Abstract
The aim of this prospective cohort study was to evaluate the efficacy of maternal screening for congenital cytomegalovirus infection (CCI) using cytomegalovirus (CMV) immunoglobulin G (IgG) and the IgG avidity index (AI).
Pregnant women underwent screening of CMV IgG and AI measurements. IgG-negative women underwent remeasurement of IgG after educational intervention. Women with an AI ≤45% received further examinations, including measurement of CMV IgM. All newborns received polymerase chain reaction analyses of the urine, and CCI was diagnosed by the detection of CMV-DNA in the urine. Primary infection was defined as an AI <35% and/or positive IgM (>1.20 index). Serum samples from women with an AI >45% were stored, and the IgM levels were measured after delivery. The efficacy of AI and IgM for CCI screening was compared.
A total of 1562 (71.2%) women tested positive for IgG. In this study, 10 newborns with CCI were detected. The presence of infection in 3 newborns from mothers with primary infection was predicted by screening of IgG and AI <35%. However, infection in 7 newborns from women with nonprimary infection could not be predicted by screening of CMV IgG, AI <35%, or IgM. The application of an AI <35% for CCI screening yielded 22.2% sensitivity, 95.0% specificity, 2.5% positive predictive value, and 99.5% negative predictive value and was similar to that of IgM (11.1% sensitivity, 93.2% specificity, 0.9% positive predictive value, and 92.7% negative predictive value).
Maternal screening using CMV IgG and AI can identify pregnancies with CCI from primary infection, but overlooks a number of those from nonprimary infection.
(See the and .)
Human cytomegalovirus (CMV) is one of the most common causes of mother-to-child infection worldwide. The prevalence of congenital CMV infection is 0.2%–2.4% in newborns in developed countries [1–3], and 10%–15% of infected fetuses show symptomatic congenital CMV infection at birth. The clinical manifestations of CMV infection, including fetal growth restriction, low birth weight, and central nervous system and multiple organ involvement, can be so severe that they lead to a high perinatal mortality rate and major neurological sequelae in approximately 90% of surviving infants. In addition, 10%–15% of infants with asymptomatic congenital infection develop long-term sequelae, including progressive sensorineural hearing difficulty and mental retardation [4]. It was recently demonstrated that early diagnosis and early intervention with antiviral drugs can improve neurological outcomes in symptomatic congenital CMV–infected infants [5, 6]. Therefore, prenatal detection of pregnancies at high risk of congenital CMV infection may be important for the accurate diagnosis of infected newborns and early commencement of antiviral therapy on symptomatic newborns.
The risk of virus transmission to the fetus is highest during pregnancy for women with primary CMV infection [7, 8]. Maternal serological screening is considered effective for detecting pregnant women with primary CMV infection [9]. For this purpose, maternal blood tests for CMV-specific immunoglobulin G (IgG) and CMV-specific immunoglobulin M (IgM) are widely employed [9]. In particular, testing for maternal serum CMV IgM is commonly used to identify primary infection. Positive results of CMV IgM yield high sensitivity, low specificity, and low positive predictive value for the detection of primary infection [1, 10]. Pregnant women can produce CMV IgM during reinfection and reactivation of the virus [10]. In addition, CMV IgM may persist for months after primary infection and has been detected in some pregnant women 6–9 months after the primary infection [11]. However, serum CMV IgG avidity tests are highly specific and sensitive for the detection of primary infection [12], and CMV IgG avidity is considered the gold standard for the identification of primary infection during pregnancy [13–16]. Nonetheless, no studies have performed universal screening based on CMV IgG avidity during pregnancy, and the efficacy of this parameter for predicting congenital CMV infection is still undetermined. A prospective cohort study was conducted to evaluate the efficacy of maternal universal screening based on CMV IgG avidity for the prediction of congenital CMV infection, compared with that based on CMV IgM.
METHODS
Study Design and Participants
The institutional review board at Kobe University Hospital approved this prospective cohort study (reference number 922), and written informed consent was obtained from all participants. All pregnant women who visited or were referred to Kobe University Hospital between February 2010 and April 2016 underwent maternal serological CMV screening.
These women underwent initial blood screening for CMV IgG before 22 gestational weeks (GW). CMV IgG–negative women received educational intervention to prevent primary infection during the rest of their pregnancies and CMV IgG levels were measured again at 34–36 GW. CMV IgG seroconversion in IgG-negative women was regarded as primary CMV infection during pregnancy. All CMV IgG–positive women were tested for serum IgG avidity 2–4 weeks after the initial CMV IgG measurements. Pregnant women who were referred after 22 GW received simultaneous CMV IgG and IgG avidity testing at the first visit.
All women who had a CMV IgG avidity index (AI) ≤ 45% underwent a series of examinations for CMV infection. These examinations included the measurement of serum CMV IgM levels, blood antigenemia (C7-HRP) testing (TEIJIN CMV antigen test, Teijin, Tokyo, Japan), and quantitative polymerase chain reaction (PCR) analyses for CMV-DNA in maternal serum, urine, and uterine cervical secretion. In the present study, women who had an AI <35% and/or a positive result for CMV IgM were considered to have primary CMV infection during pregnancy. Women with an AI >45% were considered to have nonprimary infection during pregnancy, and their serum samples were stored at −80°C. Serum CMV IgM levels were later measured to evaluate their efficacy for the prediction of congenital CMV infection compared with the efficacy of IgG avidity.
All newborns received polymerase chain reaction (PCR) analyses of the urine, and congenital infection was diagnosed with the detection of CMV-DNA in the urine. Women who were referred because of positive results for CMV IgM, those who were referred after 37 GW, and those who had deliveries in other hospitals were excluded from the present study.
Procedures
Serological tests for CMV-specific IgG (negative: <230, borderline: 230–240, and positive: >240) were performed using the Enzygnost assay (Siemens Healthcare Diagnostics, Tokyo, Japan). CMV-specific IgM (negative: <0.8, borderline: 0.8–1.2, and positive: >1.2 index) was measured using enzyme immunoassay kits produced by Denka Seiken (Tokyo, Japan). CMV-specific IgG avidity was measured as described previously (Aisenkai Nichinan Hospital, Miyazaki, Japan) [17]. The C7-HRP assay (TEIJIN CMV antigen test) (positive: ≥1 positive cell/slide) was used for CMV antigenemia testing. Real-time PCR analyses for CMV-DNA in maternal serum, urine, and uterine cervical secretion were performed (Special References Laboratories, Tokyo, Japan) (positive, ≥ 1.0 × 102 copies/mL).
Urine samples were collected from newborns onto filter paper within 1 week after birth and the presence of CMV-DNA was assessed as described previously [18]. Liquid urine samples were obtained from CMV-positive newborns, and the CMV-DNA copy number was determined using real-time quantitative PCR. The presence of congenital CMV infection was confirmed by positive PCR results for CMV-DNA in the liquid urine samples of the newborns [19]. All newborns with positive PCR results for CMV-DNA in the urine received a workup to identify the symptoms of congenital CMV infection. Ophthalmoscopy, cerebral ultrasound, physical and neurological examinations, head computed tomography, head magnetic resonance imaging, and an auditory brainstem response (ABR) test were performed. Symptomatic congenital CMV infection was diagnosed when newborns with positive PCR results had microcephaly, hepatosplenomegaly/hepatitis, thrombocytopenia, abnormality of brain images, CMV-associated retinopathy, or abnormal ABR [19, 20].
The efficacy of maternal serological screening based on IgG AI (cutoff values: ≤45% or <35%) and positive results for CMV IgM (>1.2 index) for the prediction of congenital CMV infection was compared.
Statistical Analysis
Clinical characteristics and laboratory findings were compared between pregnancies with congenital CMV infection and those without congenital infection. Differences between the 2 groups were analyzed using the Mann-Whitney U test, Fisher exact test, and the χ2 test. Statistical significance was considered present at a P value <.05. All statistical analyses were performed using SPSS software, version 19 (SPSS Inc, Chicago, Illinois).
RESULTS
A flowchart of the subjects in this prospective cohort study is shown in Figure 1. Four newborns with symptomatic congenital CMV infections (0.18%) and 6 with asymptomatic congenital infections (0.27%) were identified in the study.
Flow diagram for the study participants, results of maternal serological screening, and occurrence of congenital cytomegalovirus (CMV) infection among all subjects. During the study period, 2634 pregnant women underwent blood screening for CMV immunoglobulin G (IgG) and CMV IgG avidity. After applying the exclusion criteria, 2193 pregnant women (GW, gestational weeks) were enrolled in this study. A total of 631 (28.8%) pregnant women had negative results for CMV IgG, and 5 (0.8%) of these women had IgG seroconversion. One of the 5 women with IgG seroconversion had asymptomatic congenital CMV infection. A total of 1562 of the 2193 (71.2%) pregnant women had positive results for CMV IgG. A total of 183 pregnant women with an avidity index (AI) ≤45% underwent measurements of CMV immunoglobulin M (IgM) and C7-HRP, and polymerase chain reaction (PCR) analyses for CMV-DNA in maternal serum, urine, and uterine cervical secretion. Eighty-eight of the 183 (48.1%) pregnant women with an AI ≤45% had an AI <35% and/or positive results for CMV IgM, and 2 had congenital infection. A total of 1379 women had an AI >45%, and 7 of these women had congenital CMV infection.
Table 1 shows the clinical characteristics of the 2193 women evaluated. Seven hundred forty-six (34.0%) were referrals. These referred women had obstetric complications including multiple pregnancies, low-lying placenta, placenta previa, threatened premature labor, and fetal growth restriction (n = 339 [45.4%]), maternal complications including hypertensive disorders, diabetes mellitus, autoimmune diseases, and thyroid diseases (n = 218 [29.2%]), and their move to parent’s home (n = 189 [25.3%]).
Clinical Characteristics of 2193 Women
| Characteristic | All (n = 2193) | Congenital Infection (n = 10) | No Congenital Infection (n = 2183) | P Value |
|---|---|---|---|---|
| Age, y | 33.2 ± 5.5 | 32 (28–40) | 34 (14–49) | .5 |
| Gravidity | 1.3 ± 1.5 | 1 (0–3) | 1 (0–13) | .3 |
| Parity | 0.6 ± 0.8 | 0 (0–2) | 0 (0–6) | .4 |
| BMI prior to pregnancy, kg/m2 | 21.4 ± 3.7 | 22.9 (17.4–27.7) | 20.5 (13.8–44.9) | .1 |
| Referral | 34.0% | 60.0% | 33.9% | .1 |
| Maternal fever or flulike symptoms | 16.6% | 50.0% | 16.4% | .02 |
| Gestational weeks at the initial CMV IgG measurements | 17.9 ± 8.5 | 17 (8–32) | 16 (4–36) | .4 |
| Gestational weeks at delivery | 37.4 ± 2.9 | 36 (26–40) | 38 (22–42) | .1 |
| Birth weight, g | 2727.7 ± 634.2 | 2362 (1080–3320) | 2820 (314–4748) | .06 |
| Characteristic | All (n = 2193) | Congenital Infection (n = 10) | No Congenital Infection (n = 2183) | P Value |
|---|---|---|---|---|
| Age, y | 33.2 ± 5.5 | 32 (28–40) | 34 (14–49) | .5 |
| Gravidity | 1.3 ± 1.5 | 1 (0–3) | 1 (0–13) | .3 |
| Parity | 0.6 ± 0.8 | 0 (0–2) | 0 (0–6) | .4 |
| BMI prior to pregnancy, kg/m2 | 21.4 ± 3.7 | 22.9 (17.4–27.7) | 20.5 (13.8–44.9) | .1 |
| Referral | 34.0% | 60.0% | 33.9% | .1 |
| Maternal fever or flulike symptoms | 16.6% | 50.0% | 16.4% | .02 |
| Gestational weeks at the initial CMV IgG measurements | 17.9 ± 8.5 | 17 (8–32) | 16 (4–36) | .4 |
| Gestational weeks at delivery | 37.4 ± 2.9 | 36 (26–40) | 38 (22–42) | .1 |
| Birth weight, g | 2727.7 ± 634.2 | 2362 (1080–3320) | 2820 (314–4748) | .06 |
Data are expressed as the mean ± standard deviation, median (range), or percentage.
Abbreviations: BMI, body mass index; CMV, cytomegalovirus; IgG, immunoglobulin G.
Clinical Characteristics of 2193 Women
| Characteristic | All (n = 2193) | Congenital Infection (n = 10) | No Congenital Infection (n = 2183) | P Value |
|---|---|---|---|---|
| Age, y | 33.2 ± 5.5 | 32 (28–40) | 34 (14–49) | .5 |
| Gravidity | 1.3 ± 1.5 | 1 (0–3) | 1 (0–13) | .3 |
| Parity | 0.6 ± 0.8 | 0 (0–2) | 0 (0–6) | .4 |
| BMI prior to pregnancy, kg/m2 | 21.4 ± 3.7 | 22.9 (17.4–27.7) | 20.5 (13.8–44.9) | .1 |
| Referral | 34.0% | 60.0% | 33.9% | .1 |
| Maternal fever or flulike symptoms | 16.6% | 50.0% | 16.4% | .02 |
| Gestational weeks at the initial CMV IgG measurements | 17.9 ± 8.5 | 17 (8–32) | 16 (4–36) | .4 |
| Gestational weeks at delivery | 37.4 ± 2.9 | 36 (26–40) | 38 (22–42) | .1 |
| Birth weight, g | 2727.7 ± 634.2 | 2362 (1080–3320) | 2820 (314–4748) | .06 |
| Characteristic | All (n = 2193) | Congenital Infection (n = 10) | No Congenital Infection (n = 2183) | P Value |
|---|---|---|---|---|
| Age, y | 33.2 ± 5.5 | 32 (28–40) | 34 (14–49) | .5 |
| Gravidity | 1.3 ± 1.5 | 1 (0–3) | 1 (0–13) | .3 |
| Parity | 0.6 ± 0.8 | 0 (0–2) | 0 (0–6) | .4 |
| BMI prior to pregnancy, kg/m2 | 21.4 ± 3.7 | 22.9 (17.4–27.7) | 20.5 (13.8–44.9) | .1 |
| Referral | 34.0% | 60.0% | 33.9% | .1 |
| Maternal fever or flulike symptoms | 16.6% | 50.0% | 16.4% | .02 |
| Gestational weeks at the initial CMV IgG measurements | 17.9 ± 8.5 | 17 (8–32) | 16 (4–36) | .4 |
| Gestational weeks at delivery | 37.4 ± 2.9 | 36 (26–40) | 38 (22–42) | .1 |
| Birth weight, g | 2727.7 ± 634.2 | 2362 (1080–3320) | 2820 (314–4748) | .06 |
Data are expressed as the mean ± standard deviation, median (range), or percentage.
Abbreviations: BMI, body mass index; CMV, cytomegalovirus; IgG, immunoglobulin G.
The proportion of the presence of maternal fever or flu-like symptoms in 10 women who had congenital CMV infection was higher than that in 2183 women with no congenital infection (P < .05). Maternal age, gravidity or parity, body mass index prior to pregnancy, the proportion of referrals, GW at the initial CMV IgG measurement, GW at delivery, and the birth weight of newborns were not significantly different between the 2 groups.
The results of maternal serological screening are shown in Figure 1. Six hundred thirty-one pregnant women (28.8%) had negative results for CMV IgG during the initial tests, and 5 of the 631 (0.79%) women experienced IgG seroconversion during pregnancy. Among the 5 women with IgG seroconversion, 1 woman delivered a newborn with asymptomatic congenital CMV infection. Five of 455 (1.1%) women with negative results for CMV IgG before 25 GW experienced IgG seroconversion between 25 GW and 36 GW.
A total of 1562 pregnant women of the 2193 (71.2%) had positive results for CMV IgG, and underwent measurements for CMV IgG AI. One hundred eighty-three pregnant women with an AI ≤45% underwent CMV IgM measurements, C7-HRP tests, and PCR analyses for CMV-DNA in maternal serum, urine, and uterine cervical secretion.
Among 183 women with an AI ≤45%, 79 had an AI <35%, 15 had positive results for CMV IgM, 3 had positive PCR results for CMV-DNA in uterine cervical secretion, 2 had positive PCR results in maternal serum, and none had a positive PCR result in maternal urine or a positive C7-HRP result. Eighty-eight of the 183 (48.1%) pregnant women with an AI ≤45% had an AI <35% and/or positive results for CMV IgM, and were classified as having primary CMV infection during pregnancy. Two women had newborns with congenital CMV infection. One mother with symptomatic congenital infection had an AI <35%, positive CMV IgM, and a positive PCR result in uterine cervical secretion. The other woman with asymptomatic congenital infection had an AI <35% and a borderline result for CMV IgM.
Table 2 shows the clinical characteristics and laboratory findings for 3 women who had newborns with congenital CMV infection due to primary CMV infection during pregnancy.
Three Women Who Had Newborns With Congenital Cytomegalovirus (CMV) Infection due to Primary CMV Infection
| Case | Age, y | Gravidity/ Parity | GW at Flulike Symptoms | CMV IgG (GW at Measurements) | IgG Avidity Index, % | CMV IgM (Index) | CMV-DNA in Maternal Serum/Urine/Uterine Cervical Secretion | C7-HRP Tests | Gestational Age | Birth Weight, g | Symptoms of the Newborn |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 35 | 1/0 | 28 | 24290 (30) | 23.1 | 1.05 | –/ –/ – | – | 39 w 6 d | 2784 | None |
| 2 | 28 | 1/1 | 16 | 1726 (27) | 3.6 | 6.6 | –/ –/+ | – | 32 w 6 d | 1396 | Abnormal ABR, SGA |
| 3 | 31 | 1/0 | None | – (17)/9600 (35) | NA | NA | NA | NA | 36 w 3 d | 3238 | None |
| Case | Age, y | Gravidity/ Parity | GW at Flulike Symptoms | CMV IgG (GW at Measurements) | IgG Avidity Index, % | CMV IgM (Index) | CMV-DNA in Maternal Serum/Urine/Uterine Cervical Secretion | C7-HRP Tests | Gestational Age | Birth Weight, g | Symptoms of the Newborn |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 35 | 1/0 | 28 | 24290 (30) | 23.1 | 1.05 | –/ –/ – | – | 39 w 6 d | 2784 | None |
| 2 | 28 | 1/1 | 16 | 1726 (27) | 3.6 | 6.6 | –/ –/+ | – | 32 w 6 d | 1396 | Abnormal ABR, SGA |
| 3 | 31 | 1/0 | None | – (17)/9600 (35) | NA | NA | NA | NA | 36 w 3 d | 3238 | None |
Abbreviations: –, negative; +, positive; ABR, auditory brainstem response; CMV, cytomegalovirus; GW, gestational weeks; IgG, immunoglobulin G; IgM, immunoglobulin M; NA, not applicable; SGA, small for gestational age.
Three Women Who Had Newborns With Congenital Cytomegalovirus (CMV) Infection due to Primary CMV Infection
| Case | Age, y | Gravidity/ Parity | GW at Flulike Symptoms | CMV IgG (GW at Measurements) | IgG Avidity Index, % | CMV IgM (Index) | CMV-DNA in Maternal Serum/Urine/Uterine Cervical Secretion | C7-HRP Tests | Gestational Age | Birth Weight, g | Symptoms of the Newborn |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 35 | 1/0 | 28 | 24290 (30) | 23.1 | 1.05 | –/ –/ – | – | 39 w 6 d | 2784 | None |
| 2 | 28 | 1/1 | 16 | 1726 (27) | 3.6 | 6.6 | –/ –/+ | – | 32 w 6 d | 1396 | Abnormal ABR, SGA |
| 3 | 31 | 1/0 | None | – (17)/9600 (35) | NA | NA | NA | NA | 36 w 3 d | 3238 | None |
| Case | Age, y | Gravidity/ Parity | GW at Flulike Symptoms | CMV IgG (GW at Measurements) | IgG Avidity Index, % | CMV IgM (Index) | CMV-DNA in Maternal Serum/Urine/Uterine Cervical Secretion | C7-HRP Tests | Gestational Age | Birth Weight, g | Symptoms of the Newborn |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 35 | 1/0 | 28 | 24290 (30) | 23.1 | 1.05 | –/ –/ – | – | 39 w 6 d | 2784 | None |
| 2 | 28 | 1/1 | 16 | 1726 (27) | 3.6 | 6.6 | –/ –/+ | – | 32 w 6 d | 1396 | Abnormal ABR, SGA |
| 3 | 31 | 1/0 | None | – (17)/9600 (35) | NA | NA | NA | NA | 36 w 3 d | 3238 | None |
Abbreviations: –, negative; +, positive; ABR, auditory brainstem response; CMV, cytomegalovirus; GW, gestational weeks; IgG, immunoglobulin G; IgM, immunoglobulin M; NA, not applicable; SGA, small for gestational age.
Among 1379 women with an AI >45%, 92 had positive results for CMV IgM. The remaining 1287 women with an AI >45% and a CMV IgM ≤1.2 index were classified as having nonprimary CMV infection (Figure 1). Seven of the 1287 (0.54%) women with an AI >45% together with a CMV IgM ≤1.2 index had newborns with congenital CMV infection, including 3 with symptomatic and 4 with asymptomatic infections.
Table 3 shows a comparison of the efficacy of maternal blood screening for the prediction of congenital CMV infection. Screening modalities included an AI ≤45%, an AI <35%, and positive results for CMV IgM (>1.2 index). The efficacy of maternal screening based on AI <35% was similar to that based on CMV IgM.
Comparison of Efficacy of Maternal Blood Screening for Prediction of Congenital Cytomegalovirus Infection
| Maternal Blood Screening | Sensitivity, % | Specificity, % | PPV, % | NPV, % | Accuracy, % |
|---|---|---|---|---|---|
| CMV IgG avidity index ≤45% | 22.2 | 88.4 | 1.1 | 99.5 | 87.9 |
| CMV IgG avidity index <35% | 22.2 | 95.0 | 2.5 | 99.5 | 94.6 |
| CMV IgM >1.2 index | 11.1 | 93.2 | 0.9 | 99.5 | 92.7 |
| Maternal Blood Screening | Sensitivity, % | Specificity, % | PPV, % | NPV, % | Accuracy, % |
|---|---|---|---|---|---|
| CMV IgG avidity index ≤45% | 22.2 | 88.4 | 1.1 | 99.5 | 87.9 |
| CMV IgG avidity index <35% | 22.2 | 95.0 | 2.5 | 99.5 | 94.6 |
| CMV IgM >1.2 index | 11.1 | 93.2 | 0.9 | 99.5 | 92.7 |
Accuracy (%) is calculated by the following formula: (number of true positive population + number of true negative population) / number of all populations.
Abbreviations: CMV, cytomegalovirus; IgG, immunoglobulin G; IgM, immunoglobulin M; PPV, positive predictive value; NPV, negative predictive value.
Comparison of Efficacy of Maternal Blood Screening for Prediction of Congenital Cytomegalovirus Infection
| Maternal Blood Screening | Sensitivity, % | Specificity, % | PPV, % | NPV, % | Accuracy, % |
|---|---|---|---|---|---|
| CMV IgG avidity index ≤45% | 22.2 | 88.4 | 1.1 | 99.5 | 87.9 |
| CMV IgG avidity index <35% | 22.2 | 95.0 | 2.5 | 99.5 | 94.6 |
| CMV IgM >1.2 index | 11.1 | 93.2 | 0.9 | 99.5 | 92.7 |
| Maternal Blood Screening | Sensitivity, % | Specificity, % | PPV, % | NPV, % | Accuracy, % |
|---|---|---|---|---|---|
| CMV IgG avidity index ≤45% | 22.2 | 88.4 | 1.1 | 99.5 | 87.9 |
| CMV IgG avidity index <35% | 22.2 | 95.0 | 2.5 | 99.5 | 94.6 |
| CMV IgM >1.2 index | 11.1 | 93.2 | 0.9 | 99.5 | 92.7 |
Accuracy (%) is calculated by the following formula: (number of true positive population + number of true negative population) / number of all populations.
Abbreviations: CMV, cytomegalovirus; IgG, immunoglobulin G; IgM, immunoglobulin M; PPV, positive predictive value; NPV, negative predictive value.
Figure 2 shows the serological status of 2193 pregnant women and 10 newborns with symptomatic/asymptomatic congenital CMV infection.
Serological status of 2193 pregnant women and 10 newborns with symptomatic/asymptomatic congenital cytomegalovirus (CMV) infection. Numbers in the Venn diagram represent the number of pregnant women who had unique or nonunique results in serological tests. ● indicates newborns with symptomatic congenital CMV infection (n = 4); ◌ indicates newborns with asymptomatic congenital infection (n = 6). Abbreviations: AI, avidity index; CMV, cytomegalovirus; IgG, immunoglobulin G; IgM, immunoglobulin M.
DISCUSSION
This study evaluated 2193 pregnant women. One of the 5 (20.0%) women who experienced IgG seroconversion had a newborn with congenital CMV infection, which was caused by primary CMV infection during pregnancy. Two of the 88 (2.3%) women who had an AI <35% and/or positive results for CMV IgM had newborns with congenital infection, likely caused by primary CMV infection. In addition, 7 of the 1287 (0.5%) women who had an AI >45% and a CMV IgM ≤1.2 index had newborns with congenital infection, probably caused by nonprimary CMV infection (Figure 1). It has been reported that 30%–80% of women who experience IgG seroconversion during pregnancy have newborns with congenital CMV infection [21–23], and 0.5%–1.0% of women with nonprimary CMV infection have newborns with congenital CMV infection [24, 25]. These values for the transmission rates of CMV to the fetus are compatible with the results of the present cohort study.
Many investigators have reported that the measurement of CMV IgM and CMV IgG avidity is useful for the detection of maternal primary CMV infection during pregnancy [9, 15, 26–28]. The level of CMV IgM and/or CMV IgG avidity for the prediction of congenital CMV infection has also been assessed in high-risk pregnant women who experience IgG seroconversion, women who have positive results for CMV IgM, and women who have prenatal ultrasound abnormalities suggestive of congenital CMV infection [16, 29–31]. The investigators have concluded that positive CMV IgM and low AI in maternal blood yield positive predictive values of 9.8% [30] and 16.7%–50% [16, 28, 30], respectively, for congenital CMV infection.
In this prospective cohort study, we determined for the first time the efficacy of maternal universal screening based on CMV IgG avidity for the prediction of congenital CMV infection in low-risk women. A cutoff value of CMV IgG AI <35% yielded 22.2% sensitivity, 95.0% specificity, 2.5% positive predictive value, 99.5% negative predictive value, and 94.6% accuracy for the occurrence of congenital CMV infection. The efficacy of maternal screening based on an AI <35% was similar to that based on the CMV IgM (>1.2 index). The positive predictive values of low AI for predicting congenital CMV infection in the present study are much lower than those in previous studies [16, 27–30, 32–35]. These studies enrolled <200 women, of whom 30%–80% had primary CMV infection during their pregnancies. CMV IgG avidity measurements may not be useful as a universal test for the prediction of congenital CMV infection in low-risk pregnant women.
Nearly all symptomatic congenital CMV infections are believed to be caused by a primary infection either during or just before pregnancy [8]. In the present study, 7 of the 1287 (0.5%) pregnant women with an AI >45% coupled with either negative or equivocal results for CMV IgM (≤1.2 index), defined as nonprimary CMV infection, had newborns with congenital infection (Figure 2). Three of the 7 (42.9%) newborns from mothers with nonprimary CMV infection had symptomatic CMV infection. Mothers with nonprimary CMV infection were also at risk of giving birth to newborns with symptomatic congenital infection.
Nearly all pregnant women with primary infection seem to have both positive results for CMV IgM and low IgG AIs [28, 35]. Our previous study targeted high-risk pregnancies, and found that 45 of the 300 (15%) women with positive results for CMV IgM showed low AI (<35%) [36]. The present study enrolling low-risk pregnancies excluded referrals with positive results for CMV IgM, and showed that only 6 of 180 (3.3%) women with positive CMV IgM (>1.2 index) had low AI (<35%). Therefore, the number of pregnant women who truly experienced primary infection during pregnancy in the present study might be much fewer than that in previous studies.
Prophylactic education for women who are not immune to CMV during pregnancy reduces the incidence of IgG seroconversion [37]. In the present study, although prophylactic education was provided to pregnant women with negative results for CMV IgG, 1.1% of these women experienced IgG seroconversion during pregnancy. The incidence of CMV IgG seroconversion was similar to that (1%–2%) reported in previous studies [38].
Recently, we reported that the presence of CMV-DNA in uterine cervical secretion was predictive of the occurrence of congenital CMV infection in high-risk pregnant women who were positive for CMV IgM [36]. In this study, 3 of the 183 pregnant women with an AI ≤45% had positive PCR results for CMV-DNA in uterine cervical secretion, and 1 of these 3 (33.3%) had congenital CMV infection. The maternal screening of congenital CMV infection based on PCR assays for CMV-DNA in the uterine cervical secretion may be inefficient in the low-risk population.
The present study demonstrates, for the first time, that the efficacy of maternal universal screening based on CMV IgG avidity for congenital CMV infection was similar to that based on CMV IgM. However, newborns with congenital infection from mothers with nonprimary infection and those from mothers who had negative results but experienced CMV IgG seroconversion during pregnancy were overlooked. The universal screening of CMV IgG avidity in low-risk women is inefficient for the prediction of congenital CMV infection, as the majority of newborns with congenital infection cannot be anticipated by measuring a high AI, which is not considered a source of reassurance. Universal screening of PCR tests for CMV-DNA in newborn urine can identify all newborns with congenital CMV infection. If universal CMV screening of newborn urine is not possible, PCR tests for CMV-DNA in the urine should be performed for newborns delivered by women who carry a high risk of congenital CMV infection. A high risk of congenital CMV infection is indicated by CMV IgG seroconversion, an AI <35%, positive CMV IgM, and/or ultrasound abnormality. However, it is still difficult to detect newborns with congenital infection born to mother with nonprimary infection.
There are some potential limitations associated with the present study. Kobe University Hospital has a maternofetal center where many pregnant women with a variety of complications are referred from clinics and hospitals. Therefore, the gestational age at the first CMV IgG measurement varied widely. This might have partially influenced the results of the present study.
Notes
Acknowledgments. We are grateful for participation of the subjects and care provided by staff at Kobe University Hospital. We thank the clinical and laboratory personnel who supported this study at Kobe University Hospital and Aisenkai Nichinan Hospital.
Financial support. This work was supported by the Ministry of Health, Labour and Welfare of Japan (grant numbers H23-Jisedai-Ippan-001, AM55708030).
Potential conflicts of interest. All authors: No potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the Editors consider relevant to the content of the manuscript have been disclosed.
References
Author notes
K. T. and S. T. contributed equally to this work.
