A PK-PD modelling and simulation-based evaluation of tedizolid dosing strategies for immunocompromised patients
Khalid Iqbal (1), Sebastian G. Wicha (1)
(1) Dept. of Clinical Pharmacy, Institute of Pharmacy, University of Hamburg, Hamburg, Germany
Objectives: Tedizolid (TDZ) is a clinically useful antibiotic with efficacy against Gram-positive bacteria. The currently recommended dose to treat skin and skin structure infections is 200 mg/day. Multiple pharmacokinetic-pharmacodynamic (PK-PD) studies indicate the contribution of the immune system to the bacterial killing of TDZ. Hence, and due to lack of data, TDZ is not recommended for neutropenic patients. The aim of this study was to (i) to characterize the “pure” PK-PD relationship for TDZ against E. faecalis in a hollow fiber infection model (HFIM) and (ii) to develop a pharmacokinetic (PK) -pharmacodynamic (PD) model to evaluate the currently approved and/or predict potential alternative human dosing regimens with high bacterial killing and resistance suppression against E. faecalis in immunocompromised patients.
Methods: The unbound concentration-time profiles of TDZ at doses of 200-5000 mg/day were simulated using a pharmacokinetic (PK) model in the HFIM over 120 h against E. faecalis (MIC=0.25 mg/L). High-performance liquid chromatography was used to confirm TDZ concentrations in the bacterial compartment. Colony forming units (CFU) were quantified on drug-free and TDZ-containing (3x, 5x MIC) Muller-Hinton Agar (MHA) plates for total and resistant subpopulations, respectively. A PK-PD model was developed consisting of compartments for a susceptible and resistant (preexisting population with reduced susceptibility) subpopulation to describe total CFU, and compartments to model CFU of subpopulations resistant to 3x, 5x MIC of TDZ. For clinical trial simulations, a PK model was linked to the developed PK-PD model and utilized to predict the CFU in plasma and in target tissues (adipose, muscle, epithelial lining fluid (ELF) and sputum) at 24 h and 120 h of therapy. Numerical and graphical model diagnostic criteria (goodness of fit plots and visual predictive checks (VPC’s)) were used to guide model building. NONMEM® 7.4 was used for all estimation and simulation tasks.
Results: The measured PK concentrations in the HFIM were in line with the planned PK profiles. The PK-PD model successfully described the bacterial growth kinetics for all bacterial populations after TDZ exposure.
At the human recommended dose of 200 mg/day, killing was predicted in ELF only, whereas growth was predicted in plasma and the other target tissues at 24 h. Bacteriostasis in all target tissues was observed only at a higher dose of 800 mg/day at 24 h. Except for ELF, subpopulations resistant to 3x MIC were amplified in plasma and the other target tissues at doses of 200-800 mg/day, whereas a dose of >800 mg/day suppressed the resistance development. Resistance selection at 5x MIC plates was sparse.
Significant differences in the magnitude of bacterial killing at 24 h and 120 h were observed. A median of 0.78 log CFU growth (10th percentile-90th percentile: -1.26 to 2.21) was reached in sputum at 24 h as compared to 2.56 log CFU growth (-1.52 to 2.63) at 120 h at a dose of 200 mg/day. Similarly, at 800 mg/day, a median of 0.15 log CFU kill (-0.51 to 1.46) was predicted in plasma at 24 h as compared to 0.30 log CFU growth (-2.03 to 2.20) at 120 h.
Conclusions: At the currently recommended standard dose of 200 mg/day, no reliable killing effect was obtained in the HFIM. Higher doses of 800 mg/day resulted in bacteriostasis at 24 h. Hence, the results indicate that higher doses of TDZ are required for immunocompromised patients. The discrepancies in magnitude of bacterial killing at 24 h and 120 h necessitate careful consideration of the PK/PD indices derived at earlier time points in preclinical studies when implied to predict endpoints in clinical trials. Application of the presented approach to further PK-PD studies is warranted.