Fetal Infection.

Clinically significant infection with CMV occurs during intrauterine life by transplacental passage of the virus. The organism is transmitted to the fetus usually during a primary maternal infection (less commonly during recurrent infection) with viremia and subsequent placentitis. Recent data, however, suggest that maternal immunity before pregnancy cannot be viewed as protective in terms of altering long-term outcome and that the outcome of infants infected following primary and nonprimary infections is remarkably similar. The maternal infection is usually asymptomatic but may be manifested by a mononucleosis-like illness (≈10%) or a more serious systemic illness. Maternal infection is very common; cytomegaloviruria occurs in 3% to 6% of unselected pregnant women. Cervical CMV infection is several times more common than cytomegaloviruria but tends to occur late in pregnancy and is probably less likely to result in significant fetal infection. Clinically significant fetal CMV infection probably occurs principally in the first or second trimesters, particularly if CNS disease is the outcome measure. The possibility of CNS involvement after CMV infection in the third trimester was suggested by a study of seven children but the exact timing of the fetal infection was established in only one case (at 27 weeks of gestation). Moreover, the nature of the neuropathological features in some infected infants is also consistent with CNS involvement secondary to infection relatively late in pregnancy (see later discussion).

Two recent studies evaluated intrauterine transmission rates following primary CMV infection in the pre- and periconceptional period, first, second, or third trimester in 248 and 238 pregnancies respectively ( Table 34.2 ). The overall transmission rates were very similar in both studies and significantly increased with the trimester of pregnancy, with the highest transmission rate in the third trimester. There was a significantly higher risk of cranial ultrasound abnormalities when maternal infection occurred during the preconceptional or periconceptional periods and the first trimester, compared with risk with infection acquired in later trimesters. No symptomatic neonatal infection was noted when maternal infection occurred after 14 weeks of gestation. Hearing loss developed in 5% to 10% of asymptomatic infants.

Approximately 30% to 40% of infants whose mothers experience primary infection during pregnancy develop congenital infection. Cytomegaloviruria has been observed in approximately 0.5% to 2% of infants in the neonatal period. Because a period of approximately 4 to 8 weeks is required between the time of infection and the viruria, these neonatal examples reflect intrauterine infection and not perinatal acquisition from parturitional or postnatal exposure. In these cases of congenital CMV infection, involvement of the CNS may be overt in the neonatal period or may not become apparent for months or years thereafter (see later discussion). In a prospective series of more than 117,986 infants screened, the overall CMV birth prevalence estimate was 0.7%. The percentage of infected children with CMV-specific symptoms at birth was 12.7%. The percentage of symptomatic children with permanent sequelae was 40% to 58%. The percentage of children without symptoms at birth who developed permanent sequelae was estimated to be 13.5%. The true burden of congenital CMV infection is unclear because data on important outcomes, such as visual impairment, are lacking and follow-up of infected children has been too short to fully identify late-onset sequelae. Clinically apparent congenital infection following recurrent (nonprimary) maternal infection (i.e., infection in women with preexisting seroimmunity) is no longer considered a rare event because of the large prevalence of latent maternal CMV infection among women of childbearing age and the failure of maternal antibodies to prevent transmission during pregnancy. This phenomenon of intrauterine transmission in the presence of substantial maternal immunity has been attributed to reactivation of endogenous virus in some cases and to reinfection with different strains of CMV in other instances. A higher prevalence was found in developing countries than for Europe and North America owing to the higher maternal CMV seroprevalence. According to data derived from 11 studies from Africa, Asia, and Latin America and involving numbers of newborns tested ranging from 317 to 12,195, maternal CMV seroprevalence ranged from 84% to 100%. CMV birth prevalence varied from 0.6% to 6.1%.

Parturitional and Postnatal Infections.

Parturitional and postnatal exposures cause an additional 10% to 15% of infants to acquire CMV infection in the first 4 to 8 weeks of life. Clinical signs and symptoms of pCMV infection in very or extremely preterm infants include pneumonia, enteritis, cholestasis, hepatosplenomegaly, sepsis-like syndrome (SLS), thrombocytopenia, and neutropenia. Postnatal CMV infection has also been associated with an increased risk for bronchopulmonary dysplasia (BPD). In one study, 42% of infected infants developed clinical or laboratory abnormalities (neutropenia and thrombocytopenia). Another study showed a mean incubation time of 42 days (95% confidence interval [CI] 28 to 69) with symptoms in about 50% of the infected infants and 4 of 33 with sepsis-like symptoms. In a study from the Netherlands, the majority of CMV-infected infants (85%) did not develop any symptoms of pCMV infection. The most important independent risk factors for pCMV infection were nonnative Dutch maternal origin (OR 9.6 [95% confidence interval (CI) 4.3 to 21.5]) and breast milk (OR 13.2 [95% CI 1.7 to 104.5]). The risk of pCMV infection significantly decreased for each additional week of gestational age (GA) (OR 0.7 [95% CI 0.5 to 0.9]). Lenticulostriate vasculopathy (LSV) was significantly more often present at term-equivalent age in infants with pCMV infection (OR 4.1 [95% CI 1.9 to 8.8]). Breast milk is probably the single most important source of CMV exposure in premature infants. It has been documented that 96% of CMV-seropositive women have CMV reactivation with shedding of virus or the presence of CMV DNA in breast milk within several days after delivery. In a meta-analysis, among 299 infants fed untreated breast milk, 19% (11% to 32%) acquired pCMV infection and 4% (2% to 7%) developed pCMV-SLS. Among 212 infants fed frozen breast milk, 13% (7% to 24%) developed CMV infection and 5% (2% to 12%) an SLS, yielding slightly lower rates of breast milk–acquired CMV infection (4.4%; 2.4% to 8.2%) but similar rates of CMV-SLS (1.7%; 0.7% to 4.1%). The benefits of breast milk are still considered to outweigh the risks of severe disease from breast milk–acquired CMV infection in the neonatal period, which has so far not been associated with delayed development, SNHL, or clear cognitive impairment.

Although the results of one study raised the possibility of an increased risk of neurological sequelae in premature infants who acquire pCMV infection during the first 8 weeks of life, most data have indicated that CNS involvement does not occur with parturitional or early postnatal infection. However, recent long-term outcome studies raise some concern. Although no effect of pCMV infection was obvious at assessments between 2 and 4.5 years of age, differences in outcome were noted when 42 children were assessed again during adolescence ( n = 42, 11.6 to 16.2 years, mean = 13.9; 15 girls; 19 with and 23 without an early pCMV infection). Assessed with the German version of the Wechsler Intelligence Scale and the Developmental Test for Visual Perception, adolescents born preterm with early pCMV infection scored significantly lower than those without this infection regarding overall cognitive abilities (92.67 [14.71] vs. 102.75 [13.67], P = .03) but not visuoperceptive abilities (91.22 [10.88] vs. 98.96 [13.45], P > .05). However, the group of children was small, and the data were not adjusted for known independent risk factors such as postnatal corticosteroids, duration of mechanical ventilation, sepsis, other congenital infections, necrotizing enterocolitis, surgery, or socioeconomic status. The last of these is predictive of poor cognitive outcome and could be an important confounder. The reasons for the difference in propensity to affect the CNS between early prenatal versus natal or postnatal acquisition of CMV remain to be determined.

Blood transfusion has been a particularly important source in low-birth-weight infants, but with the introduction of transfusion of CMV-seronegative and leukoreduced blood products, transmission of CMV to very-low-birth-weight infants has been effectively prevented.

Meningoencephalitis.

Meningoencephalitis is characterized by the following features: (1) inflammatory cells in the meninges; (2) perivascular infiltrates with inflammatory cells; (3) necrosis of brain parenchyma, with all cellular elements affected, especially in the periventricular region, and often associated with calcification; (4) reactive microglial and astroglial proliferation; and (5) the occurrence of enlarged cells (neuronal and glial elements) with intranuclear inclusions ( Fig. 34.1 ). Electron microscopic studies have revealed virions in brain tissue, a finding attesting to primary infection of the CNS by the organism. Recovery of virus from brain confirmed this conclusion. A role for the inflammatory response itself in causing tissue destruction is suggested by the disparity between the detection of cytomegaloviral DNA by in situ hybridization and the extent of the tissue necrosis.

Congenital cytomegalovirus infection: encephalitis, from an infant who died at 10 weeks of age.

(A) Region of necrosis in the subependymal germinal matrix (ependyma is at the upper right). Note enlargement of cells, many with intranuclear inclusions. (B) Higher-power view of the same region showing the enlarged cells with prominent intranuclear inclusions. (C) A closer view of the inclusions; note the clear halo around the inclusions and the nuclear chromatin displaced to the periphery (Cowdry type A inclusions).

(From Bell WE, McCormick WF. Neurologic Infections in Children . Philadelphia: Saunders; 1975.)

The cellular and regional targets of the meningoencephalitis include especially the germinative ventricular–subventricular zones, radial glial cells, cerebral white matter, subplate neurons, and cerebral cortex . Germinal matrix necrosis and cysts are prominent. Subsequent matrix calcification is important in determining the periventricular distribution of calcifications . The particular tropism for progenitor cells in the ventricular-subventricular zones, radial glial cells, and subplate neurons has recently been identified. Periventricular cerebral white matter is also a site of injury, sometimes with cyst formation and subsequent calcification. A predilection for the parietal white matter may mimic periventricular leukomalacia (PVL). In addition, a predilection for anterior temporal white matter is particularly suggestive of CMV infection. Cerebrocortical atrophy is a later feature.

Microcephaly.

Microcephaly is a common feature in the neonatal period and is still more prominent later in infancy. The small size of the brain appears to relate to the encephaloclastic effects of the virus and probably also to a disturbance of cell proliferation in the developing brain. The latter disturbance relates to the propensity of the virus to affect progenitor cells in the ventricular and subventricular zones (see earlier). A predilection for involvement of proliferative cells is also suggested by the frequency of intrauterine growth retardation in congenital CMV infection and by the observation in a variety of tissues of a decrease in absolute number of cells.

Disturbances of Neuronal Migration.

Disturbances of neuronal migration have been described repeatedly in congenital CMV infection. ^a

a References .

Congenital cytomegalovirus infection: neuronal migrational disturbance.

(A) Microgyric cerebellar cortex from a preterm infant with congenital cytomegalovirus infection. (B) Section of the same microgyric cerebellar cortex shown in (A), illustrating the distribution of calcification within the malformed cerebellar cortex and suggesting the coexistence of destructive and teratogenic effects.

(From Perlman JM, Argyle C. Lethal cytomegalovirus infection in preterm infants: clinical, radiological, and neuropathological findings. Ann Neurol .1992;31:64-68.)

Cerebellar Hypoplasia.

Cerebellar hypoplasia , best detected by magnetic resonance imaging (MRI), is a feature in at least 50% of symptomatic cases. This finding likely is primarily a proliferative disturbance, although, as noted earlier, migrational disturbances may also be seen in the cerebellum. The finding of cerebellar hypoplasia in the clinical setting of an intrauterine infection is highly suggestive of congenital CMV.

Other Findings.

Porencephaly, hydranencephaly, hydrocephalus, focal subcortical cysts, impaired myelination, and more diffuse cerebral calcifications have also been described to variable extents in congenital CMV infection.

Incidence of Clinically Apparent Infection.

Although congenital CMV infection occurs frequently, its clinical manifestations do not. Indeed, available data indicate that approximately 90% of affected infants are asymptomatic in the newborn period.

Clinical Features.

The most frequent clinical features of symptomatic congenital CMV infection are shown in Table 34.5 . The most common findings relate to disturbance of the reticuloendothelial system. Hepatosplenomegaly and a petechial rash, usually related to thrombocytopenia, are encountered very frequently. Infants are often small for GA; moreover, approximately one third of affected infants have a GA of 37 weeks or less. Inguinal hernia is a helpful clinical sign when present (≈25% of cases).

The neurological syndrome is variable in presentation. Seizures may be prominent, although only approximately 10% of symptomatic patients exhibit overt neonatal seizures. Microcephaly is a consistent manifestation in patients with severe disease and appears in approximately 50% of all symptomatic patients. Cerebral calcification, usually periventricular in location, occurs in 50% to 60% of cases. Germinolytic cysts and LSV are also suggestive of CMV infection. Cerebrospinal fluid (CSF) findings of encephalitis (e.g., pleocytosis, elevated protein content) are found in the majority of patients, but precise data are not available. In a recent study, CSF β2-microglobulin levels were increased and of prognostic value for neurodevelopmental outcome.

The clinical course is most commonly that of a static process. Rare evidence of progressive encephaloclastic disease, documented by computed tomography (CT) scan, was provided by a report of two such cases. Similarly, postnatal evolution of cerebral calcification and of subependymal necrosis has been documented. Progression of hearing loss during infancy and early childhood has been clearly described (see later). ^a

a References .

Clinical Diagnosis.

The diagnosis of congenital CMV infection may be suspected with a high degree of accuracy on the basis of certain clinical features. These features include the periventricular locus of the cerebral calcification, the presence of germinolytic cysts, and LSV—best seen with ultrasonography —microcephaly, CSF pleocytosis, and intrauterine growth retardation. Cerebellar hypoplasia and neuronal migrational abnormalities are also distinctive features and are best recognized with MRI. The absence of the “salt and pepper” chorioretinitis of congenital rubella and the relative infrequency of the grossly scarring chorioretinitis of congenital toxoplasmosis are also helpful.

Diagnosis During Pregnancy.

Most CMV infections encountered during pregnancy are asymptomatic. In about 5% there is a history of a flu-like episode. Only CMV serology (negative for immunoglobulin M [IgM] and immunoglobulin G [IgG]) can exclude a congenital CMV infection. Both primary as well as nonprimary CMV infections (reinfection/reactivation) can lead to congenital CMV infection. The diagnosis of a primary CMV infection can be made by the detection of seroconversion. In most countries women are not routinely screened for CMV antibodies before their pregnancy. The presence of anti-CMV IgM antibodies is considered to be a good indicator of an acute or recent CMV infection, but IgM antibodies are present in only 70% of infected babies and in only fewer than 10% of IgM-positive women is the fetus infected. When the infection takes place before conception or very early in the pregnancy, IgM may have become negative by the time the suspicion of congenital CMV infection is raised. Pregnant women can also produce IgM during reactivation or reinfection, and false-positive results are not uncommon because IgM may be found in mothers who have another viral illness, such as Parvovirus B19 or Epstein-Barr virus. A relationship between low total IgM values and clinical symptoms in newborns with congenital CMV infection has also been reported. This relationship can be explained by the longer period since the occurrence of the CMV infection early in pregnancy, resulting in lower total IgM in symptomatic cases. Severe fetal sequelae were also reported in six fetuses despite maternal immunity for CMV, confirmed by the detection of IgG with no IgM in previous pregnancies or early in the current pregnancy. Cranial ultrasound showed ventriculomegaly, calcification, and LSV, and amniocentesis confirmed the presence of CMV by polymerase chain reaction (PCR) in all six.

The anti-CMV IgG avidity test is the most reliable procedure to identify primary infection in pregnant women. The avidity indices may vary with the tests used. Low avidity indices indicate low-avidity IgG antibodies in serum caused by acute or recent primary CMV infection, whereas high avidity indices (high-avidity serum IgG) indicate no current or recent primary infection. The determination of anti-CMV IgG avidity, performed before the 16th to 18th week of pregnancy, identifies all women who will have an infected fetus/newborn (sensitivity 100%). After 20 weeks’ gestation, sensitivity is drastically reduced (62.5%). A high avidity index during the first 12 to 16 weeks’ gestation can be considered a good indicator of past infection. The presence of true IgM combined with a low/moderate avidity index has the same diagnostic value as seroconversion. Although not quite as powerful as a high-avidity result, an intermediate-avidity result during the first trimester also indicates a low risk of intrauterine transmission. In contrast, an intermediate-avidity or high-avidity result during the second or third trimester does not rule out postconception primary infection and is associated with an increased risk of transmission.

A reliable prenatal diagnosis is obtained by performing a PCR on the amniotic fluid. Amniocentesis is best performed between the 21st and 22nd weeks of gestation. CMV is a slowly replicating virus, and 6 to 9 weeks are required after maternal infection for the virus to be eliminated in the fetal urine in amounts sufficient to be detected in the amniotic fluid. There is a risk of a false-negative test, as can occur when the amniocentesis is carried out earlier, when little amounts of virus have been shed by the fetal kidney. The sensitivity and specificity (90% to 98% and 92% to 98%, respectively) for PCR analysis in the amniotic fluid are high with respect to viral transmission from mother to fetus. The risk of a severe infection with a high risk of severe sequelae occurs when the infection is contracted in the first 12 to 16 weeks of gestation.

Diagnosis of Cytomegalovirus Infection in the Newborn

Diagnostic Studies.

CSF characteristically exhibits the findings of meningoencephalitis. In a study of 18 infants with neurological manifestations, the mean white blood cell (WBC) count was 42 cells/μL including predominantly lymphocytes, and the mean protein content was 192 mg/dL. In a later series of nine newborns with neurological involvement, all had elevated WBC counts and protein levels. In another study of 56 infants (which included 30% with no CT abnormalities), CSF protein exceeded 120 mg/dL in 50%.

Skull radiographs were formerly used to demonstrate the periventricular calcifications ( Fig. 34.3 ). CT scanning is more sensitive than skull radiography for detection of calcifications ( Fig. 34.4 ). In a series of 41 infants with symptomatic congenital CMV infection, a CT-detected abnormality was present in 78%. Of those with abnormalities, periventricular calcifications occurred in 75%, varying degrees of cortical and white matter abnormalities were seen in 30%, and ventriculomegaly was reported in 40%. CT scanning was recommended in the past as the gold standard to assess cerebral involvement in infants with cCMV infection, but CT scanning is no longer recommended in such infants. Cranial ultrasound and MRI are safer, reliable alternatives.

Congenital cytomegalovirus infection: periventricular calcification.

This skull radiograph is from an affected 5-day-old infant with microcephaly.

(From Bell WE, McCormick WF. Neurologic Infections in Children . 2nd ed. Philadelphia: Saunders; 1981.)

Congenital cytomegalovirus infection: computed tomography.

These scans are from a 5-day-old infant with congenital cytomegalovirus infection. (A) Periventricular and diffuse cerebral calcifications and ventriculomegaly are apparent. (B) In addition to calcifications and ventriculomegaly, note the cerebellar hypoplasia and large cisterna magna (arrows) .

Cranial ultrasound scans frequently demonstrate abnormalities. These findings consist of periventricular cysts—especially in the region of the subependymal germinal matrix ( Figs. 34.5 and 34.6 )—ventriculomegaly, periventricular (and more diffuse) calcifications, and periventricular echolucencies (consistent with cerebral white matter cysts) (see Fig. 34.6 ). The correlations with neuropathological findings (see previous discussion) are obvious. An additional ultrasonographic finding, overt in approximately one third of cases, is the presence of branched echodensities in basal ganglia and thalamus (see Fig. 34.6 ). That the echodensities are alongside the lenticulostriate arteries has been shown by Doppler ultrasound examination (see Fig. 34.6 ). In two series, 15% to 40% of infants with such echodensities had CMV infection. (Other diagnoses included congenital rubella, congenital syphilis, trisomy 13, trisomy 21, fetal alcohol spectrum disorder, metabolic disorders, and neonatal asphyxia .) One pathological study defined hypercellular vessel walls and a mineralizing vasculopathy, probably secondary to perivascular inflammation ( Fig. 34.7 ).

Congenital cytomegalovirus infection: periventricular cyst in subependymal germinal matrix: cranial ultrasound.

(A) This scan, performed at 1 day of age, demonstrates the bilateral cysts (small arrowheads). The ventricles are indicated by the large arrowheads. (B) Coronal section of brain shows both cysts in the subependymal regions.

(Courtesy Dr. Gary Shackelford.)

Congenital cytomegalovirus infection: cranial ultrasound.

Coronal (A) and parasagittal (B) views showing large bilateral germinolytic cysts and lenticulostriate vasculopathy. Using power Doppler (B), it is clear that the echogenic linear abnormalities follow the lenticulostriate arteries.

Congenital cytomegalovirus infection: cranial ultrasound.

Coronal (A) and parasagittal (B) views showing small bilateral germinolytic cysts, lenticulostriate vasculopathy, and mild ventriculomegaly. Magnetic resonance imaging, axial T2-weighted images of the same infant (C and D) additionally show an intraventricular hemorrhage (C), a hemorrhagic lesion in the parietal white matter (D), and extensive polymicrogyria.

(From Gunkel J, van der Knoop BJ, Nijman J et al. Congenital cytomegalovirus infection in the absence of maternal cytomegalovirus-IgM antibodies. Fetal Diagn Therapy . 2017; DOI: 10.1159/000456615.)

MRI is of particular value for detecting the disorders of neuronal migration, cerebral parenchymal destruction, delays in myelination, and cerebellar hypoplasia observed with congenital CMV ( Figs. 34.8–34.10 ). ^a

a References .

In a large study of 40 infants with cCMV infection cranial ultrasound and cardiac MRI were performed within the first month of life. Six newborns showed pathological cardiac MRI and cranial ultrasound findings (pseudocysts, ventriculomegaly, calcifications, cerebellar hypoplasia), but MRI provided additional information (white matter abnormalities in three cases, lissencephaly/polymicrogyria in one, and a cyst of the temporal lobe in another); cerebral calcifications were detected in three of six infants by cranial ultrasound but in only two of six by MRI. Four of these six infants showed severe neurodevelopmental impairment and five showed deafness on follow-up. Three newborns had a normal cranial ultrasound, but MRI documented white matter abnormalities and in one case also cerebellar hypoplasia; all showed neurodevelopmental impairment and two were deaf at follow-up. In another study, 36 infants with cCMV infection were studied, with MRI available in 20, allowing comparison of cranial ultrasound and MRI. Migrational disorders were diagnosed only with MRI in 9 of the 20 infants assessed with this technique. Of 10 infants infected during the first trimester, 7 had severe abnormalities on cranial ultrasound (5 confirmed on MRI) and adverse sequelae; 3 had no or mild abnormalities on cranial ultrasound/MRI and a normal outcome. Of seven infants infected during the second or third trimester with no/mild abnormalities on cranial ultrasound/MRI, six had a normal outcome; one with mild cranial ultrasound and MRI abnormalities developed SNHL. As expected, the worst outcome was seen in 16 of 26 symptomatic infants with severe cranial ultrasound/MRI abnormalities (neuronal migration disorders seen only on MRI: cerebellar hypoplasia, ventriculomegaly, extensive periventricular calcifications, and white matter cysts). In 1 of 16 infants with only mild abnormalities on cranial ultrasound (germinolytic cysts and LSV), occipital cysts as well as extensive polymicrogyria were noted with MRI (see Fig. 34.8 ), highlighting the need for MRI even in the absence of severe cranial ultrasound abnormalities. The study again showed that infants with cCMV infection acquired during the first trimester of pregnancy are at increased risk of symptomatic presentation with severe cerebral abnormalities and the subsequent development of adverse sequelae such as cerebral palsy, SNHL, and mortality.

Congenital cytomegalovirus infection: photomicrograph of a small vessel.

Note the thickened wall, focal globular subendothelial deposits of mineralized material, mononuclear infiltrates in adventitia, and perivascular reactive astrocytosis (hematoxylin and eosin, Luxol fast blue, ×250).

(From Teele RL, Hernanz-Schulman M, Sotrel A. Echogenic vasculature in the basal ganglia of neonates: a sonographic sign of vasculopathy. Radiology .1988;169:423-427.)

Congenital cytomegalovirus infection: magnetic resonance imaging.

Axial T2-weighted images showing ventriculomegaly, a loculated area in the right occipital horn, periventricular calcification (low signal intensity), extensive polymicrogyria (A), and cerebellar hypoplasia as well as increased signal intensity in the white matter (B).

Congenital cytomegalovirus infection, magnetic resonance imaging.

This 16-day-old infant was born after a 31-week gestation with congenital cytomegalovirus infection identified in utero. The axial T1-weighted magnetic resonance images show (A) increased signal in the periventricular regions (short arrows) , consistent with calcification, and diffuse polymicrogyria (long arrow) ; in B, note the striking cerebellar hypoplasia (arrows) . At 6 months of age, the axial T2-weighted image (C) shows diffuse frontal polymicrogyria (long arrows) , abnormal high signal intensity in cerebral white matter (short black arrows) , and marked paucity of parieto-occipital cerebral white matter (double white arrows) .

(Courtesy Dr. Omar Khwaja.)

In a carefully studied series of 11 infants with congenital CMV infection, polymicrogyria was present in 5 and lissencephaly in 4. In a series of MRI-documented cases of lissencephaly-pachygyria, CMV infection was present in 6 of 23 infants. In a related study of 10 infants with MRI-identified migrational disorders, especially polymicrogyria, 4 were found to have CMV. Delays in myelination and increased signal on T2-weighted images (see Figs. 34.8 and 34.9 ) have been observed in approximately half of the infants with CMV studied by MRI. ^a

a References .

Indeed, the predilection of abnormal cerebral white matter signal, including cystic change, for posterior parietal regions may mimic PVL. Thus, in an MRI series of 152 infants (mean age 22 months) with “static leukoencephalopathy of unknown etiology,” 10% were found to have congenital CMV, based on retrospective PCR testing of neonatal blood spots. Cerebellar hypoplasia, a finding in 40% to 70% of infants with cCMV infection, is detected best by MRI scanning (see Fig.34.10 ). Notably, MRI is less sensitive than CT and cranial ultrasound for the detection of cerebral calcifications.

Antenatal neuroimaging combining cranial ultrasound and MRI has shown characteristic findings, also seen on postnatal imaging. Many of the common findings in cCMV are illustrated in a recent review by Averill and colleagues. Dilated occipital horns of the lateral ventricles with thin septations can be well visualized with cranial ultrasound as well as MRI. These characteristic occipital cysts are usually bilateral and will become less conspicuous with time (see Figs. 34.10 and 34.11 ). Polymicrogyria is likely to be missed with cranial ultrasound and is easier to detect on postnatal MRI, but it is often seen or at least suspected on fetal MRI. In a study by Doneda et al., prenatal cranial ultrasound and MRI findings were compared in 30 fetuses with a proven cCMV infection. Fetal MRI did show higher sensitivity than cranial ultrasound in predicting symptomatic infection (83% vs. 33%). However, both modalities showed low positive predictive values (36% with MRI vs. 29% with cranial ultrasound). In another study of fetal cranial ultrasound and MRI in 38 cases of cCMV infection, MRI was shown to add important details, especially with regard to the detection of gyrational anomalies, cerebellar hypoplasia, and white matter abnormalities. In both studies and a more recent study, a negative fetal brain MRI finding was reassuring for a good clinical outcome, although the hearing loss may still develop with time.

Congenital cytomegalovirus infection, magnetic resonance imaging.

Axial T2-weighted (A) and parasagittal (B) images showing large germinolytic cysts as well as loculated areas within the occipital and temporal horns of the enlarged lateral ventricles. Signal intensity in the white matter is diffusely increased.

Hearing loss caused by cCMV infection was first reported in 1964 and is the most common sequela of cCMV infection. Hearing loss is thought to be due to cytopathic effects and localized inflammatory responses. This infection is now known to be the most common cause of nonhereditary SNHL, involving 10% to 20% of hearing-impaired children. Testing of brain stem auditory evoked responses in the neonatal period and subsequently demonstrates the high likelihood of SNHL, including the delayed onset and postnatal progression of this loss. In four series of 281 infants with symptomatic congenital CMV infections, 50% to 75% exhibited hearing loss on follow-up. Although approximately 60% of those with hearing loss had hearing loss at birth or in the neonatal period, fully 40% had delayed-onset loss (i.e., not apparent until months after the neonatal period). In addition, progressive hearing loss was noted in approximately 60% of the infants with hearing loss. Progression of hearing loss has also been observed in infants with asymptomatic CMV infection. Thus, in one series, 3% of such patients had hearing loss detected in the neonatal period, but by the age of 6 years, 11% of the previously asymptomatic patients had hearing loss. A recent meta-analysis described findings from 14 longitudinal and 13 retrospective studies. The researchers were able to show that of infants with a proven cCMV infection 12.6% (95% CI, 10.2 to 16.5) will have hearing loss; 1 of 3 symptomatic children and 1 of 10 asymptomatic children. Bilateral hearing loss was present in most children with symptomatic cCMV infection, whereas unilateral hearing was more common in those with asymptomatic cCMV infection. Hearing loss may have a delayed onset and can vary over time. Foulon et al. showed that the risk of cCMV-related SNHL was highest when the infection occurred during the first trimester (4 of 5; 80%), rare following an infection during the second trimester (1 of 12; 8%), and nonexistent in 11 children with cCMV acquired during the third trimester. Fluctuation and improvement of SNHL were seen regardless of the trimester of pregnancy during which the mother’s primary infection occurred. The risk of delayed hearing loss was shown to be associated with the presence of symptoms at birth ; children who passed initial audiological examinations but had cCMV-related symptoms at birth (e.g., jaundice, petechiae, and microcephaly) were nearly 6 times more likely to develop hearing loss than children who were asymptomatic at birth. A longer duration of viral shedding may also be a predictor of delayed hearing loss. The value and importance of serial studies throughout infancy are obvious .

Relation to Time of Onset of Fetal Infection and Antenatal Neuroimaging Findings.

In a large prospective study that enrolled 145 fetuses during pregnancy—with a primary CMV infection obtained during the first and second trimesters of pregnancy in 71 and 74 patients, respectively—the risk of an adverse outcome was significantly higher when the infection occurred during the first trimester and when imaging abnormalities were found as well ( Table 34.6 ). Abnormal prenatal findings on ultrasound examination were associated with an increased risk of sequelae. In a recent study of 121 fetuses, MRI was performed at 27 and/or 33 weeks (51 at both time points). A five-grade classification was used: (1) for normal findings, (2) the presence of isolated frontal or parieto-occipital periventricular T2-weighted signal hyperintensity, (3) the presence of isolated temporal periventricular T2-weighted signal hyperintensity, (4) the presence of cysts and/or septa in the temporal and/or occipital lobe, and (5) the presence of migration disorders, cerebellar hypoplasia, and/or microcephaly. Isolated periventricular T2-weighted signal hyperintensity is a very common finding in cCMV infection (41%) but was not associated with adverse postnatal outcome except for 3 of the 21 neonates (14.3%) with isolated hyperintensity of the temporal lobes who had SNHL. The negative predictive value (NPV) was especially high: 96% in the absence of any MRI abnormalities.

Relation to Neonatal Clinical Syndrome.

The outcome relates to the severity of the neuropathological findings, and these findings correlate with the neonatal clinical syndrome ( Table 34.7 ). Although the data depicted in Table 34.7 are based on a sample that was selected to a certain degree, the observations are useful regarding the relationship between the neonatal clinical signs and the neurological outcome in congenital CMV infection. Thus, of those infants with the overt neurological syndrome (i.e., microcephaly, intracranial calcifications, or chorioretinitis), approximately 95% had major neurological sequelae (e.g., mental retardation, seizures, deafness, and motor deficits) or died. Infants with less obvious ( other ) neurological phenomena had slightly better prognoses. Approximately 70% of these infants with neonatal neurological signs also experienced systemic phenomena. In the large series ( n = 80) of MacDonald and Tobin, of the group of infants with systemic signs but no neonatal neurological deficits, approximately 50% were normal and only 16% exhibited major neurological sequelae or died (see Table 34.7 ). Further insight into the spectrum of cCMV infection is provided by the results of a recent study of 178 infants by Dreher et al. Comparison was made between a group of 78 recognized by newborn screening and 100 infants referred with clinical symptoms leading to a diagnosis of cCMV infection. Two or more clinical findings were detected at birth in 91% of referred infants and only 58% of screened infants ( P <.001). Significantly more children in the referred group had hearing loss compared with screened infants ( P = .009). Of the screened children, 51% were free of sequelae at follow-up compared with only 28% of the referred group ( P <.003).

For a more reliable prediction of neurological sequelae, MRI is currently recommended, especially in the presence of any cranial ultrasound abnormality. MRI allows assessment of additional migrational abnormalities not recognized with cranial ultrasound (see Fig. 34.8 ). The major neurological deficits include pronounced cognitive deficits, most commonly with intelligence quotient (IQ) scores lower than 70; spastic motor deficits; seizure disorders; and bilateral hearing loss. ^a

a References .

Outcome With Asymptomatic Congenital Infection.

The asymptomatic group has been the particular focus of numerous investigators. ^a

a References .

The possibility of subtle disturbances of intellectual function was initially suggested by studies conducted by Hanshaw et al., who demonstrated a statistically significant lower mean IQ score in asymptomatic patients versus matched controls (102 vs. 112). Subsequent large-scale studies did not document definite impairment of intellectual function in asymptomatic infants, particularly when hearing-impaired children were excluded. Intellectual outcome in hearing-impaired children has not yet been studied in detail; it is important to define outcome in this setting separately because the virus clearly has entered the CNS in this subgroup of infected infants. Nevertheless, several reports suggest that an asymptomatic neonatal period may be followed by varying combinations of developmental delay, microcephaly, ataxia, SNHL, and seizures, usually recognized in the first year of life. CT or MRI has shown cerebral calcification, abnormal cerebral white matter signal, delayed myelination, polymicrogyria, focal subcortical areas of abnormality, or cerebellar hypoplasia. The possibility of late intrauterine acquisition of infection has been suggested in some of these asymptomatic infants, but most studies indicate that although the risk of intrauterine transmission following primary maternal infection in the third trimester is high, the risk of neonatal disease is low. In the study by Enders and colleagues, no symptoms were observed in infected newborns of mothers with primary infection in the preconceptional period and in the third trimester.

The important clinical point is that CMV infection should be considered later in infancy in the presence of such neurological or neuroradiological features or both, even if the neonatal period was unremarkable . More data are needed on these issues.

Prevention.

Congenital CMV infection is related to primary infection of the pregnant woman, presumably early in pregnancy. Two preventive approaches may be used: one to prevent or treat the primary infection and the other to terminate the pregnancy. Prevention of the primary maternal infection by vaccination has received initial investigation, with variable results. However, more information is needed about the effectiveness, hazards, and feasibility of this approach. Treatment of the primary maternal infection with hyperimmune gamma globulin is a possibility, but the difficulty in detecting most maternal infections has been a major problem with this approach. The results of the first randomized trial with hyperimmune gamma globulin or placebo enrolled 124 women with primary CMV infection. There was no difference in viral load in the amniotic fluid and newborn urine and no difference in cCMV infection. There was a nonsignificant increase in adverse obstetrical events, including preterm birth, preeclampsia, and fetal growth restriction. One other trial (NCT01376778) with hyperimmune gamma globulin is still in progress. Termination of a pregnancy complicated by a primary maternal infection has been difficult because the exact risks of fetal infection are not entirely known. Detection of the infected fetus by amniocentesis and identification of the virus or DNA by culture or PCR, respectively, constitute the principal approach. The sensitivity for detection of fetal infection increases markedly after 21 weeks of gestation. The fetal condition can then be assessed further by ultrasonography, which may show intracranial calcification or other evidence of parenchymal disease, and, if desired, by cordocentesis, with evaluation of fetal blood for abnormal liver function tests, CMV-specific IgM, anemia, or thrombocytopenia. In one series, the risk of identification of neonatal neurological abnormality by neurological examination, cranial ultrasonography, or hearing assessment was only 19% when no prenatal ultrasonographic abnormalities were present. Ultrasonographic abnormalities were detected prenatally in 21%, and nearly all of these pregnancies were terminated.

Supportive Therapy.

From the neonatal neurological standpoint, supportive therapy consists principally of control of seizures.

Antimicrobial Therapy.

Prenatal therapy with CMV-specific hyperimmune immunoglobulin is controversial, with a recent phase 2 clinical trial study refuting earlier claims of improved outcomes. Investigation of maternal oral treatment with valganciclovir is currently in clinical trials as well. Treatment in the neonatal period is more established and is instituted in infants with evidence of brain involvement including SNHL as well as other serious end-organ disease. Several antiviral agents—including adenine arabinoside (Ara-A), 5-iodo-2′-deoxyuridine (IDU), cytosine arabinoside (Ara-C), and acyclovir—have been studied because of their effectiveness in vitro. Ara-A and acyclovir are the least toxic of these agents, but early trials did not provide reason for optimism. Ganciclovir, an acyclovir derivative, has been shown to be effective in the prophylaxis and treatment of CMV infections in immunocompromised adults and children. Initial data with infants are promising. An earlier study of 12 infants with congenital CMV infection suggested distinct clinical benefit (e.g., loss of hepatosplenomegaly, improvement in tone) with a 3-month course of therapy. The most impressive data with ganciclovir involved a randomized controlled trial of the effect of the agent on hearing in symptomatic congenital CMV disease involving the CNS ( Table 34.9 ). The treated infants received ganciclovir, 6 mg/kg per dose, administered intravenously every 12 hours for 6 weeks. Hearing deficits either improved or remained static in 56% of the ears of treated infants versus only 17% of those of the control infants (see Table 34.9 ). Progression of deficits occurred in only 21% in the ganciclovir group versus 61% in the control group. The beneficial effect of ganciclovir was accompanied by significant neutropenia in approximately 65% of treated infants. These findings were supported by an observational study enrolling 23 asymptomatic children with cCMV infection. Twelve children were treated just after diagnosis of cCMV infection in the newborn period, with ganciclovir 10 mg/kg bodyweight for 21 days. The other 11 children were observed without therapy. All 23 children had normal sensorineural hearing at 1-year follow-up. In all, 18 children were seen over the 4- to 11-year follow-up period. SNHL occurred in 2 (11.1%) children who did not receive ganciclovir in the newborn period. None of the nine ganciclovir-treated children developed SNHL. During ganciclovir therapy, moderate neutropenia occurred as a side effect in 2 out of 12 (16.6%) treated children.

The experience with ganciclovir illustrates the need for an agent that has less toxicity and can be administered orally. Valganciclovir, shown to be effective in adults with CMV retinitis, may prove to be such an agent. In a recent randomized, placebo-controlled trial of valganciclovir therapy in 96 neonates with symptomatic cCMV disease, the effect of 6 months of therapy was compared with that resulting from 6 weeks of therapy ( Table 34.9 ). The primary end point was the change in hearing in the better ear ( best ear hearing) from baseline to 6 months. Secondary end points included the change in hearing from baseline to follow-up at 12 and 24 months and neurodevelopmental outcomes, with each end point adjusted for CNS involvement at baseline. Of the initial 96 neonates, 86 could be evaluated at 6 months. There was no difference for best-ear hearing at 6 months. Total-ear hearing (hearing in one or both ears that could be evaluated) was more likely to be improved or to remain normal at 12 months in the 6-month treatment group compared with the 6-week treatment group (73% vs. 57%, P = .01). This benefit was still present at 24 months (77% vs. 64%, P = .04). At 24 months, the 6-month group, as compared with the 6-week group, had better neurodevelopmental scores on the language-composite component ( P = 0.004) and the receptive-communication scale ( P = .003) of the Bayley Scales of Infant and Toddler Development (3rd ed.). Neutropenia was not uncommon and occurred in 27% of those in the 6-week treatment group. In the 6-month treatment group, neutropenia occurred in 19% of the infants during the first 6 weeks and in 21% during the next 4.5 months. Because migrational disorders are associated with the most severe adverse neurological sequelae and evolve in utero, it is unlikely that either 6 weeks or 6 months of treatment with valganciclovir will have a positive effect on such neurological sequelae. Preservation of hearing, however, would be of benefit, and the policy of 6 months of therapy is now recommended. Treatment for 12 months with a combination of ganciclovir and valganciclovir was recently reported. Hearing impairment was diagnosed at birth in 54 (36%) of the 149 infants diagnosed with symptomatic cCMV; it was unilateral in 31 (57%) and bilateral in 23 (43%). After 1 year of antiviral treatment and a long-term follow-up of the 77 affected ears at baseline, 50 (65%) had improved, 22 (29%) remained unchanged, and 5 (6%) had deteriorated. Most improved ears (38 of 50 = 76%) returned to normal hearing. Improvement was most likely to occur in infants born with mild or moderate hearing loss and less in those with severe impairment.

There is at present no agreement about the beneficial role of treatment with valgancyclovir in preterm infants with pCMV infection. Most clinicians are reluctant to use a potentially toxic drug in preterm infants and restrict therapy to those with a sepsis-like illness. Ultimately, treatment of asymptomatic infants would be ideal if the risk-to-benefit ratio of the agent were favorable.

Fetal Infection.

Clinically significant infection with toxoplasmosis occurs during intrauterine life by transplacental passage of the parasite. The sequence of events is (1) primary infection of the mother, (2) parasitemia, (3) placentitis, and (4) hematogenous spread to the fetus. The organism can be cultured consistently from the placenta when the fetus is infected. As with CMV, the mother infected with toxoplasmosis is usually asymptomatic. The most common clinical presentation of the mother is localized or generalized lymphadenopathy, sometimes with fever and other features suggestive of infectious mononucleosis.

The incidence of primary maternal infection during pregnancy varies around the world. In Paris, when consumption of undercooked meat was relatively common, the value was as high as 5 per 100 pregnancies. At a comparable time, the rate in the United States was approximately 1.1 per 1000 pregnancies. More recent incidences are 0.5 to 2.0 per 1000 pregnancies in Western Australia, Europe, and the United States. These rates should be contrasted with the approximately tenfold-higher rates for CMV infection during pregnancy (see earlier discussion). A study of congenital infection, based on a serological investigation of filter paper blood specimens for neonatal metabolic screening in Massachusetts, yielded an incidence of only approximately 1 per 10,000. In a study from Switzerland using enzyme-linked immunosorbent assay (ELISA) (IgM and/or IgA antibodies) of the cord blood, seroprevalence of congenital toxoplasmosis declined from 0.08% to 0.012% from 1982 to 1999. Despite increasing maternal age, seroprevalence for toxoplasmosis decreased steadily from 53% to 35% during this period. The unusual susceptibility of the human fetus and newborn to severe infection with T. gondii appears to relate in large part to inadequate cellular defenses. Mononuclear phagocytes are the principal defense against such infection, and a decreased generation of macrophage-activating material by fetal lymphocytes has been demonstrated. Moreover, the response to this activating material by macrophages in the neonate is also deficient. Uncontrolled replication of the organism is the expected result.

Importance of Time of Maternal Infection.

The likelihood and severity of congenital toxoplasmosis bear a distinct relation to the time of maternal infection ( Table 34.10 ). Only approximately 20% to 25% of infants will be infected if the maternal infection occurs in the first or second trimester, especially the second to sixth months of gestation, versus approximately 65% if maternal infection occurs in the third trimester. In a large series of 603 women with confirmed maternal toxoplasmosis, the maternal-fetal transmission rate was only 6% with infection at 13 weeks but increased to 72% with infection at 36 weeks. However, although fetal infection is less likely earlier in pregnancy, the severity of the disease is greater. Indeed, most infants infected in the first trimester exhibit severe disease, manifested by CNS and ocular involvement ( Table 34.11 ). As a result of these counterbalancing effects, in one large series the highest risk of bearing an infected infant with early clinical manifestations (10%) occurred in women who seroconverted at 24 to 30 weeks of gestation. A CT study of 31 infants further documented the severity of the CNS lesions as a function of the time of intrauterine infection. Therefore it appears that although a fetal-maternal barrier to infection may be operative early in pregnancy, once fetal infection is established at that time, it is a potentially devastating disease. Treatment of the infected mother alters both the likelihood of fetal transmission and the severity of the disease (see later discussion). Following the introduction of prenatal screening in France in 1992, a significant reduction in the rate of congenital infection and a better outcome at 3 years of age in infected children was reported. Among 2048 mother-infant pairs, 93% of mothers received prenatal treatment and 513 (25%) fetuses were infected. Probabilities of congenital infection were less than 10% for maternal infections before 12 weeks of gestation, 20% at 19 weeks, and to 52% and almost 70% at 28 and 39 weeks’ GA, respectively.

Meningoencephalitis.

The meningoencephalitis of toxoplasmosis has a striking multifocal, necrotizing, granulomatous quality and is characterized by the following: (1) inflammatory cells in the meninges, especially over focal lesions; (2) perivascular infiltrates with inflammatory cells, the latter often including eosinophils; (3) multifocal and diffuse necroses of brain parenchyma, with all cellular elements affected, involving cerebrum, brain stem, and spinal cord and often associated with calcification; (4) reactive microglial and astroglial proliferation; and (5) miliary granulomas, containing large epithelioid cells and free, intracellular, or encysted organisms ( Fig. 34.12 ).

Congenital toxoplasmosis: encephalitis.

Photomicrograph of a region of necrosis containing many free Toxoplasma organisms (note small, darkly stained nuclei to the left of larger, preserved neurons). Although this lesion was from an older child with Toxoplasma encephalitis who was receiving immunosuppressive therapy, the organisms are identical in appearance to those of the congenital form.

(From Bell WE, McCormick WF. Neurologic Infections in Children . Philadelphia: Saunders; 1975.)

Porencephaly and Hydranencephaly.

With particularly severe, diffuse, cerebral destructive disease, porencephalic cysts or hydranencephaly may develop. Of the 33 fetuses with congenital toxoplasmosis studied by Hohlfeld et al., all exhibited areas of brain necrosis, the initial lesions that evolve to porencephaly and hydranencephaly. The development of these large areas of tissue destruction is particularly likely if aqueductal block and increased intraventricular pressure are associated.

Hydrocephalus.

Two processes appear to be operative in the periventricular region with toxoplasmosis and may underlie the propensity for aqueductal block and consequent hydrocephalus in this disorder. First, the inflammation with toxoplasmosis has a predilection for the periventricular region, as with CMV infection (although in toxoplasmosis more severe diffuse disease is present elsewhere, and calcified areas of necrosis are present throughout the cerebrum). Second, it is believed that Toxoplasma organisms enter the ventricular system from the parenchymal lesions and disseminate there. This highly antigenic ventricular fluid then seeps through the damaged ependyma to periventricular blood vessels, where an antigen-antibody reaction may occur at the vessel wall, thereby causing thrombosis and periventricular infarction. This additional necrosis apparently causes the serious aqueductal block that results in hydrocephalus, the common complication. Among 33 infected fetuses identified in utero by Hohlfeld et al., 19 (58%) had ventricular dilation at autopsy after elective termination of pregnancy.

Microcephaly.

Although hydrocephalus is a more common result of congenital infection with T. gondii and is more frequent in this variety of congenital infection than in any other, microcephaly does occur in a significant percentage of patients, approximately 15% (see subsequent discussion). The microcephaly relates to the multifocal necrotizing encephalitis, particularly of the cerebral hemispheres. Indeed, even in patients with hydrocephalus, it is clear that a serious loss of brain substance, in addition to the effects of the hydrocephalus, has invariably occurred.

Incidence of Clinically Apparent Infection.

As with CMV, clinically asymptomatic cases of congenital toxoplasmosis outnumber symptomatic cases. However, a larger proportion of infants with congenital toxoplasmosis than with congenital CMV infection can be detected clinically in the newborn period.

Of 156 children with congenital toxoplasmosis who were monitored prospectively from the time of maternal infection, approximately 18% had CNS and ocular involvement, 2% had CNS involvement without ocular involvement, 12% had ocular involvement only, and 68% were asymptomatic. Thus 20% of infants with congenital toxoplasmosis had observable CNS involvement in the newborn period in this study. The incidence of subclinical infection is higher in infants of women treated during pregnancy than in infants of women not treated (see later discussion). The findings are quite similar when compared with a more recent prospective study covering a 20-year period (1985–2005), where all mothers received spiramycin, alone or associated with pyrimethamine-sulfadoxine, and underwent amniocentesis and monthly ultrasound screening. Of 666 liveborn children, 112 (17%) had congenital toxoplasmosis and 107 were followed for 12 to 250 months: 79 were asymptomatic (74%) and 28 had chorioretinitis (26%). There was only one infant with serious neurological involvement.

Clinical Features.

Symptomatic patients (not treated in utero) can often be divided into those with predominantly neurological syndromes and those with predominantly systemic syndromes ( Table 34.12 shows combined data for both syndromes). The neurological syndrome accounts for approximately two thirds of the cases and consists principally of abnormal CSF and other signs of meningoencephalitis, seizures, diffuse intracranial calcification, hydrocephalus, or, less commonly, microcephaly (see Table 34.12 ). At least 90% of these patients exhibit chorioretinitis (also termed retinochoroiditis). In congenital toxoplasmosis, chorioretinitis is typically bilateral and prominent in the macular regions ( Fig. 34.13 ) and is of major diagnostic importance. Initially, the lesion appears in the fundus as yellowish white, cotton-like patches with indistinct margins. These patches evolve over the ensuing months into sharply demarcated, punched – out , pigmented lesions, often accompanied by optic atrophy. Although the chorioretinitis is most commonly apparent in the newborn period, particularly by indirect ophthalmoscopy, it may not develop for months or even years. These ocular lesions may relapse after birth despite pre- and postnatal treatment. In a recent study by Wallon and colleagues, 477 cases of confirmed congenital toxoplasmosis were followed for a median of 10.5 years (75th percentile: 15.0 years). Almost one third (29.8%) showed at least one ocular lesion. The lesion was unilateral in about two thirds (69.0%) and lesions were first manifested at a median age of 3.1 (0.0 to 20.7) years. In one third (33.8%) of the children, recurrences or new ocular lesions occurred up to 12 years after the appearance of the first lesion. Early maternal infection, prematurity, and nonocular congenital toxoplasmosis lesions at the time of diagnosis of congenital toxoplasmosis were associated with a higher risk of chorioretinitis.

Congenital toxoplasmosis: chorioretinitis.

Note the striking lesion at the macula (right) as well as optic atrophy (left).

(From Bell WE, McCormick WF. Neurologic Infections in Children . Philadelphia: Saunders; 1975.)

The systemic syndrome of congenital toxoplasmosis is dominated by signs referable to the reticuloendothelial system, especially hepatosplenomegaly, hyperbilirubinemia, and anemia (see Table 34.12 ). A petechial rash may occur but is less common than with CMV infection. Overt clinical evidence of neurological involvement is often lacking in these patients, but CSF abnormalities, frequently with a disproportionately elevated protein content for the degree of pleocytosis, occur in approximately 85% and reflect concomitant meningoencephalitis. Chorioretinitis is observed in at least two thirds of patients with systemic infections, a finding that underscores the importance of careful evaluation of the fundus, especially by indirect ophthalmoscopy, when congenital toxoplasmosis is possible.

An unknown but probably very considerable number of infants with no neurological or systemic signs of congenital toxoplasmosis (i.e., asymptomatic disease ) will exhibit chorioretinitis, detectable in the newborn period by indirect ophthalmoscopy. Although most of the retinal lesions observed later probably develop in the weeks and months after delivery (see subsequent discussion), further data are needed regarding the proportion detectable in the newborn period. In a series of 48 asymptomatic infants in whom infection was detected by newborn blood screening, 2 had active chorioretinitis and 7 others had retinal scars; thus 19% had retinal disease. Moreover, approximately 20% had cerebral calcifications detectable by CT and 25% had CSF findings consistent with encephalitis. The later development of neurological deficits and visual loss is appreciable and is discussed in the section on prognosis.

The clinical course of the disease is not readily predicted; indeed, evidence of progression of retinal and cerebral disease has been presented. Because many patients with symptomatic congenital toxoplasmosis exhibit very severe neurological deficits from the neonatal period, the frequency of progression is difficult to quantitate.

Clinical Diagnosis.

Certain clinical features are helpful in suggesting the diagnosis of congenital toxoplasmosis. A particularly noteworthy constellation of features includes evidence of meningoencephalitis, focal and multifocal cerebral necroses, diffuse cerebral calcification, hydrocephalus, and scarring chorioretinopathy in the macular regions. Fetal growth restriction or prematurity is generally not a prominent feature, as in infants with congenital CMV or rubella infections, and microcephaly is less common than in congenital CMV infection. The systemic syndrome may cause confusion in differentiating toxoplasmosis from other congenital infections, but a petechial rash is relatively less common in congenital toxoplasmosis.

Isolation of Toxoplasma.

Determination of T. gondii as the responsible microbe in the newborn with congenital toxoplasmosis depends principally on serological tests rather than isolation of the organism per se. Nevertheless, the organism or associated DNA can be isolated from placental tissue, ventricular or lumbar CSF, blood, and amniotic fluid. The tissue extracts or fluids can be injected into either mice or tissue culture preparations. Parasitemia is more readily demonstrated in the first week after birth (71%) than in the second to fourth weeks (33%), and it is detected most easily in generalized rather than in neurological disease. The isolation procedures for toxoplasmosis require specialized techniques and an experienced, skilled laboratory staff. Detection of T. gondii in amniotic fluid by PCR with a sensitivity of 80% has suggested that this rapid test (result available in ≤6 hours) could become very important in the diagnosis when the technique is applied to biological fluids of the newborn infant. Even better overall sensitivity of 92.2% (95% CI 81% to 98%) was reported in another study, where PCR was performed on 261 amniotic fluid samples. There were four negative results in fetuses that were infected. PCR performed in the CSF was found to be positive in 27 of the 58 (46.5%) congenitally infected infants and negative in each of the 103 infants without congenital toxoplasmosis. The CSF PCR was positive in 70.9%, 53.3%, and 50.9% of those with hydrocephalus, cerebral calcifications, and/or eye disease, respectively. Of six infants who were negative for both IgM and IgA antibodies, three had a positive PCR in their CSF as the confirmatory test for diagnosis of congenital toxoplasmosis. IgM and IgA antibodies and CSF PCR, when combined, yielded a higher sensitivity for diagnosis of congenital toxoplasmosis when compared with the performance of each test alone. The sensitivity of PCR on placental tissue is only approximately 60%.

Serological Studies.

The identification of most cases of congenital toxoplasmosis is established by serological techniques. The two most commonly used tests are the Sabin-Feldman dye test and the IgM-fluorescent antibody test. The Sabin-Feldman dye test , perhaps the single most reliable test, is performed by mixing live organisms with the test serum (and a human serum component, the accessory factor) and then exposing the mixture to methylene blue. Parasites exposed to the antibody-containing serum are modified and stained. The antibodies for the dye test are passively transferred, and the maternal component does not decrease significantly for several months after birth. Persistence of high titers is necessary for diagnosis and obviously is time-consuming. Moreover, the dye test requires the use of hazardous live organisms, and the accessory factor is sometimes difficult to obtain. The Sabin-Feldman dye test, immunofluorescent antibody test, ELISA, IgG avidity test, and agglutination and differential agglutination test can be used for the detection of IgG antibodies. These tests are positive within 1 to 2 weeks after the infection and persist indefinitely. IgM antibodies arise within the first week of infection, rapidly increase, and thereafter decline and disappear at highly variable rates. A negative IgM test essentially rules out a recently acquired infection. However, it should be noted that commercial kits used to detect IgM antibodies in nonreference laboratories may be unreliable, with false-positive rates as high as 60%. The avidity test for IgG antibodies helps to discriminate between a recently acquired infection and one obtained in the more distant past. The presence of high-avidity antibodies essentially rules out infection acquired in the preceding 3 to 4 months. Reactivation can be seen in immunocompromised women, and this has been reported in HIV-infected women. The Toxoplasma-specific IgM-fluorescent antibody technique is faster and more specific. This test measures fetally produced IgM antibody to the organism. Killed organisms are used to bind specific IgM, which is then detected by exposure to fluorescein-antiserum to human IgM. The reliability of this test is hindered by certain factors that impair both sensitivity and specificity. The more recently developed IgM-capture ELISA , which isolates and concentrates the infant’s IgM, increases the sensitivity markedly (90% of infected infants are detected), and false-positive reactions are unusual. This test has been adapted to filter paper blood specimens. More recent data suggest value for the analysis of IgA or IgE in neonatal blood. However, sensitivities for all these last tests are very low for infants infected in the first 20 weeks of gestation, when severe disease is the most likely result. Finally, the use of a comparative analysis of mother–newborn Ig-G and Ig-M by western blot, first described by Remington et al., has also been advocated. Tissot Dupont et al. reported a sensitivity of 82.6% for the detection of IgG, IgM, and IgA by western blot within the first 3 months of life, whereas at birth the same combination had a sensitivity of 65.2%. The combination of IgG and IgM yielded the best score, whereas IgA detection was the least sensitive. The combination of western blot and conventional serological analysis increased the sensitivity at birth to 78% and within the first 3 months of life to 85%.

Neurodiagnostic Studies.

Neurodiagnostic studies that particularly suggest congenital toxoplasmosis are evaluations of CSF and brain imaging. Pleocytosis and elevated protein content of CSF indicate meningoencephalitis and may be observed in asymptomatic as well as symptomatic patients. Particularly characteristic of congenital toxoplasmosis is the finding of a very high protein content in the ventricular fluid, usually reflecting the aqueductal obstruction and stagnation of infection within the lateral ventricles. Although skull radiographs are effective in demonstrating the diffuse and periventricular cerebral calcification ( Fig. 34.14 ), CT scan is more effective and allows identification of calcification more distant from the lateral ventricles, in contrast to CMV, in which calcifications are mostly adjacent to the lateral ventricles ( Fig. 34.15 ). CT data indicate that calcifications of the basal ganglia are more common than previously suspected. The calcifications associated with congenital toxoplasmosis, especially of the periventricular type or of the basal ganglia, can also be detected by cranial ultrasound (see Figs. 34.15 and 34.16 ). In contrast to cCMV, data on MRI in congenital toxoplasmosis are limited. In a small study of eight cases with antenatal imaging, MRI was performed before and after 3 weeks of spiramycin treatment, showing similar findings. MRI scan provides the most detailed assessment of parenchymal necroses ( Fig. 34.17 ).

Congenital toxoplasmosis: diffuse calcification shown on skull radiograph.

This is a lateral skull film of an infant with hydrocephalus, chorioretinitis, and multiple, punctate calcifications scattered diffusely in brain.

(From Bell WE, McCormick WF. Neurologic Infections in Children . Philadelphia: Saunders; 1975.)

Congenital toxoplasmosis: ultrasonography and computed tomography.

Parasagittal ultrasound (A) showing widespread areas of calcification, confirmed with axial computed tomography (B), showing periventricular as well as subcortical and cortical calcifications.

Congenital toxoplasmosis: cranial ultrasound.

Coronal (A) and parasagittal (B) ultrasound scans showing severely enlarged ventricles and a large left-sided porencephaly. The germinal layer is also very echogenic bilaterally, most likely due to calcification.

Congenital toxoplasmosis: antenatal and postnatal magnetic resonance imaging.

Antenatal T2-weighted images obtained at 32 weeks of gestation in axial (A) and coronal (B) planes showing severe ventriculomegaly, subcortical cysts in the right frontal and left temporal lobe, and migrational disturbances (arrows) . The postnatal image (C) obtained at 34 weeks of gestation confirmed the antenatal findings and showed more subcortical cysts and areas of calcification. (D) Fundoscopy (left eye) shows severe vitritis; the active lesions are seen as whitish foci of retinochoroiditis. (E) Fundoscopy of the right eye shows “headlight in the fog” appearance as a result of intense vitritis.

(D and E, courtesy Elsbeth Voskuil-Kerkhof.)

Relation to the Neonatal Clinical Syndrome.

As with CMV, the outcome of untreated congenital toxoplasmosis relates to the severity of the neuropathology, which correlates to a modest extent with the neonatal clinical syndrome ( Table 34.13 ). Infants with congenital toxoplasmosis with prominent neonatal neurological features do poorly; only 9% are normal on follow-up. Most of the remaining infants exhibit serious disturbances of cerebral function (i.e., mental retardation, seizures, and spastic motor deficits). Essentially all such patients have chorioretinitis and may also have optic atrophy; as a consequence, approximately 70% have severe visual impairment.

Somewhat unlike congenital CMV infection, congenital toxoplasmosis with a neonatal syndrome characterized by prominent systemic signs, if untreated, also results in a poor neurological outcome. Approximately 50% of such patients with CMV are normal on follow-up, whereas only approximately 16% of patients with congenital toxoplasmosis are normal (see Table 34.13 ). Although the nature of the study populations differs, it seems reasonable to conclude that CNS involvement in congenital toxoplasmosis is more prominent than in congenital CMV infection when nonneurological features dominate the neonatal syndrome. Again, chorioretinitis is found in the majority of such patients with toxoplasmosis, and severe visual impairment occurs in approximately 40%. Antiparasitic treatment begun in the first months of life has a beneficial effect on outcome in symptomatic congenital toxoplasmosis (see later discussion).

Outcome With Asymptomatic Congenital Infection.

Infants with subclinical infection (i.e., the majority [about 85%] of cases of congenital toxoplasmosis) comprise an important group. For example, in the United States, approximately 400 to 4000 such infants are affected yearly. Previous studies have emphasized that such infants had a relatively high frequency of chorioretinitis and a modest impairment of intellect. A prospective study of 13 asymptomatic infants identified by serological screening in the newborn period and evaluated by particularly detailed serial, ocular, neurological, and audiological follow-up studies indicated that few such asymptomatic children survive without deficits ( Table 34.14 ). Thus 11 infants in this study developed chorioretinitis (3 with unilateral blindness) and 5 had neurological sequelae (1 with severe mental retardation and microcephaly). The neurological sequelae were always associated with retinochoroiditis. The mean IQ score for the group was only 89. SNHL occurred in 3 of 10 infants tested, although in none was moderate or severe bilateral loss observed. However, diagnosis by neonatal screening and prompt institution of therapy can result in a markedly better outcome (see later discussion).

Prevention.

Three major approaches to prevention include (1) avoidance of primary maternal infection, (2) treatment of maternal infection, (3) abortion in the presence of maternal infection, and (4) treatment of the affected fetus. The first of these approaches is the most important. Pregnant women who have seronegative test results must avoid primary acquisition of Toxoplasma infection; the two measures necessary are avoiding the ingestion of infective cysts (e.g., in raw meat) and contact with sporulating oocysts (e.g., in animal intestine and feces). Ingestion of infective cysts occurs when infected meat is undercooked. It is recommended that consumption of raw or undercooked meat be completely avoided and that handling of raw meat be done with gloves on or followed by careful hand washing. Contact with sporulating oocysts is principally through household cats that carry oocysts in their intestines. It is recommended that pregnant women avoid contact with cat feces and that contact with soil or other materials potentially contaminated with cat feces be avoided or performed while wearing gloves. Cost-to-benefit analyses (relative to other approaches to prevention) demonstrate the particular desirability of a health education campaign to encourage these practices. In a study performed in Finland, the total annual costs of congenital toxoplasmosis without screening amounted, in US dollars, to $128 per pregnancy per year; with systematic serological screening, the cost was $95 per pregnancy per year, thus reducing the cost by 25%. Thus screening for Toxoplasma infections during pregnancy is economically worthwhile even in a country with a low incidence. The investigators recommended systematic screening for maternal primary Toxoplasma infections combined with health education for prevention.

Primary maternal infection has been treated with the antibiotic spiramycin. In one large series, a significant reduction in cases of congenital infection was observed in treated (24%) versus untreated (45%) mothers. The approximately 50% decrease in the incidence of congenital toxoplasmosis was confirmed by subsequent data. Moreover, a multicenter study showed a marked decrease in the incidence of neonatal chorioretinitis and intracranial lesions when prenatal treatment (spiramycin) was instituted within but not after 4 weeks of diagnosis of maternal infection. Treatment varies across countries. Spiramycin is continued throughout pregnancy in the United States and France. In Austria and Germany, spiramycin prophylaxis is followed by a 4-week course of pyrimethamine plus sulfadiazine at 17 weeks of gestation.

Abortion has been performed in women who have exhibited serological evidence of primary infection during early pregnancy. This approach is less desirable for several reasons, one of which is the finding that only 17% to 25% of women infected in the first and second trimesters transmit the infection to the fetus (see Table 34.10 ). However, initial work by Desmonts et al. and subsequently by others demonstrated the feasibility of prenatal diagnosis of congenital toxoplasmosis. Sampling of fetal blood by cordocentesis under ultrasound guidance for serological indicators of infection is less useful, and evaluation of pregnant women for possible fetal infection with toxoplasmosis is based principally on the sampling of amniotic fluid by amniocentesis and detection of the organism’s DNA by PCR. It is recommended to carry out amniocentesis after 18 weeks of gestation and at least 4 weeks after the estimated date of maternal infection to minimize the risk of a false-negative result because of the late passage of the parasite across the placenta into the fetus. Ultrasonography of the fetal cranium at approximately 19 to 20 weeks of gestation provides information concerning cerebral abnormality. Fetal cranial ultrasonography may show ventriculomegaly, evidence of tissue necrosis, and cerebral calcifications. The diagnosis of fetal infection has also been made by isolation of the organism from fetal blood or from amniotic fluid and by identification in fetal blood of hematological abnormalities (e.g., eosinophilia, thrombocytopenia), elevated gamma-glutamyltransferase activity, and Toxoplasma -specific IgM.

The positive identification of fetal infection provides the possibility of treatment of the fetus . In the classic study of Hohlfeld et al., fetal treatment consisted of administration to the mother of alternating 3-week courses of spiramycin and of pyrimethamine, sulfadiazine, and folinic acid (see later discussion). (The last three agents are not used before the 18th week of gestation because of the teratogenic potential of pyrimethamine.) The beneficial effect on the severity of fetal infection was dramatic ( Table 34.15 ). Although, as in pretreatment years, the incidence of severe fetal infection increased the earlier in pregnancy the infection was acquired, only 11% of first-trimester infections treated in utero resulted in severe manifestations in the neonatal period. Moreover, of the third-trimester fetal infections treated in utero, all resulted in asymptomatic newborns. Continuation of antimicrobial therapy postnatally in 53 infants was accompanied, after relatively short-term follow-up, by normal neurological development and examination in 52 (98%) infants and by the development of peripheral chorioretinitis with no visual impairment in 5 (9%) infants. These favorable outcomes represent a dramatic improvement as compared with outcomes in the pretreatment era (see Table 34.13 ). A larger study ( n = 112) also showed a beneficial effect of fetal treatment. The mothers were treated during pregnancy with pyrimethamine-sulfadoxine. Follow-up was available in 107 for 12 to 250 months; 79 were asymptomatic (74%) and 28 had chorioretinitis (26%). Only one child had a serious neurological involvement. A longer-term study confirmed the favorable effects of combined fetal and postnatal therapy, although delayed onset of chorioretinitis was shown ( Table 34.16 ). During follow-up, almost one third of the 142 patients (29.8%) manifested at least 1 ocular lesion. Lesions were unilateral in 98 individuals (69%) and caused no visual loss in 81%. Lesions were first manifested at a median age of 3.1 (0.0 to 20.7) years. In 48 (34%) of the children, recurrences or new ocular lesions were seen up to 12 years after the appearance of the first lesion. However, severe bilateral visual impairment did not occur. Nevertheless, careful long-term follow-up is imperative.

Supportive Therapy.

Such therapy is carried out as described for congenital CMV infection.

Antimicrobial Therapy.

Although significant injury has already occurred in many cases of untreated congenital toxoplasmosis by the time of birth, good evidence indicates that some of this injury is reversible and that continuing postnatal injury is preventable by therapy directed against the organism. ^a

a References .

Fetal Infection.

Clinically significant infection with rubella virus occurs during intrauterine life by transplacental passage of the virus. As with CMV infection and toxoplasmosis, the sequence of events is primary maternal infection, viremia, placental infection, and, finally, fetal infection. Cases of asymptomatic maternal infection are common, outnumbering those of symptomatic infection by nearly 2 to 1. Viremia occurs during the week before the onset of clinical manifestations, which include fever, cervical adenopathy, and a maculopapular rash lasting 3 days.

Importance of Time of Maternal Infection.

The likelihood and severity of fetal infection are functions of the time of maternal infection. The risk to the fetus begins when the rash in the mother appears at least 12 days after the last menstrual period (i.e., the likely time of conception); in a series of 38 carefully studied pregnancies in the periconceptional period, no cases of fetal infection occurred when the rash appeared at 11 days or less after the last menstrual period. In general, both the frequency of occurrence of infection and the severity of clinical disease are greater the earlier in pregnancy the maternal infection occurs ( Table 34.18 ). Thus it differs from the situation with toxoplasmosis, in which the likelihood of infection is less but the severity of disease greater when it is acquired early in pregnancy. With congenital rubella, ocular and cardiac defects are particularly common when infection occurs in the first and second months, but they become essentially nonexistent when infection occurs after the first trimester. However, hearing loss , although most common with early infection, is still found in approximately half of infants infected in the fourth month; later maternal infection appears not to be dangerous in this regard. Neurological deficits , especially intellectual retardation and motor deficits, are most common with infection in the first 2 months and are not observed with infection past the fourth month.

The most critical gestational periods concerning the major defects have been defined particularly closely. In a series of 55 children from carefully dated, affected pregnancies, cataracts were observed with maternal infection between 26 and 57 days of GA; heart disease occurred in maternal infection between 25 and 93 days of GA; deafness occurred in maternal infection between 16 and 131 days of GA; and severe mental retardation occurred in maternal infection between 26 and 45 days of GA. The placenta may play the greatest role in determining the decreasing incidence of fetal infection with progression of gestation. Maturational factors of host tissue may be most important in determining the concomitant changes in organ susceptibility.

Meningoencephalitis.

The meningoencephalitis of rubella infection is similar in certain respects to the other neonatal encephalitides and is characterized by the following: (1) inflammatory cells in the meninges; (2) perivascular infiltrates with inflammatory cells; (3) necrosis of brain parenchyma, with all cellular elements affected; and (4) reactive microglial and astroglial proliferation.

Vasculopathy.

An additional, prominent, and distinctive feature of rubella infection is vasculopathy. Involvement of blood vessels is observed in many organs and prominently in the brain. In the well-studied series of Rorke and Spiro, involvement of large leptomeningeal vessels and, particularly, smaller intraparenchymal vessels and capillaries was defined. Destruction of one or more layers of the vessel wall occurs, with replacement by deposits of amorphous granular material ( Fig. 34.18 ). Associated with these vascular lesions are focal areas of ischemic necrosis, especially in the cerebral white matter (centrum semiovale, periventricular regions, and corpus callosum) and in the basal ganglia. The vascular abnormalities may account for the echogenic vessels observable on cranial ultrasonography of the affected newborn (see later discussion) ( Fig. 34.19 ).

Congenital rubella infection: vasculopathy.

(A and B) Cerebral vessels from a 10-week-old infant with a birth weight of 2250 g and involvement of multiple organs. Note the destruction of vessel walls with replacement by deposits of amorphous granular material, which is evident especially in A. Surrounding ischemic changes are also present, especially in B.

(From Bell WE, McCormick WF. Neurologic Infections in Children . Philadelphia: Saunders; 1975.)

Rubella: cranial ultrasound.

Coronal (A) and parasagittal (B) scans showing lenticulostriate vasculopathy in both thalami (A) and calcification within the corpus callosum (B).

Microcephaly and Impaired Myelination.

Two additional features of congenital rubella (i.e., microcephaly and impaired myelination) may relate to the effect of the virus on cellular replication. Microcephaly does not appear to be accounted for readily by destructive disease and, indeed, is often not prominent until months after birth. The possibility that the decreased brain mass is related to a decrease in the number of neurons and glia is supported by the observations that rubella disturbs mitotic activity of human fetal cells in culture and also causes a reduced number of cells in a variety of organs in affected infants. In addition to the microcephaly, a cellular deficit may account for the moderately impaired myelination observed by Rorke and Spiro and by Kemper et al. Indeed, although quantitative data are lacking, an apparent decrease in oligodendrocytes has been observed in association with the delay in myelination.

Incidence of Clinically Apparent Infection.

The devastating rubella pandemic of the mid-1960s made it possible to define the enormous clinical spectrum of congenital rubella. Because the largest portion of available data was derived from studies of infants identified at birth, the spectrum of manifestations as a function of maternal infection has been more difficult to define than for congenital CMV infection or toxoplasmosis. Nevertheless, it does appear that the likelihood of asymptomatic congenital rubella infection is more nearly comparable to that of congenital toxoplasmosis than to that of congenital CMV infection. Thus approximately two thirds of patients are asymptomatic in the neonatal period. However, most of these infants do develop evidence of disease in the first several years of life, a finding imparting clinical significance to observations that prolonged viral replication is an important feature of this disease.

Clinical Features.

The clinical features in symptomatic patients are shown in Table 34.20 . Intrauterine growth retardation (followed by postnatal growth failure) is a particularly common feature. Disturbances of the reticuloendothelial system are also prominent and are characterized particularly by hepatosplenomegaly and thrombocytopenia with or without purpura. The purpura should be distinguished from the peculiar dermal erythropoiesis that results in the small purple lesions of the “blueberry muffin” syndrome. Cardiovascular defects are characteristic and consist principally of peripheral pulmonary stenoses and patent ductus arteriosus. Myocardial injury can be demonstrated in a few patients by abnormal electrocardiographic findings as well as pathologically, and it may contribute to the occurrence of congestive heart failure. Other lesions, apparent in 20% to 50% of patients, are linear areas of radiolucency of the metaphyses of long bones (i.e., celery stalk lesions ), prominent especially around the knee, and interstitial pneumonitis. These abnormalities usually subside in the first few months of life.

Neurological phenomena in the newborn period are prominent in approximately 50% to 75% of cases (see Table 34.20 ). The most common manifestations relate to meningoencephalitis, seen most clearly in most patients by elevated levels of CSF protein and mononuclear cells. The anterior fontanelle is full in 25% to 50% of patients. The most common initial neurological features are “lethargy” and hypotonia, accompanied and followed shortly by prominent irritability. The irritability may relate to meningeal irritation, which probably also accounts for the occurrence of retrocollis and opisthotonos. These signs of meningeal irritation may worsen in the first weeks or months of life. Seizures appear in approximately 10% to 15% of infants. Definite microcephaly is unusual at birth. Most of the acute clinical features subside over the first several months and evolve to the sequelae outlined subsequently.

The ocular lesions consist principally of cataracts, usually white or pearly, especially centrally, and chorioretinitis, which may be more common than was previously appreciated. Indirect ophthalmoscopy is especially helpful to demonstrate the characteristic spotty pigmentation (i.e., “salt and pepper”) appearance, which may be particularly prominent peripherally. Microphthalmia is sometimes difficult to appreciate when it is bilateral and is associated particularly with cataract.

The auditory lesion may be difficult to demonstrate in the newborn, although the application of brain stem evoked response audiometry (BERA) has improved detection. In one series, approximately 20% of infants had suspected or definite hearing loss by behavioral testing in the neonatal period. The basis for the hearing loss in congenital rubella is a cochlear inflammatory and destructive lesion. A significant minority of infants will subsequently exhibit disturbances in response to sound that appear to be on a central basis, although the locus of this central pathology is unclear. In a detailed study of hearing loss in children with congenital rubella, the hearing deficit was usually uniform over all frequencies, symmetrical, and severe (mean threshold, 93 dB). As many as 60% to 80% of infants with congenital rubella are found later to have hearing loss as children. This increase in incidence from early infancy to childhood relates to a combination of inadequate testing in infancy with delayed diagnosis and progression of disease in the auditory apparatus; the relative importance of each of these factors remains unclear.

Clinical Diagnosis.

Clinical features that favor the diagnosis of congenital rubella are intrauterine growth retardation, CSF pleocytosis, salt-and-pepper chorioretinopathy, cataracts, cardiovascular defects, and skeletal lesions. The absences of prominent cerebral calcification, hydrocephalus, overt microcephaly, and vesicular rash are the clinical features that best differentiate congenital rubella from congenital toxoplasmosis, CMV infection, and HSV infection. Following a recent outbreak in Vietnam, a peak incidence of 7.8 per 1000 live births was seen, and 38 infants could be studied in detail. Low birth weight (71%), cardiovascular defects (72%), suspected hearing impairment (93%), hepatosplenomegaly (68%), thrombocytopenia (76%), and developmental delays (73%) were noted. Fully 84% of the patients presented with characteristic hemorrhagic purpuric eruptions, the “blueberry muffin baby” syndrome. Twenty-four of the infants (67%) had a significant persistent ductus arteriosus, and notably this finding was associated with pulmonary hypertension in 16 of the 24 infants. Thirteen infants (34%) died. Pulmonary hypertension, hepatosplenomegaly, and severe thrombocytopenia were more frequently observed among those who died.

Isolation of Virus.

Determination of rubella virus as the responsible microbe depends particularly on isolation of the virus and serological tests. The virus can be isolated best from the nasopharynx and urine, but it can also be isolated from stool and CSF (and various tissues, including lens and brain). Approximately 55% to 85% of patients exhibit positive cultures. Performance of multiple cultures increases the yield appreciably. The virus has been isolated from approximately 30% to 45% of CSF samples examined. This relatively high frequency of isolation of virus from CSF is unlike other congenital viral encephalitides. The chronicity of rubella infection is emphasized by the finding of positive CSF cultures in infants as old as 18 months. A similar conclusion can be derived from isolation of virus from the cataractous lens of a child aged 2 years, 11 months. Moreover, as many as one third of infants with congenital rubella are still excreting virus at 8 months of age.

Serological Studies.

All women of childbearing age should have been vaccinated against rubella as children or before conception. Women who have been vaccinated should be considered immune. As seroconversion is not 100%, serological testing is indicated in vaccinated women who have a known exposure or a rash and illness consistent with rubella to rule out acute primary infection or reinfection. Serological diagnosis during pregnancy may be based on maternal blood studies. If negative for IgM (IgM−), the IgG results determine if the woman is seropositive (immune) or seronegative (not immune). If a pregnant woman is IgG-negative at the first visit, she should be retested monthly for seroconversion until the end of the fifth month of pregnancy. If the maternal blood is positive for IgM (IgM+) and IgG (IgG+), the next step would be an IgG avidity assay on the same blood sample to estimate the time of infection, with low avidity indicative of recent primary infection. The same tests should be repeated on a second blood sample obtained 2 to 3 weeks later. If the results remain the same (IgM+ IgG−), the IgM result is considered nonspecific, indicating that the woman has not been infected; however, she is seronegative and should be followed until the end of the fifth month. If the woman has seroconverted (IgM+ IgG+), recent primary infection is confirmed and a prenatal evaluation should be performed, followed by a discussion to see whether the woman wishes to continue her pregnancy. After birth, the fastest and most useful test is determination of IgM-specific antibody; this is the most definitive serological diagnostic test in the first few weeks of life. IgM antibody can be detected in the infant’s cord blood or serum and persists for about 6 to 12 months.

Neurodiagnostic Studies.

No specific neurodiagnostic tests are available, although a high rate of viral isolation from the CSF and the frequency of CSF signs of inflammation are very helpful in diagnosis. CT and ultrasound are useful in the detection of the focal areas of ischemic necrosis secondary to the vasculopathy and the less common calcification in the basal ganglia. CT scans obtained between 1 and 3 years of age have demonstrated cerebral white matter hypodensity and multiple calcified nodules in the centrum semiovale, presumably reflecting the impaired myelination and focal ischemic lesions described earlier in the section on neuropathology. MRI is most useful for detection of the presumed ischemic lesions in cerebral white matter and impairment of myelination. Cranial ultrasound may show focal areas of calcification, subependymal cysts, and echogenic vessels in the basal ganglia and thalamus (see Fig. 34.19 ). In a series of 12 infants with echogenic vessels in basal ganglia, 2 patients had congenital rubella.

Relation to Neonatal Clinical Syndrome.

Although outcome is related to neonatal clinical features, the relationships are not as obvious as with congenital CMV infection and toxoplasmosis. This fact may relate to the particular chronicity of congenital rubella as well as to the relative infrequency of completely asymptomatic neonatal disease. In a population of 100 carefully studied infants, 90% of whom were overtly symptomatic in the neonatal period and very early infancy, only 9% appeared to be free of deficits at 18 months ( Table 34.21 ). Neuromotor deficits (i.e., spastic motor deficits and delayed neurological development) were severe in approximately 50% of the patients. Fully 81% of these infants had microcephaly, and 72% had definite hearing loss or other apparent disturbances related to auditory perception. In a subsequent report on the same population, of patients followed to 16 to 18 years of age, 28% exhibited mental retardation, and an additional 25% exhibited low-average intelligence. Of 14 children with suspected hearing loss at 18 months, 13 were definitely hearing impaired. In another prospective series of infants with congenital rubella, approximately similar outcomes were observed; approximately 45% of such infants exhibited psychomotor retardation , and 50% of these infants were moderately or severely affected.

Long-Term Hearing Deficits and Other Sequelae.

Even those infants who appear to be less severely affected often show evolution of disabling auditory, motor, behavioral, and learning deficits as they grow older ( Table 34.22 ). A multidisciplinary longitudinal study of 29 nonretarded infants with congenital rubella demonstrated definite hearing loss in 1 infant in the first 2 months, in 12 infants by 12 months, in 22 infants by 24 months, in 25 infants by 48 months, and in an additional 2 infants by 11 years, for a total of 27, or 93% of the children. This accretion of patients with definite hearing loss may reflect continuing cochlear injury. The analogy with the delayed onset of hearing loss in congenital CMV infection and of chorioretinopathy in congenital toxoplasmosis is apparent.

In the longitudinal study just mentioned, early disturbances of motor development and of tone were followed by impairments of motor coordination and muscle weakness in approximately 50% of the children. Behavioral disturbances, which in the early years included impaired attention span and hyperkinesis, evolved to emotional irritability and persisting distractibility in approximately 50%. A propensity for congenital rubella infection to lead to impairment of behavioral and emotional development is also apparent in other studies by a 6% incidence of subsequent autism. Moreover, although IQ scores remained within the normal range in the study of Desmond et al., learning deficits and visual-perceptual-motor deficits were prominent in approximately 50%. These abnormalities had major impacts on the children’s adaptation to educational and home environments and underscore the necessity for careful follow-up and appropriate interventions in infants with congenital rubella.

An interesting relationship between the rate of linear growth and cognitive outcome was apparent in a 20-year follow-up of 105 cases of CRS. Children with normal growth had normal cognitive development, and those whose linear growth was at less than the fifth percentile exhibited moderate to severe mental retardation.

Late Progressive Panencephalitis.

A rare complication of congenital rubella, the precise frequency and importance of which are not clear, is the occurrence of progressive panencephalitis with onset, usually in the second decade, of intellectual and motor deterioration, elevated CSF protein levels, and elevated antibody titers to rubella virus in serum and CSF. The virus has been isolated from the brain. Whether this disorder represents reactivated infection, bears a relation to the role of measles virus in subacute sclerosing panencephalitis, or both remains to be determined.

Prevention.

Preventive measures present the realistic hope of eradicating CRS. Of the three major approaches to prevention (avoidance of maternal infection, treatment of maternal infection, and abortion in the presence of maternal infection), the first has been accomplished in large part through vaccination.

Active immunization with a live attenuated rubella vaccine has been accomplished by two major approaches. In the United States, mass vaccination of all children aged 1 year to puberty has been used to limit the spread of infection to the pregnant woman by curtailing circulation of virus in the community. In the United Kingdom and in many other European countries, selected immunization, especially of girls from ages 11 to 14 years, was used initially to provide protection for the childbearing years. Mass vaccination of all children in the second year of life was instituted in the United Kingdom in 1988. The policy in the United States has been effective; the incidence of congenital rubella declined by approximately fivefold in the decade following the initiation of this vaccination regimen. Since 2004, the rubella incidence has remained below 1 case per 10 million population, and the incidence of the CRS has been below 1 case per 5 million births. However, some questions still remain concerning how long vaccine-induced protection will last or whether inapparent reinfection of the mother with transmission to the fetus may occur. In 1996, it was estimated that 110,000 infants with CRS were born annually in developing countries. In 2000, the World Health Organization (WHO) published the first rubella vaccine position paper to guide introduction of RCV in national childhood immunization schedules. From 1996 to 2009, the number of countries that introduced RCV into their national routine childhood immunization programs increased by 57%, from 83 countries in 1996 to 130 countries in 2009. Also during this time period, the number of rubella cases reported decreased from 670,894 in 2000 to 121,344 in 2009. Rubella control and prevention of CRS can be accelerated by integrating with current global measles mortality reduction and regional elimination activities. All women identified to be seronegative during pregnancy should be vaccinated postpartum.

Passive immunization with immune globulin may be useful in the special case of a susceptible pregnant woman exposed to rubella. The effectiveness of this approach is controversial, but it is necessary to recognize that passive immunization is useful to prevent viremia and fetal infection and therefore must be given promptly.

Abortion in the woman infected with rubella requires understanding of the risks of fetal infection as they relate to the timing of the infection in pregnancy. The demonstration of prenatal diagnosis by fetal blood sampling in the 20th week of gestation may help to prevent abortion of the unaffected fetus.

Supportive Therapy.

Supportive therapy is carried out principally as described for congenital CMV infection. In addition, recognition and prompt control of cardiac failure are critical, particularly in view of cerebral vasculopathy and therefore already compromised cerebral perfusion. Careful auditory assessment, with brain stem evoked response audiometry as well as with behavioral studies, is critical to detect hearing loss and to provide appropriate intervention as early as possible. Similarly, detection of cataract is important because delay of surgery into the second and third years of life prevents useful vision. Opinions differ about the optimal time of therapy. Nevertheless the infant with auditory and visual deficits is at great risk for subsequent disturbances of language and other aspects of neurological development, and the earliest interventions regarding vision and audition are critical.

Antimicrobial Therapy.

No known effective chemotherapeutic agents of value exist in the treatment of congenital (or postnatal) rubella infection.

Parturitional and Ascending Infection.

Most infants with neonatal HSV infection acquire the infection during passage through an infected birth canal near or at the time of birth (85%). The markedly higher rate of neonatal HSV infection in infants born to mothers shedding virus at delivery when delivery is by the vaginal route than by cesarean section is consistent with this notion ( Table 34.23 ). In the classic case, the virus is acquired by direct contact of the infant’s skin, eye, or oral cavity with the virus in the mother’s birth canal. Overt maternal herpetic lesions are uncommonly present. Because of the importance of direct contact, it is understandable that the vesicular lesions of the infant’s infection are usually over the scalp and face in cephalic presentations and over the buttocks in breech presentations. Vivid demonstrations of the relation of contact with the sites of herpetic lesions are provided by several reports of vesicular lesions (and serious neonatal disease) at the sites of placement of fetal scalp electrodes for intrapartum monitoring. In a large study of women with asymptomatic HSV infection at delivery, fetal scalp electrodes were used in 25 infants, of whom 5 developed neonatal herpes, versus zero cases of neonatal herpes in the 25 deliveries in which fetal scalp electrodes were not used. A larger study later confirmed the role of invasive monitoring. The risk of transmission to the newborn is more than 10-fold greater when the mother is shedding HSV-1 versus HSV-2 at delivery (see Table 34.23 ). The five most important factors known to have an effect on transmission of HSV from mother to neonate are type of maternal infection (primary vs. recurrent), maternal antibody status, duration of rupture of membranes, integrity of mucocutaneous barriers (use of scalp electrodes, for instance), and mode of delivery (cesarean vs. vaginal).

Large-scale studies of neonatal HSV infection in relation to the type of maternal infection have been of great interest. It is clear that in most cases neonatal infection occurs in children of asymptomatic rather than symptomatic mothers. Moreover, the serological status of the mother is a critical determinant of the risk of neonatal infection ( Fig. 34.20 ). It is important to make a distinction between a primary and a recurrent infection. A first-episode primary HSV infection occurs when a person with no prior HSV-1 or HSV-2 antibodies acquires either virus in the genital tract. If a person with preexisting HSV-1 antibodies acquires an HSV-2 genital infection, a first-episode nonprimary infection ensues. A recurrent infection can occur when reactivation of the virus occurs, affecting the skin and mucosal membranes. Of 177 women with positive HSV cultures within 48 hours of delivery, 26 (15%) had evidence of a recently acquired first episode of infection. Nearly one third of these pregnancies resulted in neonatal HSV infection, and fully 80% of neonatal HSV infection in the entire cohort occurred in this subset of infected women . No difference in risk of transmission was noted in the first-episode group whether the infection was primary (mother had no antibodies to HSV) or nonprimary (mother had antibodies to HSV-1 with HSV-2 infection or to HSV-2 with HSV-1 infection). In the large group of women ( n = 151) with reactivated infection, or 85% of the total group, the risk of transmission of neonatal HSV infection was very low (i.e., only ≈1%). Although most of the reactivated maternal infections involved HSV-2, the risk of transmission of virus to the infant was confined to the small group with reactivated HSV-1 (see Table 34.23 ). Indeed, none of 140 reactivated maternal HSV-2 infections resulted in neonatal HSV, whereas 2 of 11 reactivated maternal HSV-1 infections were transmitted.

Neonatal herpes simplex virus (HSV) infection in relation to asymptomatic maternal infection at the time of labor.

Type of maternal infection and risk of HSV transmission to the neonate. A total of 39,949 samples from women without clinical evidence of genital HSV infection were cultured within 48 hours of delivery; 121 were found to be shedding. For the latter, sera were available for analysis. Outcomes for these 121 women are illustrated.

(Modified from Brown ZA, Wald A, Morrow RA, et al. Effect of serologic status and cesarean delivery on transmission rates of herpes simplex virus from mother to infant. JAMA. 2003;289:203-209. See also Kimberlin 2005; reference 292.)

Symptomatic primary infection is characterized by fever, pain, and vesiculoulcerative lesions of the vagina and cervix. The risk of infection for a newborn delivered vaginally in the presence of clinically visible maternal genital infection in an antibody-negative woman may approach 50% to 60% (a much higher risk than in the much more common circumstance of an antibody-positive woman), a 25% risk in infants born to women with first-episode nonprimary infection, and a 2% risk increase in those born to women with recurrent genital herpes.

Ascending infection of the fetus can occur after rupture of the membranes. Indeed, available data suggest that the risk of fetal infection in the presence of clinically visible maternal genital infection is greater by the ascending route than by parturitional contact (see Table 34.24 ). This finding may relate to a larger inoculum of virus and exposure at multiple sites with ascending versus parturitional infection. The current rarity of this clinical situation is illustrated by a study of 58 cases of neonatal HSV infection in which only 1 mother had intrapartum genital lesions.

Fetal (Transplacental) Infection.

Prenatal acquisition of HSV through transplacental mechanisms is a rare cause of fetal infection (5%). Isolated examples have been reported in single case reports. However, in a study of 155 cases of neonatal HSV infection, 8 (5%) had evidence of acquisition of infection during early gestation and all were premature. Clinical manifestations have included skin lesions at birth, chorioretinitis, microphthalmia, microcephaly, hydranencephaly, multicystic encephalomalacia, cerebral calcifications, and other CNS abnormalities evident on CT scans. The clinical features of intrauterine HSV infection as a function of estimated time of acquisition of the virus are summarized in Table 34.25 . A definable clinical triad includes (1) cutaneous findings (active lesions, scarring, aplasia cutis, hyperpigmentation or hypopigmentation), (2) neurologic findings (microcephaly, intracranial calcifications, hydranencephaly), and (3) ocular findings (chorioretinitis, microphthalmia, optic atrophy).

In a recent review of 64 cases of intrauterine herpes infection, more than two thirds did not have this triad and the cutaneous lesions were the most common presentation. Interestingly the cutaneous lesions seen were not restricted to vesicles or bullae and included aplasia cutis in three. Two thirds ( n = 43) had CNS manifestations and 29 of the 43 had more than one neurological finding. Ocular findings were found in 25 (39%), mostly retinal disease (18) as well as microphthalmia (4) and cataracts (4). Subsequent outcome was poor, with death in 29 (45%), including 4 intrauterine deaths. Outcome was not reported in all, but 13 were said to have developmental delay and 8 were doing well at the age of 6 months.

Postnatal Infection.

Postnatal acquisition of HSV by the newborn is an uncommon but documented occurrence (10%) and is almost always due to HSV-1. Of 24 infants described in a review as having acquired infection shortly after birth, 13 acquired the infection from mothers with oral herpetic lesions, 9 from other adults (including hospital personnel), and 2 from other infected infants. As in other examples of HSV infection of the newborn, the infection was serious, and 67% of the children died. These data have important implications for the prevention of exposure of the newborn to sources of HSV.

Role of Host Factors.

Host factors must play some role in explaining the malignancy of HSV infection in the perinatal period. Indeed, in older patients who are immunologically competent, severe disseminated disease is rare. The likelihood of neonatal disturbances in response to HSV infection has been delineated. Defects in the infant’s response to herpes simplex infection can be divided into those response mechanisms involved in the initial “containment” phase, during which the virus is localized to a limited anatomical area, and those mechanisms active in the later specific curative phase, during which localized infection is eliminated. Defects in the initial containment phase involve the operation of the alveolar-macrophage barrier, expression of natural killer cytotoxicity, and production of interferon-alpha and tumor necrosis factor. Defects in the later elimination phase involve antibody production, both of the neutralizing type and that responsible for antibody-dependent cellular (leukocyte) cytotoxicity; T-cell proliferation; and the production of interferon-gamma.

Although the two serotypes of HSV are genetically closely related and share many clinical features, their ability to modulate host responses can differ substantially. The systemic inflammatory response was evaluated in a study of 19 infants with HSV infection, 9 of whom had a DIS. Concentrations of inflammatory cytokines and markers of apoptosis were noted to be significantly higher in infants with disseminated HSV infection and were correlated with HSV load. Toll-like receptors (TLRs) constitute a family of innate immune receptors that recognize and respond to a wide spectrum of microorganisms, including viruses (see Chapters 13 and 14 ). These receptors are located on the plasma membrane or intracellular membranes. TLRs 3, 7, 8, and 9 are localized almost exclusively on intracellular membranes and are activated by nucleic acids of viral and bacterial origin. TLR3 is triggered by viral double-stranded RNA, leading to the activation of specific transcription factors that stimulate the production of antiviral interferons and other cytokines. Human TLR3 deficiency has been associated with childhood herpes simplex encephalitis (HSE), with mutations in TLR3 and TLR3 pathway genes identified in a number of patients, in particular those with HSE recurrence. Cord blood natural killer (NK) cells have been reported to have deficient TLR3 expression, which might explain why newborns are especially sensitive to neonatal HSV infections. Ligands for TLR3 (and TLR9) induce potent innate immune antiviral responses against herpes simplex virus type 2 (HSV-2). Among these responses is induction of the expression of type I interferon genes, which are cytokines known to inhibit viral replication. Production of beta interferon but not of alpha interferon, gamma interferon, or tumor necrosis factor alpha was found to be correlated with innate immune protection against HSV-2.

Meningoencephalitis.

Meningoencephalitis is characterized by (1) inflammatory cells in the meninges, (2) perivascular infiltrates with inflammatory cells, (3) severe multifocal necrosis of all cellular elements of brain parenchyma, often with some degree of hemorrhage, (4) reactive microglial and astroglial proliferation, and (5) occurrence of Cowdry type A intranuclear inclusions in neuronal and glial, especially oligodendroglial, cells. The nucleus containing the inclusions is characteristically distorted, with clumping of nuclear chromatin and undulation of the nuclear membrane. These pathological findings are often accompanied by a considerable degree of brain swelling, and hemorrhage in the areas of necrosis may occur, in part because of associated endothelial involvement. A detailed MRI study of 12 infants with neonatal HSV encephalitis showed multilobar cerebral involvement, including temporal and frontal regions and deep gray matter structures.

Neuropathological Sequelae.

The result of HSV infection of the perinatal brain most commonly is a devastating effect on neural structure and function. Subsequent failure of brain growth and microcephaly (after the neonatal period) are the rule. Multicystic encephalomalacia has been documented repeatedly ( Fig. 34.21 ). Indeed, the destruction may be so complete that hydranencephaly is the result. These lesions can be demonstrated readily by cranial ultrasound and especially MRI (see later).

Neonatal herpes simplex infection: multicystic encephalomalacia in an infant who died at 24 weeks of age.

Onset of recognized disease at approximately 21 days of age with seizures was followed by rapid progression to electrocerebral silence at 34 days of age and subsequent vegetative state. (A) Left lateral view of brain showing thickened covering membranes and severe destruction of cerebral hemispheres. (B) Coronal section of cerebral hemispheres showing parenchymal destruction with cavitation. (C) Photomicrograph of coronal section of cerebral hemispheres stained with hematoxylin and eosin. Note destruction of cerebral cortex and subcortical white matter and replacement by astrocytic glial stroma and dramatic masses of foamy macrophages. The macrophages are particularly striking in the superior aspects of the cerebral convexity. (D) Higher-power view showing the cavitation, necrotic debris, and masses of macrophages, especially around blood vessels.

(From Young GF, Knox DL, Dodge PR. Necrotizing encephalitis and chorioretinitis in a young infant: report of a case with rising herpes simplex antibody titers. Arch Neurol . 1965;13:15-24.)

Incidence of Clinically Apparent Infection.

Neonatal HSV infection is distinctive among the diseases caused by organisms of the TORCH complex in the essentially uniform occurrence of symptomatic disease. The clinical spectrum varies from infections localized to a few vesicles on the skin to those involving dissemination to every major organ ( Table 34.27 ). ^a

a References .

A major distinction is made between disseminated and localized HSV disease. Disseminated disease is associated with evidence of involvement of multiple systems, particularly the reticuloendothelial system. Hepatoadrenal necrosis is the hallmark of the disorder. Localized disease is associated with involvement confined to a single site; if multiple sites are involved, the term localized is still used if the reticuloendothelial system and other visceral organs are not included. A distinction is made between disease localized to the CNS (a serious form) and that localized to the skin, eye, or mouth—that is, SEM disease (a less serious form). Ten or more years ago, disseminated disease accounted for approximately 40% to 70% of all cases. More recently, the approximate distribution of clinical types has been as follows: disseminated disease, about 25%; localized CNS disease, 30%; and localized SEM disease, 45%. Overlap of the latter two localized forms is not uncommon. The relative decrease in disseminated disease and the increase in SEM disease appear to relate to earlier diagnosis and the use of antiviral therapy.

Clinical Features of Disseminated Disease.

The early signs of disseminated HSV infection occur in most cases by 10 to 12 days of life. ^a

a References .

Disseminated HSV infection may involve multiple organ systems including the liver, lungs, adrenals, gastrointestinal tract, and the SEM. Hepatomegaly, hyperbilirubinemia, and bleeding are common (see Table 34.28 ). The bleeding relates to a combination of hepatic disease and, often, disseminated intravascular coagulation, and it may be very severe. It is important to realize that more than 20% of neonates with disseminated HSV infection do not develop any cutaneous vesicles during their illness; the vesicles often appear after the clinical onset of the disease and evolve from macules to papules before the vesicles form, which often resemble pustules. In a recent study, only 2 of 13 infants (15%) who died had skin lesions at presentation. Disseminated HSV infection is a devastating disorder (see Table 34.27 ); before the era of therapy, approximately 80% of infected children died, frequently within a few days, and approximately 50% of the survivors exhibited serious neurological sequelae, predictable on the basis of the neuropathology (see the previous section). Mortality has declined to approximately 29% and morbidity to 17% since the introduction of antiviral therapy (see Table 34.27 ).

Clinical Features of Localized Disease.

Localized HSV infection is characterized by the absence of clinical or laboratory evidence of visceral involvement. The sites most commonly affected are the CNS and SEM (see Table 34.29 ). The age of onset of localized involvement of the CNS is later than for disseminated disease (day 16 to 19). Indeed, because relatively more term infants exhibit localized CNS disease than disseminated disease, the infant with this variety of neonatal HSV infection has usually been discharged from the hospital before the illness begins. The usual signs are stupor and irritability, which evolve to seizures (often focal) and, perhaps, coma. As with disseminated disease, many infants (≈35%) do not exhibit mucocutaneous lesions. CSF pleocytosis and elevated protein are characteristic; depressed CSF glucose is also common. The outcome is unfavorable but better than with disseminated disease. Before antiviral therapy, approximately 50% of these infants died, a rate that has declined to 4% following the introduction of antiviral therapy. Similarly, whereas about 70% to 80% of survivors exhibited serious neurological sequelae before antiviral therapy, this number has declined to about 40% (see Table 34.27 ). The morbidity among infants with CNS HSV infection is higher among those with HSV-2 infection than among those with HSV-1 infection and may include developmental delay, epilepsy, blindness, and cognitive disabilities.

HSV infection clinically localized to the SEM usually presents in the second week of life (days 10 to 12). The progression from macules to vesicles occurs over 24 to 48 hours, often at sites of trauma (e.g., site of scalp electrodes or presenting body part). The vesicles may be obscured by overlying hair. The presence of a vesicle in a newborn should be seen as indicating HSV infection until ruled out as soon as possible by appropriate diagnostic studies (see later). Progression to CNS or disseminated disease occurs in 75% of untreated cases. Indeed, clinical localization of lesions to the SEM does not imply that the CNS is not also affected (see Table 34.27 ). Before antiviral therapy, approximately 30% to 40% of such patients exhibited neurological sequelae on follow-up. Currently, with therapy, sequelae are rare.

Clinical Diagnosis.

Of the congenital infections, neonatal HSV infection may be the most distinctive. The presence of a vesicular rash, keratoconjunctivitis, seizures, tense anterior fontanelle, and CSF evidence of meningoencephalitis is characteristic. However, the skin and ocular manifestations are present at the onset of the illness in only the minority of cases. Even in disseminated HSV, classic signs may not be present on admission. Of 49 infants (16%) with HSV, 18 lacked classic signs at hospitalization, 3 had disseminated disease, 4 had CNS involvement, and 1 uncategorized ; most of these 18 infants developed signs suggestive of HSV within 24 hours. The majority (84%) presented with seizure, vesicular rash, or critical illness on admission.

Additional information of value is epidemiological, regarding HSV infection in the mother or her sexual contact. Important negative differential diagnostic information includes the rarity in the neonatal period of microcephaly, hydrocephalus, intracranial calcification, and cardiovascular defects.

Cytological Techniques.

Identification of HSV infection is based principally on cytological techniques, isolation of the virus, or detection of viral DNA. A high index of suspicion of the disease is crucial. Cultures may be taken before any clinical symptoms—for instance, in the presence of genital herpes noted during delivery. However, only 20% to 40% of mothers whose children develop HSV infection have had symptomatic genital herpes or sexual contact with a partner with recognized HSV infection during or before the pregnancy. Certainly any infant with a vesicular lesion should be considered to have an HSV infection and should be evaluated appropriately. However, nonspecific signs (e.g., lethargy, cessation of feeding, or other features suggestive of bacterial sepsis) should raise the possibility of neonatal HSV infection. Cytological techniques are readily available and are a rapid means of establishing a presumptive diagnosis of neonatal HSV infection. Scrapings can be obtained from the base of vesicular lesions of the skin or oral cavity or from conjunctival lesions. Smears are fixed in alcohol and stained immediately, according to the Papanicolaou method. The typical morphological changes are multinucleated giant cells and intranuclear inclusions. Viral particles may also be observed by electron microscopic examination of material from lesions and from urine and CSF. HSV antigens have also been detected in CSF leukocytes through the use of immunofluorescence techniques.

Isolation of Virus and Detection of Viral DNA.

Isolation of the virus is a definitive means for establishing the diagnosis and is best accomplished from observable lesions, but isolation is also possible from throat, stool, urine, and CSF. In one series, a pharyngeal swab was the source with the highest detection rate (79%). In appropriate media, samples can be transported at room temperature. Cytopathic changes in inoculated tissue cultures are usually detectable within 1 to 3 days. The virus not uncommonly is difficult to isolate from CSF very early in localized CNS disease.

A major advance in the diagnosis of neonatal HSV infection is application of PCR to amplify the very small quantities of viral DNA , which then can be detected by conventional methods (e.g., DNA hybridization). PCR assay has been shown in the newborn to be clearly superior to culture and is highly sensitive and specific for CNS involvement when CSF is studied. However, HSV DNA in CSF may be undetectable in 30% of cases early in the disease, and serial sampling is required to reach nearly 100% sensitivity in HSV encephalitis. In a recent study, HSV PCR results were obtained from plasma ( n = 47), CSF ( n = 56), or both ( n = 40) at the time of diagnosis in 63 infants with SEM ( n = 26), CNS ( n = 18), or disseminated disease ( n = 19) ( Table 34.30 ). Plasma HSV PCR was 100% positive only in infants with disseminated disease. Even in infants with CNS infection, only 72% had positive CSF PCR. This discrepancy could be explained by the pathogenesis of neonatal HSV CNS disease, which is thought to be neuronal transport to the brain resulting in an initially localized encephalitis before involvement of the meninges. Thus early sampling of the CSF may not detect HSV and repeating a lumbar puncture should be considered when the diagnosis is suspected. Mean plasma viral level was 2.8 log10 copies per milliliter in SEM, 2.2 log10 copies per milliliter in CNS, and 7.2 log10 copies per milliliter in infants with disseminated disease. The HSV levels were higher among infants who died compared with surviving infants, 8.1 log10 copies per milliliter (range 7.7 to 8.6) versus 3.8 log10 copies per milliliter (range 0.0 to 8.6), P = .001. However, the level of HSV DNA in the CSF or in plasma did not correlate with neurological outcome.

Importance of Maternal Evaluation.

Isolation of the virus from mothers (or their sexual contacts) with genital infection is a valuable adjunct to diagnosis of the neonatal disorder. Because most cases of genital herpes in pregnant women are subclinical, cultures, PCR assays, or both are usually necessary to establish the diagnosis. Viral cultures of genital secretions and serological studies of maternal serum by western blot are the most convenient and effective approaches. Detection of viral DNA in genital specimens by PCR is a highly sensitive approach, but its value for routine use in obstetrics is not yet established.

Serological Studies.

Serological studies for the diagnosis of neonatal infection with HSV are less useful than for other congenital infections. This finding relates to the masking of the infant’s own IgG response by passively transferred maternal antibody and the 1- to 2-week delay in the rise in specific IgM antibody generated by the infant. There are two type-specific antibody assays that allow a distinction between HSV types 1 and 2. Specific HSV IgM antibody can be detected rapidly by an immunofluorescence test. Because the antibody persists for 6 to 12 months, this test is useful in survivors of the neonatal infection.

Neurodiagnostic Studies.

Neurodiagnostic studies of value include particularly examination of the CSF, electroencephalography (EEG), and brain imaging. Brain biopsy is no longer considered for making a diagnosis.

The CSF exhibits the findings of meningoencephalitis (i.e., pleocytosis and elevated protein content). Polymorphonuclear cells are occasionally predominant; in severely affected cases, the pleocytosis includes many red blood cells. The CSF glucose level may be depressed. In one series, mean CSF glucose in the first week of the disease was 39 mg/dL; in the second week, it was 32 mg/dL; and in the third to fifth weeks, it was 28 mg/dL. Protein level is elevated consistently and often exceeds 100 to 150 mg/dL as the disease progresses. Any infant with a CSF formula suggestive of encephalitis (i.e., pleocytosis and elevated protein) should be considered to have HSV encephalitis until proven otherwise. The virus can be isolated from the CSF, but as noted in the previous subsection, the cultures are often negative early in the disease. PCR assay clearly is the optimal method to identify the virus in CSF rapidly. As noted earlier, the sensitivity early in the disease is approximately 50% to 75% and increases to nearly 100% with later CSF samples.

CT was formerly used to demonstrate the extent and severity of the brain injury ( Fig. 34.22 ). Progression of abnormalities to multicystic encephalomalacia has readily been documented by serial studies. Cranial ultrasound is also useful in the detection of parenchymal changes and evolution into multicystic encephalomalacia, but it is likely to miss cortical and brain stem injury. MRI is the most useful approach for the identification of parenchymal lesions. Diffusion-weighted imaging (DWI) MRI is the most sensitive imaging modality for the early detection of CNS disease ( Fig. 34.23 ). The rapid evolution to multicystic encephalomalacia is clearly delineated by MRI ( Figs. 34.24 and 34.25 ). In a relatively large retrospective study of 29 infants with neonatal HSV infection, bilateral multilobar ( n = 8), pontine ( n = 3), thalamic ( n = 6), or internal capsule and corticospinal tract ( n = 5) involvement was noted on MRI. Additional information, not yet apparent on conventional T1- and T2-weighted imaging, was seen in 6 infants who had DWI. As might be expected, neurodevelopmental outcome was found to correlate with MRI abnormalities. Notably, somewhat reminiscent of adult HSV encephalitis, the temporal lobe was involved in 67% of cases, and in 25% this cerebral structure was exclusively involved.

Neonatal herpes simplex infection: multicystic encephalomalacia.

This computed tomography scan is from an affected 6-week-old infant who had onset of neurological signs at 7 days of age. Note the large lucent areas, representing regions of cystic necrosis, scattered throughout the cerebral hemispheres.

(Courtesy Dr. Charles Abramson.)

Magnetic resonance imaging: axial T2-weighted image (A) and diffusion-weighted image (B) showing restricted diffusion extending from the cortex to the posterior limb of the internal capsule (B), not seen on the T2 sequence (A). The child developed mild unilateral spastic cerebral palsy.

Neonatal herpes simplex infection: magnetic resonance imaging.

Axial T2-weighted image obtained on the second postnatal day (A) showed no definite abnormality (although the diffusion-weighted image exhibited abnormal signal in the basal ganglia). By 1 month of age (B), the axial T2-weighted image showed evidence of diffuse cerebral cortical and white matter injury with early cystic changes as well as lesions in putamen (P) and thalamus (T). At 6 months of age, an axial T2-weighted image (C) showed multicystic encephalomalacia, most prominent in the right frontal region.

(Courtesy Dr. Omar Khwaja.)

Neonatal herpes simplex infection: magnetic resonance imaging.

Axial apparent diffusion coefficient (ADC) map (A) 3 days after onset of symptoms and T2-weighted image at 2 months after disease onset (B), showing symmetrical involvement of the temporal lobes, initially seen as restricted diffusion (A) with subsequent cystic evolution.

The EEG usually shows striking and characteristic changes ( Fig. 34.26 ). These are principally focal or multifocal paroxysmal, periodic, or quasiperiodic discharges consisting of repetitive sharp slow-wave complexes. The occurrence of prolonged periods (1 to 2 minutes) of such discharges was associated with death or major neurological sequelae in 9 of 10 affected infants in a reported series. A normal EEG was associated with normal outcome in all five infants in another series. The EEG is one of the most sensitive noninvasive laboratory studies in the diagnosis of herpes infection ( Table 34.31 ). In the study of Mikati et al., the EEG was abnormal when CT and ultrasonographic studies were normal in the first 4 days of the illness in seven infants who had both EEG and an imaging study.

Electroencephalographic tracings from an infant with neonatal herpes simplex infection.

The tracings were obtained with two different electrode arrays. The upper two tracings were obtained from one array, and the lower two tracings from a second array. Foci of periodic lateralized epileptiform activity are apparent in both hemispheres.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Prosencephalic Development Encephalopathy of Prematurity: Pathophysiology Stroke in the Newborn Cerebellar Hemorrhage Passive Addiction and Teratogenic Effects Bacterial and Fungal Intracranial Infections

Stay updated, free articles. Join our Telegram channel

Join