International Journal of Pediatric Otorhinolaryngology (2005) 69, 1191—1198

www.elsevier.com/locate/ijporl

Audiometric, clinical and educational outcomesin a pediatric symptomatic congenitalcytomegalovirus (CMV) population withsensorineural hearing loss

Colm Madden a, Susan Wiley b, Mark Schleiss c, Corning Benton d,Jareen Meinzen-Derr e, John Greinwald a, Daniel Choo a,*

aCenter of Hearing and Deafness Research, Division of Otolaryngology-Head and Neck Surgery,Department of Pediatric Otolaryngology, OSB-3, Cincinnati Children’s Hospital Medical Center,3333 Burnet Ave., Cincinnati, OH 45229, USAbDivision of Developmental Disabilities, Cincinnati Children’s Hospital Medical Center,3333 Burnet Ave., Cincinnati, OH 45229, USAcDivision of Infectious Disease, Cincinnati Children’s Hospital Medical Center,3333 Burnet Ave., Cincinnati, OH 45229, USAdDivision of Radiology, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave.,Cincinnati, OH 45229, USAeCenter for Epidemiology and Biostatistics, Cincinnati Children’s Hospital Medical Center,3333 Burnet Ave., Cincinnati, OH 45229, USA

Received 6 December 2004; accepted 17 March 2005

KEYWORDSCongenital;Cytomegalovirus;Sensorineural;Hearing loss;Neurology

Summary

Objective: To correlate audiometric findings and outcomes with the clinical, radi-ological and educational findings in a symptomatic congenital cytomegalovirus (CMV)population with sensorineural hearing loss.Methods: A retrospective review of data from 21 symptomatic congenital CMVpatients identified in a pediatric hearing impaired database of 1500 patients. Clinicaldata, audiometric thresholds and outcomes, radiographic abnormalities, communi-cation and educational achievements were used as outcome measures.Results: Twenty-one patients were identified with symptomatic congenital CMVinfection at birth; 5 with unilateral hearing loss and 16 with bilateral hearing loss.The median initial pure-tone average (PTA) for the 21 subjects was 86 dB and themedian final PTA was 100 dB. Progression of hearing loss was seen in 9 patients (43%).Neurological and radiological sequelae of symptomatic CMV infection were seen in

* Corresponding author. Tel.: +1 513 636 4870; fax: +1 513 636 2886.E-mail address: daniel.choo@cchmc.org (D. Choo).

0165-5876/$ — see front matter # 2005 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.ijporl.2005.03.011

1192 C. Madden et al.

81% of affected patients. Children with neurological dysfunction were significantlymore likely to rely on special education ( p = 0.045). There was a significant correla-tion between the severity of the initial PTA and the development of a progressivehearing loss ( p = 0.0058). Initial hearing thresholds were significantly better in thosechildren with a history of jaundice ( p = 0.002), hepatosplenomegaly (HSM)( p = 0.022) and cerebral palsy (CP) ( p = 0.013). There was a significant correlationbetween a less severe final PTA and the presence of CP ( p = 0.005). A history of mentalretardation in children was significantly associated with poorer communication skills( p = 0.043).Conclusions: The severity of neurological manifestations in congenital symptomaticCMV infection was positively correlated with the need for total and manual commu-nication and the reliance on special education. Statistical associations betweenclinical findings such as hepatic dysfunction, CP and hearing level were identifiedhowever plausible mechanisms explaining these associations remain ambiguous andare discussed in the context of this complex population of children with congenitalsymptomatic CMV.# 2005 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Congenital cytomegalovirus (CMV) infection is aleading cause of neurological disease and hearingloss in children. Most of the infants with congenitalCMV infection are without apparent clinical mani-festations at birth (asymptomatic) and only 5—10%of the neonates exhibit findings suggestive of con-genital infection (symptomatic infection). The aver-age infection rate is 1% of all live births [1—3]. Anestimated 40,000 infants in the USA are born eachyear with congenital CMV infection and about 7000of these children will develop permanent centralnervous system defects [1,4,5].

Symptomatic congenital CMV infection manifestsitself as a triad of jaundice (62—76%), petechiae(58—79%) and HSM (50—88%). Microcephaly (21—53%), intra-uterine growth retardation (33—50%)and/or premature birth (25—34%) are also found[6—10].

The most common neurological defect (30—65%prevalence) is sensorineural hearing loss (SNHL).Other sequelae of CMV infection are mental retar-dation, cerebral palsy (CP), seizures and visualdefects [6,7]. The presence of specific clinical fea-tures, associated with CMV infection, has beenassessed to see if they can predict hearing andneurodevelopmental outcomes. The presence ofan abnormal cranial computed tomography (CT)scan, microcephaly, chorioretinitis or other neuro-logical abnormalities at birth have been shown toforecast an adverse hearing and neurodevelopmen-tal outcome [8,9,11].

CMV infection is, at this moment, the only rele-vant viral cause of perinatal hearing loss as rubella,measles and mumps have become rare due to vac-cination. CMV is the leading cause of congenital non-hereditary sensorineural hearing loss in children

[2,12]. Peckham et al. estimated that 12% of allcases of congenital bilateral sensorineural deafnesswere caused by congenital CMV infection [13]. Awide variety of audiometric thresholds and out-comes have been described in this patient popula-tion.

Onset of SNHL after congenital CMV infection maybe immediate (at birth) or delayed, with variabilityin the severity of the hearing loss in the affectedchildren. This affects the early identification ofhearing loss in this population and has implicationsfor the implementation of newborn hearing screen-ings [14]. Screening for CMV infection in all neonateswould improve the identification of neonatal hear-ing loss and identify those at risk of developing ahearing loss. Children with symptomatic infectionhave hearing loss at an earlier age and with greaterseverity than children with asymptomatic infection.Further deterioration or progression of the hearingloss will occur in a proportion of children withcongenital CMV [3,15—18]. Long-term follow-upand serial audiometric testing and assessment iscrucial in this population.

Based upon temporal bone studies, CMV-relatedSNHL is postulated to arise from a labyrinthitisinduced by CMV that enters the endolymph viathe stria vascularis [19]. Infected epithelial cellshave been found in the walls of the inner ear andstria vascularis, with pronounced hydrops of thecochlear duct and saccule also noted [20]. CMVhas also been cultured from perilymph and viralantigen production widely demonstrated in thespiral ganglion neurons and organ of corti [21,22].A review of published temporal bone studies demon-strated the typical inclusion cells of CMV infection inthe scala media, Reissner’s membrane and the striavascularis of children who died of symptomaticcongenital CMV infection [22,23].

Audiometric finding in pediatric CMV population with sensorineural hearing loss 1193

The purpose of this study was to examine theaudiometric outcomes in a symptomatic congenitalCMV population with SNHL and to compare thesedata with the clinical, radiological and educationalfindings in this population. We further sought toidentify the factors associated with communication,educational and audiological outcomes in thesechildren. We hypothesize that CMV-associated neu-rologic injury in newborns will positively correlatewith the presence of poorer hearing, language,developmental and educational outcomes as com-pared to children without neurologic manifestationsof CMV.

2. Patients and methods

An otologic database consisting of over 1500 chil-dren diagnosed with a permanent sensorineuralhearing loss was queried for children with a provensymptomatic congenital cytomegalovirus infection.Twenty-one children in the database had a diagnosisof CMV, based on the protocol of isolating the virusfrom urine or saliva during the first 2 weeks of theirlife [24,25]. The charts of these 21 children werethen reviewed for additional clinical informationincluding presenting history, neonatal and past med-ical histories, imaging data as well as their labora-tory and clinical examination findings. The presenceor absence of prematurity, microcephaly, neonataljaundice, hepatosplenomegaly (HSM), cerebralpalsy, seizure activity andmental retardation (intel-ligence quotient <70) was noted. Communicationmode was categorized as oral, manual or totalcommunication in nature, and the educationalenvironment of the child was noted.

2.1. Audiological testing

Audiometric testing was performed on all childrenincluded in this report. The test protocol wasselected according to the developmental age ofthe child; thresholds for auditory brainstemresponses, soundfield or earphone visual reinforce-ment audiometry, play audiometry or conventionalaudiometric testing from 250 to 8000 Hz. A pure-tone average (PTA) was derived by averaging theaudiometric thresholds at 500, 1000, 2000 and4000 Hz. For the sake of this study, the PTAs fromthe better-hearing ear are reported. Hearing out-come was defined as being stable, fluctuating, pro-gressive or fluctuating and progressive [16,18]. Theslope and configuration of the audiogram was cate-gorized as flat, down-sloping, up-sloping or cookie-bite shaped. We defined mild hearing loss as a loss of�45 dB, moderate loss as a loss of 46—70 dB, severe

loss as a loss of 71—90, and profound hearing loss as aloss >90 dB.

2.2. Radiological protocol

Non-contrast cranial computed tomography andmagnetic resonance imaging (MRI) scans of the chil-dren were analyzed for the presence or absence ofintracerebral calcification(s), ventricular dilata-tion, white matter loss, demyelination and neuronalmigrational abnormalities.

2.3. Statistical analysis

All data were analyzed using SAS, version 8.2. A p-value of 0.05 or less was considered significant. TheFisher’s exact test was used to assess whether pro-gression, configuration, education, treatment andcommunication outcomes were associated with thepresence of the following medical conditions: pre-maturity, jaundice, HSM, microcephaly, CP, seizureactivity, cerebral calcification, ventricular dilata-tion, mental retardation and white matter loss.Pathologically similar conditions were groupedtogether and then analyzed using the same proce-dure as described above. These included evidenceof CNS infection, abnormal radiological findings andevidence of non-neurological infection. To deter-mine if the median initial or median final pure toneaverage was different in those with specific present-ing features (see above medical conditions), theWilcoxon rank-sum test was used.

3. Results

3.1. Study population

Twenty-one children with a documented sympto-matic congenital CMV infection were identified.The hearing loss was bilateral in 16 of the 21 chil-dren (76%), while 5 children (24%) had a unilateralloss. The population was largely Caucasian (n = 16).Four were African American and 1 Hispanic. Theaverage age of our patient population was 12 yearsold, ranging from 3.4 years to 17.1 years and 61%(n = 13) weremale. None of the children in our studyreceived anti-viral or immunosuppressant therapyfor their CMV infection.

Patients were followed for a median of 97 months(range 8—158 months). The median initial pure-toneaverage was 77 dB (range of 20—130 dB); the lastknown PTA was 99 dB (35—130 dB) (Table 1). Themedian age of initial PTA was 35 months (6—151.6months). In the group that showed progressive loss,5 children had a final PTA in the profound range

1194 C. Madden et al.

Table 1 Audiologic findings of 21 subjects

Subject Age(months)at initial PTA

Age(months)at last PTA

PTAinitial

PTAlast

Laterality Typeof lossa

Audioconfigb

Managementof HL

Initialseverityof HL

Finalseverityof HL

1 27 164 91 122 Bi P D HA Profound Profound2 99 50 35 Uni F F Observe Mild Mild3 70 70 Uni S F Observe Moderate Moderate4 13 167 77 130 Bi P F HA Severe Profound5 46 157 60 67 Bi FP U, C HA Moderate Moderate6 99 114 121 120 Bi S D, F HA CI Profound Profound7 112 125 121 122 Bi S D HA Profound Profound8 92 100 67 65 Uni S D HA Moderate Moderate9 23 26 59 72 Bi P F HA Moderate Severe

10 152 153 101 100 Uni S D Pref. FM Profound Profound11 30 70 122 126 Bi P D, F HA Profound Profound12 43 140 86 99 Bi P D HA Severe Profound13 31 118 25 60 Bi P D HA Mild Moderate14 31 78 130 130 Bi S F HA Profound Profound15 35 101 66 85 Uni P D Observe Moderate Severe16 42 20 97 Bi P D HA Mild Profound17 33 34 44 41 Bi S D HA Mild Mild18 51 57 70 59 Bi S F Observe Moderate Moderate19 40 62 130 130 Bi S F, D HA Profound Profound20 6 41 130 130 Bi S F HA Profound Profound21 29 164 130 130 Bi S F HA Profound Profounda P, progressive; F, fluctuating; S, stable; FP, fluctuating progressive.b D, down-sloping; F, flat; C, cookie bite, U, upslanting: children with bilateral hearing loss may have different configurations for

each ear.

(>90 dB). However, 1 patient was already in theprofound range (initial PTA >90 dB) and the clinicalsignificance of such ‘‘progression’’ is questionable.In contrast, 1 patient had an initial PTA of 20 dB andprogressed to a profound hearing loss while the 3other patients started with an initial PTA in thesevere range and progressed onto profound hearingloss levels. Audiometric follow-up showed a stablehearing level in 11 patients (52%), a progressivehearing loss in 8 patients (38%), with a fluctuatingloss in 1 patient (5%) and a fluctuating and progres-sive loss in 1 patient (5%).

The median initial PTA for those with a progres-sive loss was 66 dB and for those with non-progres-sive loss was 104 dB (range of 20—122 dB and 44—130 dB, respectively). The last PTA value recordedfor those with a progressive loss was 97 dB and thosewith a non-progressive loss was 100 dB (range of 45—130 dB and 35—130 dB, respectively). Seven chil-dren with non-progressive hearing loss had slightimprovement when the final PTA was recorded.However, this improvement (ranging from 1 to15 dB) was not considered a clinical improvement(Table 1).

The severity of the initial PTA was a significantrisk for the development of a progressive hearingloss, as those patients with progressive hearing losshad better initial PTA thresholds (p = 0.059). The

most common audiometric configurations weredownsloping in 18 ears (48%) and flat in 17 ears(46%). A cookie-bite shaped audiogram and anupsloping audiogram were less frequent. Progres-sive hearing loss seemed to be significantly asso-ciated with a downsloping audiogram configuration(9 of 17 ears; p = 0.04). Among children with a flataudiogram configuration, 54% had a history of micro-cephaly (p = 0.021).

Of the five children with unilateral loss, one had amild fluctuating loss, two had a stable moderateloss, one had a stable profound loss and one childhad a moderate hearing loss that progressed tosevere levels (Table 1).

3.2. Clinical and radiologic features

Table 2 contains the clinical features and radiologicfindings of the study subjects. A history of prema-turity was found in 7 children (33%). Neonatal jaun-dice was found in 6 children (28%). Five children(24%) had hepatosplenomegaly (HSM). Cerebralpalsy (CP) was documented in 7 children (33%). Fourchildren (19%) had a history of a seizure disorder.Fourteen children (66%) had mental retardation andmicrocephaly was documented in 6 children (28%).Three (14%) had CMV retinitis. Most children hadmore than one clinical problem; 2 had hearing loss as

Audiometric finding in pediatric CMV population with sensorineural hearing loss 1195

Table 2 General characteristics and clinical features of 21 subjects

Subject Gender Race Age of HLidentification(months)

Clinical featuresa Treatment Educationalsettingb

Communicationmode reported

1 F W 19.3 HA Deaf Manual2 M B 27.2 MC, CP, MR Observe Sp Ed Manual3 M H 23.4 MC, CP, SZ, MR, EN Observe Sp Ed Total4 F B 13.1 J, HS, MR HA Deaf Manual5 F B 0.7 Prem HA Main Oral6 M W CI & HA Deaf Total7 M W CP HA Sp Ed Manual8 M W J, V, EN HA Main Oral9 M W 5.7 Prem, J, HS, MC,

CP, SZ, MR, ENHA Sp Ed Total

10 M W MR Pref. FM Main Oral11 M W 12.0 Prem HA Deaf Manual12 M W 38.7 V, MR HA Deaf Total13 M B 12.0 Prem, J, CP, MR HA Sp Ed Oral14 M W 12.8 MR HA Sp Ed Manual15 F W 3.8 HS, CP Observe Main Oral16 F W 41.9 Prem, J, HS, Cal, MR HA HI class Total17 M W 32.9 J, HS, MR HA Sp Ed Total18 F W 12.0 Prem, MC, CP, Cal, MR, EN Observe Sp Ed Total19 F W 12.7 SZ, MR, EN HA HI class Total20 F W 1.6 Prem, MC, Cal, MR, EN HA Sp Ed Total21 M W 0.4 MC, SZ, Cal, MR HA Unknown Manuala MC, microcephaly; CP, cerebral palsy; MR, mental retardation; SZ, seizures; EN, encephalopathy; J, jaundice; HS, hepatosple-

nomegaly; Prem, premature; V, ventricular dilatation; Cal, calcification; RT, CMV retinitis; WM, white matter loss.b Deaf, school for the deaf; Sp Ed, special education class; HI class, hearing impaired class; Main, mainstream school.

their only clinical problem. The range of differentclinical features a child could have in this populationranged from 1 to 8 (1 child had 8 different features).Children with only one clinical problem had a med-ian age at initial PTA of 73 months as compared to 33months in children with greater than one clinicalproblem, although this finding was not statisticallydifferent ( p = 0.12). There was also no statisticaldifference between the initial or final PTA thresh-olds based on the number of clinical features a childhad (initial PTA with 1 problem verses >1 problem:104 dB versus 70 dB, respectively).

Imaging was available in 15 of the children. Sixchildren had cerebral calcifications noted on theirimaging (40%) and 5 children had ventricular dilata-tion (33%). White matter loss was noted in 3 of thechildren (20%). Overall, an abnormal radiologicalscan was noted in 10 of the children (67% of thosewith scans). Three children hadfindings in addition tocalcifications, ventricular dilatation, and white mat-ter loss. These included a migrational abnormality,encephalomalacia, and a Chiari I malformation withhydromyelia. No abnormality of the cochleovestibu-lar systemwas seen in thenine childrenwhohadhigh-resolution temporal bone CT scans (Table 2).

Table 3 describes the audiometric findings (of thebetter hearing ear) grouped by clinical features.

Fifteen children (71%) had more than one clinicalfinding. When comparing the presence or absence ofthese clinical features with the audiometric out-comes, a history of jaundice or HSM was associatedwith statistically significantly better hearing thresh-olds at initial assessment ( p < 0.05), while a historyof CP was associated with statistically significantlybetter hearing thresholds at final assessment. How-ever, of the seven children with CP, one also hadHSM, one had history of jaundice, and one child hadboth HSM and history of jaundice. Five children hadHSM and six children had a history of jaundice; fourof these children had a diagnosis of HSM and ahistory of jaundice. CP was the only clinical featuresstatistically associated with better hearing thresh-olds at the final assessment. Although not statisti-cally significant, there seemed to be a trend towardsworse hearing loss at both the initial and final PTArecordings with having either a seizure disorder or adocumented abnormal scan.

3.3. Rehabilitation, communication andeducational outcomes

Ten children (48%) used solely a hearing aid; 5children (24%) also used a frequency modulation(FM) system at school in conjunction with a hearing

1196 C. Madden et al.

Table 3 Clinical features and audiometric findings

Median dB initial PTA Median dB final PTA

Yes No p-Value Yes No p-Value

Prematurity 65 91 0.16 72 120 0.24Neonatal jaundice 59 114 0.002 72 121 0.08Hepatosplenomegaly 59 104 0.023 85 117 0.42Seizure disorder 124 85 0.29 114 100 0.43Microcephaly 70 87 0.74 72 114 0.74Mental retardation 77 99 0.99 100 121 0.96Cerebral palsy 66 104 .013 70 122 0.005CMV retinitis 70 86.5 0.76 60 107 0.59Abnormal scan 124 77 0.05 114 99.5 0.32>1 Clinical feature 70 101 0.44 91 120 0.61

aid. Four children (19%) were observed without anyspecific treatment for their hearing loss (majoritywith unilateral loss), 1 child used an FM system atschool alone (5%) and 1 child derived no benefit fromhearing aids and went on to have a cochlear implant(5%).

Observation and FM system alone were morelikely to occur in children with unilateral hearingloss (3/5). Two of these three were oral commu-nicators. Of the remaining two children with obser-vation alone, one had CPand one had significant CNSfindings–—microcephaly and mental retardation.Both relied on total communication.

There was significant variability in educationalsettings and communication modes. Seventy-sixpercent required specialized educational supportin non-mainstream settings. The majority reliedon total communication or manual communication(Table 4). The use of non-oral communication wassignificantly associated with the presence of mentalretardation ( p = 0.043).

Table 4 Hearing, communication and educationaloutcomes

Outcomes n = 21

Amplification optionsObservation 4 (19%)FM system alone 1 (5%)Hearing aids 10 (48%)Hearing aid and FM system 5 (24%)Cochlear implant 1 (5%)

Communication modeOral 5 (24%)TC 9 (43%)Manual 7 (33%)

Educational settingMainstreamed 4 (19%)School for the deaf 5 (24%)Hearing impaired classroom 2 (9%)Special needs setting (MH) 10 (48%)

Of the four children in mainstream settings, allwere oral. Five children were oral communicators,only one of which was in a special education setting.The remainder of the children was mainstreamed.Three of the oral communicators had unilateral loss,two with bilateral loss in the mild to moderaterange. Two of the oral communicators had bothmental retardation and cerebral palsy.

Children with evidence of neurological disease,characteristic of symptomatic congenital CMV infec-tion, were significantly associated with the need forgreater educational support (63% of CNS disease inspecial education setting versus 0; p = 0.03). Thiswas particularly true of children with a history ofmicrocephaly (100% versus 27%; p = 0.02), cerebralpalsy (86% versus 29%; p = 0.03), and mental retar-dation (64% versus 14%; p = 0.05). There was nosignificant relationship between audiological find-ings and educational and communication outcomes.

Thechildrenwith cerebral palsywereof particularinterest. They were a largely heterogeneous group inregards to clinical features; however, audiologically,they seemedtohave less severeoutcomes.Twoof thesix childrenwith CPhad improvements in their audio-grams, one improved from moderate to mild and theother from severe to moderate. One had a stableaudiogram in the profound range. Three had a uni-lateral loss. The tendency towards preservation ofhearing was significantly associated with a history ofCP (p = 0.005).

4. Discussion

Children with sensorineural hearing loss from symp-tomatic congenital CMV vary widely in their clinicalfeatures, audiometric outcomes, communicationmode, and need for educational support [5,9,10].

A progressive hearing loss was found in 43% of ourpatient population, similar to other reports[7,15,18]. As the majority of children with progres-

Audiometric finding in pediatric CMV population with sensorineural hearing loss 1197

sive loss began with relatively good audiometricthresholds and progressed to severe or profoundhearing loss, this group warrants close monitoringand aggressive management. It is important torecognize that our descriptive data do not providea mechanism that explains why children with betterinitial PTAs tended to display progressive hearinglosses over time. One plausible hypothesis is thatthe natural history of hearing in congenital CMV isprogressive hearing loss. Those children with betterinitial PTAs essentially have more testable hearingthat can demonstrate deterioriation over time(when compared to those children already showinga profound hearing loss). An equally plausiblehypothesis is that CMV-related hearing loss is uni-formly progressive but over varying timeframes indifferent patients. In such a model, our medianfollow up period of 97 months may not be adequateto define the variable course of hearing loss in thispopulation.

Regardless of mechanism, however, these chil-dren with initially serviceable hearing should bemonitored for progression as they may be appro-priate candidates for cochlear implantation if theirhearing loss does progress. The average time overwhich progressive hearing loss was observed in ourpatients was 60.5 months (range 35—118 months).Such data suggest that long-term follow up of CMV-affected children is required in order to avoid miss-ing ‘‘later’’ progression of hearing loss. Audiogramconfiguration was also significant. A history ofmicrocephaly was associated with a flat configura-tion which may suggest insult to the central auditorypathways by CMV infection.

Although, congenital CMV infection has been pos-tulated to cause anomalies or growth disturbances ofthe temporal bone [26], a review of the high-resolu-tion temporal bone imaging in our patient populationfound no evidence of any cochleovestibular anoma-lies. Additionally, in our population, the presence ofradiological abnormalities was not associated withpoorer audiometric, communication and educationaloutcomes; however a large number of patients didnot have documented scans.

There is often an assumption that symptomaticCMV infection is associated with a poor educationaloutcome [7]. However, some of the children in ourpopulation were appropriately functioning in main-stream educational settings. Those with milderhearing loss and unilateral loss were more likelyto be served in this setting. As might be suspected,children with multiple clinical problems and/ormore severe hearing loss were more likely to requiregreater levels of educational support.

Neurological involvement and subsequent globalneurological dysfunction were significantly asso-

ciated with enrolment special education settings.There was a greater dependency on non-oral meth-ods of communication in children with mental retar-dation.

The finding of a trend towards preservation ofhearing in the children with cerebral palsy group isintriguing. Our data support prior reports that dis-cuss the lack of predictive value of microcephalyand neurological findings for progressive SNHL [27].Intracranial calcifications have been associated withprogressive hearing loss on univariate analysis [27].Both an abnormal CT and microcephaly have beenassociated with motor disability and mental retar-dation [11,28]. We only had imaging data on twopatients with cerebral palsy. Therefore, it is difficultto postulate about these features related to hearingloss in our group. Themechanism or pathophysiologyfor this finding in our population is still uncertain. Inaddition, the progression of hearing loss in ourpopulation may not be complete, as our range offollow-up is variable, with some of our populationstill fairly young.

The pathophysiology of CMV on neuro-develop-ment and hearing continues to raise questions. Thegreater risk for hearing loss and an adverse neuro-developmental outcome in the presence of brainabnormalities on neonatal imaging, supports theidea that inner ear infection with CMV is associatedwith central nervous system infection [11].

It is also of particular interest that in childrenwith only one clinical feature of symptomatic con-genital CMV were more likely to have a much laterage of initial audiometric testing (69 months ascompared with 31 months). These children maybe less aggressively managed based on the percep-tion that they have less severe disease. The delayedidentification in less severely affected children hasimplications for universal CMV screening.

Our small sample size (n = 21) limits our ability tofind statistically significant differences in some ofour comparisons and this study is retrospective indesign. A possible referral bias may be present, aschildren referred to our tertiary care center may bemore severely affected than those initially diag-nosed at our center. We did not see a differencein the severity of hearing loss between these twogroups (data not shown).

Children with symptomatic congenital CMV infec-tion exhibited diverse audiological, clinical andeducational outcomes. In our study, the presenceof central nervous system infection is associatedwith a flat audiogram configuration, non-oral com-munication, increased reliance on special educationand poorer audiological rehabilitation, but notdecreased audiometric thresholds or the risk ofprogression.

1198 C. Madden et al.

Acknowledgements

The authors would like to thank Stacy Poe, M.S., andJudy Bean, Ph.D., of the Department of Biostatisticsat Children’s Hospital Medical Center, Cincinnati,Ohio, for their statistical analysis of the data.

References

[1] G.J. Demmler, Summary of a workshop on surveillance forcongenital cytomegalovirus disease, Rev. Infect. Dis. 13(1991) 315—329.

[2] T. Hicks, K. Fowler, M. Richardson, A. Dahle, L. Adams, R.Pass, Congenital cytomegalovirus infection and neonatalauditory screening, J. Pediatr. 123 (1993) 779—782.

[3] S. Stagno, R.F. Pass, M.E. Dworsky, C.A. Alford, Congenitaland perinatal cytomegalovirus infections, Semin. Perinatol.7 (1993) 31—42.

[4] S. Stagno, R.F. Pass, M.E. Dworsky, C.A. Alford, Maternalcytomegalovirus infection and perinatal transmission, Clin.Obtset. Gynaecol. 25 (1982) 563—576.

[5] F.P. McCollister, L.C. Simpson, A.J. Dahle, R.F. Pass, K.B.Fowler, A.S. Amos, et al. Hearing loss and congenitalsymptomatic cytomegalovirus infection: a case report ofmultidisciplinary longitudinal assessment and intervention,J. Am. Acad. Audiol. 7 (1996) 57—62.

[6] R.F. Pass, S. Stagno, G.J. Myers, C.A. Alford, Outcome ofsymptomatic congenital cytomegalovirus infection: resultsof long-term longitudinal follow-up, Pediatrics 66 (1980)758—762.

[7] W.D. Williamson, M.M. Desmond, N. LaFevers, L.H. Taber,F.I. Catlin, T.G. Weaver, Symptomatic congenital cytomega-lovirus: disorders of language, learning and hearing, Am. J.Dis. Child 136 (1982) 902—905.

[8] T.J. Conboy, R.F. Pass, S. Stagno, C.A. Alford, G.J. Myers,W.J. Britt, et al. Early clinical manifestations and intellec-tual outcome in children with symptomatic congenital cyto-megalovirus infection, J. Pediatr. 111 (1987) 343—348.

[9] S.B. Boppana, R.F. Pass, W.J. Britt, S. Stagno, C.A. Alford,Symptomatic congenital cytomegalovirus infection: neona-tal morbidity and mortality, Pediatr. Infect Dis. J. 11 (1992)93—99.

[10] K.B. Fowler, S. Stagno, R.F. Pass, W.J. Britt, T.J. Boll, C.A.Alford, The outcome of congenital cytomegalovirus infec-tion in relation to maternal antibody status, N. Engl. J. Med.326 (1992) 663—667.

[11] S.B. Boppana, K.B. Fowler, K. Yaid, G. Hedlund, S. Stagno,W.J. Britt, et al. Neuroradiographic findings in the newbornperiod and long-term outcome in children with symptomaticcongenital cytomegalovirus infection, Pediatrics 99 (1997)409—414.

[12] S. Harris, K. Ahlfors, S. Ivarrson, B. Lemmark, L. Svanberg,Congenital cytomegalovirus infection and sensorineuralhearing loss, Ear Hear 6 (1984) 354—355.

[13] C.S. Peckham, O. Stark, J.A. Dudgeon, Martin JAM, G.Hawkins, Congenital cytomegalovirus infection: a cause of

sensorineural hearing loss, Arch. Dis. Child 62 (1987) 1233—1237.

[14] K.B. Fowler, A.J. Dahle, S.B. Boppana, R.F. Pass, Newbornhearing screening: will children with hearing loss caused bycongenital cytomegalovirus infection be missed? J. Pediatr.135 (1999) 60—64.

[15] A.J. Dahle, F.P. McCollister, S. Stagno, D.W. Reynolds, H.E.Hoffman, Progressive hearing impairment in children withcongenital cytomegalovirus infection, J. Speech Hear Dis-ord. 44 (1979) 220—229.

[16] K.B. Fowler, F.P. McCollister, A.J. Dahle, S. Boppana, W.J.Britt, R.F. Pass, Progressive and fluctuating sensorineuralhearing loss in children with asymptomatic congenital cyto-megalovirus infection, J. Pediatr. 130 (1997) 624—630.

[17] W.D. Williamson, G.J. Demmler, A.K. Percy, F.I. Catlin,Progressive hearing loss in infants with asymptomatic con-genital cytomegalovirus infection, Pediatrics 90 (1992) 862—866.

[18] A.J. Dahle, K.B. Fowler, J.D. Wright, S.B. Boppana, W.J.Britt, R.F. Pass, Longitudinal investigation of hearing dis-orders in children with congenital cytomegalovirus, J. Am.Acad. Audiol. 11 (2000) 283—290.

[19] M. Strauss, G.L. Davis, Viral disease of the labyrinth. Part I:Review of the literature and discussion of the role of cyto-megalovirus in congenital deafness, Ann. Otol. 82 (1973)577—583.

[20] E.N. Myers, S. Stool, Cytomegalic inclusion disease of theinner ear, Laryngoscope 78 (1968) 1904—1914.

[21] L.E. Davis, L. Johnsson, M. Kornfeld, Cytomegalovirus labyr-inthitis in an infant: morphological, virological, and immu-nofluorescent studies, J. Neuropathol. Exp. Neurol. 40(1981) 9—19.

[22] S. Stagno, D.W. Reynolds, C.S. Amos, A.J. Dahle, F.P. McColl-ister, I. Mohindra, et al. Auditory and visual defects resultingfrom symptomatic and subclinical congenital cytomegalo-viral and toxoplasma infection, Pediatrics 59 (1977) 669—678.

[23] M. Strauss, Human cytomegalovirus labyrinthitis, Am. J.Otolaryngol. 5 (1990) 292—298.

[24] S. Stagno, R.F. Pass, G. Cloud, W.J. Britt, R.E. Henderson,P.D. Walton, et al. Primary cytomegalovirus infection inpregnancy: incidence, transmission to fetus and clinicaloutcome, JAMA 256 (1986) 1904—1908.

[25] K.B. Balcarek, W. Warren, R.J. Smith, M.D. Lyon, R.F. Pass,Neonatal screening for congenital cytomegalovirus infectionby detection of virus in saliva, J. Infect. Dis. 167 (1993)1433—1436.

[26] N.M. Bauman, L.J. Kirby-Keyser, K.D. Dolan, D. Wexler, B.J.Gantz, B.F. McCabe, et al. Mondini dysplasia and congenitalcytomegalovirus infection, J. Pediatr. 124 (1994) 71—78.

[27] L.B. Rivera, S.B. Boppana, K.B. Fowler, W.J. Britt, S. Stagno,R.F. Pass, Predictors of hearing loss in children with sympto-matic congenital cytomegalovirus infection, Pediatrics 110(2002) 762—767.

[28] D.E. Noyola, G.J. Demmler, C.T. Nelson, C. Griesser, W.D.Williamson, J.T. Atkins, The Houston Congenital CMV Long-itudinal Study Group, Early predictors of neurodevelopmen-tal outcome in symptomatic congenital cytomegalovirusinfection, J. Pediatr. 138 (2001) 325—331.