Taking a step beyond serology: progress in the search for a biomarker predicting the risk of maternal-fetal transmission of cytomegalovirus (CMV)

Congenital cytomegalovirus infection (cCMV) is the most important infectious cause of newborn disability, including sensorineural hearing loss (SNHL), in the United States and Europe. Pregnancies affected by cCMV are at an increased risk of miscarriage and intrauterine fetal demise.¹ CMV infections acquired during pregnancy can be either primary or recurrent, and typically do not produce symptoms in the pregnant patient. Although controversial,² most evidence supports the view that primary maternal infections are more likely to lead to transmission to the fetus, and in particular are more likely to produce disease and disability in the newborn. Because of their insidious and minimally symptomatic nature, most CMV infections unfortunately escape detection in the pregnant patient. When a newborn is discovered to have cCMV, the diagnosis is typically met with surprise and frustration on the part of the post-partum parent, and a sense of disappointment is expressed over the fact that information about CMV wasn’t provided as a part of routine prenatal care. Education is provided about less common diagnoses, such as congenital toxoplasmosis, and awareness of that infection is high in pregnant patients, but maternal awareness of CMV lags far behind.

To improve care and minimize the risk of cCMV infection, there is interest in developing strategies to counsel women about these risks during their pregnancies. A component of prenatal care that has considerable appeal is to offer CMV screening during pregnancy—but, unfortunately, there is no consensus on how best to do this. Serological screening is one approach. Surveys have shown that women of child-bearing age strongly support implementation of CMV sero-screening as a component of obstetrical care.³ CMV serological screening, best performed in the first trimester of pregnancy, is based on IgG and IgM testing, followed by determination of the IgG avidity index in cases with a positive IgM and/or a documented seroconversion.⁴ The Society of Obstetricians and Gynaecologists of Canada (SOGC) notes that investigation for CMV infection should be performed in any pregnant patient with a mononucleosis-like illness or undifferentiated hepatitis. Furthermore, SOGC suggests that CMV serological screening can be offered to women living in provinces where follow-up avidity testing is readily available.⁵ Seroconversion for IgG antibodies on serial testing, and/or a positive IgM with a low antibody avidity index, strongly suggests a primary maternal CMV infection, and SOGC notes that, in concert with fetal ultrasound findings, the information gleaned from a high-risk serological profile can be provided to patients along with an option offered for amniocentesis (at least eight weeks after the estimated time of primary maternal infection) to confirm the presence of a fetal CMV infection. Effort to make the diagnosis of a primary maternal CMV infection has also taken on heightened interest in light of data that the administration of high-dose oral valacyclovir to the pregnant patient with primary infection can mitigate the risk of fetal transmission for infections acquired periconceptually or in the first trimester.⁶ The American College of Obstetrics and Gynecologists (ACOG), however, has not embraced these approaches. Thought leaders opposed to maternal sero-screening during pregnancy point out that serological methods provide, at best, indirect evidence of maternal infection status during pregnancy; that amniocentesis carries attendant risks that create hesitancy on the part of patients and practitioners alike; that the benefits suggested by studies of valacyclovir therapy, though intriguing, need to be validated in larger studies; and, importantly, that serological screening does not provide insight into the phenomenon of maternal re-infection in pregnancy, which carries the risk of both vertical transmission and subsequent disability in the infant, in spite of serologic documentation of pre-conception immunity. Hence, it is argued by some experts that maternal serological testing does not provide adequate clarity to the question of how to counsel a pregnant patient about cCMV transmission.

Against this backdrop, there is therefore a compelling need for novel strategies—beyond serology—to gauge the risk of CMV transmission during pregnancy. In an intriguing paper in the February 2024 issue of eBioMedicine from Faas et al., the investigative team exploited the fact that non-invasive prenatal testing (NIPT) for fetal aneuploidy screening early in pregnancy is a standard-of-care test used in obstetrical care in many countries.⁷ NIPT is commonly used in prenatal care—indeed, ACOG has recommended that NIPT be offered to all pregnant patients, regardless of gestational history or risk factors.⁸ To examine for evidence of CMV infection using samples obtained in the context of NIPT for aneuploidy testing, the authors in the current paper examined data from 204,818 pregnant women for the presence of CMV-specific cell free DNA (cfDNA). For validation of positive results, diagnostic CMV-qPCR and serology were performed on a subset of 112 cfDNA CMV-positive samples, and 127 cfDNA-negative controls. CMV DNA was detected in just under 1% of all samples in the study. The samples with higher cfDNA CMV viral loads were positive with standard CMV PCR testing. In addition, serologic assessment was consistent with a recent primary infection in 32/112-positive samples (28.6%). Reassuringly, in cfDNA-negative samples, serology did not suggest recent primary infection. The authors suggest that incorporation of CMV testing into the NIPT testing algorithm could provide a useful screening strategy for evaluating the risk of a primary CMV infection in early pregnancy.

These exciting data could lead to assays that redefine the paradigm of how we evaluate for maternal CMV infection during pregnancy. Notably, the study did not require new experiments to examine for CMV DNA signal in the initial analyses of these 204,818 samples, since the sequence information for CMV could be obtained using bioinformatic approaches, culled directly from the NIPT-derived sequence data. For the validation of the initial informatics-driven analyses, DNA was then extracted from leftover plasma of 239 samples and amplified using a commercially validated quantitative PCR (qPCR) assay. All cfDNA CMV-negative samples were negative by qPCR. Of the cfDNA CMV-positive cases, qPCR confirmed the presence of CMV DNA in many instances, although there was not complete concordance between the assays. In fact, 67.8% (76/112) cfDNA-positive samples tested negative with the qPCR assay; furthermore, six cfDNA CMV-positive cases were believed to be falsely positive due to a sequence analysis artifact. On the other hand, of the 76 cfDNA CMV-positive samples with a negative qPCR, six samples (7.9%) demonstrated a serological profile compatible with a primary infection. Thus, more work needs to be done to optimize the cfDNA approach—in particular, to examine the utility of continuing to obtain serologic assays that demonstrate recent primary infection in the absence of a cfDNA signal. Still, a real benefit of the technique is that, once it is optimized, serology should hopefully no longer be the key laboratory test that informs and directs prenatal cCMV counselling. This is because, as noted above,² if a pregnant patient is CMV IgG positive, it does not eliminate the risk of cCMV for the fetus. Therefore, a positive cfDNA for CMV might be a valuable result in future clinical practice precisely because it skirts the issue of interpreting (and potentially falsely minimizing) cCMV risk in an IgG-seropositive pregnant patient who has pre-conception immunity, but still has risk facors for infection. In such a setting, a high-level positive cfDNA result for CMV DNA in a sero-immune patient may reflect a re-infection, could portend risk for the fetus, and may identify a patient who could benefit from more aggressive intervention, irrespective of her serological profile. Sensitivity of the cfDNA assay will therefore need to be further clarified. The authors noted that the sensitivity of the assay was higher in samples with higher cfDNA viral load; a cut-off threshold of 0.3 fragments per million reads (FPMs) had a 100% correlation with qPCR results. This requires further validation. Another issue to consider is that the finding of CMV sequences by cfDNA analysis does not necessarily mean that “infectious virus” is present. Studies demonstrating that this signal correlates with the presence of CMV mRNA transcripts could provide support for the hypothesis that a positive cfDNA signal could correlate with actively replicating virus and, hence, could portend an increased risk of cCMV transmission. Thus, more work needs to be done to further define the sensitivity and predictive value of the assay (including cCMV surveillance screening studies in the corresponding newborn population), to pursue clinically relevant molecular correlates (such as transcriptomics), and to further validate the methodology, before this assay is incorporated into clinical practice.

These challenges notwithstanding, the work reported in this manuscript could be a real “game-changer” in how maternal CMV screening and counselling is implemented in the clinic. It is easy to envision a future where, at the time of NIPT, nucleic acid-based screening—not only for CMV, but for other pathogens important in obstetrical care, such as T. pallidum—could become a standard-of-care. It is highly appealing to incorporate CMV cfDNA testing into the NIPT test. A positive cfDNA test for CMV DNA could identify pregnancies that require meticulous additional evaluation to ascertain the level of risk for cCMV infection. It remains to be demonstrated that a positive cfDNA screen in pregnancy correlates with an increased risk of cCMV transmission, and such studies are required. Fortunately, future studies that correlate cfDNA testing for CMV DNA with cCMV transmission are now feasible in US states (Minnesota) and Canadian provinces (Ontario, Saskatchewan) that perform universal cCMV screening on all newborns, and studies of maternal cfDNA screening performed in parallel with newborn cCMV screening should be conducted.⁹ Testing for cfDNA can also provide insights into the impact and potential effectiveness of pre-conception maternal vaccination aimed at prevention of cCMV¹⁰–once such a vaccine is licensed.

Assuming that prenatal infectious diseases diagnostics continue to evolve toward an “omics”-based approach, clinical care will have taken a big step beyond serology. Repurposing NIPT testing to leverage the advantages of bioinformatics will aid in assessing risk factors for not only active maternal CMV infection, but other pathogens of perinatal importance. Meanwhile, every effort should be expended on encouraging more pregnant patients to obtain NIPT testing. Fewer than 25% of women in most European countries currently use NIPT, although there is considerable variability from one country to another; similarly, in most US and Australian states, only 25%–50% of the pregnant patients have NIPT performed.⁸ These percentages need to increase in order to make cfDNA CMV DNA testing a useful component of clinical practice. It will also be important, even as these exciting advances in diagnostics move forward, to consider health equity issues in prenatal CMV testing. Individuals who live in regions who have lower socioeconomic status, are born in countries outside of the US and Europe, or who are from communities of colour are more likely to experience cCMV infection, making this a disease of considerable health disparity. cCMV is a global health issue, and even as advances in bioinformatics lead to improved prenatal care in high-income countries, strategies are needed for implementation science that will ensure equitable world-wide access to these novel technologies.

Contributors

Literature search: M.R.S.; Data collection: M.R.S.; Data interpretation: M.R.S.; Writing: M.R.S. The author prepared and approved the manuscript.

Declaration of interests

M.R.S. declares consulting fees from GSK Vaccines. M.R.S. declares grant support, but no personal honoraria, from Moderna vaccines.