Lessons derived from a 3-year congenital cytomegalovirus screening programme in Israel: a prospective population-based cohort study

Background: Congenital cytomegalovirus (cCMV) is a leading cause of paediatric neurological and hearing deficits, yet there is currently no uniform public health strategy for cCMV screening. We investigated key outcomes of a 3-year universal cCMV screening programme using our newly developed saliva-pooling setup. The study objectives were to: assess the performance and feasibility of the pooled-saliva testing over time; determine the true burden of cCMV and the fraction of infants with cCMV missed by expanded-targeted screening; and define the attribution of cCMV and cCMV-related sequelae to primary or non-primary maternal infection. Methods: A prospective study was conducted in two hospitals in Jerusalem, Israel (from April 1, 2022, to March 31, 2025). All newborns whose parents provided written informed consent were screened for cCMV in pooled-saliva real-time PCR (rtPCR) testing as part of a routine newborn screening policy. Infants with positive saliva tests had confirmatory urine rtPCR tests. Pooling efficiency (number of samples tested per single rtPCR) and loss of sensitivity (pool cycle threshold [Ct] vs individual positive sample Ct) were calculated. Detection by universal screening was compared with the expanded-targeted screening strategy, focusing on infants with failed hearing screen, clinically suspected cCMV, or a history of maternal infection. Infants with cCMV were evaluated at birth and at 1 year. Maternal CMV infection type was defined by prenatal serology. Findings: Overall, 48 556 infants (94·7% of all live newborns) were screened for cCMV with the use of the pooled approach. cCMV was identified in 176 newborns, with a birth prevalence of 3·6 per 1000 (95% CI 3·1–4·2). The pooling efficiency was 5·82 (95% CI 5·69–5·95), with 3·7 Ct loss of sensitivity. 100 (57%) of 176 infants with cCMV identified by universal screening would have been missed by expanded-targeted screening. Of these, eight (8%) were classified as moderately to severely symptomatic and three (3%) as asymptomatic with sensorineural hearing loss, and 11 (11%) received valganciclovir. 84 (53%) of 158 cCMV cases with maternal infection type available were born to mothers with non-primary infection and 74 (47%) to mothers with primary infection; these infants exhibited similar rates of moderate-to-severe symptoms, sensorineural hearing loss, and 1-year hearing or developmental sequelae. Interpretation: The study demonstrated the benefits and feasibility of pooled-saliva testing, which is a sensitive approach to screening for cCMV that could enable the implementation of universal cCMV screening in diverse settings; however, further cost–benefit analyses are needed. Beyond the clinical implications of universal screening for cCMV, including enabling early diagnosis and treatment, data derived from universal screening could serve to inform public health guidelines. Funding: Israel Science Foundation and Moderna Therapeutics.