Laboratory and clinical evaluation of a commercial DNA probe for the detection of Legionella spp.

We prospectively compared a commercially available Legionella DNA probe with culture and direct immunofluorescence. The analytical sensitivities of the DNA probe and direct immunofluorescence were equal. Both tests detected 4 X 10(3) CFU of Legionella pneumophila or Legionella micdadei per ml in the pulmonary secretions of experimentally infected guinea pigs. The diagnostic sensitivity of the reagent was evaluated by using 809 samples of respiratory secretions. Of 51 DNA probe-positive specimens, 31 came from patients with culture-confirmed legionellosis. Two culture-positive specimens had negative DNA probe tests. The sensitivity and specificity of the DNA probe were 93.9 and 97.4%, respectively. The sensitivity and specificity of direct immunofluorescence were 68.9 and 99.6%, respectively. The low specificity of the DNA probe resulted in an unacceptable positive predictive value (60.8%). False-positive DNA probe tests were not due to nonspecific binding of the probe or to technical problems but were associated with one lot of probe reagent. Most of the false-positive probe tests had values near the threshold value of greater than or equal to 4.0 suggested by the manufacturer. Raising the threshold value for a positive test to 7 lowered the sensitivity to 69.2% but raised the specificity to 99.2%. At this level, the performances of the DNA probe and direct fluorescent-antibody testing were equivalent. Respiratory secretions from patients receiving therapy for culture-confirmed Legionella infection remained DNA probe positive for up to 8 days, even though cultures and/or direct immunofluorescence tests often became negative. The DNA probe test is a satisfactory replacement for direct immunofluorescence but cannot replace culture for the laboratory diagnosis of Legionella infections.