Long term bivalent mRNA vaccine effectiveness against COVID-19 hospitalisations and deaths in Portugal: a cohort study based on electronic health records.

BACKGROUND

In Autumn 2022, there were recommendations for a COVID-19 booster vaccination with adapted bivalent vaccines to eligible population. Evaluating vaccine effectiveness (VE), in a short period after the vaccination, is key to guide public health decisions on the vaccine performance, allowing implementation of mitigation strategies promptly. However, to assess long-term protection post-vaccination and evaluate the need for additional boosters, it is crucial to conduct studies that span the maximum duration of the vaccination program. This study aims to estimate the VE of bivalent mRNA vaccines against COVID-19-related hospitalisation and death in the Portuguese population aged 65 years or older, from September 2022 to May 2023.

METHODS

We used a cohort approach to analyse six electronic health registries using deterministic linkage, with a follow-up period of eight months. Severe outcomes included COVID-19-related hospitalisations and death, classified using discharge ICD-10 codes as proxies. The exposure of interest was the bivalent mRNA vaccine. VE was estimated for 14-97, 98-181 and 182-240 days after bivalent vaccination. Confounder-adjusted hazard ratio (aHR) was obtained by fitting a time-dependent Cox regression model with time-dependent vaccination status, adjusted for sociodemographic, history of influenza and pneumococcus vaccination, previous SARS-CoV-2 tests and infection, and comorbidities. VE was estimated by one minus the aHR between vaccinated with bivalent vaccine person-years versus those without bivalent vaccine person-years.

RESULTS

The cohort included 2,151,531 individuals aged 65 or older (27.8% with 80 or more years). In the ≥ 80 years old, VE was 41.3% (95%CI: 34.5-47.5%) and 50.3% (95%CI: 44.6-55.3%) against COVID-19-related hospitalisation and death, respectively. In the 65-79 years old, VE was 58.5% (95%CI: 51.9-64.2%) against COVID-19-related hospitalisation, and 65.1% (95%CI: 59.0-70.4%) against COVID-19-related death. VE waned for both age groups and outcomes. Among adults aged 65 years or older, we observed long-term moderate VE estimates against severe COVID-19-related outcomes.

CONCLUSIONS

These results support the need for yearly boosters of COVID-19 vaccination to maximise the protection of the senior population against COVID-19 severe disease. Additional (spring boosters) during a vaccination campaign should be evaluated considering the epidemiological context and results from long-term VE studies.