Dados do Trabalho

Título

Whole Genome Sequencing-Based Prediction of Antimicrobial Susceptibilities (WGS-AST) in Multidrug-Resistant Klebsiella pneumoniae from Salvador, Brazil

Introdução

Klebsiella pneumoniae is a priority pathogen for genomic surveillance of antimicrobial resistance. Although phenotypic antimicrobial susceptibility tests (AST) remain as the most traditional methods to guide therapeutic interventions, their integration with whole genome sequencing (WGS) might improve the decision-making process, then reducing the lethality of infections. 

Objetivo (s)

The aim of this study was to evaluate the reliability of using WGS to predict phenotypic antimicrobial susceptibility profiles in clinical isolates of K. pneumoniae collected in tertiary hospitals in the city of Salvador, Brazil. 

Material e Métodos

Thirteen samples of K. pneumoniae collected between 2010 and 2022 were isolated from blood, urine, cerebrospinal and abdominal fluids. The study has been approved by an institutional Ethics Committee (FACFAR/UFBA nº 2.170.080/2017). The antimicrobial susceptibilities of the isolates were tested by Vitek-2 and disc-diffusion method, to the following antibiotics: amikacin, aztreonam, cephalosporins (cefepime, ceftazidime, ceftriaxone, cefuroxime, cefoxitin), ciprofloxacin, gentamicin, piperacillin + tazobactam, imipenem, meropenem, and ertapenem. WGS was performed by Illumina technology, using paired-end reads. Genome assemblies and annotation were achieved through the BV-BRC bioinformatics platform. Genome-based prediction of AMR of all isolates was accomplished through ResFinder v.4.4. 

Resultados e Conclusão

The assembled genomes presented a mean estimated length of 5.5Mbp (5.3Mbp-5.7Mbp), with G+C contents between 57.05% and 57.30%, and contig N50s varying from 168,872bp to 499,608 bp. Plasmid sequences were detected by the mlPlasmids tool in all sequenced isolates. ResFinder predicted resistance to amikacin, aztreonam, carbapenem and cephalosporins in 38.5% of the isolates; to piperacillin + tazobactam, in 46.1% of the isolates; and to gentamicin in 23.1%. Phenotype and genotype integration revealed that blaCTX-M-15, aac(3)-IId and aac(6')-Ib-cr correctly predicted resistance for aztreonam, gentamicin and amikacin; blaNDM-1 was a good predictor for imipenem, ertapenem, meropenem and piperacillin + tazobactam; and blaCTX-M-15, blaNDM-1, blaTLA-1, blaTEM-1B, blaOXA-1, to cephalosporins; qnrB, to ciprofloxacin (accuracy 76.9%-100%). Although OqxA and OqxB were present in all isolates, these genes were correctly associated with phenotypic resistance to ciprofloxacin in only 38.5% of the studied samples, showing a very poor specificity for fluoroquinolone resistance prediction.  

Palavras Chave

antimicrobial resistance; genomic surveillance; ResFinder 4