Seroprevalence, seroconversion, and seroreversion of infection-induced SARS-CoV-2 antibodies among a cohort of children and adolescents in Montreal, Canada

Abstract

Importance: Repeated serological testing for SARS-CoV-2 allows the monitoring of antibody dynamics in populations, including detecting infections that are missed by RT-PCR or antigen testing. Understanding the factors associated with seroconversion and seroreversion as well as the duration of infection-induced antibodies can also inform public health recommendations regarding disease prevention and mitigation efforts. Objective: To use serological testing to assess the prevalence, seroconversion, and seroreversion of infection-induced SARS-CoV-2 antibodies in children and adolescents in Montreal, Canada. Design: This analysis reports on three rounds of data collection from a prospective cohort study (Enfants et COVID-19: Étude de séroprévalence [EnCORE]). The study rounds occurred as follows: Round 1 October 2020-March 2021, Round 2 May to July 2021, and Round 3 November 2021 to January 2022. Most Round 3 samples were collected prior to the spread of the Omicron BA.1 variant in Quebec. Setting: Population-based sample. Participants: Children and adolescents aged 2 to 17 years in Montreal, Canada. Exposure: Potential exposure to SARS-CoV-2. Main Outcomes and Measures: Participants provided dried blood spots (DBS) for antibody detection and parents completed online questionnaires for sociodemographics and COVID-19 symptoms and testing history. The serostatus of participants was determined by enzyme-linked immunosorbent assays (ELISAs) using the receptor-binding domain (RBD) from the spike protein and the nucleocapsid protein (N) as antigens. We estimated seroprevalence for each round of data collection and by participant and household characteristics. Seroconversion rates were calculated as were the likelihoods of remaining seropositive at six months and one year. Results: The study included DBS samples from 1 632, 936, and 723 participants in the first, second, and third rounds of data collection, respectively. The baseline seroprevalence was 5.8% (95% CI 4.8-7.1), which increased to 10.5% and 10.9% for the respective follow-ups (95% CI 8.6-12.7; 95% CI 8.8-13.5). The overall average crude rate of seroconversion over the study period was 12.7 per 100 person-years (95% CI 10.9-14.5). Adjusted hazard rates of seroconversion by child and household characteristics showed higher rates in children who were female, whose parent identified as a racial or ethnic minority, and in households with incomes less than 100K. The likelihood of remaining seropositive at six months was 67% (95% CI 59-76) and dropped to 19% (95% CI 11%-33%) at one year. Conclusions and Relevance: The data reported here provide estimates of pre-Omicron seroprevalence, seroconversion rates and time to seroreversion in a population-based cohort of children and adolescents. Serological studies continue to provide valuable contributions for infection prevalence estimates and help us better understand the dynamics of antibody levels following infection. Continued study of seroconversion and seroreversion can inform public health recommendations such as COVID-19 vaccination and booster schedules.

Infection‐induced seroconversion and seroprevalence of SARS‐CoV‐2 among a cohort of children and youth in Montreal, Canada

“…Differences in seroconversion rates were observed by age group. Consistent with other studies, we found that young children were more affected by Omicron's surge with an estimated 36% higher rate of seroconverting, which was not seen in pre‐Omicron rounds 4–6 . This may be explained by the high vaccine uptake in the older age groups.…”

“…We found that having a parent self‐identify as a racial or ethnic minority was associated with a 44% higher rate of seroconverting during this period than children with white parents. This was also an important risk factor for seroconversion in EnCORE's earlier rounds and has been highlighted on multiple occasions since the beginning of the pandemic as a major risk factor of infection and infection severity 4,5,14 . Our results support that even in the context of Omicron‐enhanced contagiousness, health inequalities still play a key role in SARS‐CoV‐2 incidence.…”

“…A limitation of our study was the significant loss to follow‐up and based on anecdotal reports from parents, this was largely associated with the parent–child discomfort in providing the DBS. However, we corrected estimates using inverse probability of censoring weights and we found little evidence of differential patterns in loss to follow‐up 5 . Nevertheless, the analysis was not powered to detect meaningful differences in seroconversion rates.…”

“…This could indicate a reinfection or the persistence of a detectable antibody response from an earlier infection. We previously found that the likelihood of remaining seropositive was 68% at 6 months versus 41% at 1 year; estimates vary in other studies of children and adults, with as many as 96% of children and 83% of adults having detectable antibodies at 11–12 months after infection 5,15 …”

“…Unadjusted seroprevalence estimates for Round 4 by age, sex, vaccination status and prior seropositivity status were also estimated with correction by inverse probability of censoring weights. For further details on the study design and laboratory testing, please refer to Zinszer et al and Zinszer et al, 3–5 and for the statistical methods, see Appendix S1.…”