Transitioning cytomegalovirus viral load testing from a laboratory developed test to the cobas^® CMV quantitative nucleic acid assay

Commutability between human cytomegalovirus (CMV) viral load assays (VLA) is poor, despite the development of a WHO CMV International Standard (CMV IS). We evaluated a new CMV VLA, cobas^® CMV, as compared to our current laboratory developed CMV VLA (LDT), for clinical use. Both the LDT and cobas^® CMV were run in parallel for 109 patient samples. In addition, 104 replicates, over 8 dilutions, of the CMV IS were tested. Conversion factors and correlation between the two assays were calculated. The correlation coefficient between the LDT and cobas^® CMV was 0.91 for patient samples. The Bland-Altman graph displayed a systematic bias of +0.31 log₁₀ for the cobas^® CMV as compared to the LDT. The bias was greater for lower CMV viral loads. This increase in CMV viral loads was not seen with testing of the CMV IS dilutions by both the LDT and cobas^® CMV. CMV VLA inter-assay commutability continues to be an issue when switching CMV testing platforms and requires communication between the laboratory and clinicians during the transition period to prevent misinterpretation of results.