Vincent CROIXMARIE, Adrien TESSIER and Marylore CHENEL
Pharmacometrics and Clinical Pharmacokinetics division, Servier, France
Objectives: A phase 2 dose-finding study is planned for drug S in cardiovascular area. The targeted population experiences an acute renal impairment originating from pathology. As this drug candidate is mainly excreted by kidney a dose adaptation strategy is proposed. A population PK (popPK) model and clinical trial simulation (CTS) strategy were used to (i) support selection of dose levels to test in phase 2; (ii) propose a dose adaptation to reduce the inter-individual variability (IIV) in exposure within a dose level.
Methods: A previous popPK model developed on data in healthy volunteers after intravenous (IV) administration at various doses was used. The model is a 3-compartment model with IIV estimated on V1, V2, CL and Q2 and a combined proportional and additive error model. Clearance was expressed as fraction of the passive renal clearance and was equal to: CL = fr*fu*CRCL, with fr the estimated fraction, fu the drug unbound fraction in plasma and CRCL the creatinine clearance. Two covariates – weight on CL and dose on Q2 – were also included in the model.
In phase 2 the drug is intended to be administered through IV with a bolus for 20% of the total amount followed by 6-hours infusion for the remaining 80%. In the targeted population approximately 60 % of the participants are expected to have a mild (estimated glomerular filtration rate (eGFR) of [60;90[ mL/min) acute renal impairment (RI) – respectively 40% a moderate RI (eGFR of [30;60[ mL/min). A CTS approach using the phase 1 popPK model combined with the expected characteristics of the targeted phase 2 population was used. The endpoint for dose selection was based on the drug concentration at 10 hours (C10h) after drug intake in mild RI subjects (reference population), that has to be maintained in at least 90% of the reference population above a given threshold. Then a dose adaptation was proposed in order to keep exposure of moderate RI patients close to the reference population one. Three scenarios were evaluated and compared: no adaptation, adaptation for moderate RI patients as a single category (2 intervals total) or adaptation with 3 intervals (moderate RI interval being split in two: eGFR [30;45[ and [45;60[ mL/min).
The main hypotheses were (i) an absence of difference between healthy volunteers and patients when considering significant covariates (i.e. CRCL and weight); (ii) that the eGFR extrapolation from healthy volunteers for mild and moderate renal impaired function is valid.
Results: The lowest and the highest dose levels to test in phase 2 were defined through CTS in the reference population based on two different thresholds for the C10h endpoint. Two intermediate dose levels were also selected after checking that overlap in exposures across the different dose levels were minimal. For dose adaptation, the scenario with only 2 intervals did not corrected C10h distribution for the [30;45[ mL/min eGFR subgroup to the reference population level. Consequently the 3 intervals dose adjustment strategy has been proposed. The PK profiles overlap of two adjacent dose levels is markedly reduced when applying a 3 intervals adaptation scenario.
Conclusions: PopPK modelling and CTS were used to support dose selection and propose a dose adaptation strategy in patients for a future phase 2 study. The adaptation will be performed in order to limit IIV in drug exposure within a dose level and will further help phase 2 study outcome analysis including dose-response analysis.
Reference: PAGE () Abstr 9402 [www.page-meeting.org/?abstract=9402]
Poster: Drug/Disease Modelling - Other Topics