849. Whole Genome Sequencing Analysis of Klebsiella pneumoniae Isolates Reveals Diversity in Genetic Antibiotic Resistance Patterns

Abstract Background Klebsiella pneumoniae is among the leading causes of healthcare-associated infections (HAI). Multidrug-resistant (MDR) Klebsiella variants are difficult to treat and have been reported with increasing frequency in hospitals. Whole genome multi-locus sequence typing (wg-MLST) of K. pneumoniae HAI isolates was used to compare antibiotic resistance genetic patterns against epidemiologic typing. Figure 1. Prevalence of ABX resistance genes for each drug class. Figure 2. Prevalence of ABX resistance genes by hospital. Methods Forty-six clinical bacterial HAI isolates were collected from patients admitted to two disparate tertiary care hospitals in Detroit, Michigan between 2017 and 2019. Data output from wg-MLST was de novo assembled using SPAdes (version 3.7.1) assembler on the Bionumerics calculation engine. Minimum spanning tree (MST) analysis categorized isolates into unique MLST Pasteur serotypes (ST). Antimicrobial resistance genes and/or chromosomal mutations were identified using the ResFinder Database (version 3.2). K. pneumoniae isolates were compared for antibiotic-resistance patterns by hospital, unit, and ST. Figure 3. Hospital 1 clusters and strains that detected gene qnr1, arranged by unit for visual comparison. Figure 4. Sample of clinically significant genes and prevalence, excluding SHV family of genes for simplicity. Results There was significant genetic diversity among the isolates, and no predominant strain was identified. MST analysis revealed 17 unique strains. Only six strains had genetically unique resistance genes detected in more than one isolate, and only three of six were hospital-specific; none were unit-specific. Out of the 75 unique resistance genes detected, only 8 genes had a prevalence >50%: oqxA (100%), oqxB (100%), fosA (89%), blaCTX-M-15 (76%), aac(6’)-Ib-cr (61%), blaTEM-1B (52%), blaOXA-1 (52%), and catB3 (52%). No colistin resistance genes were detected. Of the remaining 69 low-prevalence resistance genes, only 8 hospital-specific genes were detected in more than one isolate (qnrB1, blaSHV-1, blaSHV-110, ac(6’)-Ib3, blaCTX-M-3, blaSHV-36, blaSHV-80, blaSHV-178) with a prevalence range of 4%-22%. Conclusion Our genetic analysis of antibiotic resistance patterns and wg-MLST revealed significant heterogeneity among the isolates, indicating no common source of transmission for either hospital. Although K. pneumoniae is a very common nosocomial pathogen, etiologic analysis suggests diverse community strains (e.g., gut colonization) may actually be responsible for previously-designated HAI. Disclosures Chetan Jinadatha, MD, MPH, AHRQ (Research Grant or Support)Department of Veterans Affairs (Other Financial or Material Support, Owner: Department of Veterans Affairs. Licensed to: Xenex Disinfection System, San Antonio, TX)Inventor (Other Financial or Material Support, Methods for organizing the disinfection of one or more items contaminated with biological agents)NiH/NINR (Research Grant or Support)NSF (Research Grant or Support)Xenex Healthcare Services (Research Grant or Support) Mark Stibich, PhD MHS, Xenex Disinfection Services, Inc (Board Member, Employee)