Transmission of Zika virus through breast milk and other breastfeeding-related bodily-fluids: A systematic review

Background Zika virus (ZIKV) infection is an emerging mosquito-borne disease, which is associated with an increase in central nervous system malformations and newborn microcephaly cases. This review investigated evidence of breastfeeding transmission from ZIKV-infected mothers to their children and the presence of ZIKV infection in breastfeeding-related fluids. Methodology/Principal findings We conducted a systematic review of observational studies, case studies, and surveillance reports involving breastfeeding women with ZIKV infection in several international databases. Data extraction and analysis were conducted following a PROSPERO-registered protocol. From 472 non-duplicate records, two case reports met criteria for inclusion. We reviewed three cases of ZIKV infection among lactating mothers near the time of delivery. Two of the three (2/3) associated newborns had evidence of ZIKV infection. ZIKV was detected in breast milk of all three mothers. Breast milk detection results were positive in all mothers (3/3) by RT-PCR, one was positive by culture (1/3), and none was tested for ZIKV-specific antibodies. Serum samples were ZIKV positive in all mothers (3/3), and sweat was not tested for ZIKV. Conclusions/Significance We describe three cases of ZIKV-infected breastfeeding mothers who were symptomatic within three days of delivery, and two cases with ZIKV-infected newborns. While ZIKV was detected in the breast milk of all three mothers, the data are not sufficient to conclude ZIKV transmission via breastfeeding. More evidence is needed to distinguish breastfeeding transmission from other perinatal transmission routes.


Introduction
Zika virus (ZIKV) infection is an emerging vector-borne disease of the Flaviviridae family, which includes dengue, yellow fever, Japanese encephalitis, and West Nile viruses [1]. ZIKV infection causes a mild, self-limiting influenza-like illness with a 10-day incubation for most cases and shares similarities with other circulating arthropod-borne viral infections like the alphavirus chikungunya [1,2]. Many cases of ZIKV infection are asymptomatic and therefore unreported.
The World Health Organization (WHO) has developed an interim case definition to classify and report cases of ZIKV infection (Fig 1). A suspect case is a person presenting with rash and/ or fever and at least one of the following: arthralgia, arthritis or conjunctivitis. A probable case is a suspected case with presence of IgM antibody against ZIKV and an epidemiological link; and a confirmed case is a person with laboratory confirmation of recent ZIKV infection: by presence of ZIKV RNA or antigen in serum or other samples or IgM antibody against ZIKV positive and plaque reduction neutralization test ! 90% (PRNT 90 ) for ZIKV with titre ! 20 and ZIKV PRNT 90 titre ratio ! 4 compared to other flaviviruses [3]. Due to the possible cross reactivity with other members of the Flaviviridae family, the presence of IgM is not enough to rule out ZIKV infection, and the PRNT 90 will determine if the in vitro inhibition of cell growth is produced by antibodies against ZIKV [4,5]. An enzyme linked immunoassay (ELISA) for ZIKV has been developed by the Centers for Disease Control and Prevention, but is only available upon request for emergency use [6].
The timing and the test performed could be crucial for detecting the ZIKV infection. During the first 7 days, viral RNA can often be identified by reverse transcriptase polymerase chain reaction (RT-PCR), but as viremia decreases, a negative RT-PCR does not exclude flavivirus infection, and serologic testing should be performed. On the other hand virus-specific IgM antibodies may be detectable >4 days after onset of illness, however a sample taken within 7 days of illness onset may not have detectable virus-specific IgM antibodies [7].
Until recently, outbreaks of ZIKV were sporadic. During the last 50 years, widespread infection throughout Africa and Southeast Asia is suspected, but the asymptomatic nature and limited diagnostics have likely hampered disease surveillance [12,[26][27][28]. In 2007, the disease migrated to Oceania where an outbreak in Yap State in the Federated States of Micronesia infected roughly 5,000 individuals, nearly 75% of the island population [2]. The next outbreaks occurred in French Polynesia (396 confirmed), New Caledonia (1,400 confirmed), and the Cook Islands (50 confirmed) in 2013-2014 [29][30][31]. The first official outbreak in the Americas arrived to Easter Island, Chile in early 2014 with 51 confirmed cases [32]. In April, 2015, Brazil reported the first confirmed autochthonous case of ZIKV infection [33]. Since then, an epidemic has rapidly expanded affecting 48 countries and territories in South and Central America [34]. The Brazilian Ministry of Health estimates the number of ZIKV cases in 2015 alone between 0.4-1.3 million [8].
During the Brazilian ZIKV epidemic, clinicians have observed a 20-fold increase in suspected cases of microcephaly in newborns [35]. Reported microcephaly and/or central nervous system malformations have affected 7,150 individuals in Brazil between 22 October 2015 and 16 April 2016 [36]. Other flaviviruses have not been known to cause microcephaly, however ZIKV has been confirmed in recent microcephaly cases, which has prompted global concern for pregnant women and a large-scale investigation [37]. A recent report from the WHO indicated that there is scientific consensus that Zika virus is a cause of microcephaly and Guillain- Barré syndrome [36,38] based on results from a systematic review [39,40]. Many mothers of infants with microcephaly reported no illness or symptoms associated with Zika infection [41]. Regardless of symptoms, pregnant women are at risk for infection and potential complication in any trimester [13,42]. At this time, the WHO recommends standard breastfeeding practices for all mothers, regardless of ZIKV infection [43], unless there is an acceptable medical reason for permanent or temporary avoidance of breastfeeding [44].
The primary objective of this systematic review was to review evidence related to the transmission of ZIKV through breastfeeding. For the purposes of this review, ZIKV infection included suspected, probable, or confirmed cases as described by the WHO interim case definition. A secondary objective assessed the available literature regarding the presence of ZIKV or ZIKV-specific antibodies in breast milk and breastfeeding-related bodily fluids (i.e. blood or sweat) of lactating women. We sought to address the following questions: Primary Outcome: Does the literature provide evidence that in ZIKV-free infants or children, breastfeeding (any or exclusive) from a ZIKV-infected lactating mother, compared to not breastfeeding, result in evidence of ZIKV infection in the infant?
Secondary Outcome: Does the literature provide evidence there are ZIKV specific antibodies present in breast milk?

Study criteria
Study characteristics, as well as inclusion and exclusion criteria, were defined by study designs, participants, ZIKV infection exposure, and outcomes.
Types of studies. We included observational studies, case studies, and surveillance reports, which include epidemiological data from outbreak investigations.
Types of participants. Study participants were limited to adolescents age 10-19 years or adult women who were lactating, or expressing milk, with ZIKV infection. This includes lactating participants who were currently breastfeeding or not, as well as those who were breastfeeding prior to a ZIKV presumptive diagnosis. Studies with populations who did not meet these criteria or who had a non-ZIKV infection were excluded.
Types of exposure. Exposure criteria were described as any mothers with ZIKV infection who were breastfeeding or expressing breast milk.
Types of outcomes. Primary outcomes included infants or children with any ZIKV infection (suspected, probable or confirmed cases), within 30 days of breastfeeding or receiving expressed breast milk from a mother with ZIKV infection. Secondary outcomes included the detection of ZIKV in breast milk, maternal blood, maternal sweat or infant saliva by the detection methods that allow identification as suspected, probable and confirmed cases. Detection methods included:

Search strategy
A search overview is provided in the S1 Appendix.
Electronic databases: Search terms included variations and permutations of United States National Library of Medicine Medical Subject Headings (MeSH) terms and text words relating to flaviviruses (Zika, West Nile, and yellow fever), breastfeeding, transmission fluids (breast milk, blood, and sweat), and participants (mother or child) (See appendix for full search strategy). Report characteristics included a time range of all years, any language, and any publication status. The following electronic databases were searched:

Study selection
Screening of search results was performed using Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia). Two authors independently screened the titles and abstracts of studies based on the inclusion criteria. A third author assessed and resolved disagreements on study selection. All irrelevant titles were excluded. For studies that met eligibility criteria, full text articles were obtained and managed using EndNote (version X7Á5 2016 Thomson Reuters), a reference management software.

Data extraction and management
A data extraction form was tailored for this review. One author extracted study characteristics and two authors extracted study outcome data according to the pre-designed data extraction form. For each study, information pertaining to the source, eligibility, methods, participants, exposures, outcomes, and results was entered into the data extraction form. When relevant, effect estimates including odds ratios, relative risks, mean differences, or summary effects were extracted for each outcome. All potential modifiers or confounders of study outcomes were included in the extraction form.
This review followed a pre-established protocol based on methods for systematic reviews described in the Cochrane Handbook for Systematic Reviews [45]. The protocol was registered in PROSPERO, the international prospective register of systematic reviews of the University of York and the National Institute for Health Research, under the number CRD42016036667. The authors followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and include a checklist in the S1 Checklist.
Quality of the evidence. We set out the main findings of the review in Summary of Findings tables prepared using GRADE profiler software (GRADEpro Guideline Development Tool, McMaster University, 2015, developed by Evidence Prime, Inc.). The primary outcomes were listed with estimates of relative effects along with the number of participants and studies contributing data for the outcomes. For each individual outcome, we assessed the quality of the evidence using the GRADE handbook for grading quality of evidence [46]. We expressed the results as one of four levels of quality (high, moderate, low, or very low).

Results
Our search strategy (11 March 2016) identified 472 records, detailed in the S1 Table, after duplicates were removed (Fig 2). No unpublished records were identified in the search nor included in the analysis. Initial screening retained 42 records. At the time of screening, inclusion criteria included terms for West Nile and yellow fever viruses. For the purposes of this review, only ZIKV was considered for quantitative analyses, which yielded 2 records for review (Fig 2). A total of 2 studies (that included three mother child pairs) were included for analysis. The main reasons for exclusion were non-ZIKV infections or ineligible populations.
A case report from the ZIKV outbreak in French Polynesia (French territory) described two mothers, who had recently given birth, with ZIKV infection (Table 1) [13]. Mother 1 initiated breastfeeding to Newborn 1 on the day of delivery. On day 2 following delivery, mother 1 had a confirmed case of ZIKV detected by serum RT-PCR and saliva RT-PCR. On day 3, the breast milk from mother 1 was found to contain ZIKV by RT-PCR, however ZIKV breast milk culture was negative. Also on day 3, Newborn 1 had confirmed ZIKV infection by serum RT-PCR and saliva RT-PCR.
Mother 2 was confirmed with ZIKV infection on days 1 and 5 post delivery by serum RT-PCR and initiated breastfeeding on day 3. On day 8, the ZIKV RT-PCR results from mother 2 were serum negative, urine positive, and breast milk positive, however ZIKV breast milk culture was negative. Newborn 2 tested negative for ZIKV on day 0 and day 3 by serum RT-PCR, but had confirmed ZIKV infection on days 4 and 7 by serum RT-PCR and on day 8 by urine RT-PCR. On day 9, newborn 2 urine was ZIKV negative by RT-PCR. These case reports confirmed ZIKV infection in 2 breastfeeding mothers and their newborns as well as detected ZIKV in serum and breast milk of both mothers. Both mothers had clinical signs of rash within days of delivery, and the authors hypothesized that the infants were probably infected in utero or intrapartum because the infants' sera were positive for the presence of Zika virus within one day of starting breastfeeding. The author of this study was contacted (M. Besnard, personal communication, 2016) and confirmed that no long-term complications were reported for either of the two infants at 2 years of age.
A second case report described a mother (referred to as case 3 in Table 1) from New Caledonia (French territory) who initiated breastfeeding on the day of delivery and developed fever and maculopapular rash in the following days [47]. On day 3 post delivery, mother 3 tested positive for ZIKV infection by serum RT-PCR, however the serum RT-PCR results for newborn 3 were reported as ambiguous. Breast milk was ZIKV positive by RT-PCR on day 4 and ZIKV breast milk culture was also positive. While vertical transmission was not described in this case, the presence of ZIKV in breast milk was confirmed. No long-term complications were reported for the child at 8 months of age (M. Dupont-Rouzeyrol, personal communication, 2016). The overall quality of the evidence was very low for all the proposed outcomes, as described in the GRADE Summary of Findings (Table 2).

Discussion
The cases presented in these two reports confirm the presence of ZIKV RNA in breast milk from three ZIKV-infected mothers. The presence of Zika-specific antibodies was not reported in these cases. Of the three newborns delivered to ZIKV-infected mothers who were receiving breast milk with confirmed presence of ZIKV, only two were confirmed to be infected with ZIKV with no reported adverse outcomes. With regard to the presence of ZIKV in breastfeeding-related fluids, ZIKV was detected by RT-PCR in breast milk and blood of the three mothers; sweat was not measured.
Like other viral infections, mother-to-child transmission of ZIKV infection can potentially occur during antepartum, intrapartum, or postnatal periods.
[48] Given the variable incubation period for ZIKV, it can be difficult to distinguish breastfeeding transmission from other perinatal routes. For the two newborns who contracted ZIKV from ZIKV-infected mothers expressing ZIKV-infected breast milk, antepartum or intrapartum transmission is suspected. Even if a newborn is ZIKV negative following delivery from a ZIKV-infected mother and contracts ZIKV infection while consuming breast milk with ZIKV, there remains a possibility for separate mosquito transmission. Identifying the time of infection and duration of an incubation period is further complicated by the asymptomatic nature of acute ZIKV infection.
There is limited evidence describing breastfeeding transmission for other flavivirus infections. West Nile virus (WNV), dengue virus, and yellow fever virus have been detected in breast milk [49,50]. Of these infections, WNV has been associated with breastfeeding transmission in a small number of cases [24]. Like ZIKV, breastfeeding transmission for other flavivirus infections is likely underreported due to asymptomatic illness and limited access to diagnostics. We intend to review breastfeeding transmission for related flavivirus infections in the near future.
Our systematic review for ZIKV breastfeeding transmission resulted in two studies and three cases of lactating women with confirmed ZIKV infection. As new data emerges from these current outbreaks, further investigation is needed to explore ZIKV breastfeeding transmission dynamics. This includes understanding the mechanics of transmission with regards to timing of infection for mother and infant, breast milk viral load, and exposure duration as well as assessing the frequency and distribution of breastfeeding transmission among affected populations. In addition to determining viral transmission risk, research should also explore the protective properties of ZIKV-specific immunoglobulin in breast milk transferred from mothers who have experienced ZIKV infection. At this time, the data are not sufficient to conclude ZIKV transmission via breastfeeding, and the authors support the WHO breastfeeding  CI: Confidence interval; RT-PCR: reverse transcription polymerase chain reaction *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effectVery low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect 1 The evidence is based on three mother-infant pairs from two case-reports. A case report of two mother- guidelines currently in place recommending initiating breastfeeding within one hour of delivery, exclusively for 6 months and extended until 2 years or beyond [51].