Towards a comprehensive PK/PD model of infliximab in inflammatory bowel diseases, with support of prior knowledge
Ana-Marija Grisic (1,2), Alexander Eser (3), Wilhelm Huisinga (4), Walter Reinisch (3), Charlotte Kloft (1)
1. Dept. of Clinical Pharmacy and Biochemistry, Institute of Pharmacy, Freie Universitaet Berlin, Germany; 2. Graduate Research Training Program PharMetrX, Germany; 3. Dept. for Gastroenterology and Hepatology, Medical University of Vienna, Austria; 4. Institute of Mathematics, Universitaet Potsdam, Germany
Objectives:
Introduction of the anti-tumour necrosis factor α monoclonal antibody (mAb) infliximab (IFX) brought about a revolution in the treatment of inflammatory bowel diseases (IBD), offering an efficacious therapeutic option for patients unresponsive to conventional treatment. However, a high variability in response rate has been reported, with up to 60% of the IBD patients losing response to IFX over time. Previously, the loss of response has been related to low IFX plasma concentrations [1]. To our knowledge, IFX concentration-effect relationship in IBD has not been comprehensively described by PKPD modelling, despite its potential to empower therapeutic drug monitoring and improve therapy success.
To gain more insight into the mechanisms underlying the loss of response and set the basis for improved therapy strategies we employed PKPD modelling approaches to analyse the dose-concentration-effect relationship.
Methods:
Clinical PK (IFX plasma concentrations) and PD (biomarker concentrations of C-reactive protein; CRP) data from IBD patients (npatients = 121) was gathered as a part of an investigator initiated trial at the outpatient clinic of the Medical University of Vienna. Patients were treated with 2-h IFX infusion of absolute doses between 100 and 1300 mg. The samples (nPK observations=388, nPD observations=339) were collected at mid-term between two maintenance infusions and at end of a dosing interval (0.6-12.4 weeks after last dose). For the data analysis R, NONMEM (7.3.0), PsN and Pirana were used. As a first step, a PK model was developed and the impact of covariates on interindividual variability (IIV) of CL investigated (considering statistical and clinical significance of available covariates). To compensate for the data sparseness, a previously reported PK model [2] was utilised, by informing PK parameter estimation using the $PRIOR functionality of NONMEM. For the covariate analysis, parameters informed by the prior model were fixed to the final estimates of the base PK model. In the next step, a sequential PKPD analysis was performed, by using the empirical Bayes estimates of the PK model. Simultaneous parameter estimation was performed as an evaluative step on the final base PKPD model. Performance of the model was thoroughly evaluated at each step of the model development process.
Results:
A 2-compartment disposition model with linear elimination (CL=0.266 L/d) resulted in adequate description of the PK data. Random effect parameters comprised of mixed proportional-additive residual unexplained variability (RUV) and IIV in V1, V2, CL and RUV. As only IIV in CL was estimable solely from the data, covariate effects on CL were investigated. Inclusion of anti-IFX antibody status, disease activity (serum albumin concentration), body weight and concomitant therapy with immunomodulators explained ~30% of IIV in CL and reduced IIV in RUV. The model exhibited a good performance as judged by standard diagnostics (incl. simulation-based methods). As extension to incorporate the PD component and in accordance with the mechanism of action of IFX, indirect response Emax model with inhibition of CRP synthesis process adequately described the biomarker data. Baseline CRP (0.63 mg/dL), Imax (0.72), IC50 (2.04 mg/L) and proportional RUV were estimated from the data alone, the CRP degradation rate constant was fixed to correspond to known CRP half-life (19h) and estimation of IIV in baseline CRP and IC50 was supported by a prior model [3]. Values of all parameters were in plausible ranges and the model demonstrated a reliable predictive performance. Estimated IIV in baseline CRP and IC50 was very high (115 and 209 % CV, respectively). Even though ability of investigated covariates to explain IIV was limited, the covariate analysis identified potentially significant covariates to be: a history of prior surgeries, age at diagnosis and smoking status.
Conclusions:
Within the study, a well-performing population PKPD model of IFX in IBD was successfully developed. Moreover, influential covariates on PK and PD parameters, and thus subpopulations at risk (presence of anti-IFX antibodies, higher disease activity, higher body weight and absence of co-therapy with immunomodulators) were identified. The developed model will be further extended to account for other biomarkers and subsequently be used to investigate potential therapy strategies to increase therapy success (i.e. ameliorate remission rate).
References:
[1] A. Hemperly, N. Vande Casteele. Clinical pharmacokinetics and pharmacodynamics of infliximab in the treatment of inflammatory bowel disease. Clin. Pharmacokinet. (2018).
[2] A.A. Fasanmade et al. Pharmacokinetic properties of infliximab in children and adults with Crohn’s disease: A retrospective analysis of data 2 phase III clinical trials. Clin. Ther. 33: 946-964 (2011).
[3] D. Ternant et al. Pharmacokinetics and concentration-effect relationship of adalimumab in rheumatoid arthritis. Br. J. Clin. Pharmacol. 79: 286-297 (2014).
