Evaluation of a monoclonal antibody-based latex agglutination test for diagnosis of cryptococcosis: comparison with two tests using polyclonal antibodies.

摘要

Cryptococcal antigen detection has become a routine biological test performed for patients with AIDS. The poor prognosis of cryptococcosis explains the need for reliable tests. We evaluated the performances of a newly commercialized agglutination test that uses a monoclonal antibody specific for cryptococcal capsular polysaccharide (Pastorex Cryptococcus; Sanofi-Diagnostics Pasteur, Marnes-la-Coquette, France) and compared them with those of tests that use polyclonal immune sera (Cryptococcal Antigen Latex Agglutination System, Meridian Diagnostics, Inc., Cincinnati, Ohio; and Crypto-LA, International Biological Labs Inc., Cranbury, N.J.). The sensitivities and specificities of the tests were compared by using purified polysaccharides and yeast suspensions. Clinical specimens (131 serum samples, 41 cerebrospinal fluid samples, 34 urine samples, and 19 bronchoalveolar lavage samples) from 87 human immunodeficiency virus-positive subjects with (40 patients) and without (47 patients) culture-proven cryptococcosis were retrospectively tested during a blinded study. The effect of pronase treatment of samples was assessed for Pastorex Cryptococcus and the Cryptococcal Antigen Latex Agglutination System, and the antigen titers were compared. Our results show that (i) during the screening, concordance among the three tests was 97%; (ii) the use of pronase enhanced both the sensitivities and specificities of the Pastorex Cryptococcus test; (iii) titers agreed for 67% of the cerebrospinal fluid samples and 60% of the serum samples; and (iv) cryptococcosis was detected equally well with Pastorex Cryptococcus and with the other tests, whatever the infecting serotype (A, B, or D). The meaning of in vitro sensitivity and the relationship between titers and sensitivity are discussed. The results show that Pastorex Cryptococcus is a rapid and reliable test for the detection of cryptococcal antigen in body fluids and suggest that kits cannot be used interchangeably to monitor antigen titers in patients.