Simultaneous pharmacokinetic modeling of unbound and total darunavir with ritonavir in adolescents: a substudy of the SMILE trial

Abstract

Darunavir (DRV) is an HIV protease inhibitor commonly used as part of antiretroviral treatment regimens globally for children and adolescents. It requires a pharmacological booster, such as ritonavir (RTV) or cobicistat. To better understand the pharmacokinetics (PK) of DRV in this younger population and the importance of the RTV boosting effect, a population PK substudy was conducted within SMILE trial, where the maintenance of HIV suppression with once daily integrate inhibitor + darunavir/ritonavir in children and adolescents is evaluated. A joint population PK model that simultaneously used total DRV, unbound DRV, and total RTV concentrations was developed. Competitive and non-competitive models were examined to define RTV's influence on DRV pharmacokinetics. Linear and non-linear equations were tested to assess DRV protein binding. A total of 443 plasma samples from 152 adolescents were included in this analysis. Darunavir PK was best described by a one-compartment model first-order absorption and elimination. The influence of RTV on DRV pharmacokinetics was best characterized by ritonavir area under the curve on DRV clearance using a power function. The association of non-linear and linear equations was used to describe DRV protein binding to alpha-1 glycoprotein and albumin, respectively. In our population, simulations indicate that 86.8% of total and unbound DRV trough concentrations were above 0.55 mg/L [10 times protein binding-adjusted EC₅₀ for wild-type (WT) HIV-1] and 0.0243 mg/L (10 times EC₉₀ for WT HIV-1) targets, respectively. Predictions were also in agreement with observed outcomes from adults receiving 800/100 mg DRV/r once a day. Administration of 800/100 mg of DRV/r once daily provides satisfactory concentrations and exposures for adolescents aged 12 years and older.

Keywords: adolescent; children; darunavir; free fraction; population pharmacokinetics; ritonavir; unbound.

Fig 1

Time point distribution of blood samples collected and used for the PK analysis.

Fig 2

Schematic representation of the final joint model. F is the bioavailability, D is the dose administered, ka is the absorption constant, V is the volume of distribution, ke is the elimination constant, [AAG] is the AAG concentration, N_AAG is the number of binding sites on AAG for DRV, K_d is the dissociation constant of AAG for DRV, [HSA] is the albumin concentration, θ_HSA is the binding constant of HSA for DRV, and AUC_RTV is the RTV AUC_0-24h inhibiting the elimination of DRV.

Fig 3

Goodness-of-fit plots from the final joint model. Red, blue, and gray points are unbound DRV, total DRV, and RTV concentrations, respectively. NPDE is the normalized prediction distribution error. Solid lines represent identity lines (top plots) or the theoretical mean of NPDE (bottom plots). Orange dashed lines are the regression lines.

Fig 4

pcVPC plots for total DRV (left), unbound DRV (middle), and RTV (right) concentrations. Black points represent observed concentrations. Solid black lines represent the 10th, 50th, and 90th percentiles of observed concentrations. Gray areas represent the 95% confidence interval of the 10th, 50th, and 90th percentiles of simulated concentrations.

Fig 5

Percentage of darunavir bound to HSA (dotted line), AAG (dashed line), and unbound fraction of darunavir (solid line) according to total DRV concentrations in different conditions of AAG and HSA levels. Bottom-middle plot depicts protein binding for average AAG and HSA levels in our population.