Maternal COVID-19 Vaccination and Prevention of Symptomatic Infection in Infants

Launched in 2021, the MOMI-Vax longitudinal cohort includes mother and infant pairs followed actively from pregnancy until the infant’s first birthday.¹⁶ Participants were recruited from 9 US academic sites, before or after maternal COVID-19 primary series or booster dose vaccination. Infants were eligible for inclusion in this analysis if they were born to individuals vaccinated during pregnancy with 2 or 3 doses of a monovalent messenger RNA COVID-19 vaccine (no boost or boosted group, respectively), and were followed prospectively from birth up to 6 months of age. Infant blood for sera was collected at delivery, and anti-Spike IgG, pseudovirus nAb against the ancestral Wuhan strain (614D), and live virus nAb levels against the D614G and omicron BA.1 and BA.5 variants were measured at this time point. Infants were tested for nucleocapsid (N) protein IgG antibodies at birth, 2 months, and 6 months of life.

The primary case definition for infant SARS-CoV-2 infection included SARS-CoV-2 infections by verified maternal-report. Infant cases were first identified by maternal interview at in-person visits at prespecified time points at birth, 2, and 6 months of age to determine presence of signs and symptoms and date of infection, and verified through review of documentation of laboratory testing (antigen testing or polymerase chain reaction) and health care utilization. The secondary case definition included both infant cases that were determined by verified maternal-report as in the primary case definition, as well as the addition of infant cases that were identified by seroconversion of N protein between time points.

Serum-binding IgG against SARS-CoV-2 Spike protein and N protein was measured using the Meso Scale Discovery V-PLEX SARS-CoV-2 Panel 2 assay (Meso Scale Discovery #K15383U), bridged to international standards and reported as binding antibody units.¹⁷ SARS-CoV-2 nAb titers were evaluated by a pseudovirus neutralizing assay using a replication-incompetent lentivirus coding for luciferase and containing the SARS-CoV-2 Spike protein (Wuhan-Hu-1) in the viral envelope (expressed as an IC50 value indicating the sample antibody titer capable of inhibiting viral entry and replication by 50%),¹⁸ and a live virus focus reduction neutralization titer assay with viruses representing SARS-CoV-2 Spike mutation D614G, and omicron BA.1 and BA.5 variants (expressed as the serum inhibitory dilution required to achieve 50% neutralization).¹⁹ Detailed assay methods are described elsewhere.¹³ 

To ensure valid comparisons between exposures of interest and potential confounders including calendar-time and infant age, we calculated the amount of person-time of follow-up per calendar month in the study and defined the analysis period to include calendar months with >0.8 person-years of follow-up (∼10 infants in follow-up for the entire month) in both the 2- and 3-dose exposure groups and by age groups 0 to <2 and 2 to 6 months. This resulted in an analysis period from December 2021 to July 2022.

Incidence rates per 100 person-years with 95% confidence intervals (CIs) were calculated by exposure group using the normal approximation. The risk reduction for COVID-19 infection in infants by booster status and antibody titers at delivery was estimated using separate calendar-time Cox regression models stratified by site. Covariates (including vaccine type, maternal age, race, ethnicity, number of maternal comorbidities, and trimester of infection) were assessed for inclusion in the models on the basis of their association with:

1. risk of SARS-CoV-2 acquisition in the infant as assessed univariably by Cox regression, adjusting for each of the immune markers and boosting status, respectively; and

2. antibody level in the risk model using linear regression models.

Variables significant at P < .10 in either the outcome or exposure assessment were included in the final regression model for that exposure. The primary end point was defined as time to the primary case definition (verified maternal-report of infant infection), and the secondary end point defined as time to the secondary case definition (verified maternal-report or seroconversion), with censoring at infant’s first infection date up to 194 days. Infants whose parents did not report infection but transitioned from negative to positive N protein between time points did not have a date of infection to include as an end point in the model. The date of infection between 0 to <2 months and 2 to 6 months old was imputed for these infants on the basis of the observed data from infants who had an infection reported by parents and also transitioned from negative to positive. The association between immune markers at delivery and time from last maternal dose to delivery was examined visually using locally estimated scatterplot smoothing splines. All analyses were conducted in R Version 4.2.2.²⁰ 

In total, 475 infants completed at least 1 follow-up visit after delivery and were eligible for analysis; this included 204 infants of boosted mothers and 271 infants of nonboosted mothers (Table 1; Supplemental Fig 4). By parental report, 76% of infants were white, 10.2% Black or African American, 8.1% Asian American, and 11.9% Hispanic or Latino. Most infants were born term (>89%) and had a normal birth weight (>82%). Retention by 6 months was >90%. Although all follow-up was during the omicron variant period, more infants were born before the first omicron wave in the nonboosted group (83.4%), reflecting the timing of the rollout of primary series vaccine implementation and booster dose recommendation for adults in the United States during 2021.

For the primary end point, 71 cases were reported in infants between birth and 6 months of life (41% in infants of boosted mothers, and 59% in infants of nonboosted mothers). For the secondary end point, 19 additional cases were identified for a total of 90 cases reported by either verified maternal report or seroconversion. The majority of infant cases in the primary end point group were laboratory confirmed by polymerase chain reaction or antigen testing (61 of 71; 85.9%) and were symptomatic (100% in infants of boosted, and 93.5% in infants of nonboosted mothers), with 70.0% reporting rhinorrhea, 55.2% reporting fever, and 55.2% cough, and the median duration of symptoms was 5 days. No infants in this analysis cohort were hospitalized.

The majority of COVID-19 infections in infants occurred between 2 and 6 months of age (n = 61 of 71; 85.9%) (Fig 1A). Infants 2 to 6 months old had higher COVID-19 incidence rates by calendar month and overall (56.8 per 100 person-years [95% CI 44.2–73.0]) as compared with infants 0 to <2 months of age (21.6 per 100 person-years [95% CI 11.6–40.1]).

Infants of nonboosted mothers had a higher overall COVID-19 incidence rate (52.1 per 100 person-years [95% CI 38.5–70.6]) compared with infants of boosted mothers (39.6 per 100 person-years [95% CI 27.5–57.0]) (Fig 1B). In the multivariable regression model examining the effect of 3 vs 2 maternal COVID-19 vaccine doses for infant protection, boosting during pregnancy reduced the infant’s risk of acquiring COVID-19 in the first 6 months by 56% (95% CI 8%–79%, P = .03) relative to no boosting (Table 2).

Infants whose mothers were boosted in pregnancy had higher mean Spike IgG, pseudovirus nAb, and live nAb titers at delivery compared with the nonboosted group (Spike: 4433 vs 1093 binding antibody units per mL, P < .001; pseudovirus nAb: 1282 vs 410 IC50, P < .001; live virus D614G: 1194 vs 254, P < .001; live virus omicron BA.1: 226 vs 26, P < .001; live virus omicron BA.5: 122 vs 24, P < .001) (Table 1, Fig 2A–D boxplots).

In the multivariable regression model for the primary end point, a 10-fold increase in Spike IgG titer (unit change in the log10 scale) measured at delivery was associated with a 47% (95% CI 8%–70%, P = .02) reduction in the infant’s risk of acquiring COVID-19 in the first 6 months of life (Table 2). A similar reduction in risk with higher delivery titers was observed for each of the pseudovirus and live virus nAb models. A 10-fold increase in pseudovirus nAb against Wuhan Spike, and live virus nAb against D614G, and omicron BA.1 and BA.5 variants at delivery was associated with a 30% (95% CI −15% to 57%; P = .16), 46% (95% CI 6%–69%, P = .03), 56% (95% CI 1%–80%, P = .05), and 60% (95% CI −15% to 86%; P = .09) reduction in risk for the infant in the first 6 months of life, respectively. Regression models for the secondary end point showed a similar or higher reduction in risk with higher delivery titers (all P values <.05) for all analysis sets (Table 2).

In sensitivity analyses that subset the Spike IgG, pseudovirus, and live virus nAb models to either boosted or nonboosted cohorts, a trend toward risk reduction was identified with increase in delivery titer only in the boosted cohorts (Supplemental Tables 3–8). Although not statistically significant, this trend suggests that the maternal boost during pregnancy and resulting higher titers was likely the main contributor to the infant’s reduction in risk.

Regardless of the number of weeks between the last COVID-19 vaccine dose during pregnancy and delivery, infants of boosted mothers had higher Spike IgG, pseudovirus, and live virus nAb titers than infants of nonboosted mothers at delivery (Figs 2A–E). Delivery titers increased for the prenatally boosted group as the number of weeks to delivery decreased, until ∼4 to 5 weeks before delivery, when they began to decline precipitously to match titers from the nonboosted group. For the boosted group, peak Spike IgG, pseudovirus, and live virus nAb titers were observed ∼6 to 12 weeks before delivery.

By gestational age, infant Spike IgG levels at delivery were lowest if mothers were boosted in the first trimester of pregnancy (≤13 weeks), increased by week gestation when mothers were boosted in the second trimester (14–26 weeks), reached peak titers at delivery when mothers were boosted between 28 and 34 weeks, and then declined slightly if mothers were boosted after ∼34 weeks (Fig 3A). Similarly, infant pseudovirus nAb titers at delivery were highest if mothers were vaccinated in the early third trimester (Fig 3B). Despite declining levels in the late third trimester, titers among those boosted in the third trimester remained at or above those of participants boosted in the first and early second trimesters.

In this multisite, prospective, longitudinal cohort study of mother–infant dyads enrolled after maternal COVID-19 vaccination, higher antibody titers at birth were associated with substantially increased infant protection from infection in the first 6 months of life. Maternal boost during pregnancy resulted in notably higher infant binding and nAbs versus primary series alone, with peak delivery titers after third trimester maternal boost, and booster receipt was associated with a strong reduction in risk of infant infection. Additionally, after combining boosted and nonboosted groups, increased protection was also demonstrated with higher delivery titers for Spike IgG, pseudovirus nAb against Wuhan, and live virus nAb to D614G, and omicron BA.1 and BA.5 variants, indicating consistency across the breadth of the humoral immune response and highlighting the predictive value of neutralization antibody as it relates to disease outcomes.

Achieving optimal protection from SARS-CoV-2 infection during early infancy has been elusive, in part because of the unavailability of COVID-19 vaccines for infants <6 months old. After maternal 2-dose primary-series vaccination, 1 previous study showed that higher infant anti-Spike cord blood levels were associated with longer disease-free intervals after birth, but lacked the sample size or booster dose data to measure the degree of protection and compare titers by number of doses.¹⁵ Our study is unique in estimating the substantial reductions for infants because of higher antibody titers at delivery, as well as quantifying the additional protection that a maternal COVID-19 booster dose affords in early infancy. This is consistent with findings from a recent study of adults in Japan²¹ that found that higher anti-Spike antibody titers after 3-dose vaccine series were associated with higher protection against omicron BA.5 infection. Our results are also complementary to VE studies of maternal vaccination for pregnant persons with a booster dose,²² as well as studies that have examined VE in infants after maternal vaccination.^6,–9,23,24 

Higher delivery titers for Spike IgG, pseudovirus, and live virus nAb were associated with substantial reductions in risk in our infant cohort, indicating that both binding and functional antibodies of the humoral immune response correlate with protection from infection. As compared with antibody levels against D614G, lower omicron BA.1 and BA.5 titers were noted at delivery in both boosted and nonboosted groups, and BA.1 and BA.5 titers at delivery were much lower in the nonboosted cohort, in line with previous studies.¹³ Yet, importantly, all live virus models were strongly supportive of infant protection. Therefore, these data indicate that maternal booster vaccination with ancestral strains provided functional antibodies that protected against infection in infants during early infancy, at a time when omicron variants were circulating widely. These results are critically important, because some VE studies in children and after maternal vaccination for infants have shown lower effectiveness during periods of omicron predominance, with some estimates crossing 0.²³ Because infection-induced immunity has increased since the start of the pandemic, with seroprevalence ranging widely by age and geographic location from 11% to 92% in children <5 years old,^25,–27 the ability to measure VE has become more challenging, making direct measurement of antibody titers a crucial piece of the puzzle to understand the beneficial impact of maternal immunization.

Nationally, infants <6 months old were hospitalized at increasing rates during and after the widespread takeover of omicron BA.1 and BA.5 throughout the United States, coinciding with the COVID-19 cases in our analysis cohort. During this time of omicron predominance, the hospitalization rate in infants <6 months old surged above 70 per 100 000, reaching similar rates to those in older adults aged 65 to 75 years.² As a result, maternal booster vaccination during pregnancy is an increasingly important strategy to decrease the burden of disease in this young, vulnerable population. A number of studies have sought to understand the optimal timing of maternal immunization during pregnancy for infant protection, with some indicating that vaccination in the second or early third trimesters provides better protection for infants,^8,9,28,29 whereas others have signaled that vaccination at any point during pregnancy is likely to be beneficial.³⁰ Although our study was not specifically designed to assess infection risk by timing of maternal vaccination, evaluation of both binding IgG and nAb delivery titers by week gestation of vaccination during pregnancy indicates that higher antibodies at delivery were achieved if maternal vaccination, and particularly a booster dose, was received during the early third trimester of pregnancy. Peak titers were observed after maternal booster dose if given ∼6 to 12 weeks before delivery, corresponding with 28 to 34 weeks’ gestation during pregnancy. However, despite this variation in titer by timing of vaccination, the most striking difference in infant delivery titer for all binding and functional antibody analysis data sets was between the boosted and nonboosted groups, reinforcing the importance of booster dose receipt at any time during pregnancy for infant protection. Ultimately, recommendations on timing of vaccination will need to balance the benefits in providing protection to the pregnant person and the infant.

Our study has limitations. First, the differential timing of enrollment of boosted and nonboosted mothers of the infants in our analysis made comparisons between groups challenging, because national immunization recommendations evolved to add booster doses concurrent with the emergence of new variants throughout the United States. To account for this, we purposefully restricted the analysis to include sufficient follow-up time by dose exposure group and infant age, and our regression analyses controlled for calendar-time. Nonetheless, unmeasured confounders could remain. Second, because of no hospitalized cases, we were unable to distinguish if higher antibodies or maternal boost afforded more protection against severe disease; at the same time, few moderate to severe cases in an entirely vaccinated cohort are consistent with the known effectiveness of messenger RNA COVID-19 vaccines. Third, our study did not have enough power to examine relative risk by both infant age and booster status simultaneously. Yet, the higher COVID-19 incidence in older infants (2–6 months versus 0–<2 months old) is consistent with other studies that have demonstrated higher VE for infants within a shorter time frame after maternal vaccination.

Our study also had notable strengths. First, we designed a large prospective cohort study early in the pandemic that met participant targets and resulted in robust comparisons by vaccine group. Second, a multicenter study with racial and ethnic diversity reflective of recruitment at participating sites increases the generalizability of our findings. Third, our high retention rates at follow-up time points minimize bias. Fourth, we were able to conduct extensive, centralized laboratory testing for each of the 3 antibody assays including neutralization and paired these results with case data in young infants, allowing for increased understanding of the breadth and function of the immune response in an understudied population after maternal COVID-19 immunization.

This study provides strong evidence of the robust immune response after maternal booster COVID-19 vaccination that protects infants against SARS-CoV-2 infection, and adds to the growing evidence base of effectiveness of maternal COVID-19 vaccination for both mothers and their infants.³¹ Specifically, we show that a monovalent booster dose during pregnancy leads to higher binding and nAb titers at delivery that are effective against omicron, for an age group that has the highest COVID-19–associated hospitalization rate in pediatrics since the emergence and ubiquitous spread of omicron variants. Currently, infants are not eligible for COVID-19 vaccination until they reach 6 months of age. Until they are, maternal booster vaccination during pregnancy is an effective strategy that provides transplacentally transferred passive binding and neutralizing SARS-CoV-2 antibodies that protect effectively against infection during early infancy.

We thank Jeannie Murray (University of Maryland), Joy Miedema (FHI 360), Jeanna Piper (National Institutes of Health), and the staff at the study sites for their exceptional support without which this work would not have been possible. We also thank the participants at each site for their numerous contributions to the study.

CI

confidence interval

COVID-19

coronavirus disease 2019

IgG

immunoglobulin G

LLOQ

lower limit of quantification

N

nucleocapsid

nAb

neutralizing antibody

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

Spike

full-length spike

VE

vaccine effectiveness