Prenatal diagnosis of congenital cytomegalovirus infection in 115 cases: a 5 years single center experience

Transcription

1 DOI: /pd.5025 ORIGINAL ARTICLE Prenatal diagnosis of congenital cytomegalovirus infection in 115 cases: a 5 years single center experience M. Enders 1 *, A. Daiminger 1, S. Exler 1, K. Ertan 2, G. Enders 1 and R. Bald 2 1 Laboratory Prof. Gisela Enders & Colleagues MVZ and Institute of Virology, Infectiology and Epidemiology e.v., Stuttgart, Germany 2 Department of Obstetrics, Gynecology and Prenatal Medicine, Klinikum Leverkusen, Leverkusen, Germany *Correspondence to: M. Enders. menders@labor-enders.de ABSTRACT Objective The objective of this study is to investigate the diagnostic value of invasive prenatal diagnosis (PD) of congenital cytomegalovirus (CMV) infection from amniotic fluid (AF) and fetal blood (FB). Methods A retrospective study was conducted on 115 pregnancies with CMV primary infection. A total of 111 AF and 106 FB samples were investigated for various virological and non-virological markers. Detailed ultrasound examinations were performed at time of PD. Results Overall sensitivity of CMV PCR in FB (75.6%; 95%CI 60 87) and AF (72.7%; 95%CI 57 85) was comparable. In women with amniocentesis >8 weeks between seroconversion and PD, we did not observe significant differences between amniocentesis performed (sensitivity 90.9%; 95%CI 71 99) and gestational weeks (sensitivity 90.0%; 95%CI 68 99). Virological markers in FB were higher in symptomatic compared with asymptomatic fetuses (p < 0.05). No significant differences were observed for non-virological markers. However, platelet counts <120 10e9/L and beta-2 microglobulin values >14 mg/l were more frequently found in fetuses with severe ultrasound abnormalities compared with fetuses with no or mild abnormalities (p < 0.001). Conclusion Optimal timing of amniocentesis in women with primary infection in early gestation should be reevaluated in a prospective study. Analysis of FB markers may be beneficial in the individual management of pregnant women with confirmed congenital CMV infection. Funding sources: None Conflicts of interest: None declared INTRODUCTION Management of primary cytomegalovirus (CMV) infection in pregnancy includes targeted ultrasound (US) monitoring as well as counselling on invasive prenatal diagnosis (PD) of fetal infection. While US is indispensable for detecting CMV-derived abnormalities in infected fetuses, its value in diagnosing intrauterine viral transmission is limited. 1 Amniocentesis (AC) is considered the best option for PD. 2 In order to reduce the risk of false negative results, it is generally recommended to perform AC after completed 21 weeks of gestation and at least 6 to 8 weeks after maternal infection. 2 4 Another approach to PD of fetal CMV infection is cordocentesis. Reported sensitivities for CMV DNA detection in fetal blood (FB) range from 41% to 92%, 5 8 whereas the sensitivity for detection of CMV-specific Immunoglobulin M (IgM) is generally low. Although FB sampling does not increase sensitivity of PD, investigation of non-virological (hematological and biochemical) parameters may be of value to assess the clinical condition of the fetus In our study, we describe a single center experience on invasive PD of fetal CMV infection in a 5-year cohort of women with primary CMV infection. We reevaluated factors (e.g. type of sample used, method, and scheduling of the procedure) that influence the diagnostic value of PD. In addition, we analyzed laboratory parameters in FB in relation to intrauterine infection and to US findings at the time of invasive PD. METHODS This retrospective study investigated 132 pregnant women with primary CMV infection who underwent invasive PD at the prenatal medicine unit Leverkusen (Department of Obstetrics and Gynecology, Klinikum Leverkusen) between 2010 and During the initial presentation, all women received extensive counselling and detailed US examinations. Following written informed consent, invasive PD was made by AC and concomitant cordocentesis in 117 women; in 11 cases, only amniotic fluid (AF) and in four cases, only FB were obtained. Maternal and fetal samples were investigated at Laboratory Enders, Stuttgart. Results of invasive PD were confirmed by CMV DNA and/(or) virus detection in fetal or placental tissue samples in case of termination of pregnancy

2 390 M. Enders et al. or in neonatal urine samples obtained within the first 2 weeks of life. Congenital CMV infection may also be confirmed by positive CMV DNA or CMV-IgM ab results in cord or neonatal blood, whereas negative findings do not exclude intrauterine infection. Seventeen cases (15 with AC and simultaneous FB sampling and two with AC only) were excluded from the study because no information on virological outcome was available. Following first invasive PD, most cases with proven intrauterine infection were treated with fetal intravenous administration of CMV hyperimmune globulin (HIG) and/or materno-fetal valganciclovir. Hence, we cannot describe the natural course of fetal CMV infection. Patient data (e.g. demographic information, case history, prenatal CMV HIG treatment, US findings, and laboratory results) were retrospectively extracted from individual medical records and the laboratory information system. Prior to analysis patient identifiers were removed. In women with known virological pregnancy outcome, CMV primary infection was diagnosed by IgG seroconversion (n = 57) or by low IgG avidity in the presence of positive IgM antibodies (n = 54). In four cases, primary infection was suspected because of detection of moderate IgG avidity in presence of high positive IgM values. The following serological CMV assays were used according to the manufacturers instructions: Enzygnost Anti-CMV/IgG EIA using the BEP III processor (SIEMENS), CMV IgM ELA PKS (medac), recomblot and/or recomline CMV IgG [avidity] (MIKROGEN), and CMV IgM and IgG [avidity] ARCHITECT (ABBOTT). The VIDAS CMV IgG Avidity I/II assay (biomérieux) was performed as previously described. 13,14 Time of maternal infection was estimated by combining anamnestic and serologic data (including previous test results from other laboratories). Ultrasound examinations were performed by one experienced examiner and one dedicated specialist in US and fetal medicine. High-frequency probes of premium-class US machines (Toshiba Aplio 500 Platinium edition and Samsung UGEO WS 80A) were used for transabdominal and transvaginal scans. All patients underwent a complete sonographic examination with evaluation of the AF, as well as the localization, size, and the sonographic appearance of the placenta. A complete fetal anatomical scan including fetal echocardiography and doppler sonography was performed. The guidelines of the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) for the sonographic examination of the fetal central nervous system and the ISUOG practice guidelines for sonographic screening of the fetal heart were followed. Grading of US findings (severe or mild) at invasive PD was performed on the basis of number and localization of abnormalities as well as on the severity of isolated abnormalities. All women who underwent cordocentesis were admitted to hospital and monitored for least 24 h after intervention. AF and FB sampling were performed under continuous US guidance by one very experienced operator. Cordocentesis was usually carried out in free-hand technique at the placental cord insertion site or a free loop. For FB sampling before 21 gestational weeks (gw), a 22-G spinal needle was used. The CMV DNA was quantitatively determined using an inhouse real-time PCR 15 or the RealStar CMV PCR Kit (altona DIAGNOSTICS). The sensitivity of both assays was 125 international units per ml (IU/mL). Infectious virus was detected in AF by rapid cell culture using a quantitative shell vial assay as previously described. 16 CMV-specific IgM was determined in all 106 FB samples using the CMV IgM ELA PKS (medac). FB samples were analyzed for platelet count (n = 99) using the Sysmex KX-21 analyzer, for beta-2 microglobulin (n = 100), gamma-glutamyl transpeptidase (GGT) (n = 97), and/or aspartate aminotransferase (AST) (n = 98) on an autoanalyzer (Roche/Hitachi MODULAR P800) Figure 1 Outcome of pregnant women with primary CMV infection who underwent invasive PD at the prenatal medicine unit Leverkusen between 2010 and * indicates CMV DNA detection (PCR) in amniotic fluid and/or fetal blood; indicates one case with detection of CMV DNA (and CMV IgM) in fetal blood, but negative CMV DNA results in amniotic fluid and newborn urine. ccmv, congenital CMV infection (confirmed at TOP or birth); CMV, cytomegalovirus; PD, prenatal diagnosis; TOP, termination of pregnancy; US, ultrasound

3 Prenatal diagnosis of congenital cytomegalovirus infection 391 Table 1 Ultrasound abnormalities, virological findings, and fetal blood markers in infected symptomatic fetuses (n = 16) Case no. Maternal infection in Gestational age at invasive PD Ultrasound findings Grading of ultrasounds findings Amniotic fluid CMV DNA [ 10 3 IU/mL] CMV DNA [ 10 3 IU/mL] Fetal blood CMV-specific Beta-2 MG IgM [Index] a [mg/ml] Platelets [ 10e9/L] GGT [U/L] AST [U/L] A) Cerebral and extracerebral findings PVE, HEB Severe periconcept./ PVE, HEB, OH, intrahepatic calcification, PM Severe periconcept./ PVE, VM (14.6 mm), HSM, CM Severe periconcept./ PVE, PM Mild periconcept./ PVE, HEB, high MCA flow velocity, PM, IUGR Severe nd nd nd PVE, HEB, HSM, AS, intrahepatic calcification Severe nd trim PVE, VM (13.4 mm), OH Mild B) Cerebral findings only 40 preconceptional, periconceptional PVE Severe periconcept./ PVE Mild nd trim PVE, VM (10.7 mm) Mild C) Extracerebral findings only HEB Severe HEB, CM Severe periconcept./ HEB Severe nd HEB Severe periconcept./ HEB, HSM Severe nd st/2nd trim OH Mild AS, ascites; AST, aspartate aminotransferase; CM, cardiomegaly; GGT, gamma-glutamyl transpeptidase; HEB, hyperechogenic bowel; HSM, hepatosplenomegaly; IUGR, intrauterine growth retardation; MCA, middle cerebral artery; nd, not done; OH, oligohydrammnion; PM, placentomegaly; PVE, periventricular hyperechogenicity; trim., trimester of pregnancy; VM, ventriculomegaly. a Evaluation IgM ELA: index: <0.9 negative; 0.9 to 1.1 borderline; >1.1 positive.

4 392 M. Enders et al. using commercially available reagents (cobas, Roche). Gestational age is indicated in completed gestational weeks plus days (gw), for example, gw. Institutional review board approval was not requested given the retrospective nature of the study and the removal of patient identifiers before analysis. For statistical analysis, we used MedCalc software version for the diagnostic performance in pregnancies with short ( 8 weeks) and long (>8 weeks) intervals between estimated seroconversion and AC (Fisher s exact test) and in pregnancies with and without CMV HIG treatment (chisquared test). The software was also used for analyzing the results for FB markers in different subgroups (box-and-whisker plots, Mann Whitney test, and Fisher s exact test). RESULTS Information on virological outcome was available in 115 of 132 women who underwent invasive PD (Figure 1). Mean (range) maternal and gestational age at time of invasive PD was 33 years (19 44) and 24 gw (16 36), respectively. No pregnancy loss following AC and cordocentesis was observed in the 115 women, including 13 cases with early cordocentesis < gw. In ten of the 17 cases without suitable samples to confirm results of invasive PD, no procedure-related pregnancy loss usually defined as fetal demise within 2 weeks after intervention was observed. One intrauterine death occurred in a fetus with trisomy 21 within 1 week following intervention, and six pregnancies (with FB sampling gw) were lost to follow up beyond the inpatient 24-h observation period. Ultrasound Abnormal US findings were observed in 22% (25/115) of fetuses at initial invasive PD. In our cohort, congenital cytomegalovirus infection (ccmv) was present in 64% (16/25) of fetuses with and 34% (31/90) of fetuses without US abnormalities. Table 1 shows US abnormalities of infected fetuses ranging from mild to severe. Seven fetuses presented with both cerebral and extracerebral abnormalities, three showed only cerebral, and six only extracerebral US abnormalities. Single US abnormalities were found in six cases (37.5%). The estimated time between seroconversion and detection of cerebral abnormalities was >10 weeks in eight of ten cases. US abnormalities in the nine uninfected fetuses were isolated polyhydramnion (n = 3) and hyperechogenic bowel, either isolated (n = 2) or together with placentomegaly (n =1) or macroglossia, abnormal kidney (n = 1). One fetus presented with white spot at gw, another with mild periventricular echogenicity and borderline ventriculomegaly (10.0 mm) at gw. Invasive prenatal diagnosis Figure 2 shows the results of invasive PD and virological outcome in relation to gestational age at intervention. Invasive PD was positive in AF and/or FB in 36 of 115 pregnancies. In 35 of these cases, ccmv infection was confirmed at birth. A total of 12 newborns with negative PD showed evidence of ccmv. False negative PD results were observed at any gw. US abnormalities were present in three of six fetuses with a positive PD < gw and in 13 of 29 fetuses with a positive PD gw (see also Table 1). Figure 2 Results of CMV PD and virological outcome in relation to gestational age at time of intervention (n = 115). PD pos (neg) indicates CMV DNA positive (negative) in AF and/or FB. AF, amniotic fluid; FB, fetal blood; ccmv, congenital CMV infection; CMV, cytomegalovirus; PD, prenatal diagnosis. [Colour figure can be viewed at wileyonlinelibrary.com]

5 Prenatal diagnosis of congenital cytomegalovirus infection 393 Table 2a presents the diagnostic values of the various detection methods applied in AF and FB. CMV PCR and rapid cell culture in AF gave concordant results in 99.1% (110/111). In one pregnancy with confirmed ccmv infection, culture was negative, and PCR was low positive (1700 IU/mL). In this case, the child was asymptomatically infected. The detection of CMV DNA and CMV-specific IgM in FB showed an overall concordance of 84.9% (90/106). Lowest diagnostic values were observed for IgM detection in FB [sensitivity 48.9%; negative predictive value (NPV) 71.6%], but sensitivity increased to 81.3% (13/16), if only symptomatic fetuses were considered. CMV DNA detection in AF and FB gave concordant results in 99% (101/102). In one FB sample of an uninfected fetus, CMV DNA (<125 IU/mL) and CMV IgM (index 1.8) were detected at low level, whereas AF and neonatal urine were CMV PCRnegative and culture-negative. The diagnostic value of CMV DNA detection in AF in relation to gestational age and time between estimated seroconversion and AC is presented in Table 2b. Because the proportion of false negative PD results was very high (33%, 2/6) in women with AC gw, we selected a subgroup of women with invasive PD gw (n = 105) for this analysis. The interval between estimated time of seroconversion and invasive PD was 8 and >8 weeks in 25 and 52 pregnancies, respectively. In the remaining 28 cases (including two cases with false negative PD), patient s history and laboratory results did not allow a classification into these time categories. Specificity and positive predictive value of CMV DNA detection in AF was 100% in all subgroups. Sensitivity and NPV were significantly higher in pregnancies with a time interval of >8 weeks compared with those with 8 weeks between seroconversion and AC (p < 0.05). Restriction to AC gw did not increase sensitivity and NPV. A total of 49 women (42.6%) received CMV HIG (Cytotect CP, Biotest Dreieich, Germany) on an individual basis for prevention of fetal CMV infection before they attended the prenatal medicine unit, Leverkusen. There were great variations between treatment protocols concerning, for example, time interval between seroconversion and first dose, dosage, number of applications, time interval between applications. Sensitivity and NPV of invasive PD in women with HIG were 84.2% (16/19) and 90.9% (30/33) and in women without HIG 67.9% (19/28) and 80.4% (37/46), Table 2a Diagnostic value of methods in AF and/or FB in 115 cases Sample a Method Result Congenital infection Overall Sensitivity % Specificity % concordance Yes No (95%CI) (95%CI) PPV % NPV % % (n/n) AF (n = 111) Rapid cell culture Positive Negative (55-83) (95-100) (110/111) PCR Positive Negative (57-85) (95-100) FB (n = 106) lgm Positive b Negative (34-64) (86-99) (90/106) PCR Positive Negative (60-87) (91-100) AF, amniotic fluid; CI, confidence interval; FB, fetal blood; NPV, negative predictive value; PPV, positive predictive value. a AF only: n = 9; FB only: n = 4; AF and FB: n =102. b includes two borderline results. Table 2b Diagnostic value of CMV DNA detection in amniotic fluid in relation to gestational week and time interval between estimated seroconversion and amniocentesis Gestational weeks + days Time interval N Result PCR Congenital infection Sensitivity % Yes No (95%CI) NPV % a 8 weeks 25 Positive Negative 6 14 (17-77) >8 weeks 52 Positive Negative 2 30 (71-99) weeks 19 Positive Negative 6 10 (7.5-70) >8 weeks 37 Positive Negative 2 17 (68-99) CI, confidence interval; CMV, cytomegalovirus; NPV, negative predictive value. a Patients with amniocentesis at to (n = 21) and (n = 56).

6 394 M. Enders et al. respectively. The observed differences were statistically not significant (p > 0.05). Non-virological markers in fetal blood Figure 3 compares beta-2 microglobulin, platelet count, GGT, and AST levels in infected fetuses with positive PD and abnormal (A1) or normal (A2) US findings, in uninfected fetuses (B) and in fetuses with false negative PD (C). Infected fetuses (A1 and A2) had significantly higher beta-2 microglobulin (median 7.8 mg/ L, range ), GGT values (median 122 U/L, range ), and AST values (median 32 U/L, range ) than uninfected fetuses (beta-2 microglobulin: median 3.8 mg/l, range ; GGT: median 45 U/L, range ; AST: median 26 U/L, range 17 52). Platelet count was significantly lower in infected fetuses (median e9/L, range ) versus uninfected fetuses (median e9/L, range ). In fetuses with false negative PD, all four FB markers were mostly within the range of the uninfected group. No significant differences were observed for all four FB markers between infected fetuses with and without US abnormalities (A1 vs A2, p 0.05). However, beta-2 microglobulin values above 14 mg/l, platelet counts below 80 10e9/L, GGT levels above 350 U/L, and AST levels above 45 U/L were exclusively found in FB samples of symptomatically infected fetuses. As shown in Table 1, all five symptomatic fetuses with US abnormalities graded as mild had beta-2 microglobulin values below 14 mg/ L and platelet counts above e9/L. This was also true for all 23 asymptomatic infected fetuses (Table 3), except for one with a platelet count of 82 10e9/L. Interestingly, this fetus developed US abnormalities 7 weeks after invasive PD (data not shown). In summary, beta-2 microglobulin values above 14 mg/l or platelet counts below e9/L were significantly more often found in fetuses with severe US abnormalities (beta- 2: 6/11; platelets: 8/11) compared with fetuses with no or mild abnormalities (beta-2: 0/23; platelets: 1/23; p < 0.001). Viral markers in fetal blood and amniotic fluid As presented in Figure 4, symptomatic infected fetuses had significantly higher IgM index values (median 4.35, range ) and higher viral load (median IU/mL, range e3) in FB than fetuses with normal US findings (IgM index median 0.65, range ; viral load median 3225 IU/ ml, range e3). All symptomatic fetuses revealed a viral load of >1500 IU/mL in FB. In AF, no statistically significant difference of viral load between symptomatic (median e3, range e3 IU/mL) and asymptomatic (median e3, range e3 IU/mL) infected fetuses was observed (p = 0.21). DISCUSSION In our series of 115 pregnant women with primary CMV infection and known virological outcome, we reevaluated the factors that potentially influence the diagnostic value of PD. On the whole, US imaging was not able to reliably diagnose intrauterine CMV infection. This observation is in accordance with the findings of others. 1 However, the presence of multiple and/or cerebral US abnormalities was suggestive of intrauterine infection. In our cohort, severe fetal symptoms Figure 3 Box plots of (a) beta-2 microglobulin, (b) platelet count, (c) GGT, and (d) AST levels. A1 indicates infected fetuses (PD positive) with ultrasound abnormalities, A2 indicates infected fetuses (PD positive) with normal ultrasound, B indicates uninfected fetuses, and C indicates fetuses with false negative PD. AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; PD, prenatal diagnosis

7 Prenatal diagnosis of congenital cytomegalovirus infection 395 Table 3 Virological findings and fetal blood markers in infected asymptomatic fetuses (n = 18) Case no. Maternal infection in Gestational age at invasive PD Amniotic fluid Fetal blood CMV DNA CMV DNA CMV-specific Beta-2 MG Platelets GGT AST [ 10 3 IU/mL] [ 10 3 IU/mL] IgM [Index] a [mg/ml] [ 10e9/L] [U/L] [U/L] b st/2nd trim st/2nd trim nd trim st/2nd trim st/2nd trim st/2nd trim st/2nd trim st/2nd trim rd trim nd/3rd trim nd trim AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; trim., trimester of pregnancy. a Evaluation IgM ELA: index: <0.9 negative; borderline; >1.1 positive. b Abnormal ultrasound findings (periventricular hyperechogenicity) at gestational week (7 weeks after invasive PD). mostly occurred following periconceptional or first trimester infections. These findings are in accordance with previous studies showing that maternal infection in the first trimester is a risk factor for symptomatic fetal infection. 17,18 Nevertheless, second trimester infections may also lead to symptoms in the fetus or the newborn. 18,19 As recently reviewed, the value of invasive PD depends on several factors (i.e. method used, type of specimen, and timing of procedure). 20,21 PCR has been suggested to be more sensitive than culture for detection of CMV in AF In our series, the agreement between the two methods was 99.1% (110/111) giving a nearly equal sensitivity for PCR and culture of 72.7% and 70.5%, respectively. The discordant sample was obtained at gw and <6 weeks after maternal seroconversion. Hence, the false negative culture result can be explained by the short time interval between seroconversion and AC. A high concordance between viral culture and PCR was also reported by Leyder et al. (100%), 25 Feldman et al. (100%), 19 and Gouarin et al. (98.9%). 26 In the latter study, the only discrepancy was a PCR-positive/ culture-negative finding in a child uninfected at birth resulting in a slightly superior specificity of the culture method. Such false positive PCR results in AF have also been reported in other studies. 7,22,23,27,28 Because a positive result in PD may lead to irreversible decisions, we and others 21,29 recommend that at least two virological assays should be performed. Prenatal diagnosis of ccmv may also be attempted via cordocentesis. Detection of specific IgM in FB to diagnose congenital CMV infection has a well-known limited sensitivity, 8 which is in line with our observations. For detection of CMV DNA in FB, reported sensitivity of PCR ranged from 41% to 92%. 6 8,23 One strength of our study is a large number of FB samples (n = 102) concomitantly investigated with AF. In contrast to early studies, 6,23 we showed that the sensitivity of CMV DNA detection in AF was not superior to that in FB (72.7% vs 75.6%). However, in view of the higher complication rate associated with FB sampling, AC is the method of choice for diagnosis of fetal CMV infection. A very important factor for the sensitivity of invasive PD is the timing of AC. 7,8,30 In our series, for women with AC weeks gestation, sensitivity of PD was about twice as high in cases with an interval of >8 weeks between maternal seroconversion and AC (90.9%) compared with those with an interval of 8 weeks (45.5%). Current recommendations on timing of AC 2 4 are based on limited data. 5 7,30,31 In the only study analyzing both factors concomitantly (gestational age and time interval), the authors concluded that for high sensitivity, the delay between AC and seroconversion was more important than gestational age. 30 Cases with false negative AC < gw reported in the literature are often poorly defined with regard to the time interval between maternal infection and AC and may include a significant number of cases with early AC before In our study, we did not observe a

8 396 M. Enders et al. Figure 4 Box plots of (a) CMV-specific IgM and (b) viral load in fetal blood in infected fetuses (PD positive); (A1) with US abnormalities and (A2) with normal US. CMV, cytomegalovirus; PD, prenatal diagnosis; US, ultrasound difference in sensitivity whether the AC was performed gw or gw as long as the time interval was >8 weeks. Therefore, invasive PD might be approached gw in women with preconceptional/periconceptional or early first trimester primary infection without a significant loss of sensitivity. This option of earlier diagnosis needs further investigation, as in our study, the number of cases with intrauterine transmission and AC before gw was limited. An earlier PD is desired for several reasons: Negative results may reassure parents. In case of positive results, early measures may be taken if considered appropriate, like termination of pregnancy before the age of fetal viability or initiation of treatment before significant damage to the central nervous system has occurred. It is noteworthy that a negative PD cannot rule out congenital infection at birth with absolute certainty, even when performed under the best conditions. One possible explanation is delayed intrauterine transmission after the performance of AC. This hypothesis is supported by recently published data showing that infected children with false negative PD performed after completed 21 weeks gestation and more than 6 to 8 weeks after seroconversion were less likely to have symptoms or hearing loss at birth compared with a control group of children with positive AF. None of the children with a false negative PD had neurological sequelae at long-term follow-up. 32 It remains to be determined whether a false negative PD result at to gw is associated with the same favorable outcome. In addition to virological (CMV DNA and CMV IgM), we also investigated several non-virological (platelet count, beta-2 microglobulin, GGT, and AST) laboratory parameters in relation to fetal infection and US findings at time of PD. We could not evaluate the prognostic value of these parameters concerning the natural course of infection, because most of our infected fetuses received prenatal treatment. The outcome of pregnancy with respect to treatment is subject of a forthcoming publication. Infected fetuses with abnormal US findings had significantly higher IgM index values and a higher viral load in FB compared with asymptomatically infected fetuses. However, there was a great overlap between both groups restricting the value of these markers in differentiating symptomatic from asymptomatic infections. A viral load of <1500 IU/mL in FB was only observed in asymptomatic fetuses. As previously reported, viral load in AF was not associated with the presence of US abnormalities. 33,34 Our values for non-virological laboratory parameters in uninfected fetuses correlated well with those of others. 12,35 37 Infected fetuses presented with significantly higher beta-2 microglobulin, GGT, and AST levels and significantly lower platelet counts compared with noninfected fetuses. Similar findings have been described by Fabbri et al. 12 Best discrimination between infected and uninfected fetuses was achieved by beta-2 microglobulin. Our observation that in cases with a false negative PD, all non-virological parameters were within the normal range suggests that the fetus was either in the incubation period or that intrauterine transmission had not occurred at the time of intervention. In contrast to Fabbri et al., 12 we did not observe significant differences in the investigated non-virological parameters between infected fetuses with and without US abnormalities. This may be explained by the fact that we did not consider clinical findings at birth. Nevertheless, we found abnormal non-virological FB markers beyond a certain threshold (beta-2 microglobulin >14 mg/l, platelet counts <120 10e9/L) alone or in combination significantly more often in symptomatic fetuses with severe US abnormalities compared with fetuses with no or mild abnormalities. Generally, in case of a positive PD, the prognosis for the fetus is assessed using targeted fetal US. However, grading and evaluation of sonographic findings may vary between observers, and clinical relevance of non-specific or isolated US signs is often unclear. In addition, US findings may change with time, and various studies indicate that 7% to 52% of initially asymptomatic fetuses will develop symptoms later during gestation. 1,11,38 In this context, FB markers alone or in combination may be a helpful supplemental tool for identifying fetuses at increased risk for severe disease and for selecting those which might profit from prenatal treatment, for example, with valacyclovir as recently suggested by Lereuz-Ville et al. 39 Cutoffs and optimal combination of various virological and non-virological markers for prediction of outcome are still under investigation. 12,38,40

9 Prenatal diagnosis of congenital cytomegalovirus infection 397 CONCLUSION Our results confirm that detection of CMV in AF by PCR or culture is a reliable method to diagnose fetal infection as long as AC is performed >8 weeks after maternal seroconversion. In contrast to current recommendations, we hypothesize that AC as early as gw has an adequate diagnostic sensitivity. This has to be confirmed by larger (preferably prospective) clinical studies. In our cohort, cordocentesis does not add diagnostic sensitivity nor do FB markers enable a clear-cut differentiation between symptomatic and asymptomatic fetuses. However, they may be beneficial in the individual management of pregnant women with confirmed congenital CMV infection. ACKNOWLEDGEMENTS The excellent technical assistance of Ursula Bäder, Yvonne Schimpf, Claudia Weber, and Andrea Maurer is gratefully acknowledged. Ethics approval Not applicable (retrospective study). WHAT S ALREADY KNOWN ABOUT THIS TOPIC? According to current recommendations, amniocentesis should be performed after gestational week 21 and >6 to 8 weeks after maternal cytomegalovirus seroconversion. The role of virological and non-virological fetal blood markers as indicators of the severity of congenital disease is under investigation. WHAT DOES THIS STUDY ADD? Sensitivity of amniocentesis at to gw should be reevaluated in a prospective study. Analysis of FB markers may be helpful in individual management of pregnancies with confirmed fetal infection. REFERENCES 1. Guerra B, Simonazzi G, Puccetti C, et al. Ultrasound prediction of symptomatic congenital cytomegalovirus infection. Am J Obstet Gynecol 2008;198: Society for Maternal-Fetal Medicine (SMFM), Hughes BL, Gyamfi- Bannermann C. Diagnosis and antenatal management of congenital cytomegalovirus infection. Am J Obstet Gynecol 2016;215:B5 B Yinon Y, Farine D, Yudin MH, et al. Cytomegalovirus infection in pregnancy. J Obstet Gynaecol Can 2010;32: Public Health England. Cytomegalovirus Serology. UK Standards for Microbiology Investigations. V 28 Issue URL uk/uk-standards-for-microbiology-investigations-smi-quality-andconsistency-in-clinical-laboratories [accessed on 8 September 2016] 5. Donner C, Liesnard C, Content J, et al. Prenatal diagnosis of 52 pregnancies at risk for congenital cytomegalovirus infection. Obstet Gynecol 1993;82: Liesnard C, Donner C, Brancart F, et al. Prenatal diagnosis of congenital cytomegalovirus infection: prospective study of 237 pregnancies at risk. Obstet Gynecol 2000;95: Enders G, Bader U, Lindemann L, et al. Prenatal diagnosis of congenital cytomegalovirus infection in 189 pregnancies with known outcome. Prenat Diagn 2001;21: Revello MG, Gerna G. Pathogenesis and prenatal diagnosis of human cytomegalovirus infection. J Clin Virol 2004;29: Hohlfeld P, Maillard-Brignon C, Vaudaux B, Fawer CL. Cytomegalovirus fetal infection: prenatal diagnosis. Obstet Gynecol 1991;78: Lynch L, Daffos F, Emanuel D, et al. Prenatal diagnosis of fetal cytomegalovirus infection. Am J Obstet Gynecol 1991;165: Benoist G, Salomon LJ, Jacquemard F, et al. The prognostic value of ultrasound abnormalities and biological parameters in blood of fetuses infected with cytomegalovirus. BJOG 2008;115: Fabbri E, Revello MG, Furione M, et al. Prognostic markers of symptomatic congenital human cytomegalovirus infection in fetal blood. BJOG 2011;118: Enders G, Daiminger A, Bader U, et al. The value of CMV IgG avidity and immunoblot for timing the onset of primary CMV infection in pregnancy. J Clin Virol 2013;56: Enders M, Daiminger A, Exler S, Enders G. Author s reply to Letter to the Editor regarding Comments on an alternative calculation of cytomegalovirus IgG avidity indexes on VIDAS [J Clin Virol 60 (2014) 76 77]. J Clin Virol 2014;61: Schalasta G, Eggers M, Schmid M, Enders G. Analysis of human cytomegalovirus DNA in urines of newborns and infants by means of a new ultrarapid real-time PCR-system. J Clin Virol 2000;19: Daiminger A, Schalasta G, Betzl D, Enders G. Detection of human cytomegalovirus in urine samples by cell culture, early antigen assay and polymerase chain reaction. Infection 1994;22: Picone O, Vauloup-Fellous C, Cordier AG, et al. A series of 238 cytomegalovirus primary infections during pregnancy: description and outcome. Prenat Diagn 2013;33: Zavattoni M, Rustico M, Tassis B, et al. Risk of congenital disease in 46 infected fetuses according to gestational age of primary human cytomegalovirus infection in the mother. J Med Virol 2016;88: Feldman B, Yinon Y, Tepperberg OM, et al. Pregestational, periconceptional, and gestational primary maternal cytomegalovirus infection: prenatal diagnosis in 508 pregnancies. Am J Obstet Gynecol 2011;205: Lazzarotto T, Guerra B, Gabrielli L, et al. Update on the prevention, diagnosis and management of cytomegalovirus infection during pregnancy. Clin Microbiol Infect 2011;17: Revello MG, Gerna G. State of the art and trends in cytomegalovirus diagnostics. In Cytomegaloviruses: From Molecular Pathogenesis to Intervention, Reddehase MJ (ed)2. Caister Academic Press: Norfolk, UK; p Guerra B, Lazzarotto T, Quarta S, et al. Prenatal diagnosis of symptomatic congenital cytomegalovirus infection. Am J Obstet Gynecol 2000;183: Lazzarotto T, Guerra B, Spezzacatena P, et al. Prenatal diagnosis of congenital cytomegalovirus infection. J Clin Microbiol 1998;36: Revello MG, Baldanti F, Furione M, et al. Polymerase chain reaction for prenatal diagnosis of congenital human cytomegalovirus infection. J Med Virol 1995;47: Leyder M, Vorsselmans A, Done E, et al. Primary maternal cytomegalovirus infections: accuracy of fetal ultrasound for predicting sequelae in offspring. Am J Obstet Gynecol DOI: /j. ajog Gouarin S, Palmer P, Cointe D, et al. Congenital HCMV infection: a collaborative and comparative study of virus detection in amniotic fluid by culture and by PCR. J Clin Virol 2001;21: Lazzarotto T, Varani S, Guerra B, et al. Prenatal indicators of congenital cytomegalovirus infection. J Pediatr 2000;137: Revello MG, Lilleri D, Zavattoni M, et al. Prenatal diagnosis of congenital human cytomegalovirus infection in amniotic fluid by nucleic acid sequence-based amplification assay. J Clin Microbiol 2003;41: Lazzarotto T, Gabrielli L, Guerra B, et al. Diagnosis and prognosis of congenital CMV infection: a case report and review of the literature. Scand J Clin Lab Invest Suppl 2014;244: Bodeus M, Hubinont C, Bernard P, et al. Prenatal diagnosis of human cytomegalovirus by culture and polymerase chain reaction: 98 pregnancies leading to congenital infection. Prenat Diagn 1999;19: Donner C, Liesnard C, Brancart F, Rodesch F. Accuracy of amniotic fluid testing before 21 weeks gestation in prenatal

10 398 M. Enders et al. diagnosis of congenital cytomegalovirus infection. Prenat Diagn 1994;14: Bilavsky E, Pardo J, Attias J, et al. Clinical implications for children born with congenital cytomegalovirus infection following a negative amniocentesis. Clin Infect Dis 2016;63: Picone O, Costa JM, Leruez-Ville M, et al. Cytomegalovirus (CMV) glycoprotein B genotype and CMV DNA load in the amniotic fluid of infected fetuses. Prenat Diagn 2004;24: Benoist G, Leruez-Ville M, Magny JF, et al. Management of pregnancies with confirmed cytomegalovirus fetal infection. Fetal Diagn Ther 2013;33: Weiner CP, Sipes SL, Wenstrom K. The effect of fetal age upon normal fetal laboratory values and venous pressure. Obstet Gynecol 1992;79: Hu S, Wu J, Lu D, et al. Reference values of fetal serum beta2- microglobulin in the Chinese: evaluation of its clinical usefulness. Clin Chem Lab Med 2014;52: Van den Hof MC, Nicolaides KH. Platelet count in normal, small, and anemic fetuses. Am J Obstet Gynecol 1990;162: Leruez-Ville M, Stirnemann J, Sellier Y, et al. Feasibility of predicting the outcome of fetal infection with cytomegalovirus at the time of prenatal diagnosis. Am J Obstet Gynecol 2016;215: Leruez-Ville M, Ville Y. Optimum treatment of congenital cytomegalovirus infection. Expert Rev Anti Infect Ther 2016;14: Zavattoni M, Lombardi G, Rognoni V, et al. Maternal, fetal, and neonatal parameters for prognosis and counseling of HCMV congenital infection. J Med Virol 2014;86: