AstraZeneca AZD1222 Post-Authorisation Safety Study

Background: AstraZeneca was asked to conduct a post-authorisation safety study (PASS) for the AZD1222 COVID-19 vaccine. One of the 40 PASS outcomes was thrombosis with thrombocytopenia syndrome (TTS) for which a signal was detected in March 2021 and additional risk minimisation measures (RMM) were implemented, including age restrictions.

Objectives: To estimate the risk of TTS in subjects who received ≥1 dose of AZD1222, compared to concurrent unvaccinated subjects (primary), pre-pandemic historical comparators, or mRNA-vaccinated subjects (exploratory).

Methods: A cohort study was conducted using secondary data—CPRD (UK), VID (Valencia, Spain), SIDIAP (Catalonia, Spain), and PHARMO (The Netherlands). AZD1222-vaccinated subjects were 1:1 matched on age, sex, region, prior COVID-19, and by special populations of interest. As no TTS code exists, venous TTS (42-day risk window) was defined as a thromboembolic event (TE) and thrombocytopenia (TCP) ±10 days, and no TTS in the last year, with prior TE/TCP permitted. Incidence rates (IR) per 10,000 py and IR ratios (IRR) were estimated using propensity score-weighted Poisson regression. Subgroup, sensitivity, and meta-analyses were performed. Instead of a dichotomous interpretation based on significance testing, our interpretation considers the magnitude, precision, and potential bias.

Results: 5,321,930 AZD1222-vaccinated subjects were matched to concurrent unvaccinated subjects, 4,831,010 to historical comparators (CPRD, SIDIAP, PHARMO), and 4,028,091 to mRNA-active comparators (CPRD). In CPRD, 83% of subjects were vaccinated in Q1 2021 (pre-RMM implementation), and 64% were < 60 yrs. In VID, SIDIAP, and PHARMO, >60% were vaccinated after Q1 2021; most subjects were aged ≥60 yrs.

In the primary analysis, IRR (95% CIs) for TTS were 1.14 (0.60-2.17) in CPRD, 0.34 (0.10-1.18) in VID, and 0.66 (0.33-1.34) in SIDIAP; PHARMO identified no TTS events. Estimated IR and IRR for TTS, where available, were higher in the AZD1222 cohort than in the concurrent unvaccinated cohort in people aged < 60 yrs or when using shorter risk windows. After case validation, PPV-adjusted IRRs were below 1 in all data sources.

In the historical comparators analysis, meta-analysis yielded an IRR of 1.78 (1.12-2.82; I2=0%). In the mRNA-active comparators analysis, the IRR was 1.12 (0.61-2.05).

Conclusions: Acknowledging limitations introducing uncertainty, such as potential outcome misclassification, the totality of evidence suggests a possible increased risk of TTS post-AZD1222 vaccination that may be higher among individuals aged < 60 yrs, which is in line with the literature. Differential age distributions, resulting from country-level differences in the RMMs, may explain IRR disparities across data sources.