Immunity Acquired From the First Wave of COVID-19 Against Reinfections Up to Omicron Predominance.

The protective effects of prior infection with SARS-CoV-2 against reinfection are of great public health interest. Among the limited literature covering delta and omicron variant predominance, a nationwide study in Qatar reported that the protective effects derived from prior infection are 92% and 56% against the delta and omicron variants, respectively.1Altarawneh H.N. Chemaitelly H. Hasan M.R. et al.Protection against the omicron variant from previous SARS-CoV-2 infection.N Engl J Med. 2022; 386: 1288-1290Crossref PubMed Scopus (213) Google Scholar Malato et al2Malato J. Ribeiro R.M. Leite P.P. et al.Risk of BA.5 infection among persons exposed to previous SARS-CoV-2 variants.N Engl J Med. 2022; 387: 953-954Crossref PubMed Scopus (42) Google Scholar found that a previous infection could have 50% to 75% protective effects against the most recent BA.5 variant reinfection, with the greatest protection from previous BA.1/BA.2 omicron infection. Cerqueira-Silva et al3Cerqueira-Silva T. de Araujo Oliveira V. Paixão E.S. et al.Vaccination plus previous infection: protection during the omicron wave in Brazil.Lancet Infect Dis. 2022; 22: 945-946Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar found a relatively low (28.9%) protective effect of past infection against omicron infection, although better against severe outcomes (85.6%). However, these existing studies may have misclassified individuals having contracted SARS-CoV-2 but without symptoms and not having a diagnostic test as lacking prior infection. Therefore, we conducted this study to observe a cohort of health care workers (HCWs) in a Massachusetts-based health care system with baseline serology testing results available during the first wave of the pandemic,4Bruno-Murtha L.A. Osgood R. Lan F.Y. et al.SARS-CoV-2 antibody seroprevalence after the first wave among workers at a community healthcare system in the Greater Boston area.Pathog Glob Health. 2021; 115: 331-334Crossref PubMed Scopus (5) Google Scholar accounting for potential confounders (ie, age, sex, ethnicity,5Lan F.Y. Filler R. Mathew S. et al.Sociodemographic risk factors for coronavirus disease 2019 (COVID-19) infection among Massachusetts healthcare workers: a retrospective cohort study.Infect Control Hosp Epidemiol. 2021; 42: 1473-1478Crossref PubMed Scopus (16) Google Scholar and the time from taking serology test to receiving the first COVID-19 vaccine dose). Because the HCWs had received vaccines with various efficacies, we adjusted for their intention to be vaccinated instead of treating vaccination as a time-dependent variable. Each HCW was followed up from the date of serology testing to his or her termination date, the date of a subsequent positive result of polymerase chain reaction assay, or February 28, 2022 (about 1.5 years of follow-up). Those with a positive baseline IgG, IgM, or IgA result were considered seropositive and compared with seronegative staff. The Kaplan-Meier survival curve was used to delineate the infection-free trends across the groups (Figure) . After visually inspecting the log(–log) plot, we built the Schemper weighted Cox regression models for survival analyses to account for the nonproportionality and further adjusted for potential confounders. The study was exempted for human subjects research by the Cambridge Health Alliance Institutional Review Board (4/29/202-003). Among 176 eligible HCWs, the 86 (48.9%) seropositives were younger (44.7±11.7 years vs 48.3±11.9 years; P=.04), more likely to be men (27.9% vs 12.2%; P=.02), and ethnic minority (White: 30.2% vs 53.3%; P<.001), and they tended to receive their first COVID-19 vaccine dose later (time to first injection: 282.6±99.9 days vs 203.0±80.2 days; P<.001). After adjustment for potential confounders, HCWs with baseline positive serology results had a reduced hazard ratio of 0.13 (95% CI, 0.05 to 0.39) for contracting SARS-CoV-2 throughout the follow-up period. The findings remained robust after exclusion of those with equivocal serology results at baseline. Although limited by a lack of statistical power that prevents us from performing stratified analyses for different variant predominance, unmeasured confounding such as household situation and testing behaviors, and different serology tests to determine the baseline serology results (there were two laboratories using different assay techniques), our study highlights the protective effects of immunity conferred by infection throughout periods covering the variants of interest, in agreement with existing literature, while using serology results in the first wave as a more rigorous measure to determine prior infection status. S.N.K. has received COVID-19–related consulting fees from Open Health and has owned shares of Regeneron, Moderna, and AstraZeneca. All other authors declare no competing interests.