Population Pharmacokinetic Model and Meta-analysis of Outcomes of Amphotericin B Deoxycholate Use in Adults with Cryptococcal Meningitis

Abstract

There is a limited understanding of the population pharmacokinetics (PK) and pharmacodynamics (PD) of amphotericin B deoxycholate (DAmB) for cryptococcal meningitis. A PK study was conducted in n = 42 patients receiving DAmB (1 mg/kg of body weight every 24 h [q24h]). A 2-compartment PK model was developed. Patient weight influenced clearance and volume in the final structural model. Monte Carlo simulations estimated drug exposure associated with various DAmB dosages. A search was conducted for trials reporting outcomes of treatment of cryptococcal meningitis patients with DAmB monotherapy, and a meta-analysis was performed. The PK parameter means (standard deviations) were as follows: clearance, 0.03 (0.01) × weight + 0.67 (0.01) liters/h; volume, 0.82 (0.80) × weight + 1.76 (1.29) liters; first-order rate constant from central compartment to peripheral compartment, 5.36 (6.67) h^-1; first-order rate constant from peripheral compartment to central compartment, 9.92 (12.27) h^-1 The meta-analysis suggested that the DAmB dosage explained most of the heterogeneity in cerebrospinal fluid (CSF) sterility outcomes but not in mortality outcomes. Simulations of values corresponding to the area under concentration-time curve from h 144 to h 168 (AUC_144-168) resulted in median (interquartile range) values of 5.83 mg · h/liter (4.66 to 8.55), 10.16 mg · h/liter (8.07 to 14.55), and 14.51 mg · h/liter (11.48 to 20.42) with dosages of 0.4, 0.7, and 1.0 mg/kg q24h, respectively. DAmB PK is described adequately by a linear model that incorporates weight with clearance and volume. Interpatient PK variability is modest and unlikely to be responsible for variability in clinical outcomes. There is discordance between the impact that drug exposure has on CSF sterility and its impact on mortality outcomes, which may be due to cerebral pathology not reflected in CSF fungal burden, in addition to clinical variables.

Keywords: amphotericin B deoxycholate; cryptococcal meningitis; meta-analysis; pharmacodynamics; pharmacokinetics; population pharmacokinetics.

FIG 1

Amphotericin B serum concentrations in 42 patients. Patients received 1.0 mg/kg of amphotericin B deoxycholate (DAmB), infused over 5 to 6 h.

FIG 2

Linear regression of the relationship between (a) patient weight and (b) estimated glomerular filtration rate and Bayesian posterior estimates for clearance and volume of distribution. Circles are Bayesian estimates from each patient. Solid line: linear regression. (a) (Left panel) R² = 0.32. Clearance = 0.05 · weight − 0.2. (Right panel) R² = 0.12. Volume = 1.08 · weight − 24.8. (b) (Left panel) R² = 0.17. Clearance = 0.01 · eGFR + 1.03. (Right panel) R² = 0.36. Volume = 0.67 · eGFR − 31.13.

FIG 3

Scatter plots showing observed versus predicted values for the chosen population pharmacokinetic model after the Bayesian step (model 2). (Left panel) R² = 0.17. Intercept = 0.18 (95% CI, 0.03 to 0.32). Slope = 0.89 (95% CI, 0.70 to 1.09). (Right panel) R²= 0.74. Intercept = 0.01 (95% CI, −0.04 to 0.07). Slope = 1.01 (95% CI, 0.95 to 1.07). Circles, dashed lines, and solid lines represent individual observed-predicted data points, the line of identity, and the linear regression of observed and predicted values, respectively. All observed and predicted amphotericin B concentrations (conc) are indicated in milligrams per liter. AmB, amphotericin B; CI, confidence interval.

FIG 4

Visual predictive check of the final model. The black circles indicate observed DAmB concentrations. The continuous lines represent the 5th, 50th, and 95th percentiles of DAmB concentrations for 1,000 simulated patients. In total, 83.4% of observed DAmB concentrations fall within the 5th and 95th percentiles estimated by the final model, indicating adequate model fit.

FIG 5

Meta-analysis of clinical trials of DAmB monotherapy, showing dose-adjusted effects on CSF sterility (left panel), mortality at 2 weeks (middle panel), and mortality at 10 weeks (right panel). (Left panel) Tau value for unadjusted model: 4.22. Tau value for dose-adjusted model: 0.98. Dose adjustment accounts for (4.22 − 0.98)/4.22 = 77% of the heterogeneity in clinical outcomes. P value for dose adjustment, 0.007. (Middle panel) Tau value for unadjusted model: 1.90. Tau value for dose-adjusted model: 1.28. Dose adjustment accounts for (1.90 − 1.28)/1.90 = 33% of the heterogeneity in clinical outcomes. P value for dose adjustment, 0.14. (Right panel) Tau value for unadjusted model: 9.0. Tau value for dose-adjusted model: 4.93. Dose adjustment accounts for (9.00 − 4.93)/9.00 = 45% of heterogeneity in clinical outcomes. P value for dose adjustment, 0.07. RE, random effects. Day et al., reference ; Saag et al., reference ; Leenders et al., reference ; van der Horst et al., reference ; Brouwer et al., reference ; Mwaba et al., reference .

FIG 6

AUC distributions based on Monte Carlo simulations. Simulated dosing regimens are 0.4, 0.7, and 1.0 mg/kg q24h. Medians, 25th percentiles (P25), and 75th percentiles (P75) are displayed on each histogram.