Zika virus infection in pregnancy: a systematic review of disease course and complications

Objectives To characterize maternal Zika virus (ZIKV) infection and complement the evidence base for the WHO interim guidance on pregnancy management in the context of ZIKV infection. Methods We searched the relevant database from inception until March 2016. Two review authors independently screened and assessed full texts of eligible reports and extracted data from relevant studies. The quality of studies was assessed using the Newcastle-Ottawa Scale (NOS) and the National Institute of Health (NIH) tool for observational studies and case series/reports, respectively. Results Among 142 eligible full-text articles, 18 met the inclusion criteria (13 case series/reports and five cohort studies). Common symptoms among pregnant women with suspected/confirmed ZIKV infection were fever, rash, and arthralgia. One case of Guillain-Barré syndrome was reported among ZIKV-infected mothers, no other case of severe maternal morbidity or mortality reported. Complications reported in association with maternal ZIKV infection included a broad range of fetal and newborn neurological and ocular abnormalities; fetal growth restriction, stillbirth, and perinatal death. Microcephaly was the primary neurological complication reported in eight studies, with an incidence of about 1% among newborns of ZIKV infected women in one study. Conclusion Given the extensive and variable fetal and newborn presentations/complications associated with prenatal ZIKV infection, and the dearth of information provided, knowledge gaps are evident. Further research and comprehensive reporting may provide a better understanding of ZIKV infection in pregnancy and attendant maternal/fetal complications. This knowledge could inform the creation of effective and evidence-based strategies, guidelines and recommendations aimed at the management of maternal ZIKV infection. Adherence to current best practice guidelines for prenatal care among health providers is encouraged, in the context of maternal ZIKV infection. Electronic supplementary material The online version of this article (doi:10.1186/s12978-017-0285-6) contains supplementary material, which is available to authorized users.

Results: Among 142 eligible full-text articles, 18 met the inclusion criteria (13 case series/reports and five cohort studies). Common symptoms among pregnant women with suspected/confirmed ZIKV infection were fever, rash, and arthralgia. One case of Guillain-Barré syndrome was reported among ZIKV-infected mothers, no other case of severe maternal morbidity or mortality reported. Complications reported in association with maternal ZIKV infection included a broad range of fetal and newborn neurological and ocular abnormalities; fetal growth restriction, stillbirth, and perinatal death. Microcephaly was the primary neurological complication reported in eight studies, with an incidence of about 1% among newborns of ZIKV infected women in one study. Conclusion: Given the extensive and variable fetal and newborn presentations/complications associated with prenatal ZIKV infection, and the dearth of information provided, knowledge gaps are evident. Further research and comprehensive reporting may provide a better understanding of ZIKV infection in pregnancy and attendant maternal/fetal complications. This knowledge could inform the creation of effective and evidence-based strategies, guidelines and recommendations aimed at the management of maternal ZIKV infection. Adherence to current best practice guidelines for prenatal care among health providers is encouraged, in the context of maternal ZIKV infection.

Plain english summary
Aedes mosquitoes transmit Zika virus (ZIKV) infection, its clinical presentation in humans is often mild or asymptomatic. Due to a marked increase in the number of symptomatic or suspected cases across continents, ZIKV infection was declared a Public Health Emergency of International Concern (PHEIC) on February 1, 2016.
Pregnant women are at an exceptional risk of being affected with potential adverse effects. To describe maternal Zika virus (ZIKV) infection and complement the evidence base for the WHO interim guidance on pregnancy management and in the context of ZIKV infection, we did a systematic review.
We conducted database searches, independent screening of records and assessment of resulting full texts from which we extracted data. We further assessed the quality of the included studies.
Of the 142 eligible full-text articles, we extracted data from the 18 studies which met the inclusion criteria (13 case series/reports and five cohort studies). Among the pregnant women with suspected/confirmed ZIKV infection in these studies, fever, rash and joint-pain were the commonly reported symptoms. Neurological affectation (Guillain-Barre syndrome) was reported in one mother, no other cases of severe maternal illness or deaths were reported. Complications reported in association with maternal ZIKV infection included a broad range of fetal and newborn neurological and ocular abnormalities; fetal growth restriction, stillbirth, and perinatal death. Microcephaly was the primary neurological complication reported in eight studies, with an incidence of about 1% in one study.
In conclusion, knowledge gaps on the features characterising maternal ZIKV infection and its effects are evident. Further research and better reporting practices may help create effective strategies aimed at the management of maternal ZIKV infection and subsequent fetal complications. Pregnant women and health providers are encouraged to adhere to current best practice guidelines, in the context of maternal ZIKV infection.

Background
Since the late 1940s, there have been reports of Zika virus (ZIKV) isolated from rhesus monkeys in Uganda, and human cases in Tanzania and Nigeria [1,2]. At present, ZIKV has a broad distribution across sub-Saharan Africa, South-East Asia, and the Americas [3]. ZIKV belongs to the Flaviviridae family of viruses and is transmitted by Aedes mosquitoes.
In humans, ZIKV infection is often mild or asymptomatic. However, an exponential increase in the number of symptomatic or suspected cases across continents has intensified international concern [4]. Prior to 2007, few cases of ZIKV infection were reported [5]. Thereafter, an infectious outbreak affecting 74.6% of the Micronesian Yap population [6] and 32,000 people in the French Polynesian region in 2013-2014 were reported [7,8]. In the latest outbreak in 2015, up to 1.3 million suspected cases were identified in Brazil over a 9-month period [9,10]. In addition to transmission by Aedes mosquitoes, sexual transmission has also been reported [11,12].
Despite the international efforts to curb its spread, ZIKV infection has expanded to 62 countries and territories, and is projected to increase due to potential climate change affecting the spread of the ZIKV Aedes vector [7,[13][14][15]. Notably in 2015, a 20-fold increase in the incidence of microcephaly relative to previous years was reported with the onset of ZIKV transmission in northeastern Brazil. This trend led to the World Health Organization (WHO) declaring a Public Health Emergency of International Concern (PHEIC) on February 1, 2016 [13,16,17].
Pregnant women are at exceptional risk of being infected with ZIKV infection and exhibiting potential adverse effects, including a wide range of congenital abnormalities associated with ZIKV infection in utero. Delayed or arrested neurological development and complications common in fetuses and neonates born to ZIKV-infected women, may drain family resources due to the need for special care. Perinatal deaths may further impose an emotional burden on affected families. The health systems in most ZIKV-affected areas have limited resources to manage the current outbreak and its consequences.
Many gaps in the knowledge regarding ZIKV infection exist, these include, the clinical spectrum of disease presentation in infected pregnant women, mechanisms of vertical transmission, and the risk of complications in mother, fetus and newborn. Also, the impact of ZIKV co-infection with other flaviviruses on pregnancy as well as the long-term consequence of ZIKV infection among infants and adults of reproductive age is unclear. As part of the process to complement the evidence base for WHO interim guidance on pregnancy care in the context of ZIKV infection and close the knowledge gaps, we conducted a systematic review.

Search strategy
We searched the following electronic databases: MED-LINE, EMBASE, CINAHL, World Health Organization Global Health Library (WHOGL), and Cochrane Central Register of Controlled Trials (CENTRAL) on March 3, 2016. An information specialist configured search terms for each database according to the level of term indexation. Search terms included flavivirus, chikungunya virus, fever, arbovirus, Zika, congenital malformations, cephalometry, and nervous system abnormalities as shown in Additional file 1). There were no dates, language or study design restrictions. We registered this review in PROSPERO, the international prospective register of systematic reviews of the University of York and the National Institute for Health Research, under the number CRD4201603969.

Selection of studies
We considered all study designs including randomized controlled trials, prospective or retrospective cohorts, cross-sectional, case-control studies and case series/reports. Original reports, briefs, letters, editorials, correspondence and news reports were considered for inclusion. A minimum of two review authors assessed the title and abstract of references for relevance to review objectives. We evaluated the full text of potentially eligible records for inclusion relative to our review objectives. The review objectives included primary data on disease characteristics in ZIKV-infected pregnant women and their babies and the risk of adverse pregnancy outcomes. References of included reports were also searched for potentially eligible studies. Disagreements on the eligibility of reports, where present, were resolved by consensus with a third review author.

Types of studies
We included all studies reporting on ZIKV infection and complications in pregnant women residing in or with a history of travel to areas of ongoing ZIKV transmission or recent ZIKV outbreaks. Studies comparing pregnant and non-pregnant women with ZIKV infection were considered for inclusion where available. We excluded any study or report on non-pregnant populations or lacking primary data.

Types of participants
Pregnant women suspected of being at risk of, or diagnosed with ZIKV infection regardless of their location, or who have had sexual contact with a partner who (i) recently returned from an area of active ZIKV transmission, or (ii) was diagnosed with ZIKV infection.

Data extraction and synthesis
Two review authors independently extracted and presented data based on the review objectives in an agreed data collection form. We summarized the data in a descriptive form (Table 1).
We extracted information on study design, sample size, history of maternal and fetal ZIKV infection, symptoms, presence and type of clinical features, pregnancy-specific complications during pregnancy, childbirth or postpartum period. The absolute risk of microcephaly and other birth defects was also recorded where available.

Risk of bias assessment
Two review authors conducted independent assessments on the quality of included studies. Where discordance in ranking was observed, this was resolved by consensus or discussion with a third review author. For individual case series/reports, a quality assessment of the National Institute of Health Tool (NIH) [18] was applied. This tool includes questions based on nine criteria to which either of the binary answers (Yes/No) was allotted. Based on the number of 'Yes' answers, a rating of good (7 -9), fair (4 -6) or poor (≤3) was allocated to the individual study and differences in quality ratings amended by consensus. Studies for which the criteria were irrelevant were labelled 'not applicable' and 'cannot determine' if such information was lacking in the study.
For observational studies, we applied the Newcastle-Ottawa Scale [18] which consisted of lead questions aimed at assessing three domainsselection, comparability, and outcome. For each outcome, a maximum of three points could be allotted as previously described [20].

Search results
Search strategies identified 2316 records as shown in the PRISMA flow chart in Fig. 1. Based on the title and abstract screening, 2174 records were excluded due to lack of primary data and/or irrelevance to review objective. Of the resulting 142 eligible full texts, 18 studies met our inclusion criteria (reasons for exclusion are outlined in Additional file 2).
In some studies, ocular examinations were also conducted for the mother [34] and newborn [22,34].

Characteristics of maternal ZIKV infection Symptoms/signs/complications
No study suggested a higher risk of acquiring ZIKV infection in pregnant women compared to the non-pregnant population, after exposure to Aedes mosquitoes. Most  [21][22][23][24][25][26][27][28][29][32][33][34]. Two of the remaining three studies (one observational study, one case series) did not report clinical manifestations during pregnancy as they focused on the newborns of ZIKV-infected mothers [30,31]. Rash was absent in the third study as the mothers were asymptomatic [35]. In seven reports, the rash was further qualified as 'a pruritic cutaneous rash with itching in the back and hands', 'generalised, descending macular' or 'generalised maculopapular' [21,28,33], 'petechial' or 'generalised' [21,28,32,33,36]. Only one case-series reported on symptom occurrence after delivery in two ZIKV-infected pregnant women [21]. A 4-day duration of persistent rash in one pregnant woman and a pruritic rash on the third day after childbirth in the second case was observed. In the former, the rash started 2 days before and ended 2 days after delivery of a healthy infant. In the latter case, a mild fever and myalgia were also reported.
An observational study reported a significantly higher frequency of lymphadenopathy and conjunctival injections in ZIKV-infected compared with uninfected pregnant women [33].
Neurological complications were reported in ZIKVinfected pregnant women in one case report [23] and one observational study [33]. Guillain-Barré syndrome (GBS) at 28 weeks gestation in addition to other typical ZIKV signs and symptoms were reported in one case report only [23]. The patient had respiratory distress that prompted intensive care unit (ICU) admission. She fully recovered within 3 weeks and gave birth to a normal infant at 39 weeks of gestation.
Preterm birth necessitating emergency caesarean delivery was reported in two observational studies [27,33] and one case series [21]. In the two observational studies, intrauterine growth restriction (IUGR) was an underlying complication. In one of the studies [33], the two preterm births reported were due to IUGR accompanied by oligohydramnios, macular hypoplasia, and placental insufficiency in one fetus and anhydramnios in the other.
Miscarriages were reported in five ZIKV-infected pregnant women in two case series [27] and one observational study [33]. Stillbirths were also recorded in three pregnant women in one observational study and one case report [33,35].
In fetuses of ZIKV-infected pregnant women, placental and umbilical [32,33] and amniotic fluid [33] abnormalities were reported. IUGR was reported in five studies (two observational, three case-series/reports) [21,33] and stillbirths in two studies [33,35]. Despite normal fetal biometrics at 14 weeks in one case of an asymptomatic ZIKV-infected pregnant woman, IUGR was detected at 18 weeks and accompanied by other neurological abnormalities including microcephaly, hydranencephaly, hydrops fetalis and eventual stillbirth [35].
Fetal deaths were reported in two studies (one case series and one observational study), these deaths occurred in four fetuses at > 30 weeks gestation [27,33]. Elective terminations of pregnancies were conducted for seven pregnant women. However, the gestational age or trimester of termination was unknown [31].

Newborns
Symptoms/signs Rash described as 'transient isolated diffuse' was reported in an infant on the fourth day postdelivery [21]. Other symptoms reported included conjunctivitis or conjunctival injections [24,28].
Four studies reported the birth of healthy infants [23,33,37] and one study reported early neonatal deaths in three infants with microcephaly within 20 h of delivery [27].

Impact of ZIKV co-infection in pregnancy
There was a lack of information on co-infection with other flaviviruses and common congenital infections in some studies. Serological tests were conducted in studies to exclude possible co-infections and/or antibody tests for previous exposure to other related flaviviruses.
Four reports assessed previous exposure to other infections based on the presence of IgG [24,28,33,35] and IgM [28,35] antibody positivity to cytomegalovirus, rubella and toxoplasmosis [24,33,35] infections. IgG positivity to infections indicating exposure to dengue virus in 88% of 88 pregnant women was reported in one observational study [33], and toxoplasmosis and rubella in three of 28 women [24] in two case series/report.

Risk of bias assessment
We judged the overall risk of bias as fair or good for all the included case series/reports except one, which we rated as poor [25] due to selective reporting ( Table 2).
Four of five observational studies included in the review were assigned an average quality rating of four to five stars based on the NOS. One observational study [22] was assigned a very low quality rating of one star as it provided sparse information (Table 3).

Discussion
Our systematic review describes the complications to fetuses and newborns have been reported during ZIKV infection in pregnant women. However, no study clearly described the history of maternal ZIKV infection and its effects on newborns of affected mothers. A lack of information on factors related to ZIKV infection in pregnancy further limits our understanding. Most studies were conducted in 2016, these studies focused on the presenting symptoms and associated complications in pregnant women and their fetuses/newborns. Relative to non-pregnant women or the general population [12,39], symptoms and complications of ZIKV infection were comparable.
Maternal ZIKV infection tended to have more adverse effects on the fetus and the infant than on the mother, as most maternal symptoms were self-limiting. A broad spectrum of clinical presentation was observed in fetuses and newborns of ZIKV-infected mothers, ranging from normal to abnormal ultrasound findings during pregnancy, such as healthy newborns or newborns with an abnormality. A similar pattern was observed in ZIKVinfected pregnant women, ranging from symptomatic and asymptomatic clinical presentation to uncomplicated deliveries in most studies. Severe pregnancyspecific complications were uncommon and no maternal deaths were recorded. This highlights a challenge for both health providers and mothers at risk because of the increased likelihood of missed opportunities due to the frequent asymptomatic presentation.
The diagnostic accuracy of tests employed in the ZIKV context is still unclear [40] because existing studies are based on epidemiological models. In some women, ZIKV positivity was confirmed later in pregnancy, although the standard RT-PCR tests done in early trimesters were negative. Fetal ultrasound tests that were negative in the first and second trimesters were positive in late pregnancy stage [35]. The absolute risk of developing fetal abnormalities is also unclear, but microcephaly is unlikely being a rare condition.
ZIKV shares the same vector with other arboviruses including chikungunya and dengue. TORCH (Toxoplasmosis, Other (syphilis, varicella-zoster, parvovirus B19), Rubella, Cytomegalovirus (CMV), and Herpes) infections have been associated with congenital malformations including central nervous system anomalies. In the included reports, the absence of co-infection with or previous exposure to other viruses was established in most of the pregnant women. In some studies, IgG-positivity to dengue and toxoplasma were observed. Immunity to rubella and CMV were reported in some women with poor outcomes. Potential synergy due to the presence of immunity and/or seropositivity to other viruses could not be ascertained as clinical presentation varied from the absence of symptoms to the presence of typical symptoms seen in ZIKV-infected pregnant women. In one study, about half of the pregnant women who No description of the derivation of the cohort Selection of non-exposed cohort No non-exposed cohort No non-exposed cohort Only one cohort of exposed individuals Only one cohort of exposed individuals Only one cohort of exposed individuals Drawn from the same community as the exposed cohort a Drawn from a different source No description of the derivation of the non-exposed cohort Ascertainment of exposure Only one cohort of exposed individuals Only one cohort of exposed individuals Only one cohort of exposed individuals Only one cohort of exposed individuals Only one cohort of exposed individuals  presented with rash had lymphadenopathy [33], a condition commonly observed in dengue infection due to lymphatic infiltration [41]. Although inconclusive, the history of lymphadenopathy in some pregnant women in this study in addition to the presentation of rash may show a possible role for dengue in extending the spectrum of maternal or fetal presentation. Indirect evidence from the history of viral infections that share similar characteristics and complications with ZIKV, such as CMV, may be important in understanding ZIKV infection.
In some studies, sample selection was based on the presenting clinical features. Limitation of the study population to only pregnant women presenting with features suggestive of ZIKV infection such as rash and fetuses with microcephaly as seen in some reports could be misleading. The rigorous search strategy, the inclusion of foreign language studies and exclusion of overlapping studies add to the strength of this review.
The broad spectrum of clinical presentation and complications in fetuses and newborns of ZIKV-infected pregnant women provides an insight into the potential liability for affected women, their families and health systems in resource-constrained settings. The possibility of an asymptomatic presentation proposes an emphasis on routine antenatal care and attention to signs of fetal brain abnormalities by health providers in pregnant women with molecular or epidemiological links to ZIKV infection.
Ongoing research on live animal models may help to understand the pathogenesis of ZIKV, with a potential for application to exposed pregnant cohorts. The rapid rate of emerging data from research on Zika virus infections in pregnancy and recent years proposes a need for frequent updates. As newer studies become available, further research on adverse pregnancy outcomes and implications on long-term consequences of ZIKV during pregnancy will further enable early diagnosis and better management modalities in provider care. Furthermore, personal protective measures should be encouraged.

Conclusions
This review highlights key evidence gaps that need to be urgently prioritized by the international community, more so, with the wide variations in fetal and newborn presentations/complications associated with prenatal ZIKV infection. Further research and comprehensive reporting of maternal ZIKV infection and fetal/neonatal complications may provide a better understanding of ZIKV infection in pregnancy and its attendant maternal/ fetal complications. Such knowledge could inform the creation of effective and evidence-based strategies, guidelines, recommendations and health policies aimed at the management of maternal ZIKV infection.
Adherence to current best practices guidelines for prenatal care among health providers is encouraged. These include appropriate notification of suspected cases to the responsible authorities, long-term evaluation, monitoring and follow-up for newborns and infants exposed to ZIKV infection by a multidisciplinary team of health providers, more so, with ongoing variations in presentation and distribution of maternal ZIKV infection.