Venlafaxine belongs to the group serotonin-noradrenaline re-uptake inhibitor (SNRI) antidepressant drugs. It was first marketed in 1995 and is widely used in the management of depression. Venlafaxine is metabolized by CYP2D6 resulting in its active metabolite desmethylvenlafaxine (i.e. desvenlafaxine), which was marketed in 2012 with the aim of reducing venlafaxine’s side effects (Food and Drug Administration, 2011). Desvenlafaxine daily dose ranges between 50 and 100 mg. All the pharmaceutical forms marketed are prolonged-release tablets. After oral administration, desvenlafaxine is highly absorbed, with a bioavailability of 80%. Maximum plasma concentrations (C_max) are observed 7.5 h after oral administration (t_max) (Liebowitz and Tourian, 2010). The area under the time-concentration curve (AUC) in a 24 h dosing interval at steady state (AUC_τ) with the 100 mg dose is 6747 ng*h/mL, and the C_max is 376 ng/mL. The binding of desvenlafaxine to plasma proteins is low (30%) and the mean terminal half-life (t_1/2), is approximately 11 h (Spanish Drug Agency, 2012). It is mostly metabolized by O-glucuronide conjugation by UDP-glucuronyltransferase (UGT) enzymes, including UGT1A1, UGT1A3, UGT2B4, UGT2B15, and UGT2B17. To a lesser extent, it undergoes oxidative metabolism by the cytochrome p450 (CYP) isoform CYP3A4 (Sproule et al., 2008; Preskorn et al., 2009; DrugBank, 2010). Venlafaxine structure contains a tertiary amine and a methoxy group, which can be N- and O-demethylated, respectively. The N-demethylation to N-desmethyl venlafaxine (NDV) is catalyzed by CYP3A4 and CYP2C19, while the O-demethylation to desvenlafaxine is catalyzed by CYP2D6 and CYP2C19 (Fogelman, 1999). When desvenlafaxine is administered instead of venlafaxine, the N-demethylation to N, O-didesmethyl venlafaxine is assumed to be mediated by means of CYP3A4 and CYP2C19 (pathway available at: https://www.pharmgkb.org/pathway/PA166014758).

CYP2D6 polymorphism conditions venlafaxine exposure and response. In fact, the Royal Dutch Pharmacist’s Association (KNMP) Pharmacogenomics Working Group (DPWG) includes the CYP2D6-venlafaxine drug-gene pair in their recommendations list. Briefly, an alternative drug that is not metabolized by CYP2D6 or a dose reduction should be considered for CYP2D6 intermediate metabolizers (IMs) and poor metabolizers (PMs) and a 150% dose increased should be considered for ultrarapid metabolizers (UMs) if necessary (Nichols et al., 2009; Nichols et al., 2013; Dutch Pharmacogenetics Working Group, 2022). In contrast, desvenlafaxine is apparently not affected by CYP2D6 phenotype and scarce pharmacogenetic information is available nowadays for this drug. One study evaluated the impact of ABCB1 rs1045642, ABCC1 rs212090, and UGT1A1 rs8175347 on desvenlafaxine dose requirements (Bousman et al., 2017). However, several additional pharmacogenetic biomarkers could be relevant concerning desvenlafaxine pharmacokinetics and/or dose requirements. Hence, the aim of this study was to evaluate the impact of 29 alleles in 12 candidate genes (i.e., ABCB1, ABCG2, CYP2B6, CYP2C19, CYP2C8, CYP2C9, CYP2D6, CYP3A4, CYP3A5, SLCO1B1, UGT1A1, and UGT2B15) on desvenlafaxine pharmacokinetic variability and tolerability.

A total of 98 healthy volunteers participated in the pharmacogenetic study. Mean age was 31.0 ± 8.9 years old, with evenly distributed proportions of males (n = 52) and females (n = 46) with 26 Caucasians and 72 healthy volunteers with Mixed race (including 71 Latin-Americans and one Black) (Table 2). Females exhibited lower height (p < 0.001) than males and lower weight (p < 0.001). In addition, the body mass index (BMI) of Caucasians was significantly lower than that of Mixed-race healthy volunteers (p = 0.010) (Table 2).

To date, numerous studies compared the efficacy and safety of antidepressants. In 2018, (Cipriani et al., 2018) published a comprehensive systematic review and meta-analysis of 21 antidepressant drugs and their use for the treatment of major depressive disorder. All antidepressants were found effective vs. placebo, with ORs ranging between 1.37 (for reboxetine) and 2.13 (for amitriptyline), with desvenlafaxine and venlafaxine showing intermediate effectiveness (OR = 1.49 and 1.78, respectively). In terms of dropout rate, both drugs showed a similar profile (ORs = 1.08 and 1.04 vs. placebo, respectively). Unfortunately, data comparing desvenlafaxine with other antidepressants is scarcely available, and placebo-controlled clinical trials are mainly available. With placebo as comparator, it seems that both drugs show a similar effectiveness profile, with desvenlafaxine showing a better safety profile (i.e., lower incidence of nausea) (Coleman et al., 2012). In a 8-week phase-III clinical trial, desvenlafaxine showed a 5% discontinuation rate compared to 6% for venlafaxine, and 13% for duloxetine (Tourian et al., 2009), which is consistent with desvenlafaxine having a slightly better safety profile. In summary, little evidence is available to justify the selection of other antidepressants in preference to desvenlafaxine. Otherwise, desvenlafaxine is one more drug in the therapeutic arsenal available to treat depression, and its selection over other antidepressant drugs depends on the clinician’s preference.

When comparing venlafaxine and desvenlafaxine, at equivalent doses, they could be considered similar therapeutic alternatives, as both compounds are potent serotonin and norepinephrine reuptake inhibitors. Venlafaxine, through CYP2D6, is converted to desvenlafaxine, hence, the pharmacological effects of desvenlafaxine can be expected when venlafaxine is administered; the only difference is that venlafaxine (the precursor) also inhibits dopamine reuptake and desvenlafaxine does not (Singh and Saadabadi, 2022). However, a clear drawback of venlafaxine therapy is the variability in response and safety associated with CYP2D6 phenotype. A priori, desvenlafaxine seems to have a better pharmacokinetic profile and a better dose-response and dose-safety relationship (Norman and Olver, 2021). In this work, not only did we confirm that the occurrence of ADRs caused by desvenlafaxine is associated with the time (multiple dose administration vs. single dose administration) and/or magnitude (AUC) of drug exposure, but we also clearly observed that the exposure to desvenlafaxine does not depend on the genetic polymorphism of candidate genes. The strong dose-dependent relationship with the occurrence of ADRs confirms that the mechanism of action of desvenlafaxine is essentially dose-dependent, compared to the biphasic dose-dependent mechanism of action of venlafaxine (Harvey et al., 2000). To our knowledge, this work is the first to analyze all potentially relevant pharmacogenes comprehensively and robustly with respect to the pharmacokinetics and safety of the drug. Although this might be somewhat obvious a priori, this does not hold for all drugs. For instance, voriconazole shows nonlinear pharmacokinetics, with the dose-response (or dose-safety) relationship exhibiting wide interpatient variability. This is typically explained by its limited capacity of elimination. Furthermore, CYP2C19 phenotype is a clear predictor of drug metabolism (Moriyama et al., 2017). Therefore, a reduction or increase in dose is not directly related to an increase in exposure or in the chances of achieving efficacy (Theuretzbacher et al., 2006; Zubiaur et al., 2021). Studies such as the present one are important to confirm this pharmacokinetic characteristic, which enables more predictable dosage adjustments of the drug.

Here, the healthy volunteers presented pharmacokinetic parameters consistent with previous literature. For instance, an AUC_∞ of 5875.01 ng*h/mL was observed for the 100 mg formulation, compared to 6747 ng*h/mL and a Cmax of 251 ng/mL compared to 376 ng/mL (Spanish Drug Agency, 2012) (Spanish Drug Agency, 2012). Furthermore, the global tmax was 6.66 h compared to 7.5 h, and a t1/2 of 10.71 compared to 11 h (Spanish Drug Agency, 2012). Female healthy volunteers presented a higher Cmax/DW, which may be explained by sex-specific physiological differences in drug absorption (Soldin and Mattison, 2009). A clear dose-dependent relationship was observed in the variations of AUC and Cmax. Furthermore, several differences in pharmacokinetic variables were observed according to the clinical trial design. However, these differences were small, with the AUC/DW ranging from 86% to 115% with respect to mean AUC/DW, and the Cmax/DW ranging from 83% to 121%, except for the clinical trial C, which showed a 138% with respect to mean C_max/DW. This difference can be explained by the fact that, in clinical trial C, several desvenlafaxine doses were administered and the residual concentration from previous drug administrations needs to be considered. In contrast, AUC∞ from single dose clinical trials and AUCƬ from the multiple-dose clinical trial were equivalent, as expected. The remaining differences related to the clinical trial design can be considered spurious and related to the low sample size for each clinical trial. Finally, fed healthy volunteers showed a higher C_max/DW, an effect that is well described in desvenlafaxine’s drug label (Spanish Drug Agency, 2012).

As mentioned earlier, the incidence of ADRs was significantly higher in the only clinical trial with multiple-dose administration (clinical trial C). This may be explained by the higher observed Cmax/DW or the prolonged exposure to the drug. Furthermore, the AUC was identified as the best predictor of ADR development; the fact that drug exposure is the main predictor of drug tolerability is reassuring. Assuming linear pharmacokinetics, ADRs resulting from drug overexposure could be resolved with dose adjustments. This does not apply equally to venlafaxine, as certain patients (e.g., CYP2D6 UMs or PMs) will respond variably to similar adjustments. The association between CYP2B6 RMs and asthenia incidence can be considered spurious; should CYP2B6 metabolize desvenlafaxine, a lower rate of ADRs should be expected in individuals with a higher metabolic capacity. Alternatively, this would be explained if desvenlafaxine had an active metabolite responsible for this side effect. However, to our knowledge, this metabolite has not been described to date. Likewise, the association between ABCB1 rs203258 G/T and asthenia can only be explained as a spurious event because a heterozygous disadvantage is not possible for traits inherited in a codominant manner.

Several study limitations should be considered. The clinical trial study design did not allow to collect data on the long-term pharmacokinetics and tolerability or effectiveness. We were also unable to analyze the influence of Black race on desvenlafaxine pharmacokinetics because there was only one subject self-reported as Black. In addition, some Tier 1 variants in CYP2C9 and CYP2D6 were not included (Pratt et al., 2019; Pratt et al., 2021) thus the pharmacogenetic phenotype may not have been correctly inferred, albeit this possibility is very unlikely due to the low prevalence of these alleles (e.g., CYP2C9*8). In addition, among the UGT enzymes, only variants in UGT1A1 have been demonstrated to alter the enzyme’s function, while no variants with a functional impact are described for other important UGT candidates such as UGT1A3 or UGT2B4. Furthermore, these results may be interpreted with caution given the limited sample size, which implies that the statistical power is limited and may not be sufficient to detect pharmacogenetic effects, especially for rare polymorphisms. In addition, some associations are described that did not reach the significance level after Bonferroni correction for multiple comparisons, but these associations must be considered cautiously due to the exploratory nature of this work. In contrast, bioequivalence clinical trials offer a controlled setting for the evaluation of pharmacokinetic variability based on genetic polymorphism or demographics as confounding factors are avoided.

Desvenlafaxine pharmacokinetic parameters were unrelated to genetic polymorphism of the principal candidate pharmacogenes. In contrast, drug exposure in time and magnitude (AUC) were the principal predictors of ADR incidence, which confirms that the mechanism of action of desvenlafaxine is essentially dose-dependent. These two findings can be interpreted as therapeutic advantages of desvenlafaxine compared to venlafaxine.