Interlaboratory Comparison of Cytomegalovirus Viral Load Assays

To assess interlaboratory variability in qualitative and quantitative cytomegalovirus (CMV) viral load (VL) testing, we distributed a panel of samples to 33 laboratories in the USA, Canada and Europe who performed testing using commercial reagents (n = 17) or laboratory-developed assays (n = 18). The panel included two negatives, seven samples constructed from purified CMV nucleocapsids in plasma (2.0-6.0 log(10) copies/mL) and three clinical plasma samples. Interlaboratory variation was observed in both actual (range, 2.0-4.0 log(10) copies/mL) and self-reported lower limits of detection (range, 1.0-4.0 log(10) copies/mL). Variation observed in reported results for individual samples ranged from 2.0 log(10) (minimum) to 4.3 log(10) (maximum)(.) Variation was greatest at low VLs. Assuming +/- 0.5 log(10) relative to the expected result represents an acceptable result, 57.6% of results fell within this range. Use of commercially available reagents and procedures was associated with less variability compared with laboratory-developed assays. Interlaboratory variability on replicate samples was significantly greater than intralaboratory variability (p < 0.0001). The significant interlaboratory variability in CMV VL observed may be impacting patient care and limiting interinstitutional comparisons. The creation of an international reference standard for CMV VL assay calibration would be an important step in quality improvement of this laboratory tool.