Advancing paroxetine treatment in depression: predicting remission and plasma concentration, and validating and updating therapeutic reference ranges.

Abstract

Optimizing paroxetine therapy for major depressive disorder (MDD) requires effective prediction models for treatment efficacy and therapeutic drug monitoring (TDM). This study aimed to develop prediction models for treatment remission and steady-state concentration (Css) of paroxetine, elucidate the role of CYP2D6 activity score (AS) in predicting Css, establish associations between adverse drug reactions (ADRs) and Css, and validate and update the therapeutic reference range (TRR) for patients with MDD in the Han Chinese population. We conducted a post-hoc analysis of an 8-week multicenter prospective cohort study involving 530 Han Chinese patients with MDD. Logistic regression models were developed to predict treatment remission at the eighth week and Css as a binary variable (within/outside TRR of 20-65 ng/ml). The model for predicting treatment remission demonstrated an AUC of 0.707, while the model for Css achieved an AUC of 0.615. Associations between ADRs and Css were assessed using logistic regression, adjusted for sex and age. Patients with Css within 20-65 ng/ml were more likely to achieve remission (OR = 1.655, 95% CI: 1.109-2.489) and less likely to experience ADRs (OR = 0.460, 95% CI: 0.203-0.961). Additionally, those with lower AS were more likely to maintain Css within this range (OR = 0.638, 95% CI: 0.461-0.878). ROC analysis further established an updated TRR of 20.8-52.5 ng/ml considering both treatment remission and ADRs. Our findings enhance paroxetine treatment and monitoring, underscoring the potential of CYP2D6 AS and Css as predictors for Css and treatment remission, respectively.