Implementation of pooled saliva tests for universal screening of cCMV infection

Congenital cytomegalovirus (cCMV) infection is the most prevalent intrauterine infection globally, affecting 6.4–7 per 1,000 live births. Approximately 20% of affected infants develop neurodevelopmental disabilities or sensorineural hearing loss (SNHL), often with asymptomatic onset. Early diagnosis via universal newborn screening is crucial, as early antiviral treatment with valganciclovir improves outcomes. Current targeted screening strategies, focusing on high-risk infants, miss the majority of asymptomatic cases. While PCR of dried blood spots (DBS) is convenient, its sensitivity is variable, and even recent improvements leave a substantial portion of cases undetected. Saliva PCR, though highly sensitive, has high false-positive rates requiring urine confirmation, hindering large-scale universal screening due to cost and complexity. Sample pooling, a cost-effective approach, offers a potential solution. While a few small studies have demonstrated the feasibility of pooled saliva testing, large-scale implementation and comprehensive evaluation are lacking. This study aims to validate and implement large-scale pooled saliva PCR testing for universal cCMV screening, addressing sensitivity, efficiency, and feasibility.

The optimal screening strategy for cCMV remains uncertain. Targeted screening of high-risk newborns is common, focusing on those with failed hearing screens, maternal infection history, or clinical suspicion. However, this approach misses most asymptomatic infants at risk of late-onset sequelae. Universal screening is increasingly advocated for early identification of all infected infants. The accuracy and feasibility of the screening test are key considerations. DBS PCR, while using existing infrastructure, shows variable sensitivity. Saliva PCR is highly sensitive but yields high false-positive rates, mainly due to CMV DNA contamination from breast milk, requiring urine confirmation, thus limiting its suitability for universal screening. Sample pooling offers a promising solution, increasing throughput while saving resources. The low prevalence of cCMV makes it suitable for pooled screening due to the expected efficiency gain. Proof-of-concept studies have indicated the sensitivity of pooling but involved only small cohorts. A larger study from 2023 reported feasibility but lacked detailed analysis. Previous experience with large-scale sample pooling for SARS-CoV-2 detection informed this study’s methodology.

This prospective, population-based cohort study was conducted at two Hadassah Medical Center hospitals from April 2022 to April 2023. Following validation of the pooled saliva RT-PCR testing method, universal screening was implemented. Informed consent was obtained from parents. Saliva samples were collected via buccal swabs and placed in universal transport medium. Urine samples were also collected for confirmatory testing. Samples were tested using a Tecan robot for pooling, DNA extraction, and RT-PCR. Dorfman pooling was used; if a pool tested negative, all samples within the pool were declared negative. Positive pools were opened and all samples re-tested individually. A confirmatory urine test was conducted for saliva-positive samples. Sensitivity was assessed by comparing pool Ct values with individual sample Ct values. Negative predictive value (NPV) was assessed by randomly retesting a subset of negative pools. Empirical efficiency was calculated as the number of samples tested per PCR reaction. The study also retrospectively analyzed data from 2014–2021 to predict pooling accuracy. The study used primers and probes derived from CMV glycoprotein B (gB) and immediate early 1 (IE1) genes, along with a human ERV3 housekeeping gene. The assay was linear over a 6-log range with a sensitivity of 50–100 copies per ml. The IT system supported the pooling process, including sample tracking, reporting, and data analysis.

During the 13-month study period, 15,805 infants (93.6% of all live births) were screened for cCMV. The empirical efficiency of pooling was six, saving 83% of saliva tests. A minor 3.05 PCR cycle loss of sensitivity was observed, consistent with theoretical predictions. cCMV was identified in 54 infants (3.4 per 1,000), with 30 (55.6%) being asymptomatic at birth and missed by targeted screening. One asymptomatic infant with intracranial involvement detected via screening received early valganciclovir treatment. Of 1990 pools, 76 (3.82%) tested positive for either IE or gB genes. Most positive pools (85.5%) contained at least one positive sample. Random retesting of 5% of negative pools showed no missed true positives, supporting the method's high NPV. A comparison of pool Ct values and individual sample Ct values in pools with a single positive sample showed a 3.05-cycle increase, consistent with the theoretical prediction.

This study demonstrates the high feasibility and significant benefits of pooled saliva testing for universal cCMV screening. The high screening rate (94%) and minimal loss of sensitivity outweigh the benefits of cost and efficiency savings. The pooled assay also reduced false positives, lowering the need for confirmatory urine tests. The observed prevalence of cCMV (3.4 per 1,000) is within the range reported in Western European countries. The high parental acceptance rate further supports the widespread adoption of this approach. Importantly, the study identified a substantial number of asymptomatic infants who would have been missed by targeted screening, highlighting the advantages of universal screening.

The implementation of pooled saliva testing significantly improved the efficiency and feasibility of universal cCMV screening. The approach demonstrated high sensitivity and acceptable negative predictive value with a substantial reduction in testing costs. More than half of the infants identified with cCMV would have been missed by the previous targeted approach. Future research should focus on cost-effectiveness analysis and the long-term outcomes of universally screened infants.

The study could not fully assess the false-negative rate as not all samples in pools were tested individually; however, random retesting of negative pools helped mitigate this limitation. The study did not address cost-effectiveness. The prevalence estimates may not represent the entire Israeli population, being limited to Jerusalem's two hospitals, although the hospitals serve a diverse population.