Beyond plasma: defining the translation value of epithelial lining fluid exposure profiles for the treatment of pneumonia

Background Plasma pharmacokinetic/pharmacodynamic (PK/PD) targets are a surrogate of pulmonary exposures for pneumonia, despite potential differences in exposures. This study defined free plasma and pulmonary epithelial lining fluid (ELF) PK/PD targets for different antibiotics against Klebsiella pneumoniae (KP) in a preclinical pneumonia model, and assessed the predicted cfu/lung response of human exposures relative to MIC breakpoints. Method Secondary analyses were performed on previously published in vivo cfu/lung data generated using 17 KP after administration of different dosing regimens of meropenem, cefiderocol, levofloxacin and tobramycin. Efficacy was evaluated as change in log10 cfu/lung at 24 h. Emax models were fitted to composite cfu data, and plasma and ELF targets were calculated using the Hill equation. The average human plasma and ELF exposures achieved clinically relative to MIC breakpoints were overlaid on the Emax curve to predict corresponding cfu/lung efficacy. Results All Emax models had an R2 > 0.85. ELF targets differed from plasma targets and were generally lower (meropenem, cefiderocol and tobramycin), but higher for levofloxacin. Based on average human ELF exposures, the predicted cfu/lung responses of all four antibiotics were ≥1 log10 cfu/lung reduction at susceptible MICs, variable at intermediate MICs and limited at resistant MICs. Based on plasma targets, both meropenem and cefiderocol predicted cfu reduction at resistant breakpoints, whereas ELF exposures predicted net growth. Conclusion These data demonstrate the clinical translation of the pneumonia model relative to established breakpoints and emphasizes the potential limitations of relying solely on plasma exposure targets in predicting clinical efficacy for pneumonia.