Differences between the quantitative antigenemia assay and the COBAS Amplicor Monitor quantitative PCR assay for detecting CMV viraemia in bone marrow and solid organ transplant patients

The relationship between quantitative PCR (COBAS Amplicor CMV Monitor, Roch e Diagnostics) and quantitative antigenemia (Monofluor pp65, Sanofi Diagnos tics) was examined for monitoring CMV viraemia. A total of 469 specimens fr om immunocompromised haematology and solid organ transplant patients were t ested by quantitative antigenemia and qualitative PCR. Quantitative PCR (QP CR) was performed on the 245 specimens in which CMV DNA was detected by qua litative PCR, To exclude any effect due to specific anti-CMV treatment, ana lysis of antigenemia and QPCR results was only performed on the 164 of 245 specimens collected from patients not on ganciclovir or foscarnet treatment . Forty seven specimens had < 400 CMV copies/mL and a negative antigen resu lt, four specimens were antigen positive (all between 1 to 10 positive CMV cells/2x10(5) leucocytes) and had < 400 CMV copies/mL. Fifty-one specimens had a CMV viral load greater than or equal to 400 copies/mL and a negative antigen result and 62 specimens had a CMV viral load greater than or equal to 400 copies/mL and a positive antigen. The viral load was shown to be as high as 43,000 copies/mL in some patients with a negative antigen and occur red in non-neutropenic patients. The correlation coefficient for antigen an d QPCR results for specimens from bone marrow transplant patients, was 0.69 with an average CMV viral load of 3,200 copies/mL (SEM=800) and an average antigen of nine positive CMV cells/ 2 x 10(5) leucocytes (SEM = 3). In the corresponding solid organ transplant group, the correlation coefficient fo r antigen and QPCR results was 0.71 with an average CMV viral load of 9,900 copies/mL (SEM = 2,100) and an average antigen of 26 positive CMV cells/2 x 10(5) leucocytes (SEM = 6). Both the average viral load and the average a ntigen result in specimens from solid organ transplant patients, were signi ficantly higher than the average viral load and antigen result in the corre sponding group of bone marrow transplant patients (Two-Sample-for-Means z-T est, P=0.001 and P=0.003, respectively). The differences in the kinetics of the two assays in monitoring CMV and their ability to predict CMV disease was also assessed in a sub-group of patients. in conclusion, the two assays used in this study do not always show parallel changes in CMV viral load, but may be complementary for the diagnosis and management of CMV disease. T he observation that non-neutropenic patients can have a high viral load in plasma and a negative antigenemia has implications for laboratories using a ntigenemia alone to monitor patients for CMV disease. J, Med. Virol, 64:275 -282, 2001, (C) 2001 Wiley-Liss, Inc.