Congenital cerebral malformations and dysfunction in fetuses and newborns following the 2013 to 2014 Zika virus epidemic in French Polynesia

M Besnard 1 , D Eyrolle-Guignot 2 , P Guillemette-Artur 3 , S Lastère 4 , F Bost-Bezeaud 5 , L Marcelis 6 , V Abadie 7 , C Garel 8 , M Moutard 9 , J Jouannic 10 , F Rozenberg 11 12 , I Leparc-Goffart 12 13 , H Mallet 6 1. Service de réanimation néonatale, Centre Hospitalier de Polynésie Française (CHPF), Pirae, Polynésie Française 2. Service d’obstétrique, CHPF, Pirae, Polynésie Française 3. Service de radiologie, CHPF, Pirae, Polynésie Française 4. Laboratoire de biologie médicale, CHPF, Pirae, Polynésie Française 5. Laboratoire d’anatomie pathologique, CHPF, Pirae, Polynésie Française 6. Bureau de Veille Sanitaire, Direction de la Santé, Papeete, Polynésie Française 7. Service de pédiatrie générale, Assistance Publique-Hôpitaux de Paris, (AP-HP), Hôpital Necker, Paris, France 8. Service de radiologie, AP-HP, Hôpital Armand-Trousseau, Paris, France 9. Service de neuro-pédiatrie, AP-HP, Hôpital Armand-Trousseau, Paris, France 10. Service de Médecine Fœtale, AP-HP, Hôpital Armand-Trousseau, Paris, France 11. Service de virologie, AP-HP, Hôpital Cochin, Paris, France 12. These authors contributed equally to the manuscript 13. Centre National de Référence des arbovirus, IRBA, Marseille, France


Introduction
A Zika virus (ZIKV) outbreak in French Polynesia from October 2013 to March 2014 resulted in 8,750 suspected cases reported through the general practitioners-based (25 to 45 sentinel practitioners) sentinel surveillance system for infectious diseases.The system exists since 2009 and syndromic cases definitions are basis for weekly reporting.An estimated 32,000 suspected cases sought medical care and more than half of the population might have been infected [1,2].ZIKV is an emerging arbovirus that before 2013 was considered to cause only mild disease, characterised by fever, rash, joint pain, and conjunctivitis.
From 2014 to 2015, following the ZIKV epidemic, we observed an unusual increase in annual congenital cerebral malformations (two-fold), brainstem dysfunction (31-fold), and severe microcephaly (14-fold) among fetuses and newborns (data not shown).Following the announcement by the Brazilian Government in November 2015 of a dramatic increase in the incidence of microcephaly possibly associated with an ongoing ZIKV outbreak [3,4], we notified the World Health Organization (WHO) of this unusual cluster of congenital neurological abnormalities in our islands [5].
French Polynesia consists of five archipelagos in the South Pacific Ocean, with 118 islands, of which 76 are inhabited.The total population was 271,800 in 2014 and 70% reside on Tahiti.Approximately 4,200 births are recorded per year (2014: 4,161 births), mainly in Tahiti, with 60% at the Centre Hospitalier de Polynésie Française (CHPF), 30% in two private clinics in Tahiti, and 10% in two distant district hospitals.
We report here a retrospective case series of 19 fetal and newborn cases with congenital cerebral malformations and dysfunction and detail the neurological lesions identified and the corresponding virological results.

Methods
We conducted a retrospective review of congenital cerebral malformations and dysfunction, detected in a prenatal and neonatal population in French Polynesia from March 2014 to May 2015.We included pregnancies beginning between June 2013 and August 2014, the period which corresponded to the largest estimated circulation period of ZIKV.There is no territorial register of congenital malformations, thus the prenatal cases were collected through the non-computerised charts of the prenatal diagnostic unit of the CHPF and the neonatal cases were recorded through the hospital Programme de Médicalisation des Systèmes d'Information (PMSI).

Prenatal monitoring
According to the recommendations of the Haute Autorité de Santé, France [6], three prenatal ultrasounds (US) scans, trisomy screening and serological assays for, hepatitis B, HIV, rubella, toxoplasmosis and syphilis are performed for all pregnant mothers.When fetal anomalies are detected by routine antenatal US scan, the pregnant woman is referred to the CHPF prenatal diagnostic unit, where a second US evaluation is conducted by an expert obstetrician and a nurse experienced in prenatal diagnosis.Once cerebral congenital malformations, for example, are confirmed, they perform an amniocentesis and prescribe magnetic resonance imaging (MRI) of the fetal brain.Then, as all cases of severe congenital abnormalities, they are reviewed by the Prenatal Multidisciplinary Diagnostic Centre of East Paris, France.According to the French law, the committee determines if termination of pregnancy (TOP) can be performed, until full gestational term, upon parental request.An average of 20 annual TOPs are performed (range 12-26) for various severe fetal malformations that constitute a substantial risk of serious motor or cognitive disabilities.After the TOP procedure, pictures of the fetus are taken and skeletal X-rays are performed.Weight, height and head circumference are measured.No autopsy is done as there is no fetopathologist in French Polynesia.Except for  formalin-fixed, paraffin-embedded placenta samples, no fetal tissue is stored.

Postnatal diagnosis
There is only one neonatal intensive care unit (ICU) in French Polynesia.It is situated at CHPF and takes care of all severe neurological malformations and dysfunction.We collected all cases by reviewing the coding of hospitalised patients [7].When congenital cerebral malformations are detected in utero, systematic control after birth of brain imaging (US, MRI or CT-scan) are performed to confirm and assess the brain lesions.

Case definition
We included cases with congenital microcephaly, defined by a head circumference below the third percentile for gestational age and sex, according to the Association des Utilisateurs de Dossiers Informatisés en Pédiatrie, Obstétrique et Gynécologie (AUDIPOG) charts [8], cases with brain lesions without microcephaly and cases with congenital brainstem dysfunction characterised by a deficiency of coordination between sucking, swallowing and breathing, with no visible cerebral malformation and a normal birth head circumference.We assigned them to three groups: Group 1: fetuses (Group 1a) and newborns (Group 1b) with severe cerebral lesions and microcephaly; Group 2: fetuses with severe cerebral lesions without microcephaly; Group 3: newborns with congenital brainstem dysfunction.
Cases with proven aetiology and usual and isolated neurological anomalies without brain damage were excluded.
We retrospectively reviewed the medical records of all mothers and fetuses or newborns with their clinical, serological and radiological (US scans and MRI) data.The imaging examinations were reviewed by radiologists of CHPF and of Hospital Armand-Trousseau, Paris.
We contacted mothers of cases identified and asked about symptoms of ZIKV infection during their pregnancy.Informed written consent was obtained from mothers for all investigations regarding available samples, and publication.Our study was approved by the Ethics Committee of French Polynesia.

Laboratory investigations
Amniotic fluids were analysed for karyotype (comparative genomic hybridisation (CGH) array was not routinely performed) and PCR for cytomegalovirus (CMV).Between June 2013 and August 2014, no virological testing for ZIKV was performed in amniotic fluids and pregnant women.

Results
There were 4,787 births during the study period March 2014 to May 2015, and we observed 33 cases with congenital brain malformations or dysfunction.We excluded four cases with proven aetiology (one toxoplasmosis, two chromosomal abnormalities, one myopathy) and ten with usual and isolated neurological anomalies without brain damage (three intraventricular haemorrhages, three corpus callosum agenesis, one neural tube defect and three polymalformative syndromes during the first trimester of pregnancy).The remaining unusual 19 congenital cerebral malformations and dysfunction were included in our study: eight in Group 1 i.e. five fetuses in Group 1a and three newborns in Group 1b; six fetuses in Group 2 and five newborns in Group 3 (Table 1).
The mean age of the mothers in Group 1 was 29.7 years (range 22.8-38.9),26.9 years in Group 2 (range 15.7-38.0),and 30.8 years in Group 3 (range 20.2-39.0).There was no relevant medical or family genetic history in any of the mothers.Alcohol or maternal cocaine use during pregnancy was denied in our series.There was no seroconversion for CMV, hepatitis B, HIV, rubella, syphilis and toxoplasmosis in any of the mothers; for 7 of 13 samples of amniotic fluid, PCR for EV, rubella, LCMV, HSV, VZV and DENV was negative.
Imaging findings for all cases are summarised in Table 2. Skeletal X-rays were done for all fetuses and revealed an intracranial calcification in one case (Case 5), which was also observed on prenatal US scan.
Below we describe in detail the findings in the different Groups.

Group 1a
The five fetuses in this group exhibited severe signs of brain injury associated with microcephaly, visualised within the second trimester of pregnancy (after 20 weeks of gestation (WG)) by US scan and confirmed by MRI in four cases.In all cases, in addition to microcephaly, imaging findings (fetal US or MRI) showed multiple severe cerebral injuries including profound destruction of median structures and evidence of interruption of brain development (Figures 1, 2, 3) including absence (or rupture) of the corpus callosum (n = 3) and of cavum septi pellucidi (n = 3), ventriculomegaly superior or equal to 12 mm (norm: <10mm, mild ventriculomegaly 10-12mm) (n = 3), occipital subependymal pseudocysts (n = 2), opercular dysplasia (agyria, Information about ZIKV infection symptoms in mothers was obtained retrospectively in December 2015. polymicrogyria) (n = 4), vermian dysgenesis (n = 3), enlarged pericerebral space (n = 4) and parenchymal calcifications (n = 5).In each of the five cases at least five of these lesions were present and all pregnancies were terminated.Extracranial findings were micropenis (Case 1) and hypotrophy below the third percentile with intestinal hyperechogenicity (Case 5).At delivery, head circumferences were below the third percentile, but not measured on two fetuses.Placental microcalcifications were observed in three cases either on US scan and histology (Case 3) or histologically only (Cases 4 and 5).
Retrospectively, ZIKV RNA was detected by RT-PCR and infectious ZIKV was isolated in four of five amniotic fluid samples (Cases 2, 3, 4, 5).When interviewed, the mothers of Cases 2, 4 and 5 reported clinical infection in the first trimester of pregnancy, the mother of Case 3 could not be reached.The chronological sequence of maternal symptoms, fetal malformations discovery and PCR results are summarised in Figure 4.

Group 1b
Microcephaly was diagnosed in utero for two of the three cases (Case 7 and 8) but mothers did not wish
Viral tests were done on two amniotic fluids available (cases 12 and 13) and PCR was negative for ZIKV.Two mothers (Cases 11 and 14) reported symptoms of ZIKV infection in the first trimester (Figure 4).

Group 3
The five newborns in this group had clinical evidence of brainstem dysfunction, manifest by absence of sucking and swallowing, but without microcephaly or severe cerebral radiological anomalies.Prenatal US scan revealed in three cases polyhydramnios as consequence of early swallowing deficiency ( Cases 17,18,19).Head circumference was normal at birth for all cases.They were admitted to the neonatal ICU, due to severe feeding disorders, need for frequent aspiration and cardiac dysautonomia.Two cases (Cases 15 and 17) had an associated Pierre Robin Sequence (PRS) with retrognathia, glossoptosis and posterior U-shaped No amniotic fluid was available for any of the five cases.Pregnancies were uneventful, besides urinary tract infection for one mother (Case 16).Mothers reported no ZIKV infection symptoms during pregnancy.In-house ELISA flavivirus serology revealed IgG in four of the five mothers.Two of them had neutralising antibodies for dengue serotypes 1-4, but no ZIKV neutralising antibodies.In the two others, no neutralising antibodies were detected for ZIKV, DEN, West Nile and Japanese encephalitis virus.

Discussion
Congenital cerebral malformations are rather rare in French Polynesia and are not collected in a register of congenital malformations.After 2011, when the prenatal diagnosis unit records started, about 15 congenital cerebral malformations were observed in 2012 and 2013, respectively and according to PMSI data, between 2001 and 2013, only two cases of brainstem dysfunction were noted in 2009, leading to an average annual incidence rate of 0.34 cases per 10,000 births.Seven cases of congenital microcephaly were reported in the same period, equivalent to an average annual incidence rate of 1.2 cases per 10,000 live births.The present cluster of severe cerebral malformations with a 14-fold increase in congenital microcephaly and 31-fold in brainstem dysfunction was spatially and temporally associated with a large outbreak of ZIKV in French Polynesia.
In Brazil, the detection of ZIKV genome in amniotic fluid [9][10][11] and in fetal brains of children with microcephaly [10,12,13], concurrent with widespread local ZIKV transmission [3,4,9] strengthened the hypothesis of teratogenicity of ZIKV.In our series, four fetuses who were infected in utero with ZIKV had typical symptoms of viral fetopathy, with microcephaly, severe brain lesions, intrauterine growth retardation and placental calcifications.Infectious virus was isolated from the amniotic fluid of four microcephaly cases, and clinical symptoms of ZIKV infection in the first trimester of pregnancy were reported by three of their mothers.However, ZIKV was not detected in the available amniotic fluids of three other cases tested (one microcephalic, two non-microcephalic), and their mothers were asymptomatic for ZIKV infection, as 80% of the population infected with ZIKV [2].This finding raises the question about the possible correlation between maternal symptoms, probably associated with high viraemia, and the risk of transplacental transmission.
In the three microcephalic newborns in Group 1b, no further malformations were associated, besides micropenis.Another case of micropenis was found in Group 1a; a possible explanation for this finding could be central hormonal deficiency.Although there was no laboratory evidence for ZIKV infection in all three cases in Group 1b and no ZIKV infection symptoms reported in two of three mothers, the severity of the microcephaly and of the radiologic brain lesions similar to those of group 1a, supports the hypotheses of probable ZIKV infection in the context of the ZIKV outbreak.
Of six cases with cerebral congenital malformations without microcephaly in Group 2, two mothers reported ZIKV infection symptoms in pregnancy.Unlike the other mothers of the same group, none of these mothers had fetuses with extracerebral malformation, besides fetal akinesia which could have been due to severe brain damages or peripheral neuropathy.One case with hydrops and brain lesions had not been explored for ZIKV infection.In Brazil, a fetus with hydrops, hydranencephaly and cranial calcifications was associated with a positive RT-PCR of ZIKV in brain and amniotic fluid [13].
The wide spectrum of clinical and neuroradiological findings in our microcephalic and non-microcephalic foetuses and newborns series suspected to be associated with ZIKV infection, may reflect the timing and severity of fetal viral infection.Early viral infection within the first trimester of pregnancy may interfere with neurocorticogenesis by impairing neuronal proliferation and migration between 12 and 24 WG, which leads to severe microcephaly [14] and neuronal cell migration disorders, whereas second semester infection may cause brainstem dysfunction without obvious lesions and intellectual disabilities.Microcephaly has been diagnosed as early as 21 WG in one fetus.By 25 to 30 WG, cortical gyrification can be analysed by MRI and in microcephalic fetuses, migration disorders were constantly present.The first signs of cerebral damage can be identified by the second trimester of pregnancy by US scan [15] when it reveals ventriculomegaly, destruction of the corpus callosum and/or of the cavum septi pellucidi and cerebral cysts.Occipital subependymal pseudocysts, which seemed to be another frequent hallmark visible on US scans in four of seven of our explored microcephalic cases, may suggest the consequences of the destruction of the germinative zone by the ZIKV, similar to what was observed in CMV infection [16].
Congenital brainstem dysfunction includes several symptoms such as sucking deficiency, aspiration, upper airways obstruction, and a possible PRS with a posterior and vertical position of the tongue [17].It may be isolated or syndromic (CHARGE syndrome, etc.) [18].
Besides genetic causes, it could be due to an injury of the brainstem during embryonic or fetal life, such as an ischaemic stroke, misoprostol teratogenicity, maternal cocaine use or later viral infection [19] of the brainstem.The latter can be suspected for two of our cases with microcalcifications in the pons region.There was no evidence of ZIKV congenital infection, nor other proven aetiology, in any of our patients with congenital brainstem dysfunction (Group 3), and their mothers did not report any viral symptoms consistent with ZIKV infection during pregnancy.
The main limitations of our study are those of a retrospective review with bias of collecting data and loss of biological fluids which are stored for one year only.We have chosen to include only the severe brain malformations or dysfunction in our study.A surveillance artefact could result from missing data due to lack of an official register of congenital malformations, or due to non-exhaustive PMSI data and the risk that less severe cases who are not hospitalised were not notified which would have led to less cases having been noted.As concerns severe cases, the nature of the symptoms seems to make it unlikely that we have missed such cases while as pointed out above, milder cases could have been missed.Other postnatal lesions such as hearing, visual, cognitive impairments and epilepsy that could manifest later in life might be reported in further prospective studies in populations exposed to ZIKV.

Conclusion
ZIKV was for many years thought to be a benign febrile illness.In line with findings in Brazil, we retrospectively identified an unusual and heterogeneous cluster of congenital brain malformations and brainstem dysfunction in fetuses and newborns over a limited period following a ZIKV epidemic in French Polynesia.Except for four cases of fetal microcephaly, presenting with infectious virus in amniotic fluid and further confirming the Brazilian experience [9][10][11][12][13], we found no evidence of ZIKV vertical transmission for additional 15 cases; however, we were only able to perform ZIKV RT-PCR for three of them and we have no more opportunity to prove or to exclude the ZIKV infection in the 12 remaining cases.The spacio-temporal association might explain brain malformations in context of a viral outbreak.A case-control study is planned to look for other risk factors for these unexplained fetal and newborn abnormalities.
Given the potentially substantial public health implications of our findings, we strongly encourage researchers and clinicians to ensure that non-microcephalic congenital abnormalities, extra-cerebral malformations and brainstem dysfunction are included in their viral investigations for the potential teratogenicity of ZIKV.

Figure 1
Figure 1 Prenatal MRI T2-weighted performed at 30 weeks plus 5 days of gestation on fetus with congenital cerebral malformations following the 2013-2014 Zika virus outbreak, French Polynesia, 2014 to 2015

Figure 2
Figure 2 Prenatal MRI T2-weighted performed at 28 weeks plus 4 days of gestation on fetus with congenital cerebral malformations following the 2013-2014 Zika virus outbreak French Polynesia, 2014 to 2015

Figure 3
Figure 3 Prenatal ultrasound (a) performed at 27 weeks of gestation and prenatal MRI T2-weighted (b) performed at 30 weeks plus 1 day of gestation on fetus with congenital cerebral malformations following the 2013-2014 Zika virus outbreak, French Polynesia, 2014 to 2015

Figure 4 US
Figure 4 Chronological sequence of Zika virus infection and pregnancy in 19 cases, French Polynesia, 2014 to 2015

Table 1
Main characteristics of fetuses and newborns with congenital cerebral malformations and dysfunction following the 2013-2014 Zika virus outbreak, French Polynesia, 2014 to 2015 (n=19) Laboratory tests were normal for Cases 7 and 8, for Case 6, none were available.No amniotic fluids nor placental samples were taken from any of the three cases in this group.One mother (Case 6) reported clinical signs of ZIKV infection in early pregnancy, but no : day; NA: not available; Neg: negative; Pos: positive; TOP: termination of pregnancy; WG: weeks of gestation.a Term (WG+D) at TOP.

Table 2
Imaging findings in fetuses and newborns with congenital cerebral malformations and dysfunction following the 2013-2014 Zika virus outbreak, French Polynesia, 2014-2015 (n=19) palate, and a Moebius syndrome i.e. palsy of sixth and seventh cranial nerves.One of them (Case 17) required a tracheostomy and the other also had clubfeet and epilepsy (Case 15).One case had renal hypoplasia (Case 16).Cardiac dysautonomia manifested by attacks of bradycardy or tachycardy, was constantly present in all five cases.