Homology surveillance of carbapenem-resistant Klebsiella pneumoniae (CRKP) is critical to monitor and prevent outbreaks of nosocomial infections. In the present study, a matrix-assisted laser desorption/ionisation-time of flight (MALDI-TOF MS)-based method was evaluated as a rapid tool for typing CRKP in comparison with pulsed-field gel electrophoresis (PFGE) and multi locus sequence typing (MLST). Drug-resistant phenotypes and genotypes of 44 CRKP isolates were detected by microdilution broth method and polymerase chain reaction, and typed by PFGE, MLST and MALDI-TOF MS. Simpson's Index of Diversity was used to evaluate taxonomic diversity, Adjusted Rand Index (ARI) for congruence between the typing methods an... More
Homology surveillance of carbapenem-resistant Klebsiella pneumoniae (CRKP) is critical to monitor and prevent outbreaks of nosocomial infections. In the present study, a matrix-assisted laser desorption/ionisation-time of flight (MALDI-TOF MS)-based method was evaluated as a rapid tool for typing CRKP in comparison with pulsed-field gel electrophoresis (PFGE) and multi locus sequence typing (MLST). Drug-resistant phenotypes and genotypes of 44 CRKP isolates were detected by microdilution broth method and polymerase chain reaction, and typed by PFGE, MLST and MALDI-TOF MS. Simpson's Index of Diversity was used to evaluate taxonomic diversity, Adjusted Rand Index (ARI) for congruence between the typing methods and Wallace coefficients (W) for the ability of either method to predict each other. Forty-four CRKP isolates of 15 sequence types (STs) produced either NDM-1 (n = 16), NDM-5 (n = 9) or KPC-2 (n = 19) carbapenemases. PFGE differentiated these isolates into 16 distinct types, and two deoxyribonucleic acid profiles were assigned to ST337 and ST11, respectively. MALDI-TOF MS failed to clearly delineate between clusters on dendrograms based on principal components analysis and main spectrum profile. The chosen parameters resulted in a maximum ARI of 0.310 (95% CI 0.088-0.531) between MALDI-TOF MS typing and the PFGE reference, indicating a low ability of the former to correctly identify related isolates. Likewise, the maximum W coefficient of 0.367 (95% CI 0.203-0.532) showed that MALDI-TOF MS had a lower predictive power than PFGE. We conclude that MALDI-TOF MS lacks the discriminatory power necessary for clone assignment of CRKP isolates and consequently cannot be considered as a rapid and creditable method for this purpose.
