Correlation between Different Commercially Available Real-Time Polymerase Chain Reaction Tests for Diagnosis of CMV Viremia

Background: Blood CMV viral load (VL) has widely been used for diagnosis of CMV infection/disease. Recently, the real-time polymerase chain reaction (r-PCR) test has commercially been available but the data correlating the result between different tests is limited.

Methods: A retrospective study of 186 saved blood specimens of adult patients (pts) who were tested for blood CMV viral load from October 2015 to January 2016. We compared the blood VL value using the Roche COBAS AmpliPrep/COBAS Taqman (CAP/CTM) assay and the Abbott RealTime CMV assay (ARC). The results of r-PCR using the ARC assay were compared to the CAP/CTM assay for the test performance as well as test linearity. Patients’ medical records were reviewed for demography, underlying disease, and clinical diagnosis.

Results:

A total of 186 specimens were tested. Patients’ mean age was 46.4 ± 19.4 years old and 59.3 % were female. Major underlying disease included solid organ transplantation (31.1%), autoimmune disease (30.4%), hematologic malignancy (17%). By CAP/CTM, 60 (44.4%) had detectable blood VL [asymptomatic in 28.9% and symptomatic in 15.6%]. By ARC, 132 (71%) had detectable blood VL [asymptomatic in 73 %, and symptomatic in 27%]. A total of 31(16.7%) episodes were detectable viremia by ARC but undetectable viremia by CAP/CTM. One with CMV retinitis had undetectable blood VL by CAP/CTM but detectable blood VL (1.9 log copies/ml) by ARC. Among cases with detectable viremia, median blood VL by CAP/CTM and ARC assay was 3.25 (2.67-4.23) and 3.69 (3.06-4.70) log copies/ml respectively. The median blood VL of symptomatic cases was significantly higher than asymptomatic cases, both by CAP/CTM and ARC test [(4.36 (3.45-5.19) VS. 2.89 (2.60-3.85) log copies/ml, p<0.001 by CAP/CTM and (4.70 (3.90-5.48) VS. 3.36 (2.82-3.93) log copies/mL, p<0.001 by ARC, respectively]. Between the 2 tests, the median blood VL tested by the CAP/CTM assay was 0.09 log copies/mL (SD 0.63) higher than the ARC assay. Concordance between the two assays was 98.4%. The assays showed a good linearity (R²= 86.4%) with a correlation coefficient of 0.93, p<0.001.

Conclusion: The concordance between the CAP/CTM and ARC was excellent but the latter detected less clinically significant low graded viremia more frequently than the CAP/CTM.