A diagnostic algorithm for accurate identification of non-fermentative Gram-negative rods and epidemic strains of Pseudomonas aeruginosa from cystic fibrosis patients

Abstract number: P1971

Warwick S., Duke B., Soleimanian S., Wareham D.

Objectives: Identification of non-fermentative Gram-negative bacteria using phenotypic tests such as biochemical profiling and chromogenic agar is difficult. Accurate identification is most important in individuals with cystic fibrosis (CF) as some species e.g. P. aeruginosa are associated with chronic respiratory infection whilst the role of others e.g. A. xylosoxidans is less well defined. Even amongst P. aeruginosa some strains such as the Liverpool Epidemic Strain (LES) are associated with heightened transmissibility and virulence. We developed a diagnostic algorhithm involving specific PCR and 16S rDNA sequencing for the accurate identification of CF isolates

Methods: Non-fermentative Gram-negative rods were isolated from CF sputum by standard techniques. Phenotypic identification of oxidase positive organisms was attempted by colonial morphology using Pseudomonas isolation agar and API 20 NE. PCR for P. aeruginosa specific genes (eta 396 Bp) and LES specific sequences (LES9 431 bp and PS21 400 bp) was applied to 597 consecutive isolates. Isolates negative by P. aeruginosa specific PCR were identified by sequence analysis of the 16S rDNA gene.

Results: Phenotypic identification yielded P. aeruginosa (n = 446), Burkholderia spp. (n = 6), S. maltophilia (n = 5), A. xylosoxidans (n = 2) and R. pickettii (n = 1). 139 isolates could only be identified as 'non-fermenters'. 525 isolates were positive for P. aeruginosa specific PCR. Sequence analysis of the 16S rDNA gene of the remaining isolates yielded; A. xylosoxidans (n = 19), Burkholderia cepacia complex (n = 10), P. aeruginosa (n = 10), P. putida (n = 9), S. maltophilia (n = 9), Chryseobacterium spp. (n = 7) along with Agrobacterium, Bacillus licheniformus, Bacillus simplex, Pseudomonas fulva, Proteus mirabilis, Neisseria elongata, Enterobacteriaciae, Sphingomonas multivorum (all n = 1). The LES strain of P. aeruginosa was identified in 76 isolates from 21 patients.

Conclusions: Phenotypic identification of non-fermentative Gram-negative bacteria from CF patients is extremely unreliable. A combination of P. aeruginosa specific PCR with 16S sequence analysis of negative isolates is a useful approach for the identification and detection of classical and 'emerging' pathogens in CF.