Development of an in vitro pharmacokinetic/pharmacodynamic model in the presence of serum for studying micafungin activity against Candida albicans: a need for revision of CLSI susceptibility breakpoints.

Abstract

The CLSI breakpoint for micafungin and Candida albicans is 0.25 mg/L, higher than the CLSI epidemiological cut-off value (0.03 mg/L) whereas the EUCAST values are identical (0.016 mg/L). We developed a novel in vitro dialysis-diffusion pharmacokinetic/pharmacodynamic (PK/PD) model, confirmed correlation to in vivo outcome and studied micafungin pharmacodynamics against Canida albicans. Four C. albicans isolates, including a weak (F641L) and a strong (R647G) fks1 mutants, were studied using a 104 cfu/mL inoculum and RPMI medium with and without 10% pooled human serum. The exposure-effect relationship fAUC0-24/MIC was described for CLSI and EUCAST methodology. Monte Carlo simulation analysis included standard (100 mg i.v.) and higher (150-300 mg) doses q24h to determine the corresponding probability of target attainment (PTA). The in vitro PK/PD targets for stasis/1-log kill were 36/57 fAUC0-24/MIC in absence and 2.8/9.2 fAUC0-24/MIC in the presence of serum, and similar for wild-type and fks mutant isolates. The PTAs for both PK/PD targets were high (>95%) for EUCAST susceptible isolates but not for CLSI susceptible non-wild-type isolates (CLSI MICs 0.06-0.25 mg/L). 300 mg q24h was needed to attain PK/PD targets for non-wild-type isolates with CLSI MICs 0.06-0.125 mg/L and EUCAST MICs 0.03-0.06 mg/L. The in vitro 1-log kill effect corresponded to stasis in animal model and mycological response in patients with invasive candidiasis, thereby validating the model for studying pharmacodynamics of echinocandins in vitro. EUCAST breakpoints were well supported by our findings but our data questions whether the current CLSI breakpoint, which is higher than the epidemiological cut-off values, is appropriate.