We found a significantly lower prevalence of initiation of breastmilk feeding among people with confirmed SARS-CoV-2 infection within two weeks before delivery. After controlling for known factors associated with breastmilk feeding, including maternal age, race/ethnicity, gestational age, and education, the association between timing of infection before delivery and breastmilk feeding was attenuated towards the null among mothers who roomed-in with their infants. Among mothers who did not room-in, however, we continued to see a significant reduction in prevalence of breastmilk feeding for those with infection closer to delivery. Given how little was known about the impact of SARS-CoV-2 infection and potential concern of adverse impacts on the infant, separation of mothers from infants for IPC precautions could have led to less breastmilk feeding initiation or less support to mothers with SARS-CoV-2 infection earlier in the pandemic and among mothers with infection closer to delivery. Breastmilk feeding and rooming-in appeared to improve later in the epidemic when the recommendations were updated to emphasize the importance of breastmilk feeding while following safety precautions.
Statewide breastmilk feeding prevalence reported during the majority of 2020 were similar to baseline estimates of the five included states from SET-NET. 83% of infants were ever fed any breastmilk [13]. People with lower education status, younger age, and of Black race are less likely to initiate feeding breastmilk [1, 13, 14]. These characteristics are also associated with an increased risk of SARS-CoV-2 infection [15]. We found that breastmilk feeding was also lowest among people with fewer years of education and of Black race which emphasizes the continued need for focused and culturally relevant breastfeeding promotion efforts in tandem with education regarding the breastmilk feeding practices in the setting of COVID-19. Variations in frequency of breastmilk feeding followed known demographic and birth characteristics, including a lower frequency among mothers enrolled in Medicaid and lower education levels, as well as infants born preterm, admitted to the NICU, or who did not room-in [1].
Previous studies have demonstrated infants rooming in with mother encourages initiation of the infant being fed breastmilk, and rooming-in is recommended by numerous public health and clinical organizations to support breastfeeding [1, 16, 17, 18, 19]. While early recommendations included consideration for temporary separation of mothers with COVID-19 from their newborns, multiple studies have now found low incidence of SARS-CoV-2 infection among infants born to people with SARS-CoV-2 infection [20] and low risk of transmission from mother to infant when appropriate IPC is followed [21]. The AAP and CDC recommend that mothers with COVID-19 room-in with their infants and use appropriate IPC measures (e.g., masks, hand hygiene) even if asymptomatic [6, 7].
While most infants who test positive for SARS-CoV-2 have mild or asymptomatic infection, severe disease does occur rarely [22, 23]. In addition to the numerous well-documented benefits of breastfeeding to both mothers and infants [1], accumulating evidence suggests antibodies against COVID-19 are present in breastmilk of mothers with SARS-CoV-2 infection. Multiple reports have now described detection of SARS-CoV-2 specific IgA and IgG in breastmilk after infection [24, 25]. Breastmilk has not been found to contain any SARS-CoV-2 that is replication competent, and thus, there has not been documented risk of transmission through breastmilk [24]. While it is unclear exactly what level of protection against postnatal SARS-CoV-2 infection these antibodies may provide for infants receiving breastmilk, there is a large body of existing evidence that breastmilk feeding reduces infants’ risk of respiratory tract infections [26, 27].
There are several limitations to note. First, important predictors of maternal breastmilk feeding initiation such as marital status and prior initiation of breastfeeding are not included in the SET-NET dataset, and about half of the people did not have enough information available to be able to classify disease severity, which may be an important factor in the ability or decision to initiate breastmilk feeding. Thus, residual confounding may exist. Second, we were not able to assess or control for facility-level factors that may influence breastfeeding initiation. Third, SARS-CoV-2 testing or screening practices have varied over the course of the pandemic, complicating interpretation of trends over time. Finally, this analysis is limited to breastmilk feeding practices at birth hospitalization, does not account for initiation after birth hospitalization, and further follow up of this cohort is needed to assess the full impact of SARS-CoV-2 infection on breastmilk feeding practices overtime.