Laboratory evaluation of serological tests for systemic candidiasis: a cooperative study

Abstract

Three serological tests for candidiasis, agar gel diffusion (AGD-1), whole cell agglutination (AGGL-1), and latex agglutination (LAT), were evaluated by six laboratories with 100 coded sera. In addition, each of six laboratories performed a test of its choice, either the AGD-2, the AGD-3, the AGGL-2, or one of three counterimmunoelectrophoresis (CEP) methods (CEP-1, CEP-2, and CEP-3). Results are presented by laboratory for a group of 53 "candida-involved" cases (33 proven, 14 presumptive, and 6 probable) and 47 negative controls (41 normal and 6 other disease states). The AGD-1 test produced an overall average of 85.1% positive results in the candida-involved group and 5.0% positives in the control group. The LAT produced an overall average of 89.0% positives in the candida-involved group and 17.4% positives in the controls. The AGGL-1 test produced an overall average of 63.8% positives in the candida-involved group, with 12.3% positives in the controls. In the individual tests, the best performance was shown by the CEP-3 test (92.5% positives in the candida-involved group and 2.1% positives in controls) and the CEP-1 test (88.7% positive in the candida-involved group and no positives in the controls). The tests with the highest sensitivity were the AGGL-2 and CEP-2 (94.3 and 96.2%, respectively). These tests were also the least specific (80.9 and 76.6%, respectively). In the three common tests, the AGD-1 was the most reproducible, whereas the AGGL-1 produced considerable laboratory-to-laboratory variation. Since cell-free extracts of mechanically disrupted C. albicans were used for the LAT and all the AGD and CEP tests, the difference in performance was considered to be mainly due to antigenic composition and the conditions of the test. The results of this study confirm the value of serological tests for the diagnosis of systemic candidiasis, but point out the need for standardized reagents.