Ba-Hai LE1,2, Nadège Néant1, Benoit Blanchet 3, François Goldwasser3, Joseph Ciccolini1, Florence Gattacceca1
1: Aix Marseille Univ, INSERM, CNRS, CRCM SMARTc, F-13005 Marseille, France; 2: Hanoi University of Pharmacy, Ha Noi, Viet Nam; 3: CERIA, Paris, France
Introduction: High inter- (61-65%) and intra-individual (44-47%) variability in pharmacokinetics (PK) of sorafenib was observed during phase I studies, as well as in some population pharmacokinetic (pop-PK) studies, which could explain the unstable response and the unintended toxicities that occur in some patients under a recommended dosage of sorafenib [ 1-4]. Therefore, we aimed to develop a sorafenib pop-PK model based on a large population treated by sorafenib during a long period. This pop-PK model will be used to predict sorafenib plasma PK based on sparse therapeutic drug monitoring using Bayesian approach, then to explore the relationship between sorafenib plasma exposure and toxicity outcome in patients from La Timone hospital.
Method: Sparse PK data available from 267 patients treated with sorafenib between 2008 and 2018 in have been included in this multicentric study (10 French hospitals). The daily dosing of sorafenib doses ranged from 200 to 5200 mg in a b.i.d regimen. The PK data were analyzed using nonlinear mixed-effect modeling (NONMEM software version 7.3). Model evaluation was performed using standard goodness-of-fit plots, and simulation-based tools such as visual predictive check (VPC).
The final pop-PK model was applied to provide explanations for the onset of severe sorafenib-related toxicities in one patient, then to evaluate the rationale of the associated empirical dose reductions in this patient. The individual parameters were estimated using Bayesian method in NONMEM and used to simulate the exposure to sorafenib over the course of the treatment in R studio.
Results: All 1310 plasma concentrations came from the clinical routine practice and were collected at steady-state. The follow-up of patients lasted between 15 and 1997 days (5.5 years). A 1-compartment structural model with first-order absorption and linear elimination described the data satisfactorily. Bioavailability (F1) was found to vary as a function of the dose and to decline over time on sorafenib treatment, as described by the following equation: F1=(Fmax-(Fmax*DOSE^n)/(D50^n+DOSE^n))*exp(-lambda*TIME/720). Typical values (RSE %) of the final model parameters were as follows: clearance (CL) 4.91 L/h (4%), distribution volume (V) 177 L (12%), absorption rate constant (ka) 0.727 h-1 (28%). The high inter-patient variability was confirmed in this study with 46.5 % for CL (46.5%), V (73.2%) and ka (150%). The lack of information regarding the absorption phase and covariates such as food intake may account for the extremely large variability of ka.
Based on the final pop-PK model, the individual parameters of the patient were estimated and used to describe the PK of sorafenib over the course of the treatment. The steady-state AUC from 0 to 12h (AUC0-12) post dose was calculated: values of 99.85, 82.83 and 57.70 mg*h/L were obtained with the dosages 800mg/24h, 400mg/24h and 200mg/24h respectively. These values are particularly high when compared to the mean AUC0-12 of 57.7±28.6 mg* h/L obtained in patients experiencing grade 3–4 adverse events in a previous study [4] and are consequently consistent with the observed severe toxicities. These results document the decision to continue reducing the sorafenib dose by expanding the time between doses with the 200mg dosage.
Conclusion and perspectives: The high unexplained inter-individual variability in our population could be partly explained by covariates that will need to be collected in further prospective studies. However, the established population model allows reliable prediction of pharmacokinetics based on Bayes estimation. The population model could be used in the context of therapeutic drug monitoring to support dose adjustments in patients.
References:
- Hornecker, M., et al., Saturable absorption of sorafenib in patients with solid tumors: a population model. Invest New Drugs, 2012. 30(5): p. 1991-2000.
- Jain, L., et al., Population pharmacokinetic analysis of sorafenib in patients with solid tumours. Br J Clin Pharmacol, 2011. 72(2): p. 294-305.
- Awada, A., et al., Phase I safety and pharmacokinetics of BAY 43-9006 administered for 21 days on/7 days off in patients with advanced, refractory solid tumours. Br J Cancer, 2005. 92(10): p. 1855-61.
- Boudou-Rouquette, P., et al., Variability of sorafenib toxicity and exposure over time: a pharmacokinetic/pharmacodynamic analysis. Oncologist, 2012. 17(9): p. 1204-12.
Reference: PAGE 28 (2019) Abstr 9139 [www.page-meeting.org/?abstract=9139]
Poster: Drug/Disease Modelling - Oncology