Background: The polymorphic CYP2D6 enzyme metabolises the antipsychotic drugs risperidone and aripiprazole to their active metabolites, 9OH-risperidone and dehydroaripiprazole. The aim of this study was to quantify the effect of CYP2D6 genetic variability on risperidone and aripiprazole exposure and treatment in a large patient population.
Methods: We retrospectively obtained patient data from a routine therapeutic drug monitoring database at the Center for Psychopharmacology, Diakonhjemmet Hospital, Oslo, Norway, between Jan 1, 2005, and Oct 15, 2018. Individuals included in our analyses were CYP2D6-genotyped patients treated with risperidone or aripiprazole. Inclusion criteria for measurement of pharmacokinetic parameters (drug and metabolite serum concentrations) were oral administration of risperidone or aripiprazole, information known about prescribed daily dose and comedications, and aged older than 18 years. Exclusion criteria included polypharmacy with drugs known to be CYP2D6 inhibitors or CYP3A4 inducers or inhibitors. Treatment failure was analysed in all patients treated with risperidone or aripiprazole without these criteria. The first endpoint in our analysis was the metabolism of risperidone to 9OH-risperidone and aripiprazole to dehydroaripiprazole, estimated by the log-transformed ratio between the concentrations of metabolite and parent drug (ie, the metabolic ratio for risperidone [9OH-risperidone]/[risperidone] and the metabolic ratio for aripiprazole [dehydroaripiprazole]/[aripiprazole]). Endpoint two was measurement of drug exposure, quantified by the dose-normalised sum of parent drug and active metabolite serum concentrations (ie, active moiety). The third endpoint of treatment failure was measured as the number of patients switched from risperidone or aripiprazole to another antipsychotic drug within 1 year after the last therapeutic drug monitoring analysis of risperidone or aripiprazole. Patient subgroups were defined by CYP2D6 genotype-determined metaboliser status: poor metabolisers, intermediate metabolisers, normal metabolisers, and ultrarapid metabolisers. ANOVA was used to assess the differences in metabolic ratios, active moieties, and daily doses between individual metaboliser categories, and risperidone and aripiprazole therapeutic failures were compared by logistic regression using the normal metaboliser subgroup as a reference.
Findings: 1288 risperidone-treated patients and 1334 aripiprazole-treated patients were included in the study, of whom 725 (56%) risperidone-treated and 890 (67%) aripiprazole-treated patients were eligible for the pharmacokinetic analyses. CYP2D6 genotype significantly changed risperidone and aripiprazole metabolism resulting in an approximately 1·6-times and 1·4-times increase in risperidone and aripiprazole active moiety exposure in poor and intermediate metabolisers compared with normal metabolisers, respectively (odds ratios [OR] for the risperidone dose-normalised active moiety concentration 1·568, 95% CI 1·401-1·736, and 1·373, 1·213-1·532; and for the aripiprazole dose-normalised active moiety concentration 1·585, 1·447-1·724, and 1·476, 1·263-1·688, respectively; p<0·0001 for all). Compared with doses for normal metabolisers, clinicians reduced daily doses of risperidone and aripiprazole administered to poor metabolisers by 19% (95% CI 5-35, p=0·010) and 15% (95% CI 1-28, p=0·033) respectively. The incidence of switching from risperidone to another antipsychotic was increased in ultrarapid metabolisers (OR 2·934, 95% CI 1·437-5·989, p=0·003) and poor metabolisers (1·874, 1·128-3·112, p=0·015); by contrast, the incidence of switching from aripiprazole to another antipsychotic was not significantly related to CYP2D6 metaboliser status.
Interpretation: CYP2D6 genotype had a substantial clinical effect on risperidone and aripiprazole exposure and on the therapeutic failure of risperidone. Pre-emptive CYP2D6 genotyping would be valuable for individualising risperidone and aripiprazole dosing and treatment optimisation.
Funding: H2020 program U-PGx, The Swedish Research Council, the Swedish Brain foundation, and the South-Eastern Norway Regional Health Authority.
Copyright © 2019 Elsevier Ltd. All rights reserved.