1556. Assessment of Translational In Vitro and Animal Pharmacokinetic–Pharmacodynamic Data Used to Support Drug Development of Recent Tetracycline Derivatives

BACKGROUND: Nonclinical (animal and in vitro) models are commonly used during the development of antibacterial drugs. Pharmacokinetic (PK) and pharmacodynamic (PD) data obtained from these nonclinical models are used to generate a PK-PD target, which can then be bridged to humans in a probability of PK-PD target attainment (PTA) analysis to support selection of the dose regimen for phase 3 trials and in vitro susceptibility testing criteria (breakpoints) to guide clinical usage. METHODS: Two recently approved tetracycline antibacterial drugs, eravacycline (ERV) and omadacycline (OMD), were evaluated. PK-PD data from nonclinical models and clinical microbiological response were collected from each of the respective clinical pharmacology reviews and assessments published by FDA and EMA, respectively. The highest MICs (minimum inhibitory concentrations) reflecting 80% success in the ability of the drug to inhibit growth in the target bacteria were identified from clinical and nonclinical data and termed the MIC cutoff. The nonclinical MIC cutoff was obtained from the PTA analysis using the PK-PD targets from animal studies. The clinical MIC cutoff was obtained from microbiological response (microbiological intent-to-treat population) data from clinical trial experience. The ratios of the clinical and nonclinical MIC cutoffs were calculated and used to evaluate potential discrepancies between the animal model prediction and clinical trial experience. RESULTS: The drug development programs for ERV and OMD included murine infection models and in vitro models to characterize PK-PD. The clinical to nonclinical MIC cutoff ratios ranged from 4 to 32. Higher values of the MIC cutoff signify that the drug can treat larger proportions of the bacterial population; therefore, high clinical to nonclinical MIC cutoff ratios signify that the drugs had more activity in reducing the bacterial population in clinical than in nonclinical studies. CONCLUSION: Thus, the nonclinical models for ERV and OMD under-predicted microbiological response and breakpoints. While nonclinical models are generally useful, more characterization of translational factors may be needed to allow nonclinical models to be more predictive of clinical trial outcomes. DISCLOSURES: All authors: No reported disclosures.