Pharmacokinetics of venlafaxine in treatment responders and non-responders: a retrospective analysis of a large naturalistic database

Purpose To assess in a large naturalistic sample, whether clinical response to a treatment with venlafaxine is associated with different patterns of plasma concentrations of active moiety, AM (sum of venlafaxine (VEN) and its active metabolite O-desmethylvenlafaxine (ODVEN)). Methods Applying a regression model, plasma concentrations and plasma concentrations corrected-by-dosage (C/D) for AM were included as independent variable with Clinical Global Impressions-Improvement (CGI-I) scale ratings as dependent variable. Moreover, AM, VEN, and ODVEN were compared between treatment responders and non-responders, defining response as much or very much improved on the CGI-I scale based on the non-parametric Mann-Whitney U (M-W- U ) test with a significance level of 0.05. Results No correlations were found between AM and C/D AM plasma concentrations and CGI-I ratings (regression coefficient 0.0, CI 0.000, 0.001, p = 0.492 for AM and 0.047, CI − 0.065, 0.159, p = 0.408 for C/D AM). Venlafaxine daily dosage did not differ between responders and non-responders (217.7 ± 76.9 vs. 222.0 ± 72.7 mg/day, p = 0.45 for M-W- U ). Responders displayed lower ODVEN ( p = 0.033) and AM ( p = 0.031) plasma concentrations than non-responders ( p = 0.033 and 0.031, respectively for M-W- U ). No other differences were detected. Using a cut-off level of 400 ng/mL for AM concentrations, a higher percentage of responders was reported in the group of patients with AM 400 ng/mL (8%) ( p = 0.038). Conclusions Higher ODVEN and AM concentrations in non-responders than in responders indicate that treatment escalation above upper thresholds of therapeutic reference ranges of venlafaxine is not promising. Hence, the therapeutic reference range for venlafaxine can help in improving outcomes in a measurement-based care model that takes advantage of therapeutic drug monitoring.