A67 Bloodstream infections by carbapenem-resistant Klebsiella pneumoniae subsp. pneumoniae: Bayesian phylogenetic analysis of whole genomes

Bloodstream infection (BSI) caused by carbapenemase-producing Enterobacteriaceae (CPE) is a major public health concern, particularly in the hospital setting. The rapid detection of resistance patterns is of paramount importance for establishing the proper antibiotic regime. In addition, in countries where CPE are endemic, it is also important to evaluate genetic relationship among the isolates in order to trace pathogen circulation and to improve the infection control programs. This study is an application of a rapid blood culture (BC) workflow consisting of fast reporting of Gram stain results, rapid pathogen identification (using MALDI TOF technology), and a molecular assay for the detection of the major genes conferring resistance, all of them performed directly from positive BCs. The application of phylogenetic and phylodynamic analyses to bacterial whole-genome sequencing (WGS) data have become essential in the epidemiological surveillance of multidrug-resistant nosocomial pathogens. We analyzed 40 strains of Klebsiella pneumoniae subsp. pneumoniae (KP) carrying blaKPC (KP-KPC), randomly selected among 147 CPE identified from BCs collected from consecutive patients from 2013 to 2016. The number of BSIs-related CPE were 23, 31, 43, and 50 in 2013, 2014, 2015, and 2016, respectively. Among 147 CPE isolates, 143 were KP and four were Escherichia coli (EC). The gene blaKPC was detected in 117 strains of KP and in four strains of EC. Other carbapenemase genes, such as blaVIM and blaOXA-48, were detected in four and nine different isolates of KP, respectively. Moreover, 13 KP strains carried two resistance genes: twelve vehicled blaKPC plus blaVIM and one blaKPC plus blaOXA-48. Phylogenetic analysis of bacterial WGS data was used to investigate the evolution and spatial dispersion of KP in support of hospital infection control. The maximum likelihood tree showed two main clades statistically supported, with statistical support for several subclusters within as well. The minimum spanning tree showed mixing between sequences from different years and wards with only few specific groups. Bayesian analyses are ongoing, as the aid of Bayesian genomic epidemiology in combination with active microbial surveillance is highly informative regarding the development of effective infection prevention in healthcare settings or constant strain reintroduction.