Comparison of four methods for rapid detection of Pneumocystis carinii in respiratory specimens.

Four stains for the detection of Pneumocystis carinii in respiratory specimens were compared for sensitivity, specificity, preparation time, and ease of interpretation. One hundred specimens were collected. Of these, 50 were induced sputum specimens and 50 were bronchoalveolar lavage fluid. All specimens were stained with Diff-Quik (DQ) (a modified Giemsa stain), a quick silver stain, and direct and indirect immunofluorescence stains. A positive specimen was defined as any smear positive by two or more of the methods. Fifty-eight percent of specimens were positive. Seventy-four percent of the sputum specimens and 42% of the bronchoalveolar lavages were positive. The sensitivities for detection of P. carinii in sputum were 92% with silver stain, 97% with direct immunofluorescence assay (DFA), 97% with indirect immunofluorescence assay (IFA), and 92% with DQ. The sensitivities for detection in bronchoalveolar lavage were 86% with silver stain, 90% with DFA, 86% with IFA, and 81% with DQ. Preparation times varied from 90 min for the silver stain and IFA to 3 min for DQ. Costs of the tests varied from $1.50 per slide for DQ to $10.00 per slide for the silver stain and DFA. Reading times varied from 10 to 30 min for the silver stain and DQ to less than 5 min for the immunofluorescence assays. We conclude that all of these tests are viable options for the clinical laboratory, and the choice will be influenced by factors such as clinical volume, ability to batch specimens, and expertise of technological support. A reasonable option may be to use the quick and inexpensive DQ as a screening test and to confirm negative smears with a more sensitive assay.