Objective. This study was undertaken to ascertain whether children with congenital cytomegalovirus (CMV) infection at birth, but without neurologic symptoms at 1 year of age, differed in somatic, neurologic, developmental, or intellectual status from matched control children at long-term follow-up.
Material and Methods. Congenital CMV infection, as demonstrated by isolation of the virus within the first week of life, was diagnosed in 44 (0.5%) of 9806 infants. From this basic CMV-infected population, children who developed neurologic disturbances including sensorineural deafness before 1 year of age were excluded (n = 7), as were those unwilling to participate (n = 2). Congenitally infected infants (n = 35) and matched control infants (n = 53) were followed up clinically and neurologically. At 21 months, development was assessed with the Griffiths' Developmental Scale and at 7 years of age neurologic status was assessed with the Stott test and intellectual development with the Wechsler Intelligence Scale for Children.
Results. As reported previously, 18% (8/44) of the CMV-infected infants manifested symptoms at birth, as compared to 8% (4/53) in the control group. In the congenital CMV group 7% (2/30) children tested had abnormal Stott test results, as compared to 2% (1/43) in the control group. Thirty-two CMV-infected and 51 control group children were assessed with the Griffiths' scale at 21 months of age. The two groups did not differ significantly, either in mean scores (6.3 ± 2.3 vs 6.1 ± 1.9) or in the proportion of children with scores below normal (19% [6/32] vs 16% [8/51]). Twenty-five CMV-infected and 41 control group children were assessed with the Wechsler Intelligence Scale for Children at 7 years of age (median 86 months; range, 82 to 90 and 82 to 91, respectively). The two groups did not differ significantly, either in mean scores (5.8 ± 2.0 vs 6.4 ± 1.6) or in the proportion of children with scores below normal (12% [3/25] vs 5% [2/41]).
Conclusion. Children with congenital CMV infection are unlikely to be at an increased risk of subsequent neurodevelopmental or intellectual impairment if they show normal development at 12 months of age.
Congenital cytomegalovirus (CMV) infection is present in some 0.4% to 2.3% of all liveborn infants and is the most common viral infection known to be transmitted in utero.1 In up to 10% of congenitally infected infants, the clinically symptomatic form, cytomegalic inclusion disease, is manifest in the neonatal period.1 Both neonatally overt and asymptomatic infections can result in permanent neurologic sequelae. Of infants with silent infection at birth, 5% to 17% suffer neurologic sequelae.1 The figure is greater among those manifesting neurologic symptoms already at birth.2-5 Some investigators have also suggested CMV to be a major cause of developmental and intellectual impairment as well as of mental retardation.6,7
In our prospective study, no child had a symptomatic form of CMV (ie, intrauterine growth retardation, jaundice, hepatosplenomegaly, thrombocytopenia, petechiae, pneumonia or central nervous system disease (CNS) including microcephaly, intracranial calcifications, chorioretinitis, or sensorineural deafness), and all children with neurologic sequelae including sensorineural deafness were diagnosed before 1 year of age.8 The evidence of mental retardation in the congenital symptomatic CMV population suggests that the intellectual development of the much larger asymptomatic neonatal population might be affected in a more subtle manner, a suggestion deriving further support from the finding that asymptomatic infants are at risk of hearing impairment.8 The latter finding, together with the affinity of CMV for a wide range of cells in the human CNS and the prolonged period of viral excretion in congenitally infected children, indicates the possible involvement of CMV as a cause of intellectual impairment later in life. This part of our longitudinal study was undertaken to ascertain whether children with congenital CMV infection at birth, but without neurologic symptoms at 1 year of age, differed in somatic, neurologic, developmental, or intellectual status from matched control children at long-term follow- up.
PATIENTS AND METHODS
Infants With Congenital CMV Infection
In a prospective study, reported in detail elsewhere,8 newborns were screened for the presence of CMV in urine. Congenital CMV infection, as demonstrated by isolation of the virus within the first week of life, was diagnosed in 44 (0.5%) of 9806 infants (with a successful virus isolation test), representing about 85% of all infants born in the city of Malmö, Sweden, during the 5-year period from October 1977 through September 1982.
From this basic CMV-infected population identified in the above-mentioned study, children with no symptoms at birth but who developed neurologic disturbances (ie, manifesting clinical evidence of hypotonia, spasticity, hemiparesis, delayed motor development, seizures, or sensorineural deafness during the first years of life) were excluded (n = 7), as were those unwilling to participate (n = 2). Of the remaining 35 CMV-infected infants, 30 were followed up clinically at least for 7 years, and 32 and 25 children were followed up with regard to specific developmental and intellectual tests at 21 months and 7 years of age, respectively.
Of 70 infants (ie, 2 per index case) born during the 4-year period, 1977 to 1981, with a negative CMV isolation test at birth, selected on the basis of birth order (consecutive number 1 or 25, before or after an index case), 53 infants were included in the study as controls with the consent of the parents. In this way we eliminated selection bias, and there were no differences between the congenital CMV group and the control group in terms of status at birth, socioeconomic factors, parental level of education or ethnic background. For instance, the respective CMV and control group data were 26.5 vs 27.8 years maternal age at parturition, 40.2 vs 40.2 weeks for gestational age at birth, 3440 vs 3542 g for birth weight, 20/15 (57%/43%) vs 31/22 (58%/42%) for the female/male ratio among children, 80% vs 83% for the proportion with both parents from Scandinavia, 9% vs 8% with one parent from outside Scandinavia, and 11% vs 9% with both parents from outside Scandinavia. After the end of 1981 no new controls were enrolled. Forty-three children in the control group were followed up clinically for 7 years, and 51 and 41 children underwent developmental and intellectual assessment at the same ages as the CMV-infected children (ie, at 21 months and 7 years of age, respectively).
Investigation of General Health, Somatic, and Neurologic Status
In Malmö virtually all infants are born at the hospital. The newborns are examined by pediatricians. The infants were followed up clinically and neurologically at the ages of 3, 6, and 9 months and at 1, 1½, 2½, 4, and 7 years, and the congenitally infected children also at 10 years.9,10
At 7 years of age the children underwent the Stott test (ie, examination of gross and fine motor performance, possible scores 0 to 10; 0 to 2, normal; 3 to 10, increasingly pathological),11 together with the Circle and Maze tests.
In the Circle test the subject is required to cut out a circle (Ø 10 cm) drawn on paper. The result is considered satisfactory if the cut out circle is less than 20% bigger or smaller than the drawn circle and the child uses less than 2 minutes to complete the task.
The Maze Test
In this extra task included in the Wechsler Intelligence Scale for Children (WISC),12 the child, using a pencil, is required to indicate the path out of a maze printed on paper (scored 0 to 10; 0 to 8, below normal; 8 to 10, normal). There is a time limit and an injunction against crossing lines.
Both the Circle and the Maze tests are part of a short neurodevelopmental screening assessment for deficits in attention, motor function, and perception, which contains 6 items (diadochokinesia, hopping on one leg, standing on one leg, cutting out a paper circle, associated movements when walking on the lateral sides of feet and the maze test from WISC).13 The four remaining items are included in the Stott test.
Assessment of Developmental and Intellectual Status
Developmental and intellectual status were assessed at 21 months and 7 years of age with the Griffiths' Developmental Scale14 and the WISC,12 respectively. Griffiths' Developmental Scale are the scales most widely used to assess development in Swedish preschool children. There are two scales: Scale I for children 0 to 2 years of age, and Scale II for those 2 to 8 years of age. The scales were originally in English,14 but have been translated and standardized for use in Swedish children.15 Scale I, which was used in this study, consists of 5 subscales (locomotor scale, personal social scale, language scale, hand and eye coordination scale, and performance scale). The respective subscale scores as well as the aggregate score for the developmental scale as a whole are expressed as stanine scores (ie, 7 to 9 points is above normal, 4 to 6 normal, and 1 to 3 below normal development).
The WISC has also been standardized for use in Swedish children12 and is the scale most widely used to assess overall intellectual development of school age children in Sweden. WISC results are also expressed in stanine points both for each subscale and for the scale as a whole.
Urine samples were obtained within the first week of life. If the cell cultures (human embryonic lung fibroblasts) could be maintained in good condition for at least 4 weeks, the test was considered successful. CMV isolates were verified by their characteristic cytopathic effect.
Classification of Maternal CMV Infection
The mothers of the CMV infected infants were investigated in sera collected preconceptionally (occasionally), during pregnancy (repeatedly), and at delivery (regularly). Classification of maternal CMV infections was based on principles previously described by us and others8,16,17 as follows:
Group 1 (n = 14): primary infection, (ie, seroconversion), de novo appearance of CMV-immunoglobulin G (IgG) in the second of two serum samples, both collected during pregnancy.
Group 2 (n = 3): presumed primary infection, (ie, CMV-immunoglobulin M (IgM) present in the first postconceptional serum sample).
Group 3 (n = 8): recurrent infection, (ie, CMV-IgG present before conception).
Group 4 (n = 2): presumed recurrent infection, (ie, CMV-IgG but no CMV-IgM <12 weeks after conception).
Group 5 (n = 8): type of infection cannot be determined, (ie, all sera collected after the 12th gestational week; no CMV-IgM present and no increase in the CMV-IgG titer).
Determination of the Onset of Maternal CMV Infection
In group 1, the onset of infection was indicated by seroconversion. In group 2, the occurrence of specific IgM indicated onset to have occurred between conception and the date of serum sampling (possibly just before conception). In recurrent infections, onset was sometimes indicated by an increase in CMV-IgG. None of the mothers manifested any clinical symptoms compatible with CMV infection.
CMV Antibodies in Maternal Sera
Specific IgM was studied by means of indirect immunofluorescence, and in presumed recurrent and indeterminable infections also by enzyme labelled antigen test.18 In the IgM indirect immunofluorescence test, false-positive reactions due to rheumatoid factors were precluded by the absorption of sera with aggregated human IgG. CMV-IgG was studied with an in-house enzyme-linked immunosorbent assay.19
Fisher's exact test (two-tailed) and Student's ttest were used for statistical analysis.
As reported previously, 18% (8/44) of the CMV- infected infants manifested symptoms from the reticuloendothelial system (RES) at birth, (ie, hepatosplenomegaly and/or petechiae (Table).8 However only 14% (5/35) belonged to the subgroup without neurologic disturbances at 1 year of age. At 10 years of age, none of the children had any subsequent major illnesses. Five were allergic, one had primary nocturnal enuresis, two had had repeated urinary tract infections, and one had type 1 (insulin-dependent) diabetes mellitus.20
Neurologic Findings in the Congenital CMV Group
Neurologic examinations were performed at 3, 6, 9, and 12 months and at 1½ (n = 35), 2½ (n = 34), and 4 (n = 33) years of age. The infants were examined with the Stott test and the Maze and the Circle tests at 7 years of age. Two (7%, none with symptoms at birth) of the 30 children tested had abnormal Stott test results. Of the children who underwent the Maze and Circle tests (n = 24 and n = 25, respectively), all had normal scores (Table 1).
In groups 1 and 2, all five (29%) of the 17 infants with symptoms at birth were the offspring of mothers with primary infection, as were the 2 (13%) of 16 children with abnormal Stott test results at 7 years of age. All 12 children who underwent both the Maze and Circle tests had normal scores (Table 1).
Of infants born to mothers with recurrent infection (groups 3 and 4, n = 10), none manifested symptoms at birth. Normal test results were obtained by all eight children undergoing the Stott test, by all seven undergoing the Maze test, and by all six undergoing the Circle test (Table 1).
Of infants born to mothers with undefined infection (group 5, n = 8), none manifested symptoms at birth. All six children undergoing the Stott, Maze, and Circle tests had normal results (Table 1).
Four (8%) of the 53 infants in the control group manifested symptoms [ie, petechiae (n = 3) or icterus (n = 1) at birth (Table 1)]. At 7 years of age two children had been operated on due to strabismus, two were allergic, one had febrile convulsions, one had hydronephrosis, and one had rheumatic fever and suspected myocarditis. One girl had facial paralysis due to surgery for a fistula in her ear, one was operated on for a granuloma telangiectaticum, one due to a fibroma in front of her ear, and one was operated on twice due to a dermoid cyst.
Neurologic Findings in the Control Group
Neurologic examinations were performed at 3, 6, 9, and 12 months and at 1½ (n = 53), 2½ (n = 52), and 4 (n = 49) years of age. The children underwent the Stott test and the Maze and Circle tests at 7 years of age. One (2%) of 43 children had abnormal Stott test results. In the Maze and Circle tests, undergone by 41 of the children, one child (2%) had an abnormal Maze test result and two children (4%) failed the Circle test (Table 1).
Developmental Assessment of Congenital CMV and Control Groups
Thirty-two CMV-infected and 51 control group children were assessed with the Griffiths' Developmental Scale at 21 months of age (median 89 weeks; range, 81 to 107 and 82 to 121 weeks, respectively). The two groups did not differ significantly, either in mean scores (6.3 ± 2.3 vs 6.1 ± 1.9) or in the proportion of children with scores below normal (19% [6/32] vs 16% [8/51]).
Scores below normal were obtained by 13% (2/16) of the offspring of mothers with primary infections, 22% (2/9) of the offspring of mothers with recurrent infections, and 29% (2/7) of the offspring of mothers with undefinable infections.
Intellectual Assessment of Congenital CMV and Control Groups
Twenty-five CMV-infected and 41 control group children were assessed with the WISC at 7 years of age (median, 86 months; range, 82 to 90 and 82 to 91, respectively). The two groups did not differ significantly, either in mean scores (5.8 ± 2.0 vs 6.4 ± 1.6) or in the proportion of children with scores below normal (12% [3/25] vs 5% [2/41]).
Scores below normal were obtained by 17% (2/12) of the offspring of mothers with primary infections and by 14% (1/7) of the offspring of mothers with recurrent infections. All six offspring of mothers with undefinable infections had normal test results.
The present prospective study of congenital CMV infection was based on screening of infants with a virus isolation test in urine within the first week of life. It is unique in the sense that none of the 44 congenitally infected children had any neurologic symptoms at birth. Eight had transient, usually mild neonatal symptoms from the RES. There was no significant impact of neonatal RES symptoms on the development of neurologic sequelae. The results of the present study (where children with neurologic sequelae diagnosed within the first year of life were excluded) suggest the presence of congenital CMV infection to be unaccompanied by any increase in the risk of developmental deficits at 21 months of age. This is in accord with previously reported findings.2
Although more subtle problems may emerge at school age, we found no evidence of this in our series. Unlike Kumar and coworkers,21 who retested their series at ages from 4½ to 10 years, and Conboy and colleagues,22 who tested children at ages ranging from 6½ to 12½ years, all subjects in our congenital CMV series were tested at the same age (7 years). Nonetheless, our findings were consistent with theirs in yielding no evidence of intellectual impairment.
It is also salutary to compare the present results with inconsistent results obtained by others.6,7 There are several possible reasons for the discrepancies in results. For instance, in the above-mentioned studies where the developmental and intellectual status of children with congenital CMV infection has been assessed, the series have comprised children with normal hearing and varying proportions of hearing impaired children. As hearing impairment may be indicative of viral penetration and lesions to the CNS at some level, it is reasonable to suppose children with hearing loss to be at greater risk of intellectual impairment. Accordingly, tests appropriate for use in deaf children should be used to investigate the possibility of intellectual impairment in children with congenital CMV infection and hearing loss. Another possible source of discrepancy is inadequate matching of controls. In both of the foregoing studies,6,7for example, the congenital CMV group was of a lower socioeconomic status than the control group, besides including children with hearing loss, as mentioned above.
Perhaps the most obvious source of discrepancy in results is the difference in the criteria used to identify children with CMV infection. In both of the foregoing studies, the criterion was an increased IgM or CMV-IgM concentration in cord blood, though the CMV infection so identified is more severe than that identified by virus isolation.
Unlike rubella and toxoplasmosis, CMV can be transmitted to the fetus even if there is preconceptional maternal immunity. The type of maternal infection might be an important determinant of developmental and intellectual impairment in the offspring. In contrast to other follow-up studies where the developmental and intellectual status of children with congenital CMV infection has been investigated, we had access to maternal preconceptional sera in most cases. Comparison of maternal infection subgroups with regard to outcome in the offspring yielded no evidence that maternal primary infection was associated with an increased risk of developmental or intellectual deficit in the offspring. However, owing to the small size of the present series (16 primary and 8 recurrent maternal infectious), it cannot be excluded that the existence of such a relationship might have been elicited in a larger series.
To sum up, in the present series of children with congenital CMV infection but no neurologic problems, there were no difference in the frequency of developmental impairment or intellectual deficit.
The findings constitute a source of reassurance for the parents of newborns with CMV infection inasmuch as they suggest that, if no neurologic sequelae emerge within the first year of life, the risk of subsequent intellectual or developmental sequelae is negligible.
This study was supported by grants from the Faculty of Medicine, University of Lund; and the Health Services Administration, MalmöUniversity Hospital.
- Received July 31, 1996.
- Accepted November 19, 1996.
Reprint requests to (S-A.I.). Department of Pediatrics, University Hospital, Malmö, 205 02 Malmö, Sweden.
- CMV =
- congenital cytomegalovirus •
- CNS =
- central nervous system •
- WISC =
- Wechsler Intelligence Scale for Children •
- IgG =
- immunoglobulin G •
- IgM =
- immunoglobulin M •
- RES =
- reticuloendothelial system
- Pearl KN,
- Preece PM,
- Ades A,
- Peckham CS
- Ramsey MEB,
- Miller E,
- Peckham CS
- Pass RF,
- Stagno S,
- Myers G,
- Alford CA
- ↵Illingworth RS. The Development of the Infant and Young Child. London, England: Churchill Livingstone; 1980
- ↵Touwen B. Neurological development in infancy. Clinics in Developmental Medicine. No 58. Spastics International Medical Publications. London, England: W. Heineman Med Books Ltd; Philadelphia, PA: JB Lippincott Co; 1976
- ↵Stott DH, Moyes FA, Hendersen SE. Test of Motor Impairment. Hendersen Revision. Guelph, Ontario: Brook Educational Publishing Limited; 1984
- ↵Manual till WISC (Wechsler Intelligence Scale for Children). Stockholm, Sweden: Psykologiförlaget; 1977
- Gillberg C,
- Carlström G,
- Rasmussen P,
- Waldenström E
- ↵Griffiths R. The abilities of babies, a study in mental measurement. London, England: University of London; 1954
- ↵Alin-Åkerman B, Norberg L. Griffiths' utvecklingsskalor I och II. Stockholm, Sweden: Psykologiförlaget; 1991
- Conboy TJ,
- Pass RF,
- Stagno S,
- et al.
- Copyright © 1997 American Academy of Pediatrics