Inter- and Intraobserver Agreement for the Interpretation of an In-house CMV Antigenemia Assay.

H. GUPTA, R. CHUA, C. PORTER, S. WOOD, D. KITCHING, B. LEGERE, T. MAZZULLI. Department of Microbiology, Mount Sinai and Princess Margaret Hospitals, University of Toronto, Toronto, Ontario.

Objective: We currently perform an in-house CMV antigenemia assay using an immunofluorescence technique. Because the reading of slides may be subjective, particularly those with few fluorescent foci, we evaluated the reproducibility with which technologists interpret the results of this assay.

Methods: 71 CMV antigenemia slides prepared in the routine clinical virology laboratory were interpreted in triplicate by three technologists who were blinded to each others results. Inter- and intraobserver agreements were assessed using the kappa statistic. Results were also compared to the routine virology report issued under non-test conditions.

Results: Inter- and intraobserver agreement were very good with kappa values >0.61 [reflecting substantial to almost perfect agreement]. When results of the study technologists were compared to the reported result issued by the routine clinical virology laboratory, agreement was moderate to substantial with kappa values in the 0.503-0.597 range.

Conclusion: There is very good inter- and intraobserver agreement for the reading and interpretation of immunofluorescent-stained CMV antigenemia slides. Although agreement between the study technologists and the routine virology laboratory was moderate to substantial, loss of fluorescence or antigen due to freezing and storage of slides as well as mislabeling may account for some of the variability observed.

Presented at:

CANADIAN ASSOCIATION FOR CLINICAL MICROBIOLOGY AND INFECTIOUS DISEASES (CACMID), 66^th Conjoint Meeting on Infectious Diseases , Toronto, ON, Canada, Nov 8-12, 1998.