http://www.nejm.org/doi/full/10.

1056/NEJM199203053261003#t=article

The Outcome of Congenital Cytomegalovirus Infection in Relation to Maternal Antibody Status

Abstract Article References Citing Articles (194) Letters

CONGENITAL cytomegalovirus (CMV) infection is the most common intrauterine infection, affecting from 0.4 to 2.3 percent of live-born infants.1 , 2 In the United States it has been estimated that 1 percent of newborns are infected prenatally with CMV; this rate is equivalent to approximately 40,000 new cases each year. Although over 90 percent of infants infected with CMV are free of symptoms at birth, sensorineural hearing loss, chorioretinitis, mental retardation, and neurologic deficits subsequently develop in 5 to 17 percent of newborns with asymptomatic CMV infection.3 4 56 7 8 9 10 11 12 More frequent and more severe sequelae occur in infants with symptomatic infection; nearly 90 percent have one or more abnormalities caused by damage to the central nervous system or the organs of perception.3 4 5 6 7 8 9 10 11 12 The natural history of CMV in pregnancy is complex. Cytomegalovirus differs from other wellknown causes of fetal infection such as rubella and toxoplasmosis, which produce congenital infection only if the mother acquires the infection immediately before or during pregnancy.13 CMV, in contrast, can be transmitted from mother to fetus even when the mother is known to have been infected months or years before conception, as well as when primary infection occurs during pregnancy.13 Congenital CMV infections in the children of women with immunity acquired before pregnancy are less likely to be clinically apparent at birth than those resulting from a primary infection.14 , 15 However, two reports from Europe have described newborns with symptoms of congenital CMV infection who were born to mothers with immunity to CMV.16 , 17 The presence of maternal antibody before conception does not prevent the transmission of CMV to the fetus, but it helps prevent serious injury.18 The extent to which maternal antibody protects against symptoms in the newborn and against the late sequelae of congenital CMV infection is not known. Determining the effectiveness of maternal antibody in preventing the damaging consequences of congenital CMV infection is of major importance in assessing the potential benefit of a vaccine to prevent primary infection during pregnancy. We evaluated the effectiveness of antibody acquired naturally before conception in preventing injury to the fetus in utero and

sequelae in children with congenital CMV infection. We compared the outcomes of children born to immune mothers with those of children whose mothers had a primary infection during pregnancy.

METHODS
Study Population The study population consisted of 197 children with congenital CMV infection whose mothers had either primary or recurrent CMV infection during pregnancy. Maternal infections were classified on the basis of our analysis of serum samples collected and stored before pregnancy (usually cord serum from previous deliveries) and before delivery. Primary maternal CMV infection was defined by evidence of seroconversion (the first appearance of IgG antibodies) or by the detection of CMVspecific IgM antibodies during pregnancy.14 , 19 , 20 Seventy-four primary maternal infections were documented by evidence of seroconversion, and 58 were identified by the detection of IgM antibody to CMV in prenatal serum samples. Mothers who had antibody to CMV before conception, as indicated by the IgG antibody level in stored serum samples, were classified as having recurrent infection. In women for whom preconception serum samples were not available, recurrent infection was defined by the presence of CMV-specific IgG antibodies without CMV-specific IgM antibodies within the first 12 weeks of gestation.14 , 19 , 20 Thirty-nine of the mothers in the recurrent-infection group were known to have been immune at least nine months before conception, and 26 mothers were classified as having recurrent infection on the basis of the absence of IgM antibody to CMV early in gestation. The infants and children we studied were born between 1972 and July 1, 1990. The criteria for inclusion in this study were congenital CMV infection, demonstrated by the isolation of virus during the newborn period, and maternal infection that could be classified as primary or recurrent according to the methods described above. Of the children in the study, 172 were identified through the obstetrical services at hospitals where we screened newborns for viruria to detect congenital CMV infection. Twenty-five additional congenitally infected newborns were referred to us by other hospitals or physicians; these infants were identified as having CMV infection because the mother had evidence of infection on serologic screening or had illness during pregnancy or because elevated cord-blood levels of IgM or rheumatoid factor or symptoms of congenital infection in the newborn led to virologic testing. Congenital infection in the referred patients was confirmed by the isolation of virus in our laboratory, and maternal infection was classified in the same way as for the babies born in the study hospitals. Serologic and Virologic Testing CMV-specific IgG antibodies were measured by anticomplement immunofluorescence assay, enzyme-linked immunosorbent assay (ELISA), or both; CMV-specific IgM antibodies were measured by ELISA or radioimmunoassay, as previously described.19 , 20 All congenital infections were confirmed by the isolation of CMV from urine collected within the first or second week of life. Urine specimens were processed and inoculated onto monolayers of human foreskin fibroblasts, as previously described.21 Clinical Assessment and Follow-up All the children's medical records for the first month of life were reviewed at the time of enrollment. An infant was classified as having symptomatic infection if any of the following were observed:

petechiae, hepatosplenomegaly, microcephaly, thrombocytopenia, or jaundice with conjugated hyperbilirubinemia. All the children were followed in a special clinic; serial assessments of their general health and neurologic, developmental, audiologic, and visual status were performed according to methods that have been described previously.11 , 12 , 22 The children were routinely evaluated three or four times during the first year of life, twice during the second year, and annually thereafter. Developmental and intellectual assessments were performed with use of standard psychometric tests suitable for each child's age, degree of perceptual function, and psychomotor ability. 12 The tests we used were the Bayley Scales of Infant Development, StanfordBinet Intelligence Scale, AlpernBoll Developmental Profile, Leiter International Performance Scale, Wechsler Preschool and Primary Scale of Intelligence, and Wechsler Intelligence Scale for Children Revised. Psychometric assessments were routinely performed when the child was four years old and again when he or she reached six years of age. Statistical Analysis Chi-square and Fisher's exact tests were used to assess statistical significance. The association between sequelae in children with congenital CMV infection and the type of maternal infection was assessed with KaplanMeier survival curves, compared by means of a log-rank test.23 All analyses were performed on a PC computer with SAS software and KMSURV, a microcomputer program for the statistical analysis of survival data.24 , 25

RESULTS

Characteristics of the study population are shown in Table 1

T ABLE 1

Characteristics of

the Study Population, According to Type of Maternal Infection. . The children of mothers with primary

infections during pregnancy (primary-infection group) and the children of mothers with immunity to CMV (recurrent-infection group) differed with respect to race and the proportion of infants referred. In the primary-infection group, 57 percent of the children were white, as compared with 22 percent of the children in the recurrent-infection group. Referrals accounted for 17 percent of the primaryinfection group and only 6 percent of the recurrent-infection group. The two groups did not differ with respect to sex or enrollment in our follow-up program. Only children of mothers with primary infections during pregnancy (primary-infection group) had symptomatic congenital CMV infections (18 percent). The signs of infection observed during the

newborn period are shown in Table 2

T ABL E 2

Clinical Findings in the First Month of Life

in 24 Newborns with Symptomatic CMV Infection after a Primary Maternal Infection. . The most common

clinical findings in newborns with symptomatic CMV disease were jaundice, petechiae, and hepatosplenomegaly; 83 percent of the newborns had more than one of these signs.

The mean age at the most recent clinic visit for the 189 children enrolled in the follow-up study did

not differ significantly between the two groups (Table 3

T ABLE 3

Evaluation of Children

Enrolled in the Follow-up Study, According to Type of Maternal Infection.). The proportions of children

in the two groups who had tests of hearing and vision and psychometric evaluations were also similar. The children in both groups had multiple examinations of hearing and vision and psychometric assessments. The sequelae of congenital CMV infection in the study children are shown inTable 4
T ABLE 4

Sequelae in Children with Congenital CMV Infection, According to Type of Maternal Infection.. Sensorineural hearing loss was observed in 15 percent of the primary-infection group, as

compared with 5 percent of the recurrent-infection group. Bilateral hearing loss occurred in 8 percent of the primary-infection group; 8 percent had moderate-to-severe hearing loss (speech threshold, 60 dB). No bilateral hearing loss was observed in the recurrent-infection group. Children in the primary-infection group were also significantly more likely to have an IQ 70 than those in the recurrent-infection group (13 percent in the primary-infection group and none in the recurrent-infection group). Chorioretinitis was observed in 6 percent of the children in the primaryinfection group, whereas only one child in the recurrent-infection group (2 percent) had detectable chorioretinitis. Other neurologic sequelae were observed in 6 percent of the children in the primaryinfection group and in only 2 percent of those in the recurrent-infection group. The only neurologic sequela observed in the recurrent-infection group was microcephaly. The primary-infection group had more neurologic sequelae than the recurrent-infection group, but when each abnormality was examined individually, no significant difference was observed between the two groups. Overall, 25 percent of the children in the primary-infection group and only 8 percent of those in the recurrent-infection group had sequelae (P = 0.003). In the primary-infection group, one sequela was observed in 18 percent of the children, and an additional 6 percent had more than one problem. No child in the recurrent-infection group had multiple sequelae. The proportion of children in both groups who remained free of sequelae over time is shown

in Figure 1

F IGUR E 1

Percentage of Children with Congenital CMV Infection Who Remained

Free of Sequelae, According to the Type of Maternal Infection. . In the primary-infection group, 67

percent of the children were free of sequelae at 72 months of age, as compared with 86 percent of the children in the recurrent-infection group. Sequelae in the primary-infection group were most frequently observed during the first 12 months of life, although sequelae were identified up to 72 months of age. In the recurrent-infection group all the sequelae we observed were evident before

48 months of age. The proportion of children who were free of sequelae was significantly lower in the primary-infection group than in the recurrent-infection group (P = 0.02).

DISCUSSION
Our findings indicate that preexisting maternal antibody to CMV protects the fetus and lessens the severity of the sequelae of congenital CMV infection. Children in the primary-infection group were more likely to have symptoms at birth. They were also more likely to have functionally important bilateral sensorineural hearing loss and to be mentally retarded than the children in the recurrentinfection group. Not only were the children in the primary-infection group more likely than those in the recurrent-infection group to have at least one sequela (18 percent vs. 8 percent), but those in the primary-infection group were also more severely damaged, as indicated by the presence of multiple sequelae in 6 percent (as compared with none in the recurrent-infection group). Although the primary-infection group and the recurrent-infection group differed in the proportions that were black or were identified by referral rather than screening, it is unlikely that these differences accounted for the greater frequency of sequelae observed in the primary-infection group. Large-scale surveys indicate that the expected rate of mental retardation (IQ 70) in the general population is around 3 percent.28 A higher proportion of the children in the recurrentinfection group were black and came from low-income, urban families; more of the children in the primary-infection group were from white families with middle-to-upper-level incomes. These demographic differences would have been expected to bias us against finding a higher rate of mental retardation in the primary-infection group. However, no child in the recurrent-infection group had an IQ70, whereas 13 percent of the children in the primary-infection group were mentally retarded. The greater frequency and severity of sequelae in the primary-infection group could not be attributed to differences in follow-up. Both groups had similar proportions of children who had assessments of hearing and vision and psychometric tests. Comparisons of the numbers of followup evaluations per subject revealed no differences between the two groups, implying equivalent ascertainment of sequelae. We believe our results have probably underestimated the greater severity of sequelae in the primary-infection group through the misclassification of some maternal infections as recurrent when they were in fact primary infections. This could have occurred since we presumptively classified maternal infections as recurrent if no IgM antibody to CMV was detected. We used this criterion only for seropositive women for whom no preconception serum samples were available and whose first prenatal visit took place before 12 weeks of gestation. Studies have shown that in women in whom IgM antibody develops with seroconversion, IgM antibody persists for more than 12 weeks in 24 percent.20 The persistence of IgM antibody varies, however, and it is likely that some maternal infections classified as recurrent by this approach may have been recent primary infections. As an example, one of the mothers in our study, a day-care worker, was tested as part of our earlier study of day-care workers29 and found to be seronegative. She happened to receive care from a private obstetrical group involved in this study. At her initial prenatal visit at 11 weeks' gestation, 7 months after her last negative antibody test, she was found to be positive for IgG antibody against CMV but negative for IgM antibody. This woman clearly fit our criteria for recurrent infection, and her infant

was classified in that group. It is very likely, however, that she acquired CMV around the time of conception. Her infant was asymptomatic at birth, but sensorineural hearing loss later developed. In addition to the variable persistence of IgM antibody after primary infection, identifying the time of CMV infection is complicated by the fact that not all primary infections are accompanied by the development of IgM antibody to CMV that is detectable by the assays we used. The ELISA rand radioimmunoassay methods have sensitivities of 73 percent and 78 percent, respectively, in detecting primary CMV infections in pregnant women who have recently seroconverted.19 , 20 Thus, 22 to 27 percent of recent primary infections could be misclassified as recurrent because of the absence of IgM antibody. When we reexamined our data, removing the 26 children whose mothers' infections were classified as recurrent on the basis of the absence of IgM antibody, the proportion of children with sequelae from confirmed recurrent infection dropped from 8 percent to 5 percent. Our findings indicate that the type of maternal CMV infection is a major determinant of outcome for infants with congenital CMV infection. Since our results provide evidence that preexisting maternal antibody protects the fetus and the newborn from serious damage, it is interesting to estimate the public health impact of a vaccine that could mimic natural immunity. Each year, about 40,000 infants are born with congenital CMV infection, and about 8000 of them are injured as a result

(Table 5

T AB LE 5

Public Health Consequences of Congenital CMV Infection.*). If we

assume that primary maternal infection accounts for all symptomatic congenital CMV infection and for approximately 65 percent of the asymptomatic infections, a vaccine that could be provided to all seronegative women of childbearing age could prevent about 6000 of the damaging congenital CMV infections that occur each year. Much must be learned before we can intervene to prevent maternal CMV infections, but the potential benefit appears to justify increased efforts in this direction. Supported in part by grants from the National Institute of Child Health and Human Development (HD10699), the General Clinical Research Center (RR0032), the Deafness Research Foundation, and the March of Dimes Birth Defects Foundation (6490). We are indebted to Dr. Ronald E. Henderson and Dr. Philip D. Walton of Henderson and Walton Women's Center and Dr. Robert L. Goldenberg of the Department of Obstetrics and Gynecology at the University of Alabama at Birmingham for enrolling their patients in the study, and to Arthur J. Dahle, Faye P. McCollister, and Catherine S. Amos for their technical assistance in the evaluations of hearing and vision.

SOURCE INFORMATION
From the Departments of Pediatrics (K.B.F., S.S., R.F.P., W.J.B., C.A.A.) and Psychology (T.J.B.), University of Alabama at Birmingham. Address reprint requests to Dr. Fowler at the Department of Pediatrics, University of Alabama at Birmingham, 1600 Seventh Ave. S., Suite 752, Birmingham, AL 352940011.