Risk Factors Associated With the Ophthalmoscopic Findings Identified in Infants With Presumed Zika Virus Congenital Infection

Importance  The Zika virus (ZIKV) might cause microcephaly and ophthalmoscopic findings in infants of mothers infected during pregnancy.

Objective  To assess and identify possible risk factors for ophthalmoscopic findings in infants born with microcephaly and a presumed clinical diagnosis of ZIKV intrauterine infection.

Design, Setting, and Participants  We conducted a cross-sectional study at the Altino Ventura Foundation in Recife, Brazil, that included 40 infants with microcephaly born in Pernambuco state, Brazil, between May and December 2015. Toxoplasmosis, rubella, cytomegalovirus, syphilis, and human immunodeficiency virus were ruled out in all of them. Testing of cerebrospinal fluid for ZIKV using IgM antibody-capture enzyme-linked immunosorbent assay was performed in 24 of 40 infants (60.0%). The infants and mothers underwent ocular examinations. The infants were divided into 2 groups, those with and without ophthalmoscopic alterations, for comparison.

Main Outcomes and Measures  Identification of risk factors for ophthalmoscopic findings in infants born with microcephaly and ZIKV intrauterine infection.

Results  Among the 40 infants, the mean (SD) age was 2.2 (1.2) months (range, 0.1-7.3 months). Of the 24 infants tested, 100% had positive results for ZIKV infection: 14 of 22 infants (63.6%) from the group with ophthalmoscopic findings and 10 of 18 infants (55.6%) from the group without ophthalmoscopic findings. The major symptoms reported in both groups were rash by 26 mothers (65.0%), fever by 9 mothers (22.5%), headache by 9 mothers (22.5%), and arthralgia by 8 mothers (20.0%). No mothers reported conjunctivitis or other ocular symptoms during pregnancy or presented signs of uveitis at the time of examination. Thirty-seven eyes (46.3%) of 22 infants (55.0%) had ophthalmoscopic alterations. Ten mothers (71.4%) of infants with ocular findings reported symptoms during the first trimester (frequency, 0.48; 95% CI, 0.02-0.67; P = .04). A difference was also observed between the groups of infants with and without ocular findings regarding the cephalic perimeter: mean (SD) of 28.8 (1.7) and 30.3 (1.5), respectively (frequency, −1.50; 95% CI, −2.56 to −0.51; P = .004).

Conclusions and Relevance  Ocular involvement in infants with presumed ZIKV congenital infection were more often seen in infants with smaller cephalic diameter at birth and in infants whose mothers reported symptoms during the first trimester.

The Zika virus (ZIKV) is a neurotropic flavivirus associated with the dengue, yellow fever, and West Nile viruses.¹ It was first identified in a rhesus monkey in 1947¹ and in humans from Nigeria and Uganda 5 years later.^2,3 However, it was not until April 2007 that a ZIKV outbreak was reported outside of Africa, on Yap Island in the Federated States of Micronesia.⁴ This outbreak was characterized by rash, conjunctivitis, and arthralgia.⁴ Another ZIKV outbreak occurred in French Polynesia from September 2013 to March 2014, where evidence of perinatal viral transmission was identified.⁵ Although intrauterine infection was documented during this outbreak, no case of microcephaly was associated with this viral infection.^4,5

In April 2015, the first ZIKV autochthonous transmission was confirmed in Brazil, followed by an epidemic in the country.⁶ It was estimated that more than 1 million Brazilians have had the ZIKV infection since then, reflecting the virus’ capacity to cause large-scale outbreaks where the vector is present.^6,7 After the Brazilian ZIKV outbreak, an unexpected increase in the number of newborns with microcephaly was identified. An update issued by the Brazilian Ministry of Health (BMH) released in January 2016 reported 3174 suspect cases.⁸

The virus has been detected in the amniotic fluid of 2 pregnant women of babies born with microcephaly in Brazil and in the tissue of a newborn with microcephaly who died after birth.⁶ Thus, the BMH and the World Health Organization have associated this malformation with the ZIKV intrauterine infection, and an epidemiologic alert regarding the consequences of ZIKV infection was released in December 2015.^6,7 After this alert, a retrospective analysis of the French Polynesia outbreak identified an unusual increase in central nervous system malformations in fetuses and infants born from 2014 to 2015 in 17 cases; the malformations included cerebral malformations, polymalformative syndromes, and brainstem dysfunction.⁹

In January 2016, our group published the first report in the literature, to our knowledge, on the ocular findings of 3 infants with microcephaly and presumed ZIKV vertical infection, followed by 10 other cases published in February 2016 describing the optic nerve and macular abnormalities.^10,11 De Paula Freitas et al¹² also published similar ophthalmoscopic findings in infants from Bahia state, Brazil, in February 2016. In the current study, we analyzed the variables associated with the ophthalmoscopic findings of infants born with microcephaly during the Brazilian ZIKV epidemic and presumed ZIKV intrauterine infections.

Box Section Ref ID

The institutional review board of the Altino Ventura Foundation approved this cross-sectional study, which followed the tenets of the Declaration of Helsinki.¹³ The parents of the patients received explanation concerning the study and provided written informed consent before their children were enrolled.

The inclusion criteria included children born between May and December 2015 in the state of Pernambuco, Brazil, who had suspected or confirmed microcephaly, referred by maternities and pediatric hospitals in Recife, Pernambuco, and children with microcephaly, defined as an occipitofrontal circumference more than 2 SDs below the mean for age and sex, based on the definition of the BMH.⁷

The serological technique used for testing toxoplasmosis, rubella, cytomegalovirus, and human immunodeficiency virus was the chemiluminescent microparticle immunoassay (CMIA); for syphilis, the chemiluminescent microparticle immunoassay and fluorescent treponemal antibody absorption; and for herpes simplex virus, the enzyme-linked immunosorbent assay (ELISA) test.

Patients with incomplete or positive serology results for other congenital infectious disease, such as toxoplasmosis, rubella, syphilis, cytomegalovirus, herpes simplex virus, and human immunodeficiency virus, were excluded from the study.

Twenty-four infants from this study had their cerebrospinal fluid (CSF) tested for ZIKV and for dengue fever virus (DFV) using the IgM antibody-capture ELISA (MAC-ELISA) test. Owing to the unavailability of the test in December 2015, it was not performed in the remaining 16 infants. The method used to perform this test followed the Centers for Disease Control and Prevention protocol as described by Martin et al.¹⁴

All infants and their mothers underwent ophthalmologic assessment at the Altino Ventura Foundation in Recife, Brazil. The fundus of each infant was examined by indirect ophthalmoscopy and fundus findings were documented using the Retcam Shuttle (Clarity Medical Systems Inc). The axial length was measured by a pseudoimmersion technique using the Alcon Ultrascan ultrasound (Alcon Laboratories). Mothers underwent ocular biomicroscopy and fundus examination.

Mothers were interviewed individually by a trained educational psychologist using a standardized and previously used questionnaire¹⁵ regarding their pregnancy history including pregnancy duration, type of delivery, and symptoms during pregnancy. A list of possible systemic and ocular symptoms and signs using colloquial language was presented and read out loud to the mothers so that they could refer if they were present during pregnancy. The time signs and/or symptoms appeared were specifically asked by the psychologist. The trimesters were defined as follows: first trimester, until week 13, or until third month of pregnancy, and second trimester, weeks 14 to 27, or from the fourth to sixth month of pregnancy. Demographic variables, such as mothers’ and infants’ ages at examination, infants’ sex, infants’ cephalic perimeter, and infants’ weight at birth, were analyzed, as well as the presence of cerebral calcifications in previous computerized tomographic (CT) scans.

The statistical software package SPSS for Windows (version 16.0, SPSS Inc) was used for data analyses. Categorical variables were expressed as relative and absolute frequencies. Continuous variables were expressed as the mean (SD) and the maximal and minimal values. For statistical purposes, infants were divided into 2 groups: those with and without ophthalmoscopic alterations. For continuous data, the Mann-Whitney U test was used to analyze any differences between the groups.

Of the 55 infants and their respective mothers who underwent ophthalmic assessment, 40 infants fulfilled the inclusion criteria. One infant (1.8%) was excluded because of positive serology results for syphilis, 1 (1.8%) for positive IgM serology results for toxoplasmosis, 6 (10.9%) for incomplete serology results owing to inadequate specimen, and 7 (12.7%) for not fulfilling the microcephaly criteria. Consequently, the infants and their mothers were also excluded from this study.

Results from IgM MAC-ELISA for ZIKV were positive in the CSF of all 24 infants tested: 14 of 22 infants (63.6%) from the group with ophthalmoscopic findings and 10 of 18 infants (55.6%) from the group without ophthalmoscopic findings. All infants who were tested for DFV in the CSF had negative results, with the exception of 2 (8.3%).

The mean (SD) patient age at examination was 2.2 (1.2) months (range, 0.1-7.3 months), and 21 infants (52.5%) were male. The mean (SD) maternal age at the examination was 25.6 (7.0) years (range, 16-42 years). At birth, the infants’ mean (SD) cephalic perimeter was 29.5 (1.8) cm (range, 26-32 cm), and the mean (SD) birth weight was 2674 (464.9) g (range, 1690-3665 g). Thirty patients (75.0%) underwent CT scans, of which 29 (96.7%) had cerebral calcifications. Twenty-eight infants (70.0%) were born at term, 11 (27.5%) were preterm, and 1 (2.5%) was postterm. One caregiver from each group did not provide the type of delivery. Twenty-one of 38 mothers (55.3%) reported cesarean delivery.

When the pregnancy duration and type of delivery of the infants were analyzed according to the presence or absence of ophthalmoscopic alterations, no significant differences were found. Seventeen mothers (77.3%) of infants with ocular findings and 11 mothers (61.1%) of infants without ocular findings had their pregnancy at term (frequency, 0.16; 95% CI, −0.20 to 0.58; P = .75). As for the type of delivery, 11 mothers (55%) of infants with ocular findings and 10 mothers (55.6%) of infants without ocular findings had a cesarean delivery (frequency, −0.01; 95% CI, −0.33 to 0.32; P = .97). There were also no significant differences between the groups with and without ocular findings regarding the mean (SD) maternal age at examination (26.4 [7.8] years and 24.5 [6.0] years, respectively; P = .41), infants’ mean (SD) age at examination (2.0 [0.7] months and 2.5 [1.6] months, respectively; P = .15), the male to female ratio (9 males [40.9%] and 10 males [55.6%], respectively; P = .36), and birth weight (2688.1 [538.4] g and 2655.7 [363.6] g, respectively; P = .83). There was a difference between the groups (with or without fundus alterations) regarding the cephalic perimeter at birth (frequency, −1.50; 95% CI, −2.56 to −0.51; P = .004) (Table 1).

Twenty-seven mothers (67.5%) reported symptoms during pregnancy (Table 2). The main symptoms were rash (65.0%), fever (22.5%), headache (22.5%), and arthralgia (20.0%). Fourteen of the symptomatic mothers (63.6%) had infants with ophthalmoscopic abnormalities (frequency,  −0.09; 95% CI, −0.28 to 0.48; P = .56). A comparison of the specific symptoms reported by mothers of infants with fundus abnormalities and those with a normal ocular examination showed similar incidence rates in both groups (P > .05 for all comparisons) (Table 2). However, the groups differed significantly regarding the point at which the symptoms occurred, ie, 10 of 14 mothers (71.4%) with infants with ocular findings reported symptoms during the first trimester, whereas 8 of 13 mothers (61.5%) with infants with a normal ocular examination had symptoms during the second trimester (frequency, 0.36; 95% CI, 0.02-0.67; P = .04) (Table 2).

When analyzing the macular and optic nerve findings according to the trimester of the mothers’ symptoms, the difference (95% CI, 0.03-0.76; P = .008) was associated with the macular alterations and first trimester of pregnancy (Table 3). No mothers reported conjunctivitis or ocular symptoms during pregnancy. All of them had normal ocular examinations, except for 1 mother with a pale optic disc in the right eye; however, the decreased vision was present before the pregnancy, according to the mother’s report.

Fundus alterations were seen in 37 eyes (46.3%) of 22 infants (55.0%), and they were bilateral in 15 (68.2%) of them (Table 4). In 5 infants (22.7%), only the optic nerve was affected; in 7 (31.8%), only the macula; and in 10 (45.5%), both structures (Table 4). Optic nerve abnormalities were seen in 25 eyes (31.3%) of 15 infants (37.5%) and consisted of hypoplasia with the double-ring sign in 14 eyes (17.5%), optic disc pallor in 6 eyes (7.5%), and increased cup-to-disc ratio in 9 eyes (11.3%).

Macular abnormalities were present in 24 eyes (30.0%) of 17 infants (42.5%) and were unilateral in 10 infants (58.8%) (Table 4). The macular abnormalities included foveal reflex loss in 24 eyes (30.0%), mild pigment mottling in 13 eyes (16.3%), gross pigment mottling in 9 eyes (11.3%), and sharply demarcated circular areas of chorioretinal atrophy in 6 eyes (7.5%) (Table 4; eFigure in the Supplement).

The mean (SD) axial length of the eyes was 18.5 (0.9) mm (range, 16.8-21 mm) and did not differ significantly (mean, −0.60; 95% CI, −1.13 to −0.05; P = .055) compared with the group with ocular abnormalities with and without fundus alterations (Table 1). In the eyes with macular alterations, no posterior coloboma was detected by ocular ultrasonography. In all eyes, the retina and choroid were attached, and no signs of uveitis or ocular calcifications were detected. Furthermore, no anterior coloboma was identified during the ocular examination.

The current study reports the results of the ophthalmoscopic assessments of 40 infants with microcephaly and presumed congenital ZIKV infection examined in Recife, Pernambuco. Pernambuco state is considered the epicenter of the ZIKV infection, with the highest number of microcephaly cases in the country. According to the BMH report released in April 2016, 1849 cases were under investigation in this state.⁸

The presumed diagnosis of ZIKV congenital infection of these 40 infants was based on the criteria established by the BMH, which includes microcephaly (defined by an occipitofrontal circumference more than 2 SDs below the mean for age and sex; negative serology results for toxoplasmosis, rubella, syphilis, cytomegalovirus, and human immunodeficiency virus; and infants born in Brazil after May 2015).⁷ It is important to rule out these other agents because they are known to cause fetal infection, brain insults such as microcephaly, cerebral calcifications, and similar ocular findings.^16,17

In December 2015, when these infants started being examined, the only available blood examination available to test ZIKV was reverse-transcriptase-polymerase-chain-reaction assays, which is useful for acute illness.^7,18 Recently, IgM MAC-ELISA for ZKV became commercially available in Brazil and we were able to perform the test in 60.0% infants of our study. Despite not being able to perform ZIKV serology in all 40 infants, the similar distribution of positive results for ZIKV in both groups (with and without ophthalmoscopic findings) reinforces the clinical diagnosis of ZIKV congenital infection in all 40 infants.

There were 2 infants those results were also positive for DFV infection, but in both cases, the IgM value for DFV in the CSF was 3 times less than the IgM value for ZIKV, which makes us conclude that these 2 patients were infected by ZIKV alone. Zika virus serology, especially in regions where DFV is highly frequent, can cause false-positive results and that is why its reliability is questioned for accurate ZKV prenatal and antenatal infection diagnosis.¹⁸ In addition, the first case of DVF in Brazil was reported in 1982.¹⁹ Ever since, Brazil has reported cases of all 4 DVF serotypes and thousands of cases of DVF are reported every year.¹⁹ Nevertheless, no neurological complications, such as Guillain-Barré Syndrome or congenital malformations such as microcephaly, cerebral calcifications, and ocular findings, ever were associated with this virus.¹⁹

The literature contains little information about the ophthalmologic findings of infants with a clinical diagnosis of ZIKV-related microcephaly.^10-12 Furthermore, the current study is the first, to our knowledge, to assess the effects of different variables, such as the trimester of the maternal symptoms on the occurrence of ophthalmoscopic abnormalities in these infants and demographic and pregnancy history variables (a Medline literature search on March 8, 2016, did not reveal any publications regarding these variables).

Previous studies have supported the hypothesis that the 20-fold increase in microcephaly in Brazil in 2015 was associated with vertical transmission of the ZIKV.^6,7,16 Moreover, a Slovenian group published a study in which the complete genome sequence of ZIKV was isolated in the brain of an aborted fetus with microcephaly, which reemphasizes the association of ZIKV and microcephaly.²⁰ Additionally, Rasmussen et al²¹ used Shepard’s criteria, proposed for assessment of potential teratogens, to state that there is sufficient evidence to infer a causal association between prenatal ZIKV infection and the adverse outcomes, such as microcephaly and other brain anomalies, detected in these infants.

The presence of cerebral calcifications, as observed in 96.7% of our 30 patients who had CT scans, supports the assumption of vertical infection transmission because it seems to be a frequent sign observed in congenital infections and in the current patients.¹⁶ Our group previously published an assessment of 13 infants born with microcephaly and presumed ZIKV vertical transmission in the state of Pernambuco, and all had cerebral calcifications.^10,11 Other authors also have reported cerebral calcifications in babies with microcephaly associated with ZIKV.^16,20

Because the Brazilian ZIKV epidemic is the first to associate intrauterine infection with infants born with microcephaly and ocular lesions, and owing to the pandemic pattern of the disease, the ZIKV infection became a major public health problem.^6,10-12 To control the spread and damage of this disease, new strategies are being adopted worldwide. The Pan American Health Organization/World Health Organization has reemphasized surveillance recommendations and prevention control measures such as actions for vector (Aedes aegypti) control, encouragement of pregnant women to use mosquito repellents, and are supporting research to develop antiviral therapy and vaccines against the ZIKV.⁶ The Centers for Disease Control and Prevention has also recommended that pregnant women should not travel to countries with ZIKV cases and the World Health Organization recently declared ZIKV as a global health emergency.^22,23

When subdividing all infants based on the presence or absence of fundus alterations, both groups were compared for the demographic and pregnancy history variables; the small cephalic perimeter at birth was the only variable that was correlated significantly with the ophthalmoscopic findings.

According to previous reports, it was estimated that 20% of patients with ZIKV actually develop symptoms and that the absence of symptoms during pregnancy does not eliminate the possibility of ZIKV-related microcephaly in babies.⁷ In the current study, almost 70% of the mothers reported symptoms during pregnancy, similar to our previous study.¹¹ De Paula Freitas et al¹² also reported this high frequency of symptoms described by almost 80% of the mothers from Salvador, Bahia, Brazil.

In the current study, mothers with symptoms during the first trimester were more likely to have an infant with ophthalmoscopic alterations, which resembles the natural history of other congenital infections such as toxoplasmosis.²⁴ When analyzing the gestational period in which the symptoms occurred, they were significantly associated with the fundus findings in general and with the macular findings isolated, but not with the optic nerve findings. This suggests that ZIKV ocular involvement and, more specifically, the macular alterations, are related to the trimester of vertical transmission.

No mothers reported conjunctivitis or ocular symptoms during pregnancy, which differs from the findings encountered during the Micronesia outbreak.⁴ Furthermore, the last country before Brazil to have a ZIKV outbreak was French Polynesia and it did not report ocular abnormalities in infants, which also differs from Brazil’s outbreak.^10-12

In the current study, the main optic nerve and macular findings, respectively, were optic disc hypoplasia and mild pigment mottling with foveal reflex loss. Most ocular abnormalities were bilateral, as reported previously.^11,12 Chorioretinal abnormalities and central nervous system malformations have been found in association with another flavivirus, the West Nile virus, but they have not been previously associated with ZIKV.²⁵ In the current study, macular lesions were detected in 55% of the studied population. Optic nerve hypoplasia with the double-ring sign was detected in 17.5% of infants. This finding has already been correlated with central nervous system abnormalities similar to congenital cytomegalovirus infection, which was ruled out by serology.²⁵

A comparison of the mean axial lengths of both groups showed no significant difference. No posterior coloboma was identified using ocular ultrasonography, which led us to hypothesize that ZIKV vertical infection does not interfere with the period of ocular organogenesis, although additional research is necessary to support this hypothesis. It remains unclear whether the ocular lesions are a result of the microcephaly and/or to the ZIKV. It is also possible that a dual-mechanism process might be the pathogenesis of these ocular findings.

A limitation of the current study was the absence of specific serology to confirm the ZIKV infection in all cases. Another limitation was regarding the data collected about mothers’ symptoms during pregnancy. All data analyzed were based on what was reported by them; therefore, we had to rely on what they recalled of their symptoms and period of pregnancy; most of them could not recall their precise their symptoms by weeks of gestational age.

Fundus abnormalities in infants with presumed ZIKV congenital infection were associated with smaller cephalic diameters at birth and with those infants whose mothers reported symptoms during the first trimester.

Corresponding Author: Rubens Belfort Jr, MD, PhD, Departamento de Oftalmologia, Escola Paulista de Medicina, Rua Botucatú 831, São Paulo, Brazil 04023-062 (clinbelf@uol.com.br).

Submitted for Publication: March 12, 2016; final revision received April 18, 2016; accepted April 24, 2016.

Published Online: May 26, 2016. doi:10.1001/jamaophthalmol.2016.1784.

Author Contributions: Dr C. V. Ventura had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: C. V. Ventura, Maia, L. O. Ventura, Belfort.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: C. V. Ventura, Maia, Torres, L. O. Ventura, Belfort.

Critical revision of the manuscript for important intellectual content: C. V. Ventura, Maia, Travassos, Martins, Patriota, Nunes, Agra, van der Linden, Ramos, Rocha, Silva, L. O. Ventura, Belfort.

Statistical analysis: C. V. Ventura, Maia, L. O. Ventura.

Obtained funding: C. V. Ventura, L. O. Ventura.

Administrative, technical, or material support: C. V. Ventura, Maia, Torres, L. O. Ventura, Belfort.

Study supervision: C. V. Ventura, Maia, L. O. Ventura, Belfort.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Additional Contributions: We are thankful to our ophthalmology colleagues at the Altino Ventura Foundation: Vasco Bravo Filho, MD, Adriana L. Góis, MD, Bruna V. Ventura, MD, Erika Anjos, MD, Silvio Di Biase, MD, Leandro Pimentel, MD, Paulo Escarião, MD, PhD, and Mônica Rocha, MD, who contributed greatly in this study during the collection of data. They did not receive compensation for the contributions.