Background: Using data from electronic health registries, this study intended to estimate the COVID-19 vaccine effectiveness (VE) in the population aged 65 years and more, against symptomatic infection, COVID-19-related hospitalizations, and deaths, overall and by time since complete vaccination for the period February to September 2021.
Methods: We established a cohort of individuals aged 65 and more years old, resident in Portugal mainland, using the National Health Service User number to link eight electronic health registries. Outcomes included were symptomatic SARS-CoV-2 infections, COVID-19-related hospitalizations or deaths. The exposures of interest were the mRNA vaccines (Comirnaty or Spikevax) and the viral vector (Vaxzevria) vaccine. Complete schedule VE was estimated as one minus the confounder adjusted hazard ratio, for each outcome, estimated by time-dependent Cox regression with time-dependent vaccine exposure.
Results: For the cohort of individuals aged 65-79 years, complete scheme VE against symptomatic infection varied 43 (95%CI: 37-49) (Vaxzevria) and 65 (95%CI: 62-68) (mRNA vaccines). This estimate was slightly lower in the ≥80 years cohort (53, 95%CI: 45-60) for mRNA vaccines). VE against COVID-19 hospitalization varied between 89% (95%CI: 52-94) for Vaxzevria and 95% (95%CI: 93-97) for mRNA vaccines for the cohort aged 65-79 years and was 76% (95%CI: 67-83) for mRNA vaccines in the ≥80 years cohort. High VE against COVID-19-related deaths was estimated, for both vaccine types, 95% and 81 (95%CI:76-86) for the 65-79 years and the ≥80 years cohort, respectively. We observed a significant waning of VE against symptomatic infection, with VE estimates reaching approximately 34% for both vaccine types and cohorts. Significant waning was observed for the COVID-19 hospitalizations in the ≥80 years cohort (decay from 83% (95%CI:68 to 91) 14-41 days to 63% (95%CI:37 to 78) 124 days after mRNA second dose). No significant waning effect was observed for COVID-19-related deaths in the period of follow-up of either cohort.
Conclusions: In a population with a high risk of SARS-CoV-2 complications, we observed higher overall VE estimates against more severe outcomes for both age cohorts when compared to symptomatic infections. Considering the analysis of VE according to time since complete vaccination, the results showed a waning effect for both age cohorts in symptomatic infection and COVID-19 hospitalization for the 80 and more years cohort.