Population pharmacokinetics and pharmacodynamics of eravacycline against pulmonary infections

Xiwei Ji(1), Wenyao Mak(2), Xiaoqiang Xiang(2), Yun Li(1), Xiao Zhu(2)

(1)Institute of Clinical Pharmacology, Peking University First Hospital, Beijing 100034, China. (2)School of Pharmacy, Fudan University, Shanghai 201203, China

Introduction:

Eravacycline is a broad-spectrum fluorocycline approved to treat complicated intra-abdominal infections. In lung-infected animal models, it was effective against methicillin-resistant S. aureus (MRSA) and tetracycline-resistant MRSA[1]. When evaluated in healthy volunteers, it showed extensive disposition in the epithelial lining fluid (ELF) and alveolar macrophages[2].

To assess eravacycline’s profile in pulmonary infections, we sought to integrate previously collected in vitro, in vivo (IVIV) and clinical data into a quantitative framework through modelling and simulation (M&S), and to predict its efficacy against main causative pathogens of community-acquired pneumonia (CAP). The study should aid decision-making in expanding its indication to include CAP.

Objectives:

1. Develop a population PK model to describe the kinetics of free eravacycline in serum and ELF
2. Assess the probability of target attainment (PTA) of five dosing regimens based on predetermined PK/PD target values
3. Determine the PK/PD breakpoint against four clinically important microorganisms

Methods:

Eravacycline data came from a phase I trial on healthy volunteers. As concentrations in ELF could not be measured directly, these were estimated with bronchoalveolar lavage (BAL) via the formula: C_ELF=C_BAL*(Urea_plasma/Urea_BAL) where C_ELF and C_BAL were eravacycline concentration in ELF or BAL,  and Urea_plasma and Urea_BAL were urea concentration in plasma or BAL.  

Nonlinear mixed-effects modeling software (NONMEM®, ICON) was used for population PK modeling. First-order conditional estimation with interaction was used for estimation. Model selection was based on changes in the AIC, standard GoF plots, and relative standard error (RSE) of parameter estimates. Stepwise covariate modeling building procedure was used. The performance of the final model was evaluated by visual predictive checks and bootstrap.

Preclinical studies had determined the ratio of steady-state AUC at ELF site over MIC (AUC_ELF/MIC) was a suitable PK/PD index for eravacycline. In this study, the pharmacodynamics for four microbes (E. coli, K. pneumoniae, A. baumannii, S. aureus) were assessed via Monte Carlo simulation under five different dosing regimens:

1. 1h infusion, 1.0mg/kg q12h for 7 days
2. 2h infusion, 1.5mg/kg q12h for 7 days
3. 2h infusion, 1.5mg/kg q24h for 7 days
4. 2h infusion, 1.5mg/kg q12h on 1^st day, 1.5mg/kg q24h 2^nd day onwards
5. 2h infusion, 1.5mg/kg q12h on 1^st day, 1.0mg/kg q12h 2^nd day onwards

PTA under different MIC values was calculated, and PK/PD breakpoints were defined as MIC values that corresponded to PTA=90%. Cumulative fraction of response (CRF), weighted by MIC distribution determined by Peking University First Hospital, was calculated for each microbe. A CFR of at least 90% was determined as optimal.

Results:

Evaracycline PK was best described with a 3-compartment model with allometric scaling, and an ELF compartment parameterized as the ELF distribution ratio (Ratio=(C_free,ELF/C_free,central) as sparse BAL samples led to imprecise estimates for k_cl and k_lc between the two compartments. Also, free drug concentration in plasma and ELF was highly correlated (R=0.66), thus providing a statistical basis for model simplification.

Eravacycline PK parameters (estimate, interindividual variability) were precisely estimated: central (V_c=3.88L, ω_Vc=115.8%) and peripheral (V_p1=39L, ω_Vp1=55%, V_p2=122L) volumes of distribution, and systemic (CL=16.3L/h, ω_CL=13.6%) and intercompartmental (Q₁=33.9L/h, Q₂=23.5L/h) clearances, Ratio=8.26 (RSE=12%). Beside allometrically scaled weight, no other significant covariate was found.

MIC₉₀ was 2mg/L (E. coli and K. pneumoniae), 0.5mg/L (A. baunmannii), and 0.05mg/L (S. aureus). All 5 regimens achieved PTA>90% for E. coli when PK/PD target=3h and MIC= 4mg/L; For K. pneumoniae, A. baumannii and S. aureus, only regimen 2 achieved PTA>90% when the PK/PD target=6h and MIC=2mg/L respectively. Regimen 2 showed superior efficacy against all microbes, while regimens 3 and 4 were inferior.  With regimen 2, the PK/PD breakpoints were: 4mg/L (E. coli), and 1mg/L (K. pneumoniae, A. baumannii, S. aureus) respectively, and CFR was >90% against all microbes.

Conclusions:

Eravacycline was widely distributed into the lungs and was effective against all tested microbes. M&S has allowed the carry-over and integration of previous knowledge and the resultant findings support the proposal to expand eravacycline indication to include CAP.

References:
[1] Grossman, T.H., et al., Eravacycline (TP-434) is efficacious in animal models of infection. Antimicrob Agents Chemother, 2015. 59(5): p. 2567-71.
[2] TConnors, K.P., et al., Phase I, open-label, safety and pharmacokinetic study to assess bronchopulmonary disposition of intravenous eravacycline in healthy men and women. Antimicrob Agents Chemother, 2014. 58(4): p. 2113-8

Reference: PAGE 31 (2023) Abstr 10320 [www.page-meeting.org/?abstract=10320]

Poster: Drug/Disease Modelling - Infection