Background: Previous research suggests that although (some) COVID-19 vaccines could increase rare thrombosis risk, the disease leads to a much higher risk. The effects of vaccination on post-COVID thrombosis and cardiovascular risks remain unknown.
Objectives: To study the effect of Pfizer/BioNTech (PF) and Oxford/AstraZeneca (AZ) COVID-19 vaccines on the risk of thromboembolic and cardiovascular complications in up to 30 days after SARS-CoV-2 infection.
Methods: We used UK primary care records from the Clinical Practice Research Datalink (CPRD) AURUM database. All individuals alive in January 2021 and aged >= 75 were eligible for inclusion. People who received a first vaccine dose between 4th and 27th January 2021 (enrolment period) were assigned to either PF or AZ cohort depending on the vaccine brand they received. Next, the comparator (UV) cohort was comprised by people who were unvaccinated during the enrolment period. Index date was defined as first dose date for vaccinated people, and for unvaccinated individuals it was sorted within enrolment period based on distribution of index in the AZ and PF cohorts respectively. Lastly, people with a history of COVID-19 or vaccination against it before index date were excluded. People were followed up until the first occurrence of the event of interest, death, end of data, or first vaccine dose (only UV). Outcomes of interest were venous thromboembolism (VTE), arterial thromboembolism (ATE), and heart failure (HF), in the 30 days after COVID-19 infection. Propensity Scores (PS) were used for overlap weighting to minimise confounding. Variables forced into PS were age, location, prior observation time, and numbers of previous visits and PCR tests. In addition, covariates were selected with LASSO regression where variables included had prevalence >0.5%. PS were estimated for all vaccinated individuals. Unmeasured confounding was assessed using 43 Negative Control Outcomes. Fine-Gray models were used accounting for death as competing risk. Sub-distribution Hazard Ratios (sHR) were calculated for each outcome, and subsequently calibrated.
Results: A total of 680,842 individuals were included for PF comparison (332,790 vaccinated + 348,052 unvaccinated); and 564,491 for AZ comparison (219,804 + 344,687), respectively. Results showed reduced risk for VTE and ATE in the 0 to 30 days after infection. sHR were of 0.36 [0.22 – 0.60], and 0.34 [0.19 – 0.60] against VTE, for PF and AZ respectively. sHR for ATE were 0.46 [0.23 - 0.91] and 0.47 [0.22 - 0.99] for PF and AZ. No significant effect was seen for post-COVID HF.
Conclusions: COVID-19 vaccines were associated with a reduced risk of VTE and potentially of ATE in the immediate time following SARS-CoV-2 infection.
