Analytical and Clinical Performance of a Real-time Screening PCR Assay Identifying Congenital CMV Infection

Background Congenital CMV infection (cCMV) is the most common identifiable cause of mental retardation in the United States but requires early diagnosis to define the infection and to institute effective antiviral therapy. Traditional identification strategies including hearing screens and physical exams likely miss many patients with cCMV. We therefore developed and evaluated the performance of a PCR assay optimized for low cost, specimen collection at time of dried blood spot collection, and detection thresholds below the salivary CMV concentrations known to occur in cCMV patients. Methods We utilized a real-time CMV PCR assay (SimplexaTM CMV)(DiaSorin, Cypress CA) amplifying the UL83 gene and the 3M Integrated Cycler. Saliva was collected from volunteers (Copan swab), and spiked with known concentrations of CMV culture supernatant quantified by COBAS AmpliprepTM. (Roche Diagnostics). Additionally, saliva was collected by copan swab from all births within a single multi-hospital system from 3/21/16 – 5/4/17. Newborns who were initial screen PCR positive were subsequently evaluated by urine CMV PCR by an outside laboratory for confirmation of cCMV. Results Analytical threshold of detection was well below 4 log copies/ML, with 100% of samples testing positive at 3.5 log copies/ML (Fig 1). 6127 newborn saliva samples were evaluated and 61 were PCR positive (£40 CT). 47 of these tests were confirmed by urine PCR (Fig 2) (PPV 0.9792, NPV 0.9988, Sens 0.8704, Spec 0.9998). Screen positive tests which were not confirmed by urine PCR had CT values £36. Adjusting the definition of a positive to CT £36 further improved the performance (PPV >0.9999, NPV 0.9997, Sens 0.9592, Spec >0.9999). Conclusion We demonstrate good performance of a congenital CMV methodology thus facilitating an effective universal newborn screening program Disclosures J. P. Devincenzo, AstraZeneca/MedImmune: Investigator, Research support