Cytomegalovirus infection in pregnancy

INTRODUCTION — Cytomegalovirus (CMV) is a ubiquitous DNA herpesvirus. As with other herpesviruses, it becomes latent after a primary infection but can reactivate with renewed viral shedding. Shedding can occur from multiple sites and for prolonged periods of time. Women can also become infected with a different viral strain.

CMV is the most common congenital viral infection, with birth prevalence of 0.48 to 1.30 percent in recent decades [1-5]. Most infants with congenital CMV infection are asymptomatic, but approximately 10 to 15 percent of infected infants experience symptomatic infection, which can be severe and life-threatening. Both asymptomatic and symptomatic newborns are at risk of developing long-term neurodevelopmental morbidity, particularly sensorineural hearing loss.

This topic will discuss issues specific to CMV in pregnant and breastfeeding women. General issues related to CMV infection and CMV infections in other populations are reviewed separately.

●(See "Epidemiology, clinical manifestations, and treatment of cytomegalovirus infection in immunocompetent adults".)

●(See "Overview of cytomegalovirus infections in children".)

●(See "Congenital cytomegalovirus infection: Clinical features and diagnosis".)

●(See "Congenital cytomegalovirus infection: Management and outcome".)

MATERNAL CMV INFECTION

Routes of transmission — Maternal acquisition of CMV infection can occur via multiple routes, including close nonsexual contact (including household and occupational exposure [especially contact with young children]), sexual exposure, transfusion, and organ transplant. CMV has been cultured from multiple body fluids, including urine, saliva, blood, nasopharyngeal secretions, tears, cervical and vaginal secretions, semen, and breast milk. Transmission from respiratory droplets or aerosolized droplets is unlikely [6]. (See "Epidemiology, clinical manifestations, and treatment of cytomegalovirus infection in immunocompetent adults", section on 'Transmission'.)

Seroprevalence — CMV infection is common. A systematic survey of the literature estimated global CMV seroprevalence in women of childbearing age was 86 percent, with the highest seroprevalences in the Eastern Mediterranean, Western Pacific, African, and Southeast Asian regions (seroprevalence range 89 to 92 percent) and the lowest in the European region and the Americas (seroprevalence 70 and 79 percent, respectively) [7].

Seroprevalence increases with age and varies by geographic residence, race, ethnicity, and demographic factors [1-3,7,8]. The following characteristics are predictive of positive CMV serology:

●Lower socioeconomic strata.

●Contact with children under age 3 years, especially if they are in daycare [2,9,10].

●Non-Hispanic Black or Mexican American versus non-Hispanic White race [3].

●Age older than 25 to 30 years [2,11].

●Higher parity [12].

●Residence in a developing country [7].

Risk of seroconversion — The likelihood of seroconversion depends on social, behavioral, and environmental factors. Annual rates of maternal seroconversion are generally reported to be 1 to 7 percent worldwide in populations with low-to-intermediate CMV seroprevalence [13]. However, seronegative pregnant women living in high seroprevalence areas are at higher risk; estimated annualized rates of seroconversion of 14.8 to 22.5 percent have been reported [14-16].

In a systematic review of studies that measured rates of CMV seroconversion, summary annual rates of seroconversion in specific groups were as follows [13]:

●Pregnant women – 2.3 percent (95% CI 2.1-2.4 percent)

●Health care workers, including those caring for infants and children – 2.3 percent (95% CI 1.9-2.9 percent)

●Daycare providers – 8.5 percent (95% CI 6.1-11.6 percent)

●Parents of a child:

•Not shedding CMV – 2.1 percent (95% CI 0.3-6.8 percent)

•Shedding CMV – 24 percent (95% CI 18-30 percent)

Other groups with elevated risk of seroconversion included families with a CMV-shedding member, female minority adolescents, women attending sexually transmitted disease clinics, and immunocompromised individuals. The risk related to parenthood was illustrated in a study in which 15.6 percent of women seronegative at their first pregnancy seroconverted by the time of a subsequent pregnancy that occurred within two years [17]. Twenty-nine percent of the seroconversions occurred periconceptionally or in the first trimester, thereby placing the subsequent child at risk of sequelae from primary maternal CMV infection.

However, seroconversion rates were reduced to 0.6 to 1.3 percent in some large studies with use of preventive behavioral interventions [18]. (See 'Behavioral risk reduction interventions' below.)

Clinical findings — Primary CMV infection may cause a mild febrile illness and other nonspecific symptoms (rhinitis, pharyngitis, myalgia, arthralgia, headache, fatigue), but is not clinically apparent in approximately 90 percent of cases. CMV mononucleosis can be accompanied by dermatologic manifestations in approximately one-third of patients including macular, papular, maculopapular, rubelliform, morbilliform, and scarlatiniform eruptions.

Reinfection with a different strain of CMV or reactivation of virus in women with preexisting antibody generally does not result in maternal symptoms [19].

Pregnancy does not appear to affect clinical severity, but the integrity of the host immune system affects the spectrum of disease: Hosts with impaired cellular immunity are at risk for severe and disseminated infection.

The clinical manifestations of CMV infection in adults are discussed in detail separately. (See "Epidemiology, clinical manifestations, and treatment of cytomegalovirus infection in immunocompetent adults".)

Diagnosis

Whom to test — Testing pregnant women for CMV is indicated [20]:

●As part of a diagnostic evaluation of a mononucleosis-like or influenza-like illness, especially in the setting of negative Epstein-Barr and influenza virus tests (see "Infectious mononucleosis", section on 'Diagnosis' and "Seasonal influenza in children: Clinical features and diagnosis", section on 'Diagnosis') or as part of the diagnostic evaluation of patients with symptoms of hepatitis (eg, fever, fatigue, loss of appetite, nausea, vomiting, abdominal pain, dark urine, light-colored stools, joint pain, jaundice) in the setting of negative hepatitis A, B, and C serologies.

●When a fetal anomaly suggestive of congenital CMV infection is detected on prenatal ultrasound examination. (See 'Ultrasound markers and monitoring' below.)

How to test — The diagnosis of clinically suspected maternal primary CMV infection is based on serology (table 1). Seroconversion of CMV-specific immunoglobulin G (IgG) in paired acute and convalescent sera collected 3 to 4 weeks apart is diagnostic of a new acute infection. The presence of CMV immunoglobulin M (IgM) is not helpful for timing the onset of infection because (1) it is present in only 75 to 90 percent of women with acute infection, (2) it can remain positive for over one year after an acute infection, (3) it can revert from negative to positive in women with CMV reactivation or reinfection with a different strain, and (4) it can become positive in response to other viral infections, such as Epstein-Barr virus.

In the absence of documented recent seroconversion, it is difficult to distinguish between primary infection, reactivation, reinfection, and quiescent disease since all are associated with IgG and IgM antibodies, and rising titers alone are not diagnostic. Determination of IgG avidity is helpful to better assess the acuity of the infection and thus the risk of in utero transmission [21-24]. High anti-CMV IgG avidity suggests that the primary infection occurred more than six months in the past; low avidity anti-CMV IgG suggests a recent primary infection (within two to four months) [25]. It should be noted that commercial avidity antibody assays have varying performance characteristics [26], the interpretation of intermediate avidity and the optimal cutoffs for low and high avidity are not well established [27,28], and the cutoffs for low and high avidity vary among laboratories. The diagnosis of CMV is discussed in detail separately. (See "Overview of diagnostic tests for cytomegalovirus infection".)

Serologic tests cannot differentiate among the many strains of CMV.

Classification — CMV infections in pregnant women are generally classified as primary or nonprimary.

●Primary – Primary maternal CMV infection is defined as initial acquisition of virus during pregnancy. Seroconversion from negative to positive is diagnostic, if available. Primary infection is strongly suspected if IgM and IgG are positive and IgG has low avidity.

It is estimated that 25 percent of congenital CMV infections in the United States result from primary maternal infection and the remaining three-quarters of infants with congenital CMV infection are born to women with preexisting seroimmunity to CMV (nonprimary maternal infection) [29]. This estimate was based on summary data on maternal-fetal transmission rates from a comprehensive literature review and previously reported nationally representative data on age and race-specific CMV seroprevalence and seroincidence.

Among pregnant women with a primary CMV infection, the overall risk of severe newborn sequelae based on maternal serologic diagnosis alone is approximately 3 percent and risk of any adverse outcome is approximately 8 percent [30]. These risks are obviously higher if the fetus is confirmed to be infected. (See 'Frequency of maternal-fetal transmission' below and 'Clinical features and sequelae' below.)

●Nonprimary – Nonprimary maternal CMV infection occurs in women with initial acquisition of virus before pregnancy and is characterized by presence of maternal anti-CMV antibodies before conception. Like other herpesviruses, CMV establishes latency after the host is initially infected. Nonprimary infection, also sometimes called recurrent or secondary infection, may be due to reactivation of latent virus or reinfection with a new strain.

Because the diagnosis of nonprimary CMV infection is difficult, it is not known how many women have reactivation or reinfection during pregnancy, how many congenital infections result from reactivation or reinfection, and what the exact frequency and full spectrum of clinical sequelae are in offspring after reactivation or reinfection [31]. In a prospective study that collected saliva, urine, vaginal swabs, and blood from 120 CMV-seropositive pregnant women in the first, second, and third trimesters and one month postpartum, CMV shedding detected by polymerase chain reaction was documented at least once in approximately one-third of the women [32]. In another prospective study that collected urine across gestation from 240 CMV-seropositive pregnant women, CMV shedding detected by the CMV immediate-early 1 gene was documented at least once in 75 percent of women [33]. The difference in the CMV shedding rates between the two studies could be due to differences in population characteristics or the difference in sample processing and PCR technology used in the studies.

Available data on pregnancy risks of nonprimary CMV infection are based on women with preconception immunity. These data suggest that the frequency of fetal infection is low in women who seroconverted prior to conception, but infected fetuses are still at risk for symptomatic and asymptomatic newborn disease and long-term sequelae [4,34-38]. In one study, the frequency of hearing loss in infants with congenital CMV infection was approximately 10 percent, regardless of whether seroconversion occurred before or after conception [39]. However, the number of children with severe-to-profound hearing loss was significantly lower in children whose mothers seroconverted before pregnancy.

Treatment — Immunocompetent pregnant women with CMV infection should be offered supportive care for symptomatic relief, as needed (eg, acetaminophen for fever). Use of antiviral drugs for treatment of CMV infections in immunocompetent adults, including pregnant women, is rarely indicated. Several medications (eg, ganciclovir, foscarnet, cidofovir) are available to treat severe end-organ CMV disease, but none have been shown to decrease perinatal transmission, and information on fetal risks is limited. (See "Epidemiology, clinical manifestations, and treatment of cytomegalovirus infection in immunocompetent adults", section on 'Therapy'.)

CONGENITAL INFECTION

Route of transmission — Maternal viremia that leads to placental infection, followed by transplacental transmission to the fetus is the major source of fetal infection. Placental cytotrophoblasts are permissive to CMV replication.

Postnatal infection can occur via intrapartum exposure to cervical/vaginal viral shedding or via consumption of infected breast milk and is unlikely to cause severe disease in healthy full-term neonates [40,41].

Frequency of maternal-fetal transmission

Seroconversion during pregnancy — Women who seroconvert during pregnancy (primary infection) are at highest risk for maternal-fetal transmission and the rate of transmission appears to increase with advancing gestational age.

A review that pooled data from 10 studies (2942 fetuses) of maternal-fetal CMV transmission in women who seroconverted just before or during pregnancy reported the following rates of transmission [42]:

●Preconception period (up to 12 weeks before the last menstrual period) – 5.5 percent

●Periconceptional period (4 weeks before to 6 weeks after the last menstrual period) – 21 percent

●First trimester – 36.8 percent

●Second trimester – 40.3 percent

●Third trimester – 66.2 percent

Although the risk for maternal-fetal viral transmission is lower in early compared with late pregnancy, the risk for symptomatic disease at birth and long-term sequelae is higher when infection occurs in early pregnancy. (See 'Symptomatic newborns' below.)

Seroconversion remote from conception — As with other herpesvirus infections, periodic reactivation of latent virus is possible, especially in immunocompromised hosts (eg, organ transplant recipients, individuals with human immunodeficiency virus [HIV] infection). Reinfection with different viral strains is also possible. Although maternal antibodies to CMV formed as a result of primary infection provide protection [11], they do not prevent reactivation or reinfection with a different strain, which is relatively common and leads to transient excretion of virus during pregnancy [32]. Thus, seropositivity before pregnancy does not eliminate the risk of congenital infection [43].

The overall risk of fetal infection among seropositive women is low: 0.15 to 2 percent [14,44]; however, because the population of seropositive women is large, most congenitally infected infants are born to seropositive mothers. The frequency of intrauterine transmission in women with preexisting seroimmunity (nonprimary maternal infection) in relation to the timing of maternal infection (reinfection/reactivation) is not known.

Women with HIV appear to be an exception. A higher birth prevalence of congenital CMV infection has been associated with advanced maternal immunosuppression and has been observed despite maternal antiretroviral prophylaxis [45-49]. A study from the United States reported congenital CMV rates remained high in HIV-exposed but uninfected infants born during the era of antiretroviral therapy [50], while a French study reported congenital CMV prevalence decreased in infants born to women on antiretroviral therapy [51].

Clinical features and sequelae — The following key principles are important for understanding congenital infection:

●The occurrence of fetal infection increases with advancing gestational age in women with primary CMV infection; however, the occurrence of symptomatic disease in offspring decreases with advancing gestational age and is unlikely if a primary maternal infection occurs near term [34].

●Preconception seroimmunity provides substantial protection against the occurrence of fetal infection compared with seroconversion during early pregnancy (approximately 0.15 to 2 percent versus 40 percent) [11,34,44].

●Once fetal infection occurs, the frequency of newborn disease and long-term sequelae is similar for infants of mothers with primary CMV infection during pregnancy and those who were CMV seropositive prior to pregnancy [37,39,52-54].

●Congenital infection may be symptomatic or asymptomatic in newborns. Most congenital infections are asymptomatic in the neonatal period.

●Both symptomatic and asymptomatic infected newborns are at risk for development of adverse sequelae in early childhood, but symptomatic newborns are at higher risk (eg, death 5 versus 0 percent, sensorineural hearing loss 50 versus 10 percent in one review) [55].

Asymptomatic newborns — Most congenitally infected newborns are initially asymptomatic. However, 15 to 25 percent of these initially asymptomatic newborns go on to develop neurodevelopmental abnormalities, most commonly sensorineural hearing loss, within the first three years of life [44,56].

Symptomatic newborns

●Clinical findings – Clinical findings in symptomatic newborns include small for gestational age, microcephaly, ventriculomegaly, chorioretinitis, jaundice, hepatosplenomegaly, thrombocytopenia, and petechiae. These findings are thought to result from a combination of viral cytopathic effects and immune responses to the virus in different organs (eg, salivary gland, lung, liver, kidney, intestine, adrenal gland, placenta, central nervous system).

The mortality rate among symptomatic newborns is approximately 5 percent [57,58], and 50 to 60 percent of survivors develop serious long-term neurologic morbidity (eg, progressive sensorineural hearing loss and/or visual impairment, motor/cognitive impairment) [44,59]. (See "Congenital cytomegalovirus infection: Clinical features and diagnosis".)

●Frequency of symptomatic infection after maternal seroconversion – The frequency of symptomatic newborn disease has been well studied in women with primary CMV infections in pregnancy. Symptomatic disease at birth and severe sequelae occur significantly more frequently among offspring of women who seroconvert in the first half of pregnancy, particularly the first trimester [34,60,61].

A review that pooled data from 10 studies (796 fetuses) reported the following rates of fetal/newborn impairment (ie, neurologic symptoms at birth or termination of pregnancy because of CMV-associated findings in the central nervous system on ultrasonography or magnetic resonance imaging) when seroconversion occurred during the [42]:

•Periconceptional period (4 weeks before to 6 weeks after the last menstrual period) – 28.8 percent (95% CI 2.4-55.1)

•First trimester – 19.3 percent (95% CI 12.2-26.4)

•Second trimester – 0.9 percent (95% CI 0-2.4)

•Third trimester – 0.4 percent (95% CI 0-1.5)

The frequency of symptomatic newborns is not well established in nonprimary infection. Data are limited to case reports and small case series of symptomatic newborns of mothers with known preconceptional immunity, not proven reactivation or reinfection during pregnancy [37,39,52-54].

Placental histopathology — The classic histopathologic placental findings include the following, although not all may be present:

●Lymphoplasmacytic villitis (diffuse chronic villitis with plasma cells)

●Sclerosis of the villous capillaries

●Chorionic vessel thromboses

●Necrotizing villitis

●Hemosiderin deposition in the villous stroma

●Normoblastemia

Virus replication has been demonstrated in smooth muscle cells of arteries and veins in floating villi and the chorion [62]. Large fibrinoids with many avascular villi and edematous villi and inflammation, changes that likely impair placental transport, have been observed in births complicated by intrauterine growth restriction and primary or recurrent CMV infection. Progressive fetal thrombotic vasculopathy has been observed in stillbirths.

Viral inclusions (picture 1), which can be subtle, are observed in 10 percent of cases in fetal infection but are more often visible in cases associated with stillbirth. Immunohistochemistry will detect many inclusions not identified with routine hematoxylin-eosin stains [63].

PRENATAL (FETAL) DIAGNOSIS

Our approach — We suggest offering amniocentesis for prenatal (fetal) diagnosis when fetal infection is suspected because of primary maternal infection or findings on ultrasound (see 'Ultrasound markers and monitoring' below). The rationale for offering prenatal diagnosis is to determine whether the fetus is infected (vertical transmission is not 100 percent). Some couples may use this information in decision-making regarding termination of pregnancy. It also helps couples prepare for the birth of an infected and possibly clinically affected child. Lastly, it may change the intensity of fetal monitoring (eg, frequency of ultrasound examination) and is informative for the pediatricians caring for the child. Maternal drug therapy has not been proven to improve outcome. (See 'Prenatal drug therapy' below and "Congenital cytomegalovirus infection: Clinical features and diagnosis".)

Amniocentesis — Amniocentesis to perform polymerase chain reaction (PCR) for CMV DNA in amniotic fluid is the preferred diagnostic approach for identifying an infected fetus; viral culture is less desirable because of several limitations. Reported sensitivity of PCR ranges from 70 to 100 percent [64-68]. (See "Overview of diagnostic tests for cytomegalovirus infection".)

Technically, amniocentesis can be performed at any time ≥15 weeks of gestation. The interval from maternal infection to amniocentesis impacts the ability to detect CMV DNA in amniotic fluid; gestational age also appears to play a role. Since it takes six to eight weeks for placental infection and replication, transmission to the fetus, viral replication in the fetal kidney, and excretion into amniotic fluid (fetal urination begins near the end of the first trimester and increases with advancing gestation), an interval of at least eight weeks between maternal infection and amniocentesis is desirable to increase diagnostic sensitivity.

Traditionally, it has been reported that sensitivity is highest after 21 weeks of gestation; however, in a retrospective study that reviewed 264 pregnancies between 17+0 weeks and 22+6 weeks of gestation with an interval >8 weeks between seroconversion and amniocentesis, diagnostic sensitivity was similar before versus after 21 weeks (87.2 and 92.1 percent, respectively), as was negative predictive value (93.6 and 96.8 percent, respectively). This finding highlights the importance of the lag time between maternal infection and amniocentesis. Women infected early in the first trimester may have positive PCR for CMV DNA in amniotic fluid before 21 weeks because 6 to 8 weeks may have elapsed since the maternal infection and thus allowed time for placental infection, fetal transmission, viral replication in the fetal kidney, and excretion.

Therefore, when amniocentesis is performed before 21 weeks or soon after diagnosis of maternal infection, positive results are reliable evidence of fetal infection, but negative results do not reliably exclude fetal infection, so amniocentesis should be repeated later in gestation to allow time for placental to fetal transmission and fetal excretion (algorithm 1). Rarely, false positive results occur from contamination of the amniotic fluid sample by maternal fluids. The first 1 mL of fluid obtained should be discarded to reduce the risk of contamination [64-66,69].

Presence of CMV DNA in maternal blood at the time of amniocentesis does not appear to be a significant risk factor for iatrogenic antepartum transmission [70].

Ultrasound markers and monitoring — The following ultrasonographic markers are suggestive, but not diagnostic, of fetal CMV infection in the presence of a known maternal infection [2,64-66,71-75].

●Periventricular calcifications (image 1)

●Cerebral ventriculomegaly (image 2)

●Microcephaly

●Pseudocysts, periventricular or adjacent to the occipital or temporal horn

●Hyperechogenic fetal bowel (image 3)

●Fetal growth restriction

●Ascites (image 4)

●Pleural and/or pericardial effusion (image 4 and image 5)

●Hepatosplenomegaly

●Hepatic calcifications

●Polymicrogyria

●Cerebellar hypoplasia

●Large cisterna magna

●Amniotic fluid abnormalities (oligohydramnios or polyhydramnios)

●Hydrops

●Placental thickening and enlargement, heterogeneous appearance, calcifications

The most characteristic sonographic finding of fetal CMV infection is bilateral periventricular hyperechogenicities (calcifications) (image 1) [76,77]. These calcifications or hyperechoic foci can be highly reflective and may not cast an acoustic shadow [78]. Branching linear echogenic areas in the thalami also occur and correspond to arteries in the basal ganglia and thalamus [79,80]. The presence of intraventricular filmy, thin adhesions and linear edges traversing the ventricle is typical in CMV infection of the brain [74,81-89].

In infected fetuses, serial ultrasound examinations at two- to four-week intervals can be useful to detect development of sonographic abnormalities. Ultrasound abnormalities can appear 12 or more weeks after maternal infection [68,90].

Although there is a considerable overlap in imaging findings between congenital CMV infection and congenital Zika syndrome, severe microcephaly, evidence of fetal brain collapse (fetal brain disruption sequence), and evidence of contractures are most characteristic of intrauterine Zika virus infection [91]. In addition, the distribution of intracerebral calcifications is subcortical in congenital Zika syndrome but periventricular in congenital CMV infection.

Prenatal prediction of offspring outcome — Accurate prediction of poor prognosis for infected fetuses is difficult and based primarily on the gestational age when the infection occurred (ie, early maternal infection leads to a lower risk of congenital infection than late pregnancy maternal infection but a high risk of clinical consequences in infected offspring) and the presence and type of fetal cerebral abnormalities.

●Abnormal ultrasound – Persistent fetal abnormalities, such as ventriculomegaly, periventricular calcifications, microcephaly, growth restriction, and hydrops, suggest the presence of severe disease and high risk of long-term neurodevelopmental impairment. If an abnormality is suspected on ultrasound and clarification is needed, magnetic resonance imaging (MRI) may helpful. There are no abnormal ultrasound or MRI findings specifically predictive of postnatal hearing loss. (See 'Magnetic resonance imaging' below.)

●Normal ultrasound – A normal ultrasound examination is reassuring but does not completely exclude the possibility of an infected fetus, symptomatic neonate, or development of long-term neurologic morbidity. Likewise, the combination of a normal ultrasound and a negative PCR on amniotic fluid does not completely exclude the possibility of congenital infection, because PCR sensitivity is not 100 percent and maternal-fetal transmission may occur after the amniocentesis.

A systematic review including 1178 fetuses with a normal ultrasound examination at the time of maternal diagnosis of CMV infection reported the following major findings [92]:

•An associated CNS anomaly was detected on follow-up ultrasound in 4.4 percent of cases (95% CI, 1.4-8.8 percent).

•An associated extra-CNS anomaly was detected on follow-up ultrasound in 2.9 percent of cases (95% CI, 0.8-6.3 percent).

•Symptomatic infection occurred in 1.5 percent of cases (95% CI, 0.7-2.7 percent), and the overall rate of a neurodevelopmental anomaly in these cases was 3.1 percent (95% CI, 1.6-5.1 percent).

•Hearing problems affected 6.5 percent of children (95% CI 3.8-10 percent).

•Motor and cognitive delays and visual problems were reported in 2.3, 1.1, and 1 percent of children, respectively.

Importantly, when trimester of maternal infection was considered:

•An anomaly was subsequently detected only among fetuses infected in the first trimester.

•Abnormal childhood neurodevelopment outcome only occurred among fetuses infected in the first trimester, and the rate was 5.4 percent in this group.

•Hearing problems occurred in 11.4 percent of children infected in the first trimester, 7 percent of those infected in the second trimester, and none of those infected in the third trimester.

If the fetus is infected and the ultrasound is normal, determination of viral load in amniotic fluid may help to distinguish those who are infected but likely to be asymptomatic at birth from those who are likely to be symptomatic and at high risk of developing serious sequelae [2,93-95]. For example, a study of 456 women at weeks 21 to 23 of pregnancy found that higher viral loads in amniotic fluid (eg, greater than 100,000 copies/mL) were associated with symptomatic newborns [2,93]. In another small study of 21 fetuses, the median DNA level in amniotic fluid was higher in symptomatic newborns, but the difference was not statistically significant [94]. Lastly, a study of 82 fetuses primarily exposed to first-trimester maternal infection reported the negative predictive value for symptoms at birth or at termination of pregnancy was 93 percent for ultrasound alone versus 95 percent for ultrasound and viral load in amniotic fluid [95].

Other tests

Fetal blood sampling — Cordocentesis for the purpose of evaluating for fetal CMV disease is not recommended. Cordocentesis to test fetal blood does not significantly increase diagnostic sensitivity or specificity compared with amniotic fluid tests, but increases the risk of fetal loss. Although the presence of abnormal liver function and hematologic tests (especially thrombocytopenia) and elevated beta-2 microglobulin level are signs of severe disease, individual marker levels are not sufficiently reliable for distinguishing between asymptomatic and symptomatic disease at birth and predicting the likelihood of an unfavorable long-term outcome [96,97].

In a study that looked at combined markers, the negative predictive value for symptoms at birth or at termination of pregnancy was 100 percent when the combination of ultrasound findings, viral load in amniotic fluid, and fetal blood parameters were considered [95]. The positive predictive value was 79 percent when the combination of ultrasound findings, amniotic fluid viral load, fetal blood parameters were considered. The authors concluded that a small increase in predictive value with fetal blood sampling did not clearly warrant exposure to the additional risks of fetal blood sampling, but this decision should be made based on individual patient values.

Magnetic resonance imaging — MRI may provide additional information about anomalies, particularly neurologic abnormalities. In a systematic review, an anomaly was detected on MRI in approximately 6 percent of fetuses with a normal ultrasound examination [92]. However, subtle nonprogressive findings on MRI, particularly when the ultrasound is normal, are not necessarily predictive of an affected neonate [98], and a normal ultrasound and MRI do not completely exclude the possibility of development of postnatal hearing loss [90,99]. The cost of the additional testing and likelihood of gaining information that will alter management should be considered prior to obtaining MRI in the setting of known or suspected fetal CMV infection.

PRENATAL CARE AND DELIVERY — In addition to routine prenatal care, symptomatic women are offered supportive therapy. Antiviral therapy for maternal indications is rarely indicated. (See 'Treatment' above.)

Prenatal diagnosis is offered to women at high risk of having an affected fetus (primary maternal infection or suggestive findings on ultrasound) and their fetuses are monitored with serial ultrasound examinations. (See 'Prenatal (fetal) diagnosis' above.)

Hyperimmunoglobulin therapy of pregnant women with primary CMV infection in early pregnancy is an investigational approach for reducing symptomatic infection in offspring (see 'CMV immune globulin' below). Antiviral therapy is not indicated for fetal therapy. (See 'Antiviral drug therapy' below.)

The timing and route of delivery are determined by standard maternal and fetal indications. Recovery of CMV from the cervix or urine is not an indication for cesarean delivery.

STRATEGIES FOR PREVENTION OF MATERNAL AND/OR FETAL INFECTION

Screening — We suggest not screening all pregnant women for primary CMV infection.

In our practice, we obtain CMV serology in HIV-infected women at their initial prenatal visit and repeat the serology in screen-negative women who develop signs/symptoms suggestive of CMV infection or have suggestive fetal ultrasound findings. For any woman with a known exposure to CMV, we obtain baseline serology at the time of the exposure and, if negative, repeat serology several weeks later to assess for seroconversion.

Since CMV infection is usually asymptomatic and transmissible to the fetus, some authorities suggest that all women of childbearing age should know their CMV serostatus, although there is no consensus [64,65,100,101]. Others, including the American College of Obstetricians and Gynecologists [102] and Society for Maternal-Fetal Medicine [30], recommend against routine serologic screening for CMV for several reasons:

●No vaccine is available to prevent infection in seronegative women.

●In seropositive pregnant women, it is difficult to distinguish between primary and nonprimary infection or determine the timing of the infection, which could have occurred many months before conception.

●Seropositive women remain at risk of fetal infection from reactivation of latent virus and/or reinfection with a new viral strain. In a study in Japan, universal screening of pregnant women using CMV-IgG and IgG avidity identified only 3 of the 10 infants with congenital CMV infection [103]. The three infants were born to mothers with primary CMV; the other seven congenitally infected infants were born to mothers with nonprimary CMV.

●There is no evidence that antiviral drug treatment of primary infection in pregnant women prevents or mitigates sequelae of CMV infection in the neonate. (See 'Antiviral drug therapy' below.)

●Randomized trials of use of hyperimmune globulin to prevent congenital infection have not established a benefit, in contrast to observational studies. (See 'CMV immune globulin' below.)

●Although fetal infection can be detected, there is no way to accurately predict whether or not the fetus will develop significant sequelae.

●Routine screening can lead to unnecessary, and potentially harmful, intervention.

On the other hand, proponents of universal screening argue that knowing that the patient's serology is negative for CMV antibodies and CMV counseling increase some patients' motivation to practice good hygiene and thus decrease the risk of seroconversion during pregnancy. Knowledge of CMV seronegativity and education about routes of transmission appears to change maternal behavior and decrease seroconversion in pregnant women at high risk [71,104-106]. However, this information does not appear to significantly affect seroconversion rates in nonpregnant women, even in those trying to conceive [105].

Repeated serologic screening during pregnancy to detect seroconversion is not commonly performed because of cost, lack of effective treatment, and generally poor ability of positive serology to predict the fetus's long-term outcome.

Screening children for CMV or excluding CMV-excreting children from schools or institutions is unnecessary because the virus is frequently found in many healthy children and adults.

Behavioral risk reduction interventions

Pregnant women — All pregnant women should be aware of CMV prevention measures; however, no actions can eliminate all risks of becoming infected with CMV. The following measures may reduce the risk of transmission [18]:

●Practice good personal hygiene throughout pregnancy, especially hand washing with soap and water after contact with diapers or oral and nasal secretions (particularly with a child who is in daycare). Wash well for at least 15 to 20 seconds.

●Avoid kissing children under age 6 on the mouth or cheek; instead, kiss them on the head or give them a hug.

●Do not share food, drinks, or oral utensils (eg, fork, spoon, toothbrush, pacifier) with young children.

●Clean toys, countertops, and other surfaces that come into contact with children's urine or saliva.

Women who care for young children

●Female childcare employees should be educated concerning CMV, its transmission, and hygienic practices, such as handwashing, which minimize the risk of infection.

●Pregnant employees working with infants and young children should be informed of their increased risk of acquiring CMV infection and the possible effects on the unborn child.

●Routine laboratory screening for CMV antibody in female childcare workers is not recommended. Susceptible women working with infants and young children do not have to be transferred to other work situations but may reduce their risk of infection by practicing good hygiene.

Health care workers — The risk of CMV infection among health care workers appears to be no greater than that among the general public. This is probably due, at least in part, to adherence to standard precautions by health care providers when handling body fluids and less personal contact in the health care setting than in the family setting.

Women with recent infection — Because CMV DNA has been detected in blood of 20 percent of immunocompetent patients as long as six months after diagnosis of primary infection, some authorities suggest that a woman wait at least six months after a primary infection before attempting to conceive; however, data are limited and other experts suggest waiting a minimum of three or four months [2,107].

Breastfeeding women — The demonstrated benefits of breastfeeding outweigh the minimal risk of acquiring CMV from an infected breastfeeding mother.

HIV-infected mothers in the areas where formula is readily available are advised to not breastfeed because of a possible risk of HIV transmission. (See "Prenatal evaluation of women with HIV in resource-rich settings", section on 'Counseling'.)

Transfusion of CMV-negative blood — CMV seronegative pregnant women, fetuses, and newborns should be transfused, when necessary, with blood from CMV seronegative donors. (See "Practical aspects of red blood cell transfusion in adults: Storage, processing, modifications, and infusion", section on 'Specialized modifications and products'.)

Prenatal drug therapy

Antiviral drug therapy — Maternal antiviral therapy is not recommended clinically as there are only limited data on its effects on the prevention [108,109] or outcome of congenital infection [110,111]. If such therapy is offered to patients who seroconvert during pregnancy, it should be done in the setting of a research protocol or with the understanding that it remains an investigational approach.

Strategies involving antiviral therapy for prevention of vertical transmission are challenging given that most maternal infections are asymptomatic and pregnancy CMV screening is not performed. However, small studies in clinical settings with well-established routine maternal serum screening policies in early pregnancy have reported that patients with primary CMV infection diagnosed in the first trimester who began treatment with high-dosage oral valacyclovir (4 grams orally two times daily) before 14 weeks of gestation had a 70 percent reduction in fetal infection, as assessed by amniocentesis at 17 to 22 weeks of gestation for polymerase chain reaction (PCR) of amniotic fluid [108,109].

Strategies for antiviral treatment of infected fetuses identified by ultrasound and diagnostic testing are also promising. In a multicenter, open-label, phase II study of oral valacyclovir treatment of pregnant women carrying a CMV-infected fetus with measurable extracerebral or mild cerebral ultrasound findings, the intervention improved neonatal outcome of infected fetuses without severe brain abnormalities [111]. Compared with a historical cohort obtained by a meta-analysis, the use of valacyclovir increased the proportion of asymptomatic neonates from 43 percent without treatment to 82 percent with treatment. Valacyclovir 8 g daily was initiated at a median of 25.9 weeks of gestation and continued until delivery or termination of pregnancy. Adherence to treatment was >90 percent despite the need to take 16 pills/day, and the high dose was well tolerated. However, the findings are limited by the open-label study design and should be confirmed in a randomized trial to better determine the efficacy of in utero treatment before it can be recommended.

CMV immune globulin — CMV-specific hyperimmune globulin therapy of pregnant women with primary CMV infection in early pregnancy is an investigational approach to reducing symptomatic infection in offspring, but two randomized trials did not show efficacy [112,113]:

●In the largest of these trials, which randomly assigned 394 pregnant women diagnosed with primary CMV infection before 24 weeks of gestation to a monthly infusion of CMV hyperimmune globulin (100 mg/kg of body weight) or matching placebo until delivery, the composite outcome (congenital CMV infection or fetal/neonatal death in the absence of CMV testing) was similar in both groups (22.7 versus 19.4 percent; RR 1.17, 95% CI 0.80-1.72) [113]. There was also no clear difference in individual outcomes, such as death (4.9 versus 2.6 percent; RR 1.88, 95% CI 0.66-5.41), preterm birth (12.2 versus 8.3 percent; RR 1.47, 95% CI 0.81-2.67), and birth weight below the fifth percentile (10.3 versus 5.4 percent; RR 1.92, 95% CI 0.92-3.99).

●In the Congenital Human CMV Infection Prevention (CHIP) trial, 124 pregnant women at 5 to 26 weeks of gestation with recent onset of primary CMV infection were randomly assigned to receive CMV-specific hyperimmune globulin or placebo every four weeks until 36 weeks of gestation or detection of CMV in amniotic fluid [112]. The overall rate of congenital infection was similar for both groups (30 versus 44 percent in the placebo group, 95% CI -3 to 31 percentage points). The proportion of infected infants symptomatic at birth was also similar for both groups (3 of 10 CMV-infected newborns in the hyperimmune globulin group [30 percent] versus 4 of 17 CMV-infected newborns in the placebo group [24 percent]). There were also no differences in the viral or immune characteristics of infected infants between the groups. In addition, the number of obstetric adverse events trended higher in the hyperimmune globulin group who remained in the study until delivery than in the placebo group (7 of 53 [13 percent] versus 1 of 51 [2 percent]). These events consisted of preterm birth, preeclampsia, and fetal growth restriction.

An important limitation of these trials is that the investigators did not evaluate longterm hearing loss, which is an important clinical outcome that potentially could be affected by this therapy. Other limitations are the possibilities that CMV hyperimmune globulin was administered to some women after their fetuses were already infected and that some women became infected after being screened. Lastly, the sample size of the largest trial would detect a statistically significant reduction in the primary outcome only if greater than 50 percent.

Given these findings, practitioners should emphasize the behavioral preventive measures noted above. CMV hyperimmune globulin therapy should only be used in a research setting until more data are available. (See 'Strategies for prevention of maternal and/or fetal infection' above.)

Vaccine development — No CMV vaccine is available for use in humans, although several candidate vaccines have been developed and tested in clinical trials. In a phase 2 trial that included 464 CMV-seronegative women of childbearing age, an MF59-adjuvanted CMV glycoprotein B subunit vaccine had 50 percent efficacy at preventing CMV infection, which was the primary endpoint [114]. Congenital CMV infection was diagnosed in one infant of a vaccinated subject versus three infants of subjects who received placebo. The overall benefits were modest and the study was not powered to assess efficacy in preventing maternal-fetal transmission [115]. The MF59-adjuvanted CMV glycoprotein B vaccine has also been studied in seronegative adolescent girls for the prevention of primary CMV infection; vaccine efficacy was 43 percent and higher than placebo [116]. The same vaccine has also been studied in solid organ transplant recipients and showed a significant reduction in the duration of viremia and the number of days of ganciclovir treatment in seronegative recipients with seropositive donors in the vaccine group [117].

It is unlikely that a CMV vaccine will be available for several years. In addition, the strategy of preventing primary maternal infection does not address the CMV-associated hearing loss and other neurologic sequelae in congenitally infected children born to women with preexisting CMV immunity, who account for a larger proportion of congenital infection. Ideally, a CMV vaccine that induces high titers of cross-neutralizing antibodies will be developed and will protect individuals from infection with antigenically different CMV strains [118]. Although the exact nature of protective immune responses against CMV has not been characterized, a recombinant CMV glycoprotein B vaccine with MF59 adjuvant appeared to boost both antibody and CD4 T-cell responses in previously CMV-seropositive women, thereby raising the possibility that these boosted responses may prevent mother-to-child transmission of CMV [119]. (See "Society guideline links: Cytomegalovirus in solid organ transplant recipients".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: Avoiding infections in pregnancy (The Basics)" and "Patient education: Cytomegalovirus (The Basics)")

●Beyond the Basics topic (see "Patient education: Avoiding infections in pregnancy (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Prevalence in pregnancy – Approximately 2 percent of seronegative pregnant women will develop a CMV infection during pregnancy, but the risk can be reduced with use of preventive behavioral interventions. CMV infection may cause a mild maternal febrile illness and other nonspecific symptoms but is not apparent in 90 percent of women. (See 'Risk of seroconversion' above and 'Clinical findings' above and 'Behavioral risk reduction interventions' above.)

●Primary versus nonprimary infection – CMV infections in pregnant women are classified as primary if the initial acquisition of virus occurs during pregnancy, and nonprimary if maternal antibody to CMV was present before conception. Nonprimary infection, also sometimes called recurrent or secondary infection, may be due to reactivation of latent virus or reinfection with a new strain. (See 'Classification' above.)

●Risk of congenital infection – The timing of primary maternal infection is the most important determinant of fetal/newborn sequelae. The occurrence of fetal infection increases with advancing gestational age at the time of maternal seroconversion, but the occurrence of symptomatic newborns decreases with advancing gestational age and is unlikely near term. (See 'Frequency of maternal-fetal transmission' above.)

Preconception seroimmunity provides substantial protection against the occurrence of fetal infection compared with seroconversion during pregnancy but does not completely protect the fetus from infection (fetal infection: approximately 1 percent versus 40 percent) (algorithm 2). However, the vast majority of infants with congenital CMV infection are born to mothers with nonprimary CMV infection. (See 'Frequency of maternal-fetal transmission' above.)

Once fetal infection occurs, preexisting immunity provides limited protection against newborn disease and adverse outcome. (See 'Clinical features and sequelae' above.)

●Maternal screening and diagnosis

•We suggest not screening all pregnant women for CMV infection (Grade 2C). (See 'Screening' above.)

•Testing pregnant women for CMV is indicated as part of the diagnostic evaluation of women with mononucleosis-like illnesses or when a fetal anomaly consistent with congenital CMV infection is detected on prenatal ultrasound examination. (See 'Whom to test' above.)

•The gold standard for maternal diagnosis of clinically suspected primary CMV infection is seroconversion (table 1). In the absence of documented seroconversion, the presence of anti-CMV immunoglobulin G (IgG) and anti-CMV immunoglobulin M (IgM) may represent primary infection, reactivation, reinfection, or latent disease. In these cases, avidity testing is useful: High antibody avidity suggests infection occurred more than six months in the past, while low avidity suggests recent infection. (See 'How to test' above.)

●Fetal diagnosis and prognosis – Prenatal (fetal) diagnosis may be offered when fetal infection is suspected because of primary maternal infection or findings on ultrasound, given the risk of severe sequelae in offspring. Testing includes (1) amniocentesis ideally at least six to eight weeks after the presumed time of maternal infection to perform polymerase chain reaction (PCR) for CMV DNA and (2) ultrasound assessment of the infected fetus to detect stigmata suggestive of fetal sequelae (algorithm 1). (See 'Prenatal (fetal) diagnosis' above.)

Although fetal infection can be detected by PCR, fetal prognosis is difficult to predict. An abnormal ultrasound examination suggests a poor prognosis, while a normal ultrasound examination does not exclude the possibility of a symptomatic neonate or development of long-term neurologic morbidity. (See 'Prenatal prediction of offspring outcome' above.)

●Sequelae of congenital infection

•Congenital cytomegalovirus (CMV) infection is a leading nonhereditary cause of hearing loss and other neurodevelopmental disabilities. (See 'Clinical features and sequelae' above.)

•Congenital infection may be symptomatic or asymptomatic in newborns. Most congenital infections are asymptomatic in the neonatal period. (See 'Clinical features and sequelae' above.)

Both symptomatic and asymptomatic infected newborns are at risk for development of adverse sequelae in early childhood, but symptomatic newborns are at higher risk (eg, death 5 versus 0 percent, deafness 50 versus 10 percent). (See 'Symptomatic newborns' above and 'Asymptomatic newborns' above.)

●Maternal therapy – During pregnancy, neither antiviral therapy (oral valacyclovir) nor hyperimmune globulin has been effective for prevention of fetal disease or reduction in risk of sequelae. These experimental treatments are not recommended except in a research setting. (See 'CMV immune globulin' above and 'Antiviral drug therapy' above.)

●Prevention – Measures to prevent CMV infection during pregnancy are based on good personal hygiene and using CMV-negative blood products when transfusing seronegative pregnant women. (See 'Strategies for prevention of maternal and/or fetal infection' above.)