Background: Long-term effectiveness data on bivalent COVID-19 boosters are limited.

Aim: We evaluated the long-term protection of bivalent boosters against severe COVID-19 among >= 65-year-olds in Finland.

Methods: In this register-based cohort analysis, we compared the risk of three severe COVID-19 outcomes among >= 65-year-olds who received a bivalent booster (Original/Omicron BA.1 or Original/BA.4-5; exposed group) between 1/9/2022 and 31/8/2023 to those who did not (unexposed). We included individuals vaccinated with at least two monovalent COVID-19 vaccine doses before 1/9/2022 and >= 3months ago. The analysis was divided into two periods: 1/9/2022-28/2/2023 (BA.5 and BQ.1.X predominating) and 1/3/2023- 31/8/2023 (XBB predominating). The hazards for the outcomes between exposed and unexposed individuals were compared with Cox regression.

Results: We included 1,191,871 individuals. From 1/9/2022 to 28/2/2023, bivalent boosters were associated with a reduced risk of hospitalisation due to COVID19 (hazard ratio (HR):0.45; 95% confidence interval (CI):0.37-0.55), death due to COVID-19 (HR:0.49; 95%CI:0.38-0.62), and death in which COVID-19 was a contributing factor (HR:0.40; 95%CI:0.31-0.51) during 14-60days since vaccination. From 1/3/2023 to 31/8/2023, bivalent boosters were associated with lower risks of all three severe COVID-19 outcomes during 61-120days since a bivalent booster (e.g. HR:0.53; 95% CI: 0.39-0.71 for hospitalisation due to COVID-19); thereafter no notable risk reduction was observed. No difference was found between Original/Omicron BA.1 and Original/BA.4-5 boosters.

Conclusion: Bivalent boosters initially reduced the risk of severe COVID-19 outcomes by ca 50% among >= 65-year-olds, but protection waned over time. These findings help guide vaccine development and vaccination programmes.