Objective Investigating the clinical implications of isolated white matter abnormalities on neonatal brain MRI in congenital cytomegalovirus (CMV).
Design Prospective, observational.
Patients/interventions Two paediatric radiologists, blinded to clinical data, independently scored the white matter in 286 newborns with congenital CMV. After assessing interobserver variability, mean score was used to categorise white matter (normal, doubtful or abnormal). Patients with other brain abnormalities were excluded.
Main outcome measures Hearing and neuromotor evaluation.
Results Cohen’s weighted kappa was 0.79 (95% CI 0.73 to 0.84). White matter was normal in 121 patients, doubtful in 62, abnormal in 28. Median clinical follow-up was 12.0 months (IQR 12.0–27.7 months). Neonatal hearing loss occurred in 4/27 patients (14.8%) with abnormal, 1/118 patients (0.8%) with normal and 1/62 patients (1.6%) with doubtful white matter (p<0.01). Impaired cognitive development was seen in 3/27 patients (11.1%) with abnormal, 3/114 patients (2.6%) with normal and 1/59 patients (1.7%) with doubtful white matter (p=0.104). Alberta Infant Motor Scale (AIMS) was below P75 in 21/26 patients (80.8%) with abnormal, 73/114 patients (64.0%) with normal and 36/57 patients (63.2%) with doubtful white matter (p=0.231). In a subgroup of patients with minimal clinical follow-up of 18 months, AIMS score was below P75 in 10/13 patients (76.9%) with abnormal, 13/34 patients (38.2%) with normal and 7/20 patients (35.0%) with doubtful white matter (p<0.05).
Conclusions Abnormal white matter was associated with neonatal hearing loss and mild, lower motor scores. A tendency towards impaired cognitive development was seen. Patients with doubtful white matter did not show worse clinical outcome.
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
WHAT IS ALREADY KNOWN ON THIS TOPIC
White matter signal abnormalities are one of the most characteristic findings on brain MRI in congenital cytomegalovirus infection, but the prognostic role, especially when detected as an isolated abnormality, remains unclear.
WHAT THIS STUDY ADDS
Isolated abnormal white matter was associated with hearing loss at birth and mildly lower motor scores at follow-up and could also imply a higher risk of impaired cognitive development. Patients with discrete or doubtful white matter signal alterations had similar neuromotor outcome compared to patients with normal brain MRI.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
MRI with white matter scoring is indicated in newborns with and without clinical symptoms of cytomegalovirus. Although larger and longer follow-up studies are needed, this study can help refine therapeutic guidelines and prognostic counseling for parents.
Congenital cytomegalovirus (cCMV) infection is the leading congenital infection worldwide, with a birth prevalence rate ranging from 0.2 to 2%. Infants born with CMV have a significant risk of developing long-term impairments, such as sensorineural hearing loss and neurodevelopmental abnormalities.1–3
At birth, only 10% of children will show clinical signs of infection. Diagnosis in asymptomatic patients is often the result of screening for maternal seroconversion during pregnancy. Symptomatic newborns have a high risk of permanent sequelae (40%–58%); however, important impairments are also found in 13% of initially asymptomatic children.4–10
cCMV infection is responsible for a wide spectrum of brain anomalies, the type of lesion being dependent on the gestational age at the time of infection.11 12 Since infection can persist over a relatively long period and different parts of the brain mature at different gestational ages, combinations of malformative and destructive lesions are often seen.13 14 White matter abnormalities are one of the most characteristic findings of cCMV, occurring in up to 75% of clinically symptomatic and 30% of asymptomatic newborns. These lesions are often seen in association with other brain abnormalities, however, they are also frequently observed as an isolated finding.15 16
According to the European consensus, patients with neuroimaging consistent with cCMV are considered ‘severely’ symptomatic, and 6 months of antiviral therapy is recommended.17 Neonatal MRI has shown a lot of potential in addition to ultrasound in detecting cCMV-related lesions, however, there still is no European consensus for its role in screening infected newborns, especially in the absence of clinical symptoms.14 16–18
While some brain abnormalities, like cortical malformations and hippocampal dysplasia, seem strong predictors of poor neurological outcome, the prognostic role of others, such as white matter abnormalities on MRI or lenticulostriate vasculopathy on ultrasound, are less well known.19 20
Unlike for lenticulostriate vasculopathy, patients with ‘white matter change’ are considered severely symptomatic. MRI is better than ultrasound in detecting these lesions, but the prognostic impact, especially when detected as an isolated abnormality, remains unclear. Reported studies are often small and mainly limited to severely affected children.11 14 15 17–19 21–25 To our knowledge, no previous studies focused on white matter abnormalities alone.
The aim of this study was to investigate whether isolated white matter abnormalities—detected on neonatal brain MRI—were associated with neurodevelopmental issues in children with cCMV.
Materials and methods
Between April 2007 and March 2020, 286 patients with confirmed diagnosis of cCMV received a brain MRI at the Ghent University Hospital (‘Centre One’ of the Flemish Congenital Cytomegalovirus Infection registry). Confirmation of diagnosis in the neonatal period was made by viral isolation and/or PCR on urine or saliva within the first 3 weeks of life. Retrospective diagnosis (after age of 21 days) was made by PCR on dried blood spots.
At enrolment, the following data were collected: reason for diagnostic workup, clinical features at birth, hearing, ophthalmological examination and brain MRI.
Brain MRI was performed with a headcoil or neonatal headcoil in a 1.5T scanner (Avanto/Symphony/Aera, Siemens Healthcare). Sequences included: 3 mm sagittal T1-weighted spin echo, 3 mm axial T2-weighted fat saturated turbo spin echo, 3 mm axial T1-weighted inversion recovery turbo spin echo, 3–4 mm axial T2-weighted fluid-attenuated inversion recovery (FLAIR) turbo spin echo, 3–4 mm axial diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping, 3 mm axial T2* weighted gradient echo (T2*) and 3 mm coronal T2-weighted turbo spin echo. Detailed information is provided in online supplemental table E1.
Images were separately reviewed by two paediatric radiologists (17 years and 10 years of experience), blinded to clinical and other imaging findings.
MRIs were scored for the presence of white matter abnormalities. White matter abnormalities were defined as ‘increased signal of the white matter on T2-weighted imaging and DWI (ADC) with corresponding decreased signal on T1 and FLAIR images’.
A five-point measurement scale was developed to score the signal of the white matter in each patient:
0: No signal abnormalities.
1: Minimal signal alterations, probably normal.
2: Doubtful signal.
3: Moderate signal alterations, probably abnormal.
4: Severe signal alterations, abnormal.
Five-point measurement was performed in various locations of the brain (frontal superficial, deep frontal, parietal superficial, deep parietal, occipital, anterior temporal and periventricular white matter) before giving the final score. Both severity of focal signal alterations and extension of multifocal lesions were considered in the decision process.
After assessing interobserver variability between the two radiologists, a simplified 3-point score was calculated using the mean score of the individual measurements.
0: Normal white matter (mean score 0–1) (figure 1).
1: Doubtful white matter (mean score 1.5–2.5) (figure 2).
2: Abnormal white matter (mean score 3–4) (figure 3).
Associated brain abnormalities (cortical malformations, cerebellar anomalies, cerebral calcifications, ventricular dilatation, ventricular adhesions and subependymal cysts) were noted as well.
Patients with comorbidities that could interfere with imaging findings (preterm birth, syndromes with known brain abnormalities), MRI not performed in the first 45 days of life, insufficient image quality or when MRI showed other abnormalities than white matter lesions (cortical malformations, cerebellar anomalies, cerebral calcifications, ventricular dilatation, ventricular adhesions or subependymal cysts).
Definition of ‘clinically’ symptomatic cCMV at birth
We did not use the full definition of symptomatic cCMV infection suggested by the European expert consensus statement on diagnosis and management in cCMV,17 because brain imaging findings were part of this definition and this would interfere with the purpose of the study (circular reasoning).
Instead, patients were defined ‘clinically symptomatic’ when the confirmed diagnosis of cCMV was associated with one or more of the following features: clinically detectable symptoms/signs at physical examination, abnormal laboratory results (blood test), abnormal hearing, abnormal ophthalmological testing. Definition of these abnormalities was largely in accordance to the European consensus statement.17 Details are listed in online supplemental table E2.
Clinical follow -up
Clinical follow-up consisted of hearing evaluation, vision assessment by funduscopy and neuromotor evaluation at the centre for developmental disorders. Periodicity of these follow-up visits was conform the guidelines of the Flemish consensus 2018.8 Details are provided in online supplemental table E3.
Neonatal hearing evaluation was carried out using automated auditory brainstem response (automated ABR). In case of refer, diagnostic click-evoked ABR was performed within the first month after birth, combined with otomicroscopy and tympanometry to rule out otitis media. Neonatal ABR thresholds ≤40 dB nHL were considered normal.
Cognitive development was assessed using the Bayley scales of Infant and Toddler Development Third Edition. Delayed cognitive development was categorised according to the definition of the test: normal (Bayley above 86), mild (70–85), moderate (50–69), severe impairment (below 50).
Motor development was assessed using the Alberta Infant Motor Scale (AIMS) during the first 18 months of life, using the Bayley scales Third edition between the age of 18 and 42 months and using the Movement Assessment Battery for Children after the age of 42 months. Based on the AIMS-percentile ranks, motor development was categorised into the following groups: above P75, between P25 and 75, between P10 and P25, below P10.
Categorical variables were expressed in number and percentage, numerical variables in median and IQR.
Cohen’s weighted kappa analysis with quadratic weights was used to assess interobserver variability.
The following kappa classification was used for interpretation of agreement: <0.20: poor, 0.21–0.40: fair, 0.41–0.6: moderate, 0.61–0.8: good, 0.81–1.00: very good.
Last observation carried forward analysis was used to assess outcome measures.
Tests of independence were performed using Pearson χ2 and Fisher’s exact test. Binary and multinomial regression analyses were performed. A significance level of 0.05 was accepted.
Analyses were performed in SPSS (V.26.0).
Patient and public involvement
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
Brain MRI was performed in 286 newborns with cCMV. Seventy-five patients were excluded from further evaluation: 6 due to comorbidities (Möbius (1), Down syndrome (1), tuberous sclerosis (1), prematurity (3)), 1 because of seizures during MRI, 3 because MRI was performed after the age of 45 days. Sixty-five patients were excluded because MRI showed other brain abnormalities than white matter lesions. Neonatal brain MRI findings of these 65 patients are provided in online supplemental table E4.
Of the remaining 211 included patients, 101 were boys (47.9%), 110 girls (52.1%). Median gestational age was 39.0 weeks (IQR 38.0–39.0 weeks, missing in 13 patients). Median age at MRI was 21.0 days (IQR 13.0–27.0 days).
Neonatal brain MRI
Results of the 5-point scoring system, given by the individual reviewers, are listed in table 1. Cohen’s weighted kappa value for the scoring of white matter lesions was 0.79 (SE 0.03, 95% CI 0.73 to 0.84, p<0.001).
According to the simplified, 3-point scoring system, white matter was normal in 121 patients (57.3%), doubtful in 62 (29.4%) and abnormal in 28 patients (13.3%).
Median clinical follow-up was 12.0 months (IQR 12–27.7 months). Detailed information on reasons for cCMV testing, clinical findings at birth and during follow-up are provided in table 2.
At birth, 13/207 patients (6.3%) were ‘clinically’ symptomatic: 5/27 patients (18.5%) with abnormal white matter, compared with 4/118 patients (3.4%) with normal and 4/62 patients (6.5%) with doubtful white matter (Fisher’s exact p=0.025).
Clinically detectable symptoms/signs at birth (physical examination) were present in 5/207 patients (2.4%): in 1/27 patients (3.7%) with abnormal, 2/118 patients (1.7%) with normal and 2/62 patients (3.2%) with doubtful white matter (Fisher’s exact p=0.523).
Figure 4 shows the clustered bar charts of white matter scores on neonatal brain MRI by several clinical findings, both at birth and during clinical follow-up. Results of the regression analyses are displayed in table 3.
Neonatal hearing was abnormal in 6/207 patients (2.8%). In two of these patients, hearing returned to normal during follow-up. At follow-up, hearing was abnormal in 13/202 patients (6.2%): in 4 patients, this was due to persistence of neonatally detected hearing loss, while in 9 patients, late-onset hearing loss had occurred.
Neonatal hearing loss was seen in 4/27 patients (14.8%) with abnormal white matter, compared with 1/118 patients (0.8%) with normal and 1/62 patients (1.6%) with doubtful white matter (Fisher’s exact p=0.005). All nine patients with late-onset hearing loss had normal white matter on neonatal brain MRI.
Impaired cognitive development (Bayley score below 86) was seen in 7/200 patients (3.5%). This was noted in 3/27 patients (11.1%) with abnormal, 3/114 patients (2.6%) with normal and 1/59 patients (1.7%) with doubtful white matter (Fisher’s exact p=0.104).
Concerning motor development, AIMS score was above P75 in 66 patients (31.5%), between P25 and P75 in 80 patients (40.6%), between P10 and P25 in 33 (16.8%) and below P10 in 18 (9.1%).
AIMS score below P75 was seen in 21/26 patients (80.8%) with abnormal, 73/114 patients (64.0%) with normal and 36/57 patients (63.2%) with doubtful white matter (χ2 p=0.231).
Longer clinical follow-up
Sixty-seven patients had clinical follow-up of at least 18 months. Motor development was further observed in this subgroup of patients (figure 5). In this group, AIMS score was above P75 in 37 patients (55.2%), between P25 and P75 in 17 (25.4%), below P25 in 9 (13.4%) and below P10 in 4 patients (6.0%).
AIMS below P75 was seen in 10/13 patients (76.9%) with abnormal white matter, compared with 13/34 patients (38.2%) with normal and 7/20 patients (35.0%) with doubtful white matter (χ2 p=0.033).
This cohort study prospectively examined the neurodevelopmental outcome in children with isolated white matter lesions. In this study, grading of the white matter was feasible with good interobserver agreement. At birth, abnormal white matter was associated with neonatal hearing loss. During follow-up, more patients with abnormal white matter had impaired cognitive development, although not statistically significant. Patients with abnormal white matter had mildly lower motor scores, but only statistically significant in a subgroup of patients with minimal clinical follow-up of 18 months. Patients with discrete or doubtful white matter signal alterations did not show worse hearing or neurodevelopmental outcome, compared with children with normal brain MRI.
In this study, white matter abnormalities were seen in 15% of patients with neonatal hearing loss, but only in 1% of patients with normal neonatal hearing (p<0.01). This association strengthens the belief that white matter lesions are indeed part of the symptomatic spectrum of cCMV,17 as such confirming the recent study, where white matter lesions were observed as a concomitant finding in nearly all patients with other CMV-related brain abnormalities (gyral abnormalities, cysts, ventriculomegaly,…).16
The mechanism of hearing loss in cCMV is not well understood. Hearing loss can be present at birth, but can also develop later in life, both in initially symptomatic and asymptomatic children.9 None of the patients with late-onset hearing loss had abnormal white matter at birth, in line with previous studies.21 26
Cognitive developmental problems occur in a minority of patients with cCMV, however, they are a serious complication of the disease.6 7 In this cohort, a Bayley score below 86 was seen in 11% of patients with definite white matter signal alterations, compared with 2% of patients with normal or doubtful white matter, but this finding was not statistically significant. Although the numbers in this cohort were small, and the median follow-up sometimes short, concern should still be raised that isolated white matter lesions could lead to long-term cognitive problems—this in contrast to some previous studies, where white matter lesions were considered either unreliable or used in combination with other brain abnormalities to predict neurodevelopmental outcome,19 21 however, in accordance to others.12 25 27
Patients with abnormal white matter also had mildly lower motor developmental scores, with an AIMS score below P75 occurring in 81% of patients, compared with 64% with normal or doubtful white matter. Again, this finding was not statistically significant. In interpreting AIMS results we need to take into account the rather short follow-up of several patients in this study. In the first 18 months of life, a large variance in normal motor development can be observed. Problems later in life might not get picked up in the first few months. On the other hand, mild hypotonia early in life can influence the AIMS score, without persistent problems at later age. Therefore, we looked at the AIMS results in the subgroup of patients with a follow-up longer than 18 months. In this subgroup, we did not see a significant increase of serious motor problems depending on white matter score. However, a significantly larger group of patients with white matter lesions had mildly lower AIMS scores compared with patients with normal or doubtful white matter (p=0.01). This finding suggests that white matter lesions could be related to—be it rather mild—motor problems. However, larger and longer follow-up studies are needed to confirm this.
Grading the white matter can be challenging in newborns.19 Mainly discrete or isolated lesions can be hard to distinguish from normal variation in the unmyelinated white matter. At present, there is no expert consensus on defining ‘white matter’ signal alterations as pathological, leaving the decision up to the expertise of the reader. In this study, excluding patients with other brain abnormalities, the independent scoring of the white matter showed good interobserver agreement. While the white matter was considered doubtful in almost 30% of patients, no significant clinical differences were found between patients with normal versus doubtful white matter, both at birth or during follow-up. This finding can be reassuring for radiologists, treating physicians and parents. On the one hand, this could mean that observed discrete signal alterations are still part of the normal, unmyelinated neonatal brain. In other cases, these could be associated with very mild disease, causing no or only very discrete developmental problems.
Our study had some limitations. Limitations were certainly the number of patients (power) and the short length of follow-up. During follow-up, only the last observation carried forward was noted. Longer and larger cohorts, ideally after power analysis, are necessary to further evaluate the clinical meaning of white matter abnormalities and to provide accurate counselling to parents.
The white matter scoring system we used showed good interobserver agreement, however, further studies are needed to implement standardised cut-off values when evaluating the white matter, as well as investigating the meaning of the location and the extent of white matter lesions. Optimising MRI sequence parameters to the specific T1 and T2 relaxation times of the unmyelinated white matter in newborns and perhaps the quantitative measurement of ADC values might lead to a better detection of abnormalities.28 29 The use of rather short echo times for T2-weighted sequences in this study might have complicated interpretation.
Therapy, as well as the strict exclusion of other brain abnormalities, could have caused a bias in hearing and developmental outcome. Another limitation was the lack of a control group of normal newborns. However, getting ethical approval and moreover, finding parents to consent with an MRI without any clinical need, appeared merely impossible.
Grading of the white matter was feasible with good interobserver agreement. Abnormal white matter signal alterations were associated with neonatal hearing loss and mildly lower motor scores at follow-up. A tendency towards impaired cognitive development was seen. Patients with discrete or doubtful white matter did not show worse outcome, compared with children with normal brain MRI.
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
Patient consent for publication
This study involves human participants and was approved by Health, innovation and research institute, University Hospital Ghent, Belgium. Reference number: 2019/0199 BC-04369. Participants gave informed consent to participate in the study before taking part.
Contributors CVW: scientific guarantor, conceptualisation, methodology, formal analysis, investigation, data curation, writing—original draft, review and editing; AK, AO, ES, IJD and KS: conceptualisation, methodology, investigation, data curation, writing—review and editing; NH: conceptualisation, methodology, investigation, data curation, writing—review and editing, supervision; Deschepper Ellen, Wallaert Steven (biostatistics unit, Ghent University): served as scientific advisors.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.