Senior Director, Epidemiology Medical Evidence, Vaccines & Immune Therapies, AstraZeneca, Cambridge, UK, United Kingdom
Background: Many studies on COVID-19 vaccine effectiveness (VE) estimate overall VE, spanning weeks to months after vaccination. Due to waning protection and differences in VE against hospitalisation due to different SARS-CoV-2 variants, estimates may over-/underestimate protection during a specific period, potentially masking important information for decision-making at individual/population level.
Objectives: Using results from an interim analysis of AZD1222 (ChAdOx1 nCoV-19) VE in the European COVIDRIVE test-negative case-control study (EUPAS42328), we aimed to demonstrate the importance of contextualising VE, particularly regarding time since last dose.
Methods: Through COVIDRIVE, a public-private partnership evaluating brand-specific VE against COVID-19 hospitalisation, 7 hospitals in 4 EU countries recruited hospitalised severe acute respiratory infection (SARI) patients aged ≥18 years from 01 June 2021–05 Sept. 2022 (when Delta and Omicron variants were predominant). Cases had RT-PCR-confirmed SARS-CoV-2, controls did not. Fully vaccinated had received two AZD1222 doses (4–12 weeks apart, 2nd dose ≥14 days before symptom onset, no booster) at symptom onset; unvaccinated had not received any dose. VE was estimated overall (primary objective), and by time since 2nd dose (using discrete time intervals and a spline-based-curve approach, secondary objective), and adjusted for symptom onset date, age, sex, and number of chronic conditions.
Results: The analysis included 761 patients (561 test-positive cases, 200 test-negative controls). Median age was 60 years; 40% had ≥2 chronic conditions. Among fully vaccinated, 90% received their 2nd dose in June-July 2021, none received their 2nd dose after Sept. 2021. Overall VE across the 15 months was 72.8% (95% CI [53.4–84.1]); the median time since 2nd dose was 20 weeks. Analysis of VE stratified by discrete intervals showed high VE (93.8%, 95% CI [48.6–99.3]) in the interval 2 to ≤8 weeks since 2nd dose, with lower VE (67.2%, 95% CI [30.8–84.4]) in the interval >16 to ≤24 weeks. Estimates for other time intervals since 2nd dose had wide CIs due to a limited number of events. Spline-curve showed sustained protection with potential waning: VE 88% at 10 weeks (95% CI [69-95]), 67% at 20 weeks (95% CI [37-82]), and 60% at 30 weeks (95%CI [10-81]); 95% CIs crossed 0 after 35 weeks.
Conclusions: We show durable protection against COVID-19 hospitalisation of AZD1222 primary series with enduring VE through ≥6 months, but the overall VE estimate under- and overestimated VE at specific time points after vaccination. Long term-follow VE studies must be interpreted in light of context information, including incidence rates, circulating variants and time since last dose.
