Fetal Infections

Virology

Cytomegalovirus is the largest member of the Herpetoviridae family. Its genome is made of a double-stranded DNA. Human CMV is highly species specific. Humans are its only reservoir. Several strains have been described based on genomic definition, possibly causing reinfections. Like other members of the Herpesvirus group, CMV remains latent in several locations after acute infection. Reactivation (recurrences) can therefore occur during latent infection.

The virus is acquired at mucosal sites (community exposure) or by bloodborne transmission (blood transfusion or organ transplant). In community exposure, cell-free virus is transmitted by contact with saliva, tears, urine, nasal or genital tract secretion. Cell-free virus is abundant in breast milk of infected women. Cell-mediated spread of the virus begins after a replication phase. The main cells infected by CMV are endothelial cells and polymorphic nuclear leukocytes (PMLs). Dissemination of the virus is then haematogenous. This viremic phase can be diagnosed by laboratory testing. The secondary sites of replication are the spleen and the liver. Dissemination and replication is not completely controlled by host immunity.

Epidemiology

Cytomegalovirus infection is endemic and shows no seasonal variation. It is spread worldwide.

Transmission of the virus occurs by direct or indirect person-to-person contact via urine; oropharyngeal, cervical and vaginal secretions; semen; milk; tears; blood products; or organ transplants. This is caused by prolonged shedding of the virus after primary infection (PI). CMV infection requires intimate contact.

The patterns of seropositivity in the population vary greatly with geographic, ethnic and socioeconomic conditions. The prevalence of specific antibodies to CMV increases with age and in lower socioeconomic strata of developed countries as well as in developing countries. Seroprevalence among women of childbearing age also varies accordingly with these epidemiologic factors. The seropositivity ranges from 50% to 85% in the United States and in Western Europe. The incidence of CMV PI in pregnancy also varies with socioeconomic conditions, from 1% to 6% per year.

Congenital infection is the result of transplacental transmission. In the United States, around 1% of all newborns are infected when screened at birth. Nevertheless, this rate varies greatly among geographic areas and with seroprevalence. The rate of transplacental transmission varies between 5% and 50% for PI and 2% and 3% for nonprimary infection (NPI).

Congenital infections are mainly caused by PI, but several reports show a possible fetal transmission after reinfection with another strain of the virus or after reactivation of a latent infection. Irrespective of the type of maternal infection, the rate of maternal-fetal transmission is considered to be lower (around 2%).

Maternal Infection

Clinical symptoms and nonspecific biological markers are more often present in PI than in recurrences. Most PI in immunocompetent hosts are nevertheless subclinical. Nigro and coworkers reported fever in 42.1% of PI and 17.1% of recurrences, asthenia (31.4% and 11.4%,), myalgia (21.5% and 6.7%), rhino-pharyngo-tracheo-bronchitis (42.1% and 29.5%) and flulike syndrome defined as the simultaneous occurrence of fever and at least one of these signs (24.5% and 9.5%), lymphocytosis of 40% or greater (39.2% and 5.7%) and increased aminotransferases blood levels (one or both >40 iu/L) (35.3% and 3.9%). The platelet count was significantly lower in PI but within normal range.

The diagnosis of PI can be easily confirmed by documenting seroconversion ( de novo appearance of virus-specific immunoglobulin (Ig) G in a pregnant woman who was seronegative before the onset of pregnancy). Nevertheless, without a screening program adopted by public health authorities where seronegative women would be prospectively monitored during their pregnancy, this remains a rare situation. More often, serologic testing is performed when contamination is suspected with maternal clinical symptoms or when ultrasound fetal abnormalities are visualised. In this context, the presence of IgG leads to assess type of CMV infection (past, primary or recurrent infection), and to date the time of maternal PI as precisely as possible. IgM and IgG avidity assays have been developed for these purposes.

IgM antibody response begins within days after maternal contamination, reaching a peak in the first month after maternal contamination. High to medium levels of IgM antibodies can therefore be detected during the first 1 to 3 months after the onset of infection, after which the titres start declining. Nevertheless, IgM detection can remain positive more than 1 year after PI.

IgG avidity is indicative of the low functional affinity of the recently produced IgG class antibody. Soon after PI, antibodies show a low avidity for the antigen, though increasing progressively. This characteristic is used to discriminate between recent and over 3 months old PI. From the publication of Macé and coworkers, an avidity index (AI) greater than 70% reflects PI for longer than 3 months, and AI below 30% is highly suggestive of a recent PI (<3 months). AI between 30% and 70% is more difficult to interpret. This method is only applicable when IgG levels are not too low. Nevertheless, one limitation of IgG AI testing is the lack of standardisation. The best-published results reported 100% negative predictive value (NPV) for a moderate to high IgG AI obtained before the 18th week of gestation and an NPV dropping to 91% when performed between 21 and 23 weeks’ gestation. Overall, Mace and coworkers reported that dating maternal PI using IgG AI associated with IgM antibody detection failed to date the onset of infection in only 1% of cases.

One of the most challenging situations is the presence of both IgM and IgG in the first trimester of pregnancy. Avidity and maternal viremia help sorting out 90% of these cases using an incremental risk algorithm for vertical transmission of between less than 1% and 40%. Leruez-Ville and coworkers reported a significant association between the risk for vertical transmission and the AI combined with CMV polymerase chain reaction (PCR) in maternal serum or IgG titres. In this series, a total of 4931 consecutive women were screened; 201 presented with positive or equivocal IgM and with high, intermediate or low IgG avidity in 58.7%, 18.9% and 22.3%, respectively. In 72 women with low or intermediate AI, fetal transmission was 23.6%. In multivariate analysis, positive CMV PCR in maternal serum, decreasing AI and low IgG titres were all associated with fetal transmission.

After PI, both the virus and viral products can be recovered from various fluids. However, viral shedding from these sites can occur after recurrences as well. It has also been shown that detection of CMV-DNA in blood is diagnostic of PI in immunocompetent individuals, but in immunocompromised patients, it is indicative of either PI or NPI.

Quantification of viremia (infectious CMV particles in blood, evaluated by culture or rapid shell vial method), antigenaemia (pp65-positive peripheral blood leukocytes), quantification of leukoDNAemia (CMV DNA in whole blood), leukocytes or plasma and more recently RNAemia (CMV mRNA) are available. Revello and coworkers reported the diagnostic ability of these methods in 52 immunocompetent individuals comprising 40 pregnant women. Antigenaemia was detected in 57.1%, 25% and 0% of the patients examined at 1, 2 and 3 months after onset of infection, respectively. Viremia was detected in 26.3% of cases during the first month only. LeuKoDNAemia, in 100%, 89.5% and 27.3% at each of the three first months, 26.6% were still positive at between 4 and 6 months, but none were positive after 6 months. At the same time, none of the patients with recurrent infection had positive test results. These results provide with a helpful method for dating maternal infection.

Congenital Infection

Approximately 10% of the congenitally infected newborns have signs and symptoms at birth. Half of them present the typical cytomegalic inclusion disease (CID) with a high mortality rate. The other half present with atypical or no symptoms. Of these, 90% are asymptomatic, although infected as shown by the presence of the virus in their urine during the first weeks of life.

Symptomatic infected newborns are defined as presenting at least one of these abnormalities: prematurity, hypotrophy, petechiae, jaundice, hepatosplenomegaly, purpura, neurological findings (microcephaly, hypotonia, seizures), elevated alanine aminotransferases levels, thrombocytopenia, conjugated hyperbilirubinemia, haemolysis and increased cerebrospinal fluid proteins. Long-term follow-up of these children enabled researchers to establish the occurrence of at least one sequela in 90% of the subjects in this subgroup. These complications were psychomotor delay (70%), sensorineural hearing loss (SNHL; 50%) and chorioretinitis (54%). The mortality rate consecutive to congenital CMV infection was estimated to be around 6%.

The best predictor for adverse neurodevelopmental outcome in these infants is the presence of intracranial abnormalities on computed tomography within the first month of life. These abnormalities were also associated with SNHL at birth or with deterioration of audiometric status during the first months of life.

Among asymptomatic neonates, who are known to have a better long-term prognosis than symptomatic ones, 10% to 15% will still develop sequelae, more often during the first 2 years of life. These sequelae include SNHL in 7%, chorioretinitis in 2%, intellectual deficit in 4% and microcephaly in 2%. SNHL is the most frequent deficit related to congenital CMV infection in asymptomatic neonates. Fowler and coworkers have reported that 50% of the audiometric deficits were bilateral; 50% worsened during the first years of life; and in 18% of them, the audiologic deficit was diagnosed on average only at 27 months. CMV congenital infection could be the cause of one third of cases of SNHL in childhood.

Public health authorities have not adopted screening programs in the majority of developed countries. Therefore abnormal ultrasound findings related to CMV congenital infection are more likely to be diagnosed during systematic ultrasound examination rather than during the follow-up of maternal seroconversion. This may also explain why severe abnormalities are described more often than subtle findings and more often after infection in the first trimester of pregnancy.

One should know the natural history of the infection to understand which ultrasound features are evocative and should lead to offer invasive prenatal diagnosis. It is also important to remember that the correlation between sonographic abnormal findings and evidence of maternal infection is made several weeks apart. In PI, around 25% to 50% of infected fetuses can be diagnosed by ultrasound examination. This is more likely to reflect the proportion of papers published on this particular aspect than the performance of ultrasound as a screening test. In the literature, description of fetal CMV infection ultrasound features are twofold: gross abnormalities leading to the diagnosis of fetal CMV infection and subtle findings discovered after thorough serial ultrasound examination of fetuses at high risk after vertical transmission of the virus has been shown. This at least partly explains that the performance of ultrasound as a screening test could not be demonstrated in a low-risk population. Ultrasound findings are summarised in Table 42.1 . Case reports or series of abnormalities mainly reported fetal ventriculomegaly or hydrocephalus, either obstructive (aqueductal stenosis) or a vacuo with or without microcephaly, posterior fossa cysts, cerebellar hypoplasia, severe intrauterine growth restriction or even hydrops fetalis. Cases diagnosed because of maternal seroconversion followed by serial fetal ultrasound examination often lead to diagnose more subtle findings, mainly extracerebral findings including hyperechogenic bowel and oligohydramnios.

Hyperechogenic bowel grade 2 is often a transient finding. However, in a series comprising of 175 fetuses with hyperechogenic bowel, only 1 case was related to CMV infection. It is the result of viral enterocolitis, and it can present as meconium ileus or peritonitis in conjunction with ascites.

Oligohydramnios is more often reported than polyhydramnios, and considering the affinity of the CMV for the kidney, it is the result of a fetal nephritis.

The fetal heart can also be affected, showing cardiomegaly with a thick myocardium, which may contain punctuate calcifications. As a functional consequence, Drose and coworkers have also described tachyarrhythmia. This is a rare finding that could participate in the development of fetal hydrops.

Generalised oedema and ascites may also suggest anaemia-related hydrops caused by the combined effect of liver failure and bone marrow infection. This striking presentation may eventually be transient with both ultrasound and biological normalisation at follow-up.

Mild or unilateral ventriculomegaly, increased pericerebral spaces, echogenic vessels in the thalami and basal ganglia or punctuate echogenicities in the periventricular area are subtle findings especially if they are isolated and are therefore likely to slip through nontargeted ultrasound screening. The development of fetal MRI has become an asset in the assessment of infected fetuses. MRI using both T1- and T2-weighted sequences could help define the onset of fetal infection. Lissencephaly may reflect injury before 16 or 18 weeks, but polymicrogyria is likely to follow injury at 18 to 24 weeks’ gestation, and cases with normal gyral patterns would have probably been injured during the third trimester, showing diffuse heterogeneity within the white matter.

Fetal infection is diagnosed when the virus or the viral DNA is found in the fetal compartment. CMV can be detected in the AF by conventional viral isolation, rapid culture or molecular assays. Virus isolation has a high specificity but has a lower sensitivity than PCR. In recent years, PCR has been established as a reliable technique in reference laboratories. Various PCR assays have been described with multiple modalities, including single step, nested, nested modified using higher volume or multiples aliquots, and the most recent real-time PCR. The efficacy of these methods has been evaluated in several studies and is dependent on the virologic method used: sensitivity and specificity range between 75% and 100% and between 67.3% and 100%, respectively.

The false-negative results reported were explained in most cases by inappropriate timing of amniocentesis. After seroconversion or reactivation, the process leading to CMV excretion in the fetal urine takes an average of 6 to 8 weeks, and this interval should be recognised to avoid false-negative prenatal diagnosis. Amniocentesis should also be performed after fetal urination is well established and therefore not before 21 weeks. When the conditions of sampling are ideal, the sensitivity of prenatal diagnosis by PCR has been reported to be close to 100%. Another explanation for false-negative results is a late transmission of the virus through the placenta. This occurs in 8% to 15% of cases, and infected newborns have a good prognosis.

False-positive PCR results have also been reported when the neonate was not infected; these false-positive diagnoses may be explained by contamination of the AF with maternal blood during amniocentesis if the mother had a positive CMV DNAemia result at the time of sampling. Indeed, Revello and coworkers showed that CMV DNA may be recovered in the blood of nearly 50% of immunocompetent patients up to 3 months after CMV PI. Another explanation could be laboratory contamination occurring during PCR testing. Indeed, in some of these studies, a nested CMV PCR was used, which is known to be a very sensitive technique but at high risk for contamination. Generalisation of semiautomated real time PCR might help to overcome the risk for contamination and achieve quasi-absolute specificity for prenatal diagnosis of CMV infection.

Management

Several antiviral drugs are active against CMV, and the three licensed anti-CMV drugs (ganciclovir, cidofovir and foscarnet) are being used successfully in immunocompromised patients. However, their potential teratogenic effects and their well-known toxicity do not support their use in pregnancy.

Anti-CMV compounds are currently at different stages of development; several of these compounds are very promising in term of efficacy and lack of toxicity. To date, preliminary results on treatment of CMV congenital infection during pregnancy are available from two studies with promising results.

The results of a first prospective nonrandomised trial using intravenous CMV hyperimmune globulin (HIG) for CMV maternal PI were published in 2005. Nigro and coworkers selected 181 pregnant women with primary CMV infection. Two study groups were formed. In the first group, called the ‘therapeutic group’, comprising 79 patients, amniocentesis was performed. PCR results were positive in amniotic fluid (AF) in 55. Immunoglobulin therapy was given in 31 of them, a medical termination of pregnancy was performed in 10 patients and no treatment was given in 14. The proportion of symptomatic children at birth was significantly lower in the group in patients who received treatment (1 of 31 vs 7 of 14; OR, 0.02; P <.001). In the ‘prevention’ group, 102 women were included. An HIG treatment was initiated in 37 of them. No treatment was taken in the 65 remaining patients. In the latter group, the pregnancy was terminated in 18 cases. The vertical transmission rate was significantly lower in the subgroup of patients who received treatment (6 of 37 vs 19/47; OR, 0.32; P = .04). The results of this study suggested that the immunoglobulin treatment was effective in reducing the proportion of symptomatic children at birth in case of confirmed fetal infection and could reduce vertical transmission in case of maternal infection from 40% to 16%.

Following these encouraging results, a randomised trial against placebo was conducted by an Italian team. The results of the CHIP trial were published in 2014. In this study involving 123 women with PI, 61 received treatment with immunoglobulin, and 63 were treated with placebo. The congenital infection rates were 30% in the treated group and 44% in the placebo group ( P = .13). Besides this negative result, the authors also observed an adverse event rate of 13% versus 2%, including premature births, cholestasis of pregnancy to intrauterine growth retardation and one case of eclampsia. To date, immunoglobulins have failed to change the prognosis of maternal CMV infection, both in terms of vertical transmission and severity of fetal infection.

Jacquemard and coworkers have shown the pharmacologic efficacy of valaciclovir in a pilot study in 21 cases of CMV congenital infections with ultrasound abnormalities. These results have encouraged the same French team to lead a first randomised double-blind study against placebo and methodology modified to be interventional without randomisation with exclusive use of valaciclovir. The results of this study titled CYMEVAL2 were published in 2016. In this study, among the 43 fetuses infected with CMV and with moderate ultrasound abnormalities or abnormal fetal blood parameters, the proportion of asymptomatic children was 34 of 43 (above the threshold of 31, children required to consider treatment as effective). Comparing these results with data from literature, valacyclovir increased the proportion of asymptomatic neonates from 43% without treatment to 82% with treatment. Despite the lack of randomisation, these results allow us to consider valaciclovir as a control to subsequently test different antivirals.

Prevention

No vaccine is currently available in the routine practice. Nevertheless, several studies have shown encouraging results, mainly with gB recombining vaccines.

Furthermore, Picone and coworkers have shown that if clear information on CMV infection during pregnancy is given, patients frequently agree to screening and this counselling can result in a decreasing rate of seroconversion after information.

Virology

The taxonomy of the Parvoviridae ’s family includes the Densovirinae (insect viruses) and the Parvovirinae (vertebral viruses), which is composed of three genders (dependovirus, parvovirus and erythrovirus). PVB19 is a nonenveloped single-stranded DNA virus classified as an erythrovirus. It is the only parvovirus that can cause human disease.

The primary target for PVB19 appears to be erythroid precursor cells. Host cell receptor is the globoside or P-antigen (a blood group antigen), a glycosphingolipid; therefore, patients without this antigen on their red blood cells are naturally protected against PVB19 infection. Globoside is situated on the surface of erythrocyte progenitor cells (erythroblasts) but also on that of other cells (endothelial, myocardial and placental cells, as well as mature erythrocytes and megakaryocytes). Inside the host cells, PVB19 replicates and induces apoptosis and toxic cell injury.

Epidemiology

PVB19 infection occurs worldwide, and the characteristics of the disease are constant. Neither antigenic nor specific viral genotypes are related to the forms of the disease. Transmission of the virus continues throughout the year with winter and spring outbreaks.

Seroprevalence of PVB19 increases with age. In children younger than 5 years of age, the prevalence of IgG antibodies is less than 5%, increasing to reach a median of 45% in young adults and more than 85% in the geriatric population. Some studies have shown a greater risk for PVB19 infection in women. Seroprevalence is higher in white populations.

The global incidence of PVB19 infection has been reported to be 1 to 2 per 10,000 individuals. In women of childbearing age, risk factors for seroconversion are elementary school workers, contact with 5- to 11-year-old children at home or 5- to 18-year-old children at work, and women younger than 30 years of age. Overall, it can be estimated that 1% to 2% of seronegative women at the onset of pregnancy would become infected during pregnancy in endemic periods and more than 10% in epidemic periods. Risk factors of PVB19 in pregnant women are exposure to PVB19 in their own children, elementary school teachers and daycare workers.

Infection with PVB19 usually occurs through contact with respiratory droplets, but PVB19 can also be transmitted by blood and blood-derived products and can be transmitted vertically from mother to fetus. Presence of IgG antibodies gives a lifelong protection against reinfection with PVB19.

Maternal Infection

Approximately 20% of infected immunocompetent individuals are asymptomatic. Erythema infectiosum (fifth disease) is the most common clinical manifestation during childhood. It is characterised by a rash consisting of maculae that undergo central fading over in 1 to 4 days, mainly on the trunk and limbs. Symptoms such as erythema infectiosum, mild fever, arthralgia and headache start approximately 10 to 14 days after contamination and in about 50% of infected women. Arthralgia and arthritis are common in the adult form of the infections. It affects females more often than males (60% vs 30%) and children (10%). Others possible symptoms include thrombocytopenia, meningoencephalitis, hepatitis, myocarditis and vasculitis. Immunocompromised hosts can also present transient aplastic crisis, chronic red blood cell aplasia and virus-associated haematophagocytic syndrome. Symptoms reported by pregnant women are nonspecific, and serologic confirmation is required. The most characteristic symptom is symmetrical arthralgia, sometimes arthritis often involving small joints of the hands, wrists and feet. The proportion of asymptomatic women is around 30%.

Specific IgM, IgG and IgA immunoglobulins are produced after maternal infection. Specific IgM are the first antibodies to rise around 10 days post infection. They sharply peak at 10 to 14 days. The IgM response persists for 1 month to several months.

Specific IgG rises considerably more slowly about 3 weeks postinfection to reach a plateau at around 4 weeks after infection. These antibodies probably last for life.

Maternal serum should be tested when there is evidence of maternal exposure or clinical symptoms of PVB19 infection. A booking sample with positive IgG but negative IgM indicates previous maternal infection. When the sample is negative for both IgM and IgG, absence of maternal infection can be confirmed. The presence of IgM enables to conclude that a recent maternal infection has occurred regardless of IgG antibody levels. The absence of IgG associated with the presence of IgM is indicative of a recent infection before IgG can be detected. Nevertheless, it is important to remember that after a recent contact, there will be a window of 7 days, during which both IgG and IgM will remain negative. Furthermore, at the time of clinically overt hydrops fetalis, IgM levels may sometimes already have become undetectable. In these cases, PCR analysis of the same blood sample can be informative. In a series of 41 fetuses with PVB19 infection and anaemia, at the time of fetal blood sampling, it has been reported that all mothers were PVB19-DNA positive, PVB19-IgG positive and B19-IgM were detected in 95% of cases.

Serologic testing should assess pregnant women who have been exposed to PVB19 infections and those developing symptoms compatible with this infection. Seronegative women should be retested 2 weeks later. They can be reassured in the absence of seroconversion. In cases with PI, serial ultrasound examination including middle cerebral artery peak systolic velocity (MCA-PSV) measurements should be performed every fortnight up until 12 weeks after exposure).

Fetal Infection

Vertical transmission occurs in approximately one third of the cases of maternal infection (17%–33%).

Fetal infection with PVB19 is associated with intrauterine fetal death (IUFD), nonimmune hydrops fetalis (NIHF) and less often brain anomalies. Fetal infection can also be asymptomatic. Fetal manifestations caused by PVB19 are summarised in Table 42.3 .

The consequences of fetal infection are not univocal throughout the pregnancy. In the first trimester, it is controversial whether PVB19 maternal infection increases the rate of miscarriage. NIHF develops after maternal infection mainly in the first half of the pregnancy. The proportion of all NIHF caused by PVB19 infection is around 10% to 15%. In a series of 50 cases of NIHF, 4 were related to PVB19 infection. Its rate of occurrence is 3.9% after maternal infection throughout pregnancy, with a maximum of 7.1% when infection occurs between 13 and 20 weeks’ gestation. The incidence of PVB19 infection associated NIHF peaks between 17 and 24 weeks’ gestation. The interval between maternal PVB19 infection and the development of NIHF ranges from 2 to 6 weeks.

Cases of IUFD have been described mainly at around 20 to 24 weeks’ gestation but as early as 10 weeks and as late as 41 weeks. Furthermore, IUFD could occur without NIHF. NIHF is mainly related to severe anaemia, which can also lead to high-output cardiogenic heart failure. NIHF is more frequent during the hepatic stage (8–20 weeks of gestation) of the haematopoietic activity when the half-life of erythrocytes is shorter than later during the bone marrow and splenic haematopoietic stages.

Among 63 cases of laboratory-confirmed maternal PVB19 infection, Puccetti and coworkers reported a vertical transmission rate of 31.7% (20 of 63). Of the 20 infected, 8 had NIHF, 1 had signs suggestive of meconium peritonitis and 1 had an isolated hydrothorax. Three of eight fetuses presenting with hydrops were treated with intrauterine blood transfusion. Two of them died, and the last showed resolution of anaemia. Among the five untreated hydropic fetuses, one resolved spontaneously, two died and two had also cardiomegaly and elective termination of pregnancy was performed. All the anaemic fetuses had MCA-PSV values more than 1.8 MoM. No stillbirth occurred. Overall, in this series, the outcome of uncomplicated cases with PVB19 infection was good, but in the presence of NIHF, the prognosis was very poor.

In 539 cases reported by Soothill and coworkers, 30% preceded IUFD, 34% resolved spontaneously, 29% resolved after intrauterine transfusion and 6% died despite intrauterine transfusion.

In 20 cases of fetal PVB19 infection complicated by anaemia or NIHF, Mace and coworkers reported that the survival rates were 70% (14 of 20) and 76% (13 of 17) for fetuses with one or more transfusions. When fetal effusion regressed after the transfusion, all 11 fetuses survived, and neonatal outcome was favourable for all.

Nonimmune hydrops fetalis is diagnosed by ultrasound examination with the association of marked ascites, cardiomegaly and pericardial effusion and, in advanced stages, generalised oedema and a thick hydropic placenta. Hydrops related to anaemia usually presents with tense ascites, as well as thin-wall cardiomegaly. Pleural effusion is a late finding in anaemia-related hydrops and can be sometimes observed as isolated finding.

Polyhydramnios is rare with PVB19 infections. Pasquini and coworkers evaluated the rate of women with polyhydramnios who were screened positive to infectious disease by serum screening testing for TORCH (toxoplasmosis, other [syphilis, varicella-zoster, parvovirus B19] rubella, CMV and herpes infections and viral serologies) and PVB19. In this series of 290 cases, only two women were positive for parvovirus B19 and one for toxoplasmosis infection. In none of them the fetus was affected. Authors concluded that infectious disease screening does not seem beneficial in pregnancies with isolated polyhydramnios.

Pasquini and coworkers reviewed 141 cases of mild ventriculomegaly. Screening for infections, including TORCH, PVB19 and syphilis, was carried out in all cases. Maternal IgM for PVB19 resulted positive in 4.6% of cases, and one neonate was infected without any fetal or neonatal adverse consequence. Recent CMV infection was documented in 4.4% of cases. Only in one case was the infection transmitted to the fetus.

Isolated hyperechogenic bowel has been reported as the sole ultrasonographic sign of fetal B19 infection. It is likely to reflect resolving ascites.

The involvement of the fetal heart can be limited to dilation of the cardiac cavities related to anaemia and hydrops or present as a hypertrophic cardiomyopathy or myocarditis that can also develop autonomously after spontaneous resolution of the NIHF.

Fetal PVB19 infection has also been associated with paediatric stroke, neonatal encephalitis or meningitis with perivascular calcifications in the fetal cerebral cortex, basal ganglia, thalamus and germinal layers, meconium peritonitis, fetal liver calcifications, eye abnormalities such as cornea opacification and aphakic eyes or rare bones lesions.

Structural defects (cleft lip and palate, micrognathia, arthrogryposis, hypospadias) reported are likely to be coincidental findings. During a large community-wide outbreak of PVB19 infection, there was no increase in congenital malformations rate compared with the periods before and after the epidemic. This may lead to the conclusion that the association of birth defects could be fortuitous or the consequence of anaemia-related hypoxia or thrombocytopenia-related haemorrhage.

Neurological impairment can be observed related to severe and prolonged fetal anaemia accompanied by thrombocytopenia that can lead to intraventricular brain haemorrhage. Cerebellar haemorrhage, polymicrogyria, infarctions, calcifications and obstructive hydrocephalus have also been reported in association with PVB19 infection during the first half of pregnancy.

Rarely, maternal symptoms of ‘mirror hydrops’ occur secondary to lysis of the hydropic villi of the placenta. Maternal mirror syndrome is a maternal preeclampsia-like syndrome with oedema, hypertension, proteinuria and anaemia and seems to reflect the intensity and persistence of the fetal anaemia.

After confirmation of maternal infection with a suggestive serologic profile, fetal ultrasound examination should be performed to exclude the presence of fetal anaemia and hydrops. However, in most cases, NIHF is a coincidental finding during routine ultrasound examination. Thus fetal anaemia should be suspected when the MCA-PSV Doppler is increased. These changes in blood flow are the result of increased cardiac output and decreased viscosity of fetal blood. The prediction of fetal anaemia by MCA-PSV measurements also allows evaluating the severity of anaemia. Elevated values indicate the need for fetal blood sampling and intrauterine transfusion. Severe NIHF together with normal MCA-PSV in parvovirus B19 infection indicates either spontaneous resolution of fetal anaemia or progressive and autonomous myocarditis.

Increased nuchal translucency and reversed a-wave ductus venosus Doppler in the first trimester have also been reported to be associated with B19 infection as early signs of cardiac failure.

The virus or its genome can be retrieved in AF by PCR with a high sensitivity. This can also be applied in pregnant women lacking an adequate antibody-mediated immune response or who are immunocompromised or immunosuppressed in whom serologic testing for PVB19 is unreliable. Detection of PVB19-specific IgM in fetal blood has a significantly lower sensitivity than PCR. Otherwise, B19 PCR in the AF is part of the etiologic diagnosis when NIHF is observed, as well as CMV PCR and karyotype.

Management

Management of PVB19 infection with fetal transfusion can correct anaemia and is likely to significantly reduce perinatal mortality rate. It should be restricted to cases with increased MCA-PSV. IUT has been reported to be effective as early as 13 weeks of gestation. Timely IUT of anaemic fetuses with severe hydrops reduces the risk for fetal death. In most cases, one transfusion is sufficient for fetal recovery, and high reticulocyte levels indicate that anaemia is being corrected spontaneously. Hydropic changes can take up to several weeks to resolve, and MCA-PSV should be used to evaluate the correction of fetal anaemia.

In a consecutive series of 27 cases in a 15-year period, Chauvet and coworkers reported that among the 19 fetuses treated by transfusions, 11 were liveborn compared with 2 of the 6 not treated (57.8 vs 33.3%, not significant). The survival rate was higher during the second half of the study period (23.1 vs 71.4%; P = .02) and for less severe anaemia ( P = .03). In this series, all 13 liveborn children appeared healthy at the age of 1 year.

Overall, children who survived a successful IUT in this situation have a good neurodevelopmental prognosis. Nevertheless, the small number of cases limits the conclusions of these series.

Prevention

Exclusion of pregnant women from the workplace during endemic periods is not recommended. If pregnant women are exposed to individuals who are suspected or known to be infected with PVB19, they should report this exposure to their obstetrician. Serologic testing is required.

Ballou and coworkers described a recombinant parvovirus B19 vaccine composed of VP1 and VP2 capsid proteins, which proved to be immunogenic and safe to use in human volunteers. Vaccination of nonimmune pregnant women could be a highly effective method to prevent fetal infection with parvovirus B19, but the cost-effectiveness of this strategy in the general population remains to be determined.