Population Pharmacodynamics of Amphotericin B Deoxycholate for Disseminated Infection Caused by Talaromyces marneffei

Abstract

Amphotericin B deoxycholate (DAmB) is a first-line agent for the initial treatment of talaromycosis. However, little is known about the population pharmacokinetics and pharmacodynamics of DAmB for talaromycosis. Pharmacokinetic data were obtained from 78 patients; among them, 55 patients had serial fungal CFU counts in blood also available for analysis. A population pharmacokinetic-pharmacodynamic model was fitted to the data. The relationships between the area under the concentration-time curve (AUC)/MIC and the time to blood culture sterilization and the time to death were investigated. There was only modest pharmacokinetic variability in the average AUC, with a mean ± standard deviation of 11.51 ± 3.39 mg·h/liter. The maximal rate of drug-induced kill was 0.133 log₁₀ CFU/ml/h, and the plasma concentration of the DAmB that induced the half-maximal rate of kill was 0.02 mg/liter. Fifty percent of patients sterilized their bloodstreams by 83.16 h (range, 13 to 264 h). A higher initial fungal burden was associated with a longer time to sterilization (hazard ratio [HR], 0.51; 95% confidence interval [CI], 0.36 to 0.70; P < 0.001). There was a weak relationship between AUC/MIC and the time to sterilization, although this did not reach statistical significance (HR, 1.03; 95% CI, 1.00 to 1.06, P = 0.091). Furthermore, there was no relationship between the AUC/MIC and time to death (HR, 0.97; 95% CI, 0.88 to 1.08; P = 0.607) or early fungicidal activity {slope = log[(0.500 - 0.003·(AUC/MIC)]; P = 0.319} adjusted for the initial fungal burden. The population pharmacokinetics of DAmB are surprisingly consistent. The time to sterilization of the bloodstream may be a useful pharmacodynamic endpoint for future studies. (This study has been registered at the ISRCTN registry under no. ISRCTN59144167.).

Keywords: PK-PD; Penicillium; Talaromyces; amphotericin; antifungal; pharmacodynamics; population pharmacokinetics; talaromycosis.

FIG 1

Observed-predicted values for the pharmacokinetic and pharmacodynamic models. The Bayesian posterior (i.e., individual predicted) values are shown in both panels. (A) Observed-predicted values for plasma concentrations of amphotericin B deoxycholate; (B) observed-predicted values for the fungal density in blood.

FIG 2

Histogram of the distribution of the average AUC_0–24 for the 78 patients in the pharmacokinetic study (calculated as the entire AUC over the treatment course divided by the number of days of treatment). The mean ± standard deviation AUC was 11.84 ± 3.54 mg·h/liter, and the median AUC was 11.16 mg·h/liter.

FIG 3

(A) Kaplan-Meier plot of the time to sterilization. (B) Time course of the reduction of the fungal burden for each of the 55 patients in the pharmacokinetic-pharmacodynamic study. The open triangles represent data points from individual patients, and the solid lines are the model estimates for each of the 55 individual patients.

FIG 4

Predictions from the statistical models fitted to the pharmacodynamic data. (A) Cox model of AUC/MIC versus time to sterilization adjusted for the initial fungal burden. The change in the hazard ratio with AUC/MIC was estimated for the median initial fungal burden of 2.5 log₁₀ CFU/ml. The hazard ratio adjusted for the initial fungal burden was 1.03 (95% CI, 1.00 to 1.06; P = 0.091). (B) Cox model of AUC/MIC versus time to death for the median initial fungal burden of 2.5 log₁₀ CFU/ml. The hazard ratio adjusted for the initial fungal burden was 0.97 (95% CI, 0.88 to 1.08; P = 0.607). (C) Relationship between AUC/MIC and early fungicidal activity for the median initial fungal burden of 2.5 log₁₀ CFU/ml such that the slope was equal to log[(0.500 − 0.003·(AUC/MIC)]. A one-unit increase in AUC/MIC decreases the exp(slope) by 0.003 units (0.3%) (95% CI, −0.008 to 0.003).