The Diversilab system versus pulsed-field gel electrophoresis: characterization of extended-spectrum lactamase producing Klebsiella pneumoniae

Abstract number: P1726

Brolund A., Haeggman S., Edquist P.J., Gezelius L., Olsson-Liljequist B., Tegmark Wisell K., Giske C.G.

Objectives: Nosocomial and community associated spread of ESBL-producing Klebsiella pneumoniae is of great concern. Fast and reliable epidemiological typing methods for rapid identification of outbreaks and epidemic strains are of value. For these purposes pulsed-field gel electrophoresis (PFGE) is considered the gold standard method. The main limitations of PFGE are that it is time-consuming and requires rigorous standardization and experienced personnel in order to achieve reproducible results. Hence, faster, simpler and more easily standardized tools are needed. The DiversiLab system (DL) has been proposed for these purposes. We here, to our knowledge, show the first systematic comparison of DL and PFGE on K. pneumoniae.

Methods: We compared DL to PFGE on a national collection of 48 K. pneumoniae producing extended spectrum b-lactamases, as defined by clavulanic acid-reversible resistance to oxyiminocephalosporins collected in 26 clinical laboratories in Sweden during February – April 2007. Different cut offs for the Dice coefficients were evaluated for PFGE; 90, 85 and 80%. The cut-off for DL was set at 97% identity in addition to electropherogram overlay analysis, where no peak differences were accepted for isolates to be considered identical.

Results: Simpson's index of diversity was higher for DL (88.6%) than for PFGE (85.6 and 83.9% for Dice 90/85 and 80, respectively). The directional information about the partition relations by Wallace coefficients demonstrated that the probability of two strains being classified as the same type by PFGE having the same DL type was 79.6% (95% CI: 57.6–100) to 70.3% (95% CI: 46.9–94.8) (Dice 90/85 and 80). Conversely the probability was 100% regardless of cut-off. Only four of 48 isolates had discordant results with the two methods.

Conclusion: Although the number of isolates investigated in this study was relatively low and the results therefore should be interpreted with some caution, DL and PFGE showed a high degree of concordance. However, if one considers PFGE the gold standard method for probable close epidemiological relatedness in outbreak investigations the apparent higher discriminatory power of DL could represent a challenge for laboratories using this method for primary screening of clonal relatedness.