Clinical characteristics , audiological and neurodevelopmental outcomes of newborns with congenital cytomegalovirus infection

Kobas Manuelaa, Bickle Graz Myriambc, Truttmann Anita Carmenc, Giannoni Ericc, Meylan Pascalde, Asner Sandra Andreaef a Faculty of Biology and Medicine, University of Lausanne, Switzerland b Follow-up Unit, Clinic of Neonatology, Department Woman-Mother-Child, Lausanne University Hospital, Switzerland c Clinic of Neonatology, Department Woman-Mother-Child, Lausanne University Hospital, Switzerland d Institute of Microbiology, Department of Laboratories, Lausanne University Hospital, Switzerland e Infectious Diseases Service, Department of Medicine, Lausanne University Hospital, Switzerland f Paediatric Infectious Diseases and Vaccinology Unit, Department Woman-Mother-Child, Lausanne University Hospital, Switzerland

Diagnosis of cCMV infection relies on the documentation of a positive CMV polymerase chain-reaction (PCR) in urine or saliva during the first 3 weeks of life [11][12][13].Whereas some centres have introduced universal neonatal screening for cCMV [9], others, such as ours, perform targeted neonatal testing [14] in newborns with microcephaly, subependymal pseudocysts or thrombocytopenia [15][16][17].Documentation of cCMV disease in newborns is crucial as prompt initiation of oral valganciclovir treatment improves audiological and, possibly, neurodevelopmental outcomes [16,18,19].Universal screening programmes of newborns have not been implemented in part because of a lack of epidemiological data [20].As a consequence, asymptomatic newborns will be diagnosed late through identification of SNHL [21].
From recent data, we estimate that cCMV infection could affect 500 newborns each year in Switzerland, of whom up to 100 could suffer from lifelong disabilities [9,22,23].We aimed to provide information on neurodevelopmental and audiological outcomes of newborns with cCMV infection diagnosed at the Lausanne University Hospital.In addition, clinical outcomes of infants born to mothers with primary infection were compared with those born to mothers with nonprimary infection.

Study design and definitions
This was a single-centre retrospective study of neonates born between 1999 and 2014 presenting with cCMV infection and treated at the Lausanne University Hospital, Switzerland.Congenital CMV infection was defined from 1999 to 2004 as a positive CMV viral culture [24], and from 2005 to 2014 by quantitative measures of viral load with molecular assays (detection limit 100-1000 copies/ ml [25]) in urine, collected within the first 3 weeks of life [11][12][13].Infants born during the study period and diagnosed retrospectively with CMV PCR testing on stored dried blood spots (Guthrie cards) were also included [26].
Indications for neonatal CMV testing included documented maternal seroconversion during pregnancy, abnormal prenatal ultrasounds, documentation of CMV by molecular assays from amniotic fluid (detection limit 100-1000 copies/ml) and postnatal indications such as being small for gestational age (SGA), microcephaly, subependymal pseudocysts and/or thrombocytopenia [15-17, 20, 25].Screening of pregnant women and newborns relied on physician-based practice and not on predefined screening policies.Maternal seroconversion during pregnancy was defined either as documentation of CMV-specific immunoglobulins in a patient previously known to be seronegative, or a newborn with proven cCMV infection born to a mother without any demonstrable CMV-specific antibodies during pregnancy [1,5,27].A nonprimary maternal infection was defined as positive IgM and IgG serology with an IgG avidity of ≥70% determined within the first 12 weeks of pregnancy.The assay used to measure IgG avidity was an in-house assay based on the Enzygnost CMV IgG assay (Siemens) using an 8 M urea wash, as described elsewhere [28].The avidity threshold defining remote infection was set using in-house seroconversion panels, and by comparison with the work of Grangeot-Keros et al. [29].

Collection of clinical information, audiological and neurodevelopmental assessments
Retrospective chart review and use of the database from the neonatology clinic follow-up unit provided information on relevant baseline characteristics and outcomes.This study was approved by the ethics committee on research involving humans of the canton of Vaud (protocol number 426/14).

Audiological assessment
BSER was assessed at birth and between 3 and 24 months by an ear-nose-throat (ENT) specialist.Hearing thresholds were defined as follows: 0 to 30 dB for normal hearing, 31 to 45 dB for mild hearing loss, 46 to 70 dB for moderate hearing loss and above 71 dB for severe hearing loss [16,18].We used total and best ear analyses as described [18].Audiological outcomes were adjudicated on the latest BSER.

Neurodevelopmental assessments
Neurodevelopmental follow-up was offered to neurologically symptomatic patients from 2001.Infants were examined with the revised Griffiths Mental Development Scales (GMDS) at 6 months of age [37], and the Bayley Scales of Infant and Toddler Development second and third editions (Bayley-II and III) at 12 months and 18 months of age [38].Older infants were tested with the Kaufman Assessment Battery for Children (K-ABC) first edition [39], or with the Wechsler Preschool and Primary Scale of Intelligence (WPPSI-R) [40].Neurodevelopmental outcomes were based on the latest scores determined with the GMDS, Bayley-II and III, WPPSI-R, K-ABC, all with an expected mean of 100 and a standard deviation of 15 within the normal population [37][38][39][40].Hence test results were classified as abnormal if one test component was below 85.Sequelae were defined as an abnormal audiological and/or an abnormal neurodevelopmental outcome.

Statistical analysis
Descriptive statistics were used.Z-scores of the growth parameters weight, height and head circumference of newborns were calculated using the Fenton 2013 Growth Calculator for Preterm Infants [41].

Study population
From 1999 to 2014, 38,423 newborns were born at Lausanne University Hospital of whom 2274 were tested for CMV in urine samples.Of these 2274, 48 were found positive by viral culture or PCR.Among the 48 newborns diagnosed with cCMV infection, 24 (50%) were tested for CMV as a result of maternal seroconversion, 21 (44%) because they were SGA, 2 (4%) because of thrombocytopenia and 1 (2%) as a result of abnormal antenatal ultrasounds.Two additional patients (twins) with SNHL (and subependymal pseudocysts on neonatal brain ultrasound) tested positive by PCR on stored dried blood-spots at the age of 3.5 years.

Clinical characteristics at birth, audiological and neurodevelopmental outcomes of all infants born to mothers with a primary compared with those born to mothers with a nonprimary infection
Newborns born to mothers with nonprimary infection were more often symptomatic and SGA at birth than those born to mothers with a primary infection.In contrast, petechiae were more frequent in newborns born to mothers with primary infection than in newborns born to mothers with a nonprimary infection.In the two groups, the rates of microcephaly and abnormal audiological outcome at birth were similar (table 3).
Results of hearing assessments and neurodevelopmental outcomes at 3 months of age and onwards were similar in children born to mothers with a primary infection and those with a nonprimary infection (table 4).

Discussion
In this first exploratory study to address the burden of cCMV infections in Switzerland, three important observations were made: (1) cCMV infections were detected equally among newborns born to mothers with primary infection and nonprimary infection; (2) most infants who presented with SNHL or abnormal neurodevelopment were neurologically symptomatic at birth; (3) newborns of

Original article
mothers with a primary infection had similar audiological and neurodevelopmental outcomes as those born to mothers with a nonprimary infection.
We documented a higher proportion of infants born to mothers with a primary infection (76%) versus nonprimary infection, compared with a previous report of countries with a low seroprevalence, like Switzerland (44%) [5].This finding most likely resulted from the systematic testing for CMV among newborns born to mothers with se-roconversion at our centre.We report a lower proportion of symptoms among symptomatic newborns with cCMV infection compared with others studies [15,16], whereas our proportion of neurological symptoms among symptomatic newborns was similar.These discrepancies could result from a less rigorous assessment of clinical and laboratory outcomes than of neurological outcomes in our setting compared with other studies [15,16].Careful clinical and laboratory assessment is warranted in all newborns with  cCMV infection, as petechiae and SGA, as well as thrombocytopenia, were identified as significant predictors of subsequent SNHL [42].
Our study reported a similar rate of SNHL or abnormal neurodevelopmental outcomes at 3 months of age and onwards compared with other studies (24% vs 17-20%) [9].All our children with subsequent SNHL or abnormal neurodevelopmental scores were symptomatic at birth, in contrast to a systematic review [9] that reported that only one third of children with subsequent SNHL or abnormal neurodevelopmental outcomes were symptomatic at birth.As supported by the systematic review [9], neurodevelop-mental and audiological follow-up should be offered to all children with cCMV infection identified through universal newborn screening [9,43].
Confirming findings of prior reports [4,7], our study indicated that infants born to mothers with a nonprimary infection are at least as often symptomatic at birth as those born to mothers with a primary infection.As documented in other studies [4,7], we report equivalent rates of abnormal audiological and neurodevelopmental outcomes of children born to mothers with a primary infection compared with those born to mothers with a nonprimary infection.Our exploratory study represents an important first step in reporting audiological and neurodevelopmental outcomes of newborns with cCMV infection in Switzerland.An important strength relates to the precise documentation of BSER measures, including best ears analysis and total ears analysis, as well as the use of standardised developmental tests [16].Limitations of our study include the retrospective design, inhomogeneous timeframe of the audiological follow-up and the absence of follow-up of asymptomatic newborns, which may have biased our findings towards more severe outcomes among symptomatic children with longer follow-up.This limitation is inherent to all retrospective studies on cCMV infections conducted before the establishment of recent guidelines on auditory outcomes measures [3], as reflected in a recent European surveillance study [44].However, at each visit at the neurodevelopmental clinic, hearing was examined by the paediatrician and those with any abnormal findings were systematically referred to the ENT specialist.Therefore, we believe that the wide timeframe of auditory assessment conducted by the ENT specialist minimally affected our findings.Other shortcomings of our study relate to the small sample size inherent to monocentric retrospective studies conducted in a country with a low prevalence of cCMV infection [5].In addition, the absence of universal screening of newborns may have biased our study towards selecting mothers with a primary infection and mostly symptomatic newborns, who presented mostly with normal BSER and neurodevelopmental scores.
In conclusion, our study supports current findings on the burden and severity of cCMV infections [4,7,9] and provides national data on outcomes of newborns with cCMV infection.Combined information on clinical outcomes from the current study and epidemiological data provided by the Swiss Paediatric Surveillance Study (SPSU) will hopefully encourage the establishment of a national registry, which will in turn serve as foundation for future prospective studies.

Figure 1 :
Figure 1: Clinical characteristics at birth and audiological and neurodevelopmental outcomes.

Audiological and neurodevelopmental outcomes of all patients at 3 months of age and onwards
52%) infants born between 2001 and 2014.Of these 26, 21 (81%) were offered follow-up as a result of neurologically symptomatic cCMV infection.Of those 26, 6 (23%) patients presented with abnormal neurodevelopment of whom 3 had SNHL (table S2, appendix 1).

Table 1 :
Baseline characteristics at birth of all newborns with congenital cytomegalovirus infection.

Table 2 :
Baseline characteristics, audiological and neurodevelopmental outcomes of children with sequelae.

Table 3 :
Clinical characteristics at birth categorised by newborns born to mothers with primary versus nonprimary infection.

Table 4 :
Audiological and neurodevelopmental outcome at 3 months of age and onwards, categorised by newborns born to mothers with a primary versus a nonprimary infection.
Swiss Medical Weekly • PDF of the online version • www.smw.chPublished under the copyright license "Attribution -Non-Commercial -No Derivatives 4.0".No commercial reuse without permission.See http://emh.ch/en/services/permissions.html.