Background We previously reported a 2.2% rate of infants born with sensorineural hearing loss (SNHL) due to congenital cytomegalovirus (cCMV) infection identified by universal neonatal screen for cCMV using saliva.
Objective To evaluate the contribution of targeted saliva screening for cCMV to the detection of infants born with cCMV-related SNHL who failed universal newborn hearing screening (UNHS).
Methods We retrospectively reviewed the audiological and medical records of infants who failed UNHS and were tested for cCMV using saliva sample prior to discharge at Sheba Medical Center between 2014 and 2015. Positive cases were confirmed by urine sample.
Results Two hundred (1%) of the 19 830 infants tested during the study period failed in-hospital hearing screening. A saliva specimen was obtained prior to discharge in 187 infants (93.5% of those who failed UNHS). In 178 infants saliva testing was performed at ≤21 days of chronological age and yielded results. cCMV infection was identified in 4/178 tested infants (2.25%, 95% CI 0.8% to 5.3%), of whom three were diagnosed with SNHL (1.7%, 95% CI 0.5% to 4.4%) and offered antiviral treatment. Two of the tested infants (1.12%, 95% CI 0.2% to 3.6%) were diagnosed with cCMV solely due to failure in UNHS. Occult central nervous system (CNS) symptoms of cCMV infection were detected in 2/4 infants following targeted investigation.
Conclusions Targeted cCMV screening in newborns who failed UNHS contributed to the early detection of infants born with cCMV-related isolated SNHL or with occult CNS symptoms who could potentially benefit from antiviral treatment.
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?
Real-time PCR in saliva samples has been shown to be a feasible and reliable tool for detecting congenital cytomegalovirus (cCMV) infection.
Targeted screening for cCMV in infants who fail universal newborn hearing screen (UNHS) was shown to be feasible.
What this study adds?
Targeted cCMV screening in newborns who fail UNHS contributed to the detection of cCMV-related sensorineural hearing loss.
Targeted cCMV screening contributed to the detection of occult central nervous system symptoms.
Targeted screening contributed to the early detection of infants who could potentially benefit from antiviral treatment.
Congenital cytomegalovirus (cCMV) infection is the most common congenital infection, with an estimated prevalence of 0.5%–0.7% in developed countries.1–5 Sensorineural hearing loss (SNHL) is the most common sequelae of cCMV infection, with an overall rate of 10%–15%,3 4 which may be present at birth or delayed in onset.2–4 6 7 The frequency of SNHL among children born with an ’asymptomatic' infection (~10%)3 4 is lower than that among symptomatic children (~30%).4 7 8 Nevertheless, as the majority of infants born with cCMV infection are ’asymptomatic' at birth (~85%–90%),3 4 most cases with cCMV-related SNHL (cCMV-SNHL) belong to this group.3
Currently, universal cCMV screening programmes are not implemented; thus, ’asymptomatic' infants cannot be routinely diagnosed. Some of these infants can be detected by testing for cCMV infection following the confirmation of SNHL.9 Nevertheless, for accurate diagnosis and treatment, it is important to test for cCMV within the first 3 weeks of life in order to distinguish a congenital from a postnatal acquired infection.10 A CMV-positive result soon after birth can encourage parents to proceed immediately with the audiological evaluation following failure on universal newborn hearing screening (UNHS).11 Furthermore, following the confirmation of cCMV, it is important to perform a complete medical workup in order to determine whether there is evidence of other occult cCMV-related central nervous system (CNS) findings such as periventricular calcifications.12 Finally, although still controversial, there is accumulating evidence for the potential benefit of antiviral treatment in symptomatic infants with cCMV-SNHL when started within the first month of life,13 14 and it is therefore offered nowadays to symptomatic infants.15 Currently, there is no evidence from controlled studies to support the benefit of such treatment to infants with isolated cCMV-SNHL or those with subtle CNS signs.
Thus, in order to detect infants with cCMV-SNHL soon after birth and offer treatment if necessary, it was proposed to test for cCMV infants who fail hearing screening prior to discharge as part of UNHS programmes.9 11 16 17 An easily applicable and reliable tool for detecting cCMV is real-time PCR (rt-PCR) in saliva.1 18 19 A few studies of targeted cCMV saliva screening in infants who fail UNHS reported that the percentage of infants identified with cCMV-SNHL varied between 0.5% and 3.9%.20–22 One of the shortcomings of these studies was that the rate of cCMV infection in the studied populations was unknown, thus making it difficult to assess the contribution of such targeted screening to the detection of cCMV-SNHL. Recently, we conducted a universal newborn saliva cCMV screening during 1 year in our medical centre. Results indicated an incidence of 0.5% (48/9845) of CMV-positive infants of all live births.1 Since January 2014, a targeted saliva cCMV screening protocol in infants who fail UNHS prior to discharge is currently routinely operating at our medical centre. The aim of the present study was to evaluate the contribution of targeted saliva cCMV screening to the detection of infants born with cCMV-SNHL who failed UNHS.
The cohort consisted of all infants born at Sheba Medical Center between January 2014 and December 2015 who failed UNHS prior to discharge. The hospital serves Jewish population, mainly secular, with ~10 000 deliveries per year. All infants were cared for in the well-baby nursery (WBN), in the intermediate care unit or in the tertiary-level neonatal intensive care unit (NICU) with or without additional risk factors for SNHL according to the Joint Committee on Infant Hearing9 (ie, family history of permanent childhood hearing loss, NICU stay >5 days, hyperbilirubinaemia requiring exchange transfusion, in utero infections, craniofacial anomalies and syndromes associated with hearing loss). Excluded were infants who were tested for cCMV at age >21 days.
Hearing screening protocol
As part of the national UNHS programme, all infants were tested using a combined protocol of transient evoked otoacoustic emissions (TEOAE) and automated auditory brainstem response (A-ABR). The hearing screening protocols for infants in the WBN and those admitted to the NICU >5 days are detailed in figures 1 and 2, respectively. TEOAE tests were performed using an Otoport Screener (Otodynamics, London, UK) with a click intensity level set at 82±3 dB peak equivalent sound pressure level (SPL). TEOAE pass criterion was set to signal-to-noise ratio levels ≥6 dB at 2.0, 3.0 and 4.0 kHz frequency bands and response level >0 dB SPL in both ears. A-ABR test was performed using an ABaer system (Biologic, Mundelein, Illinois, USA) with a click intensity level set at 35 dB normalised hearing level.
Infants who failed two TEOAE and A-ABR or failed two TEOAE tests were immediately tested for cCMV and were referred for an outpatient audiological rescreen or evaluation (figures 1 and 2).
At our medical centre, all infants with a history of maternal seroconversion to CMV during pregnancy and/or with clinical suspicion of congenital infection at birth (ie, small for gestational age (SGA), microcephaly, petechiae and thrombocytopaenia; see definitions in table 1), are routinely screened for cCMV prior to discharge. Since January 2014, cCMV saliva screening is also routinely performed in all infants who fail UNHS by nurses trained by our virology laboratory experts. Saliva specimens are collected by swabbing the infant’s buccal mucosa using a sterile open cell bud Sigma-Swab (Sigma-Virocult) and placed in 2.0 mL of Virocult transport medium. The specimens are stored at 4°C until they were transported to the Central Virology Laboratory within 72 hours. Viral DNA is extracted by MagNA Pure LC (Roche) and tested by rt-PCR assay, developed at the laboratory, for the presence of CMV IE and gB genes.
If the saliva sample is positive for CMV, parents are immediately contacted and a urine sample is collected for rt-PCR. Confirmation of the diagnosis of cCMV infection is based on a positive urine test within 21 days of birth.
Medical and audiological assessment and treatment
All infants with confirmed cCMV infection underwent the following workup within 2 weeks: complete blood count, serum liver enzymes, head ultrasound, retinal examination, and an audiological evaluation including diagnostic ABR to clicks and tonal stimuli, behavioural audiometry, tympanometry and/or otoscopic examination. SNHL was defined as elevated ABR air-conduction thresholds >25 dBnHL and bone-conduction thresholds >15 dBnHL for 1 and 4 kHz tone bursts stimuli. The severity of SNHL was determined based on Clark’s classification.23
Infants were classified as symptomatic (other than hearing) if one or more of the findings detailed in table 1 were detected. Infants were referred to the paediatric infectious disease clinic (not later than 4 weeks of age) to consider treatment and for long-term follow-up.
According to our medical centre protocol, symptomatic infants with or without CNS symptoms (eg, microcephaly, abnormal neurological examination) and those with isolated SNHL or with CNS findings in brain imaging (eg, calcifications, multiple germinal matrix cysts or lenticulostriate vasculopathy (LSV)) were offered antiviral treatment with valganciclovir 16 mg/kg/dose twice daily for 3 months and thereafter once daily for 6–12 months. A complete blood count was performed every 2 weeks in infants receiving antiviral treatment.
This retrospective study was approved by the institutional review board of Sheba Medical Center.
The rates are presented as percentages with 95% CI. The χ2 test was used to compare the rate of cCMV-SNHL between targeted and universal cCMV screening.
During the 2-year study period, 19 887 infants were born at Sheba Medical Center and survived to discharge. Of these, 19 866 infants (99.9%) underwent hearing screening tests prior to discharge. Thirty-six infants (0.18%) failed one TEOAE test and were not retested due to an early discharge, and thus were not screened for cCMV and excluded from further analysis. A total of 19 830 infants (99.7% of all discharged infants) completed hearing screening. Of those, 18 836 were screened according to the WBN protocol (figure 1) and 994 infants (5%) according to the NICU protocol (figure 2). Our cohort comprised 200 infants (1%) who failed UNHS and were eligible for a saliva screening test (figure 3).
A saliva specimen for CMV testing was obtained in 187 infants (93.5% of infants who failed UNHS), of whom 169 infants were in the WBN and 18 infants were in the NICU. In seven infants who were in the NICU >5 days (3.5% of the infants who failed UNHS), saliva test was performed at a chronological age of >21 days. All these cases had a negative saliva assay and were excluded. In two additional cases, the saliva test was not performed due to technical reasons. Thus, cCMV saliva samples yielded a result in 178 infants (89% of all infants who failed UNHS). The mean age at testing was 3.3 days (SD=2.12). In 4/178 (2.25%), saliva was borderline CMV-positive (viral loads between 145 and 2700 copies/mL). All these infants were CMV-negative in a urine sample and thus diagnosis of cCMV was excluded. In four additional infants (4/178, 2.25%, 95% CI 0.8% to 5.3%), saliva rt-PCR tests were CMV-positive (viral loads between 2 760 000 and 25 426 000 copies/mL). All were confirmed by a positive urine rt-PCR (figure 3). The clinical characteristics of the four infants diagnosed with cCMV are detailed in table 2.
SNHL was diagnosed in 3/4 infants who were cCMV-positive (table 2). Thus, CMV-related SNHL was diagnosed in 3/178 (1.69%, 95% CI 0.5% to 4.4%) of the infants who failed UNHS. Of all 178 infants who failed UNHS, 164 (92%) underwent audiological evaluation. Of those, 36 infants were diagnosed with SNHL. Thus, the rate of cCMV identified among children who failed UNHS and were confirmed to have SNHL was 3/36 (8.3%, 95% CI 2.4% to 20.6%) while among those who were not confirmed to have SNHL was 1/151 (0.7%, 95% CI 0.1% to 3.1%).
In 2/4 cCMV-positive infants, maternal seroconversion during pregnancy prompted clinical investigation (table 2). In the remaining two cases, cCMV was diagnosed solely due to failure in UNHS (2/178, 1.12%, 95% CI 0.2% to 3.6%). Further investigation revealed in 2/4 infants occult findings related to cCMV infection that were detected only following targeted investigation (ie, LSV, germinal matrix cysts). None of the four infants had additional known risk factors for SNHL.9
All four cCMV-positive infants were offered treatment with oral valganciclovir. In the normal-hearing infant, treatment was offered due to developmental delay. Currently, all three infants with SNHL are receiving hearing rehabilitation.
In the present study, saliva samples were collected in the majority of the infants who failed UNHS (93.5%) prior to discharge. Thus, integrating a targeted cCMV saliva screening to the UNHS programme was easy and feasible. Four of the 178 infants who failed UNHS and were screened for cCMV within 21 days (2.25%, 95% CI 0.8% to 5.3%) were cCMV-positive. These data are in accordance with previous studies reporting cCMV rates of 1%–4.7% (detected in urine or saliva samples) among infants who failed UNHS.17 20–22 24 In one infant (table 2, case 4) occult symptoms were detected following targeted investigation for cCMV symptoms only due to our targeted cCMV screening. On the other hand, in two infants (table 2, cases 2 and 3), CMV infection would have been diagnosed via our routine targeted cCMV screening for infants with history of maternal CMV seroconversion during pregnancy and/or clinical symptoms associated with congenital infection. Thus, UNHS contributed to the detection of two CMV-positive infants (1.12%) (table 2, cases 1 and 4) in whom there was no previous suspicion of congenital infection.
The rate of cCMV-SNHL at birth detected in the present cohort (3/178, 1.7%, 95% CI 0.5% to 4.4%) is in accordance with previous reports, in which 0.5%–3.9% of the infants who failed UNHS were subsequently diagnosed with cCMV-SNHL.17 20–22 24 Furthermore, the rate of CMV-positive cases among infants who failed UNHS and diagnosed with SNHL (3/36, 8.3%, 95% CI 2.4% to 20.6%) and those who failed UNHS but were not confirmed to have SNHL (1/151, 0.7%, 95% CI 0.1% to 3.1%) are similar to previous reports on targeted cCMV screening in infants who fail UNHS.17 20 Two of the three infants diagnosed with SNHL were born to preconception seropositive mothers. This finding is in agreement with previous reports showing that infants with cCMV born following maternal non-primary infection are not fully protected from SNHL.25 26
Although the contribution of targeted cCMV screening to the detection of CMV-SNHL in infants who fail UNHS was previously described,17 20–22 the actual rate of cCMV in the tested populations was unknown. The novelty of the current study is that it was conducted in a population where the rate of cCMV has recently been reported using universal newborn CMV saliva screening.1 Thus, we could indirectly assess the impact of UNHS to the detection of infants born with cCMV-SNHL on the basis of modelling. The rate of cCMV infection detected via our universal cCMV screening programme was 0.5%, of whom one child was diagnosed with SNHL at birth (1/46, 2.2%: 95% CI 0.2% to 9.7%).1 If the rate of 0.5% of cCMV should be applied to the present birth cohort (n=19 887), we would expect to diagnose ~100 CMV-positive infants. As approximately ~8%–10% of these infants are expected to suffer from SNHL greater than 25 dBHL and in about half of the cases at birth,7 we would expect to identify four to five infants with SNHL at birth. In the current study, three infants were identified with cCMV-SNHL at birth. It might be the case that not all expected cases were identified due to different definitions of SNHL. Specifically the current technology for UNHS (ie, TEOAE and A-ABR) cannot detect mild degrees of SNHL. Thus, asymptomatic infants with cCMV who failed TEOAE and passed A-ABR testing may have a mild degree of SNHL that could not be detected.24 27 These asymptomatic infants would be missed by targeted cCMV screening in infants who fail UNHS but not by universal cCMV screening. Furthermore, no significant difference was found in the rate of cCMV-SNHL at birth found in the present cohort using targeted screening compared with that found in our universal cCMV screening cohort1 (3/19 830 and 1/9845, χ2=0.12, p=0.73, for targeted and universal screening, respectively). This comparison should, however, be interpreted cautiously as the sample size of both studies may not be sufficient to make an accurate comparison. If, as expected from Fowler et al’s study,7 half of all cCMV cases will suffer from a late-onset SNHL, targeted screening would miss a relatively large proportion of these cases. On the other hand, if the rate of cCMV cases with late-onset SNHL will be less than expected, then the contribution of universal versus targeted screening would be less significant. This issue will need to be investigated in future studies.
Currently, there is no consensus whether to offer antiviral treatment to children with isolated cCMV-SNHL, with some clinicians arguing that there is no evidence from controlled studies to support the benefit of such a treatment28 while others advocating for such a protocol.10 29 30 In our universal cCMV screening programme, the majority of the identified infants were asymptomatic at birth with normal hearing,1 and thus were not offered any medical intervention as there are currently no controlled data to support the benefit from such treatment. In contrast, our targeted cCMV screening in infants who failed UNHS enabled us to identify infants with cCMV-SNHL who were offered antiviral treatment. Nevertheless, the challenge of verifying the effectiveness of such treatment in children with isolated cCMV-SNHL is still warranted.
In addition to the above challenge, several other challenges need to be addressed while implementing cCMV screening. One such challenge is parental anxiety. In the present cohort, only a minority of the infants (1%) were screened for cCMV compared with universal screening, thus diminishing maternal anxiety. Although not tested in the present study, the finding that in 170/178 (96%) of the infants saliva screening was negative may cause maternal anxiety. Williams et al20 recently reported that targeted cCMV saliva screening in infants who failed UNHS did not increase maternal anxiety. As only 40% of the infants who failed UNHS participated in their study, this issue needs to be further addressed using parent anxiety questionnaire. Another challenge is the cost–benefit of screening for cCMV infection.31 32 While newborn cCMV screening was recently reported to be cost-effective,32 the precise long-term benefits of antiviral therapy as well as the benefit of earlier identification of late-onset SNHL are not yet well defined. Furthermore, the costs associated with hearing loss are yet to be determined. Finally, while the majority of infants who fail UNHS are screened for cCMV within 21 days of age, infants who are in the NICU may not be screened on time due to a prolonged stay. In our study, in 7/18 NICU infants who failed UNHS (39%), cCMV screening was performed at a chronological age >21 days and were therefore excluded. As infants in the NICU are at higher risk for SNHL and the rate of cCMV infection in this population is unknown, universal saliva screening immediately after birth should be considered for this population.
This retrospective study has several limitations. First, saliva samples were not obtained in all cases that failed UNHS. In 36 cases who failed one TEAOE test and were discharged early, cCMV screening was not performed. Thus, we may have missed CMV-positive infants. Second, the current hearing screening tools cannot detect mild degrees of SNHL. Third, due to the current screening protocol, infants with cCMV-SNHL involving the acoustic nerve alone will not be identified if screened only by TEOAE.16 Finally, the relatively small cohort in the present study may not have enabled us to precisely estimate the rate of cCMV-SNHL.
The findings of the present study further ascertain the feasibility of integrating targeted saliva cCMV screening to UNHS programmes. Targeted cCMV screening, in which only a minority of the infants were tested compared with universal screening, contributed to the early detection of infants born with isolated cCMV-SNHL or with occult CNS symptoms who were offered antiviral treatment. Whether this protocol will provide a clear benefit remains to be investigated in future controlled studies. If a significant proportion of infants with cCMV-SNHL are expected to be delayed in onset, such a targeted screening programme could not detect those ‘silent at birth’ infants.33 While there is no evidence at the present time that antiviral treatment could improve their outcomes, these ‘missed’ children could benefit from early hearing rehabilitation. Furthermore, the cost–benefit of universal versus targeted cCMV screening needs to be further assessed.
We wish to thank our UNHS team of audiologists and screening personnel, as well as the team at the Department of Neonatology, for their devoted work.
Contributors DAER conceptualised and designed the study, including data acquisition and analysis, drafted the initial draft of the manuscript and revised the manuscript. DL made a significant contribution to the conception of the study and revised the manuscript. JK critically revised the draft for important intellectual content and revised the manuscript. MT collected data, participated in the interpretation of the data and revised the manuscript. EM and TS participated in the interpretation of the data and revised the manuscript. GB conceptualised and designed the study, conducted data analysis and revised the manuscript. All authors approved the final submitted manuscript.
Competing interests None declared.
Ethics approval Institutional review board of the Sheba Medical Center.
Provenance and peer review Not commissioned; externally peer reviewed.