Abstract
BACKGROUND: During the COVID-19 vaccination campaign in Sweden, pulmonary embolism (PE) was a frequently reported suspected serious adverse drug reaction. The aim was to estimate risk of PE following vaccination for COVID-19 in the Swedish population aged 18 to 84 years. METHODS: Population-based cohort study using the CoVacSafe-SE established platform including national registers. PE-case definition: Individuals discharged from inpatient-care or visiting specialized outpatient-care with a main diagnosis of PE occurring between 27-Dec-2020 and 31-Dec-2022 without simultaneous diagnosis of COVID-19 infection. Time-to-event analysis was performed using multi-variable Cox' proportional hazard's models. Hazard Ratios (HR) adjusted for age, sex and co-morbidities were modelled.The vaccines were BNT162b2/Comirnaty®, mRNA1273/Spikevax® and ChAdOx1 nCoV-19/Vaxzevria® without regard to variants. Doses number one to five were studied. RESULTS: Eighty percent of the study-population (≈6.1 million people) received at least two doses of COVID-19 vaccine. A total of 12,456 cases of PE were identified. Twenty-eight days after vaccinations we observed 99 cases after 701,455 1st doses of ChAdOx1 nCoV-19, HR(adj), 1.29 (95%-CI, 1.05-1.59). Corresponding for BNT162b2 was 361 cases after 4,708,284 1st doses of BNT162b2 HR(adj) of 1.19 (95%-CI, 1.06-1.34) driven by age group 65-84; HR (adj), 1.24 (95%-CI, 1.07-1.44). No increased risks were observed for mRNA1273. CONCLUSION: In this nation-wide study, no strong associations were found between COVID-19 vaccinations and pulmonary embolism. Small increases in relative risk for the earliest doses of vaccines may be associated with prioritizing the frailest groups of people in the vaccination campaign, thus selection bias or unmeasured residual confounding is possible.