Abstract
BACKGROUND: This study evaluated the inter-method agreement of an in-house qualitative CMV real-time PCR assay for the detection of cytomegalovirus (CMV) DNA in various non-plasma clinical specimen types, in comparison with a commercially available comparator assay. METHODS: In this prospective comparative study, 186 clinical specimens-including bronchoalveolar lavage fluid (BALF), stool, urine, colonoscopic biopsy, amniotic fluid, and intraocular fluid-were analyzed. A total of 166 samples with valid results from both test systems were included in the inter-method comparison. CMV DNA was detected using the in-house qualitative PCR assay in parallel with the comparator assay (artus(®) CMV QS-RGQ kit). Agreement was assessed using positive percent agreement (PPA), negative percent agreement (NPA), overall percent agreement (OPA), and Cohen's kappa coefficient (κ), in accordance with CLSI EP12-A2 recommendations. RESULTS: Substantial overall inter-method agreement was observed when all specimens were evaluated collectively (κ = 0.66). Agreement metrics were highest in stool, urine, and invasive specimens, whereas BALF samples demonstrated comparatively lower agreement, reflecting potential matrix-related analytical variability. CONCLUSION: The laboratory-developed qualitative CMV PCR assay demonstrated substantial inter-method agreement with the comparator assay across multiple non-plasma specimen types. The findings highlight specimen-specific variability in qualitative CMV DNA detection and represent analytical concordance between two molecular assays rather than definitive clinical diagnostic accuracy or viral load quantification.