We recovered an unusual bacterial strain from blood or sputum of three patients with septicemia, endocarditis, and/or respiratory failure. The three isolates were thin, curved, gram-negative, light brown, pigment-producing bacilli with variable catalase activity. They were asaccharolytic, oxidase-negative, nonmotile, and fastidious. Identification was not possible on the basis of these characteristics alone or in combination with cellular fatty acid profiles. Nucleic acid amplification and sequence analysis of the 16S rRNA gene revealed that all three isolates were identical and most closely related to the emerging pathogen Bordetella holmesii, diverging from the published sequence at three nucleotide positions (99.8% similarity). Isolation of a B. holmesii-like pathogen from sputum suggests that, in addition to producing septicemia, the organism may inhabit the respiratory tract like other Bordetella species.

The detection of cryptococcal antigen by means of the agglutination of antibody-coated latex particles is an important aid in the diagnosis of cryptococcosis. A commercially available latex agglutination test (IBL kit) was compared with the latex agglutination test from the Center for Disease Control in regard to sensitivity, specificity, and height of antigen titer. Over a 13-month period, 335 specimens were tested with both kits. There was one false-positive reaction with both kits (0.4%) and one false-negative reaction only with the CDC kit, among 18 patients who had meningitis or disseminated infection due to Cryptococcus neoformans. Sera from patients who had localized pulmonary cryptococcosis showed negative results with both kits. The antigen titers measured by the two kits were the same or within two dilutions in 22 of 26 specimens of cerebrospinal fluid or serum from patients who had proved cryptococcosis. Overall, the IBL kit compared favorably with the CDC kit for the detection of cryptococcal antigen in cerebrospinal fluid or serum.

Lysostaphin sensitivity was evaluated as a rapid screening test to differentiate Staphylococcus aureus from other species of staphylococci and micrococci. A total of 168 strains of staphylococci, 108 of which were S. aureus, were cultured overnight in brain infusion broth. Gram stains were peformed before and after a 1:10 dilution of the culture was exposed to 2 micrograms of lysostaphin per ml at 37 degrees C for 30 min. A reduction of 90% or greater in the number of organisms seen on comparison of the pre- and posttreatment Gram stains was considered a "positive" test result and was found in 106 of 108 S. aureus strains; 60 of 60 non-S. aureus staphylococci had a negative test result, showing no difference between the pre- and posttreatment Gram stains. Identical results were obtained using commerical blood culture media in place of brain heart infusion broth. Also studied prospectively were 100 blood or broth cultures which the clinical microbiology laboratory identified as containing gram-positive cocci suggestive of staphylococci. All 33 cultures later found to contain S. aureus gave positive test results; 67 of 67 non-S. aureus staphylococci, micrococci, and steptococci were negative.