A light in the darkness: Predicting outcomes for congenital cytomegalovirus infections

Abstract

See related article, p. 90. Cytomegalovirus is a near ubiquitous infection of humankind and is the most common cause of intrauterine viral infection, estimated to affect 1% of births. About 10% to 15% of infected infants will be born with clinically apparent CMV disease consisting of some combination of microcephaly, intracranial calcifications, chorioretinitis, jaundice, hepatosplenomegaly, and petechiae or purpura. The mortality rate in affected infants can reach as high as 30%. Long-term neurologic sequelae including hearing loss, visual impairment, psychomotor delay, and mental retardation are expected in a large proportion of survivors.1Nelson C Demmler G Cytomegalovirus infection in the pregnant mother, fetus, and newborn infant.Clin Perinatol. 1997; 24: 151-160PubMed Google Scholar, 2Boppana S Pass R Britt W Stagno S Alford C Symptomatic congenital cytomegalovirus infection: neonatal morbidity and mortality.Pediatr Infect Dis J. 1992; 11: 93-99Crossref PubMed Scopus (484) Google Scholar Despite the high prevalence of CMV in the population and its potentially devastating consequences to the fetus, obstetricians and pediatricians have had no reliable tools with which to detect maternal CMV infection or to predict the outcome of a pregnancy known to be complicated by CMV. Clinical diagnosis of maternal primary CMV infection is unlikely because most CMV infections are asymptomatic and go unnoticed. Routine serologic screening of women during pregnancy is fraught with uncertainty because IgM can be present with CMV reactivation, a circumstance posing minimal risk to the fetus, and demonstration of seroconversion is not possible without serial testing beginning in early pregnancy. Positive maternal viral cultures can also result from reactivation of CMV. Management is further complicated by the fact that maternal CMV infection does not equal fetal infection, because only about 50% of fetuses will become infected as a result of maternal primary infection. Furthermore, detection of fetal infection by virologic or serologic testing of amniotic fluid or fetal blood does not predict which infected fetuses will be clinically affected at birth. The majority of infected infants will be free of symptoms at birth and have no long-term sequelae. If the goal of prenatal testing is to make accurate predictions of outcome to inform decision making, current technology is woefully inadequate.3Pass R Commentary: Is there a role for prenatal diagnosis of congenital cytomegalovirus infection?.Pediatr Infect Dis J. 1992; 11: 608-609PubMed Google Scholar In this issue of The Journal, Lazzarotto et al4Lazzarotto T Varani S Guerra B Nicolosi A Lanari M Landini MP Prenatal indicators of congenital cytomegalovirus infection.J Pediatr. 2000; 137: 90-95Abstract Full Text Full Text PDF PubMed Scopus (193) Google Scholar report on the application of a stepwise strategy to address the 3 areas critical to the diagnosis and prediction of outcomes of intrauterine CMV infection: maternal infection, fetal infection, and symptomatic newborn disease. In the first step, a screening program identified maternal primary infection before 16 weeks’ gestation by use of traditional serologic assays for detecting CMV-specific IgM and IgG together with methods more recently developed by the authors and others for IgM immunoblotting5Lazzarotto T Ripalti A Bergamini G Battista M Spezzacatena P Campanini F Development of new cytomegalovirus (CMV) immunoglobulin M (IgM) immunoblot for the detection of CMV-specific IgM.J Clin Microbiol. 1998; 36: 3337-3341PubMed Google Scholar and for determining the avidity index for CMV IgG antibodies.6Blackburn N Besselaar T Schoub B O’Connell K Differentiation of primary cytomegalovirus infection from reactivation using the urea denaturation test for measuring antibody avidity.J Med Virol. 1991; 33: 6-9Crossref PubMed Scopus (114) Google Scholar, 7Grangeot-Keros L Mayaux M Lebon P Freymuth F Eugene G Stickler R et al.Value of cytomegalovirus IgG avidity index for the diagnosis of primary CMV infection in pregnant women.J Infect Dis. 1997; 175: 944-946Crossref PubMed Scopus (221) Google Scholar, 8Lazzarotto T Spezzacatena P Pradelli P Abate D Varani S Landini M Avidity of immunoglobulin G directed against human cytomegalovirus during primary and secondary infections in immunocompetent and immunocompromised subjects.Clin Diagn Lab Immunol. 1997; 4: 469-473PubMed Google Scholar In the second step, women identified as having primary CMV were offered amniocentesis for viral culture and qualitative polymerase chain reaction testing and a fetal ultrasound examination. The final step consisted of performance of a quantitative PCR test on those amniotic fluid samples that were found to be positive by qualitative PCR test for CMV. Newborns and aborted fetuses were evaluated for evidence of symptomatic or asymptomatic CMV infection with clinical, histologic, and virologic criteria. Of 456 pregnant women screened for CMV, 110 were classified as having primary infection. It is unclear whether referral bias contributed to the large number of primary infections detected in this population. An additional 20 women with an undefined infection status were offered the same evaluation as for primary infection; 83% of these pregnancies had a confirmed outcome, although virologic data were available only for 44%. Pregnancy outcome was obtained by report for 87% of the 378 women for whom amniocentesis was not performed. Interestingly, 56% of the 130 women with primary or undetermined CMV infection refused amniocentesis, primarily because of its unproven predictive value. The diagnostic strategy incorporating quantitative PCR testing, with a cutoff of 105 genome equivalents per milliliter of amniotic fluid, correctly identified 7 symptomatically infected infants with a positive predictive value of 100%. Sixteen of 17 infants were correctly identified as not having symptomatic CMV, yielding a negative predictive value of 94%. One infant with <105 genome equivalents was mildly symptomatic. Not surprisingly, the value of the qualitative PCR for predicting symptomatic disease was quite poor (28%). However, enthusiasm must be tempered by acknowledging the small number of symptomatic infants, the low percentage of mothers who underwent amniocentesis, and the broad case definition used to define symptomatic disease. Mild hepatitis was the only finding noted at birth for one symptomatic infant with an amniotic fluid viral load of >105. It was also the only finding ascribed to a symptomatic infant who had <105 genome equivalents/mL of amniotic fluid. The presence of otherwise unexplained growth retardation or prematurity (<38 weeks) would have been sufficient for a diagnosis of symptomatic CMV disease, although no infant in this study met the case definition on this basis alone. The strategy may prove much less useful if it cannot discriminate between a severely affected infant with multiple findings, including central nervous system disease, and those with symptoms such as mild hepatitis or growth retardation, for whom the long-term prognosis is more favorable.9Conboy T Pass R Stagno S et al.Early clinical manifestations and intellectual outcome in children with symptomatic congenital cytomegalovirus infection.J Pediatr. 1987; 111: 343-348Abstract Full Text PDF PubMed Scopus (119) Google Scholar It may be necessary to devise a severity scale for symptomatic infants to address this issue. The study raises other important questions. Would non-Italian women be as likely to refuse invasive diagnostic testing? Would high maternal viral load predict pregnancy outcome just as high viral loads in transplant recipients predict CMV disease or transmission of human immunodeficiency virus? If a fetus is identified as likely to have symptomatic CMV disease, would currently available antiviral agents active against CMV be effective and would their toxicities be justified, especially if termination of pregnancy was not under consideration? Would a larger study with long-term follow-up of the asymptomatically infected cohort assess the value of the quantitative PCR results to predict the 10% to 15% of this population likely to have late sequelae such as hearing loss and cognitive impairment? There is currently no treatment for congenital CMV, although a National Institutes of Health–sponsored, placebo-controlled, phase III trial of ganciclovir therapy for newborns with central nervous system involvement was recently closed and the data are under analysis (D. Kimberlin, oral communication, February 24, 2000). The best hope for eliminating congenital CMV infection is to prevent prenatal infection. With no vaccine on the immediate horizon, low-tech methods such as good hygiene and handwashing must be promoted, particularly when a pregnant woman already has a young child attending day care. In such settings, the risk of acquisition by susceptible mothers is 25%.10Adler S Cytomegalovirus transmission among children in daycare, their mothers and caretakers.Pediatr Infect Dis J. 1988; 7: 279-285Crossref PubMed Scopus (42) Google Scholar The concept of quantitative virology is gaining importance in the management of other chronic viral infections such as hepatitis B and C and human immunodeficiency virus. Quantitative virology for CMV has shown some utility in predicting CMV disease in transplant recipients. For those of us who have long suffered with the uncertainties surrounding prenatal diagnosis of CMV and its implications, the work of Lazzarotto et al4Lazzarotto T Varani S Guerra B Nicolosi A Lanari M Landini MP Prenatal indicators of congenital cytomegalovirus infection.J Pediatr. 2000; 137: 90-95Abstract Full Text Full Text PDF PubMed Scopus (193) Google Scholar is a light in the darkness. We look forward to the application and validation of this work in a larger trial. Prenatal indicators of congenital cytomegalovirus infectionThe Journal of PediatricsVol. 137Issue 1PreviewObjective: To assess the validity of a diagnostic protocol designed to predict the outcome of newborns of mothers suspected to have primary cytomegalovirus (CMV) infection during the first 4 months of pregnancy. Study design: Anti-CMV immunoglobulin (Ig) M detection by enzyme immunoassay and immunoblot together with the determination of anti-CMV IgG avidity allowed us to classify 456 women as (1) uninfected, (2) undergoing either a primary or a recurrent infection, or (3) having an undefined serologic condition. Full-Text PDF