Microcephaly and Central Nervous System Abnormalities During the Zika Outbreak in Rio de Janeiro

1. Introduction

The Zika virus (ZIKV) has gone from a virus associated with mild infections to one of the most studied viruses worldwide in the past six years. Between 2015 and 2016, the ZIKV was responsible for the Americas' outbreak, especially in Brazil. During this period, the first reports of pregnant women with confirmed or suspected ZIKV infection showing fetuses and newborns with severe congenital malformations, especially microcephaly and central nervous system (CNS) abnormalities, led to a strong suspicion of their association with the virus. Since then, evidence has favored this association [1,2,3].

As of January 2018, more than 3,700 cases of congenital Zika Syndrome (CZS) have been notified in the Americas [4]. Local transmission of the ZIKV has been reported in 87 countries and territories worldwide, both in tropical and subtropical areas. The transmission of ZIKV has significantly dropped since the end of 2016, from more than 500,000 cases reported during 2016 to under 30,000 cases in 2018 [5].

ZIKV outbreaks continue to occur in different regions of the world, such as India and Southeast Asia, where a large population of women and their babies are susceptible to ZIKV infection [6], still representing a critical health concern since ZIKV transmission can occur in non-epidemic periods and where most of the cases in pregnant women can be asymptomatic but at risk for their concepts [6,7]. Our study sought to understand better the ZIKV outbreak in the State of Rio de Janeiro and its impacts on pregnant women and their babies.

2. Materials and Methods

2.1. Study Design and Setting

This retrospective cohort study reports data from Public Health databases between February 2015 and December 2018, containing information about pregnant women and their concepts during the ZIKV outbreak in Rio de Janeiro, Brazil. We used five databases in our study: GAL, FORMSUS, RESP, SINASC, and SIM. GAL is an online laboratory system that provides the results of Real-time reverse-transcriptase-polymerase-chain-reaction (RT-PCR) for ZIKV tests and some patient characteristics. FORMSUS is an online database that collects, stores, and creates health data reports, including information about pregnant women. RESP is an online notification used for suspected microcephaly or CNS congenital abnormalities during the ZIKV outbreak in Brazil. SINASC contains all the information about the registered newborns in Brazil, and SIM about deaths in Brazil. All the databases were provided by the Healthy Secretary of Rio de Janeiro, and the study protocol received ethical approval from The Code of Ethics of the World Medical Association (Declaration of Helsinki) (Institutional Review Board number 87402618.3.0000.5275).

We first used the GAL database to identify the two cohorts of pregnant women. Eligibility for confirmed ZIKV infection cohort required at least one positive RT-PCR assay for ZIKV in blood or urine during pregnancy. In the non-ZIKV infection cohort, we included pregnant women with all negative tests, all collected up to five days after the onset of the symptoms for blood samples and up to fourteen days for urine samples. The RESP database was reviewed, case by case, to identify any congenital CNS abnormalities and microcephaly cases. This revision excluded cases initially reported but did not meet the criteria for congenital microcephaly (Intergrowth21st method), as different definitions were used during the ZIKV outbreak in Brazil, especially in its beginning. Finally, a probabilistic linkage was performed among the five databases.

We obtained maternal covariates from this final database, including age, race, marital status, education level, residence region, the presence of twin pregnancy, typo of labor, number of medical appointments during pregnancy, and need for hospitalization signs and symptoms described during pregnancy. For the concepts, besides the presence of microcephaly and CNS abnormalities, we also searched for gender, birth length, weight and cephalic perimeter, death, prematurity, other serologic tests than ZIKV, RT-PCR collected in newborns, type of CNS abnormality when present, signs and symptoms in physical examination, and eye and ears anomalies.

2.2. Statistical Analysis

Descriptive statistics were generated for all baseline data; we calculated prevalence rates by dividing the number of microcephaly/CNS congenital abnormalities by the number of pregnant women in both groups. Outcomes are reported overall and according to the trimester of symptom onset. Continuous variables are presented with a median and interquartile range (IQR), and categorical variables are presented with proportions (%).

Also, categorical variables were compared among women and infants with positive and negative ZIKV RT-PCR results using the Chi-square test, Mann-Whitney test, and Fisher’s exact test, as appropriate. Values of p lower than 0.05 were considered statistically significant. Propensity-adjusted analyses control for measured confounding variables were also performed. The covariate adjustment using the propensity score method was used, whereas the outcome variable is regressed on an indicator variable denoting exposure status and the estimated propensity score. As we have a dichotomous outcome, a regression model was utilized. For the logistic model, the exposure effect is an adjusted odds ratio. We used all measured baseline covariates for inclusion in the propensity score model. Values of odds ratio and 95% confidence intervals (CI) were obtained, and values of p lower than 0.05 were also considered statistically significant.

3. Results

3.1. Study Population

From February 2015 to December 2018, we identified 2 269 pregnant women with at least one positive result for ZIKV infection on RT-PCR in blood, urine, or both (ZIKV+), and 5 601 pregnant women with only negative ZIKV infection tests (ZIKV-). Among ZIKV+ women, 758 had positive RT-PCR results in serum specimens, 88 in urine, 270 in serum and urine specimens, and 24 in more than one serum specimen. In those women with positive and negative results, 728 were only positive in serum, and 401 were only in urine. The most common signs and symptoms among ZIKV+ pregnant women were rash, pruritus, headache, arthralgia, myalgia, and fever. Despite that, ZIKV- pregnant women show a predominance of headaches, arthralgia, myalgia, fever, edema and coryza (Table 1).

Of the ZIKV+ pregnant women, 49 cases of congenital microcephaly or CNS abnormalities were identified (2.16%). In the ZIKV- group, 44 cases were identified (0.78%). The prevalence among live births was 193.9 in 10,000 exposed to ZIKV (five fetal losses) and 71.4 in 10,000 not exposed to ZIKV (four fetal losses). The prevalence varied among trimesters when the ZIKV infection occurred: 5.5% in the first trimester, 1.5% in the second, and 0.79% in the third (Table 2).

The baseline characteristics were quite similar between the women with and without ZIKV infection. Most women were aged 21-31, single, lived in the metropolitan area, and had high education. They also had natural labor and at least six medical appointments during pregnancy (Table 3).

3.2. Multivariable Analysis

The odds ratio of congenital microcephaly or CNS abnormalities caused by exposure to ZIKV during pregnancy was 2.46 (95% CI 1.30-4.64). Pregnant women were also stratified in the three different trimesters when the ZIKV infection occurred or was suspected: 4.29 (95% CI 1.93-9.53) in the first trimester, 5.29 (95% CI 1.08-25.95) in the second trimester and 0.68 (95% CI 0.21-2.14) in the third trimester (Table 4).

3.3. Adverse Prenatal and Early Postnatal Infant Outcomes

Among the 93 cases of congenital microcephaly or CNS abnormalities among ZIKV+ and ZIKV- pregnant women, the median maternal age was quite the same in both groups. We did not see statistical differences in maternal race, residence, sex, weight, and length of the concepts. The median and IQR of head circumference were also very similar, as were prematurity, twinning, and fetal loss (Table 5). Maternal symptoms, such as rash and pruritus, were more prevalent in the ZIKV+ group (Table 6).

Thirty-five of the 49 cases of ZIKV+ pregnant women (71.4%) and 20 of the 44 cases of ZIKV- women (45,5%) underwent at least one imaging exam: fetal or cranial ultrasound, cranial computed tomography, or magnetic resonance imaging (MRI). The most common findings among ZIKV+ cases were intracranial calcifications, ventriculomegaly, posterior fossa malformations, reduced brain volume, corpus callosum malformations, cortex dysplasia, lissencephaly, and pachygyria. ZIKV- cases had very similar findings, although ventriculomegaly was more frequent in this group. Arthrogryposis was the most common physical examination finding. Ophthalmologic exams were performed in nine ZIKV+ newborns (55.5% abnormal) and ten ZIKV- newborns (30% abnormal). Brainstem auditory evoked potentials (BAEPs) were reported in nine ZIKV+ cases (33.3% abnormal) and in six ZIKV- cases (16.6% abnormal) (Table 7).

4. Discussion

Our main evidence is that pregnant women infected with ZIKV are at increased risk of fetuses or newborns with microcephaly or CNS congenital abnormalities when compared to pregnant women with no ZIKV infection, with six to seven times the risk when the infection occurs in the first and second trimesters compared to those in the third trimester.

ZIKV infection signs and symptoms among pregnant women in our cohort were very similar to other studies. Garcell et al. [8], analyzing 1,541 patients with clinical suspicion of arbovirosis, and Tozetto-Mendonza et al. [9], 94 patients with ZIKV acute infection confirmed by RT-PCR, observed a majority of rash, pruritus, arthralgia, headache, and myalgia. In published cohorts of pregnant women with confirmed or probable ZIKV infection, we also observe the predominance of rash, pruritus, arthralgia, headache, and myalgia, besides fever, conjunctivitis, and retro-ocular pain [10,11,12,13,14,15,16,17]. Braga et al. [18] developed a predictive score model to differentiate the symptoms of ZIKV acute infection in regions where other arboviruses coexist. These authors found a sensitivity of 86.6% and specificity of 78.3% in those with rash associated with pruritus or conjunctival hyperemia, without fever, petechiae, or anorexia. Besides, ZIKV infection can be asymptomatic in 50 to 73% of cases in the general population [7,19] and about 69% among pregnant women [12]. Therefore, CZS can occur even in those asymptomatic pregnant women. Meneses et al. [20] describe that 24% of children with CZS, with maternal infection confirmed by RT-PCR, had no reports of symptoms during pregnancy. Vianna et al. [21] also report that 17% of children referred for CZS investigation had no history of maternal symptoms during pregnancy.

Microcephaly or CNS congenital abnormalities has a broad differential etiological diagnosis beyond ZIKV and must include other infections such as dengue, chikungunya, toxoplasmosis, cytomegalovirus (CMV), syphilis, human immunodeficiency virus (HIV), parvovirus B19, measles, rubella, chickenpox, herpes virus, Epstein-Barr and enterovirus [22]. Although less common, coinfection should always be considered. We observed a predominance of arthralgia, headache, coryza, edema, fever, and myalgia among pregnant women without ZIKV infection. Dengue typically presents a more exuberant fever associated with rash, headache, and myalgia, while chikungunya is characterized by high fever with rash, polyarthralgia/polyarthritis, myalgia, and edema [23]. Vertical transmission of the dengue virus can occur in pregnant women with viremia at the time of delivery, and it can also be found in breast milk [24,25]. The Chikungunya virus's Neonatal infection can also occur in up to 50% of pregnant women with viremia in the peripartum period [26], ranging from mild to severe cases and lethality in around 2.8% [26,27]. The association between dengue virus and congenital microcephaly or CNS abnormalities is not described. In a case series of 25 brain MRIs performed in congenital infections by the Chikungunya virus, intraparenchymal hemorrhages and white matter abnormalities were described in 14 [28], with no congenital microcephaly. However, postnatal microcephaly is described [29]. Infectious mononucleosis is another possibility, with a higher fetal risk of death and congenital malformations, especially those associated with toxoplasmosis and CMV. Parvovirus B19, rubella, measles, and enteroviruses (mainly Coxsackie A and B) also increase the risk of fetal mortality and congenital malformations [22].

Our cohort's 2.16% prevalence in congenital microcephaly or CNS abnormalities is lower than other studies [30,31]. We must note that (1) only pregnant women with a positive ZIKV RT-PCR were included; (2) the majority of the women were symptomatic (83.7%); therefore, a smaller proportion of asymptomatic women who had potential risk for congenital abnormalities in their fetuses; (3) use of retrospective data; (4) the lack of follow-up data may also underestimate the prevalence since there is a possibility of postnatal signs and symptoms [32,33].

The chance of congenital microcephaly or CNS abnormalities exposed to ZIKV is inversely proportional to the trimester in which the infection occurs [10,34,35,36,37,38]. As seen in other congenital infections, where infections in the first trimester confer a greater risk of congenital CNS injuries, ZIKV infection appears to result in a greater extent of cell loss in both placental and fetal tissues. Histopathological findings in fatal cases in fetuses with microcephaly and a maternal history of ZIKV infection during the first trimester include edema of the villi, an increase in the number of Hofbauer cells, besides the presence of antigens in the chorionic villi and necrotic fetal nerve cells or degeneration process, as well as in glial cells [39,40]. The presence of neuronal necrosis suggests a process of continuous cell injury, which could include the period of maternal infection and subsequent injuries to the developing brain [39]. The loss of nerve cells in the early stages of CNS development can result in less brain volume and impair the formation of cortical gyri [41,42]. A recently published Brazilian cohort showed a greater chance of developmental abnormalities between three and five months of age in pregnant women with ZIKV infection occurring during the first trimester [32]. The lower odds ratio found in the first trimester compared to the second may be related to the higher prevalence of outcomes in negative ZIKV pregnant women, possibly caused by other possible infections, environmental factors, or even abnormalities of genetic origin.

Laboratory confirmation of CZS is more challenging since the duration of viremia and viruria and whether viremia is constant or intermittent is still unknown, especially in those infected in the first trimester of pregnancy. These could explain the significant variation in ZIKV RT-PCR's positivity in different body fluids (blood, urine, and liquor) in other cohorts, ranging from no positive tests to 65% positivity [10,35,38,43,44,45]. The serology of newborns for ZIKV in the postnatal period, confirmed by the PRNT (Plaque Reduction Neutralization Test), seems to have a sensitivity which varies from 7.1 to 90.5% [46,47,48] and can be influenced by the time of its realization, with a considerable reduction after the first month of life [46]. The fetal immune response to ZIKV may be similar to other congenital infections, such as rubella and CMV, which have a more cellular and less humoral immune response [45].

The predominant intracranial calcifications, ventriculomegaly, reduced brain volume, and cortical defects (lissencephaly and pachygyria) are also described by other authors [49,50]. Intracranial calcifications due to CZS are typically located in the subcortical regions. Still, recent reports show other potential sites such as infratentorial, base nuclei, periventricular region, and the cortex itself [51]. Ventriculomegaly was a common finding among our CZS cases but even more prevalent among those born from women without laboratory evidence of ZIKV infection. Only 5% of congenital ventriculomegaly cases are secondary to congenital infections such as CMV, toxoplasmosis, and ZIKV [52]; therefore, the higher frequency may represent a result of other non-infectious etiologies.

The most common ophthalmological findings in CZS are pigmentary abnormalities and chorioretinal atrophy, similar to congenital toxoplasmosis [53]. Approximately 34 to 55% of children with CZS with microcephaly have at least one ophthalmic abnormality [54]. Despite the low number of ophthalmic exams, 55.5% were abnormal: optic nerve hypoplasia, incomplete vascularization, and pigmentary abnormalities. Despite not being the most common finding, Optic nerve hypoplasia has been described by some authors [54,55]. Other etiologies should be considered in those negative cases of ZIKV, such as other congenital infections (CMV and toxoplasmosis) or genetic and metabolic diseases [56,57]. Incomplete vascularization was described in a premature newborn since complete retinal vascularization occurs only at around 40 to 42 weeks of gestation [58].

Regarding auditory abnormalities, we observed that newborns exposed to ZIKV presented a higher frequency of abnormalities than those not exposed (33.3 vs 16.6%). Children exposed to ZIKV show wide variation in its frequency, whereas abnormalities in otoacoustic emissions vary from 0 to 75%, while BAEPs (Brainstem Auditory Evoked Potentials) vary from 0 to 29.2% [59]. Traditionally, other congenital infections (CMV and rubella) and genetic abnormalities are the leading causes of congenital hearing loss [60].

Arthrogryposis has also been described by other authors [61,62,63,64]. Approximately 80% of congenital arthrogryposis have a neurogenic origin, either due to abnormalities in the formation, structure, or function of central and peripheral nervous systems [61]. Imaging and electroneuromyography studies have demonstrated this involvement of both nervous systems, with a volumetric reduction of the anterior medullary tracts in children with CZS, reducing fetal mobility and, consequently, causing deformities [63,65].

Our study's strengths should be highlighted: large population sample, ZIKV infection in pregnant women confirmed by ZIKV RT-PCR, control group, an extensive search for potential covariates, use of multivariate analysis, and comprehensive review of all cases. Some limiting points are: (1) use of retrospective data; (2) lack of information in a part of the sample, including serologic tests for dengue, chikungunya, CMV, toxoplasmosis, rubella, herpes; (3) lack of complementary exams among the identified cases; (5) most of the pregnant women with ZIKV infection were symptomatic; therefore, we are not sure with our results can be extrapolated to all pregnant women (symptomatic and asymptomatic).

5. Conclusions

It is essential to highlight that ZIKV infection can lead to a spectrum of structural or functional anomalies. As there is no specific treatment for CZS, assistance should be focused on monitoring neurological, motor, auditory, visual, and orthopedic disorders in all children intrauterine exposed to ZIKV, regardless of the identification of congenital anomalies in the prenatal period, especially to those exposed during the first and the second trimesters of pregnancy. ZIKV infections are still endemic in some countries globally, including Brazil, reinforcing the preventive measures and continuous monitoring of potential fetal anomalies.