COMPARISON OF QUANTITATIVE CYTOMEGALOVIRUS ANTIGENEMIA ASSAY WITH CULTURE METHODS AND CORRELATION WITH CLINICAL-DISEASE

Blood samples, obtained predominantly from human immunodeficiency viru s-infected patients and solid-organ and bone marrow transplant recipie nts, were submitted to the clinical laboratory for detection of cytome galovirus (CMV) and were processed by three methods: conventional cult ure, centrifugation culture, and CMV antigenemia assay with monoclonal antibodies (Clonab CMV; Biotest Diagnostic Corporation, Denville, N.J .) to CMV antigens. Of 496 blood samples tested, 107 were positive by one or more methods: 56 were positive by conventional culture, 27 were positive by centrifugation culture, and 97 were positive for CMV anti gen (Ag) by the antigenemia assay. Forty-seven samples were positive b y the CMV antigenemia assay only; in these samples, a mean of 12 Ag-po sitive cells was detected per 200,000 polymorphonuclear leukocytes exa mined. In contrast, samples positive by the CMV antigenemia assay and both culture methods had a mean of 193 Ag-positive cells, and samples positive by the CMV antigenemia assay and conventional culture alone h ad a mean of 157 Ag-positive cells. In the antigenemia assay, paraform aldehyde fixation resulted in superior cell morphology when compared w ith acetone fixation. Use of immunofluorescence staining reduced sampl e processing time and the complexity of reagent preparation in compari son with immunoperoxidase staining. Differences in the sensitivities b etween the immunofluorescence and immunoperoxidase staining techniques for detection of antigenemia were minor, with discrepant samples show ing only one or two Ag-positive cells. Clinical disease was generally associated with high-level antigenemia, but exceptions were noted. The CMV antigenemia test is a rapid, quantitative assay that greatly faci litated the rapid diagnosis of CMV infection. However, quantitation of antigenemia is labor-intensive, requires processing of samples soon a fter collection, and does not always correlate with clinical disease i n the individual patient.