Abstract
Legionella longbeachae is the commonest Legionella species identified in patients with community-acquired pneumonia in New Zealand. Isolation of the organism on culture is the gold standard for the diagnosis of Legionnaires disease, but it has poor sensitivity (40%) compared with quantitative PCR (qPCR). We have developed a selective decontamination process using glycine, vancomycin, polymyxin, and cycloheximide (GVPC) with immunomagnetic separation (IMS) for culturing L. longbeachae. A polyclonal antibody specific for L. longbeachae was produced from New Zealand White rabbits and coupled to tosyl-activated magnetic beads. Stored L. longbeachae qPCR-positive respiratory samples were retrieved from -80 degrees C storage for testing. One portion of test samples was mixed with GVPC and the antibody bead complex, separated, washed, and cultured on modified Wadowsky and Yee agar (MWY) agar. Another portion was exposed to HCl-KCl acidic buffer (pH 2.2) before incubation on MWY agar. qPCR used probes specific for the ITS (internal transcribed spacer) region of the L longbeachae genome. Cultures were positive in 10/53 (19%) samples after acid wash and 26/53 (49%) after GVPC-IMS (P = 0.001). Growth of contaminants was rare. The mean qPCR threshold cycle values were lower in culture-positive samples after acid wash than in the culture-negative samples (mean, 29.9 versus 34.8; difference, 4.9; 95% confidence interval [CI], +/- 2.9; P = 0.001) but not after GVPC-IMS (mean, 33.0 versus 34.7; difference, 1.7; 95% CI, +/- 2.48; P = 0.16). The sensitivity of culture for L. longbeachae in respiratory specimens may be improved by using GVPC-IMS rather than acid wash for decontamination, but this should be confirmed in a prospective study of fresh specimens.