SARS‐CoV‐2 and human milk: What is the evidence?

Abstract The novel coronavirus SARS‐CoV‐2 has emerged as one of the most compelling and concerning public health challenges of our time. To address the myriad issues generated by this pandemic, an interdisciplinary breadth of research, clinical and public health communities has rapidly engaged to collectively find answers and solutions. One area of active inquiry is understanding the mode(s) of SARS‐CoV‐2 transmission. Although respiratory droplets are a known mechanism of transmission, other mechanisms are likely. Of particular importance to global health is the possibility of vertical transmission from infected mothers to infants through breastfeeding or consumption of human milk. However, there is limited published literature related to vertical transmission of any human coronaviruses (including SARS‐CoV‐2) via human milk and/or breastfeeding. Results of the literature search reported here (finalized on 17 April 2020) revealed a single study providing some evidence of vertical transmission of human coronavirus 229E; a single study evaluating presence of SARS‐CoV in human milk (it was negative); and no published data on MERS‐CoV and human milk. We identified 13 studies reporting human milk tested for SARS‐CoV‐2; one study (a non‐peer‐reviewed preprint) detected the virus in one milk sample, and another study detected SARS‐CoV‐2 specific IgG in milk. Importantly, none of the studies on coronaviruses and human milk report validation of their collection and analytical methods for use in human milk. These reports are evaluated here, and their implications related to the possibility of vertical transmission of coronaviruses (in particular, SARS‐CoV‐2) during breastfeeding are discussed.


| INTRODUCTION
The global pandemic caused by the SARS-CoV-2 virus is one of the most compelling and concerning global health crises of our time.
Fortunately, this pandemic has rapidly mobilized the full range of expertise represented by researchers, clinicians and public health officials. Although our understanding of the biology, clinical implications and strategies for mitigation continues to evolve, one issue that has received limited attention is the implication of this pandemic for infant feeding practices. This lack of attention has resulted in mixed messages regarding guidance about optimal infant feeding practices (e.g., American Academy of Pediatrics, 2020; Centers for Disease Control and Prevention, 2020a;World Health Organization, 2020a; United Nations Children's Fund [UNICEF], 2020) and a consequent lack of confidence about best approaches to infant feeding in the face of this growing pandemic. Even when a mother is positive for COVID-19, the World Health Organization (WHO) recommends breastfeeding be initiated within 1 h of birth, exclusive breastfeeding be continued for 6 months and breastfeeding be continued for up to 2 years. They suggest use of appropriate respiratory hygiene, hand hygiene and environmental cleaning precautions. The UNICEF recommends that COVID-19-positive mothers continue breastfeeding while applying precautions, such as wearing a mask and handwashing before and after feeding (UNICEF, 2020). The U.S. Centers for Disease Control and Prevention (CDC) neither recommends nor discourages breastfeeding but advises that decisions be made by the mother and family in consultation with their health care providers (Centers for Disease Control and Prevention, 2020a). They recommend that during temporary separation (should that occur), mothers who intend to breastfeed should express their milk using proper hand hygiene and that the expressed milk should be fed to the newborn by a healthy caregiver. Further, if a mother and newborn do room-in and the mother wishes to feed at the breast, the CDC recommends that she should wear a facemask and practice hand hygiene before each feeding.
It is well established that viral transmission through human milk can occur (Jones, 2001;Lawrence & Lawrence, 2004). Notable examples include human immunodeficiency virus (HIV; Black, 1996;Ziegler, Johnson, Cooper, & Gold, 1985), cytomegalovirus (CMV; Stagno & Cloud, 1994) and human T-cell lymphotropic virus type 1 (HTLV-1; Boostani, Sadeghi, Sabouri, & Ghabeli-Juibary, 2018). Perhaps the most prominent example of mother-to-child viral transmission via breastfeeding is HIV infection, during which higher milk and serum viral loads are associated with an increased risk of transmission (Davis et al., 2016;Semba et al., 1999;Willumsen et al., 2003). The risk of postnatal infection for breastfed infants of HIV+ mothers is ≈10-20% over the first 2 years of life without the use of antiretroviral therapies (ART; Dunn, Newell, Ades, & Peckham, 1992;Nduati et al., 2001). However, compared with mixed feeding, exclusive breastfeeding is associated with lower risk of transmission of HIV infection to infants (Coutsoudis et al., 2001;Iliff et al., 2005). In many high-income nations, breastfeeding is contraindicated in the case of maternal HIV infection with or without maternal ART (e.g., American Academy of Pediatrics, 2012; Centers for Disease Control and Prevention, 2020b).
Conversely, in low-and-middle-income nations, infant mortality from malnutrition and infectious disease may outweigh the risk of acquiring HIV via vertical transmission during breastfeeding. As such, breastfeeding is recommended (World Health Organization, 2016).
With respect to CMV, it is estimated that 60-70% of breastfed infants of CMV-seropositive mothers become infected with CMV (Dworsky, Yow, Stagno, Pass, & Alford, 1983;Minamishima et al., 1994). The risk of CMV infection in neonates is highest in preterm and very low birthweight (less than 1,500 g) infants (Hamprecht & Goelz, 2017;Lanzieri, Dollard, Josephson, Schmid, & Bialek, 2013). A small percentage of infected infants develop a severe complication known as CMV sepsis-like syndrome, which can be fatal (Fischer et al., 2010). Nonetheless, breastfeeding is not contraindicated in CMV-seropositive women with healthy, term infants For HTLV-1, breastfeeding is considered the major route of infection for infants (Moriuchi, Masuzaki, Doi, & Katamine, 2013). HTLV-1 infection is lifelong, and although most infected individuals remain asymptomatic, approximately 10% develop severe disease, including adult T-cell leukaemia, a highly aggressive and usually fatal malignancy (Rosadas & Taylor, 2019). Some organizations and agencies list maternal HTLV-1 as a contraindication for breastfeeding (American Academy of Pediatrics, 2012; Centers for Disease Control and Prevention, 2019a), whereas others do not (World Health Organization, 2009).
There are seven identified strains known to infect humans. Four of the strains (alphacoronaviruses 229E, NL63 and OC43 and betacoronavirus HKU1) are ubiquitous in humans and cause the common cold. There is limited evidence that one of these (229E) may be vertically transmitted from mothers to infants, although the mechanism remains unclear (Gagneur et al., 2008). The presence of 229E in neonatal gastric samples suggests that one possible mechanism for infection is through human milk, although this study did not evaluate human milk specifically (Gagneur et al., 2008).
In light of the emergence of the novel coronavirus SARS-CoV-2, several issues related to human milk and coronavirus infection demand immediate attention, the first and foremost being whether or

Key messages
• Very little is known about coronaviruses in human milk and whether breastfeeding is a possible mode of vertical transmission.
• Limited, weak evidence suggests that some coronaviruses (including SARS-CoV-2) may be present in human milk, but these studies do not report methods of sample collection and validation of reverse transcription polymerase chain reaction (RT-PCR) assays for human milk.
• Nothing is known about the timing of the antibody response in human milk to SARS-CoV-2 infection.
• Future research should utilize validated methods and focus on both potential risks and protective effects of breastfeeding.
not the virus is present in human milk produced by infected or exposed women. Of particular interest in this context are (1) the potential role that breastfeeding could play in vertical transmission of SARS-CoV-2 from women to infants via human milk and (2) the potential protective effects of targeted antibodies and other immunoprotective components in human milk against COVID-19. The goal of this review was to evaluate the published evidence regarding the presence of this and other human coronaviruses in human milk.

| METHODS
We used a variety of databases to identify relevant literature published as of 17 April 2020, and the list of databases and search terms used can be found in Table 1. It is noteworthy that in addition to using standard scientific databases (e.g., PubMed), we also used a general Google search and a search of preprint servers to identify reports that had not yet been published in refereed journals (i.e., grey literature).
Any research in which human milk was collected and tested for a human coronavirus was included in this review. See Addendum in the supporting information online.

| Ethical considerations
This review does not constitute human subject research, and as such, ethical approval was not required.

| MERS-CoV
The deadliest of the human coronaviruses to date is MERS-CoV, which emerged in Saudi Arabia in 2012. The disease caused by MERS-CoV, Middle Eastern respiratory syndrome (MERS), is characterized by severe respiratory illness with symptoms of fever, cough and shortness of breath. MERS-CoV is a betacoronavirus, and the case fatality rate of MERS is 34% (Mahase, 2020). There are no reports of the presence or absence of MERS-CoV in human milk.
However, there are reports of the presence of MERS-CoV in the milk of dromedary camels (Camelus dromedaries; Conzade et al., 2018;Hemida et al., 2015;Reusken et al., 2014), and there is one report of a human likely infected through the consumption of raw (unpasteurized) camel milk (Memish et al., 2014). In camel milk samples spiked with  Currently, there is one report in which human milk was tested for SARS-CoV (Robertson et al., 2004) and two reports of human milk being tested for SARS-CoV antibodies (Robertson et al., 2004;Stockman, Lowther, Coy, Saw, & Parashar, 2004). Robertson and colleagues described a woman infected during the second trimester of pregnancy (19 weeks). A single milk sample was collected 131 days after the onset of symptoms, but no additional detail on the collection methodologies was provided. Milk was submitted to the CDC, where it was analysed using reverse transcription polymerase chain reaction (RT-

| SARS-CoV
Search terms, databases and preprint servers used to identify existing literature reporting the possibility of vertical transmission of coronaviruses from mother to infant during breastfeeding as of 17 April 2020 The infant in this study was never tested for SARS-CoV infection.

Stockman and colleagues described a 38-year-old woman infected in
the first trimester of pregnancy (7 weeks

| SARS-CoV-2
The novel coronavirus SARS-CoV-2 was named after SARS-CoV due to its shared sequence homology (77.9%; Kim et al., 2020) and similar clinical characteristics. The first reported cases of SARS-CoV-2 infection emerged in late 2019 in China. Although the current estimated case fatality rate for COVID-19 (the disease caused by the SARS-CoV-2 virus) is much lower than those of SARS and MERS at roughly 2% (Mahase, 2020), the spread of this pathogen has been much more rapid and extensive.
At the time of writing, there were 13 studies (seven case studies and six case series; three of which were preprints or preliminary reports that had not been formally peer-reviewed; Table 2) reporting direct testing of milk produced by women infected with SARS-CoV-2 Dong et al., 2020;Fan et al., 2020;Liu, Wang, Li, et al., 2020;Liu, Wang, Zhang, et al., 2020;Salvatori et al., 2020;Wang et al., 2020;Yu et al., 2020;Wu et al., 2020) or by women whose infants were infected Kam et al., 2020;Yuehua et al., 2020). In total, 48 milk samples produced by 32 women had been tested; all but one sample (Wu et al., 2020;non-peer-reviewed preprint) were negative for the presence of the virus. Two milk samples produced by a single woman were tested for SARS-CoV-2 specific antibodies; IgG but not IgM was identified in both samples non-peer-reviewed preprint). A description of the relevant characteristics for the women and infants in these studies can be found in Table 2. Investigators conducting nine of the 13 studies analysed milk samples collected at birth or shortly thereafter, reporting only findings in colostrum or transitional milk. Those same nine studies reported on the milk produced by women who were infected during the third trimester of pregnancy, whereas the other four reported findings from milk produced by mothers of infants infected at 1.5, 3, 6 and 13 months of age Kam et al., 2020;Yu et al., 2020;Yuehua et al., 2020). For the infants born to women infected during pregnancy, most were immediately separated from their mothers post-delivery and were not breastfed for the duration of the period observed in their respective reports. Fourteen of the 32 infants described in these reports were born via caesarean section; only two were specified as vaginal births.
Repeated milk samples, collected up to 27 days apart, were analysed for eight of the women. All the studies were conducted in China Cui et al., 2020;Dong et al., 2020;Fan et al., 2020;Liu, Wang, Li, et al., 2020;Liu, Wang, Zhang, et al., 2020;Wang et al., 2020;Yuehua et al., 2020Wu et al., 2020), Singapore (Kam et al., 2020) or Italy (Salvatori et al., 2020). pneumonia were collected after their first lactation' and that milk was collected following WHO guidelines, but they did not provide a citation for this collection method. All milk tested negative for the virus, but no information was provided on the methods used for analysis.
In a report by Liu, Wang, Zhang, et al. (2020), milk produced by two women was tested. One woman was 34 years old and at In a case series by Liu, Wang, Li, et al. (2020), milk produced by 10 women infected during late pregnancy was tested via RT-PCR for SARS-CoV-2; all samples tested negative. It is noteworthy that this report included data from 19 women, but milk was collected from only 10 of them. The authors did not specify for which of the women milk was collected. None of the 19 infants reported in this study tested positive for SARS-CoV-2 via RT-PCR. The only detail available on collection or testing methods is that RT-PCR was used to test the samples and that 'milk was collected after the first lactation' . Despite their results, the authors concluded that delivery should occur in an isolation room and that infants be separated from infected mothers.
Li et al. (2020) described a 30-year-old woman at 35-week gestation who was positive for SARS-CoV-2 and who delivered a male infant via emergency caesarean section. The infant was tested immediately upon delivery via oropharyngeal swab, which was negative.
After delivery, the infant was kept in isolation away from his mother.
Milk was collected immediately after delivery and on Days 2 and 3 post-partum; all samples were negative. Again, no information on the collection or testing methods for the milk sample is available in this report. The infant's breastfeeding status was not specified in the report, but it is presumed that he was at least partially breastfed as the mother was producing milk at 6-month post-partum.
d Study presented data from three women but only presented data on the milk produced by two women.
e Study presented data from 19 women but only presented data on the milk produced by 10 women.
f Study presented data from 13 women but only presented data on the milk produced by three women. While the previous reports focused on infected women, there are also three case studies focused on infected infants. In these studies, milk produced by the infants' mothers was tested for SARS-CoV-2. The youngest of these infants was reported by Cui et al. (2020). After being exposed to infected family members, the 55-day-old female was admitted to the hospital with symptoms of COVID-19 and diagnosed based on clinical data and exposure history. The infant was 'mixed fed'. Her mother's milk was collected on the first three consecutive days of her hospitalization; all were negative for SARS-CoV-2. No information on the collection or testing methods for the milk sample is included in this report. Yuehua et al. (2020) reported on a 3-month-old, breastfed female who was hospitalized and tested via throat swab for SARS-CoV-2; the swab was positive. A single milk sample was collected from the infant's mother; it tested negative. The authors provided no information on the collection or testing methods for the milk. Importantly, this infant developed symptoms of COVID-19 7 days before her parents became ill. As such, one possibility is that she was infected first and passed the infection to them. Another case report on a mature milk sample comes from Singapore (Kam et al., 2020). This report is particularly interesting as the infant had no symptoms but was hospitalized and tested because his caregivers were all hospitalized with COVID-19, and there was no one to care for him. The infant was 6 months old and presumably at least partially human milk fed as a sample of milk was successfully collected from his mother. Despite being asymptomatic, a nasopharyngeal swab taken from the infant was positive for SARS-CoV-2. The authors reported that milk produced by the mother on a single day tested negative for the virus but did not specify how many samples were taken. This report provided no data on the methods used for the collection and analysis of these sample(s).

Despite the devastating clinical manifestations of MERS-CoV, SARS-
CoV and SARS-CoV-2, there remains much to be learned about their modes of transmission. Respiratory droplets are a documented source of the virus (World Health Organization, 2020b), but other sources such as breastfeeding and/or human milk may exist. The primary purpose of this review was to examine the evidence (or lack, thereof) for the vertical transmission of SARS-CoV-2 from mother to infant via breastfeeding considering what is known about other human coronaviruses. We also examined the evidence presented in the same reports related to maternal/infant antibody production to the virus.
In total, we identified 14 studies that had tested human milk for human coronaviruses directly. Thirteen of these studies were newly published reports on SARS-CoV-2 and human milk, which collectively encompassed 48 milk samples. All but one of these samples tested negative for SARS-CoV-2, and that result was reported in a non-peerreviewed, online preprint. We identified no comparable data for MERS; a single case report for SARS, which yielded a negative result for the presence of the virus but positive results for antibodies specific to SARS-CoV; and no reports of human milk tested for other human coronaviruses. There was one report of antibody tests in milk specific to SARS-CoV-2, which identified IgG but not IgM .
The single milk sample that was supposedly positive for SARS-CoV-2 (Wu et al., 2020) is clearly anomalous to the larger body of evidence. Importantly, since no methodologies were provided and the study has not been published in a refereed journal, any interpretation of this result must be made with extreme caution. This underscores the fact that it is critical for the research community to focus its efforts on analysing appropriately collected human milk using laboratory methods that have been validated and optimized for the human milk matrix. In addition, results should be submitted to high-quality, refereed journals for peer review. Only then can methods and results be rigorously assessed by scientists with the knowledge base needed to judge them. The dearth of high-quality evidence substantially compromises the ability to effectively respond to this pandemic and provide guidance to some of the most vulnerable individuals: pregnant and lactating women and infants.
Limited and weak data suggest MERS may be present in camel milk, but the relevance to SARS-CoV-2 in human milk is unclear. Notably, Reusken et al. (2014) reported that milk analysed in the camelid studies was not collected aseptically; rather, samples were obtained according to local milking customs. As such, it is possible that the presence of MERS-CoV in camel milk could be due to contamination from the milker, the calf or the environment, rather than milk representing an endogenous source of the virus. This is likely an issue with all the studies on SARS-CoV-2, where only one reported cleaning of the breast prior to sample collection (Wu et al., 2020). However, the limited data available on all three of these viruses (and human coronaviruses, in general) leave many questions unanswered with respect to the role, if any, of human milk in vertical transmission of coronaviruses.
Thus, validation of methods using human milk is needed (see Box 1).
In addition, other than general statements about the timing of collection (e.g., 'milk was collected after the first lactation') and brief descriptions of the RT-PCR assays used for nasal and throat swabs, none of the studies to date has described the methods of collection or how the milk was handled and stored in any detail. In addition, nothing is known about stability of SARS-CoV-2, if present, in human milk and how quickly (or at what temperature) it must be frozen to preserve fidelity. Information on sample collection, handling and storage is critical to evaluating whether the negative results described in these studies could be due to inadequate methods used.

Box 1. Key points of assay validation
Some factors to consider when validating methods for human milk testing of coronaviruses.
• Method of milk collection: use of manual milk expression versus electric pump, cleaning procedures of breast and pump, partial versus full breast expression and foremilk versus hindmilk.
• Sample handling and storage: container material, temperature and duration of refrigeration/freezing.
• Assay validation: nucleic acid extraction protocols, amplification protocols, reagent selection, proper positive and negative controls and fresh versus frozen milk.
• Viral quantification and viability: infectious dose and biologically relevant concentrations.
Another possibility is that there is low abundance of the virus in human milk, and it is often not captured in the limited samples tested so far. For example, in the report on other human coronaviruses by Gagneur et al. (2008), 159 maternal-infant dyads were tested (including 161 infants, two sets of twins). In this report, 229E was present in both maternal and infant samples in only two dyads. Additionally, in the milk of dromedary camels, MERS-CoV appears to be present at very low abundance (Reusken et al., 2014). This suggests the possibility that a very low viral load in milk might also lead to an inflation of false negatives. Limited evidence from two patients suggests that SARS-CoV-2 shedding in respiratory samples peaks at 6 days after onset of symptoms (Pan, Zhang, Yang, Poon, & Wang, 2020), indicating that timing of sample collection also plays an important role in virus detection.
Only one study has investigated antibodies in milk specific to SARS-CoV-2 non-peer-reviewed preprint), and these researchers identified IgG in one milk sample produced by a woman at 13-month post-partum. Although limited to a single study, this finding combined with a large body of literature documenting targeted antibodies in human milk indicates that there may be a protective effect of breastfeeding when the mother is COVID-19 positive. The infant in this study was older than all the other infants described here, was likely not exclusively breastfed (based on reported age) and likely had a more mature immune system than the youngest infants described in other reports. Still, further investigation into this finding is a critical next step in understanding how breastfeeding and/or the infant's con-  (Robertson et al., 2004). Together, these observations suggest that infant infection may occur in utero but that the virus may simply be absent from the upper respiratory tract immediately after birth and therefore undetectable on pharyngeal swabs.
ACE2 is expressed across many body sites and tissue types, including the oral cavity (e.g., tongue and oral mucosa) and in mammary tissue . If mammary epithelial cells express this receptor, then it follows that viable virus could exist in milk. If it does, then the introduction of virus-containing human milk could represent a mechanism of entry for SARS-CoV-2 and COVID-19 infection for infants.
Another observation worth considering is that, in at least one of the reports (Yuehua et al., 2020), the infant was infected and symptomatic 7 days prior to the infant's parents. This suggests the possibility that a 'reverse' vertical transmission from infant to mother could occur, a phenomenon that has been observed for other pathogens, such as HIV (Belitsky, 1989;Little et al., 2012) and Ebola virus (Sissoko et al., 2016). One possible mechanism for maternal infection in this case is through retrograde flow, where milk and saliva move back into the mammary gland from the infant's mouth during suckling (Ramsay, Kent, Owens, & Hartmann, 2004). Although this mechanism is speculative, it represents a possible route whereby an infant could theoretically transfer a pathogen it has encountered in the environment to the mother. It is also possible that maternal infection could occur through other mechanisms, such as infant respiratory droplets (World Health Organization, 2020a) or via faecal matter .
To date, all reports on SARS-CoV-2 and human milk have originated in Asia (China and Singapore) or Europe (Italy). Although this limited geography makes sense given the fact that the initial epicentre of this pandemic was in Asia and this was followed by a large outbreak throughout Italy, studies from other globally representative populations are needed to make definitive conclusions regarding the possible presence and/or role of SARS-CoV-2 in human milk. Additionally, the importance of a coordinated, international effort by scientists, clinicians and public health officials to elucidate answers to the many remaining questions related to SARS-CoV-2 and breastfeeding cannot be overemphasized.

| CONCLUSIONS
Human milk is the gold standard for infant feeding. However, confidence as to its safety and best practices around breastfeeding during maternal COVID-19 infection has been compromised by the lack of rigorous evidence as to whether SARS-COV-2 can be vertically transmitted in milk and/or during breastfeeding. As such, there exists an immediate need to rapidly generate rigorous evidence for the role Substantial interdisciplinary research on this topic is required and should be performed rigorously and rapidly to best inform policies regarding early feeding choices and clinical management of breastfeeding mothers infected with SARS-CoV-2 and their infants.

Box 2. Future needs
To understand the role of human milk and SARS-CoV-2 infection, the following points must be rapidly addressed.
• Optimization of human milk collection and storage protocols for SARS-CoV-2 research.
• Validation of assays for identification of SARS-CoV-2 RNA and SARS-CoV-2-specific immune components in human milk.
• Multinational population studies documenting presence or absence of SARS-CoV-2 virus and immune factors (including antibodies) in milk produced by infected women, women with infected infants and women who have been exposed to SARS-CoV-2; if the virus is identified in milk, its viability must be verified.
• Multinational population studies documenting (or not documenting) risk of COVID-19 infections in breastfed versus nonbreastfed infants whose mothers are COVID-19 positive.
• Research delineating implications of skin-to-skin breastfeeding versus consumption of pumped human milk.