Helena Edlund (1,2,3), Casper Steenholdt (3), Mark A Ainsworth (3), Jørn Brynskov (3), Ole Ø Thomsen (3), Wilhelm Huisinga (4), Charlotte Kloft (1)
1. Dept. of Clinical Pharmacy and Biochemistry, Institute of Pharmacy, Freie Universitaet Berlin, Germany, 2. and Graduate Research Training Program PharMetrX, Germany, 3. Dept. of Gastroenterology, Herlev Hospital, Denmark, 4. Institute of Mathematics, Universitaet Potsdam, Germany
Objectives: A substantial proportion of patients with Crohn’s disease (CD) lose response to the monoclonal antibody infliximab (IFX) [1]. The pharmacokinetic/pharmacodynamic (PK/PD) relationship of IFX in CD has primarily been assessed using minimal IFX concentrations (Cmin), limiting the predictive performance [Edlund et al. submitted]. Currently, two population PK models for IFX in CD are available implementing different, non-overlapping covariates [2,3]. This study aimed to explore the PK/PD relationship, starting by assessing covariates affecting IFX clearance (CL) exploiting the prior knowledge in [2,3].
Methods: PK/PD data originated from a 20-week investigator-initiated clinical study including 69 CD patients with treatment failure to IFX maintenance therapy [4]. As dose intervention the patients received either (i) IFX with shortened dosing interval, or (ii) were treated according to an algorithm based on IFX and anti-drug antibody (ADA) concentrations. IFX and ADA samples were drawn at weeks 0, 12 and 20 (all Cmin). Crohn’s disease activity index, C-reactive protein and patient demographics were measured at weeks 0, 4, 8, 12 and 20. The two available PK models were re-parameterised and compared with respect to predictive performance, plausibility (e.g. structural model and covariate implementation). Based on structure and estimated parameter values of the priormodel, further covariate analysis were performed on CL using NONMEM 7.3.
Results: The available PK models showed good agreement of estimated parameters values of volumes of distribution and CL, both when compared internally as well as with PK traits of the monoclonal antibodies drug class [5]. The resulting prior-model was a 2-compartment model with body weight implemented according to allometric principles on volumes of distribution and CL. Due to the sparse data situation the population estimates of V1, V2 and Q as well as their covariate relationships to body weight were fixed. The estimate of CL and the IIV parameters were implemented using the PRIOR functionality. The covariate analysis confirmed characteristics earlier identified to predict differences in CL.
Conclusions: Based on data from an investigator-initiated clinical study and two published PK models we confirmed covariates identified to affect IFX CL. The resulting PK model will be used to investigate the PK/PD relationship of IFX in CD, aiming to identify the effective therapeutic range.
References:
[1] Gisbert, Panes. Am J Gastroenterol 104:760 (2009)
[2] Fasanmade et al. Clin Ther. 33:946 (2011)
[3] Ternant et al. Clin Pharmacokinet; online first Jan 17 (2014)
[4] Steenholdt et al. Gut 63:919 (2014)
[5] Dirks et al. Clin Pharmacokinet 49:633 (2010)
Reference: PAGE 24 (2015) Abstr 3380 [www.page-meeting.org/?abstract=3380]
Poster: Drug/Disease modeling - Other topics