1593. Analysis of Hospital Antimicrobial Susceptibility Test Results for Patterns of Antibiotic Resistance

BACKGROUND: Antimicrobial susceptibility tests (ASTs) are routinely performed on pathogens isolated from clinical samples. ASTs are used by clinicians to select the most appropriate treatment for antibiotic-resistant microorganisms. In aggregate, ASTs offer insight into the rise and spread of antibiotic resistance across hospitals. Here, we used ASTs to identify patterns of antibiotic resistance across drugs and microorganisms. METHODS: We conducted a retrospective analysis of 364,813 AST results from the University of Pittsburgh Medical Center from 2015 to 2018. Data regarding infection site, hospital laboratory testing, organism identification, and antibiotic susceptibilities were extracted from the laboratory information system and anonymized prior to use. The pathogens studied included Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii, Staphylococcus aureus, Proteus mirabilis, and Enterococcus faecalis. RESULTS: We identified 21 antibiotic-pathogen combinations where resistance was found in less than 1% of AST results. Concordant susceptibility results of levofloxacin and ciprofloxacin occurred the most frequently among antibiotic pairs. Additionally, concordant susceptibility results were more common within antibiotics belonging to the same antibiotic class than between classes. P. aeruginosa had the highest rate of overall concordant results with concordance occurring within all -lactam classes. In contrast, K. pneumoniae and P. mirabilis showed the least concordance, suggesting that their resistance profiles are less predictable. Notably, we did not identify any pairs of antibiotics that strongly exhibited discordant susceptibility results regardless of the microorganism. CONCLUSION: Using routinely collected clinical microbiological data, we were able to characterize pathogen-antibiotic combinations where resistance is rarely seen. Additionally, we identified pairs of antibiotics that frequently exhibited concordance susceptibilities both within and between classes. Lastly, we were unable to find evidence of discordant susceptibility results, indicating that more clinical research is needed to determine the efficacy of collateral sensitivity treatment techniques. DISCLOSURES: All authors: No reported disclosures.