Management of congenital cytomegalovirus infection: an evidence-based approach


RS Gandhi, MRCPCH, Brighton and Sussex University Hospitals NHS Trust, 5, Barnfield Gardens, Brighton BN2 0HQ, UK.
Tel: +44-1273 248774 |
Fax: +44-1273 664435 |


inline image


Despite cytomegalovirus being the most common congenital infection leading to psychomotor impairment and sensori-neural hearing loss, little is known about early identification and management of congenitally infected neonates. This article reviews the literature and devises an algorithm for identification and management of these neonates.

Conclusion:  Application of the current knowledge in the management of congenital cytomegalovirus infected neonates could be beneficial, until further evidence is available.


Cytomegalovirus (CMV) is a virus that causes infection in people of all ages. It is the commonest cause of infection related congenital abnormality in newborns and also the leading cause of non-hereditary sensori-neural hearing loss (SNHL) in early childhood.

Human CMV is an enveloped DNA virus and like other members of the Herpes virus family establishes life-long latency following primary infection. Identification of neonates with congenital CMV (CCMV) infection can be difficult as to date universal screening of pregnant females for CMV infection is not in place and most neonates are asymptomatic at birth. Females of childbearing age are most at risk of acquiring CMV from exposure to saliva and urine of young children, as they tend to shed the virus for many months to years after their first infection, which is often asymptomatic. Congenital infection can result from primary or pre-existing infection during pregnancy. Pregnant females can transmit the virus to the foetus either transplacentally or during birth and post-natally via breast milk (1,2).

The overall birth prevalence of CCMV is 0.64% worldwide; 10–15% of the newborns with CCMV are symptomatic (1). More than 70% of these will develop permanent long-term sequelae at follow-up, namely: cerebral palsy, delayed psychomotor development, mental retardation, expressive language delay and learning disability, epilepsy, optic atrophy and SNHL (1,3–5). Of the asymptomatic children, 6–25% will develop late-onset sequelae (SNHL and neurological deficits) (1,2,4–7).

Diagnostic and therapeutic options for CCMV are increasingly been investigated, but evidence-based consensus recommendations are lacking.


We tried to develop an evidence-based structured approach in the management and follow-up of neonates with CCMV to help clinicians in their day to day care of these newborns.


MEDLINE/OVID database, Embase and Cochrane Collaboration Library were searched for original articles and reviews containing the following keywords: congenital cytomegalovirus, newborn, neonate, diagnosis, treatment, outcome, ganciclovir and valganciclovir. Articles were checked for references not retrieved by initial search and cross-referencing. Identified articles were analysed and graded for their level of evidence according to the Oxford Centre for Evidence-Based Medicine system (Tables 1 and 2) (8).

Table 1.   Levels of evidence (8)
Oxford Centre for Evidence-Based Medicine
Level 1Systematic review (SR) of randomized controlled trials (RCT), SR of prospective cohort studies, individual RCT, prospective cohort study with good follow-up
Level 2SR of cohort studies, SR of either retrospective cohort studies or untreated control groups in RCT, individual cohort study, retrospective cohort study or poor follow-up, retrospective cohort study or follow-up of untreated control patients in an RCT
Level 3SR of case-control studies, individual case–control study
Level 4Case series
Level 5Expert opinion without explicit critical appraisal
Table 2.   Grades of recommendation (8)
Oxford Centre for Evidence-Based Medicine
AConsistent level 1 studies
BConsistent level 2 or 3 studies or extrapolations from level 1 studies
CLevel 4 studies or extrapolations from level 2 or 3 studies
DLevel 5 evidence or troublingly inconsistent or inconclusive studies of any level


A total of 52 articles were included: 10 reviews, 26 retrospective and prospective case–control studies, four case reports, nine prospective comparative studies, two randomized control studies and one company product information for professionals were identified. Highest level of evidence was 1. Based on the analysis of these data, we designed an algorithm (Fig. 1), drug dosages (Table 3) and follow-up scheme (Table 4) to help with the management of CCMV infection in the newborn.

Figure 1.

 Algorithm for management of CCMV infection.

Table 3.   Recommendations for prescription of antiviral medication based on the following studies (30,31,35,38–48)
DrugGanciclovir – IntravenousValganciclovir – Oral
Drugs for treatment of CCMV infection
Dose6 mg/kg twice a day15 mg/kg twice a day
Time of levelsJust before and 1 h after second doseJust before and 2 h after third dose
Peak level6–8 mg/L6–8 mg/L
Trough level<0.5 mg/L<0.5 mg/L
Monitoring• Three times weekly full blood count, liver function tests, creatinine, urea and electrolytes• Three times weekly full blood count, liver function tests, creatinine, urea and electrolytes
• Once weekly drug levels• Once weekly drug levels
Indication for suspension of treatment• Absolute neutrophil count <500 cells/μL• Absolute neutrophil count <500 cells/μL
• Platelet count <25 000 cells/μL• Platelet count <25 000 cells/μL
Table 4.   Recommendations for follow-up based on these studies (4–6,13,15,51)
EvaluationAge recommendation
Follow-up for neonates with CCMV infection
AudiometryNewborn, 3, 6, 9, 12, 18, 24, 30 and 36 months and then annually to school age
Indirect ophthalmoscopy and visual functionNewborn, 12 months, 3 years and preschool age
Neurological examination and developmental assessmentAt each paediatric review


Identification of neonate with CCMV infection

Maternal CMV infection is mainly asymptomatic with only <10% of them developing non-specific symptoms like myalgia, asthenia with or without fever or flu-like symptoms. Primary CMV infection is defined as CMV IgG seroconversion during pregnancy or being positive for CMV-specific IgM with low anti-CMV IgG avidity (1,3). One level 2 study (9) has summarized various other studies and shown anti-CMV IgG avidity testing to have sensitivity between 92.8% and 100% to detect a recent primary CMV infection before 16–18th week of pregnancy. Three level 2 studies (1,2,9) have shown that 32% of infants born to women with primary infection had CCMV and 1.4% born to mothers with recurrent/past infection.

Recommendation: Grade B: A newborn should be evaluated for CCMV if the antenatal history or serological status of the mother is suggestive of primary or recent recurrent CMV infection.

Epidemiological studies have shown on clinical examination in newborns who are CMV positive, that 40–50% have intra-uterine growth restriction (IUGR), microcephaly, hepatosplenomegaly, petechiae, thrombocytopenia or conjugated hyperbilirubinaemia on initial blood tests or evidence of hearing loss on newborn hearing screen (10).

Recommendation: Grade B: A newborn should be evaluated for CCMV if any of these signs and symptoms is found.

In seven level 2 studies summarized in one level 2 study (1) urinary or salivary culture was used to isolate CMV. Until now this has been the gold standard test for diagnosing CCMV in newborns within the first 2–3 weeks of life. The disadvantage of this method is the time factor as it can take up to 3 weeks for the results to come back. Three level 2 studies (9,11,12) have used urinary PCR to isolate CMV with sensitivity and specificity reaching nearly 100% in diagnosing CMV. One level 1 (13) and level 2 studies (14,15) have used blood PCR either from dried blood spot or serum to isolate CMV and have achieved similar sensitivity and specificity. Saliva and umbilical cord blood have also been used in the detection of CMV (16–18). Other methods for diagnosing CMV have also been described: urine DEAFF test, CMV-specific IgM, CMV DNAemia in peripheral blood leucocytes, and CMV pp65-antigenaemia test (3,9,14,19).

Recommendation: Grade B: Urinary PCR for isolation of CMV in newborns within the first 2–3 weeks of life can be used as a first line diagnostic tool, as it is reliable, easy to collect, non-invasive and it can be stored for further evaluation for the detection of viral load.

Assessment of neonate with confirmed CCMV infection

Once CCMV has been diagnosed, the neonate should have a complete clinical, laboratory and imaging evaluation performed to determine the extent of the disease as stated in the algorithm (Fig. 1) (20–22). A complete neurological examination should be performed to evaluate for isolated microcephaly, symmetric IUGR, muscular hypotonia and early onset seizures.

A full blood count, liver function test, cranial imaging, ophthalmological screen and hearing assessment should be performed (see Table 5 for definitions of abnormal findings).

Table 5.   Level of evidence 2 for all definitions
  1. *Isolated single ventricular dilatation, subependymal pseudo cyst or lenticulostriatal vasculopathy are not considered pathognomic of CCMV infection.

Definitions of symptoms and signs
Microcephaly: head circumference <2 SD below the mean for age or <2nd centile (10,21,25,26)
Symmetric IUGR: birth weight and head circumference <2 SD below mean for age. (10,21,25,26)
Thrombocytopenia: <100 000/μL (10,21,26,27)
Conjugated hyperbilirubinaemia: >66 micromol/L (>3 mg/dL) (10,26,27)
Abnormal cranial US: moderate to severe ventriculomegaly and intracerebral calcifications* (10,21,25–28)
Abnormal cranial CT: cortical atrophy, cortical dysgenesis/dysplasia, moderate to severe ventriculomegaly, cerebellar hypoplasia/asymmetry, migration abnormalities and intracranial calcifications (10,21,25–27)
Abnormal ophthalmology screen: chorioretinitis, retinal detachment, optic atrophy, cataract, retinal scarring or sight-threatening infection (10,15,27)
Abnormal hearing assessment: unequivocally failed or >30 dB hearing loss on two or more age-appropriate audiologic tests (4,10,13,14,20–22,26,28,31,51)

A cranial ultrasound should be performed as soon as possible after birth. Cranial CT and/or MRI should also be considered in the neonatal period if there is a high index of suspicion of neurological involvement with CCMV infection despite normal cranial ultrasound scan (23,24).

Management of neonate with symptomatic CCMV infection

Symptomatic CCMV is defined as positive CMV in any secretions within the first three weeks of life with clinical manifestations of intra-uterine infection involving central nervous system (CNS) or lymphoreticular system (1).

Neonate with life-threatening infection

Mortality rate for CCMV ranges between 5% and 10%. CMV can cause pneumonitis, oesophagitis, colitis and severe thrombocytopenia requiring repeated platelet transfusions (10,15). Individual case reports have shown that infected neonates could benefit clinically and virologically from treatment with ganciclovir.

Recommendation: Grade D: Neonates with life-threatening infection should be considered for immediate treatment.

Neonate with central nervous system involvement

Infected neonates have CNS involvement if there is evidence of one of the following findings: isolated microcephaly, symmetric intrauterine growth restriction, abnormal CNS examination, abnormal cranial US/CT/MRI, abnormal ophthalmology or hearing screen. Four level 1 (25–28) and two level 2 (10,21) studies have shown that prospects for normal development are poor, as >90% of surviving infants develop significant CNS sequelae, perceptual defects or both within the first 2 years of life.

One level 1 (26) and two level 2 (10,21) studies have shown that microcephaly has 100% specificity and positive predictive value for prediction of mental retardation (IQ < 70) and/or major motor deficits. Absence of microcephaly is a sensitive marker for normal cognitive function.

Normal neuroimaging at birth in symptomatic CCMV predicts a good long-term neurologic outcome (10,21,25–28). Intracranial lesions on neuroimaging are associated with severe intellectual impairment in >80% of cases (27). The most common findings are intracranial calcifications (21). MRI provides important additional information, especially with regard to associated polymicrogyria, hippocampal dysplasia and cerebellar hypoplasia (24,29). Calcifications are better seen using US or CT. One level 1 study (21) has demonstrated a good correlation between cerebral US abnormalities and prediction of outcome in symptomatic CCMV infants. CT has been mainly recommended in the literature, but a combination of ultrasound and/or MRI might be a useful alternative.

Chorioretinitis can cause sight-threatening infection and hence warrants prompt treatment. Anecdotal reports during the CASG study suggest clinical improvement after treatment (20). In two level 1 studies (30,31) nearly 60% of the patients with retinitis had complete normal ophthalmological assessment after treatment.

Sensori-neural hearing loss is the most frequent neurological disability caused by CCMV. It can be both early-onset and late-onset. The deafness caused can be severe enough to require hearing aids and rehabilitation. Various studies have tried to identify the predictors for hearing loss, i.e. presence of petechiae, thrombocytopenia, intra-uterine growth restriction, microcephaly, abnormalities on cranial imaging. SNHL is defined as unequivocally failed or >30 dB hearing loss on two or more audiologic tests suitable for newborns and middle ear disease ruled out with normal bone conduction or use of hearing aids in one or both ears (10,32).

Two level 1 studies (30,33) have shown that treatment with ganciclovir improved neurodevelopmental outcome in 24%, improved or stabilized hearing impairment in 84% and improved visual acuity in 70% of the cases. A phase III non-blinded controlled trial of ganciclovir for CCMV reported that treated infants experienced a greater increase in median head circumference from baseline to 6 weeks evaluation although the long-term follow-up data were not available (33).

Recommendation: Grade A: Neonates with CNS involvement should be treated.

Neonate with disseminated CCMV infection and high viral load or petechiae or thrombocytopenia

Disseminated CCMV infection at birth is defined as presence of asymmetric IUGR, hepatosplenomegaly, hepatitis, anaemia and conjugated hyperbilirubinaemia (10,20,25,26); 40–50% of neonates will show evidence of disseminated infection at birth. Three level 1 (19,22,34) and two level 2 (12,13) studies have shown that the quantum of CMV copies in the blood correlates with neurological outcome irrespective of whether children are considered symptomatic or asymptomatic at birth. Four level 1 studies (13,22,26,27) have demonstrated that a high viral load in early infancy expressed by a high amount of virus in the urine is highly predictive of audiologic impairment. Greater than 70% of symptomatic (with or without CNS symptoms) infants with viruria of >5 × 104 pfu/mL will have poor neurodevelopmental outcome when compared with only 4% with viruria of <3.5 × 103 pfu/mL (27). In an animal model for human CMV infection, the number of detected DNA copies of CMV by PCR was compared with an established semi-quantitative plaque assay (35). It was concluded that 1 plaque-forming unit (pfu) is the equivalent of 1500 viral genome (35). In a study presented as abstract at the ESPID meeting 2009 in Belgium the average viral load in the urine in symptomatic infants was 1 × 107.5 copies/mL (17). In a group of four symptomatic patients from our unit the urinary viral load was 5.6 × 107 copies/mL (unpublished data). One level 2 study (36) has shown that congenitally infected newborns have higher urinary CMV viral load (1 × 106.3 copies/mL) when compared with post-natally acquired CMV infants.

Studies summarized in an article (37) have shown to decrease the viral load after commencing treatment which may lead to better hearing outcome and cognitive profile.

Two level 1 (22,26) and one level 2 (10) studies have demonstrated that presence of petechiae and thrombocytopenia with or without disseminated infection are associated with an increased likelihood of SNHL and adverse neurodevelopmental outcome. Moreover, petechiae or thrombocytopenia and IUGR alone correlate well with development of hearing loss.

Recommendation: Grade B: At this point in time, a treatment cut-off has not been established as a result of the differences in techniques used to determine viral loads. Neonates with disseminated CCMV infection and high viral load as described above or thrombocytopenia may be considered for treatment as it may affect their long-term neurodevelopmental outcome.

Management of neonate with asymptomatic CCMV infection

Asymptomatic CCMV infection is defined as positive CMV in any secretions within the first 3 weeks of life, but with normal clinical, laboratory and imaging evaluations (1). The majority of infected neonates with CMV will be asymptomatic at birth. Four level 2 studies (1,2,6,38) have demonstrated that 6–25% of these patients are likely to develop SNHL and 13.5% will develop permanent neurological sequelae. SNHL tends to be late-onset and progressive. Risk factors in asymptomatic neonates who will develop late-onset sequelae have not been determined. One level 1 study (22) has shown the presence of high viral burden in the first month of life to be associated with SNHL. There has been only one recent level 1 study (39) where 12 asymptomatic CCMV infants were treated with IV ganciclovir for a period of 21 days and a 4-10-year follow-up demonstrated no hearing loss in this group when compared with 11.1% hearing loss occurring in the asymptomatic untreated group. Although there is a risk of these asymptomatic neonates to develop long-term sequelae, the risks and side-effects related to treatment with ganciclovir are thought to be greater and its long-term benefit unknown.

Recommendation: Grade B: Currently treatment of asymptomatic neonates is not recommended.

Treatment of neonates with CCMV infection

Clinicians have been treating symptomatic CCMV in the past with various anti-viral agents. Most of the treatments involve potentially toxic drugs. Ganciclovir has been studied most and has the best treatment efficacy.

Two level 1 studies (30,31) have used intravenous (IV) ganciclovir 6 mg/kg twice a day for 6 weeks. There is no evidence to support start of treatment after the first month of life.

Recently valganciclovir, which is the orally bioavailable pro-drug of ganciclovir, has received much attention. Three level 2 (40–42) and two level 2 (43,44) studies have shown that valganciclovir has 10 times greater oral bioavailability than oral ganciclovir (53.6% vs. 4.8%). Once absorbed, it is rapidly converted by gastrointestinal and hepatic esterases into ganciclovir. Evidence in neonates with valganciclovir is growing and pharmacokinetic data recommend a dose of 15 mg/kg twice a day, which translates to an equivalent IV ganciclovir bioavailable dose (40–49).

In two level 2 studies (30,31) the predominant acute side-effect was neutropaenia (grade three or four) in up to 63% with ganciclovir and up to 38% with valganciclovir. Despite this, neutropaenia related sepsis was rarely a problem. Other rare side-effects are bone marrow suppression, raised liver enzymes, hypokalaemia and renal impairment. All these side-effects are reversible after stopping the drug for at least 3–7 days or decreasing the dose of the drug according to company product information (30,31,45). The US-FDA has listed ganciclovir as a pregnancy Category C drug. Animal experiments with high dose of ganciclovir showed that short-term exposure induces testicular damage, affects sperm viability and may have carcinogenic effects. Ganciclovir has been used in the treatment of CMV retinitis and prevention of CMV disease in paediatric transplant patients for many decades. Despite prolonged treatment in these patients reports of malignancy associated with its use have not been reported (50).

It is unclear whether prolonged or repeated treatment of CCMV more than 6 weeks is necessary as viral shedding returns at a lower viral load level after stopping treatment (34). There are very few small studies to date that have administered ganciclovir for prolonged periods without conclusive results (37).

Recommendation Grade B: All infants receiving ganciclovir or valganciclovir should initially be monitored very closely for development of side-effects for the first few weeks until a steady state of concentration is achieved (Table 3). Intravenous ganciclovir 6 mg/kg twice a day can be used in the initial management. Oral valganciclovir can be used once a steady state concentration is achieved. Infants should be closely monitored for evidence for toxicity as stated in Table 3. Treatment for more than 6 weeks cannot be recommended at this stage.


Long-term epidemiological studies have shown that 40–58% of infants with symptomatic and 13.5% of asymptomatic CCMV will have permanent neurological sequelae (1–6,38). CMV is known to cause progressive and late-onset hearing loss. The prevalence of SNHL caused by CCMV (symptomatic and asymptomatic) at birth is 5.2% and late-onset hearing loss at 6 years is 15.4% (4–6,13). The exact mechanism for the progressive hearing loss is not as yet ascertained. Universal hearing screen helps in identifying the infants with early-onset hearing loss but fails to diagnose children who develop the late-onset hearing loss that continues to occur during the first 6 years of life (51). CCMV can also lead to long-term visual impairment. Studies have estimated a prevalence ranging from 22% to 58% in symptomatic infants (5,31). One level 2 study (21) has proposed a follow-up of these infants.

Recommendation: Grade B: All infants irrespective of being treated or not should have regular auditory, ophthalmological and neurological evaluation performed for detection of cognitive and neurodevelopmental impairment as well as late-onset or progressive SNHL as stated in Table 4.


CCMV infection can cause significant impact and burden on the individual and the society. Effective prevention strategies in the high risk population, i.e. seronegative women in their child-bearing age are needed. Education about the mode of transmission and the risks associated with CCMV might be helpful in pregnant women. Women with known CMV infection during pregnancy should be counselled about the relevant risks to the foetus so that informed decisions considering risks and benefits can be made. Currently various studies are being carried out to develop CMV vaccine (52). The institute of medicine has ranked the development of CMV vaccine to prevent CCMV as a top priority (3). Evidence for neonates who would benefit from treatment is growing, and continuous advances are being made to understand the long-term role of anti-viral therapy in congenitally infected neonates. Studies looking at identification of risk factors in asymptomatic newborns who will develop long-term sequelae are required. Moreover, randomized controlled trials evaluating treatment and long-term follow-up of infants with both symptomatic and asymptomatic CCMV are necessary to help identify neonates who would benefit the most from treatment.