ABSTRACT Cytomegalovirus (CMV) viral loads in hematopoietic stem cell transplant (HSCT) recipients are typically monitored using quantitative molecular assays. The Roche Cobas AmpliPrep/Cobas TaqMan CMV test (Cobas CMV) has recently been cleared by the FDA for the monitoring of CMV viral loads in plasma samples from transplant patients. In this study, we compare and correlate the viral loads obtained by a laboratory-developed test (LC CMV) (using Roche analyte-specific reagents [ASR] on the LightCycler 2.0) on whole-blood specimens with those obtained on corresponding plasma and whole-blood specimens by the Cobas CMV assay. Testing was performed on 773 archived patient specimens. The strength of the agreement was good for the two assays performed on whole blood (κ = 0.6; 95% confidence interval [CI], 0.51 to 0.7) and moderate when the tests were performed on different sample types (κ = 0.54; 95% CI, 0.47 to 0.62 for the LC CMV whole blood [WB] assay versus Cobas plasma [PL], and κ = 0.57; 95% CI, 0.5 to 0.65 for the Cobas WB assay versus Cobas PL), although the difference was not statistically significant. Using a combination gold standard (i.e., a true positive was a specimen that was positive by two or more methods), the sensitivity and specificity of the assays were 78.8% and 99.3% for the LC CMV assay, 85.2% and 98.1% for the Cobas CMV WB assay, and 100% and 90.5% for Cobas CMV PL assay, respectively. A comparison of the CMV viral load trends in both plasma and whole blood from a few patients with multiple positive successive samples showed similar slopes, with differences in the slope ranging from 0.01 to 0.22. However, the absolute value for individual viral load differed markedly with whole-blood viral loads, being on average 0.5- to 1.22-log higher than those in plasma. The Cobas CMV assay provides a valid option for the monitoring of viral loads in transplant patients. Due to its increased sensitivity, the detection of CMV DNA in patients with low viral loads (i.e., those below limit of quantification [LOQ]) is increased with the Cobas CMV assay in plasma specimens. Longitudinal prospective studies will be needed to examine the clinical significance of these low-level viral loads.