Use of a clinically derived exposure–response relationship to evaluate potential tigecycline-Enterobacteriaceae susceptibility breakpoints

Abstract

Potential tigecycline-Enterobacteriaceae susceptibility breakpoints were evaluated using 2 approaches, which differed in the nature of the probabilities assessed by MIC value. Using a previously derived tigecycline population pharmacokinetic model and Monte Carlo simulation, a probability density function of steady-state area under the concentration–time curve for 24 h (AUC SS(0-24)) values for 9999 patients was generated. AUC SS(0-24) values were divided by clinically relevant fixed MIC values to derive AUC SS(0-24)/MIC ratios, which were used to calculate the clinical response expectation by MIC value based upon a logistic regression model for efficacy (1st approach). For the 2nd approach, the probability of pharmacokinetic–pharmacodynamic (PK-PD) target attainment was calculated as the proportion of patients with AUC SS(0-24)/MIC ratios greater than the threshold value of 6.96, the PK-PD target associated with optimal clinical response. Probabilities of clinical response and PK-PD target attainment were poorly correlated at MIC values >0.25 mg/L. For instance, the median probability of clinical success was 0.76, whereas the probability of PK-PD target attainment was 0.27 at an MIC value of 1 mg/L, suggesting that the probability of PK-PD target attainment metrics underestimates the clinical performance of tigecycline at higher MIC values.