A Precise Activity System for Pharmacogenetics Based on Estimation of CYP2D6 A llelic Percentage Activity from the International CEPAM Database

The polymorphic enzyme CYP2D6 contributes to the metabolism of many frequently prescribed drugs, causing substantial variability in drug exposure and treatment outcomes. Genotype-based dosing recommendations currently rely on drug-agnostic categorical CYP2D6 allele activity values (AVs), potentially leading to imprecise CYP2D6 phenotype classification and under-/overdosing of numerous medications. To further advance precision dosing, we developed a CYP2D6 percentage activity (CYP2D6-PA) framework by estimating CYP2D6 allelic percentage activity (CYP2D6-APA) from 0% (no activity) to 100% (CYP2D6*1-normalized reference activity) using tamoxifen and its active metabolite Z-endoxifen as a model substrate. By leveraging nonlinear mixed-effects modeling of the international CEPAM database (>7,000 patients from 31 studies), bias-corrected CYP2D6-APAs were estimated, considering global allele distributions, additional covariates, and limited CYP2D6 genotyping. Estimated CYP2D6-APAs spanned 0%-100%, several deviating substantially from their categorical AVs. In particular, CYP2D6*2 and *35 (normal function, AV = 1), and *17 and *29 (decreased function, AV = 0.5) showed lower CYP2D6-APAs of 64-68% and 29%, respectively. Conversely, CYP2D6*9 (decreased function, AV = 0.25) showed a higher CYP2D6-APA (52%). Each additional allele copy exhibited 68% of the activity of the first copy, less than the twofold increase assumed by AVs. Neglecting limited genotyping (alleles misclassified as reference CYP2D6*1), caused CYP2D6-APA overestimation by 1.1-fold. Using CYP2D6-APAs instead of AVs translated to changes in predicted individual Z-endoxifen concentrations (-22.3% to 40.0%). While substrate specificity of empirically derived robust tamoxifen CYP2D6-APAs requires evaluation, CYP2D6-PA provides a framework applicable to other substrates/pharmacogenes. Thus, CYP2D6-PA lays the foundation for revising genotype-to-phenotype translation systems, improving precision of pharmacogenetics-guided dosing and optimizing treatment outcomes upon prospective validation.