Prenatal Diagnosis of Congenital Human Cytomegalovirus Infection in Amniotic Fluid by Nucleic Acid Sequence-Based Amplification Assay

Prenatal diagnosis of congenital human cytomegalovirus (HCMV) infection is mostly performed by rapid virus isolation from, and/or PCR detection of viral DNA in, amniotic fluid (AF) samples (1, 3, 11, 12, 14). In general, both techniques have been shown to provide high specificity (96 to 100%) and good sensitivity (70 to 90%) (4, 8, 10, 14, 15), with the exception of a few reports from a single group showing low specificity and positive predictive value for PCR on AF samples (5, 9). Since irrevocable decisions are often made on the basis of prenatal diagnosis, the use of a confirmatory assay is highly recommended. Thus, given the lack of cell culture facilities in many laboratories, molecular assays which may confirm PCR results for AF samples, thereby functioning as alternatives to HCMV isolation, must be identified.

There are two commercially available assays for detection of HCMV immediate-early (IE) and late pp67 mRNA using the nucleic acid sequence-based amplification (NASBA) technique (BioMérieux, Boxtel, The Netherlands). The assay for pp67 mRNA is available as a complete kit (NucliSens CMV pp67), whereas a basic kit (NucliSens Basic Kit) is available for IE mRNA detection; IE mRNA-specific primer sequences can be obtained from BioMérieux. Currently, the NucliSens CMV pp67 assay is approved by the Food and Drug Administration for mRNA detection in blood, while the IE mRNA NASBA basic kit is to be used for research purposes.

Overall, 65 AF samples from as many pregnant women with primary HCMV infection during pregnancy were tested. Primary infection was diagnosed by either HCMV-specific immunoglobulin G (IgG) seroconversion or the presence of HCMV-specific IgM associated with a low IgG avidity index (15). The virologic outcome of pregnancy was ascertained in newborns by HCMV isolation within the first 2 weeks of life and in aborted fetuses by virus isolation from and histologic examination of fetal tissues. All AF samples had been previously characterized for HCMV presence by both rapid virus isolation on shell vial cultures and viral DNA amplification by PCR (16).

NASBA assays were performed retrospectively on AF samples according to the instructions given by the manufacturer for pp67 mRNA detection in blood samples (NucliSens CMV pp67). In particular, 49 AF samples collected from September 1991 through December 1999 and stored undiluted at −80°C were diluted 1:10 in NucliSens lysis buffer (BioMérieux) upon thawing. These samples were tested by NASBA for U1A mRNA, the transcript of a cellular gene with a low rate of transcriptional activity, to verify whether RNA was degraded during storage (7). The remaining 16 samples, collected after January 2000, were stored at −80°C as 1:10 dilutions in lysis buffer. Nucleic acids were extracted from 100 μl of AF by using the NucliSens isolation kit (BioMérieux).

No RNA degradation was detected in the 49 samples stored undiluted, as determined by U1A mRNA detection. Overall, of the 65 samples tested, 29 were positive for HCMV by both culture and DNA detection (reference assays), whereas 36 were HCMV negative by both techniques. As shown in Table 1, all 29 HCMV-positive samples were scored positive by both IE and pp67 mRNA NASBA assays. On the other hand, of the 36 HCMV-negative samples, 36 were confirmed negative by pp67 mRNA detection, whereas one sample tested positive for IE mRNA. Thus, compared to virus isolation and DNA detection, IE and pp67 mRNA NASBA assays showed 98.5 (64 of 65) and 100% (65 of 65) concordance, respectively. However, while all 29 congenital HCMV infections diagnosed antenatally were confirmed either at birth (17 newborns) or at autopsy following voluntary termination (12 fetuses), 2 out of 36 newborns with a prenatal diagnosis of no HCMV infection by virus isolation and PCR were found HCMV infected at birth (Table 1). In both cases congenital infection was asymptomatic. Neither of the two NASBA assays revealed the presence of HCMV in AF samples from these two newborns (the only IE mRNA-positive result obtained in this group did not concern either one of these two congenitally infected newborns). In addition, no U1A mRNA degradation was observed in these two samples. Therefore, when referred to the virological outcome, sensitivity, specificity, and negative and positive predictive values for virus isolation, PCR, and pp67 NASBA were found to be identical, i.e., 93.5, 100, 94.4, and 100%, respectively (Table 2). As for IE mRNA, specificity and positive predictive value were slightly lower (97.1 and 96.9%, respectively), given the false-positive result for the group of uninfected fetuses.

The two false-negative results pertained to two fetuses whose mothers suffered from primary HCMV infection at 11 and 16 weeks of gestation, respectively. In both cases, amniocentesis was performed at 23 weeks of gestation, and fetal blood was negative for antigenemia, viremia, and DNAemia as well virus-specific IgM.

This study shows that the presence of HCMV in AF samples can be reliably detected by two commercial NASBA assays. In fact, pp67 mRNA and IE mRNA NASBA assays showed 100 and 98.5% concordance, respectively, to DNA detection by PCR and virus isolation by the shell vial assay. In particular, all positive results obtained antenatally, with the exception of one IE mRNA-positive result, were confirmed at birth, thus supporting the high specificity of PCR results as previously reported by us and other groups (3, 4, 6, 10, 11, 14-17). The discordant result obtained by the IE mRNA NASBA assay in the group of negative AF samples is difficult to explain. The sample, stored at −80°C since 1998, was tested two additional times with discordant results, thus indicating a low-level positivity.

On the other hand, none of the assays included in the present study provided 100% sensitivity. False-negative results represent a well-known problem of prenatal diagnosis of congenital HCMV infection. This problem has been shown to be only partially overcome by (i) increasing the sensitivity of the PCR techniques employed for HCMV detection in AF samples (16), (ii) waiting at least 6 to 8 weeks after the onset of maternal infection before fetal sampling (1, 10, 11, 13), and (iii) performing the procedure after 21 weeks of gestation (2, 10, 11). Indeed, notwithstanding the adoption of these precautions, false-negative results due to the delayed transmission of the infection may still be encountered, although infrequently. However, since false-negative results were associated with lack of symptoms at birth, a sensitivity of greater than 90% (the top now achievable) may be considered satisfactory.

In conclusion, this study shows that both IE and pp67 mRNA NASBA assays can be reliably performed to detect HCMV in AF samples. Therefore, they can be used as first-choice assays or assays to confirm results obtained by PCR in those laboratories where virus isolation is not performed. In fact, all but one of the positive mRNA results obtained antenatally in this study were confirmed at birth. In addition, it has been shown that HCMV mRNA can be successfully detected in AF samples stored at −80°C for more than 10 years.

• Copyright © 2003 American Society for Microbiology