Precision and accuracy of recovery of Legionella pneumophila from seeded tap water by filtration and centrifugation

Abstract

Determination of the concentration of Legionella pneumophila in environmental water sites may be useful for the prediction of the risk of a particular site's causing Legionnaires' disease as well as for experimental studies of environmental growth or remediation. The precision and accuracy of recovery of two different L. pneumophila strains from seeded tap water samples were studied, with either filtration or centrifugation used to concentrate the bacteria. L. pneumophila grown on BCYEα agar or in Acanthamoeba castellanii was used to seed sterile tap water. Water samples were then either filtered (0.2-μm pore size) or centrifuged. An average of 53% (95% confidence interval [CI], 47 to 58%; n = 45) of the seeded L. pneumophila organisms were recovered by filtration with fiat polycarbonate membranes. This recovery was significantly higher (P < 0.01) than that obtained by filtration with cast membranes (mean, 13%; 95% CI, 11 to 38%; n = 4) or by centrifugation at 3,800 x g for 30 min (mean, 14%; 95% CI, 2 to 25%; n = 9) or at 8,150 x g for 15 min (mean, 32%; 95% CI, 28 to 36%; n = 19). Recovery of L. pneumophila was not significantly different whether the bacteria were grown on plates or in amoebae. Use of a selective medium did not decrease the recovery efficiency, but preplating acid treatment of specimens caused an ≃30% bacterial loss. The lower limit of detection of L. pneumophila from 50 ml of seeded tap water was highly variable and was in the range of 1 to 10 CFU/ml when filtration was performed with single-layer membranes and the membranes were either plated directly or first minced and suspended in water before plating. Filtration with single-layer membranes is the superior method for the recovery of L. pneumophila from seeded tap water. All concentration methods have highly variable recovery rates, making accurate quantification of low concentrations of L. pneumophila difficult. Even the most efficient recovery method can be very inaccurate.