Sáez-Belló M1,2,3, Mangas-Sanjuán V3,4, López Montenegro Soria MA5, Climente-Martà M3,6, Merino-Sanjuán M3,4
1 Department of Pharmacy, VITHAS Hospital 9 de Octubre, Spain. 2 Foundation for the Promotion of Health and Biomedical Research of Valencia (FISABIO), Valencia, Spain. 3 Department of Pharmacy Technology and Parasitology, Faculty of Pharmacy, University of Valencia, Valencia, Spain. 4 Interuniversity Institute of Recognition Research Molecular and Technological Development. 5 LluÃs AlcañÃs Hospital, Spain. 6 Department of Pharmacy, University Hospital Doctor Peset of Valencia, Spain.
Objectives: The aim of this study was (i) to assess the pharmacokinetics (PK) of capecitabine and its metabolites 5′-deoxy-5-fluorouridine (5′-DFUR) and 5-fluorouracil (5-FU) in a population of colorectal cancer patients, and (ii) to analyze whether single nucleotide polymorphisms in the ABC transporter gene may explain inter-individual variability of pharmacokinetic parameters.
Methods: A prospective observational post-authorization study between February 2015 and August 2016 was carried out in the Doctor Peset University Hospital of Valencia in patients with colorectal cancer. Capecitabine was administered in different schedules with doses between 850-1250 mg/m2 orally twice a day. Three blood samples were obtained at 1h, 2h and 3h post-administration and the plasma concentrations were determined by high-performance liquid chromatography (HPLC) [1]. Capecitabine and its metabolites plasma levels were described with compartmental models parameterized in first order rate constant, apparent volumes of distribution, and first-order distribution and elimination clearances. Several models to describe the possible delay in absorption were assessed. Model parameters were estimated using Monolix® (Suite-2018R1) [2]. The subject inter-individual variability (IIV) on pharmacokinetic (PK) parameters was modeled exponentially, and residual variability was described proportionally with one residual error for each analyte. Relationship between individual pharmacokinetic parameters (IPK) and covariates such as: polymorphisms of the ABC gene with prevalence higher than 20% in the population, age, sex, body surface, bilirubin, creatinine, creatinine clearance, and concomitant oxaliplatin therapy, were assessed using forward inclusion and backward elimination criteria manually. The predictive performance of the model was evaluated using a visual predictive check (VPC) based on 500 simulated replicates of the development dataset. Parameter precision was evaluated through relative standard error and non-parametric 95% confidence intervals.
Results: 48 patients were included in our study in which 432 plasma samples were collected, 12.7% (55/432) below the limit of quantification (BLQ). Different strategies were followed to account for BLQ (M1, M3, M4 and M6) [3-4] The methodology best described the experimental data in terms of parameter precision and stability was the M1 method, which discards BLQ data and applies extended least squares to the remaining observations. The absorption process was described by a first order process with absorption lag time (0.28 h). The clearance values of capecitabine, 5’-DFUR and 5-FU were 294, 26,8 and 8,97 L/h, respectively. The apparent volume of distribution of capecitabine (V2) was 449 L, while V3 (5’-DFUR) and V4 (5FU) were fixed at 1L. A statistical relationship (p-value) between IPK and 85 covariates (78 polymorphism and 7 biochemical, treatment-related and demographic covariates) was performed. From the total statistically significant covariates (36), highly-correlated covariates were removed for model evaluation (28). Each relationship was incorporated manually in the model. The final model incorporates the following covariates: oxaliplatin with absorption lag time, rs6720173 with clearance of 5’-DFUR and rs2271862 with clearance of 5-FU. A reduction of 14%, 19%, and 66% in the IIV of each parameter was estimated when base and final parameter estimates were compared.
Conclusions: A population pharmacokinetic model which allows to describe concentrations of capecitabine and its metabolites in plasma has been developed, showing the effect of ABC gene single nucleotide polymorphisms rs6720173 and rs2271862 on clearance of metabolites of capecitabine in patients with colorectal cancer.
References:
[1] Zufía L, Aldaz A, Giráldez J. Simple determination of capecitabine and its metabolites by liquid chromatography with ultraviolet detection in a single injection. J Chromatogr B Analyt Technol Biomed Life Sci. 2004;809(1):51-8.
[2] Monolix Suite 2018. Lixoft.
[3] Bergstrand M, Karlsson MO. Handling data below the limit of quantification in mixed effect models. AAPS. 2009. 11(2):371-80.
[4] Ahn JE, Karlsson MO, Dunne A, Ludden TM. Likelihood based approaches to handling data below the quantification limit using NONMEM VI. J Pharmacokinet Pharmacodyn (2008) 35:401–421.
Reference: PAGE 28 (2019) Abstr 8990 [www.page-meeting.org/?abstract=8990]
Poster: Drug/Disease Modelling - Oncology