Abstract
BACKGROUND:
Both neighborhood disadvantage and close contact with children have been associated with seroprevalence of cytomegalovirus in pregnancy. However, it is unknown which individual factors influence whether seropositive women are likely to have ongoing viral shedding.
OBJECTIVE:
This study aimed to define the frequency of and risk factors for ongoing maternal cytomegalovirus shedding across gestation among seropositive pregnant women.
STUDY DESIGN:
This was a prospective cohort study of women who were cytomegalovirus seropositive at a single tertiary care hospital between September 1, 2018, and September 1, 2020. The participants were eligible if positive for cytomegalovirus immunoglobulin G during the first trimester of pregnancy. Urine samples were planned to be collected from each trimester. DNA was isolated in urine samples to detect and quantitate cytomegalovirus immediate-early 1 gene. Participants were classified as “ever shedder” if cytomegalovirus was detected in any urine sample and “never shedder” if cytomegalovirus was never detected. Patient demographics and characteristics were compared between groups. Stochastic search variable selection (with a posterior probability of inclusion of >0.5) was used to identify predictors of cytomegalovirus shedding at any time point. Forward selection modeling was used as a sensitivity check for independent risks.
RESULTS:
A total of 240 participants who were cytomegalovirus immunoglobulin G seropositive were enrolled, with 567 urine samples analyzed across gestation. Fifty-eight participants (24.2%) were “never shedders”, and 182 participants (75.8%) were “ever shedders.” The characteristics and demographics were similar between cohorts. With stochastic search variable selection, nulliparity was the only variable selected (odds ratio, 1.82; 95% credible interval, 1.00–4.09; Bayes factor, 2.22). Furthermore, nulliparity was selected with standard logistic regression, with an odds ratio and 95% confidence interval of 1.89 (1.00–3.58). Sociodemographic characteristics, such as age, race, education level, occupation, children at home, children in daycare, housing type, insurance type, income, and concurrent infections, were not associated with shedding. The only positive neonatal sample (0.42%) was detected from a participant who had cytomegalovirus detected in all 3 time points.
CONCLUSION:
Approximately 75% of women who were positive for cytomegalovirus immunoglobulin G shed virus at some point during gestation. Nulliparity was the only variable selected that was associated with shedding.
Keywords: antiviral agents, congenital cytomegalovirus, congenital viral infection, cytomegalovirus, cytomegalovirus immunoglobulin G, fetal infection, secondary maternal cytomegalovirus infection, seroconversion
Introduction
Traditionally, it has been thought that primary maternal cytomegalovirus (CMV) infection is associated with congenital disease, but emerging data in highly seropositive populations demonstrate that up to 75% of all congenital infections result from maternal reinfection or reactivation.1 Women with a previous CMV infection may experience either reinfection with different strains or reactivation of an endogenous latent strain. However, the incidence of reinfection or reactivation is unknown, and the risk factors for ongoing shedding are poorly understood. Although living in urban low-income neighborhoods and occupational or household exposure to children have been associated with CMV positivity and primary infection, the impact of these and other demographic risk factors on ongoing viral shedding is unknown.2,3 The American College of Obstetricians and Gynecologists has called for further research to identify seropositive women at risk.4
The objective of this study was 2-fold: (1) to estimate the frequency of CMV viruria among seropositive women and (2) to identify the clinical factors associated with ongoing shedding. After resolution of the primary infection, CMV becomes quiescent and remains latent in myeloid cells. With changes to the host immunity, CMV can reactivate and cause clinical disease, with viral particles spreading hematogenously to several organ systems. In in vitro models, CMV readily infects urogenital cells. Therefore, it follows that viruria may be a marker of viral shedding, although it is unclear whether the detection in the urine is because of reinfection or reactivation.
Materials and Methods
This was a prospective cohort study of women who were CMV seropositive at a single tertiary care hospital between September 1, 2018, and September 1, 2020. Pregnant women in the first trimester of pregnancy were eligible for serologic screening for a previous CMV infection. The participants were offered inclusion if positive for CMV immunoglobulin G (IgG) during the first trimester of pregnancy and excluded if concurrently HIV positive. CMV IgG antibodies (without avidity) were measured by the Department of Microbiology at Duke University with the use of a commercially available chemiluminescent microparticle immunoassay (enzyme immunoassay, Viracor). Immunoglobulin M (IgM) was not used for the inclusion criteria because of the large false-positive rate for commercially available IgM assays and because IgM may persist for many months after acute infection.5 Urine samples were planned to be collected each trimester, and viral load (VL) was analyzed from each sample. Urine samples were first concentrated by ultracentrifugation to create 500 μL aliquots for nucleic acid extraction (HighPure Viral Nucleic Acid Kit; Roche, Basel, Switzerland). Primers specific for CMV immediate-early 1 (IE1) genes were used for viral gene amplification and detection using SYBR Green. Plasmids expressing CMV IE1 genes were used as a reference standard for absolute quantification quantitative polymerase chain reaction (qPCR).6 After amplification, the average cycle threshold (CT) was calculated for each plasmid dilution and plotted as a scatterplot. A linear regression analysis of the plot was used as a standard curve, and sample positivity was determined according to CT values: CT values >35.0 were considered negative, whereas CT values <35.0 were considered positive. A detailed qPCR reaction protocol may be found as an Online Addendum (Detailed Quantitative Polymerase Chain Reaction Protocol).
Participants were classified as “ever shedders” if any urine sample had a detectable CMV VL from at least 1 time point or “never shedders” if the participant never had any detectable VL. Social demographics were self-reported at enrollment using a Research Electronic Data Capture tool hosted at Duke University, and clinical data were abstracted by trained research coordinators. Baseline demographics were compared between ever shedders and never shedders. Continuous variables were summarized as median (Q1 [25th percentile] to Q3 [75th percentile]). Categorical variables were presented as frequency(percentage). All baseline variables were initially considered for predictors of shedding using the stochastic search variable selection (SSVS) algorithm. The log-odds ratios were assumed to be independently distributed as βk = { 0 w.p. 0.50~N(0, 1) w.p. 0.50, so each predictor is equally likely to be included or excluded in the model and a noninformative prior (σ2 = 1) was used.7 Although all variables collected were considered, the levels of categorical variables were grouped if there were low cell counts, and the square footage of home was excluded because of a low reporting (35.0%). We performed a sensitivity check using forward selection for the baseline variables that were considered using the SSVS. Analyses were performed using Statistical Analysis System (version 9.4; SAS Institute Inc, Cary, NC) and WinBUGS (version 1.4.3; MRC Biostatistics Unit, Cambridge, United Kingdom).
This study was approved by the Duke University Institutional Review Board. Oral and written informed consents were obtained from all participants before serologic screening. The authors take the responsibility for the accuracy and completeness of the data and fidelity of the study.
Results
Overall, 240 participants who were CMV seropositive were included in the study, with 567 urine samples collected and analyzed from September 1, 2018, to September 1, 2020. Because of COVID-19 research restrictions that precluded planned collection from 3 time points for each participant, the participants were included for analysis if at least 1 sample was obtained across gestation. Overall, 128 women (53.3%) had 3 samples collected, 71 women (29.6%) had 2 samples collected, and 41 women (17.1%) had 1 sample collected. There was an even distribution of VL from samples collected across the first, second, and third trimesters of pregnancy (Figure 1).
FIGURE 1.
Viral load distribution per trimester
Of the participants, 58 (24.2%) were never shedders, and 182 (75.8%) were ever shedders. The patient demographics and characteristics at enrollment are represented in the Table. The median(Q1–Q3) age for all participants was 32.0 (28.0–35.0), and approximately half of the participants were White. Among all participants, the highest level of education completed skewed toward advanced degrees: overall, 72 participants (30.0%) completed college, and 87 participants (36.3%) completed postgraduate studies. There was a wider range among total household income but still skewed toward upper categories: 18 (31.0%) never shedders and 42 (23.1%) ever shedders reported >$150,000 income. The largest occupation was healthcare, as 19 (32.8%) never shedders and 61 (33.5%) ever shedders were employed in healthcare. Furthermore, occupations with high exposure to children, such as babysitters, nannies, daycare workers, and teachers, were similar between groups. The number of participants with children living at home was similar: 39 (67.2%) never shedders and 109 (59.9%) ever shedders lived with children, with a median (Q1–Q3) of 1 (1–2) at home for both groups. Of note, 19 (48.7%) never shedders and 52 (47.7%) ever shedders had children in daycare. Although there was a high rate of missing data for square footage of homes (35.0%), the median (Q1–Q3) was similar between cohorts: 1800.0 (1380–2500) for never shedders vs 1922.0 (1500–2500) for ever shedders. All variables presented in the Table were initially considered for selection, and grouping of categorical variables may be found in the Online Addendum.
TABLE.
Patient demographics and characteristics at enrollment
| Characteristic | Never shedder (n=58) | Ever shedder (n=182) | Total (N=240) |
|---|---|---|---|
| Age (y) | 32 (30−36) | 32 (27−35) | 32 (28−35) |
| Race | |||
| Asian | 5 (8.6) | 16 (8.8) | 21 (8.8) |
| Black or African American | 14 (24.1) | 45 (24.7) | 59 (24.6) |
| White | 30 (51.7) | 102 (56.0) | 132 (55.0) |
| Unknown or not reported | 2 (3.4) | 12 (6.6) | 14 (5.8) |
| >1 race | 7 (12.1) | 7 (3.8) | 14 (5.8) |
| Ethnicity | |||
| Hispanic or Latino | 9 (15.5) | 21 (11.5) | 30 (12.5) |
| Non-Hispanic or non-Latino | 48 (82.8) | 158 (86.8) | 206 (85.8) |
| Unknown or not reported | 1 (1.7) | 3 (1.6) | 4 (1.7) |
| BMI (kg/m2) | 26.6 (23.0−32.0) | 27.2 (22.8−32.9) | 27.1 (22.9−32.7) |
| Highest level of education completed | |||
| Primary or middle school | 1 (1.7) | 0 (0.0) | 1 (0.4) |
| Some high school | 4 (6.9) | 9 (4.9) | 13 (5.4) |
| High school graduate or GED | 7 (12.1) | 22 (12.1) | 29 (12.1) |
| Some college | 9 (15.5) | 29 (15.9) | 38 (15.8) |
| College | 17 (29.3) | 55 (30.2) | 72 (30.0) |
| Postgraduate | 20 (34.5) | 67 (36.8) | 87 (36.3) |
| Occupation | |||
| Babysitter or nanny | 0 (0.0) | 1 (0.5) | 1 (0.4) |
| Daycare worker | 0 (0.0) | 2 (1.1) | 2 (0.8) |
| Healthcare worker | 19 (32.8) | 61 (33.5) | 80 (33.3) |
| Other | 25 (43.1) | 78 (42.9) | 103 (42.9) |
| Student (high school and college) | 0 (0.0) | 2 (1.1) | 2 (0.8) |
| Teacher | 2 (3.4) | 8 (4.4) | 10 (4.2) |
| Unknown or not reported | 12 (20.7) | 30 (16.5) | 42 (17.5) |
| Children living in home | 39 (67.2) | 109 (59.9) | 148 (61.7) |
| Number of children in home (among those with children living at home) | 1.0 (1−2) | 1.0 (1−2) | 1.0 (1−2) |
| Any children in daycare (among those with children living at home) | 19 (48.7) | 52 (47.7) | 71 (48.0) |
| Number of people living in home | 3 (2−4) | 3 (2−3) | 3 (2−4) |
| Square footage of home | 1800 (1380−2500) | 1922 (1500−2500) | 1903 (1475−2500) |
| Missing | 19 (32.8) | 65 (35.7) | 84 (35.0) |
| Square feet of home or number of people in home | 733.3 (450.0−950.0) | 700.0 (533.3−980.0) | 705.0 (500.0−955.5) |
| Missing | 19 (32.8) | 65 (35.7) | 84 (35.0) |
| Does anyone living in your household work in any of the following areas | |||
| Daycare | 2 (3.4) | 5 (2.7) | 7 (2.9) |
| School | 3 (5.2) | 17 (9.3) | 20 (8.3) |
| Healthcare | 24 (41.4) | 69 (37.9) | 93 (38.8) |
| None of the above | 31 (53.4) | 97 (53.3) | 128 (53.3) |
| Primary insurance type | |||
| Government assisted (Medicaid or Medicare) | 20 (34.5) | 44 (24.2) | 64 (26.7) |
| Private or commercial | 37 (63.8) | 132 (72.5) | 169 (70.4) |
| None or self-pay | 1 (1.7) | 6 (3.3) | 7 (2.9) |
| Total household income ($) | |||
| <10,000 | 1 (1.7) | 15 (8.2) | 16 (6.7) |
| 10,000−24,999 | 7 (12.1) | 15 (8.2) | 22 (9.2) |
| 25,000−49,999 | 10 (17.2) | 27 (14.8) | 37 (15.4) |
| 50,000−74,999 | 6 (10.3) | 19 (10.4) | 25 (10.4) |
| 75,000−99,999 | 7 (12.1) | 23 (12.6) | 30 (12.5) |
| 100,000−149,999 | 4 (6.9) | 35 (19.2) | 39 (16.3) |
| ≥150,000 | 18 (31.0) | 42 (23.1) | 60 (25.0) |
| Not reported (or refused to answer) | 5 (8.6) | 6 (3.3) | 11 (4.6) |
| Primary language spoken in home | |||
| English | 48 (82.8) | 161 (88.5) | 209 (87.1) |
| Spanish | 8 (13.8) | 10 (5.5) | 18 (7.5) |
| Other | 2 (3.4) | 11 (6.0) | 13 (5.4) |
| Cigarette use during pregnancy | 5 (8.6) | 10 (5.5) | 15 (6.3) |
| Alcohol use during pregnancy | 16 (27.6) | 57 (31.3) | 73 (30.4) |
| Street drug use during pregnancy | 4 (6.9) | 6 (3.3) | 10 (4.2) |
| Multiple sexual partners during the past 5 y | 13 (22.4) | 32 (17.6) | 45 (18.8) |
| Infections this pregnancy | |||
| Hepatitis C | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Herpes, active | 1 (1.7) | 2 (1.1) | 3 (1.3) |
| Urinary tract infection and/or pyelonephritis | 3 (5.2) | 16 (8.8) | 19 (7.9) |
| Parity | |||
| 0 | 17 (29.3) | 80 (44.0) | 97 (40.4) |
| 1 | 23 (39.7) | 67 (36.8) | 90 (37.5) |
| 2 | 11 (19.0) | 21 (11.5) | 32 (13.3) |
| 3 | 6 (10.3) | 7 (3.8) | 13 (5.4) |
| 4 | 0 (0.0) | 3 (1.6) | 3 (1.3) |
| >4 | 1 (1.7) | 4 (2.2) | 5 (2.1) |
| Household status | |||
| Own single-family home, townhouse, or condominium | 31 (53.4) | 99 (54.4) | 130 (54.2) |
| Rent | 21 (36.2) | 63 (34.6) | 84 (35.0) |
| Lives with parents or other adults | 6 (10.3) | 20 (11.0) | 26 (10.8) |
| Housing instability | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Housing type | |||
| Single-family home | 43 (74.1) | 125 (68.7) | 168 (70.0) |
| Townhouse | 5 (8.6) | 16 (8.8) | 21 (8.8) |
| Condominium | 0 (0.0) | 2 (1.1) | 2 (0.8) |
| Apartment | 10 (17.2) | 39 (21.4) | 49 (20.4) |
Data are presented as number (percentage), unless otherwise indicated. All continuous variables are presented as median (interquartile range).
BMI, body mass index; GED, General Equivalency Diploma.
From the 240 participants included in the study, there was 1 neonate (0.42%) identified with clinical criteria of CMV and CMV PCR detected in saliva. This neonate was born to a woman who had positive viruria in all 3 time points collected; thus, the neonatal positivity rate among ever shedders was 1 of 182 (0.55%).
The SSVS approach was used, and nearly all variables had posterior inclusion probabilities of approximately 0.5. In other words, none of the variables were informative in predicting shedding. The only exception was nulliparity, which was the sole predictor of shedding and had the highest probability, with a posterior inclusion probability of 0.68. The posterior distribution of the adjusted log-odds for nulliparity is shown in Figure 2. The posterior odds ratio (OR) and 95% credible interval for nulliparity was 1.82 (1.00–4.09), with a Bayes factor of 2.22. As a sensitivity check, a standard logistic regression was used, and in this model, nulliparity was also predictive (OR with 95% confidence interval [CI] was 1.89 [1.00–3.58]).
FIGURE 2.
Posterior density for nulliparity log-odds ratio
Principal findings
Here, we found that approximately 75% of patients who were CMV seropositive early in pregnancy shed CMV viral DNA during pregnancy. However, none of the known risk factors for primary infection were found to be predictive of viral shedding in this population.
Results in the context of what is known
Observational data have previously correlated both CMV seropositivity and seroconversion with socioeconomic status. For example, the prevalence of previous CMV exposure in women of childbearing age varies widely by both region and income, ranging from 40% to 83%.8 Using geospatial analysis of urban neighborhoods within Durham, North Carolina, we have previously demonstrated that clusters with elevated odds of CMV seropositivity had a higher frequency of primary CMV infection.2 In the aforementioned study, neighborhood median family income was associated inversely with the prevalence of chronic CMV infection. However, its impact on ongoing shedding, via either reinfection or reactivation, is largely unknown. Using CMV viruria as a proxy for shedding, our findings indicated a high rate of ongoing shedding—specifically 75.8%—during pregnancy among women with serologic evidence of previous CMV exposure. With up to 83% of reproductive-aged women seropositive,9 it is clear that the prevalence of ongoing viral shedding is similarly pervasive, contributing to CMV’s status as the most common congenital infection. This high rate of shedding was similar to a 3-year prospective study of the University of Alabama (UAB) of 205 women who were CMV seropositive, with most women (171/205, [83.4%]) demonstrating CMV viruria at 1 point across the study.10 Of note, the aforementioned study recruited participants who were postpartum, although it demonstrated a similarly high rate of shedding. However, in contrast to the UAB study and ours, a prospective study of 120 Brazilian women who were CMV seropositive demonstrated a drastically much lower rate (13/120 [13.3%]) of CMV viruria.11 The discrepancy in shedding outcomes may be attributed to a difference in methods: the latter study used different DNA extraction kits, with different primers used. The ultracentrifugation method used in the present study was expected to be more sensitive in CMV DNA detection.
The results from our study underscored the socioeconomic risk factors that belie primary infection and seroconversion rates. Here, using the SSVS, which essentially searches through 2p dimensional model space (where p is the number of variables within models consisting of a subset of the predictors), only nulliparity had an inclusion probability of >0.50. Nulliparity had a very wide credible interval and a small Bayes factor. Furthermore, nulliparity was selected using the traditional logistical regression, with an OR that included unity (1.89; 95% CI, 1.00–3.58). All other variables had posterior inclusion probabilities between 0.01 and 0.47.
Clinical implications
Findings from this study have supported the current Society for Maternal-Fetal Medicine’s recommendations against routine CMV serologic or virologic screening during pregnancy.5 For a screening method to be effective, an intervention must be available to alter the course of the disease and reduce its burden or transmission. This study has found that among those who would have screened seropositive, there is a high prevalence of viral shedding—without a currently available vaccine or proven therapy for CMV. Current clinical management focuses on reducing the risk of primary infection. Prospective data have shown that interventions to decrease viral exposure may decrease seroconversion risk among seronegative pregnant women.12 Clinical factors affecting CMV exposure risks, such as at-risk jobs, children at home, and children at daycare, were not predictive of shedding among those previously infected. Furthermore, CMV viruria followed a similar trend to serologies—highly prevalent, but with a low rate of congenital transmission—making it a poor predictor for risk stratification of disease transmission.
Research implications
There remained a gap in knowledge concerning how often women who were CMV seropositive shed the virus during pregnancy, the magnitude of viral shedding, and whether there are maternal, clinical, or viral characteristics associated with transmission. As previous immunity to CMV does not preclude congenital infection, establishing predictors of reinfection would aid in designing prevention or treatment trials, using maternal shedding as a surrogate endpoint. Furthermore, it remains unknown whether ongoing shedding is because of reinfection or reactivation. Ongoing research should seek to profile shed viral populations and serologic responses of pregnant women who were CMV seropositive using next-generation sequencing to differentiate reinfection vs reactivation. Lastly, although CMV viruria may be indicative of shedding, the predictive value of CMV viruria on neonatal transmission is unknown but likely to be low.
Strengths and limitations
This prospective study included a pragmatic design that would mimic a screening approach that would be performed if universal CMV screening were to be included in routine prenatal care. A wide array of demographic variables were used as proxies for CMV exposure, including occupational risk and housing structures. In addition, rigorous Bayesian statistical modeling was used, and a sensitivity analysis using standard logistic regression was performed, yielding the same results.
There were several limitations involved with this study. First, the study design intended for trimester urine samples from each participant. As COVID-19 research restrictions precluded planned collection, 71 participants (29.6%) had 2 samples collected, and 41 participants (17.1%) had 1 sample collected. Thus, the total participant positivity rate (182 [75.8%]) may be an underestimate, as participants considered never shedders who had <3 samples collected may have had viruria in a trimester that inevitably went uncollected. In addition, as the timing of urinary sample collection was not evenly distributed, we could not comment on the frequency of viruria across gestation. Furthermore, it was unclear whether the shedding was attributed either to a reinfection or a reactivation with a different strain. This distinction may be revealed by the profile of shed viral populations. Moreover, it was unclear to what degree urine shedding was a proxy for fetal exposure. Another limitation was that women were included with 1 positive IgG serology without avidity or CMV IgM testing that can be used to define a primary infection within 3 to 6 months after the occurrence; thus, the interval between primary infection and detected viruria was unknown.
Conclusions
Approximately 75% of women who were positive for CMV IgG shed CMV at some point during gestation. Only nulliparity, and no other patient demographic, was defined as an independent risk factor for shedding.
Supplementary Material
AJOG MFM at a Glance.
Why was this study conducted?
Women with previous cytomegalovirus (CMV) infection may experience reinfection or reactivation during pregnancy, potentially leading to congenital transmission. Risk factors for viral shedding are poorly understood. Understanding these risks may enhance our ability to develop and test preventive interventions.
Key findings
Approximately 75% of women who were CMV immunoglobulin G seropositive shed the virus at 1 of 3 measured time points during gestation. Sociodemographic characteristics were not predictive of shedding risk.
What does this add to what is known?
Although there are known sociodemographic and hygienic risk factors for primary CMV infection in pregnancy, we did not find these to be predictive of ongoing shedding among women who were previously infected.
Acknowledgments
B.L.H. is a scientific advisor for the Merck cytomegalovirus (CMV) program. S.P. provides individual consulting services to Moderna, Merck, Dynavax, and Pfizer on their CMV vaccine programs. MerckVaccines and Moderna have provided grants and contacts for her institutional sponsored programs.
This manuscript was funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (grant number 5R21HD096867–02; principal investigator B.L.H.).
Footnotes
The authors report no conflict of interest.
Data from this manuscript were presented at the 2021 annual meeting of the Infectious Diseases Society of America, held virtually, July 29–30, 2021.
Supplementary materials
Supplementary material associated with this article can be found in the online version at doi:10.1016/j.ajogmf.2021.100560.
Contributor Information
Luke A. Gatta, Department of Obstetrics and Gynecology, Duke University Hospital, Durham, NC; Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Duke University Hospital, Durham, NC.
Eric Rochat, Duke Human Vaccine Institute, Duke University Hospital, Durham, NC.
Jeremy M. Weber, Department of Biostatistics and Bioinformatics, Duke University Hospital, Durham, NC.
Sarah Valencia, Duke Human Vaccine Institute, Duke University Hospital, Durham, NC.
Alaattin Erkanli, Department of Biostatistics and Bioinformatics, Duke University Hospital, Durham, NC.
Sarah K. Dotters-Katz, Department of Obstetrics and Gynecology, Duke University Hospital, Durham, NC; Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Duke University Hospital, Durham, NC.
Sallie Permar, Duke Human Vaccine Institute, Duke University Hospital, Durham, NC; Department of Pediatrics, Weill Cornell Medical Center, New York, NY.
Brenna L. Hughes, Department of Obstetrics and Gynecology, Duke University Hospital, Durham, NC; Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Duke University Hospital, Durham, NC.
References
- 1.Wang C, Zhang X, Bialek S, Cannon MJ. Attribution of congenital cytomegalovirus infection to primary versus non-primary maternal infection. Clin Infect Dis 2011;52:e11–3. [DOI] [PubMed] [Google Scholar]
- 2.Lantos PM, Hoffman K, Permar SR, et al. Neighborhood disadvantage is associated with high cytomegalovirus seroprevalence in pregnancy. J Racial Ethn Health Disparities 2018;5: 782–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lantos PM, Hoffman K, Permar SR, Jackson P, Hughes BL, Swamy GK. Geographic disparities in cytomegalovirus infection during pregnancy. J Pediatric Infect Dis Soc 2017;6:e55–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Practice Bulletin No. 151: cytomegalovirus, parvovirus B19, varicella zoster, and toxoplasmosis in pregnancy. Obstet Gynecol 2015;125:1510–25. [DOI] [PubMed] [Google Scholar]
- 5.Society for Maternal-Fetal Medicine (SMFM). Hughes BL, Gyamfi-Bannerman. Diagnosis and antenatal management of congenital cytomegalovirus infection. Am J Obstet Gynecol 2016;214:B5–11. [DOI] [PubMed] [Google Scholar]
- 6.Kalpoe JS, Kroes AC, de Jong MD, et al. Validation of clinical application of cytomegalovirus plasma DNA load measurement and definition of treatment criteria by analysis of correlation to antigen detection. J Clin Microbiol 2004;42:1498–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.George EI, McCulloch RE. Variable selection via Gibbs sampling. J Am Stat Assoc 1993;88:881–9. [Google Scholar]
- 8.Stagno S, Pass RF, Cloud G, et al. Primary cytomegalovirus infection in pregnancy. Incidence, transmission to fetus, and clinical outcome. JAMA 1986;256:1904–8. [PubMed] [Google Scholar]
- 9.Johnson J, Anderson B. Screening, prevention, and treatment of congenital cytomegalovirus. Obstet Gynecol Clin North Am 2014;41:593–9. [DOI] [PubMed] [Google Scholar]
- 10.Arora N, Novak Z, Fowler KB, Boppana SB, Ross SA. Cytomegalovirus viruria and DNAemia in healthy seropositive women. J Infect Dis 2010;202:1800–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Barbosa NG, Yamamoto AY, Duarte G, et al. Cytomegalovirus shedding in seropositive pregnant women from a high-seroprevalence population: the Brazilian Cytomegalovirus Hearing and Maternal Secondary Infection Study. Clin Infect Dis 2018;67:743–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Vauloup-Fellous C, Picone O, Cordier AG, et al. Does hygiene counseling have an impact on the rate of CMV primary infection during pregnancy? Results of a 3-year prospective study in a French hospital. J Clin Virol 2009;46 (Suppl4):S49–53. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.