II-52 Gauri Rao

Characterization of Oseltamivir Carboxylate (OC) Disposition using a Reduced Population (POP) Pharmacokinetic (PK) Model

Gauri G. Rao (1) Neang S. Ly (1) Mohamed A. Kamal (2) Patrick Smith (1, 3) Alan Forrest (1)

(1) University at Buffalo, Buffalo, New York, USA (2) Hoffmann La-Roche, Inc., New York, New York, USA, (3) D3 Medicine, Parsippany, New Jersey, USA

Objectives: Oseltamivir (Tamiflu) is a neuraminidase inhibitor, administered as a phosphate prodrug (OP).  The non-toxic and inactive pro-drug, OP is almost completely converted to the active metabolite OC. Our aim was to develop a reduced PK model describing the disposition of OC only, from pediatrics through geriatrics.

Methods: OC plasma concentrations (conc), dosing history and demographic information was pooled from 13 clinical trials for 388 healthy and infected subjects; ages from 1 – 78 yrs and OP doses from 20 – 1,000 mg.  Nonlinear mixed-effects modeling was conducted in SADAPT facilitated by SADAPT-TRANS. PK parameters were assumed to be log-normally distributed. Data below the limit of quantification was handled using Beal M3. Model discrimination was done on basis of the objective function and examination of goodness-of-fit plots. 

Results: The reduced PK model characterized the OC disposition reasonably well (R2 of 0.970 for individual; 0.799 for POP fitted) compared with a full PK model [1] for OP and OC (R2 of 0.969 for individual; 0.741 for POP fitted). The model was described by a two-compartment (CMT) model with linear clearance. Two transit CMTs accounted for the delay in conversion of OP to OC (transit rate constant, Kt), and OC was absorbed via a first order process (Ka). Residual variability was modeled using the additive plus proportional variance model. Weight, modeled allometrically with a fixed power function (pWt) and CLcr modeled with a fitted power function (pCLcr) , were significant covariates for the apparent OC clearance (CLm/F).  The geometric mean (%SE) parameter estimates for OC were Vc/F 19.4L (8.88), Vp/F 190L (4.47), CL/F 15.3L/h (2.31), Kt 0.631h-1(2.54), Ka 1.68 h-1 (8.01), CLd 53.5 (7.04), pWT 0.462(6.64).

Conclusions: The final model described OC PK well and performed as well as the full PK model. This simple model will be useful for PK/PD analysis allowing for better PK/PD parameter estimation with sparser data in future studies.

References:
[1] Kamal et al. Antimicrob. Agents Chemother. 2013, 57(8):3470

Reference: PAGE 23 () Abstr 3048 [www.page-meeting.org/?abstract=3048]

Poster: Drug/Disease modeling - Infection