Micafungin Sequestration and Late Release During Continuous Renal Replacement Therapy with Polyacrylonitrile- and Polysulfone-Derived Filters.

Abstract

Introduction: Adsorption within continuous renal replacement therapy (CRRT) circuits may reduce exposure to echinocandins. Because micafungin is highly protein bound, the behavior of its unbound fraction during CRRT remains difficult to characterize. We assessed unbound micafungin disappearance from a central compartment and late release/desorption in a protein-free in vitro CRRT model.

Methods: Micafungin stability was assessed in a 5-L bag of Hemosol™ B0 over 8 h. In the NeckEpur model, a 5-L protein-free central compartment was circulated at 200 mL/min for 6 h through either a polyacrylonitrile hemofilter (ST™150; post-dilution continuous veno-venous hemofiltration [CVVH], 2.5 L/h) or a polysulfone hemofilter (AV™1000). For AV™1000, one run used CVVH (2.5 L/h) and one used continuous veno-venous hemodiafiltration (CVVHDF; dialysis 1.5 L/h plus filtration 1.0 L/h). Initial micafungin concentrations in the central compartment were 2.18 mg/L or approximately 6.8 mg/L. Concentrations were measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS) with a lower limit of quantification (LLOQ) of 0.1 mg/L. Apparent central-compartment clearance (Cl_CC), sieving coefficient (Sc), and extraction coefficient (EC) were used to describe disappearance from the circuit, filtration, and late release. Sensitivity analyses replaced values < LLOQ by LLOQ/2 or LLOQ/√2. Segmental sampling and within-filter mass balance were used descriptively to localize loss within the extracorporeal system.

Results: Micafungin was stable in Hemosol™ B0 over 8 h (- 5.2 ± 0.5%). With ST™150 at 2.18 mg/L, C_CC reached the LLOQ by 200 min and was below the LLOQ thereafter; apparent Cl_CC was approximately 6 L/h, effluent concentrations were not measurable, and limited late release was estimated over 120-200 min. With ST™150 at approximately 6.8 mg/L, elimination from the central compartment was 92 ± 4% at 6 h (apparent Cl_CC 4.9 ± 0.2 L/h), with within-filter contributions from measurable effluent removal (54 ± 6%) and non-effluent loss (46 ± 6%); small late release was compatible with negative EC values during 180-360 min. For AV™1000, two exploratory runs-one CVVH and one CVVHDF-showed rapid disappearance from the central compartment, with C_CC below the LLOQ by 120 min, no measurable effluent concentrations, and no negative EC values. Sensitivity analyses for values < LLOQ changed the numerical Cl_CC estimates but not the overall pattern. Segmental concentration data and within-filter mass-balance analysis localized the dominant loss to the filter module, with only limited upstream inlet-segment contribution.

Conclusion: In this exploratory protein-free in vitro model, unbound micafungin rapidly disappeared from the CRRT circuit with both tested filter systems. ST™150 showed measurable effluent removal at higher concentrations and limited late release, whereas both exploratory AV™1000 runs showed rapid disappearance without detectable late release. Segmental concentration data and within-filter mass-balance analysis localized the dominant loss to the filter module, with only limited upstream inlet-segment contribution. These findings characterize unbound micafungin-circuit interactions under the tested conditions; their clinical translation will depend on in vivo protein binding and rebinding kinetics.