E. Dahmane (1) (2), K. Zaman (3), L. Decosterd (2) (4), C.B. Eap (1) (5), L. Perey (6), A. Bodmer (7), S. Leyvraz (3), M. Galmiche (3), T. Buclin (2), M. Guidi (1) (2), C. Csajka (1) (2)
(1) School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, Geneva, Switzerland; (2) Division of Clinical Pharmacology, Department of Laboratories, University Hospital Lausanne; Lausanne, Switzerland; (3) Breast Center, University Hospital Lausanne; Lausanne, Switzerland; (4) Innovation and Development Unit, Department of Laboratories, University Hospital Lausanne; Lausanne, Switzerland; (5) Center for Psychiatric Neurosciences, University Hospital Lausanne; Prilly, Switzerland; (6) Medical Oncology, Hospital of Morges; (7) Breast Center, University Hospital Geneva, Geneva, Switzerland.
Objectives: Tamoxifen (Tam) is bioactivated into active metabolites, 4-hydroxy-Tam (4OHTam) and mainly endoxifen that involves several cytochromes (CYP). Tam and metabolites present highly variable concentrations and patients with low endoxifen levels (< 6 ng/ml) are likely to achieve less benefit from their treatment. CYP2D6 polymorphism is assumed to account for active metabolites exposure variability but other CYP or co-medications may contribute as well. The aim of this analysis was to characterize the population pharmacokinetics of Tam and 3 of its metabolites and to explore the influence of genetic and non-genetic factors on their exposure.
Methods: Patients were genotyped and/or phenotyped for CYP2D6, 3A4, 2C9, 2C19 and 2B6. Plasma levels of Tam, N-desmethyl-Tam (NDTam), 4OHTam and endoxifen were measured at baseline and at 30, 90 and 120 days. A stepwise procedure with sequential addition of metabolites was used to find the best model (NONMEM®). For identifiability issues, the volume of distribution of Tam and metabolites were assumed to be equal.
Results: 507 samples were collected from 107 patients. A 4-compartment model with 1st order absorption, elimination and linear conversion to 3 metabolites described our data. Inter-individual variability was identified on Tam apparent clearance (CL/FTam: CV 30%), Tam to NDTam (k23: 17%), Tam to 4OHTam (k24: 29%) and NDTam to Endoxifen (k35: 83%) metabolic constant rates. CYP2D6 genotype and CYP2D6 inhibitors explained 45% of k35 variability. Endoxifen formation rate was reduced by 95% and 61% in CYP2D6 poor (PM) and intermediate (IM) metabolizers and by 89% and 43% in patients with potent and moderate CYP2D6 inhibitors. Impaired CYP2D6 and 2C19 activity decreased 4OHTam formation (k24) by 24 and 17%, and CYP2C9 PM and IM by 30 and 13%, respectively. Proton-pump inhibitors and fluoxetine reduced k24 by 15 and 60%. These factors explained 33% of the variability on k24. Increasing Age, CYP3A4 activity (midazolam metabolic ratio) and reported low adherence were respectively associated with a 1% decrease, 3% and 10% increase in CL/FTam and explained 16% of the variability.
Conclusion: The conversion of Tam into its active metabolites is highly variable. Pharmacogenetics and co-medication explain a significant, although not major part of this variability. Our results suggest that endoxifen therapeutic drug monitoring might be better suited for treatment optimization than the mere determination of CYP2D6 activity.
Reference: PAGE 23 () Abstr 3157 [www.page-meeting.org/?abstract=3157]
Poster: Drug/Disease modeling - Oncology