Efficacy modelling of NS2330 in mild Alzheimerís disease patients
Lehr, T. (1), A. Staab (2), C. Tillmann (2), D. Trommeshauser (2), A. Raschig (2), H.G. Schaefer (2), C. Kloft (1,3)
(1) Dept. Clinical Pharmacy, Institute of Pharmacy, Freie Universitaet Berlin, Berlin, Germany (2) Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach a.d.R., Germany (3) Martin-Luther-Universitaet Halle-Wittenberg, Halle, Germany
Objectives & Background: NS2330 is a new drug under clinical development for the treatment of Alzheimerís disease (AD). A population pharmacokinetic (PK) model of NS2330 and its active metabolite M1 has recently been presented  based on 4 phase I studies in normal cognitive and 2 phase IIa studies in mildly impaired cognitive subjects. The PK in these two populations was not different. The objective of the current analysis was to expand the population PK model by a pharmacodynamic (PD) model using the ADAS-COG measurements of the phase IIa studies. The analysis should also include a screening for covariates that might influence the PD characteristics of NS2330 and/or its metabolite M1 in AD patients.
Methods: Modelling was performed in a sequential manner using the FOCE INTERACTION estimation method implemented in NONMEM. PK parameters were fixed to values obtained by the recent analysis  and relative potency of parent compound and metabolite was fixed to a value obtained in mice . The ADAS-COG was measured at baseline, day 1, day 28 (last day on treatment) and day 42. Forty-four patients with 176 ADAS-COG measurements and a median baseline value of 8.7 were used for PD modelling.
Results: The ADAS-COG measurements were successfully described by an extended Emax model accounting for competitive interaction between parent compound and metabolite. Effect compartments were used to link PK and PD. NS2330 and M1 achieved a maximum reduction of the ADAS-COG baseline of 39% in this patient population. ADAS-COG baseline values and age significantly influenced the efficacy in mild AD patients. A higher baseline value increased the absolute efficacy of NS2330. With increasing age the maximum effect achievable (Emax) was reduced by 1.2% per year in the population investigated. A disease progression and an elaborated placebo model could not be developed and implemented in the model.
Conclusion: NS2330 achieved good efficacy in mild AD patients. Important covariates for the therapeutic use were identified which might allow a systematic patient selection for the treatment with NS2330 in AD. Additionally, the model might help to explain non-responder sub-groups in upcoming trials. The PK/PD model should be further evaluated using data from an upcoming proof of concept trial in mild to moderate AD patients.
 Lehr. et al., PAGE 14 (2005) Abstr 729,  Lehr. et al., PAGE 14 (2005) Abstr 730