Abstract
Quantitative testing of BK virus (BKPyV) nucleic acid has become the standard of care in transplant patients. While the relationship between interassay harmonization and commutability has been well characterized for other transplant-related viruses, it has been less well studied for BKPyV, particularly regarding differences in commutability between matrices. Here, interassay agreement was evaluated among six real-time nucleic acid amplification tests (NAATs) and one digital PCR (dPCR) BKPyV assay. Differences in the commutability of three quantitative standards was examined across all assays using a variety of statistical approaches. Panels, including 40 samples each of plasma and urine samples previously positive for BKPyV, together with one previously negative plasma sample and four previously negative urine samples, were tested using all assays, with each real-time NAAT utilizing its usual quantitative calibrators. Serial dilutions of WHO, National Institute for Standards and Technology (NIST), and commercially produced (Exact/Bio-Rad) reference materials were also run by each assay as unknowns. The agreement of the clinical sample values was assessed as a group and in a pairwise manner. The commutability was estimated using both relativistic and quantitative means. The quantitative agreement across assays in the urine samples was within a single log(10) unit across all assays, while the results from the plasma samples varied by 2 to 3 log(10) IU/mL. The commutability showed a similar disparity between the matrices. Recalibration using international standards diminished the resulting discrepancies in some but not all cases. Differences in the sample matrix can affect the commutability and interassay agreement of quantitative BKPyV assays. Differences in commutability between matrices may largely be due to factors other than those such as amplicon size, previously described as important in the case of cytomegalovirus. Continued efforts to standardize viral load measurements must address multiple sources of variability and account for differences in assay systems, quantitative standards, and sample matrices.