Comparison of Overall Sensitivity and Specificity across Different Newborn Screening Algorithms for Congenital Cytomegalovirus

Abstract

Screening newborns for congenital cytomegalovirus (cCMV) infection is critical for early detection and prompt diagnosis of related long-term consequences of infection, such as sensorineural hearing loss and neurodevelopmental delays. The objective of this study was to describe the validity of different newborn cCMV infection screening approaches and compare the expected number of cCMV cases detected across targeted and universal screening algorithms. The overall sensitivity (OSn) of targeted screening algorithms that required failure of auditory brain stem response and transient evoked otoacoustic emissions (TOAE; two-fail serial testing) or TOAE only (one-fail serial testing) before diagnostic CMV testing using saliva and urine PCR tests was 79% and 88%, respectively. The OSn for two-fail serial testing with diagnostic CMV testing using dried blood spot (DBS) was 75%. In contrast, OSn was 90% for universal screening (saliva and urine PCR tests) and 86% for universal screening with DBS testing alone. Overall, specificities were 100% across all algorithms. Universal screening using DBS testing and universal screening using saliva and urine testing can potentially detect 312 and 373 more cCMV cases per 100,000 live births, respectively, than two-fail serial testing. Overall, implementing universal cCMV newborn screening would improve cCMV detection, ultimately leading to better health outcomes.