|Home | About | Journals | Submit | Contact Us | Français|
The diagnostic value of RT-PCR of amniotic fluid (AF) for prenatal diagnosis of congenital rubella virus infection in 45 pregnant women with confirmed primary infection was assessed. Specificity of RT-PCR was 100% and sensitivity ranged between 83 and 95%.
Congenital rubella syndrome is an important cause of deafness, blindness, and mental retardation (9, 11). For nearly 20 years, prenatal diagnosis of fetal rubella infection has been available to pregnant women. In the early 1980s, prenatal diagnosis was essentially based on the detection of the rubella-specific immunoglobulin M (IgM) antibody in fetal blood (3, 5, 10). More recently, methods for detecting the viral genome were developed. Very few studies have accurately addressed the use of PCR after reverse transcription (RT-PCR) for the prenatal diagnosis of congenital rubella due to the limited number of amniotic fluid (AF) samples analyzed (2, 4, 13, 14). Furthermore, the lack of an approved definition of recent primary infection has often led to misinterpretation of prenatal diagnosis results (8). The aim of our study was to compare the use of RT-PCR in AF samples to the detection of the rubella virus IgM antibody in fetal blood for the prenatal diagnosis of congenital rubella in pregnant women with confirmed primary rubella virus infection.
Between 1997 and 2002, 110 AF and 23 fetal blood samples from pregnant women with confirmed primary rubella virus infection at between 2 and 22 weeks of gestation (WG) were sent to our laboratory from all over France for prenatal diagnosis of congenital rubella. The value of RT-PCR in 45 cases for which newborns and fetuses were serologically documented was assessed. Diagnosis of primary rubella virus infection was based on seroconversion and/or the detection of rubella virus-specific IgM and rubella virus-specific low-avidity IgG. The stage of pregnancy at the time of infection was defined as the interval (in weeks) between the last menstrual period and the onset of the rash (n = 19) or the infectious contact plus 2 weeks (n = 1). For symptom-free patients, the stage of pregnancy was based on a documented seroconversion (n = 22) or estimated according to the value of the avidity index (n = 3). Informed consent was obtained from patients who underwent prenatal diagnosis.
Additional information on biological investigations, pregnancy outcome, and defects of the fetuses and infants was collected by the National Institute for Public Health Surveillance (Institut de Veille Sanitaire).
Serum samples from pregnant women and infants were tested with various commercial assays for rubella virus-specific IgG and IgM by the laboratories sending AF and fetal blood samples to our laboratory. Rubella virus-specific IgG avidity in pregnant women was measured as previously described for cytomegalovirus (CMV) IgG avidity (7) with the Enzygnost anti-rubella virus IgG kit (Dade Behring, Marburg, Germany).
In fetal blood samples, rubella virus-specific IgM was detected by a capture immunoassay (Bio-Rad, Marnes la Coquette, France). As recommended by the manufacturer, serum samples were diluted to 1/20 for fetuses and newborns, instead of 1/100 for standard serum samples, in order to increase sensitivity.
In AF, detection of the genomic RNA of rubella virus was performed in triplicate by RT-nested-PCR analysis according to a procedure derived from a previously described method (4); traditional techniques were used for extraction (proteinase K-sodium dodecyl sulfate and phenol-chloroform-isoamyl alcohol precipitation), reverse transcription (selective priming and Moloney murine leukemia virus reverse transcriptase), Southern blotting (alkaline transfer method), and detection by hybridization. The nested PCR amplified the 5′-terminal part of the E1 gene of rubella virus, producing a 320-bp amplicon. The whole process was able to detect up to 8 infectious units of the World Health Organization international standard (1st International Reference Reagent 1994 for Rubella; National Institute for Biological Standards and Control)/ml.
In 35 cases, pregnancy continued. One spontaneous abortion and five induced abortions were reported, and, in four cases, the issue of pregnancy was unknown. Clinical examination of 30 infants (28 at birth and 2 fetuses) was performed (Tables (Tables11 and and22).
In our series of 45 mothers, RT-PCR on AF samples was positive in 20 cases and negative in 21 cases (Table (Table1).1). In four cases, RT-PCR replicates were nonreproducibly positive (Table (Table22).
All of the 20 cases with positive RT-PCR were also positive for the IgM antibody. Defects compatible with rubella (heart and eye defects and deafness) were reported in 4 of 13 clinically documented cases (3 of 5 before 12 WG, 1 of 8 after 12 WG). In two cases, RT-PCR was negative for the first AF samples collected at 17 and 16 WG, 4 and 5 weeks after seroconversion, respectively, and was subsequently positive for AF samples collected 6 weeks later. In another case, RT-PCR performed on AF sent by mail at room temperature was negative, whereas it was positive when the sample was kept frozen during transportation. These three AF samples were considered positive.
Among the 21 RT-PCR-negative cases, rubella virus-specific IgM detection was negative for 17 samples and positive for 4. For the 10 clinically documented IgM-negative cases, no defect was reported. Of the four IgM-positive cases (Table (Table3),3), defects (moderate microcephaly and low birth weight) were detected in only one of the infants (patient GIR). In three of the four cases, AF sampling was performed at 17, 19, and 21 WG, while seroconversion occurred at 13, 14, and between 5 and 16 WG (patients BOR, GIR, and NZP, respectively); in the last case, seroconversion occurred at 10 WG and AF sampling was performed at 19 WG (patient OLS).
Among the four cases nonreproducibly positive by RT-PCR, blood samples were positive for rubella virus-specific IgM in three cases, without any congenital defect (patients ROC, ARI, and HEN); the last one (for patient LEB), not clinically documented, was negative for the IgM antibody at birth.
As there is no “gold standard” for prenatal diagnosis of fetal rubella virus infection, RT-PCR sensitivity and specificity were calculated with respect to IgM antibody results, since, in our experience, IgM sensitivity is ≥98% and specificity is 100% (data not shown). Nonreproducibly positive results were excluded from specificity and sensitivity calculations. Specificity of the RT-PCR, calculated for the 17 IgM antibody-negative cases, was 100%. Sensitivity of viral RNA detection was 83% (20 of 24 cases). However, if we withdraw from analysis the two or three cases for which samples were collected less than 6 weeks after maternal infection or at too early a stage of the pregnancy, sensitivity increases to 91 (20 of 22 cases) or 95% (20 of 21 cases). In the same way, the two false-negative results on a first sample, positive on a subsequent one, can be explained by sampling too early. As previously reported for CMV infection (6, 12), an interval of 6 weeks between seroconversion and amniocentesis seems to be appropriate. It has also been shown for CMV infection that, considering the stage of pregnancy at the time of AF collection, it is better to collect AF samples around the 22nd WG (1). In practice, it seems logical to follow the same rules for prenatal diagnosis of rubella virus infection. Sensitivity of prenatal diagnosis depends also on the conditions for the transportation of AF samples. Our experience leads us to recommend transportation of AF samples in dry ice and their subsequent storage at −70°C.
In three of four cases with nonreproducibly positive RT-PCR results from the replicates, IgM antibody was positive but no clinical defects were reported. This could be explained by a very low viral load.
In conclusion, our study, based on proven primary maternal rubella virus infection, confirms that RT-PCR performed on AF samples has a good sensitivity and specificity. As amniocentesis is less invasive than fetal blood sampling, we may consider that this procedure is a valuable tool for prenatal diagnosis of fetal rubella virus infection.
We are grateful to obstetricians for providing AF samples from pregnant women, blood samples from fetuses, and clinical information. We thank Pascal Dubreuil for excellent technical assistance and Richard Keros for critical reading and English language correction of the manuscript.