Antimicrobial susceptibility testing (AST) is the major driver in designing effective therapy. As multiple resistance determinants can demonstrate the same phenotype (e.g., inhibitor resistant (IR), extended spectrum (ES), and carbapenem hydrolyzing (CH) beta-lactamases), critical information provided from AST for therapy, stewardship and infection control is currently lacking. WGS provides more comprehensive genetic data, explaining phenotype, and provides insight into clonality. Efforts are in development that apply novel statistical methods (e.g., PRIMERS I-IV) and machine learning (Sci Reports, 2108, 8, 421) to interpret results accurately and anticipate AST. Using a collection of clinical strains that spanned a 3.5-year period we tested how well the detection of problematic IR, ES, and CH bla resistance genes predicted phenotype.
41 isolates were chosen for AST from a collection of 1,777 WGS K. pneumoniae (Kp). Isolates chosen possessed the following beta-lactamases: (9 isolates) NDM; (3) NDM and OXA-48; (5) KPC-8 or KPC-14; (24) with a very complex beta-lactamase background (all possessed an inhibitor resistant TEM (IRT), SHV ESBL, +/- CTX-M, and/or +/- KPC). AST was performed using CLSI methods for piperacillin/tazobactam (PIP/TAZO), ceftazidime (CAZ), aztreonam (ATM), ceftazidime/avibactam (CAZ/AVI), CAZ/AVI/ATM, and ceftolozane/tazobactam (TOL/TAZO) by disk diffusion assay.
In all cases, blaNDM-1 and blaNDM-1/OXA-48 containing isolates were resistant to CAZ/AVI; the addition of ATM fully restored susceptibility to CAZ/AVI. Surprisingly, clinical Kp isolates bearing KPC-8 (V240G) and KPC-14 did not test fully resistant to CAZ/AVI, suggesting a more complex mechanism than the D179Y variant of KPC-3. Lastly, despite the complexity of the beta-lactamase background, CAZ/AVI retained potency. Interestingly, TOL/TAZO maintained efficacy in these same complex backgrounds in the absence of NDM, KPC and SHV-12. As previously shown in PRIMERS I-II, PIP/TAZO resistance was not observed in the majority of isolates as was predicted by the genotype. WGS in Kp to predict AST results and potentially guide clinical decisions is improved for novel combinations like CAZ/AVI.
R. Olsen, None
B. N. Kreiswirth, None
S. R. Evans, None
J. Musser, None
R. Bonomo, None