Study shows that antibody levels stabilize across populations, offering a new lens for pandemic strategy and vaccine deployment.

Study: Convergence of SARS-CoV-2 spike antibody levels to a population immune setpoint. Image Credit: Juan Gaertner / Shutterstock

In a recent study published in the journal eBioMedicine, researchers conducted a two-stage anti-SARS-CoV-2 spike protein antibody immunity study in 1,045 participants (of the original 6,683 enrolled) (age ≥ 5 yrs) from the Northwest and Southeast regions of the Dominican Republic. They aimed to elucidate the patterns and processes driving population-wide immunity against high transmission pathogens eliciting transient and partial immunity following infection or vaccination.

While individual responses to COVID-19 infections and vaccinations have been intensively characterized, trends in population-wide responses remain severely underreported.

Their findings highlight that hidden immuno-ecological pressures eventually drive antibody titers (immune markers) to a single discrete, population-wide peak irrespective of the number of interim infections or vaccinations. In brief, there appears to be an ‘optimal’ population-wide antibody titer threshold (against SARS-CoV-2 spike proteins).

In individuals below the threshold, antibody concentrations gradually increase till the threshold is attained. In those above the threshold (recently recovered from an infection or recently vaccinated), titers gradually decay till they converge upon the population-wide optima.

The study’s data was collected in two phases, mid-pandemic (August 2021) and late-pandemic (November 2022), providing a comprehensive temporal view of population immunity evolution.

Background

The ideal means of combatting an emergent infectious pathogen is by accurately predicting its transmission potential/routes and by estimating human responses (e.g., immunity) to the disease outbreak. Of these, population-level immunity (called ‘immunoepidemiology’) is widely considered the foremost predictor of disease spread (‘transmission dynamics’), emphasizing the utility of immunoepidemiology in forecasting and attenuating epi- and pandemic events.

Unfortunately, accurate immunoepidemiological investigations can only be carried out under rare and specific conditions, the two most important of which are a naïve population (without preexisting immunity/antibodies against the emergent pathogen) and a pathogen eliciting transient or partial immune protection following antigen exposure (e.g., severe acute respiratory syndrome-coronavirus-2 [SARS-CoV-2]). Consequently, only a handful of studies have attempted to evaluate population-wide immune patterns, with most focusing on individual-scale immune responses.

“…individual immune responses are highly variable, influenced by the number and timing of vaccine doses or infections, severity of infection, and individual differences in the host immune response. As such, and when considered in the context of declining vaccination uptake and widespread undetected and/or unreported infections, understanding population immunity and immune markers dynamics is a considerable challenge.”

Sufficiently elucidating population-wide immune patterns and the evolution of antibody dynamics would provide epidemiologists and policymakers the understanding required to identify potential pandemic-triggering pathogens at or shortly following their emergence and, more importantly, optimally allocate resources and vaccines to counter disease spread.

About the study

The present study utilizes serological data from the Dominican Republic acquired across two stages of the coronavirus disease of the 2019 (COVID-19) pandemic to model temporal trends in the population-level evolution of immune markers (antibody titers against SARS-CoV-2 spike proteins). Data for the study was acquired first between June and October 2021 and included 6,683 participants (3,832 households from 32 national provinces) sampled across three stages. A follow-up sample was collected between October and November 2022 from 1,045 participants in the Northwest and Southeast regions.

Each sampling stage included a home visit, which included venous blood collection and electronic questionnaire completion (KoBo Toolbox data collection platform). The questionnaire included demographics (sex, age, race/ethnicity), medical and COVID-19 vaccination history, anthropometrics (weight, height), occupation, and smoking status.

An electrochemiluminescence immunoassay (Roche Elecsys) was used to measure samples’ antibody titers against the S1 subunit of the SARS-CoV-2 spike protein. Generalized Additive Models (GAMs) were used to compute the change in antibody titers across sequential sampling stages.

The researchers also employed a random forest machine learning (ML) classifier and validated their model using three independent scenarios: ‘Regional split’ (geographical clustering), ‘Random split’ (random training and validation dataset reshuffling), and ’10-fold cross-validation.’  Model performance was evaluated using mean squared error (MSE), R² score, root mean squared error (RMSE), and Pearson correlation coefficient.

Study findings

Of the 6,683 participants initially enrolled, 1,045 provided serological samples across both follow-up periods and were included in statistical modeling. COVID-19 vaccination and S-antibody prevalence were found to be nearly complete (93.0% and 100%, respectively), confirming high baseline exposure to SARS-CoV-2 antigens. Notably, attempts to verify if vaccination or infection drove observed S-antibody prevalence revealed that both were cumulatively responsible for S-antibody outcomes.

Comparisons of participant-specific S-antibody titers across follow-up time points revealed that initially, random S-antibody titers converged to a single-peak distribution by the late pandemic period – participants with initially high antibody titers depicted declines in their antibody load, while those with initially low antibody titers depicted load increases. These findings were independent of the number of sex, age, geographical region, vaccination doses, and repeat infections, highlighting their generalizability.

Since increases in antibody titers in participants with low initial titers who had neither received additional vaccination doses nor contracted COVID-19 between follow-up periods were unexpected, sensitivity analyses were carried out.

These analyses confirmed the validity of the study findings and suggested that immune-ecological pressures drive antibody titers (and potentially other immune marker levels) to a distinct population-level “immune setpoint.” This setpoint forms the convergence midpoint of antibody titers independent of vaccination and baseline levels.

“We found that if the change in titer post-infection depends on pre-infection titer and wanes for higher titers, then boosting and waning in antibody kinetics over multiple infection waves will gradually converge to a narrow distribution of set point titers. This provides a theoretical explanation for the patterns observed…”

Conclusions

The present study comprises only the second investigation of population-level immune responses to COVID-19. It reveals that, unlike in the case of individual assessments (where antibody titers wax and wane in response to infections or vaccinations), population-level responses appear to converge on a single discrete Gaussian-like “immune setpoint,” which is independent of not only antigen exposure, but also age, sex, and geography. This knowledge is indispensable to epidemiologists and policymakers as it suggests that population-level immune responses are not random (as previously assumed) but follow predictable patterns.

While further research is required to refine and theoretically explain these findings, the present study provides the first step in designing policy action and vaccination plans in preparation for future COVID-19-like pandemic events.