection. Yet, safety evaluation on CVD events following COVID-19 vaccination remains limited. This thesis aims to estimate the relative risks of three CVD major complications, including thromboembolic events, haemorrhagic stroke, and carditis associated with the administration of mRNA (BNT162b2) and inactivated (CoronaVac) vaccines. Additionally, it compares the absolute risk of these events between the COVID-19 vaccines and infection. This thesis utilised three real-world databases in Hong Kong to answer the above research questions: 1) electronic health records (EHR) for all government-subsidised public hospitals and clinics, 2) a database for the sole public mass COVID-19 vaccination program, and 3) a territory-wide COVID-19 infection recording system. The vaccine safety on CVDs was assessed using the Self-Controlled Case Series (SCCS) design. By facilitating within-individual comparison, this approach mitigates the influence of time-invariant confounding factors. However, the outcomes of interest in this thesis may violate the event-dependent exposure and event-dependent observation assumptions for a standard SCCS. To address this, a modified SCCS approach was adopted, and it identified a higher risk of haemorrhagic stroke and carditis following the BNT162b2 vaccine compared to referent windows. The CoronaVac vaccination does not carry a higher risk of these CVD complications. Importantly, the absolute risks following both vaccines were significantly lower than those observed after COVID-19 infection. A further focus of the thesis is the younger population, which was found to have a higher risk of carditis following their second dose of BNT162b2. To address the above issue, the thesis explores the impact of an interval extending between the first and second doses of BNT162b2 on carditis risk and its effectiveness. Two nested case-control study designs were utilised, and it discovered that prolonging the BNT162b2 vaccine's interdose interval can lower the incidence of carditis in the younger while preserving protection against COVID-19 infections. In summary, this thesis offers an in-depth analysis of the CVD safety profile of COVID-19 vaccinations in the Hong Kong population, including both the inactivated and mRNA platforms. Although a higher relative risk of CVD events was observed after mRNA vaccination, the absolute risk is lower after vaccination than infection. Adjusting the dosing schedule can help optimise the balance between safety and effectiveness among the younger. Besides, a modified SCCS is more appropriate for safety identification when assumption violations exist. This knowledge is vital for ensuring the safe and effective implementation of COVID-19 vaccination programs, ultimately contributing to the global efforts in combating the COVID-19 pandemic.