Human Cytomegalovirus (HCMV) DNAemia in the Mother at Amniocentesis as a Risk Factor for Iatrogenic HCMV Infection of the Fetus

Primary human cytomegalovirus (HCMV) infection during pregnancy carry a significant risk of transmission and damage to the fetus [1]. In the absence of any specific antiviral therapy which may be safely administered during pregnancy, prenatal diagnosis represents an important tool for the management of a pregnancy complicated by a primary HCMV infection. However, previous studies (reviewed in [2] and 3]) have indicated that the sensitivity of prenatal diagnosis may be affected by several factors, such as the time interval between maternal infection and prenatal diagnosis, gestational age at time of the invasive procedure, type of clinical specimen examined, and technique employed for virus detection.

It is now widely recognized that to achieve maximum sensitivity: (1) at least 6–8 weeks should elapse between onset of maternal infection and amniotic fluid (AF) sampling, (2) amniocentesis should not be performed before 21 weeks' gestation, (3) AF is the specimen of choice, and (4) molecular techniques for detection of HCMV viral nucleic acids (DNA and/or RNA) should be used. However, even when prenatal diagnosis procedures are performed under optimal conditions and the most sensitive diagnostic techniques are used, a failure to detect HCMV infection in a small proportion of fetuses that are not yet infected at amniocentesis is unavoidable.

In immunocompetent individuals suffering from primary HCMV infection, viral DNA can be detected in blood for a period of time varying from 1 to >6 months [4–6], and invasive procedures for prenatal diagnosis may, therefore, be performed in the presence of HCMV DNA in maternal blood (DNAemia). To assess whether the presence of HCMV DNAemia in the mother at the time of invasive prenatal diagnostic procedures is causally related to intrauterine HCMV transmission, retrospective data collected during more than 15 years of experience with prenatal diagnosis of fetal HCMV infection were reviewed.

Subjects, materials, and methods. A total of 194 pregnant women who underwent prenatal diagnosis because of a primary HCMV infection during the period from July 1990 through March 2007 and their 199 fetuses were included in the study. Diagnosis and date of onset of primary HCMV infection was based on the following: (1) the kinetics of virus-specific IgM antibody and IgG antibody avidity, (2) the detection of HCMV and HCMV products in the blood, and (3) the presence of clinical symptoms and/or abnormal laboratory findings [7]. HCMV-specific IgM levels and IgG avidity were determined by in-house- developed ELISA assays [7, 8]. HCMV DNA was quantified in maternal peripheral blood samples by quantitative polymerase chain reaction (PCR), as reported elsewhere [4]. Briefly, a recombinant DNA molecule containing an HCMV genome sequence relevant to exon 4 of the major IE gene was used as an external standard to establish the sensitivity of the assay. In particular, an amount of ⩾10 copies of plasmidic DNA could be amplified by using single-step PCR. DNAemia was quantified in 10⁵ peripheral blood leukocytes (PBL) until December 1999 and in 10 µL of whole blood thereafter. Samples that were negative for HCMV DNA were tested in single (until December 1999) or in triplicate (from January 2000) by nested PCR. An arbitrary value of 5 genome equivalents (GE)/10⁵ PBL (until December 1999) or 3 GE/10 µL whole blood (from January 2000) was assigned to samples that were positive for HCMV DNA by nested PCR only. However, for the present study, all samples positive for HCMV DNA by nested PCR were considered to contain 3 GE. As reported, pp65 antigenemia [9] and HCMV viremia [10] were quantified in 2 × 10⁵ PBL aliquots.

Prenatal diagnosis was performed by virus isolation and DNA detection by PCR in multiple aliquots of AF, as reported elsewhere [11]. In addition, pp65 antigenemia, viremia, and DNAemia were quantified in fetal blood samples, and the presence of virus-specific IgM was determined in fetal serum samples obtained after 20 weeks gestation. Congenital infection at birth was diagnosed by isolation of HCMV from urine within the first 2 weeks of life [10]. Statistical significance was determined by using the 2-tailed Fisher exact test.

Results. Congenital HCMV infection was diagnosed in 76 (38%) of 199 fetuses. The results of prenatal diagnosis were confirmed at birth or following pregnancy termination for 87 (92%) of 95 uninfected fetuses and 56 (100%) of 56 infected fetuses. The presence of DNAemia was investigated at amniocentesis for 132 women. DNAemia was detected in 27 (37%) of 73 of pregnant women with HCMV-negative AF and in 22 (37%) of 59 of women with HCMV-positive AF, suggesting that maternal DNAemia was unrelated to detection of virus in the fetus. Mothers of uninfected and infected fetuses had comparable levels of DNAemia (median, 3 GE in either group; range, 3–100 GE in mothers of uninfected fetuses and 3–70 GE in mothers of infected fetuses). No pregnant woman had detectable pp65 antigenemia or viremia at time of amniocentesis.

The onset of primary HCMV infection could be precisely dated and the virologic outcome of pregnancy was known for 119 (90%) of 132 women examined for viral DNAemia at the time of the prenatal diagnosis procedure (figure 1). Of the 62 mothers whose fetuses were found to be uninfected, 22 (35%) had detectable DNAemia at the time of prenatal diagnosis procedures. Similarly, 21 (37%) of 57 mothers of infected fetuses were DNAemia positive. There was no significant difference in the timing of invasive procedures between mothers of uninfected fetuses and mothers of infected fetuses (procedures were performed at 16–31 vs. 16–35 weeks of gestation, respectively; median, 22 vs. 21 weeks of gestation, respectively). The time interval between the onset of maternal infection and prenatal diagnosis in mothers of uninfected fetuses was 4–27 weeks (median, 13 weeks); in mothers of infected fetuses, it was 3–25 weeks (median, 13 weeks).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1.

Prenatal diagnosis and DNAemia results for 119 pregnant women for whom the onset of primary human cytomegalovirus (HCMV) infection could be precisely dated and the virologic outcome of pregnancy was known. Each square represents 1 patient. The number of weeks between the onset of primary HCMV infection and prenatal diagnosis procedures is shown in the central column. Dark gray square, congenital infection; light gray square, absence of congenital infection; gray gradient square, false-negative prenatal diagnosis result; asterisk, positive DNAemia at the time of invasive procedures.

Eight mothers of uninfected fetuses delivered HCMV-infected newborns (figure 1). All newborns were asymptomatic at birth. Of the 8 mothers, 4 were DNAemia positive and 4 were DNAemia negative at the time of the invasive procedures (table 1). No significant difference in the number of infected newborns was observed between the group of women who underwent amniocentesis in the presence of HCMV DNAemia (4 of 22 [18%]) and the group who underwent amniocentesis in the absence of HCMV DNAemia (4 of 40 [10%]) (odds ratio, 2.0 [95% confidence interval, 0.4–8.9]; P = .43). Five of the 8 women underwent only amniocentesis; for 3 women, both AF and fetal blood were sampled. Invasive procedures were performed at 18–31 weeks of gestation (median, 22 weeks of gestation). The only woman (subject 2, table 1) who underwent prenatal diagnosis <20 weeks of gestation suffered from a symptomatic primary HCMV infection 9 weeks before her last menstrual period. The time interval between onset of primary infection and prenatal diagnosis procedures was ⩾7 weeks (range, 7–27 weeks) for all but 1 woman (subject 1, table 1).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1.

Virologic results at time of prenatal diagnosis in 8 pregnant women with false-negative prenatal diagnosis results.

Discussion. This study indicates that invasive procedures performed in the presence of maternal DNAemia do not represent a significant risk factor for iatrogenic transmission of HCMV to the fetus and confirms that the unpredictability of HCMV transmission remains a major obstacle to the absolute sensitivity of prenatal diagnosis of congenital infection. These conclusions are important for both the management and counseling of pregnant women with primary HCMV infection.

The first conclusion relies on the observation that 4 (18%) of 22 women who underwent amniocentesis in the presence HCMV DNAemia and 4 (10%) of 40 women who underwent amniocentesis in the absence of HCMV DNAemia, and whose fetuses were found to be uninfected, eventually transmitted the infection. Although the congenital infection rate in pregnancies during which invasive procedures were performed in the presence of maternal DNAemia was about twice as high as that observed when invasive procedures were performed in the absence of detectable DNAemia, the difference was not statistically significant. Therefore, although the small number of subjects examined in this study does not allow firm conclusions to be drawn, delayed transmission to the fetus does not seem to be causally related to the presence of HCMV in maternal blood at amniocentesis.

To our knowledge, the risk of iatrogenic transmission has been mentioned in only 2 studies [4, 11]. However, neither of these studies was specifically designed to address this issue. In the first study, of the 5 mothers of uninfected fetuses who underwent prenatal diagnosis while they were DNAemia positive, 1 eventually transmitted the infection and 4 did not. In this study, numbers were too small to rule out the risk of iatrogenic transmission. In the second study, Liesnard et al. [11] reported a comparable transmission rate for women who underwent 1 antenatal sampling and for women who underwent multiple antenatal samplings. In addition, Liesnard et al. [11] reasoned that a higher transmission rate would have been observed if transmission had occurred via an antenatal invasive procedure, compared with the rate observed in retrospective studies performed in the absence of prenatal procedures.

Although iatrogenic HCMV infection of the fetus via invasive procedures that are performed in the presence of viral DNAemia in the mother appears to be an unlikely event, it should be stressed that, whenever feasible, it remains advisable to postpone invasive procedures until maternal DNAemia is no longer detectable. However, it has to be kept in mind that, because viral load in the blood of immunocompetent individuals with primaryHCMVinfection may be very low (as shown in the current study and in a previous study [4]), intermittent positivity may be observed particularly late in the convalescent phase of primary infection (unpublished observation). On the other hand, the very low levels of circulating viral DNA detected in some pregnant women at the time of amniocentesis may explain why iatrogenic transmission does not appear to be a significant risk factor for fetal HCMV infection.

The second conclusion relies on the observation that, even when prenatal diagnosis is performed with optimal timing and the virus is sought using the most sensitive techniques, falsenegative prenatal diagnosis results (i.e., transmission occurring after prenatal procedures have been performed) may occur. In fact, in the present series, prenatal procedures performed too close to the onset of maternal infection or too early during gestation may account only for 2 out of the 8 cases of false-negative antenatal results. A deficit in sensitivity of the PCR technique employed for viral DNA detection in AF represents an unlikely explanation for the false-negative antenatal results because a retrospective study from our group [10] previously demonstrated that nested PCR performed on multiple aliquots of AF reached the maximum (though not absolute) level of sensitivity. In this series, at least 3 aliquots were examined for each AF sample with the exception of a single AF specimen collected in 1991.

Finally, this study provides additional evidence that intrauterine transmission does not correlate with duration of maternal DNAemia. This observation also appears to be correct when one considers that HCMV DNAemia can be detected for months after primary infection [10, 5, 6], whereas a 6–8 week interval between the onset of maternal infection and amniocentesis is considered sufficient to achieve a reliable diagnosis of fetal infection [2, 10–15]. Two corollaries to the above considerations may be the following: the search for prognostic markers of transmission should focus on the early period after maternal infection; and interventions aimed at preventing intrauterine transmission, when performed later than 6 weeks after the onset of HCMVinfection, are likely to be ineffective in the great majority of cases.

• Received July 19, 2007.
• Accepted September 6, 2007.

• © 2008 by the Infectious Diseases Society of America