Update on the Transmission of Zika Virus Through Breast Milk and Breastfeeding: A Systematic Review of the Evidence

We systematically searched regional and international databases and screened 1658 non-duplicate records describing women with suspected or confirmed ZIKV infection, intending to breastfeed or give breast milk to an infant to examine the potential of mother-to-child transmission of Zika virus (ZIKV) through breast milk or breastfeeding-related practices. Fourteen studies met our inclusion criteria and inform this analysis. These studies reported on 97 mother–children pairs who provided breast milk for ZIKV assessment. Seventeen breast milk samples from different women were found positive for ZIKV via RT-PCR, and ZIKV replication was found in cell cultures from five out of seven breast milk samples from different women. Only three out of six infants who had ZIKV infection were breastfed, no evidence of clinical complications was found to be associated with ZIKV RNA in breast milk. This review updates our previous report by including 12 new articles, in which we found no evidence of ZIKV mother-to-child transmission through breast milk intake or breastfeeding. As the certainty of the present evidence is low, additional studies are still warranted to determine if ZIKV can be transmitted through breastfeeding.


Introduction
Zika virus (ZIKV), a virus from the Flaviviridae family, is an arthropod-borne virus.
Local ZIKV transmission has been reported in 87 countries and territories, which include the Americas, Asia, Africa, and the Western-Pacific region (1). ZIKV infection during pregnancy can lead to congenital Zika syndrome that is characterized by severe central nervous system malformations in developing fetuses such as congenital microcephaly (2,3). Several ZIKV outbreaks have occurred worldwide, most predominantly in the Western Hemisphere; which included outbreaks in Yap Island in 2007 (4), French Polynesia in 2013 (5), and most recently, Brazil and the Americas, with an exponential increase in cases between 2014 and 2016 (6). Due to the unprecedented rise in ZIKV cases and the associated risks of pregnancy complications and birth defects, the World Health Organization (WHO) declared ZIKV outbreak an international public health emergency in November 2016 (7)(8)(9)(10). In 2018, the WHO included ZIKV infection in the Research and Development Blueprint list among the priority diseases that pose the greatest public health risk due to their epidemic potential (11).
ZIKV is primarily transmitted via mosquito vectors from the Aedes genus, mostly Aedes aegypti, the same mosquito that transmits Dengue, Chikungunya, and yellow fever viruses (10).
Sexual transmission also has been identified (12), and ZIKV RNA has been found in amniotic fluid, breast milk, semen, saliva, urine, and blood (10,13,14). After exposure, the incubation period is estimated to be between 3 to 14 days, and it is often followed by either an asymptomatic or mild non-specific disease. If symptomatic, clinical manifestations include fever, pruritic maculopapular rash, arthralgia, and headache (14). To a lesser frequency, other symptoms might include myalgia, gastrointestinal distress, retroorbital pain, and lymphadenopathy.
Due the absence of clinical manifestations, ZIKV infections are often misdiagnosed with other arboviruses such as dengue fever infection, as well as other infections endemic to tropical regions. To facilitate the diagnosis of ZIKV disease, in 2016, the WHO established interim guidance for ZIKV laboratory testing (15). This WHO recommendations advise the collection of whole blood or urine samples for nucleic acid testing, via reverse transcription-polymerase chain reaction (RT-PCR), during the first seven days of symptoms onset. Thereafter, serology and/or RT-PCR is recommended for assessment. This is due to viremia dropping rapidly after seven days, at which point serology is recommended, and a negative RT-PCR result does not exclude ZIKV infection.
Despite health systems efforts to screen and counsel pregnant women for the potential risks associated with ZIKV infections, the virus remains a major public health issue for maternal and child health. ZIKV infection during pregnancy has been associated with intrauterine fetal demise and miscarriage (14,16,17). Moreover, an estimated 5-15% of infants born to mothers with ZIKV infection have been reported to have congenital complications including microcephaly and a series of congenital malformations referred to as congenital Zika syndrome (CZS) (10,17). CZS includes microcephaly, brain damage, subcortical calcifications, and a multitude of developmental disorders resulting in pulmonary, ocular, and musculoskeletal defects (18) (19).
While it has been established that mother-to-child transmission of ZIKV may occur during pregnancy or at the time of birth, less is known about transmission through breast milk and breastfeeding practices (20). In our previous rapid systematic review (13), we found limited evidence of the risk of ZIKV transmission through breast milk intake or breastfeeding. At present, WHO guidelines advise standard breastfeeding practice for all mothers regardless of ZIKV infection status (21). The current review aims to assess the available evidence of the possible transmission of ZIKV through breast milk or breastfeeding practices to update findings from the initial systematic review and contribute to the development of evidence-informed guidelines at a national, regional, and global level.

2.1.1Types of Studies
We aimed to include the following study designs in this review: randomized controlled trials (RCTs), quasi-RCTs, and all observational studies (i.e., cohort studies, case reports, and surveillance reports.

2.1.2Participants
Included participants were breastfeeding women and children with confirmed, probable or suspected ZIKV infection. This includes participants who were currently breastfeeding, as well as those who were breastfeeding before a ZIKV presumptive diagnosis. Briefly cases were defined as suspected cases: any infant or breastfeeding woman who had been vaccinated for ZIKV, traveled or lived in an endemic area within the last seven days from the start of symptoms; probable case: a suspected case with a presence of viral antibodies against ZIKV; a confirmed case: any individual with laboratory confirmation of recent ZIKV infection defined by the presence of ZIKV RNA or antigen in serum or other biological samples or IgM antibody against ZIKV positive and plaque reduction neutralization test ≥ 90% (PRNT90) for ZIKV with titer ≥ 20 and ZIKV PRNT90 titer ratio ≥ 4 compared to other flaviviruses. Studies with populations that did not meet these criteria, tested breast milk samples, or had a non-ZIKV infection, were excluded.

Types of exposure
Exposure criteria were described as any woman with ZIKV infection who was breastfeeding or intending to breastfeed an infant aged from 0 to 2 years.

Types of outcomes
Primary outcomes included infants with suspected, probable, or confirmed ZIKV infection within 30 days of breastfeeding or receiving expressed breast milk from a woman with suspected, probable, or confirmed infection. Secondary outcomes included detection of ZIKV in breast milk, maternal blood, sweat, or saliva, or infant's saliva by detection methods that identify suspected, probable, and confirmed cases. Detection methods of ZIKV infection in maternal and infant samples include: ZIKV RNA by RT-PCR, ZIKV-specific IgM antibody by ELISA, PRNT90 for ZIKV with titer > 20 and ZIKV PRNT90 titer ratio > 4 compared to other flaviviruses and, ZIKV isolation in culture.

Data extraction and management:
All included reports were screened independently by two authors using the Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia) every disagreement was resolved by consensus or a third author. A data extraction form was developed and piloted for data extraction. Two authors extracted data and discrepancies were resolved by discussion. It was not possible to calculate effect estimates since all the included studies were observational, case reports, and longitudinal studies, with a limited number of cases.

Quality of the evidence
The GRADE approach was used to ascertain the certainty of the evidence (23). Data on the primary and secondary outcomes were considered; (1) ZIKV infection in infants breastfeeding from a mother with confirmed, probable, or suspected ZIKV infection, and (2) the detection of ZIKV RNA in breast milk samples from mothers with confirmed or suspected ZIKV infection. The GRADE approach included the risk of bias, the directness of evidence, inconsistency (heterogeneity), the precision of effect estimates, and the risk of publication bias across the included studies. All the included studies are observational and with few cases, and they were downgraded one level to low-certainty of the evidence, and further study limitations led to downgrading to very-low certainty of the evidence. Considering that all included studies are observational and with few events, the evidence provided by these studies is heterogenous, and they do not allow for pooled estimates, the certainty of the evidence is described as a narrative.
Noteworthy, this updated systematic search identified four different articles that reported on the same cases: Besnard, 2014 (8) and Besnard, 2017 (24) reported on the same two mother and child pairs; similarly, Blohm, 2017 (29) and Blohm, 2018 (30) reported on a single mother and child pair. Only one of each duplicated study was considered during data extraction and analysis.
Among the studies informing our analysis, 10 were case reports (24-30, 32, 35, 37-39), and four were longitudinal studies (31,33,34,36). The records informing the present analysis included a total of 177 mother and child pairs, but only 97 women provided a breast milk sample for assessment. Our data extraction, synthesis of evidence, and analysis is solely based on the maternal-infant pairs with breast milk samples, all other cases were excluded.

Participants
Among the case reports, there were14 children were assessed for ZIKV infection. Among them there were eight neonates tested at birth (26-28, 35, 37, 38) or during the first three days of life (39,40). Additionally, there were six infants that were tested for ZIKV infection between 5 -11 months of age (26,30,32).
In one longitudinal study (33), 72 pregnant women confirmed or suspected for ZIKV infection were followed until delivery to evaluate potential adverse pregnancy outcomes. Only 15 out of the 72 these women provided a breast milk sample for ZIKV assessment within 24 hours after delivery. A different longitudinal study (34) aimed to assess the impact of ZIKV infection of breast milk viscosity. Forty pregnant women were recruited, 20 women with confirmed ZIKV infection via RT-PCR assessment, and the remaining 20 women were described as clinically healthy. Women from both groups provided one breast milk sample between 48 and 72 hours postpartum (34). The latter is the only study that included a control group during their assessment. In another study (36), people 16 years and older were recruited from two health centers if they presented with two or more symptoms associated with ZIKV infection in 14 days or less. Among the participants with ZIKV infection, there were a total of 18 pregnant women, of which seven were confirmed by RT-PCR analysis in plasma, serum, or urine. From this study, only one woman provided three breast milk samples for analysis. Another longitudinal study (31) evaluated ZIKV infection in 27 postpartum women. Limited data were available from this cohort in the form of an abstract.

Child outcomes
There were 97 mother-children pairs included in this analysis who provided breast milk for ZIKV analysis. From these mother and child dyads, the outcomes for 14 cases were described as case of reports (25-28, 30, 32, 35, 37-39) and 83 pairs in longitudinal studies (31,33,34,36). In total, this analysis included six infants with ZIKV infection confirmed by viral RNA detection and 50 children with negative RT-PCR tests for ZIKV infection. A summary description of these cases is included in Table 1, and further information is provided in the supplementary table 1.
Six out of 14 infants from case reports were confirmed with ZIKV infection through viral RNA detection with an RT-PCR assay, five cases tested positive in a blood sample by RT-PCR assay (28,30,37,39), one case was found negative in a cord blood sample but positive in urine by RT-PCR test (27). One neonate had ambiguous results from the RT-PCR assay (25).
Most mothers described in the case reports had a confirmed ZIKV infection by RT-PCR test, except one mother with negative RNA detection in plasma but positive IgM levels detected (29), another woman who presented symptoms during the first trimester and elevated IgM levels were found at 38 th gestational weeks (28). One pregnant woman had negative RT-PCR and antibody results, as well as her newborn; however, ZIKV RNA was detected in her breastmilk

Zika virus in breast milk
Of  Table 1 and more information can be found in the supplementary table Three breast milk samples from one woman were positive for ZIKV RNA and were collected 14 days after maternal symptoms onset during the 36 th gestational week. Two other samples were collected from the samples mother, the second and third samples were collected 20 two weeks and nine days after the first sample (27). All samples were positive for viral RNA.

21
Another 13 positive breast milk samples from different women were tested from 2 days to 2

22
weeks after maternal symptom onset (25,26,28,30,31,39). One breast milk sample, with detected viral RNA, from another woman was tested at birth, while maternal symptoms after birth from an asymptomatic woman who had a baby with severe microcephaly 19 months earlier from a previous pregnancy (35). In one study, stored breast milk samples from three 27 days before maternal symptoms onset were analyzed (32) resulting in the detection of ZIKV

29
Viral cell culture was attempted from seven breast milk samples corresponding to seven 30 different women (25-27, 30, 39), out of which ZIKV replication was detected in cell cultures 31 from five breast milk samples (25-27, 30, 36). Furthermore, one case report found that viral RNA 32 isolated from breast milk and child serum were genetically related to each other (28). Another 33 case report found identical viral isolates from mother and child based on the NS5 gene The search strategy did no identified any studies assessing ZIKV in breast milk after maternal sweat, tears, or skin among the mother-infant dyads providing breast milk samples.

Certainty of the evidence
The certainty of the evidence was assessed following the GRADE approach, all the number where the breast milk samples were tested. Further, the studies lacked control groups, and reporting on ZIKV infection among infants by molecular or serological tests was  exclusively breastfed children (43). Most of the studies included in this review assessed breast 100 milk samples obtained shortly after birth, without serial samples over time and sparse reporting of infant feeding practices, it will be critical to consider these aspects to further understand the dynamics of possible viral transmission by breast milk intake.
breast milk is a route of ZIKV transmission. The first review (13), was undertaken by our group and described two studies reporting on three cases of ZIKV-infected breastfeeding mothers.

106
ZIKV RNA in the breast milk samples from the three women, and two of the newborns were 107 positive for ZIKV infection. In a different systematic review, the authors investigated flavivirus 108 transmission through breast milk (44) and found five articles relevant to ZIKV transmission.

109
Although they synthesize evidence regarding the potential mother-to-child transmission of

110
ZIKV through breast milk, breastfeeding practices were not specified. Another systematic review (45) looking into the possible transmission of ZIKV through breast milk, identified ten mother-and-child pairs and and found no evidence of long-term child complications. The

113
current review expands on the number of mother-child pairs included in these analyses. reporting of the methods for obtaining and processing breast milk samples will be critical to 151 understand viral dynamics in breast milk and the potential of viral transmission. Additionally,

153
suggesting potential transmission routes due to increased contact between mother and the child 154 while caring and breastfeeding, it is important to consider the assessment of these bodily fluids 155 when evaluating the risk of mother-to-child transmission of ZIKV through breastfeeding.

156
In summary, this systematic review included evidence from 10 case reports and four 157 longitudinal studies. There were 97 mother-infant pairs from where a breast milk sample was 158 available for analysis, 77 were cases where the mother had suspected or confirmed ZIKV