BACKGROUND: Interpatient variability in cytochrome P450 2D6 (CYP2D6) enzyme activity alters the serum concentrations of most psychotropics, which often have narrow therapeutic indices. Therefore, pre-emptive knowledge of CYP2D6 activity is desired. However, accessible indicators for deficient CYP2D6 activity are necessary, as genotyping all patients prescribed CYP2D6 metabolized drugs is often not feasible or cost-effective.
METHODS: In this study, the predictive value of the ratio between a CYP2D6 substrate and its metabolite, known as the metabolic ratio (MR), the dose-corrected serum concentration of substrate (CDR), and the dose-corrected sum concentration of substrate and metabolite (Sum CDR) of venlafaxine, risperidone, aripiprazole, and nortriptyline were determined to predict the CYP2D6 poor metabolizer (PM) phenotype. The area under the receiver operator characteristic curve, as well as the sensitivity, specificity, and positive and negative predictive values of the optimal thresholds, were calculated.
RESULTS: Although the MR, CDR, and Sum CDR all predicted the CYP2D6 PM phenotype, the predictive value of the MR was most robust for venlafaxine and aripiprazole, and the Sum CDR was inferior for all three psychotropics. MRs of venlafaxine, risperidone, and aripiprazole, and CDR of nortriptyline showed an area under the receiver operator characteristics (95% confidence interval) of 97.2% (94.7.-99.6%), 93.0% (88.8.-97.2%), 97.8% (95.4.-100.0%), and 85.6% (78.0.-93.1%), respectively. Thresholds of the log(MR) of ≥0.1 for venlafaxine, ≥0.0 for risperidone, and ≥1.5 for aripiprazole, and log(CDR) ≥0.5 for nortriptyline produced >92% sensitivity and >64% specificity.
CONCLUSIONS: If therapeutic drug monitoring is available, the thresholds presented here could serve as a diagnostic tool for the CYP2D6 PM phenotype of psychiatric patients prescribed the aforementioned psychotropic medications.