RESULTS
516 CYP2D6 -genotyped patients on vortioxetine treatment were identified in the TDM database. Among these, 29 were excluded due to concomitant use of CYP inhibitors/inducers, 8 due to serum concentration measurements of vortioxetine being below the lower limit of quantification, and 21 patients lacking information about the prescribed vortioxetine dose. Thus, a total of 458 patients were included in the analysis.
The frequencies of CYP2D6 genotype-predicted PMs, IMs, NMs and UMs in the population were 7.6%, 37.8%, 52.8% and 1.7%, respectively, and all the CYP2D6 variant alleles were in Hardy-Weinberg equilibrium. There were no significant differences in patient demographics or time intervals between the last vortioxetine dose and TDM blood sampling between the CYP2D6 phenotype groups (seeTable 1 ). The median vortioxetine dose administered in the PM and IM groups were lower than those in the NM and UM groups, but the differences were not statistically significant.
The median vortioxetine exposure, measured by dose-harmonized concentration, was highest among the CYP2D6 PMs (23.9 ng/mL), followed by the IMs (12.5 ng/mL), NMs (8.1 ng/mL), and lowest for the UMs (5.9 ng/mL) (see Figure 1 ). The CYP2D6 PMs and IMs exhibited significantly higher vortioxetine exposures (P<0.001) compared to NMs with ratios of medians being 3.0 and 1.5, respectively. No significant difference in vortioxetine exposure was found between the CYP2D6 UMs and NMs (P=0.21).
In addition to the exposure differences, the frequency of patients switching from vortioxetine to an alternative antidepressant during the course of three-month follow-up was significantly higher among PMs compared to NMs (P=0.001, odds ratio (OR) 8.0, 95% CI=2.0-32.2). CYP2D6 UMs also showed a significantly higher frequency of treatment switch compared to NMs (P=0.02, OR 12.7, 95% CI=1.1-94.9), while no significant difference was found between the IMs and NMs (P=0.28, OR 1.9, 95% CI=0.6-6.8).