LAUSR

OBJECTIVES to evaluate immunogenicity and effectiveness of BNT162b2 COVID-19 vaccine in a cohort of healthcare workers (HCWs). DESIGN cohort study. SETTING AND PARTICIPANTS in a hospital in Milan (Lombardy Region, Northern Italy) HCWs without ("negative cohort") and with ("positive cohort") history of SARS-CoV-2 infection or elevated serum antibody before the vaccination campaign (27.12.2020) were included. Data collection and follow-up covered the period 27.12.2020-13.05.2022. MAIN OUTCOMES MEASURES 1. serum anti-spike-1 (anti-S1) antibody levels after vaccination; 2. vaccine effectiveness (VE) against SARS-CoV-2 infections (either symptomatic or not) in the negative cohort. Data on infections were extracted from multiple sources (laboratory, accident reports, questionnaires). Vaccination was treated as a time-dependent variable. Using unvaccinated person-time as reference, hazard ratios (HR) of infections and 95% confidence intervals (95%CI) were calculated with a Cox regression model adjusted for gender, age, and occupation. VE was calculated as (1 - HR)×100. RESULTS 5,596 HCWs were included, 4,771 in the negative and 825 in the positive cohort. In both cohorts, serum anti-S1 antibodies were high one months after the second dose, halved after six months, and returned to high levels after the third dose. In the negative cohort, 1,401 SARS-CoV-2 infections were identified. VE was 70% (95%CI 54-80; 46 infected) in the first four months after the second dose and later declined to 16% (95%CI 0-43; 97 infected). After the third dose, VE increased to 57% (95%CI 35-71; 61 infected) in the first month but rapidly declined over time, particularly after three months (24% in the fourth month and 1% afterwards). The number of infections avoided by vaccination was estimated to be 643 (95%CI 236-1,237). CONCLUSIONS in spite of rapidly declining effectiveness, vaccination helped to avoid several hundred infections in the considered hospital.