Introduction

The Zika virus (ZIKV) epidemic in Brazil, and its consequences for perinatal and childhood health outcomes, have presented lasting challenges for affected families, local health authorities, and the scientific community. In May 2015, ZIKV transmission was confirmed in Brazil [1]. Later that year, an unexpected rise in microcephaly cases among newborns prompted the identification of a link between congenital ZIKV infection and the observed birth defects [2]. Since then, researchers provided evidence showing that the full spectrum of Congenital Zika Syndrome (CZS) extends beyond the initially described characteristics of microcephaly and brain malformations to also include other structural abnormalities and functional deficits [3, 4]. In total 1,828 congenital ZIKV cases have been confirmed in Brazil, with a peak in 2015–2016 followed by a decline in 2017–2018, with sporadic cases in subsequent years, just 4 cases being reported in 2023 [5]. Zika transmission remains in several countries but has generally been at low levels throughout 2018 to the present [6].

Genomic analyses have revealed that the Asian lineage of ZIKV predominated during the Brazilian outbreak [7]. However, data on viral sequences from congenital ZIKV infections remain limited, and the contribution of specific viral mutations to microcephaly pathogenesis remains unclear [8].

Since their initiation in late 2015, individual studies in Brazil have confirmed the link between microcephaly and congenital ZIKV infection and provided evidence showing that the full spectrum of Congenital Zika Syndrome (CZS) extends beyond the initially described characteristics of microcephaly and brain malformations to also include other structural abnormalities and functional deficits [3, 4, 9].

Although the Brazilian cohorts were developed independently at the height of the epidemic, the principal investigators of the studies organized meetings, which were supported by PAHO/WHO, to harmonize protocols and data collection instruments for future pooled analyses [10]. The resultant Zika Brazilian Cohorts-Consortium (ZBC-Consortium) now brings together individual-level data from 12 cohorts of children with CZS and Zika-related microcephaly born between 2015 and 2018 in the North, Northeast and the Southeast regions of Brazil [11]. One of the main gaps in characterizing the CZS up to now is the small number of participants in individual studies. Pooling the data together overcomes the sample size-related limitations of individual studies and provides unique opportunities for investigating heterogeneity between studies and, as appropriate, to better characterize the CZS. This study focuses on the subset of children with Zika-related microcephaly in the ZBC-Consortium and conducts individual-participant data meta-analyses to describe adverse anthropometric, clinical, and neuroimaging outcomes occurring in the first three years of life.

Methods

Results

We included 843 children, of which 542 (64.3%) with definitive CZS diagnoses, 127 (15.1%) highly probable, and 35 (4.1%) moderately probable, based on França et al. criteria. It was not possible to classify 139/843 (16.5%) children with microcephaly according to these criteria. Microcephaly was diagnosed at birth in 601 of the 843 (71.3%) children. A total of 58 of the 843 (6.9%) children were diagnosed with microcephaly at first evaluation as there was no measure of HC at birth; the median age of evaluation was 31.6 months (IQR 0.4-40.1). From the 843 children, 172 (20.4%) that had no microcephaly at birth, developed microcephaly over follow-up, i.e., postnatal microcephaly. For the remaining 12 of the 843 (1.4%) children there was information on the presence of microcephaly but the HC measure was not provided. In the group of children with microcephaly diagnosed at birth 217 (36.1%) were classified as having moderate and 384 (63.9%) as having severe microcephaly. In those diagnosed at first evaluation, 8 (13.8%) had moderate and 50 (86.3%) had severe microcephaly. In the group with postnatal microcephaly, 63 (36.6%) had moderate and 109 (63.4%) had severe microcephaly. Microcephaly was not classified according the severity for the 12 children for which the HC measure was not informed.

The study participants originated from cohorts in three out of the five Brazilian Regions. The North Region comprised 2.3% of the total sample, with 19 participants, being 9 from Belém and 10 from Manaus. The Northeast Region comprised 73.5% of the total sample, with 620 participants, being 76 from Aracaju; 181 from MERG [21] and 15 from Saúde Coletiva-Universidade Federal de Pernambuco both in Recife; 106 from São Luis; 129 from Instituto de Saúde Coletiva-Universidade Federal da Bahia and 113 from Instituto Gonçalo Moniz-Fiocruz both in Salvador. The Southeast Region comprised 24.2% of the total sample, with 204 participants, being 80 from Jundiaí [22]; 34 from Ribeirão Preto; 74 from Instituto Fernandes Figueira-Fiocruz and 16 from UFRJ both in Rio de Janeiro (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Participating cohorts of the ZBC-Consortium and severity of microcephaly per site.

A. Distribution of the cohorts participating in the ZBC-Consortium by Brazilian Regions. Note: Link to the base layer of the map: https://pixabay.com/pt/vectors/brasil-geografia-mapa-estados-153881/; Link of the terms of use/license information for the base layer of the map: https://pixabay.com/pt/service/license-summary/. B. Frequency of severe microcephaly in the ZBC-Consortium per site, city, and Brazilian Region.

https://doi.org/10.1371/journal.pgph.0005425.g001

The p-value of the LR test (RE vs FE Model) for the frequency of severe microcephaly was 0,000 suggesting heterogeneity between sites. The pooled estimate of the frequency of severe microcephaly was 47.1% across study sites, ranging from 11.1 to 87.5% (Fig 1B).

The frequency of disproportionate microcephaly (i.e., classified in n = 601 children evaluated at birth) ranged from 15.8 to 93.3% overall, being 10.4 to 88.9% in children with moderate microcephaly and 64.2 to 92.8% in children with severe microcephaly. (Overall:LR test: RE vs FE Model p = 0.000, τ²= 1.376); Moderate: LR test: RE vs FE Model p = 0.000, τ²= 1.211; Severe: LR test: RE vs FE Model p = 0.000, τ²= 1.130). The site Recife-SC UFPE was excluded of the analysis for disproportionate microcephaly overall and moderate microcephaly because of the low number of children with severe microcephaly. For the analysis of disproportionate microcephaly in children with severe microcephaly the sites Manaus, Belém, Recife-SC UFPE and Ribeirão Preto were excluded because of the low number.

Although, there was heterogeneity between sites in relation to the occurrence of microcephaly at birth, severity of microcephaly, and abnormal CNS image exam, the overall pooled estimated frequency of these variables was higher in children with disproportionate than in those with proportionate microcephaly. The overall pooled estimated frequency of microcephaly at birth in children with disproportionate and proportionate microcephaly was respectively 95.7% (95% CI 84.2–98.9) and 63.6% (95% CI 42.1–80.8); of severe microcephaly in children with disproportionate and proportionate microcephaly was 66.6% (95% CI 51.8–78.7) and 25.0% (95% CI 16.9-35.3 respectively. The frequency of neurological disorders in children with disproportionate and proportionate microcephaly was 72.8% (95% CI 59.6–82.9) and 64.2% (95% CI 57.6–70.3), respectively.

Among the 680 microcephalic children with available data on congenital malformations (Table 1), more than one-fifth presented with: epicantus (40.1%), prominent occiput (39.2%), excess skin on the neck (26.7%), overlapping cranial bones (24.2%), strabismus (23.7%), retrognathia (23.5%), girata skin (22.4%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Congenital malformations in children with Zika-related microcephaly participating in the ZBC-Consortium.

https://doi.org/10.1371/journal.pgph.0005425.t001

The frequency of low birth weight was 33.2% varying from 10 to 43.8%, but for most of the sites it was between 20 and 40% (LR test: RE vs FE Model p = 0.449) (Fig 2A). For small for gestational age and large for gestational age there was also no heterogeneity (S1B Fig and S1C), while prematurity, and dysphagia presented with heterogeneity and a high variability through sites (S1A Fig and S1D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Frequency of low birth weight and neurological abnormalities of children with Zika-related microcephaly of the ZBC-Consortium.

Frequency of low birth weight (A), epilepsy (B), deficit of social attention (C) and persistence of primitive reflexes (D) in children with Zika-related microcephaly of the ZBC-Consortium.

https://doi.org/10.1371/journal.pgph.0005425.g002

The neurological manifestations among the children considered each site reported. Several neurological findings, including those that are proxy of severe injury were very frequent through studies sites; deficit of social attention (lack of social interactions) was reported in at least 50% of the children in the cohort studies with higher number of participants (Fig 2C). The same was observed on spastic hypertonia (S2B Fig). Epilepsy was reported in 30–80% (LR test: RE vs FE Model p = 0.000, overall pooled estimate = 58.3, 95% CI 44.5-71.0) (Fig 2B). Persistence of primitive reflexes (Fig 2D) was investigated in five sites and there was homogeneity between sites with an overall pooled estimate of 63.1 (CI 56.5-69.3).

The frequency of eye fundal abnormality among the children ranged from 0 to 67.1% (LR test: RE vs FE Model p = 0.000, overall pooled estimate = 49.8, 95% CI 39.0-60.7) (S2A Fig) and the frequency of optical nerve abnormality varied from 0 to 36.5% (LR test: RE vs FE Model p = 0.111, overall pooled estimate = 26.6, 95% CI 19.9-34.6) (Fig 3A), and the frequency of retinal abnormality varied from 0 to 100% (LR test: RE vs FE Model p = 1.000, overall pooled estimate = 31.1, 95% CI 26.4-36.3) (Fig 3B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Frequency of ocular abnormalities in children with Zika-related microcephaly participating in the ZBC-Consortium.

Frequency of optical nerve (A) and retinal (B) abnormalities in children with Zika-related microcephaly participating in the ZBC-Consortium.

https://doi.org/10.1371/journal.pgph.0005425.g003

In Fig 4A we observe that for five sites, all studied children had at least one abnormal CNS image exam, and for the other six, at least 80%. Calcifications and ventriculomegaly were the most consistent and frequent described abnormality, calcification ranging from 26.7 to 100% (LR test: RE vs FE Model p = 0.000, overall pooled estimate = 81.7, 95% CI 68.4-90.2) (Fig 4B) and for ventriculomegaly, the frequency was above 55% for 10 of the 11 sites (LR test: RE vs FE Model p = 0.000, overall pooled estimate = 76.8, 95% CI 57.8-88.9) (Fig 4C). Spastic hypertonia was found in seven sites in which they were investigated, with a frequency of 100% in four sites (S2B Fig). For cerebellum hypoplasia/atrophy and agenesis/dysgenesis of the corpus callosum there was heterogeneity, and the frequency of cerebellum hypoplasia/atrophy varied from 0 to 50% and of agenesis/dysgenesis of the corpus callosum, varied from 0 to 100% (S2C Fig and S2D Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Frequency of imaging and audiologic abnormalities in children with Zika-related microcephaly participating in the ZBC-Consortium.

Frequency of abnormal cerebral ultrasonography, or tomography, or magnetic resonance imaging (A), calcification (B), ventriculomegaly (C) and abnormal ABR test (D) in children with Zika-related microcephaly participating in the ZBC-Consortium.

https://doi.org/10.1371/journal.pgph.0005425.g004

Although an audiology exam was not performed in all the cohort studies, abnormalities were not as common as CNS or eye disease. The frequency of abnormal ABR test was 0% in six out of 10 sites (Fig 4D).

The forest plot graphics with a sensitive analysis, using the Freeman-Tukey double arcsine transformation, estimating the same parameters plus the I-squared (I²) statistics, are shown in supplement (S3 Fig to S7 Fig). The results of the pooled estimates were very similar in both approaches and for most of the variables the I² was above 75%, suggesting high heterogeneity.

Discussion

This study brings together individual participant data from 12 prospectively followed pediatric cohorts from three regions of Brazil that were started during the ZIKV epidemic. Although the frequencies of individual abnormalities were generally heterogeneous across the studies, most children with Zika-related microcephaly presented with a very similar spectrum of signs and symptoms, regardless of geographic origin, with the most common abnormalities related to neurologic function, ocular structure, and hearing. Similar to our findings from Brazil, previous research from the French territories in the Americas, the United States territories, and Colombia reported neurological, ophthalmological, and hearing abnormalities as the most frequent outcomes in children with prenatal ZIKV exposure [23, 24].

We observed a notable overlap between microcephaly and other congenital malformations, with the most frequent being epicantus, prominent occiput, excess skin on the neck and overlapping cranial bones. Prominent occiput and excess skin on the neck are due to the skull deformity associated with the severity of microcephaly. The other malformations appear to be part of the spectrum of manifestations of CZS.

We also observed a high frequency of children who were born with low birth weight (20–40%) and/or small for gestational age (30–65%). These findings align with a large body of evidence suggesting that congenital infection from pathogens, such as Toxoplasma gondii, rubella virus, cytomegalovirus (CMV), herpes simplex virus (HSV), and Treponema pallidum (syphilis), may result in intrauterine growth restriction (IUGR) [25]. Strikingly, a high number of children (n = 394) in this study presented with disproportionate microcephaly. As the frequency of severe microcephaly and abnormal CNS image exam was higher among children with disproportionate microcephaly compared to those proportionate microcephaly, we hypothesize that brain damage may have been the main cause of the disproportion.

A range of neurological signs/symptoms were also very frequent in children with Zika-related microcephaly. In the pooled analyses, we observed high frequencies of social attention deficits, the persistence of primitive reflexes, and epilepsy. The occurrence of neurologic abnormalities likely represents both direct (e.g., ZIKV’s tropism for neural progenitor stem cells) and indirect (e.g., inflammatory responses that compromise neurogenesis) consequences of the presence of ZIKV in the CNS during fetal development [26]. Further research is warranted to better understand the underlying mechanisms and how factors associated with vertical transmission, such as the timing of infection and/or the viral load, may influence the frequency of ZIKV-related neurologic abnormalities. Further investigation is also necessary to clarify if attention deficits in these children is due to development delay and/or autism and/or if they are sensory impaired with low vision or hearing loss.

Consistent with the high frequency of neurologic abnormalities, postnatal brain assessments based on US, CT, and/or MRI imaging found abnormalities in several sectors of the brain. In the pooled analyses, we observed high frequencies of calcifications, ventriculomegaly, and cerebral atrophy. These results are broadly consistent with MRI-based animal model studies, which showed that ZIKV-infected infant rhesus macaques had ventriculomegaly, hypotrophy/atrophy, and altered functional connectivity between brain areas (e.g., amygdala, hippocampus, cerebellum). Further, the neuropathological alterations corresponded with neuroimaging results and were consistent with the behavioral and memory deficits [27]. This type of finding is consistent with that we observed in the children exposed to ZIKV [28].

Ocular structural abnormalities were similar to those reported by other authors in Brazil and elsewhere, mainly optic nerve pallor/hypoplasia, focal pigment mottling and chorioretinal scars [29]. The hearing evaluations we showed were based just on ABR test. More detailed by Cortical Auditory Evoked Potentials in a small number of children suggested that regardless of changes in the CNS, this population appears to have, to some extent, the cortical ability to process sound stimuli, however, the presence of microcephaly may interfere with the attention capacity and auditory discrimination [30].

In the pooled analyses, dysphagia was common and observed in 46.9% of children. The substantial heterogeneity is likely to be due to differences in case ascertainment approaches across the study sites.

The findings described in this study were those that were investigated in most of the sites, as they were part of the manifestations of the CZS initially described. However, the broader description of the CZS spectrum must be complemented by the findings of specific studies that used more sophisticated instruments to investigate some abnormalities, such as manifestations of the urogenital system [31], endocrinological [32], cardiorespiratory [33, 34] and gastrointestinal [35]. Additional and equally important information should come from studies evaluating the neurodevelopment of these children as they reach different age groups [36].

Although there was heterogeneity for most of the variables, it does not seem that there is likely to be a biological difference (e.g., in relation to underlying comorbidities or consequences of congenital infection) between the children in the different cohorts within the ZBC-Consortium. There are several possible study design-related explanations for the heterogeneity in the frequency of some characteristics across studies. First, each study used different recruitment criteria (e.g., some recruited antenatally, such as from pregnancy cohorts, while others recruited from outpatient pediatric clinics, primary care settings, or from the Municipal Health Surveillance database which were subsequently contacted by the local primary care teams). These different recruitment criteria could introduce selection biases (e.g., sites recruiting as part of surveillance efforts may have had a higher sensitivity for recruiting children with milder CZS). Of note, the frequency of severe microcephaly varied from 11 to 87.5% across cohorts. We believe that this variation is probably related to different inclusion criteria among sites, as some sites primarily recruited children with phenotypes more typical for severe rather than moderate microcephaly. Second, since data were collected from studies conducted concurrently, without a previous standardization, and retrospectively harmonized, outcome definitions may have differed to some extent. Notably, each study relied on the expertise of local evaluators, and potentially differing techniques and instruments, to diagnose the abnormalities in the context of a new congenital syndrome never before described. In addition to differences between studies, the learning curve over the duration of follow-up within a given site must be considered.

An important strength of this study was the large sample size, pooling the individual primary data of 843 children with Zika-related microcephaly, a large proportion of the 1,828 confirmed CZS cases notified between 2015 and 2023 in Brazil [5]. This study has some limitations: i) children were enrolled since the start of the epidemic, and therefore the laboratory confirmation was not possible for all of them. This is the reason why we worked with different levels of evidence of ZIKV infection, the criteria adopted by Brazilian researchers [10]; ii) there may have been some degree of misclassification of CZS, but it is not likely that it may have affected substantially our results as only 4.1% of the children had a level of evidence of moderately probable. In relation to the fact that children with microcephaly born to mothers who presented ZIKV infection during pregnancy had the diagnosis of CZS categorized as ‘definitive’, it may not be excluded that they had microcephaly of another aetiology, but if it occurred was probably in a very small proportion as microcephaly is a rare event and again it is not likely that it would bias our results.

In conclusion, this study presents the findings from the largest ever analysis of children with Zika-related microcephaly. This relatively large sample size allowed us to observe that, among children with microcephaly, there exists a spectrum in the severity and types of manifestations of CZS. Despite the heterogeneity found across the different sites, structural abnormalities of the central nervous system, detected by neuroimaging, and abnormalities in neurological and ophthalmological examination were the most frequently reported outcomes. This severe CNS damage demands multidisciplinary care and assistance from different medical and other health specialties.

Supporting information

Acknowledgments

The authors also would like to acknowledge all researchers and fieldworkers who contributed to the conduction of all studies participating in the ZBC-Consortium.

The members of the Zika Brazilian Cohorts Consortium (ZBC Consortium) are: Maria de Fátima P M de Albuquerque, Maria das Graças Costa Alecrim, Melania Maria Ramos de Amorim, Marcia da Costa Castilho, Antonio Ledo A Cunha, Geraldo Duarte, Sophie H Eickmann, Flor Ernestina Martinez Espinosa, Cássia Fernanda Estofolete, Juliana Herrero-Silva, Adriana Suely de Oliveira Melo, Marisa Marcia Mussi-Pinhata, Maurício Nogueira, Arnaldo Prata-Barbosa, Deolinda Maria Felin Scalabrin, Lavínia Schuler-Faccini, Wayner Vieira de Souza, Marilia Dalva Turchi, and Ana Laura de Sene Amâncio Zara.