Poster presented at the 41st Annual International Society for Pharmacoepidemiology Conference
Background: A US post-approval safety study [EUPAS43468] is ongoing to assess the safety of the original Pfizer-BioNTech monovalent COVID-19 vaccine (BNT162b2) using data from 5 health insurers participating in the FDA’s Sentinel System.
Objectives: To assess the distribution of baseline demographic, clinical, and healthcare utilization characteristics among individuals exposed to BNT162b2 and concurrent unexposed individuals in the overall study population, immunocompromised individuals (ICs), and individuals with a recorded history of COVID-19.
Methods: Individuals aged ≥6 months who received a 1st, 2nd, and/or 3rd primary series dose of BNT162b2 were identified from 11 December 2020 through 18 April 2023. Unvaccinated individuals were matched to BNT162b2 vaccinees (in a 2:1 variable ratio) on age, sex, US state, calendar time, and propensity score. The index date for the exposed was the date of receipt of BNT162b2 and, for the unexposed, was a randomly selected date within the same 30-day calendar-time interval. Covariate balance after matching was evaluated in weighted data using standardized mean differences (SMD).
Results: Before matching, there were ~25.5 million unexposed individuals and ~8 million, ~6.6 million, and ~32.9 thousand exposed individuals in the dose 1 (D1), dose 2 (D2), and dose 3 (D3) cohorts, respectively. After matching, there were 7,196,321 exposed and 13,721,492 unexposed individuals in the D1 cohort; 5,819,324 exposed and 11,021,596 million unexposed individuals in the D2 cohort; and 32,910 exposed and 65,806 unexposed individuals in the D3 cohort. In the D1 and D2 cohorts mean age was ~45 years (~1.2% aged <5 years) and ~53% were female. In the D3 cohort mean age was 22 years (53% aged <5 years) and 52% were female.
Before and after matching in the overall D1 and D2 cohorts, all baseline characteristics were well balanced (SMD<0.2), except influenza vaccination. Before matching in the D3 cohort, the exposed group had a lower prevalence of cardiovascular disease, hypertension, and obesity, but a higher prevalence of respiratory infections, receipt of other vaccines, any healthcare encounters, and COVID-19 tests than the unexposed group. After matching, all baseline characteristics were well balanced in the overall D3 cohort and in the IC and COVID-19 history subgroups for D1 and D2. However, in the D3 cohort, there were residual imbalances (SMD>0.2) for rheumatologic/inflammatory conditions with corticosteroid use and healthcare utilization in the IC subgroup, and for healthcare encounters in the COVID-19 history subgroup.
Conclusions: The balance in baseline characteristics between the exposed and unexposed groups after matching supports the use of an unexposed comparator group for the final comparative safety analyses.
