Fowler KB, Stagno S, Pass RF. Maternal Immunity and Prevention of Congenital Cytomegalovirus Infection. JAMA. 2003;289(8):1008-1011. doi:10.1001/jama.289.8.1008
Author Affiliations: Departments of Pediatrics (Drs Fowler, Stagno, and Pass), Epidemiology (Dr Fowler), Maternal and Child Health (Dr Fowler), and Microbiology (Drs Stagno and Pass), University of Alabama, Birmingham.
Context Vaccine development to prevent congenital cytomegalovirus (CMV) infection
has been impeded by the uncertainty over whether maternal immunity protects
Objective To determine whether the presence of maternal antibodies to CMV significantly
reduces the risk of congenital CMV infection in future pregnancies.
Design, Setting, and Participants Cohort study of 3461 multiparous women from a population with a high
rate of congenital CMV infection who delivered newborns screened for congenital
CMV infection between 1993 and 1998, and whose cord serum specimen from a
previous delivery could be retrieved and tested for antibody to CMV.
Main Outcome Measure Congenital CMV infection according to maternal immune status, age, race,
parity, and socioeconomic status.
Results Of 604 newborns born to initially seronegative mothers, congenital CMV
infection occurred in 18 (3.0%). In contrast, of 2857 newborns born to immune
mothers, congenital CMV infection occurred in 29 (1.0%) Two factors, preconception
maternal immunity (adjusted risk ratio, 0.31; 95% confidence interval, 0.17-0.58)
and maternal age of 25 years or older (adjusted risk ratio, 0.19; 95% confidence
interval, 0.07-0.49), were highly protective against congenital CMV infection.
No other factors were associated with a reduction in the risk of congenital
Conclusion Naturally acquired immunity results in a 69% reduction in the risk of
congenital CMV infection in future pregnancies.
Congenital cytomegalovirus (CMV) infection remains a major public health
problem in the United States because of its frequency and its role as a cause
of sensorineural hearing loss, cognitive impairment, cerebral palsy, and visual
impairment.1,2 Congenital CMV
infection is a leading cause of sensorineural hearing loss in children and
the leading infectious cause of central nervous system damage in children
in the United States.2- 4 Development
of a vaccine for prevention of congenital CMV infection was listed as a top
priority by the Institute of Medicine of the National Academy of Sciences
in an analysis of cost of disease and impact on quality-adjusted life-years.5
Virus can be transmitted to the fetus during primary maternal infection
during pregnancy, but it also can be transmitted even when maternal infection
occurred years prior to conception.6 Whether
maternal immunity decreases the frequency of transplacental transmission of
CMV has not been previously studied. Uncertainty over the ability of naturally
acquired immunity to prevent future infection is a major impediment to development
of vaccines for prevention of congenital CMV infection. To test the hypothesis
that preconception maternal immunity reduces the risk of congenital CMV infection,
we compared rates of congenital infection according to maternal CMV serological
status determined years prior to delivery.
Women who delivered a newborn at the university hospital in the southern
part of the United States between January 1993 and December 1998 were included
in the study population if the following criteria were met: (1) the newborn
was screened for congenital CMV infection; (2) the mother had 1 or more previous
live births at the university hospital; (3) a cord serum specimen from a previous
delivery could be retrieved and tested for IgG antibody to CMV. Cord serum
specimens were stored at −20°C in a repository. The study was approved
by the institutional review board at the University of Alabama, Birmingham.
Written informed consent was obtained from the mothers of infants with congenital
CMV infection who attended follow-up clinics. The institutional review board
waived consent for those without CMV infection.
Between 1993 and 1998, 7558 multiparous women delivered at the university
hospital (Figure 1). Of these women,
4132 had a prior delivery at the university hospital, and 3461 had available
cord serum specimens. The latter group comprises the study population. The
671 women who delivered between January 1993 and December 1998 and had a previous
delivery at the university hospital but did not meet study criteria because
a cord serum specimen was not available did not differ from the study population
in either race or insurance status. However, these excluded women were slightly
The CMV-specific IgG antibodies were measured using a commercially available
enzyme immunoassay (ImX, Abbott Laboratories, North Chicago, Ill). Seroconversion
was determined by simultaneously measuring the presence of CMV IgG antibodies
in the cord serum specimen of the current delivery and the absence of CMV
IgG antibodies in the cord serum specimen from the previous delivery in women
who were initially CMV seronegative.7 Newborns
were screened for congenital CMV infection by detection of virus in saliva
using a rapid culture method.8 All CMV-positive
results on a newborn screening test were confirmed by the isolation of the
virus from urine and/or saliva at a 3- or 4-week postnatal follow-up clinic
visit. Between 1993 and 1998, newborns delivered at the university hospital
were routinely screened for congenital CMV infection with 88% screened during
Characteristics of the study population by maternal antibody status
were compared using χ2 tests or t tests
when appropriate. Rates were measured for independent and dependent variables.
The 95% confidence intervals (CIs) were calculated by exact binomial methods.
Risk ratios (RRs) were calculated from 2 × 2 tables and 95% CIs were
calculated from exact binomial methods. Multivariate unconditional logistic
regression using backward stepwise selection at a significance level of .05
was used to assess whether putative risk factors might modify the association
between preconception antibody status and risk of congenital CMV infection.
All data analyses were performed using SAS statistical software (Version 8,
SAS Institute Inc, Cary, NC).
Characteristics of the study population according to maternal CMV antibody
status at the time of the previous pregnancy are shown in Table 1. Overall, 2857 mothers (82.5%) had CMV antibodies at the
time of the previous pregnancy, and 604 (17.5%) were seronegative. The majority
of both groups were black, although seronegative mothers were more likely
to be white (33.4%) than were seropositive mothers (15.8%). Slightly more
seronegative mothers (22.0%) compared with seropositive mothers (15.3%) had
private health insurance, a difference that was statistically significant
(P<.001). Not surprisingly, the initial pregnancy
at which serological status was determined was the first pregnancy for more
of the seronegative group. Seronegative mothers were younger by less than
1 year in mean age, a difference that was statistically significant (P = .001).
Overall, congenital CMV infection occurred in 46 infants (1.3%) born
to study mothers. Of 604 newborns born to initially seronegative mothers,
congenital CMV infection occurred in 18 (3.0%). In contrast, of 2857 newborns
born to immune mothers, congenital CMV infection occurred in 29 (1.0%). In
the initially seronegative group, 142 women (23.5%) seroconverted between
deliveries. All of the congenital CMV infections among the initially seronegative
group occurred in the infants of mothers who seroconverted between pregnancies.
Risk factors for delivering an infant with congenital CMV infection
were examined for the entire study population (Table 2). Black race and lower socioeconomic status (Medicaid or
no insurance for hospital stay) were both associated with increased risk of
congenital CMV infection. However, the 95% CIs for both factors included 1.0.
Older maternal age (≥25 years) and gravidity (>2) were associated with
decreased risk of congenital CMV infection. The presence of maternal antibody
at the previous delivery was highly protective against delivering a future
newborn with congenital CMV infection (RR, 0.32; 95% CI, 0.17-0.62).
Maternal immune status at previous birth, maternal age, race, insurance
status, and gravidity were evaluated in a logistic regression model to simultaneously
assess the effects of these factors on the risk of congenital CMV infection.
Once maternal immune status at previous birth and maternal age of 25 years
or older were included in the model, none of the other potential risk factors
changed these estimates and they were not independently associated with congenital
CMV infection. Even after adjustment for age, maternal immunity remained strongly
protective against congenital CMV infection (adjusted RR, 0.31; 95% CI, 0.17-0.58).
Maternal age of 25 years or older remained strongly associated with reduced
risk of congenital CMV infection after adjustment for maternal immunity (adjusted
RR, 0.19; 95% CI, 0.07-0.49).
This study shows that young women who have immunity to CMV from naturally
acquired infection are 69% less likely to give birth to an infant with congenital
CMV infection in the future than are those who are initially CMV seronegative.
Among mothers CMV seronegative at a previous delivery, seroconversion between
deliveries is common (23.5% in this study). When seroconversion occurs between
deliveries, the risk of congenital CMV infection is high. We found a rate
of congenital CMV infection of 12.7% for offspring of mothers who seroconverted
during the interval (average, 3 years) between deliveries.
The major limitation of this study is that it included only multiparous
women. Ideally, in testing the ability of immunity from naturally acquired
CMV infection to prevent congenital CMV infection, a cohort of young women
prior to their first pregnancy would be enrolled after determining their CMV
immune status. These women would be followed up through subsequent pregnancies,
all offspring would be tested for congenital infection, and maternal seroconversions
would be defined by repeat CMV antibody testing. The expense and logistical
difficulties associated with enrollment and follow-up of such a cohort of
young women would likely make this type of study impossible to conduct. Our
study was limited to women who had at least 1 previous pregnancy because this
initial pregnancy was used as the starting point for determining immunity
to CMV prior to the subsequent pregnancy. The exclusion of younger primiparous
women may have led to an underestimation of the proportion of congenital infections
due to primary maternal infection, and therefore underestimation of the protective
effect of maternal immunity. Since the prevalence of maternal antibody increases
rapidly with age in young women, it is likely that a greater proportion of
younger women would have been seronegative near the time of their first pregnancy.
It is possible that our study has actually underestimated the ability of naturally
acquired immunity to prevent future congenital CMV infection by selecting
an older multiparous population in which fewer seroconversions will occur.
The current study is the first to compare rates of congenital CMV infection
based on maternal immunity status at a point prior to conception. The study
population, mothers who had more than 1 pregnancy with delivery of a live
newborn at the university hospital, was predominantly black (80%) and young,
with a mean age of 24 years. Previous studies of this delivery population
have reported a relatively high rate of congenital CMV infection (1.25%),
a high rate of primary infection during pregnancy, and the occurrence of congenital
CMV infection in women known to have been infected 1 year or more prior to
study population is at high risk for acquisition of CMV between pregnancies
with an annualized seroconversion rate of 7.8%, similar to rates reported
in day-care workers.11- 13 Published
reports indicate that exposure to young children and sexual activity are important
sources of CMV infection for young women.11,12,14,15 Those
exposures were not assessed in the current study and the only demographic
variable significantly associated with acquisition of CMV among seronegative
women was race. Considering the entire study population (seropositive and
seronegative women), maternal immunity and older maternal age were associated
with protection from congenital CMV infection. None of the other suggested
demographic or maternal risk factors were independently associated with congenital
CMV infection in this population.
Although the occurrence of congenital CMV infection in offspring of
women who are known to have been infected long before conception has been
previously reported, the current study probably provides the most accurate
estimate for congenital CMV infection rate. Of 2857 immune mothers from a
predominantly low-income, young, black population in the United States, 29
(1.0%) transmitted congenital CMV infection. Previous studies from this institution
reported a rate of 3.4% in women who were immune prior to conception; this
was based on a smaller sample size (n = 208) and a more highly selected study
population.6 Other studies have reported high
rates (1.2%-2.9%) of congenital CMV infection in populations in which almost
all women of childbearing age were seropositive, but they did not define congenital
CMV infection rates based on maternal immunity documented prior to conception.16- 18 However, even a rate
of congenital CMV infection of 1.0% is of great concern because the majority
of young women from the study population are immune to CMV, and recent studies
report that some congenital CMV infections in newborns born to immune mothers
are associated with central nervous system disease.19- 21 A
better understanding of the factors that influence occurrence of congenital
CMV infection and disease in offspring of immune mothers is needed, both for
planning preventive strategies and to estimate the potential effectiveness
Results reported here are of importance in planning strategies for prevention
of congenital CMV infection. Based on the reduction in risk of congenital
CMV infection associated with maternal immunity, a vaccine that could achieve
protection similar to that from immunity from naturally acquired infection
would be expected to reduce the rate of congenital CMV infection by at least
70%. Young maternal age was strongly associated with increased risk of congenital
CMV infection as has been observed in previous reports.9,22 Our
findings along with previous reports indicate that postponing pregnancy until
age 20 years or older could substantially reduce the rate of congenital CMV