The association between maternal cytomegalovirus urinary excretion and congenital infection rate

Cytomegalovirus (CMV) is a DNA herpesvirus, established as the most common congenital viral infection, with birth prevalence of approximately 0.5% [1]. Congenital CMV disease is a leading cause of birth defects, developmental disabilities and sensorineural hearing loss [2].

The most important protective factor against congenital CMV infection is preexisting maternal immunity against the virus [3], as is the timing of infection during gestation: the rate of vertical transmission in women with primary CMV infection increases with advancing gestational age [4‐7] - from approximately 40% during first trimester [6] to 73.3% for third trimester maternal infection [7]. Nevertheless, sequelae in the offspring appears to be less severe the later in gestation transmission occurs [8‐10]. Other than those, several additional risk factors, associated with vertical transmission to the fetus and severity of congenital CMV disease, have been described for both primary and non-primary maternal infections, such as high-risk occupation for viral exposure, and abnormal fetal MRI [11, 12].

Diagnosis of clinically suspected maternal primary CMV infection is based on seroconversion of negative to positive CMV IgG in maternal serum, alongside with determination of IgG avidity, which is frequently necessary in order to establish the estimated timing of the infection [13, 14]. In order to definitively determine fetal infection, the preferred prenatal analysis is PCR for CMV DNA via amniocentesis [15‐17]. In addition, sonographic monitoring for fetal anomalies that may indicate an affected fetus should be conducted, even though these anomalies are not universally diagnostic nor pathognomonic for CMV disease. In the presence of a positive PCR for CMV DNA, these finding are considered indicative for CMV congenital infection. Despite the abovementioned methods, prediction of CMV congenital infection and disease in an infected mother is limited. For that reason, we sought to investigate the role of maternal urinary CMV excretion in these entities.

The medical literature is sparse as to the association between the presence of CMV DNA in maternal urine and the rate of CMV congenital infection and disease. Delforge et al. [18] investigated a similar question and found a marginal statistically significant correlation between the two, however their study addressed only the association between CMV viruria and congenital infection rate, and did not address other implications of maternal CMV urine excretion. Therefore, the aim of our study was to establish whether maternal CMV urinary excretion is associated with congenital CMV infection and disease.

Overall, during the study period, 1632 women visited the CMV clinic. One hundred and eighty-two (11.15%) women were diagnosed with primary CMV infection. Of them, 56 were excluded: 22 did not have maternal urinary CMV analysis results, 18 had periconceptional CMV infection, 7 participated in another clinical trial, 8 had no neonatal CMV urinary analysis or amniocentesis results and one had a multiple gestation. Eventually, 126 women met the inclusion criteria and were divided into two study groups according to CMV urinary excretion: 77 (61.11%) women with positive urinary excretion and 49 (38.89%) with negative urinary excretion. Demographic and baseline maternal characteristics are presented in Table 1. Women with positive urinary CMV excretion were slightly, although significantly, older than the non-excreting group (31.2 vs. 29.7 years, p = 0.043), with no other significant differences in gravidity, parity and maternal chronic diseases.

Among infants diagnosed with in utero CMV infection (overall 36 cases), 23 were diagnosed based on positive CMV PCR, and 24 had CMV urinary excretion detected after birth in the newborn infant. 11 cases had both CMV PCR and CMV urinary excretion. We found that the congenital CMV infection rate was higher in the positive urinary excretion group (33.8% vs. 20.4% In the non-excretion group, p = 0.209), however the congenital CMV disease rate was lower in this group (18.2% vs. 22.4%, p = 0.648). These findings were not statistically significant.

Maternal CMV serology patterns are presented in Table 2. We found a significant difference in CMV Avidity between the two groups - the positive urinary CMV excretion had higher rates of low avidity compared to the non-excretion group (54.6% vs. 36.7%, p = 0.007). Non-significant trends demonstrated similar IgG (81.8% vs. 87.8% in the non-excretion group, p = 0.458) and IgM positive VIDAS (an automated enzyme-linked fluorescent immunoassay, for the detection CMV IgM antibodies) (78% vs. 85.8% in the non-excretion group, p = 0.399) rates, as well as similar gestational age at CMV infection diagnosis (17.4 vs. 16.4 weeks in the positive excretion group, p = 0.464).

We also compared separately sonographic and MRI findings that are typical for congenital CMV infection between the two study groups (Table 3). Positive MRI findings rates (8.1% vs. 5.2% in the positive excretion group, p = 0.391), and sonographic typical findings (10.2% vs. 7.8% in the positive excretion group, p = 0.746) were comparable between the groups.

An additional analysis was made to examine a possible difference in CMV excretion patterns and congenital CMV infection rate between symptomatic and asymptomatic women. Symptomatic women had a non-significant trend of higher rates of both maternal urinary excretion (70% vs. 60.5%, p = 0.388) and congenital infection rates (40.7% vs. 31.2%, p = 0.400). We investigated different obstetric parameters and found that both groups had similar rates of termination of pregnancy, gestational age at delivery and neonatal birth weight.

In the current study we aimed to investigate the association between maternal urinary CMV excretion and the rate of congenital CMV infection and disease, among women with primary CMV infection. Our findings reveal no significant association between maternal urinary CMV excretion and fetal congenital infection or disease, thereupon weakening the assumption that it might have a possible role as a predictive factor for congenital infection. Nevertheless, our study raises some interesting issues which are discussed below, and warrant further evaluations in large scale studies. CMV is shed in the urine and other bodily fluids for months after seroconversion [28], or even years in cases of infants with a diagnosis of congenital infection. The rate of CMV urinary excretion differs according to the week of infection and the week of urine analysis [29], and therefore varies between different studies. Some studies reported a rate of 3–5% [30, 31] while others reported CMV viruria as high as 13% [29, 32]. In comparison, our study demonstrated that the rate of positive CMV excreting women was significantly higher (61.11%), possibly due to our inclusion of women with exclusively primary infection. We should also be mindful to the fact that maternal CMV urinary shedding increases with gestation [28], and diagnosis of CMV infection occurred a week later among the women with positive urinary excretion (17.4 weeks of gestation vs. 16.4 in women without CMV urinary excretion). Regardless, it is possible that our results imply that CMV urinary shedding is more prevalent in women with primary infection than in those with re-infection or re-activation, however the latter were excluded from our study.

Efforts are made in order to better predict congenital CMV infection and disease in infants born to women who were infected with the virus, especially during the first trimester of pregnancy. The role of CMV urinary excretion is not well studied as a possible predictor for these outcomes and is mainly considered as non-related to intrauterine vertical transmission of the virus. Lazzarotto et al. [33], in a comprehensive review of new advances in congenital CMV infection diagnosis, described the secondary role of virological testing of maternal secretions and reported low positive prediction rates of 57.1% for congenital CMV infection and disease in infants of women who excreted CMV in the urine in the first two trimesters of pregnancy. However, in this study there was no stratification according to primary and non-primary infection.

Our results show that women with CMV viruria had a significantly higher rate of low avidity in comparison to women without presence of CMV in their urine. These results raise the possibility that urinary CMV excretion is associated with a more recent infection. Our results point to the possibility that infants of women who shed CMV in their urine were probably infected more recently and might have a higher risk for congenital CMV infection than women who do not excrete the virus in their urine.

After diagnosis of a congenital CMV infection is made, further evaluation is required in order to predict the likelihood and severity of congenital disease. This includes imaging studies, mainly sonographic but recently also with fetal MRI [34, 35], which all aim to better predict possible neurodevelopmental delays by identifying typical findings [21, 23‐25, 36].. As part of our intent to evaluate the association between maternal urinary CMV excretion and congenital disease, we investigated possible differences in imaging findings between the two study groups, under the assumption that if urinary CMV excretion increases the risk for fetal infection there might also be more sonographic and MRI abnormalities in these fetuses. Our results show no significant differences between the groups, neither when comparing the sonographic and MRI findings nor when comparing neonatal sensorineural hearing loss, therefore denoting that we could not establish an association.

Less than 5% of women infected with CMV are symptomatic [37]; however, the symptoms - such as malaise, fever and fatigue - are highly non-specific for CMV infection. Having said that, we should bear in mind that it is possible that symptoms that were reported by the women in our study were actually non-CMV related.

The strength of our study lies in the specific selection of a maternal population with primary CMV infection, and the selection of a scarcely investigated association in current literature. Additionally, the study was conducted in a single large tertiary medical center, with a high quality dedicated CMV clinic that was the key coordinator of follow up and evaluation of the patients who participated in the study. The main limitation of the study is the relatively small sample size. Due to our exclusion criteria and the sometimes-hard diagnosis of accurate timing of infection, we had to exclude a relatively large number of women. Importantly, urinary CMV shedding may be intermittent and/or prolonged [38] - therefore causing bias.

In conclusion, among women with primary CMV infection in pregnancy, we could not find an association of either urinary excretion of CMV and symptomatic maternal CMV infection, with a higher risk for congenital CMV infection and disease. However, our findings do reveal that urinary CMV excretion is associated with a more recent infection. Further studies with larger study groups are needed in order to establish these conclusions.