Evaluation of detection methods for Legionella species using seeded water samples

South African laboratories are currently using various methods in a non-standardised approach to detect Legionella species in environmental samples. In an attempt to provide guidelines for the development of a standard method, a number of currently available detection methods were evaluated, using seeded samples of sterile and non-sterile tap water, cooling water and make-up water. The samples were seeded with a type strain of L. pneumophila serogroup 1 (American Type Culture Collection 33152). The effect of sample concentration by centrifugation and membrane filtration followed by either vortex or sonication for resuspension of organisms was studied. Three currently available culture methods were evaluated: the International standard method (ISO/DIS 11731), the Australian standard method (AS 3896 - 1991) and a locally-developed adaptation of the most probable number method (MPN). In addition, the direct immunofluorescence test and a commercially available latex agglutination test kit were included in the evaluations. The usefulness of treatment with acid or heat prior to culture was also compared. Our results indicated that concentration by membrane filtration using nitro-cellulose filters with a pore size of 0.45 ?m, followed by sonication for 10 min, would be the most appropriate concentration and resuspension method for the samples. In the absence of sample pretreatment with acid or heat, organism recovery from sterile seeded samples on BCYE ranged from 85.9 - 98.7%. However, in the non-sterile samples, these figures dropped to 8.1 - 38.5%. Sample pretreatment resulted in a further loss of at least 50% of organisms in all the samples, regardless of the pretreatment method or culture medium used. In general, the ISO and AS methods were more appropriate than the MPN method for organism recovery from sterile seeded samples. However, for the nonsterile samples, the MPN method yielded better recovery. WaterSA Vol.27(4) 2001: 523-52