Synonyms: None.

Definition: Congenital cytomegalovirus infection is definitively dia­gnosed when the virus is isolated from tissue culture. The diagnosis can also be made by histologic study of typical inclusion bodies in biopsy or autopsy specimens, or less reliably by serologic study. Ascites is diagnosed by ultrasound or physical examination.

Prevalence: Congenital cyto­megalovirus infection occurs in 0.2 to 2.2% of deliveries¹². Approximately 5% of congenitally infected infants have generalized involvement¹³.  We have found reference to eight cases in which ascites, in the absence of other signs of hydrops, was diagnosed in fetuses with cytomegalovirus infection^1-8. There have been an additional three cases where similar findings were present in neonates^9-11.

Etiology: The cause of ascites in congenital cytomegalovirus infection is not certain. Contributing factors may include low serum protein levels due to hepatic dysfunction and portal obstruction resulting from periportal inflammation¹⁴.

Pathogenesis: The exact mode of transplacental passage is uncertain. The virus replicates in fetal tissues producing inflammation, tissue necrosis, and organ dysfunction. Cytomegalovirus hepatitis in the neonate can present with an intense inflammatory response involving the portal triads. In these cases, lobular disarray, degeneration of hepatocytes, and cholestasis are also seen¹⁴.

Associated anomalies: Isolated ascites is an uncommon finding in fetuses with cytomegalovirus infection. More commonly observed manifestations with prenatal ultrasound are overt hydrops, cerebral ventriculomegaly, IUGR, and oligohydramnios^1,5,6.

Differential diagnosis: Since ascites is often the first manifestation of hydrops, the differential diagnosis for fetal ascites is essentially the same as with generalized hydrops. When hepatomegaly is present, as it was in our patient, primary liver disease or extramedullary hematopoiesis should be considered.

Prognosis: In the eleven previously reported fetal and neonatal cases where ascites was the principle manifestation of cytomegalovirus, three pregnancies were terminated^4,6,7, three fetuses were stillborn or died in the first year of life^8,10,11, three were alive but had significant sequelae^1,5 and two were reportedly normal^3,9. In two cases the ascites resolved prior to delivery^2,6; however, both showed evidence of other organ damage (Table I). In general, neonates with symptomatic cytomegalovirus infection do poorly, with a high neonatal mortality rate, and a high rate of neurologic handicap in survivors¹². Cytomegalovirus hepatitis is reversible in survivors¹³.

Recurrence risk: Maternal cyto­megalovirus infection may reactivate; hence there is a small theoretical risk of recurrence of fetal infection in subsequent pregnancies¹².

Management: Fetal antiviral treatment has not been reported. Paracentesis to remove fetal ascites may help to prevent pulmonary hypoplasia³.

MESH Cytomegalic Inclusion Disease BDE 0381 POS 3220 ICD9 771.1 CDC 771.100

Address correspondence to Douglas S. Richards MD, Department of Obstetrics and Gynecology, PO Box 100294, University of Florida College of Medicine, Gainesville, Florida, 32610-0294 Ph: 904-392-2894; Fax: 904-392-6994. ¶ Dept. of Pediatrics, § Dept. of Pathology

Introduction

Congenital cytomegalovirus (CMV) infection results in considerable neonatal morbidity and mortality. The diagnosis of fetal cytomegalovirus infection has been made on several occasions when a mother has seroconverted, or after the sonographic recognition of characteristic fetal abnormalities. The diagnosis has been confirmed with varying success by performing maternal or fetal serologic tests^4,6,15,16, or by culture of the amniotic fluid^1,4,6,17. We present a case which was unusual in that the only ultrasound findings were massive ascites, hepatomegaly, and polyhydramnios. Fetal liver failure and thrombocytopenia were diagnosed prenatally by fetal blood sampling. In spite of severe fetal involvement, maternal and neonatal serologic tests did not support a diagnosis of cytomegalovirus infection, neonatal urine culture was negative, and the diagnosis was confirmed only after delivery, at autopsy.

Case report

A 14-year-old black primigravida with limited prenatal care came to the hospital at 27 weeks for evaluation following a motor vehicle accident. She had no history of hepatitis or other recent infectious diseases, and had no significant drug exposure. On physical exam she showed no evidence of trauma; however, the fundal height was greater than expected (34 cm). An ultrasound examination revealed polyhydramnios, with an amniotic fluid index of 29 cm (normal at 27 weeks 9.5-22.6 cm)¹⁸. There was a large amount of ascites, with elevation of the diaphragm and apparent shortening of the chest (fig. 1).

Fig. 1: A sagittal view through the right side of the fetus showing polyhydramnios, the fetal head (H) to the right, the small lung (Lu) with elevated diaphragm, the marked ascites (A), and enlarged liver (Li) with a blunted tip.

The liver was enlarged and had a blunted edge. There were no hepatic masses, cysts, or other visible lesions. There were no other signs of hydrops, or other fetal malformations. There were no calcified areas or cavitations in the brain. Fetal echocardiography was normal.

Maternal laboratory tests did not reveal the cause of the fetal ascites. The patient"s blood type was O positive, with a negative antibody screen. The Kleihauer-Betke stain and VDRL were negative. Maternal liver function tests (Table II) were normal except for hypoproteinemia.

To attempt to find a cause of the ascites and polyhydramnios, fetal blood sampling was performed. The fetal hematocrit was 38% (normal 38-44%), with a mean corpuscular volume of 110 fl. The platelet count was 52,000 (normal 217-289). Fetal blood was sent for a karyotype, which later returned 46,XX. Fetal liver function tests were markedly abnormal, as shown in Table I. When maternal serologic tests returned indicating the possibility of a recent herpes infection, frozen fetal serum was sent for a herpes IgM titer, which was negative.

Because of concern that pulmonary hypoplasia could develop from the marked upward displacement of the diaphragm, fetal paracentesis was performed with a 22 gauge spinal needle. After 175 ml of straw-colored ascites was aspirated, the needle became dislodged with a fetal movement. The patient requested that we stop at that point, although there was a similar quantity remaining. The ascitic fluid was sent for cytomegalovirus immunofluorescent stain and viral culture, both of which returned negative.

 At 29 weeks gestation the patient had spontaneous rupture of the membranes, with passage of yellow amniotic fluid. She soon went into labor and delivered a 1900g male infant. After the baby was intubated, 300 ml of ascites was aspirated, and he was then ventilated without difficulty. Physical exam revealed diffuse bruising and petechiae on the face, trunk and extremities. Other than ascites, there were no signs of hydrops. The liver and spleen were both markedly enlarged. Pertinent neonatal laboratory values are shown on Table I.

IU = International units; mg = milligram; dL = deciliter

The patient"s initial course was complicated by hypotension, hypoglycemia, hypocalcemia, metabolic acidosis, and disseminated intravascular coagulation. These problems improved with therapy; however, the direct bilirubin and liver function tests continued to rise. At 48 hours of age, the infant developed pulmonary edema and congestive heart failure which did not respond to therapy, and he subsequently died of these complications.

Autopsy

At autopsy, there was significant jaundice and massive ascites; however, anasarca was absent. The internal examination demonstrated 100 ml of serous ascitic fluid, pale edematous kidneys, and hepatosplenomegaly, with the liver being deep green in color. Histologic sections of the kidneys revealed diffuse viral inclusions diagnostic of cytomegalovirus infection (fig. 2). Less numerous inclusions were identified in the pancreas, lungs, adrenal glands, and the liver. The hepatic parenchyma was markedly necrotic with organizing inflammation and pronounced cholestasis.

Fig. 2: A histologic section of the fetal kidney demonstrating renal tubular epithelial cells distended by intranuclear and cytoplasmic cytomegalovirus inclusions. (Hematoxylin and eosin, 400X)

Discussion

Prenatal diagnosis

The prenatal diagnosis of cytomegalovirus infection has been made on several occasions following suspicious ultrasound findings or after maternal seroconversion^1,4-6,15,16. The most commonly described ultrasound findings suggestive of cytomegalovirus infection include hydrocephalus, microcephaly, cystic, necrotic or calcified lesions in the brain, liver or placenta, intrauterine growth retardation, and decreased amniotic fluid volume^1,5,6. Since we did not observe any of these other sonographic findings, the differential diagnosis was quite broad in our patient with fetal ascites and hepatomegaly. Our workup included tests for a number of different causes of immune and non-immune hydrops. When we documented thrombocytopenia, elevated transaminase levels, and hyperbilirubinemia with fetal cord blood sampling, we strongly suspected that a cytomegalovirus infection was the cause of the fetal problems.

In our patient the fetal lungs appeared to be very small, apparently as a result of elevation of the diaphragm by the massive ascites. Stockton reported a case in which an infant died of pulmonary hypoplasia, apparently caused by massive ascites associated with cytomegalovirus infection¹¹. Yamashita performed fetal paracentesis in an infected infant to try to avoid this problem³. Because it appeared that pulmonary hypoplasia would be a significant risk in our patient, we aspirated as much ascites as possible. After birth the infant was not thought to have pulmonary hypoplasia either by clinical grounds or at autopsy, but since delivery occurred so soon after the aspiration, it is unlikely that this procedure affected the outcome.

Several approaches have been used to confirm suspected fetal cytomegalovirus infection. Since the tubular epithelium of the kidneys is a major site of viral replication, culture of amniotic fluid or neonatal urine is probably the best way to diagnose congenital cytomegalovirus infection^1,4,13. The sensitivity of neonatal urine cultures in diagnosing cytomegalovirus infection is quite high¹³, although there have been a few cases reported in which urine cultures were negative and the virus was present at other sites¹⁹. Fetal blood cultures are negative in most, if not all, cases⁴. We cultured the ascites fluid (which did not grow the virus) but did not culture the amniotic fluid. The fact that neonatal urine cultures were negative for cytomegalovirus in our patient with severe disease is quite unusual.

There have been several reports in which infection was confirmed by the presence of cytomegalovirus specific IgM in fetal blood^4,14,16. In light of the results of a recent report by Lynch⁶, in which four of five fetal blood specimens were negative for cytomegalovirus specific IgM, it is not surprising that the result was negative in our patient. We found, as have Lynch⁶ and Hohlfeld⁴, significant thrombocytopenia and elevated transaminase levels. Since Stagno has previously reported that most severely infected neonates come from primarily infected mothers, we were surprised when maternal IgM was negative in our patient. However, Lynch recently found that maternal IgM was present in none of five cases with severely affected fetuses. Whether these cases represent recurrences or primary infections with an attenuated antibody response is unknown. Kangro reports that in most cases with reactivated cytomegalovirus, the maternal IgM titers remain negative²⁰.

In conclusion, this case illustrates that congenital cytomegalovirus may present with severe ascites and hepatomegaly. Maternal and fetal serology will not necessarily be helpful in establishing a diagnosis, even in fulminant cases near term. Fetal blood samples will often demonstrate thrombocytopenia and elevated transaminase levels. Culture of the amniotic fluid should be performed, since this may be the best way to establish the diagnosis prenatally.

References

1. Grose C, Itani O, Weiner CP. Prenatal diagnosis of fe