The variability in immunofluorescent viral capsid antigen antibody tests in population surveys of Epstein-Barr virus infections.

A comparative study of the extent of Epstein-Barr virus (EBV) infections in populations that differ with respect to the incidence of tumours associated with this virus is now in progress in different countries. In these surveys of antibody titres from the various study populations, it is of critical importance that strict comparability be maintained. Despite standardization of techniques and reagents in the cooperating laboratories, considerable variation in the results has remained. The components of the total variability in the results of the immunofluorescence test for estimating the antibody titres against viral capsid antigens (VCA) of the EBV have been investigated. With repeated tests on the same sera, four sources of variation were measured: the reading of the slides, the performance of the tests, the use of various batches of the same cell line as antigen, and the use of different cell lines. The greatest variations were due to the use of different cell lines and to differences in performing the test; the reading of the slides caused only minor variations. Both the systematic and unsystematic variations were measured. The systematic variation was great in tests between laboratories and when different cell lines were used as antigens. Most of the systematic variation resulting from the use of different cell batches from the same cell line could be accounted for by the differing proportions of brilliant fluorescent cells. Adjustments are possible to correct the systematic variation whenever this has been measured, but not the unsystematic "residual" variability, which presents the real obstacle to the comparison of results obtained in different laboratories or by different observers. To attain full comparability of VCA antibody tests the sera from the different surveys should all be tested in the same laboratory.