Vaccine- or natural infection-acquired immunity to respiratory viruses declines over time (9). Therefore, it is increasingly important to evaluate the durability of immune protection against circulating and emerging strains of SARS-CoV-2 (4–8, 10–13). Cellular and humoral immunity following natural SARS-CoV-2 infection may persist for more than 1 year (10) and immune imprinting by immunization or natural infection may affect subsequent responses to emerging strains and variants of concern (VOCs) (10, 14) (Figure 1A). Thus, preexisting immunity to early SARS-CoV-2 strains may interfere with immune priming against new SARS-CoV-2 strains, perhaps because of biased recruitment of immune memory cells specific for the previous strain. This phenomenon is also known as “original antigenic sin.” Furthermore, antibodies specific for earlier strains may have low affinity for emerging variants. Natural infection–acquired immunity may not protect against emerging virus strains but can alter the durability of immune protection following vaccination (10). The combined effects of immune imprinting, natural waning of antibody titers, and rapid viral evolution may all contribute to substantially reduced vaccine efficacy against divergent viral strains such as Omicron (B.1.1.529) and its descendent lineages. Immune imprinting following SARS-CoV-2 B.1.1.7 (alpha) infection can reduce the durability of protection against Omicron following BNT162b2 vaccination (14). In contrast, Omicron infection following 3 doses of BNT162b2 vaccine can confer durable cross-protection against earlier VOCs (14). AZD12222, BNT162b2, and mRNA-1273 are among the most widely distributed vaccines. In a small-scale trial of the mRNA-1273 vaccine, a high neutralizing antibody titer persisted through 6 months after administration of the second dose (4). The durability of vaccine efficacy can also be estimated by regression analysis of public health data sets (15); for example, in North Carolina, immune protection from mRNA-1273 and BNT162b2 vaccines lasted 7 months (11). As reported by Sobieszczyk and colleagues (1), a small number of infections with contemporaneous VOCs were observed in the vaccinated and placebo groups within the first 6 months after immunization; data collection was completed before the emergence of the Omicron strain.

Lessons and questions from a large-scale COVID-19 vaccine trial. (A) Immune imprinting and durability of protection. The immunogenic components in SARS-CoV-2 vaccines are derived from early (2019 to 2020) WT or Wuhan-like (Wu-like) viral isolates. Vaccine-stimulated immunity results in antibodies that neutralize infection by homologous or closely related viral strains (solid lines), but does not confer long-lasting protection. Immune imprinting from either vaccination or infection with early SARS-CoV-2 strains means that subsequent infections with newer SARS-CoV-2 variants stimulate existing memory cells specific for previously encountered SARS-CoV-2 variants, rather than recruiting additional immune cells better suited to the new variants. Therefore, the antibodies produced may not efficiently neutralize subsequent variants (dashed lines). Over time, protection and cross-protection are further diminished due to declining antibody levels. The Omicron variant and emerging lineages bear numerous immune escape mutations and are poorly neutralized by antisera specific for older SARS-CoV-2 strains. The durability and breadth of immune protection may be enhanced with booster dose vaccination and updated vaccines. (B) Trade-offs of placebo controls. Placebo control blinding and unblinding are a trade-off between the accuracy of RCTs’ primary and secondary outcome measures and ethical/social justice concerns.

In the current study, Sobieszczyk and colleagues update the results of an ongoing large-scale, multi-national, placebo-controlled, Phase III RCT of AZD12222 in the United States, Peru, and Chile (1). The trial, which included 32,450 participants, had a primary efficacy endpoint of symptomatic SARS-CoV-2 infection, while secondary endpoints included disease severity, infections per number of subjects who seroconverted post vaccination, and emergency visits. Here, participants were followed for 6 months after their first vaccine dose. This allowed assessment of the durability of immune protection and of changes in neutralizing antibody levels over time. Many participants were considered eligible to be unblinded and received a nonstudy COVID-19 vaccine before the end of the study (Figure 1B). Vaccine efficacy measures were stratified by age, sex at birth, race, ethnicity, BMI, comorbidities, OSHA risk categories, and participants’ global regions of origin for the period of double-blinding (days after second vaccine dose) until receipt of nonstudy COVID-19 vaccine. The AZD12222 vaccine was 67% and 95.7% effective in preventing symptomatic SARS-CoV-2 infection and severe illness during the double-blinding period, respectively. Overall, the AZD12222 vaccine durability of immune protection through 6 months is found to be 70.2% effective. There were no concerning safety issues, and most recorded adverse events (AEs) were mild to moderate. Irrespective of age, sex, comorbidities, ethnicity, or other risk covariates, AZD12222 elicited a robust humoral response through 6 months, with antibody titers declining after 180 days. Waning of immunity was relatively low among individuals seropositive for SARS-CoV-2 at baseline compared with those who were SARS-CoV-2-seronegative.