Objective: The main objective was to assess the correlation between antibiotic use and carbapenem-resistant Klebsiella pneumoniae (CRKP) and carbapenem-resistant Escherichia coli (CREC) induction by antibiotics. Materials and Methods: A retrospective cohort study was conducted from January 2017 to December 2020. This study included patients with K. pneumoniae and E. coli. Kaplan-Meier analysis and Cox proportional hazard model were used to estimate the hazard of carbapenem-resistant Enterobacterales (CRE), whereas restricted cubic spline regression was used to visualize the hazard of CRE by antibiotics at different doses. Results: Two thousand fifty-six K. pneumoniae patients and 3,243 E. coli patients were included. After Cox proportional hazard model analysis, carbapenems or 1st-cephalospoins or penicillin monotherapy, male and ICU admission were associated with CRKP. CREC was associated with quinolone monotherapy. Time-to-event analysis indicated that carbapenem, β-lactamase inhibitor mixtures, and quinolones were associated with higher 30-day CRKP hazards than other antibiotics (χ2 = 33.670, p < 0.001). Further restricted cubic spline regression analysis found that the hazard of CRKP induction decreased with the increased dose of β-lactamase inhibitor mixtures, but there was no significant change in the hazard ratio of CRKP induction with the increased dose of quinolones. Moreover, there was an obvious characteristic of "parabolic curve" for the hazard of CREC induction due to β-lactamase inhibitor mixtures, and the hazard value gradually increased with the dose, reached the maximum at 24 g, and finally gradually decreased from 26 g. Conclusions: Rational use of antibiotics should be implemented and antimicrobial stewardship policies should be adjusted according to the characteristics of each hospital.
Keywords: Escherichia coli; Klebsiella pneumoniae; antibiotic; induction effects.