Maternal SARS-CoV-2 Vaccination and Infant Protection Against SARS-CoV-2 During the First 6 Months of Life

We examined the effectiveness of maternal vaccination against SARS-CoV-2 infection in 30,288 infants born at Kaiser Permanente Northern California from December 15, 2020, to May 31, 2022. Using Cox regression, the effectiveness of maternal vaccination was 85% (95% confidence interval [CI]: 67, 93), 64% (CI: 43, 78) and 57% (CI: 36,71) during the first 2, 4 and 6 months of life, respectively, in the Delta variant period. In the Omicron variant period, the effectiveness of maternal vaccination in these three age intervals was 22% (CI: −18,48), 14% (CI: −10,32) and 12% (CI: −4,26), respectively. Over the entire study period, the incidence of hospitalization for COVID-19 was lower during the first 6 months of life among infants of vaccinated mothers compared with infants of unvaccinated mothers (21/100,000 person-years vs. 100/100,000 person-years). Maternal vaccination was protective, but protection was lower during Omicron than during Delta. Protection during both periods decreased as infants aged.

Like in uenza and Tdap vaccines 15,16 , data suggest that vaccination during pregnancy may protect infants who are not old enough to be vaccinated against COVID-19. Two recent epidemiological studies found that vaccination during pregnancy was associated with a reduced risk of SARS-CoV-2 infection in infants during their rst 4 months of life and a reduced risk of hospitalization during the rst 5 months of life 17,18 .
The objective of this study was to further evaluate the effectiveness of at least 2 doses of mRNA COVID-19 vaccination during pregnancy for preventing SARS-CoV-2 infection in infants during the rst 2, 4 and 6 months of life during the Delta and Omicron variant periods. We used two different study designs: a primary design using a cohort analysis in which infants of vaccinated pregnant persons were compared with infants of unvaccinated pregnant persons. In this design, we used Cox proportional hazards models with calendar days as the underlying scale to estimate hazard ratios and calculated vaccine effectiveness as 1 minus the hazard ratio. Secondarily, we used a Test Negative Design (TND), which is a case control study, to compare the odds of vaccination among mothers of infants who tested positive vs. the odds of vaccination among mothers of infants who tested negative. In this analysis, vaccine effectiveness was evaluated as 1 minus the odds ratio. The aim of the secondary design was to compare the results from the cohort and TND design.

Results
Descriptive statistics and characteristics Between December 15, 2020, and May 31,2022, we identi ed 62117 infants born at Kaiser Permanente Northern California (KPNC), an integrated health care delivery organization. Among these infants, we excluded 21918 (35.3%) based on maternal exclusion criteria and 10408 (16.0%) after applying infant exclusion criteria ( Figure 1). The nal study population included 30288 (48.8%) infants who were KPNC members at least 2 months after birth. The mean age at pregnancy onset was 31 years (standard deviation 4.66 years). Most mothers (66.15%) were between ages 25 and <35 years, and more than a quarter (27.26%) were of Asian race, 5.16% were Black, 24.45% were of Hispanic ethnicity and 37.57% were White. Among the infants in the cohort, 19179 (63.32%) of the mothers were unvaccinated during pregnancy, 1035 (3.42%) of the mothers received 1 dose of a mRNA COVID-19 vaccine, 9456 (31.22%) received 2 doses, and 618 (2.04%) received 3 doses during pregnancy (Table 1).
Vaccine effectiveness: Primary design cohort analyses During the Delta dominant period, the crude incidences of testing positive for SARS-CoV-2 during the rst 2, 4 and 6 months of life were lower among infants whose mothers received at least 2 doses of mRNA COVID-19 vaccines during pregnancy (0.74, 1.38, and 1.55 infants per 100 person years [PY], respectively) than those whose mothers were not vaccinated during pregnancy ( (Table 3). We observed a similar pattern in vaccine effectiveness by trimester during the Omicron dominant period, however, estimates of vaccine effectiveness by trimester during the Omicron period were imprecise and much lower than during the Delta period (Table 3).
Over the entire study period, the crude rate of hospitalization with a SARS-CoV-2 positive test was lower during the rst 6 months of life among infants whose mothers received at least 2 doses of mRNA COVID-19 vaccines during pregnancy compared with infants whose mothers were unvaccinated during pregnancy (21/100000 PY vs. 100/100000 PY). VE against hospitalization was not estimated because of the very small number of hospitalized cases. There were only 1 hospitalized case among the children of vaccinated mothers and 9 hospitalized cases among the children of unvaccinated mothers (Table 1).

Discussion
In this large study which included >30,000 infants, we found that receipt of at least 2 doses of mRNA COVID-19 vaccine during pregnancy was associated with a decreased risk of infants testing SARS-CoV-2 positive during their rst 6 months of life. Maternal vaccination reduced the infant's risk of testing SARS-CoV-2 positive initially by 85% which decreased to 57% by 6 months of life in the Delta dominant period. However, vaccination during pregnancy was less effective at protecting infants against SARS-CoV-2 infection during the Omicron period. As infants aged, protection provided by maternal vaccination decreased during both periods.
Although the study was unable to directly estimate VE against hospitalization due to the small number of hospitalized cases, it found that over the entire study period, the incidence rate of hospitalization during the rst 6 months of life was much lower among the infants whose mothers were vaccinated during pregnancy compared with those whose mothers were not vaccinated. These results suggest that in addition to providing protection against testing positive, vaccination during pregnancy may also provide protection against hospitalization (severe disease) in the infants during their rst 6 months of life as previously reported recently 18 .
Our ndings that vaccination during pregnancy was effective at protecting infants during the Delta period are similar to those reported in a recent Norwegian study showing that mRNA COVID-19 vaccination during pregnancy was associated with a 71% decreased risk of testing positive for SARS-CoV-2 in infants during their rst 4 months of life during the Delta period 17 . During the Delta period, we found that protection extended through the infant's rst 6 months of life. However, in contrast with the Norwegian study which reported that infants of mothers who were vaccinated had a 33% decreased risk of testing positive during the rst 4 months of life during the Omicron period 17 , our study found a 14% reduced risk that was not statistically signi cant. Difference between the two studies might be due to population characteristics and to the timing of follow up as ours went through May 31, 2022, while the Norwegian study's ended in April 2022.
The nding that maternal vaccination was less effective at protecting infants during the Omicron dominant period is also consistent with previous studies which have reported decreased effectiveness of mRNA COVID-19 vaccines during Omicron among children and adults 14,19 . Recently another study reported that the effectiveness of mRNA COVID-19 vaccines against infections and hospitalizations among pregnant people was higher during the Delta period than during the Omicron period 20 .
We observed that infant's protection through vaccination during pregnancy decreased as they aged from 2 months to 6 months.
These ndings are consistent with diminishing of pregnancy derived antibodies in the infants over time 21 . A recent study found that the mean titer of maternally derived antibodies in infants of vaccinated mothers were higher at ages 2 months compared with antibody titers at ages 6 months 22 .
Despite several studies showing that vaccination during pregnancy is safe for pregnant people [11][12][13][14][15] , vaccine uptake has been suboptimal in this group 23 . In the present study, the mothers of only 33% of infants in the cohort received at least 2 doses. More efforts are needed to promote COVID-19 vaccines for pregnant persons because vaccination provides protection to mothers and their infants until they are old enough to receive their own COVID-19 vaccines.
Our study was strengthened both by its large sample size and our ability to follow infants through 6 months of age. In addition, our study period included two different SARV-CoV-2 variants, which allowed estimation of the effectiveness of vaccination during pregnancy in infants during both the Delta and Omicron variant periods. Our primary cohort analysis used calendar days as the underlying scale to ensure that we compared infants of vaccinated and unvaccinated mothers on the same calendar days because vaccination status during pregnancy and risk of SARS-CoV-2 infection varied over the study period. In this primary design, all eligible infants meeting inclusion criteria were included without sampling which improved power and minimized bias related to selection. Furthermore, it was reassuring that both the cohort and the secondary TND yielded vaccine effectiveness estimates in the same direction. Although both approaches adjusted for the same confounding factors, the effectiveness estimates from the TND were higher than those from the cohort design, which is consistent with our previous analyses of in uenza vaccine effectiveness in which we also observed that the TND tended to result in higher vaccine effectiveness estimates than did our cohort analyses 24 . The TND, a case control study, has been commonly used in studies of the effectiveness of in uenza vaccines and more recently COVID-19 vaccines. It is designed to better adjust for healthcare seeking behavior [25][26][27] , although it may also introduce other biases including selection bias 28 .The TND used analysis limited to the sample of infants who were tested for SARS-CoV-2.
The study had limitations worth noting. Vaccinations were limited only to those received during pregnancy. We did not assess whether vaccines received before pregnancy or immediately after pregnancy were associated with reduced risk of testing positive for SARS-CoV-2 in infants. The study did not adjust for maternal SARS-CoV-2 infections during pregnancy. During the study period, home testing became more prevalent. It is possible that this practice may have led to some misclassi cation of outcome, and we were unable to assess whether this misclassi cation was differential between vaccinated and unvaccinated mothers. We did not have genotyping data to con rm the variant that infected infants who tested positive and instead relied on state data regarding circulating strain predominance in the Northern California region. Like all observational studies, our study results are susceptible to residual confounding.
In conclusion, in this population-based cohort study, we found that infants born to mothers who received at least 2 doses of a mRNA COVID-19 vaccine during pregnancy were at lower risk of testing positive for SARS-CoV-2 and were at lower risk of hospitalization during the rst 6 months of life compared with infants whose mothers were unvaccinated during pregnancy. Maternal vaccination was protective, but protection was lower during the Omicron period than during Delta. Protection during both periods decreased as infants aged from 2 months to 6 months. Overall, the study results support recommendations for vaccination during pregnancy to protect both mothers and their infants.

Online Method
Setting and Study Population The study setting was Kaiser Permanente Northern California (KPNC), an integrated healthcare delivery organization that provides comprehensive health care to approximately 4.4 million members as of 2019. Members receive almost all their medical care at KPNC-owned facilities, including clinics, hospitals, pharmacies, and laboratories. KPNC has a comprehensive electronic health record system that captures detailed information on all medical services, including immunization, membership enrollment including place of residence, demographics, and pregnancy related care from pregnancy onset to delivery and beyond. KPNC members are similar to the broad catchment population in Northern California in terms of sociodemographic characteristics 29 . Annually, approximately 40,000 births occur at KPNC facilities.
The study was conducted among a cohort of infants born between December 15, 2020, and May 31, 2022. From this cohort, the study excluded the following infants born to: 1) mothers who were not between ages 16 and 50 years at pregnancy onset; 2) mothers who did not have a primary KPNC facility assignment; 3) mothers who were not continuous KPNC members from December 15, 2020 until delivery 4) mothers who had a positive nasal/throat swab for SARS-CoV-2 by polymerase chain reaction (PCR) prior to pregnancy onset; 5) mothers who had a positive SARS-CoV-2 antibody test documented by KPNC prior to onset of pregnancy; 6) mothers who received one or more doses of COVID-19 vaccine prior to pregnancy onset. We excluded these infants because we were primarily interested in estimating the effectiveness of at least 2 doses of mRNA vaccines received during pregnancy; 7) mothers who received other COVID-19 vaccine than mRNA vaccine during pregnancy; 8) mothers who did not receive their mRNA vaccinations in accordance with CDC recommendations -e.g., the timing between dose 1 and dose 2 was not within the recommended intervals; and 9) infants who did not become KPNC members within two calendar months of their birth.
No other exclusion criteria were applied.
The KPNC Institutional review board approved and waived consent for this study.

Outcomes
The outcomes were infant's rst positive nasal/throat swab for SARS-CoV-2 by PCR, and rst COVID-19 related hospitalization, occurring during the rst 6 months of life and recorded in the electronic health record.

Exposure
The exposure of interest was mRNA COVID-19 vaccination status during pregnancy in the electronic health record. Mothers were classi ed as either having had ≥ 2 doses of mRNA COVID-19 vaccines during pregnancy (and completed more than 7 days prior to delivery) or not having had any COVID-19 vaccines prior to delivery. We further classi ed vaccination status by the trimester within which the 2 nd dose was received.

Covariates
For mothers of infants in the cohort we extracted from the electronic health record data: age at pregnancy onset, race/ethnicity (Asian, Black, Hispanic, Paci c Islander, Multiracial, Native American, Other, White), the primary KPNC facility at which the woman received most of their health care, insurance payor (dichotomized as "Medicare/Medicaid/other subsidized insurance" and "Other"), neighborhood deprivation index [NDI] 30 categorized into quartiles with higher values representing greater deprivation), pre-pregnancy body mass index (BMI=kg/m2; underweight <18.5, normal 18.5-24.9, overweight 25.0-29.9, obese ≥30.0), pre-pregnancy diabetes status, pre-pregnancy hypertension and parity (0, 1, 2, 3, ≥4). For infants, we included age, as a categorical time-changing variable in 30-day increments and preterm status de ned as gestational age at birth less than 37 weeks.

Statistical analysis
We conducted a descriptive analysis of the study population and calculated crude rates of SARS-CoV-2 infection and hospitalization by maternal vaccination status. In our primary analysis, we implemented a cohort study design where we used Cox proportional hazards models that allow for time-varying covariates to estimate the SARS-CoV-2 infection hazard ratio (HR) in infants of mothers vaccinated with at least 2 doses of mRNA COVID-19 vaccines during pregnancy versus mothers who were unvaccinated during pregnancy. We calculated vaccine effectiveness (VE) as 100% multiplied by 1 -HR. In all models, we used calendar days as the time scale to account for changes over time in SARS-CoV-2 circulation and vaccine uptake. Infants were followed from birth until rst positive SARS-CoV-2 test by PCR at age 2, 4 or 6 months, with censoring due to death, health plan disenrollment, or end of follow-up (May 31, 2022). Models were adjusted for covariates listed above. To account for the correlation between infants with the same mother, we t marginal Cox proportional hazards models using robust sandwich variance estimates. We ran separate models on the time periods associated with the Delta (7/01/2021 to 12/20/2021) and Omicron variants (12/21/2021 to 5/31/2022). We also conducted analyses based on the trimester during which the vaccine was received during pregnancy ( rst, second or third trimester).
We conducted secondary sensitivity analyses restricting the population to infants who received at least one SARS-CoV-2 PCR test. In this analysis, we estimated the odds ratio (OR) of vaccination of mothers of infants who tested positive for SARS-CoV-2 versus infants who tested negative using logistic regression models conditioned (strati ed) on the calendar date of the test so that infants testing positive on a certain day were compared to infants testing negative on that same day. We calculated VE as 100% multiplied by 1-OR. This case-positive, control-test-negative design also referred to as the test negative design (TND) has often been used in studies of vaccine effectiveness. The TND is designed to better control for bias related to health care seeking behavior [25][26][27] . Models in this analysis were adjusted for the same covariates included in the primary analysis. All analyses were conducted using SAS software, v9.4. and statistical signi cance was assessed at two-sided p≤0.05.