Predictors for reactogenicity and humoral immunity to SARS-CoV-2 following infection and mRNA vaccination: A regularized, mixed-effects modelling approach

Abstract

Introduction: The influence of pre-existing humoral immunity, inter-individual demographic factors, and vaccine-associated reactogenicity on immunogenicity following COVID vaccination remains poorly understood.

Methods: Ten-fold cross-validated least absolute shrinkage and selection operator (LASSO) and linear mixed effects models were used to evaluate symptoms experienced by COVID+ participants during natural infection and following SARS-CoV-2 mRNA vaccination along with demographics as predictors for antibody (AB) responses to recombinant spike protein in a longitudinal cohort study.

Results: In previously infected individuals (n=33), AB were more durable and robust following primary vaccination when compared to natural infection alone. Higher AB were associated with experiencing dyspnea during natural infection, as was the total number of symptoms reported during the COVID-19 disease course. Both local and systemic symptoms following 1^st and 2^nd dose (n=49 and 48, respectively) of SARS-CoV-2 mRNA vaccines were predictive of higher AB after vaccination. Lastly, there was a significant temporal relationship between AB and days since infection or vaccination, suggesting that vaccination in COVID+ individuals is associated with a more robust immune response.

Discussion: Experiencing systemic and local symptoms post-vaccine was suggestive of higher AB, which may confer greater protection.

Keywords: COVID-19; SARS-CoV-2; infection; protective antibodies; vaccine reactogenicity.

Figure 1

Antibody response following natural infection and vaccination Each black point represents a sample from a participant, grey lines connect points from the same participant, and the grey shaded area represents the maximum number of days between doses relative to date of full vaccination (14 days after second dose, regardless of vaccine manufacturer). (A) Days since full vaccination vs. log titers over time. t=0 on the x-axis represents the day when COVID+ participants became fully vaccinated (2 weeks after second vaccination). Bi-phasic, generalized additive model (GAM) is visualized by a blue line. (B) In unvaccinated COVID+ participants, log₂ antibody titers decay at a rate of -0.010 per day after last positive COVID-19 test result. Fitted linear model is visualized by a red line. Note that three points were excluded from the above figure due to the temporal scale used to graphically depict the data but are included in the analyses herein. (C) In vaccinated COVID+ participants, log2 antibody titers decay at a rate of -0.015 per day after full vaccination. Fitted linear model is visualized by a green line.