Nadège Néant1, Caroline Solas1, Minh Patrick Lê2, Yazdan Yazdanpanah3, Catherine Dhiver4, Sylvie Bregigeon5, Saadia Mokhtari4, Gilles Peytavin2, Catherine Tamalet4, Diane Descamps6, Bruno Lacarelle1, Florence Gattacceca7
1. Aix-Marseille Univ U105, APHM, SMARTc CRCM Inserm UMR1068 CNRS UMR7258, Hôpital La Timone, Laboratoire de Pharmacocinétique et Toxicologie, F-13005 Marseille, France 2. APHP, Hôpital Bichat-Claude Bernard, Laboratoire de Pharmaco-Toxicologie, IAME, UMR 1137, Université Paris Diderot, Sorbonne Paris Cité and INSERM, F-75018 Paris, France 3. Univ Paris Diderot, APHP, IAME-UMR 1137, Hôpital Bichat-Claude Bernard, Service des Maladies Infectieuses et Tropicales, F-75018 Paris, France 4. IHU Méditerranée Infection, Aix Marseille Univ., AP-HM, URMITE UM 63 CNRS 7278 IRD 198 INSERM 1095, F-13005 Marseille, France 5. APHM, Hôpital Sainte-Marguerite, Service d’Immuno-hématologie clinique, F-13009 Marseille, France 6. APHP, Hôpital Bichat-Claude Bernard, Laboratoire de Virologie, IAME, UMR 1137, Université Paris Diderot, Sorbonne Paris Cité and INSERM, F-75018 Paris, France 7. Aix-Marseille Univ U105, SMARTc CRCM Inserm UMR1068 CNRS UMR7258, F-13005 Marseille, France
Objectives: Rilpivirine (RPV) is a non-nucleoside reverse transcriptase inhibitor widely prescribed for the treatment of HIV-1 infection both in naïve and pre-treated patients. An important inter-individual pharmacokinetic (PK) variability has been observed [1] and RPV trough plasma concentrations (Ctrough) has been shown to be correlated with the virologic response [2]. However, until 2017, no study of RPV pharmacokinetics and pharmacodynamics (PD) had been published in a routine follow-up context. The only information available in the Summary of Product Characteristics came from the studies performed by the sponsor in drug development clinical trials (phase I healthy patients and phase III highly selected HIV-infected patients). The first aim of our study was to describe the RPV PK and its variability in adult HIV-infected patients in a routine follow-up context [3]. Second, we aimed to develop a PKPD model to establish dose-concentration-response relationships for future treatment optimization.
Methods: We conducted a multicenter, retrospective and observational study in patients treated with the once-daily RPV/tenofovir disproxil fumarate (TDF)/ emtricitabine (FTC) regimen. Ambulatory patients for which plasma RPV concentrations were determined within the context of routine therapeutic drug monitoring (TDM) were included from November 2012 to November 2015. Plasma HIV RNA load (VL), CD4 cells counts (CD4) and drug-resistance associated mutations were collected at baseline and during the monitoring. The population PK-PD (PopPKPD) analysis was performed with NONMEM VII software using Pirana interface 2.7.0. For the PD analysis, the individual Ctrough were predicted based on the PopPK model using Bayesian approach. A HIV dynamics model was developed to estimate the effect of RPV concentrations both on the infection rate of CD4 by the virus and on the VL [4,5].
Results: A total of 379 HIV-1 infected patients and 779 RPV plasma concentrations were included for the PopPK analysis. Overall, 24.4% of observed individual Ctrough were below the 50 ng/ml minimal threshold currently recommended. A one-compartment model with first-order absorption best described the data. The estimated fixed effect for plasma apparent clearance and distribution volume were 9 L/h and 321 L respectively, resulting in a half-life of 25.2 h. The inter-individual variability of clearance was 30.3 %.
Sixty-four treatment-naive patients were included in the PKPD analysis. We analyzed 492 data for VL and 487 for CD4 from initiation of treatment until a maximum of 2.5 years later. The parameters of the model were the production rate constant of uninfected target cells (So), elimination rate constant of infected cells (δ), production rate constant of free virions (c), VL at baseline (VL0), CD4 at baseline (CD40) and the RPV concentration producing 50% of the maximum effect (C50RPV). δ needed to be fixed to previously reported value of 15.2 per month [6]. The interindividual variability (IIV) on δ, So and c were inaccurately estimated, and consequently fixed to zero. The resulting model led to well-estimated parameters (Relative Standard Error under 30%) and was qualified based on the goodness of fit plots. However, a bias associated with an underestimation of the highest VL values at baseline was observed. The estimated value of C50RPV (79 ng/ml) was close but higher than the currently acknowledged target RPV Ctrough value of 50 ng/ml.
Conclusions: Our results showed that half-life of RPV is shorter in routine clinical practice than reported in the Summary of Product Characteristics, associated to a higher risk of under-exposure, in line with the simultaneous PopPK study performed within the same context of routine TDM by another group [7]. The PD model suggested that the currently used 50 ng/ml RPV Ctrough efficacy target might also be under-evaluated. Altogether, our PKPD model advocates for an increase of the RPV dose. However, our PD model showed some limitations, with a bias in VL prediction. Consequently, alternate HIV dynamics models will be tested, incorporating more biological prior knowledge (about slow viral decline and proliferation, age effect on CD4, thymic output) and a direct effect of RPV Ctrough on VL [8].
References:
[1] Brochot A, De La Rosa G, Vis P, et al. Generalised additive modelling of virologic response to the NNRTIs rilpivirine (RPV, TMC278) and efavirenz (EFV) in treatment-naïve HIV-infected patients: pooled data from ECHO and THRIVE. In: Abstracts of the Thirteenth European AIDS, Belgrade, Serbia, 2011. Abstract PS12/7.
[2] Crauwels HM, Van Schaick E, Van Heeswijk RPG, et al. Effect of intrinsic and extrinsic factors on the pharmacokinetics of TMC278 in antiretroviral-naïve, HIV-1- infected patients in ECHO and THRIVE. J Int AIDS Soc. 2010; 13: P186.
[3] Néant N, Gattacceca F, Lê MP, et al. Population pharmacokinetics of Rilpivirine in HIV-1-infected patients treated with the single-tablet regimen rilpivirine/tenofovir/emtricitabine. Eur J Clin Pharmacol. 2018. doi: 10.1007/s00228-017-2405-1.
[4] Perelson AS, Ribeiro RM. Estimating drug efficacy and the viral dynamic parameters: HIV and HCV. Stat Med. 2008; 27:4647-4957.
[5] Lavielle M, Samson A, Karima Fermin A, et al. Maximum likelihood estimation of long-term HIV dynamic models and antiviral response. Biometrics 2011; 67: 250-259.
[6] Perelson AS, Neumann AU, Markowitz M, et al. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science 1996; 271:1582–1586.
[7] Aouri M, Barcelo C, Guidi M, et al. Population pharmacokinetics and pharmacogenetics analysis of Rilpivirine in HIV-1 infected individuals. Antimicrob Agents Chemother. 2016. doi: 10.1128/AAC.00899-16.
[8] Hoare RL, Germovsek E, Cortina-Borja M, et al. A simple mechanistic pharmacometric model for HIV in children and adults. In: Abstracts of the World Conference on Pharmacometrics. Brisbane, Australia, 2016. Abstract 134.
Reference: PAGE 27 (2018) Abstr 8500 [www.page-meeting.org/?abstract=8500]
Poster: Drug/Disease Modelling - Infection