Neurodevelopmental Follow-Up in Children with Intrauterine and Perinatal Exposure to Chikungunya Virus

Abstract

Objective

To investigate the effects of intrauterine and perinatal exposure to chikungunya virus (CHIKV) on neurodevelopment in infants and toddlers.

Study design

We conducted a cohort study comparing children with intrauterine or perinatal exposure to maternal CHIKV infection with unexposed controls in Rio de Janeiro, Brazil. Neurodevelopment was assessed with General Movement Assessments in the first 6 months of life, and the Bayley-III Scales of Infant and Toddler Development and Modified Checklist for Autism in Toddlers for older children. Developmental delay (DD) was defined as a Bayley score less than 70 and risk of DD as a score less than 85.

Results

Among 60 children exposed to intrauterine or perinatal CHIKV, 20 (33%) had laboratory confirmation of CHIKV infection by reverse transcription polymerase chain reaction or immunoglobulin M serology and 40 did not; 44 exposed children (15 infected and 29 uninfected) had General Movement Assessment performed, with 19% having suboptimal or abnormal results. At 11–42 months of age, 35 exposed children and 78 unexposed controls had Bayley-III assessments. Compared with controls, exposed children had higher rates of DD (7 [20%] vs 2 [3%], P = .004) driven by the language domain, and greater risk of DD driven by motor and cognitive domains scores (10 [29%] vs 10 [13%], P = .03 and 8 [23%] vs 5 [6%], P = .02, respectively). Eight of 35 (23%), CHIKV exposed children screened positive for autism spectrum disorder. CHIKV-exposed uninfected children had 2 (9.5%) cases of DD and 5 (23.8%) cases of autism spectrum disorder.

Conclusions

Abnormal neurodevelopmental results were seen in both infected and uninfected children with intrauterine or perinatal CHIKV exposure. Infant neurodevelopment monitoring should be considered following exposure to maternal CHIKV infection in pregnancy to facilitate early interventions and to mitigate neurodevelopmental sequelae.

Chikungunya virus (CHIKV) is an arbovirus transmitted by Aedes mosquitoes and is responsible for annual epidemics in tropical countries; the virus is also capable of intrauterine and perinatal transmission. Vertical transmission of CHIKV was first observed during an outbreak on Reunion Island in 2005, in which there was vertical transmission in approximately one-half of pregnant infected individuals.¹ The disease does not appear to be more severe in pregnant women than in the general population, but is associated with poor pregnancy outcomes, particularly fetal and neonatal morbidity and mortality.² The principal risk factor associated with transmission to the fetus is maternal viremia near delivery.² In a systematic review, fetal pathology occurred in 20% and fetal loss in 2%.³ The pathology in neonates includes encephalitis in approximately one-half of newborns, as well as cerebral edema, rash, and fever. ^1,4–6

Published studies on CHIKV infection in newborns in the Reunion Island, Latin America, and India are mostly case reports with few neurodevelopmental studies. There are only 2 studies that used a validated diagnostic test, but these tests are rarely used in clinical practice today.⁷ A study on Réunion Island, nevertheless, where intrauterine and perinatal CHIKV transmission was first documented, found substantially more infected children (51%) experienced global developmental delay (DD) at 2 years of age compared with uninfected controls.⁷ The consequences and clinical course of CHIKV exposure in infants during the first 2 years of life and potentially beyond, remain poorly understood. Maternal immune activation, induced by CHIKV infection during pregnancy, is known to potentiate a harmful in-utero inflammatory environment. Although its long-term neurodevelopmental impact is still unknown, early studies suggest potential adverse effects on the fetus. Maternal infections during pregnancy can impact fetal brain development severely, increasing the risk of neurodevelopmental disorders such as cerebral palsy, autism spectrum disorder (ASD), and schizophrenia.^8–10 Maternal hyperthermia and inflammation can have severe consequences on fetal brain development, increasing the risk of neurodevelopmental disorders. CHIKV infection can directly invade the central nervous system (CNS) following vertical transmission or indirectly damage the brain through maternal inflammation, exacerbating these risks. In studies of infections during pregnancy, it appears that maternal immune activation can trigger inflammatory responses in infants, leading to adverse clinical outcomes.^11,12 In response to CHIKV exposure, the infant’s immune system also may produce inflammatory cytokines that could potentially be harmful to the endocrine or nervous system.¹³

It is crucially important to study children exposed to CHIKV because of the risk of neuro-DD and the potential of early interventions. Therefore, we evaluated neurodevelopmental performance in young children exposed to CHIKV in Rio de Janeiro, Brazil, using General Movement Assessment (GMA),¹⁴ Bayley Scales of Infant and Toddler Development, third edition,¹⁵ and Modified Checklist for Autism in Toddlers (M-CHAT)¹⁶ for assessment of cognitive, motor, and language domains.

Methods

Study Design

For the purposes of this study, CHIKV infection was defined as a positive reverse transcription polymerase chain reaction (RT-PCR) for CHIKV or anti-CHIKV immunoglobulin M (IgM) in maternal or neonatal serum specimens or both. The eligibility criteria for enrollment in the CHIKV-exposed category was being a neonate in the first 28 days of life who was CHIKV infected or exposed to a mother who was CHIKV infected during pregnancy. CHIKV-infected women were recruited prospectively by following symptomatic pregnant women with CHIKV exposure or through referrals of symptomatic infants.

We prospectively followed CHIKV-exposed children through 3 years of age. In the first month of life, newborns were assessed clinically and had brain imaging studies performed. Subsequently, we performed neurodevelopmental assessments on the CHIKV-exposed infants during 2 age periods: 2–5 months and 11–41 months. Three neurodevelopmental tests were applied: the GMAs in the first period, and the Bayley-III Scales of Infant and Toddler Development and M-CHAT in the second period.

In addition, the study design included a group of healthy children at 11–38 months of age without gestational or neonatal CHIKV exposure, called controls, in whom Bayley-III test was performed for comparison with CHIKV-exposed children.

Study Setting

CHIKV-exposed infants were recruited from March 2016 to August 2020 in the Gynecology/Obstetrics and Neonatal Intensive Care Units at the Gaffré Guinle Hospital of the Federal University of the State of Rio de Janeiro in Rio de Janeiro, Brazil. Controls were recruited from January 2018 to December 2019 at Fernandes Figueira National Institute for Women’s, Children’s, and Adolescent Health in Rio de Janeiro.

Clinical and Laboratory Data on CHIKV-Infected Pregnancies

The following data were collected from the study sample: sociodemographic data, medical and clinical history, comorbidities, and history of clinical course during pregnancy as well as delivery type and clinical management. CHIKV RT-PCR and anti-IgM and IgG testing were performed at a Flavivirus reference laboratory on maternal or infant sera or both.

Participants

Neonates were defined as CHIKV-exposed when they had a mother who was CHIKV infected during pregnancy. Neonates who were infected were defined as those with positive PCR test results. Neonates who were exposed but had negative PCR or IgM test negative were exposed uninfected. Neonatal CHIKV infection by vector transmission was defined as neonates who were CHIKV infected in the first 28 days of life with a mother without CHIKV exposure during pregnancy.

Controls were healthy children from uneventful pregnancies who demonstrated no evidence of congenital infections or genetic disease through serologic evaluations, newborn screening, and frequent clinical examinations. Control children had data on clinical characteristics during the neonatal period (sex, head circumference at birth, and gestational age at delivery) and were test-negative for dengue, Zika, CHIKV, HIV, and TORCH infections (toxoplasmosis, others, syphilis, rubella, cytomegalovirus and herpes). Healthy children were specifically enrolled as controls for our neurodevelopmental studies of ZIKV, CHIKV, and COVID-19. They were not children who normally would warrant neurodevelopmental testing.¹⁷

Clinical Data of CHIKV-Exposed Neonates

The following variables were collected during the study: infant sex, gestational age at birth, birth weight, head circumference at birth, clinical characteristics, age in days at disease onset for CHIKV infected neonates, and adverse neonatal outcomes such as neonatal death or prolonged hospitalization.

Brain Imaging of CHIKV-Exposed Neonates

Imaging of the CNS was performed via transfontanelle ultrasound with Doppler (TFU) during the neonatal period. In addition, a subgroup of participants (as requested by providers) had cerebral magnetic resonance imaging (MRI) performed. MRI was repeated for those with altered findings.

Neurodevelopmental Assessments

GMA.

Between 2 and 5 months of post-term age, we evaluated the integrity of the developing nervous system in CHIKV-exposed participants by analyzing neuromotor development by GMA, which is a validated screening test.^14,18 In brief, infants were videotaped for 2–3 minutes of active wakefulness while lying in a supine position without manipulation. Several endpoints were measured by analyzing the quantity and quality of spontaneously generated movements. These included “fidgety” and “cramped synchronized” movements. Fidgety movements are continuous, low-amplitude, moderate-velocity movements of the shoulder, wrist, hip, and ankle.¹⁹ Absence of fidgety movements is suggestive of neurological deficits. Cramped synchronized movements are rigid gestures involving rapidly contracting and relaxing the muscles of the limbs and trunk.²⁰ Presence of “cramped synchronized” movements is related to neurodevelopmental deficits. In addition, based on movement patterns, a revised motor optimality score (MOS-R) was assigned ranging from 0 to 28. The MOS-R score is predictive of normal vs neuromotor impairment such as cerebral palsy.²¹ A MOS-R score of 25–28 represents smooth and fluent movements whereas a score below 20 represents pathologic changes that could worsen and in infants who could benefit from appropriate intervention such as physical therapy, occupational therapy, and follow up with a pediatric neurologist.^14,21,22

Bayley-III Evaluations.

The Bayley-III test,¹⁵ which is a validated diagnostic test, was chosen as the main neurodevelopmental instrument as it is the gold standard for neurodevelopmental testing in children of this age group. The test includes assessments of the cognitive, language, and motor domains. In each domain, DD was defined as a Bayley score <−2SD (<70); risk of DD (rDD) score <−1SD > −2SD (<85–70). We applied the Bayley test to the CHIKV-exposed children at 11–40 months of age and to controls at 11–38 months of age. When children had one or more Bayley-III assessments, the older age bracket was used in the analysis.

M-CHAT.

Between 18 months and 30 months of gestational age, CHIKV-exposed participants underwent a screening assessment for ASD using the MCHAT-R,^16,23 which is a validated screening test. The instrument was reapplied when children were under 24 months old at the time of the first assessment and/or when the result of the first assessment showed moderate risk. For all items, the answer “no” indicates risk of ASD; except for items considered critical, in which “yes” indicates risk of ASD. Results greater than 3 (“no” in 3 items in total) or in 2 of the items considered critical were included to carry out the M-CHAT-R/F follow up with the aim of better qualifying responses and reducing false positives. The M-CHAT-R/F questions are the same as the initial instrument. At this stage, yes/no answers are converted to pass/fail.²³ Control infants (ie, CHIKV-unexposed) were planned to have M-CHAT performed but because of the SARS-CoV2 pandemic, this was not possible.

Study personnel who performed Bayley-III assessments, GMA, and M-CHAT were aware of antenatal maternal CHIKV exposure but not whether the infection was laboratory confirmed as infected.

Statistical Analysis

We compared rates of preterm delivery between CHIKV-exposed participants and controls with Fisher’s exact test. Head circumference was compared between these groups using a Mann-Whitney test. To analyze the GMA results, we used Fisher’s exact test to compare the proportion of CHIKV-exposed uninfected vs infected participants who exhibited cramped synchronized movements or failed to show fidgety movements. To analyze the Bayley-III results, we again used the Fisher exact test to compare the proportion of participants with DD or rDD in the control, exposed infected, and exposed uninfected groups.

Results

Clinical Data of CHIKV-Infected Pregnancies and Pregnancy Outcomes

The study enrolled 57 infants enrolled due to maternal infection and 3 because of postnatal concern (Figures 1 and S1, and Table S1; available at www.jpeds.com). Among the infants enrolled due to maternal infection, there were 17 with perinatal/intrapartum infection, and 16 had clinical signs at birth. The remaining 40 infants in this group were classified as exposed but uninfected. Nine of these exposed, uninfected infants presented clinical signs in the first weeks of life and the other 31 did not. In the group of infants enrolled due to postnatal concern, 2 had clinical signs at birth. Furthermore, in the group of 3 infants enrolled due to postnatal concern, there were 2 with perinatal/intrauterine infection and one with postnatal infection. Both with perinatal/intrauterine infection had clinical signs at birth. On the other hand, the infant with postnatal infection was asymptomatic at birth but developed clinical signs in the first 4 weeks of life.

Figure 1.

Flowchart of the study participants.

In all infected cases, vertical transmission occurred regardless of mode of delivery. Prematurity occurred in 11.6% of cases. Most newborns were born healthy, developing clinical manifestations in the first week of life, suggesting that vertical transmission occurred during the late intrapartum/peripartum period when mothers were viremic.

No case of spontaneous abortion, stillbirth, or maternal death was observed. The main mode of delivery was through cesarean delivery (n = 31; 53%) performed because of fetal distress in 9 cases (15%). CHIKV-exposure occurred in the first trimester of pregnancy in 15% (n = 9) of participants, in the second trimester in 13.5% (n = 8), in the third trimester 68.5% (n = 41).

Neonatal Findings of CHIKV-Exposed Participants

The 59 CHIKV-exposed pregnancies plus the infant infected postnatally by a mosquito vector resulted in a total of 60 neonates with CHIKV exposure. Among the 60 CHIKV-exposed neonates, 33.3% (n = 20) neonates had laboratory confirmed CHIKV (16 RT- PCR positive, 4 IgM positive) infection, while 40 had RT-PCR, IgM and IgG negative (66%). At birth, 18 infants had symptoms of intrauterine/perinatal infections.

The rate of preterm delivery (<37 weeks’ gestational age) and low birth weight were both (11%, n = 7). Median (IQR) head circumference was 34 (33–36) cm. There were 27 infants with symptoms of intrauterine or perinatal infection (Fig. S1; available at www.jpeds.com).

The most frequent symptoms were hypoactivity and weak sucking (15/60, 25%), fever (13/60, 22%), seizures and encephalitis (9/60, 15%), rash (11/60, 18%), and respiratory distress (11/60, 18%). Less frequent manifestations included: irritability (12%, n = 7), hyperchromia, a pigmentary change including “CHIK sign” (hyperpigmentation in the nose), were found to be the most common cutaneous finding (12%, n = 7), vesiculobullous lesions (12%, n = 7), apnea (8%, n = 5), and low weight gain (5%, n = 3). CHIKV infected children compared with exposed-uninfected neonates compared with exposed-uninfected neonates had significantly higher rates of apnea, fever, hypoactivity, irritability, rash, respiratory distress, seizures, and vesiculobullous lesions than uninfected children (Tables S2 and S3; available at www.jpeds.com).

There were 3 participants whose mothers were asymptomatic but were referred because the neonate had symptoms; 3 neonates had laboratory confirmation of CHIKV infection. Two other mothers also had laboratory confirmation of CHIKV infection, while one mother was negative. This child was classified as acquiring CHIKV infection postnatally through vector transmission.

Neonatal Findings of Control Participants

None of the 78 control subjects was born preterm. Median (IQR) head circumference was 35 (34–36) cm. Head circumference difference did not differ significantly between the controls and CHIKV-exposed groups (P = .2). Rates of preterm delivery were lower in controls than the CHIKV-exposed group (P = .02) (Table S2; available at www.jpeds.com).

Brain Imaging of CHIKV-Exposed Participants

TFU was performed on 55 participants (92%); 5 children were lost to follow up. Three participants (5.5%) had abnormal findings with indirect signs of hypoxic ischemic encephalopathy. MRI was performed on 13 newborns, and was abnormal in 5 cases (38.5%); 3 of these infants also had altered TFU (Figs S2–S4; available at www.jpeds.com). No asymptomatic participants had MRI performed. MRI performed during the acute phase of the disease (within 2 weeks of the positive PCR) showed changes consistent with restricted diffusion in the subcortical white matter of both hemispheres and the corpus callosum. In the subacute or chronic phase (more than 2 weeks after the positive PCR), imaging revealed areas of cystic cavitation in the frontal lobe with a perivascular distribution. Follow-up brain imaging in 3 participants revealed a reduction in the volume of cavitations and resolution of restricted diffusion.

GMA Assessment of CHIKV-Exposed Participants

A total of 44 infants were evaluated with GMAs. During the first 2 months of life, 19 (43%) had GMA performed. From 3 to 5 months of age, 36 (82%) had GMA performed, some of whom had also been tested in the first 2 months of life (Table I). The remaining children were lost to follow up. Sixteen of 19 children tested before 2 months of age (84%) had poor repertoire GMAs with a monotonous sequence of movements in the GMOS-R assessment. The MOS-R evaluation showed that only 6 of 36 (17%) had an optimal score; and 7 infants (19%) had a score below 20, which represents a score that could worsen, while 64% (n = 23) were in a range still considered normal, but had a reduced score potentially associated with a neurologic condition and were referred for further monitoring. Infants exposed to CHIKV showed changes in their movement repertoire, which were not adequate for their age, as well as abnormalities in posture and movement characteristics. Two of 36 children (5%) did not present “fidgety” movements during the assessment and 3 (8%) children had characteristic movements such as “cramped synchronized”, both related to neurodevelopmental deficits. Only 5 of 36 (14%) children showed normal movement characteristics, described as smooth and fluent (Table II). About 47% (n = 7) of children with laboratory confirmed CHIKV infection had abnormal MOS-R results vs 83.5% (n = 20) of exposed uninfected participants. Lack of fidgety movements was significantly more frequent in CHIKV-infected participants than CHIKV-exposed uninfected participants (P = .004). In addition, cramped synchronized movements were more frequent in CHIKV-infected participants than CHIKV-exposed, uninfected participants (P = .007).

Table I.

Motor behavior at 2–5 months of age by General Movement Assessment in CHIKV-exposed infected and CHIKV-exposed uninfected infants

Variables	CHIKV exposed* n = 44 (73%)	CHIKV exposed infected* n = 15 (34%)	Exposed not infected n = 29 (66%)
Categorical score performed	19 (43%)	6 (31·5%)	13 (68·5%)
Normal	3 (16%)	1 (16%)	2 (15%)
Poor repertoire	16 (84%)	5 (84%)	11 (85%)
Cramped synchronized	0	0	0
Chaotic	0	0	0
Sequence of movements performed	19 (43%)	6 (31·5%)	13 (68·5%)
Variable	3 (16%)	1 (16%)	2 (15%)
Monotonous	16 (84%)	5 (84%)	11 (85%)
Cramped synchronized	0	0	0
Chaotic	0	0	0
Subscore neck and trunk mean (IQR)	2 (0–4)	1 (1–4)	0·7 (0–4)
Subscore upper extremity mean (IQR)	9 (4–15)	9·5 (4–14)	7·8 (4–15)
Subscore lower extremity mean (IQR)	9 (5–16)	9 (6–13)	9 (5–16)
General movement optimality score-revised not evaluated	25 (57%)	9 (36%)	16 (64%)
Movement optimality score performed	36 (82%)	12 (33.5%)	24 (66.5%)
Optimal range 22–25	6 (17%)	2 (17%)	4 (16.5%)
Movement optimality score-Revised ≤24	23 (64%)	5 (41.5%)	18 (75%)
Movement optimality score-Revised <20	7 (19%)	5 (41.5%)	2 (8.5%)
“Fidgety” evaluated	36 (82%)	12 (33.5%)	24 (66.5%)
Normal	34 (95%)	10 (83.5%)	24 (100%)
Absent	2 (5%)	2 (16.5%)	0
Movement optimality score not evaluated	8 (18%)	3 (37.5%)	5 (62.5%)
Repertoire	36 (82%)	12 (33.5%)	24 (66.5%)
Age adequate	14 (39%)	4 (33.5%)	10 (41.5%)
Not age adequate	22 (61%)	8 (66.5%)	14 (58.5%)
Postural patterns	26 (82%)	12 (33.5%)	24 (66.5%)
More normal postures than atypical postures	5 (14%)	1 (8%)	4 (16.5%)
Equal number of normal and atypical postures	20 (56%)	7 (58%)	13 (54.3%)
Fewer normal postures than atypical postures	11 (30%)	4 (34%)	7 (29.2%)
Movement patterns	36 (82%)	12 (33.5%)	24 (66.5%)
More normal postures than atypical postures	32 (89%)	9 (75%)	23 (96%)
Movement characteristics	36 (82%)	12 (33·5%)	24 (62·5%)
Smooth and fluent	5 (14%)	2 (13%)	4 (14%)
Abnormal, but not cramped synchronized	28 (78%)	7 (47%)	20 (69%)
Cramped-synchronized	3 (8%)	3 (20%)	0
Movement optimality score not evaluated	8 (18%)	3 (37·5%)	5 (62·5%)

^*Children were enrolled in the exposed category if they and/or their mothers had a positive RT-PCR for CHIKV and/or anti-CHIKV IgM.

Table II.

Distribution of Bayley-III results in children exposed to maternal CHIKV and 78 unexposed controls

Variables	Exposed (n = 35)	Exposed infected (n = 16)	Exposed uninfected (n = 21)	Controls (n = 78)	Exposed v. controls P	Exposed infected v. control P	Exposed uninfected v. control P	Exposed infected v. exposed uninfected P
Developmentally delayed (< −2 SD), n (%)
All categories	7 (20%)	5 (31%)	2 (10%)	2 (3%)	.004	.03	<.001	.03
Cognition	1 (3%)	1 (6%)	0	0	.31	<.001	<.001	<.001
Language	7 (20%)	5 (31%)	2 (10%)	2 (3%)	.004	.03	<.001	.03
Motor	1 (3%)	1 (6%)	0	0	.31	<.001	<.001	<.001
At risk for developmental delay (< −1 and > −2 SD), n (%)
All categories	14 (40%)	7 (44%)	7 (33%)	19 (24%)	.11	.09	.49	.8
Cognition	8 (23%)	6 (38%)	2 (10%)	5 (6%)	.06	.3	.008	.04
Language	12 (34%)	7 (44%)	5 (24%)	16 (21%)	.17	.23	.99	.42
Motor	10 (29%)	4 (25%)	6 (29%)	10 (13%)	.03	.83	.42	.24

Bold indicates P < .05.

Bayley-III Evaluations of CHIKV-Exposed and Control Infants

In total, 35 CHIKV-exposed children underwent Bayley-III assessments between 11 to 41 months. The remaining children were lost to follow up. Seventy-eight control children underwent the same testing between 11 to 38 months of age. Bayley-III median composite scores for cognitive, language and motor domains in cases and controls, respectively, were 95 and 98, P = .32, 83 and 89, P = .04 and 88 and 94, P = .03 (Figure 2). The figure shows the distribution of Bayley-III scores for cognitive, motor, and language domains.

Figure 2.

Results of Bayley-III assessments for children exposed to maternal CHIKV (n = 35) compared with unexposed control children (n = 78). Total = 113. A, CHIKV exposed, 11–41 months, N = 35. B, Controls, 11–38 months, N = 78.

Children exposed to CHIKV more frequently had DD (<−2 SD) and risk of DD (<−1SD > −2SD) compared with unexposed controls (Figure 3). In the exposure group, 7 of 35 children (20%) were delayed, as opposed to 2 of 78 children (3%) in the control group, P = .004 (Table II). The lower scores were driven primarily by the language domain (n = 7, 20%) vs n = 2, 3%, in the unexposed group, P = .004. A higher percentage of children (n = 14, 40%) were at rDD in the exposure group, but this frequency was not statistically different than that observed in the unexposed group (n = 19; 24%), P = .12 (Table III). Exposed children were also significantly more likely to have lower cognitive and motor scores in the risk of DD category than unexposed children (Table II).

Figure 3.

Distribution of results Bayley-III assessments for children exposed to maternal CHIKV compared with unexposed controls. A, CHIKV exposed, 11–41 months, N = 35. B, Controls, 11–38 months, N = 78.

Table III.

Clinical characteristics and assessment of the neurodevelopment of children with ASD

ASD	Timing of maternal CHIK	Trimester	BW	Sex	Apgar	RT-PCR/IgM CHIK	Neonatal findings	MOS	Bayley	M-CHAT
1	40 W	2nd	3895	F	9/9	−/−	Asymptomatic	NA	DD	High risk
2	38 W	3rd	3125	F	7/8	+/−	Hypoactivity, glaucoma	NA	DD	High risk
3	38 W	Asymptomatic	3715	M	9/9	+/−	Fever, rash	<20	DD	High risk
4	33 W 6 d	Postpartum	1885	M	7/9	−/−	Respiratory failure	24	DD	High risk
5	37 W 3 d	Peripartum	3290	M	6/8	+/+	Thrombocytopenia hypoactivity	<20	DD	High risk
6	38 W	3rd	3330	M	9/9	−/−	Asymptomatic	NA	DD	High risk
7	39 W 5 d	1st	3450	M	9/10	−/−	Asymptomatic	NA	DD	High risk
8	37 W 6 d	Peripartum	3160	F	99	−/−	Respiratory failure hypoactivity	28	DD	High risk

GA, gestational age.

Of the 35 exposed children assessed by Bayley-III, 16 (46%) had laboratory confirmed CHIKV infection. These 16 infants comprised 76% of all infants with laboratory confirmed CHIKV infection (n = 21).

Of the infected newborns assessed, 8 (50%) were identified as being at risk of or having DD. Among 19 exposed uninfected newborns, the Bayley-III exam identified 10 (52.6%) at risk for or with DD. Rates of DD were significantly higher among CHIKV-exposed infected infants than among those who were exposed but uninfected infants (Table II). Among the exposed uninfected children, who were symptomatic there were 6 with Bayley scores, 3 of which were at rDD or DD (50%), among the exposed infected children who were asymptomatic, there were 14 with Bayley scores, 7 of which were rDD or DD (57%).

We also divided the children with CHIKV exposure into subgroups and compared them with the controls. The subgroups were children with perinatal or intrauterine infection who had no clinical signs at birth and children who were exposed but uninfected. As the subgroups were smaller than the full CHIKV exposure group, the Bayley scores had wider confidence intervals. Nevertheless, the findings were suggestive of lower language domain scores for infected asymptomatic children and exposed uninfected children, compared with controls (Figure S5; available at www.jpeds.com).

M-CHAT of CHIKV-Exposed Participants

A total of 35 children (58%) were evaluated with M-CHAT, 23% (n = 8) tested positive for ASD; 5 were CHIKV-exposed uninfected and 3 were exposed infected; of 8 children with ASD, 3 (38%) were asymptomatic at birth.

All 8 children flagged for autism by M-CHAT screening had DDon Bayley-III testing (score <70; −2SD). Four of these children had GMAs performed, with only 1 scoring in the optimal range, who was also the only 1 out of 4 exposed uninfected infants (Table III). Twenty children (33%) had both GMA assessments and Bayley-III evaluations, which allowed us to evaluate the impact of intervention therapies on improving neurodevelopment. Six children (30%) had a Bayley-III diagnosis of neuro-DD confirmed, with the most severe cases being related to the absence of fidgety movements (n = 2; 5%). Of the 11 children (55%) with normal Bayley assessments, only one (5%) presented optimal movement patterns in infancy. The others had suboptimal movements and were referred to specialists for early stimulation.

Discussion

The strength of our study is the evaluation of outcomes of children with antenatal CHIKV exposure using Bayley-III as the primary neurodevelopmental tool, which is considered the “gold standard” in the age group studied. Moreover, GMA allowed us to identify aberrant movement patterns early on, demonstrating GMA as a potentially valuable tool for predicting adverse neurodevelopmental outcomes in infants with intrauterine or perinatal exposure to CHIKV. We found that both CHIKV-exposed infected and CHIKV-exposed uninfected infants, and both asymptomatic as well as symptomatic infants at birth, had substantial neurodevelopmental sequelae.

CHIKV-exposed children exhibited poorer neurocognitive skills, compared with control groups of CHIKV-uninfected and -unexposed children. We demonstrated that the most affected domain among exposed infants was language development. In line with our results, the Reunion Island cohort study in 2006^1,7 also showed poorer neurocognitive skills in CHIKV-infected infants as evidenced by lower global DD scores and diminished functions specific for language and coordination. Unlike previous studies that limited developmental concerns to infected children,⁷ our data show that CHIKV-exposed uninfected children also are at risk for DD. In the present analysis, some children who were exposed to CHIKV but who did not have laboratory evidence of infection had neuro-DD. In such cases, the possibility of adverse inflammatory effects of maternal infection as well as false negative laboratory test results should be considered. In light of this possibility, keen neurodevelopmental monitoring of all exposed children should be performed over time.

Two noteworthy cases emerged in our study. Two children (5%) evaluated with GMA at 3–5 months of age did not exhibit normal “fidgety movements.” One child subsequently developed cerebral palsy, while the other screened positive for ASD. These cases highlight GMA’s potential as a possible tool for predicting neurodevelopmental abnormalities in infants with intrauterine or perinatal exposure to CHIKV. Similar observations have been noted in ZIKV-exposed infants,¹⁹ suggesting a broader applicability of GMA in identifying newborns at risk. Early detection would facilitate earlier intervention therapies, potentially improving long-term outcomes. In addition, of 19 CHIKV-exposed children who had GMA performed in the first 2 months of life, we observed that 16/19 (84%) exhibited monotonous movements during the neonatal period. These findings support previous research linking abnormal movement patterns in infancy with cognitive deficits at school age.¹⁹ In order to proactively address these concerns, all children in our study have begun interventional therapies to potentially mitigate future cognitive difficulties.

Neurologic manifestations are consistent with neuroradiologic findings of white matter injury and cystic cavitation observed in the cerebral frontal lobes of infants with complete follow up for CHIKV neonatal encephalopathy, a brain region crucial for coordination and language.^1,7 MRI can provide important information about prognosis and neurodevelopmental outcomes.²⁴ In our study, neuroimaging was performed only in children who had clinical neurologic manifestations and laboratory evidence of CHIKV infection, due to the difficulty in accessing MRI in the public health service. Regardless, monitoring all CHIKV-exposed babies with neurodevelopmental assessments to detect below average performance and promote timely intervention is important.

We found similarities between the clinical characteristics of our neonates and those reported previously.²⁵ Notably, the predominant neonatal clinical manifestations were hypoactivity, irritability, difficulty sucking, apnea and respiratory failure. Most newborns manifested fever, rash, vesicobullous lesions, nasal hyperpigmentation, and neurologic manifestations. Of 26 infants who manifested neurologic symptoms, all were CHIKV-infected and seizures were most common, likely due to encephalitis and intracranial hemorrhage. All of these infants required anticonvulsant medication.

The pathophysiology of CHIKV infection in the CNS still requires further study to determine whether brain involvement is due to direct action of the virus, indirect action due to maternal or infant’s inflammatory activation, or a combination of both.^8,10 A recent study from Brazil confirmed direct invasion of CHIKV into the CNS with demonstration of viral RNA in the CSF and in brains from fatal cases, with the presence of intracerebral mononuclear cell infiltration.²⁴ Additionally, substantial overproduction of proinflammatory cytokines and high level of IL-10 (anti-inflammatory cytokine) were reported.²⁶ Elevated levels of cytokines in the blood could contribute to uncontrolled inflammation and loss of homeostasis of the immune response. CHIKV infection is known to impair the integrity of the blood-brain barrier.⁸

Fever is an important indicator of CHIKV infection in pregnant women. Fever itself might trigger, or be associated with triggers for cognitive dysfunction in offspring and other neurologic outcomes, such as neural tube defect, seizures, cerebral palsy, epilepsy, autism, and schizophrenia.^8–10 Maternal hyperthermia and pro-inflammatory cytokines have been consistently demonstrated in animal models to affect apoptosis, neuronal migration and down-regulation of brain-specific genes involved in behavioral processes.^9,10 It is noteworthy that the risk of maternal immune activation is not limited to neurodevelopmental psychiatric conditions such as schizophrenia or autism but also extends to attention-deficit hyperactivity disorder.^10,24

The limitations of our study were the impossibility of performing all planned developmental tests in CHIKV-infected infants due to loss of participant follow up during the COVID-19 pandemic; there also was loss of follow up of CHIKV-exposed children who were asymptomatic at birth; and the difficulty in performing ASD screening evaluation in the control group. We showed ahigh likelihood of ASD in CHIKV-exposed infants using a screening tool with high sensitivity and high specificity for ASD. Future comparison with a CHIKV-unexposed control group using similar testing will be important to establish relative risk. Children with a positive screening result for ASD also need further follow up for a formal diagnosis of ASD to be made. Lastly, the sample size of our study was small.

The follow up of these children, essential to evaluate their developmental trajectories in the physical, cognitive, socioemotional, and language domains, has been performed since their entry in kindergarten. Brazil has made significant strides in providing care for children with global DD having established 285 such centers for children and adolescents in the public healthcare system.²⁵ Nevertheless, to meet the national goal of establishing such centers in all municipalities with more than 70 000 inhabitants, approximately 500 centers would be required. Consequently, many children do not have access to the specialized therapies to optimize reaching their full potential.

Our results are an alert to health authorities to provide conditions for intensive clinical support to mothers and infants in the acute phase of CHIKV infection and multidisciplinary care support throughout childhood for children with intrauterine or perinatal CHIKV exposure.

Glossary

ASD

Autism spectrum disorder

CHIKV

Chikungunya virus

CNS

Central nervous system

DD

Developmental delay

GMA

General Movement Assessment

IgM

Immunoglobulin M

M-CHAT

Modified Checklist for Autism in Toddlers

MOS-R

Revised motor optimality score

MRI

Magnetic resonance imaging

rDD

Risk of developmental delay

RT-PCR

Reverse transcription polymerase chain reaction

TFU

Transfontanelle ultrasound with Doppler