Optimized Flomoxef Dosing Strategy for Extended-Spectrum Beta-Lactamase-Producing Bacteria, Based on Pooled Pharmacokinetic and Pharmacodynamic Analysis

Background: Flomoxef exhibits antibacterial activity against extended-spectrum beta-lactamase (ESBL)-producing bacteria and is a potential alternative to carbapenems. However, its pharmacokinetic (PK) characteristics remain unclear, and the appropriate minimum inhibitory concentration (MIC) breakpoint for flomoxef remains controversial. This study aimed to develop a flomoxef PK model using pooled meta-analysis and to evaluate various dosing regimens for ESBL-producing bacterial infections. Methods: Flomoxef concentration data were systematically collected from PubMed and J-STAGE. PK modeling of flomoxef was performed using the collected serum concentration data with NONMEM 7.5. Patient characteristics were examined as covariates for clearance and volume of distribution. For stochastic PK/pharmacodynamic (PD) simulation, the probability of attaining the PK/PD target {70% free time above the MIC [fT > MIC]} was calculated. The target MIC was determined by systematically reviewing flomoxef susceptibility data for ESBL-producing bacteria across different regions. Results: The population PK analysis included 39 adult patients [mean age, 58.2 years; mean body weight, 52.7 kg; and mean creatinine clearance (CLcr), 44.1 mL/min]. In the PK model, CLcr was incorporated as a covariate of flomoxef clearance. A visual predictive check and goodness-of-fit plots demonstrated relatively good stability and predictive performance of the final model. For the PD evaluation, the target MIC was set at 4 µg/mL based on the systematically collected susceptibility data for ESBL-producing bacteria. PK/PD breakpoints for dosing regimens of 1 g twice daily (2 g/day) in typical patients with CLcr = 10 mL/min, 1 g three times daily (3 g/day) in patients with CLcr = 30 mL/min, and 1 g four times daily (4 g/day) with 3 h-infusion in patients with CLcr = 60 and 90 mL/min achieved the target MIC. Conclusions: Based on the stochastic PK/PD simulation results, prolonged-infusion regimens may more effectively cover isolates with a target MIC of 4 µg/mL than short-duration infusion regimens in typical patients with normal renal function (CLcr = 60 or 90 mL/min). However, clinical studies in various scenarios including severe infections are needed to identify the relationship between PK/PD results and clinical efficacy since our findings are based solely on simulation.