Insufficient discriminatory power of MALDI-TOF mass spectrometry for typing of Enterococcus faecium and Staphylococcus aureus isolates

MALDI-TOF mass spectrometry (MALDI-TOF MS) is increasingly used as a reliable technique for species identification of bacterial pathogens. In this study we investigated the question of whether MALDI-TOF MS can be used for accurate sub-differentiation of strains and isolates of two important nosocomial pathogens Enterococcus faecium and Staphylococcus aureus. For this purpose, a selection of 112 pre-characterized E. faecium isolates (clonal complexes CC2, CC5, CC9, CC17, CC22, CC25, CC26, CC92 altogether 52 multilocus sequence types) and 59 diverse S. aureus isolates (mostly methicillin resistant; CC5, CC8, CC22, CC30, CC45, CC398) were studied using a combination of MALDI-TOF MS and advanced methods of spectral data analysis. The strategy of MS data evaluation included manual peak inspection on the basis of pseudo gel views, unsupervised hierarchical cluster analysis and supervised artificial neural network (ANN) analysis. We were capable of differentiating patterns of hospital-associated E. faecium isolates (CC17) from other strains of E. faecium with 87% accuracy, but failed to identify lineage-specific biomarker peaks. For S. aureus pattern analyses we were able to confirm a number of signals described in previous studies, but often failed to identify biomarkers that would allow a consistent and reliable identification of phylogenetic lineages, clonal complexes or sequence types. Hence, the discriminatory power of MALDI-TOF MS was found to be insufficient for reliably sub-differentiating E. faecium and S. aureus isolates to the level of distinct clones or clonal complexes, such as assessed by MLST. Further, a comparison between peak patterns of susceptible and resistant isolates did not identify statistically relevant marker peaks linked to glycopeptide resistance determinants (vanA, vanB) in E. faecium, or the methicillin resistance determinant (mecA) in S. aureus.