ABSTRACT
BackgroundUS population-based data on COVID-19 vaccine effectiveness (VE) for the 3 currently FDA-authorized products is limited. Whether declines in VE are due to waning immunity, the Delta variant, or other causes, is debated. MethodsWe conducted a prospective study of 8,834,604 New York adults, comparing vaccine cohorts defined by product, age, and month of full-vaccination to age-specific unvaccinated cohorts, by linking statewide testing, hospital, and vaccine registry databases. VE was estimated from May 1, 2021 for incident laboratory-confirmed COVID-19 cases (weekly life-table hazard rates through September 3) and hospitalizations (monthly incidence rates through August 31). Results155,092 COVID-19 cases and 14,862 hospitalizations occurred. Estimated VE for cases declined contemporaneously across age, products, and time-cohorts, from high levels beginning May 1 (1.8% Delta variant prevalence), to a nadir around July 10 (85.3% Delta), with limited changes thereafter (>95% Delta). Decreases were greatest for Pfizer-BioNTech (-24.6%, -19.1%, -14.1% for 18-49, 50-64 years, and [≥]65 years, respectively), and similar for Moderna (-18.0%, -11.6%, -9.0%, respectively) and Janssen (-19.2%, -10.8, -10.9%, respectively). VE for hospitalization for adults 18-64 years was >86% across cohorts, without time trend. Among persons [≥]65 years, VE declined from May to August for Pfizer-BioNTech (95.0% to 89.2%) and Moderna (97.2% to 94.1%). VE was lower for Janssen, without trend, ranging 85.5%-82.8%. ConclusionsDeclines in VE for cases may have been primarily driven by factors other than waning. VE for hospitalizations remained high, with modest declines limited to Pfizer-BioNTech and Moderna recipients [≥]65 years, supporting targeted booster dosing recommendations.
