Objective To evaluate the outcome of infants with congenital cytomegalovirus (CMV) infection and normal fetal imaging.
Design Retrospective cohort study.
Setting Tertiary paediatric medical centre.
Patients 98 infants born to mothers with primary CMV infection in the first and second trimesters (diagnosed by positive amniotic fluid findings) and normal fetal imaging.
Methods Initial evaluation included confirmatory urine culture, complete blood count, liver and kidney function tests, funduscopy, brain ultrasound and hearing test. Follow-up included periodic neurological and developmental evaluation, hearing tests until age 5 and Bayley-III Developmental Scale (in some patients).
Main outcome measures The presence and rate of sequelae of congenital CMV.
Results 52 (53.1%) infants received early antiviral treatment for central nervous system symptoms or signs, mainly lenticulostriatal vasculopathy on postnatal ultrasonography (88.5%). Sensorineural hearing loss was found on first examination in 16 infants (25 ears), of whom 10 also had cranial ultrasound findings; another five with late-onset hearing loss were also treated. The median follow-up time was 32 (12–83) months. Most infants with moderate and severe hearing loss were infected in the first trimester (10 vs 2, p=0.053). At the last assessment, eight children (10 ears) still had hearing loss, including two with bilateral loss who underwent a cochlear implant. The mean Bayley-III score was 102.6±10.3 (range 85–127). All 98 children attended regular educational institutions.
Conclusions Congenital CMV infection acquired from primary maternal infection with normal fetal imaging is associated with a high rate of subtle signs and symptoms after birth. Overall, intermediate-term outcome is good with a low rate of sequelae.
- Infectious Diseases
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
Twenty-five per cent of infants with congenital cytomegalovirus (CMV), born to mothers with a primary infection during the first and second trimesters, have neurological sequelae.
There are few reports of the outcome of a subgroup of infants with congenital CMV with normal fetal imaging.
What this study adds
Fifty per cent of infants (46) with congenital CMV, born to mothers with primary CMV infection and who had normal fetal imaging of ultrasound and fetal MRI, had subtle abnormal postnatal ultrasound findings.
Eight of 98 children had hearing loss at the last assessment; two children with bilateral severe hearing loss who underwent a cochlear implant; the other six children all had unilateral hearing loss.
Developmental outcome was good; all children attended regular educational institutions.
Bayley-III Scale scores were normally distributed in a subgroup of patients assessed.
Cytomegalovirus (CMV) is the most frequent cause of intrauterine infections, affecting 0.2–2.2% of live births worldwide.1 ,2 The risk of primary maternal infection during pregnancy is approximately 2%,3 with intrauterine transmission rates ranging from 30% to 42% in the first trimester and from 38% to 44% in the second trimester.4–7 Primary infections are more likely to be associated with severe fetal damage than recurrent infections. In one study, 18% of infants exposed to primary maternal infection were symptomatic at birth compared with almost none exposed to a recurrent infection.8 Sensorineural hearing loss (SNHL) was found in 15% of infants born to mothers with a primary infection, which was severe and bilateral in approximately 50% of infants.8 ,9 However, in the USA, only 25% of infants with congenital CMV infection were attributable to primary maternal infection; 75% were born to mothers with non-primary infections (reactivation or reinfection with different viral strains).10 In a recent population-based prediction model, it was estimated that non-primary infections account for the majority of CMV-related hearing losses.11
In the past, owing to the perceived high rate of neurological sequelae in children infected with CMV in utero, most women in Israel preferred to terminate their pregnancy if the virus was isolated from the amniotic fluid after a primary infection. During the last decade, however, improvements in serological assays,12 ,13 fetal brain ultrasound and fetal brain MRI14–17 have led to earlier and more accurate intrauterine detection of fetal brain damage. Therefore, in Israel today, when all fetal imaging studies are normal, some women opt to continue their pregnancy even in the presence of positive CMV findings in the amniotic fluid.
The aim of this study was to compare the short-term and intermediate-term outcomes of infants with a maternal CMV infection in the first trimester versus the second trimester with normal fetal imaging.
Patients and methods
The electronic records of a tertiary paediatric medical centre were searched for all infants born during 2007–2013 to mothers who had had a primary maternal CMV infection in the first or second trimester with positive findings in the amniotic fluid. Only those with normal fetal imaging who were older than 1 year on 1 January 2014 were included in the follow-up analysis. Background, clinical and imaging data at diagnosis and follow-up were recorded. The findings were compared between infants infected in the first trimester or the second trimester.
In Israel, many women undergo serological screening for CMV before or during the first trimester. Those who are seronegative are reassessed during pregnancy. The diagnosis of primary CMV infection is based on the following criteria: seroconversion of IgG from negative to positive during pregnancy, low IgG with low avidity in the presence of specific IgM and a significant rise in IgG and avidity at a later date. The timing of the maternal infection is determined by the serological and clinical data. These data are further examined to identify cases of periconceptional infection (4 weeks before the last reported menstrual period and up to 3 weeks of gestation). Amniocentesis to detect fetal CMV is performed after 21 completed weeks of gestation and at least 7 weeks after the assumed date of infection. Fetal CMV infection is diagnosed when the virus is detected in the amniotic fluid by both PCR amplification and rapid shell-vial culture. Follow-up fetal imaging includes repeated detailed transabdominal and transvaginal ultrasound and fetal brain MRI at 33–34 weeks of gestation.
All infants undergo confirmatory urine culture (shell-vial) during the first 2 weeks of life. Clinical and laboratory studies performed immediately after birth include complete blood count, liver and kidney function tests, funduscopy and ultrasonography over the anterior and posterior fontanels (performed by a paediatric radiologist). Small head circumference is defined as less than the second centile (<2 SD) and microcephaly as <3 SD. Full physical examination and neurological and developmental assessments are performed during the neonatal period; a second examination is performed within the first 3 months of life and every 3–6 months thereafter. In addition, hearing is evaluated by brainstem evoked response audiometry (BERA) during the first 2 weeks of life and again every 3–6 months until age 2; behavioural hearing tests are performed between ages 2 and 5 years. Patients with abnormal findings on behavioural tests undergo further testing with BERA. Evoked potentials are measured using Bio-Logic (Bio-Logic Systems Corp., Mundelein, Illinois, USA). Tympanometry is performed in patients with an abnormal BERA test and suspected conduction problems. If middle ear dysfunction is found, BERA is repeated after a few weeks, and only the bone-conduction results are used. BERA results are categorised according to a modification of the threshold definition used in the University of Alabama studies18 and the Grundfast and Siparsky19 method,: <25 dB—normal hearing; 25–44 dB—mild SNHL; 45–69 dB—moderate SNHL; ≥70 dB—severe SNHL.
For the present study, we also recorded scores on the Bayley Scales of Infant and Toddler Development, 3rd Edition (Bayley-III),20 which was administered to a subgroup of patients at ages 1–3 years by a clinical psychologist well trained in the use of this instrument.
All infants with a hearing loss or abnormal cranial ultrasound findings are treated using one of two clinical protocols: (a) up to 2011, intravenous ganciclovir 5 mg/kg twice a day for 6 weeks followed by oral valganciclovir 17 mg/kg/dose in two daily doses for 6 weeks followed by one daily dose for 9 months was given21 and (b) since 2011, the protocol used oral valganciclovir 17 mg/kg/dose in two daily doses for 12 weeks followed by one daily dose for 9 months.
Groups were compared by the χ2 test or the Fisher's exact test (two-tailed) for categorical variables and the Mann–Whitney test for continuous non-parametric variables (follow-up time). p<0.05 was considered significant.
One hundred and one mothers met the study criteria: pregnancy during 2007–2013; primary CMV infection in the first or second trimester; positive CMV findings in the amniotic fluid and normal findings on fetal imaging by both ultrasound and MRI; in three cases, there were suspected abnormal findings in the brain white matter seen only on the MRI. After a review of the MRI scans, they were reported as normal; therefore, no case was excluded from the study due to a normal ultrasound and abnormal MRI. There was a steady increase over time in the number of women with positive amniotic fluid findings who opted to continue the pregnancy and gave birth: 4 during 2007 and 6, 12, 19, 32 and 28 during 2008–2012, respectively. The rate of termination of pregnancy due to CMV detection in the amniotic fluid decreased from approximately 90% in 2007 to 30% in 2012, in our centre.
The final study group consisted of 98 of these infants with congenital CMV infection. The other three were lost to follow-up. Diagnosis was confirmed by a positive urine culture for CMV (shell-vial) during the first 2 weeks of life. Fifty-two infants (53.1%) acquired the infection in the first trimester and 46 (46.9%) in the second trimester (p=0.544). There was no difference in sex distribution by time of infection (61.5% and 56.5% males, respectively, p=0.544). Ninety-one infants (92.9%) were born at term and seven (7.1%) at near-term/prematurely (34–36 weeks). Of the premature infants, six (11.5% of the cohort) were infected in the first trimester and one (2.2%) in the second trimester (p=0.063). Birth weight was within the normal range in 94 infants (95.9%). The other four were small for gestational age, two (3.9%) infected in the first trimester and two (4.4%) in the second trimester (p=0.664).
Sixty infants (61.2%) exhibited early postnatal signs or symptoms possibly related to congenital CMV infection according to our definition. Abnormal cranial ultrasound findings were observed in 46 (46.9%), mostly (44/46, 95.7%) lenticulostriatal vasculopathy (LSV). One infant each (2.1%) had periventricular hyperechoic foci and a single small calcification. Head circumference was below the second centile in six infants (6.1%). There were four cases each of elevated liver enzyme levels and thrombocytopaenia. None of the infants had a petechial rash. Splenomegaly was the most common non-central nervous system (CNS) sign found in 18 infants (18.4%). Of the 46 infants with an abnormal cranial ultrasound, 10 had SNHL, 12 splenomegaly, 4 a small head circumference, 3 thrombocytopaenia and 1 chorioretinitis. Nine of the 10 infants with an abnormal ultrasound and SNHL had only LSV. The distribution of the signs and symptoms by time of infection is shown in table 1.
SNHL was detected on initial assessment in 16/98 infants (16.3%) with 25 affected ears (12.8% of total 196 ears): mild in 13 ears, moderate in 9 and severe in 3. Although not reaching statistical significance, there were more cases of moderate and severe hearing loss among infants infected in the first trimester than infants infected in the second trimester (10 vs 2, p=0.053).
Fifty-two infants (53.1%) were treated with an antiviral agent starting in the first month of life. The main indication for treatment was an abnormal cranial ultrasound finding in 46 infants (88.5% of treated patients). Of the 16 infants with early SNHL, 10 also had an abnormal ultrasound finding and the other 6 were treated only for the hearing loss. An additional five patients were treated at a later period for SNHL, making a total of 57 treated patients in the cohort (58.2%).
The median duration of follow-up was 32 (12–83) months. Infants infected during the first trimester were followed for significantly less time than those infected during the second trimester (27 (12–83) versus 45 (12–77), p<0.001). Of the six patients with a small head circumference on initial examination (table 1), four measured above the second centile at the last follow-up and two remained below.
The findings on hearing tests are shown in table 2.
Reassessment of the 16 children (25 ears) with hearing loss at the first examination revealed that of the 13 ears with mild hearing loss, 9 showed an improvement and 4, no change. Of the nine ears with moderate hearing loss, six showed an improvement and one deteriorated to severe hearing loss. None of the three ears with initially severe hearing loss showed any change by the last visit. An additional five patients (10 ears) were found to have late-onset SNHL at age 5–14 months. All improved to normal hearing after treatment. Thus, of the 98 children in the cohort, eight children (10 ears) had hearing loss at the last assessment. Analysis by the best-ear approach18 yielded two children with bilateral severe hearing loss who underwent a cochlear implant; the other six children all had unilateral hearing loss: moderate in two, mild in four.
The patient with chorioretinitis (table 1) had normal vision with a small retinal scar. Motor development was assessed approximately every 6 months. Two children experienced mild motor delays (started walking independently at 18 and 19 months); both were diagnosed with benign hypotonia. One also had a small head circumference. The other had a brother who exhibited the same motor skills pattern. The rest of the cohort achieved normal motor milestones.
Developmental assessment by the Bayley-III was performed during 2010–2011 in a subgroup of 27 patients. The mean mental developmental index was 102.6±10.3 (range 85–127). None of the patients showed any cognitive delay.
At the end of this study, all 98 children were attending regular educational institutions. Five were in first or second grade, 40 in kindergarten and 53 in day-care centres.
In Israel, screening for CMV infection is frequently performed before or during the first trimester of pregnancy. Amniocentesis is performed in mothers with a primary infection to detect fetal infection. Over the study period, there was an increase in patients with primary CMV infection and normal fetal ultrasound who continued the pregnancy. The numbers have been steadily increasing, as demonstrated in the present study.
The neurodevelopmental impact of intrauterine CMV ranges from no adverse long-term effects to severe damage and even fetal death. The present study describes the short-term and intermediate-term outcomes of infants with congenital CMV infection diagnosed in utero by positive amniotic fluid (PCR) with normal fetal imaging (ultrasound and MRI) results. All were infected in the first or second trimester. Three cohort studies in the literature reported a good short-term prognosis for infants with these characteristics.17 ,22 ,23
Accordingly, most of the 98 infants in our study were born at term with normal birth weight. However, 61% exhibited subtle signs and symptoms of congenital CMV infection undetected on fetal imaging (table 1). About 50% had abnormal postnatal ultrasound findings, mostly LSV.
Is LSV a true sign of CNS involvement caused by CMV infection? Researchers have no definitive answer. However, in our published experience24 and a recent larger cohort study (submitted for publication), we showed that LSV is a common finding in infants with congenital CMV infection and may serve as a marker of high risk for SNHL. In the present study, 12.8% of ears had hearing loss on the first BERA. Most cases were mild, although moderate and severe hearing losses were associated more often with first-trimester infection rather than second-trimester infection (p=0.053). All infants with abnormal ultrasound findings or SNHL were treated with ganciclovir/valganciclovir. By the last hearing assessment, six children had unilateral residual hearing loss (no hearing aid); only two required cochlear implants for severe bilateral hearing loss.
Small head circumference was detected on initial examination in six infants (6%). In two of them, the findings normalised at the following clinical visit, suggesting that the abnormalities had been due either to moulding of the skull or to error. In the other four patients, head circumference remained on the second centile in two and below the second centile in two (2 SD below the mean) at the last follow-up examination. In all these cases, head circumference had been reported as normal on fetal ultrasound examination. This discrepancy may have been due to different definitions of abnormal head circumference among paediatricians (less than the second centile or <2 SD below the mean)25 and ultrasound guidelines for fetal microcephaly (3 SD below the mean).26 Accordingly, others described three infants diagnosed on follow-up with microcephaly and developmental delay in which fetal ultrasound findings had been considered normal.27 We suggest that the relationship between fetal ultrasound head circumference measurements and postnatal clinical findings needs to be reassessed.
The developmental outcome of our cohort was good, with all patients attending regular educational institutions at the end of the study. In the subgroup of patients assessed with the Bayley-III Scale, scores were normally distributed. These results may reflect our selection of patients with normal fetal imaging findings, although the antiviral therapy may also have contributed, as reported by others.28 ,29 Indeed, the rate of antiviral treatment was high (57/98 cases, 58.2%) because we treated the infants with LSV. There are no reliable data on the treatment rate of infants who were born to mothers with primary CMV infection and normal fetal imaging findings. Most studies have included mainly symptomatic patients without brain ultrasound studies. Our report is a large cohort describing this unique patient group. The results also support the relative safety of the approach used in our centre in pregnancies complicated by primary maternal CMV infection during the first and second trimesters.
The main limitations of this study were the relatively short follow-up of the infants infected during the first trimester, the absence of a control group of untreated infants with a sole abnormal finding of LSV on ultrasound and the small number of patients formally assessed for developmental delay. Although postnatal antiviral therapy is now common practice in many medical centres worldwide, the presented therapy protocol is more extensive than used in most centres treating congenital CMV. Our protocol was not studied in controlled trials; however, recent data have implied that longer treatment seems better.29 Moreover, no controlled studies have specifically assessed the benefit of treating babies with only subtle signs on cranial ultrasound or isolated SNHL or in late-onset hearing loss. Improvement in these groups may therefore be due to hearing threshold fluctuations known to occur in congenital CMV.30
In conclusion, this study demonstrates that infants with congenital CMV infection acquired from primary maternal infection during the first and second trimesters of pregnancy with normal findings on fetal imaging may have a high rate of subtle postnatal signs and symptoms. First trimester infection is associated with a trend towards more severe hearing loss than second trimester infection. Overall, the outcome of these infants is good, with a low rate of sequelae on short-term follow-up and intermediate-term follow-up.
The authors thank Mrs Phyllis Curchack Kornspan for her editorial services and Mrs Tamar Elazar for her secretarial services.
Contributors JA: designed the study, collected all follow-up data and wrote the manuscript. JA: responsible for collecting and analysing the hearing studies. NL: collected and analysed the short-term data of the newborn with congenital CMV. JP: collected and analysed the data relating to the pregnant women and contributed in interpretation of the data.
Competing interests None declared.
Ethics approval The study protocol was approved by the Rabin Medical Center's ethics committee.
Provenance and peer review Not commissioned; externally peer reviewed.