Population PK/PD model of the sedative effects of Flibanserin in healthy volunteers
Iñaki F. Trocóniz1, Katja Boland2, Alexander Staab2
1 Department of Pharmacy and Pharmaceutical Technology, School of Pharmacy, University of Navarra, Pamplona, Spain 2 Drug Metabolism and Pharmacokinetics, Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany
Background and objectives: Flibanserin (BIMT 17 BS) is being developed for the treatment of the hypoactive sexual desire disorder in females. Sedative effects have been found to be the main problematic adverse events after flibanserin administration.
The objective of the current evaluation was to establish a population PK/PD model for the sedative effects of flibanserin administered orally as an immediate release tablet to healthy volunteers.
Methods: Data were obtained from 24 healthy volunteers (14 males and 10 females) receiving flibanserin as a single oral dose of 100 mg.
Each subject was studied during two study days. During the first day the sedative effects were measured at each measuring time point using the Visual Analogue Scale (VAS) for "drowsiness" (0 cm="not in existence" to 10cm ="very strong"). In the second day the VAS response was also recorded, flibanserine was administered, and blood samples for pharmacokinetic were taken during 48 hours.
To better characterize the disposition of flibanserin in plasma, pharmacokinetic data from a clinical study where flibanserin was administered to 12 healthy male volunteers by a 30 min intravenous infusion were also included.
Three steps were followed during the population analysis performed with NONMEM VI: (i) step 1, modeling the VAS vs time profile during study day 1, (ii) step 2, pharmacokinetic analysis, and (iii) step 3, PK/PD model of the VAS effects incorporating the models developed in steps 1 and 2, plus the flibanserin effect model.
Results and conclusions:
Step 1. The time course of VAS data during the first study day (VASBaseline) was modelled empirically using linear splines with three breakpoints located at equally spaced clock times (6.65, 10.77, and 21.1 h).
Step 2. Disposition of flibanserin was best described with a three compartment model. Absorption characteristics of flibanserin were best described with a transit compartment model. Absolute bioavailability was 52%.
Step 3. Drug effects were incorporated in the model as it is shown in equation 1, where ckt refers to clock time, CP, represents the predicted drug plasma concentration, C50, is the value of CP at half of maximum VAS response [10-VASBaseline(ckt)], and n the parameter governing the steepness of the VAS vs CP curve.
VASckt = VASbaseline, ckt + (10-VASbaseline, ckt) x [(CnP,ckt) /( CnP,ckt + Cn50)] equation 1
The VAS vs CP relationship resulted to be very steep, supporting the concept of a threshold concentration (~200 ng/mL) below which plasma concentration has hardly any impact on the VAS scale. Model-based simulations showed that 20% of the subject population would show a VAS change from baseline ≥ 2 cm for a level of flibanserin in plasma of 225 ng/mL.