Interplay between phenotypic resistance to relevant antibiotics in gram‐negative urinary pathogens: A data‐driven analysis of 10 years’ worth of antibiogram data

Abstract

The global emergence of antimicrobial resistance (AMR) has become a critical issue for clinicians, as it puts the decades of developments in the medical field in jeopardy, by severely lim-iting the useful therapeutic arsenal of drugs, both in nosocomial and community‐acquired infec-tions. In the present study, a secondary analysis of taxonomic and resistance data was performed, corresponding to urinary tract infections (UTIs) caused by Gram‐negative bacteria, detected between 1 January 2008 to 31 December 2017 at the Albert Szent‐Györgyi Health Center, University of Szeged. The following were identifiable from the data collected: year of isolation; outpatient (OP)/inpa-tient (IP) origin of the isolate; taxonomy; and susceptibility/resistance to selected indicator antibiotics. Principal component analysis (PCA) and a correlation matrix were used to determine the association between the presences of resistance against indicator antibiotics in each taxonomic group. Overall, data from n = 16,240 outpatient and n = 13,964 inpatient Gram‐negative UTI isolates were included in the data analyses. In E. coli, strong positive correlations were seen between resistance to ciprofloxacin (CIP) and gentamicin (GEN) resistance (OP: r = 0.6342, p = 0.049; IP: r = 0.9602, p < 0.001), whereas strong negative correlations were shown for fosfomycin (FOS) and nitrofurantoin (NIT) resistance (OP: r= −0.7183, p = 0.019; IP: r= −0.7437; p = 0.014). For Klebsiella spp. isolates, CIP resistance showed strong positive correlation with resistance to third‐generation cephalosporins (3GC) and GEN (r = 0.7976, p = 0.006 and r = 0.7428, p = 0.014, respectively) in OP isolates, and with resistance to trimethoprim‐sulfameth-oxazole (SXT) and FOS (r = 0.8144, p = 0.004 and r = 0.7758, p < 0.001, respectively) in IP isolates. For members of the Citrobacter‐Enterobacter‐Serratia group, the resistance among indicator antibiotics showed a strong positive correlation, with the exception of FOS resistance. In the Proteus‐Providencia‐ Morganella group, the strongest association was noted between CIP and SXT resistance (OP: r = 0.9251, p < 0.001; IP: r = 0.8007; p = 0.005). In the case of OP Acinetobacter spp., CIP showed strong and significant positive correlations with most indicator antibiotics, whereas for IP isolates, strong negative correlations arose among imipenem (IMI) resistance and resistance to other drugs. For Pseudomonas spp., strong and positive correlations were noted among resistance to β‐lactam antibiotics and aminoglycosides, with the exception of ceftazidime (CEFT), showing strong, but negative correlations. Though molecular tests and sequencing‐based platforms are now considered as the gold‐standard for AMR surveillance, standard-ized collection of phenotypic resistance data and the introduction of Big Data analytic methods may be a viable alternative for molecular surveillance, especially in low‐resource settings.