Exploration of Reduced Doses and Short-Cycle Therapy for Darunavir/Cobicistat in Patients with HIV Using Population Pharmacokinetic Modeling and Simulations

Abstract

Background and objectives: Protease inhibitors such as darunavir are an important therapeutic option in the anti-human immunodeficiency virus arsenal. Current dosage guidelines recommend using cobicistat- or ritonavir-boosted darunavir 800 mg every 24 h (q24h) in protease inhibitor-naïve patients, or ritonavir-boosted darunavir 600 mg q12h in experienced patients. However, darunavir displays a large, poorly characterized, inter-individual pharmacokinetic variability. The objectives of this study were to investigate the pharmacokinetics of darunavir and to elucidate the sources of its inter-individual variability using population pharmacokinetic modeling. Then, to determine the appropriateness of current treatment guidelines and the feasibility of alternative dosing regimens in a representative cohort of adult patients using simulations.

Methods: Sparse pharmacokinetic samples were collected in 127 patients with human immunodeficiency virus type 1 infection, then supplemented with rich sampling data from a subset of 12 individuals. Data were analyzed using the nonlinear mixed-effects modeling software NONMEM. The effect of reduced doses (600 mg q24h and 400 mg q24h) or reduced frequency of administration (800 mg q24h for 5 days followed by 2 days of treatment interruption) was simulated.

Results: Our model adequately described the pharmacokinetics of darunavir. Predictors of individual exposure were CYP3A5*3 and SLCO3A1 rs8027174 genotypes, sex, and alpha-1 acid glycoprotein level. No relationship was apparent between darunavir area under the curve and treatment efficacy or safety. For reduced dose regimens, darunavir concentrations remained above the protein binding-corrected EC50 in the majority of subjects. More stringent pharmacokinetic targets were not reached in a significant proportion of patients.

Conclusions: These results add to the growing body of evidence that darunavir-based therapy could be simplified to reduce costs and toxicity, as well as to improve patient compliance. However, the heterogeneity in pharmacokinetic response should be considered when assessing whether individual patients could benefit from a particular regimen, for instance through the use of population pharmacokinetic models.

Clinical trial registration: ClinicalTrials.gov identifier: NCT03101644, date of registration: 5 April, 2017.

Conflict of interest statement

Gabriel Stillemans, Leila Belkhir, Bernard Vandercam, Anne Vincent, Vincent Haufroid, and Laure Elens have no conflicts of interest that are directly relevant to the content of this article.

Figures

Fig. 1

Goodness of fit plots. a Population predicted concentrations (PRED) vs observations. b Individual predicted concentrations (IPRED) vs observations. c Conditional weighted residuals (CWRES) vs PRED. d CWRES vs time after dose

Fig. 2

Normalized distribution prediction error (NPDE). aQ–Q plot of NPDE. b Histogram of NPDE. Shaded area represents theoretical distribution. c NPDE vs time after dose. Shaded areas represent the prediction intervals associated with the 5th, 50th, and 95th percentiles. d NPDE vs predicted concentrations