Comparison of two commercial quantitative cytomegalovirus (CMV) polymerase chain reaction tests calibrated by world health organization international CMV standard

Abstract

Cytomegalovirus (CMV) viral load quantitation is important in diagnosis, follow-up, and monitoring of antiviral therapy in transplanted patients. In this study, it was aimed to compare the results of the two commercial World Health Organization (WHO) International CMV standard calibrated polymerase chain reaction tests, CMV Cobas Ampliprep/Cobas Taqman (CMV-CAP/CTM) (Roche, Germany) and Artus CMV QIASymphony-Rotorgene (CMV-QS-RGQ) (Qiagen, Germany). Both tests were performed simultaneously on 244 plasma samples. The results were measured in copies/ml and converted to IU/ml by multiplying with 0.91 for CMV-CAP/CTM and 1.64 for CMV-QS-RGQ, as specified by the manufacturers. CMV DNA was detected in 174 (71.3%) and was not detected in 52 (21.3%) of the samples and eighteen (7.4%) samples had discordant results by both of the tests. In 16 out of 18 samples with discordance, the viral load was below the dynamic measuring ranges of both tests. In one sample, CMV DNA could not be detected by CMV-CAP/CTM but detected by CMV-QS-RGQ with 497 copies/ml, and 334 copies/ ml CMV DNA was detected by CMV-CAP/CTM in another sample where it could not be detected by CMV-QS-RGQ. A high degree of agreement was found between the qualitative results of the both tests (kappa=0.80, p<0.001). For quantitative results in the dynamic measuring range of both assays (n=129), the median viral load values measured by CMV-CAP/CTM and CMV-QS-RGQ were 1140 copies/ml (range: 151-254000) and 1826 copies/ml (range: 189-551521). When the results were converted to IU/ml, the median viral load values measured by CMV-CAP/CTM and CMV-QS-RGQ were 1037 IU/ml (range: 137-231140) and 2993 IU/ml (range: 310-904133), respectively. There was a very strong correlation (r=0.94, p<0.001; r=0.94, p<0.001, respectively) between the log 10 values of the quantitative results in the dynamic measuring ranges (n= 129) as copies/ml and IU/ml of both tests. CMV-QS-RGQ values corresponding to 150, 1000, 3000 copies/ml in CMV-CAP/CTM were as 94.5, 1571, 323.5 copies/ml and CMV-QS-RGQ values corresponding to 137, 910, 2730 IU/ml in CMV CAP/CTM were as 154, 2557.6, 6965.9 IU/ml, respectively. A variation of 0.45 log(10) was determined between these values. In a total of 131 samples; 129 of them with the result of both tests in the dynamic measuring range and two of them which CMV DNA was not detected in one of the tests; it was found that 112 (85.5%) results for copy/ml, 73 (56%) results for IU/ml were within the measurement difference of +/- 0.5 log(10) and 19 (14.5%) results for copy/ml and 58 (44%) results for IU/ml were greater than +/- 0.5 log(10). Bland-Altman analysis showed that CMV-CAP/CTM test made lower measurements than CMV-QS-RGQ and the average difference for copy/ml and IU/ml results were 0.22 log(10) copies/ml and 0.47 log(10) IU/ml. In conclusion; when the results were converted to IU/ml, the number of samples with an acceptable measurement difference between the two test results (<= 0.5 log(10)) decreased and the number of samples with a measurement difference > 0.5 log(10) increased and the difference was found as statistically significant (p<0.001).Calibrating the Roche CMV CAP/CTM and Artus CMV-QS-RGQ tests with the WHO international CMV standard did not increase comparability between quantitative results in plasma samples, on the contrary, it was found that when the results were converted to IU/ml, a measurement difference indicating biologically significant viral replication was detected between the two test results.