This retrospective naturalistic study (n=178) examined whether benzodiazepines, lithium or lamotrigine altered the antidepressant effects of subcutaneous esketamine in people with treatment-resistant depression. Depression scores fell over six weeks, but benzodiazepine use was linked to higher symptom levels during treatment, while lithium and lamotrigine were not.
Background
Ketamine and esketamine have been increasingly used as adjunctive treatments for treatment-resistant depression (TRD), yet evidence on pharmacological interactions with commonly prescribed psychotropic medications remains limited. This study evaluated the association between concomitant lamotrigine, lithium, and benzodiazepine use and antidepressant outcomes during subcutaneous esketamine treatment in patients experiencing treatment-resistant depressive episodes, including unipolar and bipolar depression.
Methods
We analyzed real-world clinical data from 178 patients treated between 2017 and 2023 at the Esketamine Clinic of the Mood Disorders Program, Universidade Federal de São Paulo. Participants received six weekly subcutaneous esketamine administrations (0.5-1.0 mg/kg), adjusted according to clinical response and tolerability. Changes in Montgomery-Åsberg Depression Rating Scale (MADRS) scores were examined using linear regression and linear mixed-effects models to assess associations between concomitant medication use and outcomes over time.
Results
MADRS scores decreased significantly from baseline to endpoint at week 6 (p < 0.001). Mixed-effects models showed a significant effect of time (β = -2.82 MADRS points per week, p < 0.001), indicating consistent symptom improvement. Concomitant benzodiazepine use was associated with higher MADRS scores across treatment weeks (β = 4.46, 95% CI [0.53-8.39], p = 0.026). Lamotrigine (β = 1.04, p = 0.777) and lithium (β = 1.70, p = 0.343) showed no significant effects, and no medication-by-time interactions were detected.
Conclusions
Subcutaneous esketamine was associated with significant symptom reduction. Benzodiazepine use was associated with higher depressive symptom severity during treatment, whereas lamotrigine and lithium were not associated with differential trajectories. Prospective studies are needed to clarify medication-specific interactions.
Papers cited by this study that are also in Blossom
Fond, G., Loundou, A., Macgregor, A. et al. · Psychopharmacology (2014)
Major depressive disorder remains a major source of disability, and a substantial proportion of patients do not achieve remission after multiple adequate antidepressant trials, leaving a persistent need for faster and more effective treatments. Ketamine and esketamine have attracted attention because they can reduce depressive symptoms rapidly through mechanisms that differ from conventional monoaminergic antidepressants. However, many patients receiving these treatments also take other psychotropic medicines, and it is still unclear whether commonly used agents such as benzodiazepines, lamotrigine, and lithium alter esketamine’s antidepressant effects. Atidio and colleagues aimed to evaluate whether concomitant benzodiazepine, lamotrigine, or lithium use was associated with different antidepressant outcomes during a course of subcutaneous esketamine in patients with treatment-resistant depressive episodes. The study specifically examined whether these medications changed the trajectory of Montgomery-Åsberg Depression Rating Scale (MADRS) scores over time in a real-world clinical cohort including both unipolar and bipolar depression. The paper addresses a practical gap in routine care, where esketamine is often used alongside other psychotropic drugs and where evidence on potential medication interactions is limited, particularly for the subcutaneous route of administration.
This was a retrospective naturalistic observational study using clinical data from the Esketamine Clinic of the Mood Disorders Programme at the Federal University of São Paulo in Brazil. The researchers reviewed records from January 2017 to December 2023. The overall sample included 178 adults aged 18-65 years with either unipolar major depressive disorder or bipolar depression, all meeting criteria for a treatment-resistant depressive episode and having a baseline MADRS score of at least 25. Diagnostic confirmation was made with the Mini International Neuropsychiatric Interview. Patients with primary psychotic disorders, active substance use disorder, significant cognitive impairment, unstable medical illness, concurrent trial participation, incomplete data, or fewer than three esketamine sessions were excluded. Subcutaneous esketamine was given once weekly for six weeks. The first dose was 0.5 mg/kg and the second 0.75 mg/kg, with increases up to 1.0 mg/kg in selected cases according to clinical response and tolerability. Patients were monitored for around two hours after each administration. Concomitant medication exposure was examined in the 162 patients with complete psychotropic medication data, and the medication categories were not mutually exclusive. Benzodiazepine exposure was standardised into diazepam-equivalent daily doses, while lithium and lamotrigine were defined by ongoing use during the treatment period. The paper notes that detailed timing of benzodiazepine use relative to esketamine sessions and serial lithium serum levels were not available in the database. The primary outcome was change in MADRS from baseline to week 6, with weekly MADRS assessments during the induction phase. Secondary outcomes were clinical response, defined as a 50% or greater reduction in MADRS, and remission, defined as MADRS of 10 or less, both assessed at week 6. The researchers used descriptive statistics, Wilcoxon rank-sum tests for between-group comparisons, linear regression to test associations between concomitant medications and change in MADRS, and linear mixed-effects models to examine longitudinal symptom trajectories with week as a fixed effect and participant as a random effect. Residual analyses were used to check model adequacy. All tests were two-tailed with p < 0.05 as the significance threshold.
The cohort included 178 patients, 61% of whom were female, with a mean age of 40.0 years. Most had unipolar depression (68%), and the remainder had bipolar depression (32%). Baseline symptom burden was high, with a mean MADRS score of 32.1, and the mean Maudsley Staging Method score suggested moderate to high treatment resistance. Concomitant psychotropic use was common: benzodiazepines were used by 109 patients (67.3%), lithium by 80 (45.0%), and lamotrigine by 30 (18.5%). Twenty-two patients (13.6%) received esketamine without any of these three medications. Across the whole sample, MADRS scores decreased significantly over the six-week induction period. In the mixed-effects model, week was associated with lower MADRS scores over time (β = -2.82 points per week, 95% CI -3.24 to -2.39, p < 0.001), indicating steady symptom improvement. When medication effects were examined, benzodiazepine use was associated with higher MADRS scores across treatment weeks (β = 4.46, 95% CI 0.53 to 8.39, p = 0.026). Lamotrigine (β = 1.04, 95% CI -6.14 to 8.21, p = 0.777) and lithium (β = 1.70, 95% CI -1.82 to 5.22, p = 0.343) were not significantly associated with MADRS scores. No significant medication-by-time interactions were detected, and the linear model also found no significant main effects or interactions, although there was a non-significant trend towards a negative benzodiazepine-lithium interaction (p = 0.068). At endpoint, 66 patients met criteria for clinical response and 34 met criteria for remission. Among benzodiazepine users, 43 of 101 patients (42.6%) responded and 19 (18.8%) remitted, compared with 23 of 52 (44.2%) and 15 (28.8%) among those not using benzodiazepines. The paper states that similar patterns were seen for lithium and lamotrigine, but no statistically significant subgroup differences were reported. Overall, the primary message was that symptom improvement occurred during subcutaneous esketamine treatment regardless of concomitant lamotrigine or lithium, whereas benzodiazepine use was associated with higher depressive symptom severity during treatment.
Atidio and colleagues interpret the findings as showing that subcutaneous esketamine was associated with meaningful improvement in depressive symptoms during the six-week induction phase in a real-world treatment-resistant population. They state that benzodiazepine co-use was linked to higher MADRS scores during treatment, suggesting attenuated antidepressant benefit, whereas lamotrigine and lithium did not appear to alter the therapeutic trajectory. The authors note that the absence of a benzodiazepine-by-time interaction implies that the rate of improvement over time was not clearly different, but rather that benzodiazepine users had higher overall symptom severity across the treatment period. The discussion places the benzodiazepine finding in the context of earlier work suggesting that benzodiazepines may blunt ketamine’s antidepressant response, while acknowledging that confounding by indication is a major alternative explanation in naturalistic care. The authors point out that benzodiazepines are often prescribed to patients with more anxiety, insomnia, agitation, or overall illness complexity, so their use may mark a more severe clinical phenotype rather than directly causing poorer response. For lithium and lamotrigine, they argue that the lack of measurable interaction suggests these mood stabilisers do not meaningfully interfere with esketamine response at typical clinical doses, despite theoretical concerns based on glutamatergic or neuroplasticity-related mechanisms. The authors also discuss implications for practice. They suggest clinicians may wish to minimise chronic or high-dose benzodiazepine co-administration when starting esketamine, particularly during induction, while the findings are somewhat reassuring for continuing lithium or lamotrigine when clinically indicated. They emphasise that the results extend real-world evidence to the subcutaneous route of administration, a comparatively under-studied formulation. Several limitations are acknowledged. The retrospective, single-centre design limits causal inference and increases the risk of selection bias and confounding by indication. The models were not adjusted for some potential confounders such as age, sex, diagnosis subtype, or baseline severity, and other background psychotropics may also have influenced outcomes. Benzodiazepine timing relative to esketamine, lithium serum levels at each time point, and detailed lamotrigine dose schedules were not available, and dose-response relationships could not be examined. The lamotrigine subgroup was relatively small, reducing power, and missing data may have introduced bias. The authors also stress that the findings apply to the acute six-week induction phase and should not be generalised to longer-term outcomes.
This was a retrospective naturalistic observational study utilizing real-world clinical data from the Esketamine Clinic at the Mood Disorders Program (PRODAF), Federal University of São Paulo (UNIFESP), Brazil. The clinic is a specialized unit dedicated to the treatment of severe, treatment-resistant mood disorders, including major depressive disorder and bipolar depression with inadequate response to conventional pharmacological interventions. Data were extracted from the clinical database spanning the period from January 2017 to December 2023, encompassing seven years of clinical experience with subcutaneous esketamine administration. This retrospective design was selected to leverage accumulated real-world clinical experience and provide pragmatic evidence regarding the efficacy and safety of esketamine in the context of complex polypharmacy, a scenario frequently encountered in clinical practice.
The study sample comprised 178 adult patients (age range: 18-65 years) diagnosed with either DSM-IV major depressive disorder (unipolar depression; n = 121, 68.0%) or bipolar depression (n = 57, 32.0%), confirmed with Mini International Neuropsychiatric Interview -MINI. All participants met diagnostic criteria for treatment-resistant depressive episode as defined by operational standards, requiring a minimum of two adequate pharmacological treatment failures, administered at therapeutic doses for a minimum duration of 6 weeks each, as assessed by clinical evaluation and documented medication records. Participants were required to present with a current major depressive episode of moderate to severe intensity at baseline, operationalized as a Montgomery-Åsberg Depression Rating Scale (MADRS) score ≥ 25, indicating substantial depressive symptomatology requiring urgent intervention. Additional inclusion criteria encompassed clinical judgment confirming the appropriateness of esketamine administration, capacity to provide informed consent, and absence of contraindications to esketamine therapy. Information on concomitant psychotropic medication use was available for 162 patients, who were included in the medication exposure analyses.
Participants were excluded if they presented with the following conditions: (1) primary psychotic disorders, including schizophrenia or schizoaffective disorder; (2) active substance use disorder or history of problematic ketamine/esketamine use; (3) significant cognitive impairment precluding informed consent; (4) unstable medical comorbidities that would contraindicate esketamine administration; or (5) participation in concurrent clinical trials. Additionally, patients with incomplete clinical data or fewer than three esketamine infusion sessions were excluded to ensure sufficient exposure to the therapeutic agent for adequate assessment of treatment response.
Esketamine was administered via subcutaneous injection once weekly for six consecutive weeks following the clinical protocol routinely used at the Esketamine Clinie (PRODAF -UNIFESP). The first administration was 0.5mg/kg and the second 0.75mg/kg. In selected cases, doses were further increased up to 1.0mg/kg according to clinical response and tolerability. Higher doses were more frequently used during the early implementation phase of the protocol, reflecting the naturalistic design of the study. This dosing strategy is consistent with emerging evidence supporting flexible, response-guided titration in subcutaneous ketamine protocols. Patients remained under clinical supervision during and after each administration for approximately two hours to monitor for potential adverse effects and changes in vital signs.
To address the primary research questions regarding pharmacological interactions, participants were stratified according to exposure to concomitant medications (lamotrigine, lithium, or benzodiazepines). These exposure categories were not mutually exclusive, as some patients received more than one concomitant medication. Analyses were conducted in the 162 patients with complete data on concomitant medication use. Four exposure categories were defined: (1) Esketamine-only group (ESK; n = 22, 13.6%), defined as participants receiving esketamine without concurrent use of benzodiazepines, lamotrigine, or lithium; (2) Esketamine with benzodiazepines (ESK+BZD; n = 109, 67.3%), encompassing participants receiving stable doses of any benzodiazepine agent (e.g., alprazolam, clonazepam, diazepam) concurrently with esketamine; (3) Esketamine with lamotrigine (ESK+LTG; n = 30, 18.5%), including participants receiving lamotrigine at therapeutic doses (≥ 50 mg/day) for mood stabilization or augmentation; and (4) Esketamine with lithium (ESK+Li; n = 70, 43.2%), comprising participants receiving lithium at therapeutic serum concentrations (0.6-1.2 mEq/L) for mood stabilization or augmentation. Benzodiazepines exposure were prescribed as part of routine clinical care, typically as scheduled daily medications rather than exclusively PRN ("as needed") use. Due to the retrospective nature of the dataset, detailed information regarding the exact timing of benzodiazepine administration relative to esketamine dosing sessions was not systematically available. For analytical purposes, benzodiazepine exposure was standardized by converting all doses to diazepam-equivalent daily doses according to established equivalence tables. Lithium exposure was defined as documented ongoing lithium treatment during the esketamine treatment period. In routine clinical practice at our center, patients receiving lithium had previously documented therapeutic serum levels in referral records; however, serum lithium values were not systematically recorded in the study database and were therefore not available for quantitative analysis. Lamotrigine exposure was defined as ongoing treatment during the esketamine induction phase. Medical records indicated stable lamotrigine dosing during this period, ranging from 50 to 400 mg/day. Individual dose values were not systematically recorded in the study database due to the retrospective design. It is important to note that some participants were concurrently exposed to multiple study medications (benzodiazepines, lamotrigine, and/or lithium); however, for the primary statistical analysis, participants were categorized into groups based on the presence or absence of each medication class independently, with sensitivity analyses conducted for individuals receiving multiple concomitant medications. The selection of these three medication classes was based on the following rationale: (1) preclinical and preliminary clinical evidence suggests potential attenuation of ketamine antidepressant effects through enhancement of GABAergic inhibition by benzodiazepines; (2) lamotrigine may theoretically interact with ketamine glutamatergic mechanism of action through modulation of glutamate release; and (3) lithium presents complex mechanisms of action involving modulation of multiple neurotransmitter systems and intracellular signaling pathways, potentially overlapping with ketamine's synaptic plasticity-promoting effects. Patients continued other psychotropic medications as part of routine clinical care. These commonly included antidepressants, antipsychotics, and other mood-stabilizing agents prescribed for treatment-resistant depressive episodes. Medication regimens were not standardized and were maintained according to clinical judgment.
The primary outcome was the change in depressive symptom severity from baseline (week 0) to week 7 6, operationalized using the Montgomery-Åsberg Depression Rating Scale (MADRS). MADRS assessments were conducted at baseline (week 0) and at weekly intervals throughout the treatment (weeks 1-6), corresponding to the six esketamine injections. Secondary outcomes included clinical response (defined as a ≥ 50% reduction in MADRS score from baseline) and remission (defined as MADRS score ≤ 10), assessed at endpoint (week 6). Additionally, the clinical course of depressive symptomatology across the study period was documented through weekly MADRS assessments, capturing symptoms change and enabling identification of differential response patterns across medication groups.
All analyses were conducted using anonymized clinical data from 178 participants treated with subcutaneous esketamine. Descriptive statistics, including age, sex, body mass index (BMI), diagnosis (unipolar or bipolar depression), and Maudsley Staging Method scores, were used to characterize sociodemographic and clinical variables. Analyses involving concomitant medication exposure were conducted in the subset of 162 patients with complete data on psychotropic medication use. Continuous variables were presented as means ± standard deviations and 95% confidence intervals, while categorical variables were reported as absolute and relative frequencies. Normality was assessed using the Shapiro-Wilk test. For comparisons between independent groups, the Wilcoxon rank-sum test was employed given the non-normal distribution of MADRS scores. Patients received different benzodiazepines at varying doses. For analytical purposes, benzodiazepine exposure was standardized by converting all doses to diazepam-equivalent daily doses, according to established equivalence tables. This approach allowed harmonization of benzodiazepine exposure across patients prior to statistical modeling. The distribution of diazepam-equivalent doses among benzodiazepine users is presente in Table. Detailed information regarding the timing of benzodiazepine administration relative to esketamine dosing sessions was not systematically available due to the retrospective nature of the dataset. -INSERT TABLE 1 ABOUT HEREA linear model (Equation) was used to explore associations between concomitant medication exposures and longitudinal changes in MADRS scores: ΔMADRS ∼ benzodiazepine × lamotrigine × lithium + ε In the linear model, no main effect or interaction reached statistical significance (p > 0.05). Although the combination of benzodiazepines and lithium showed a trend toward a negative interaction (p = 0.068), the overall model indicated that none of the concomitant medications significantly altered the antidepressant response to esketamine. A mixed-effects linear model (Equation) was then fitted to account for repeated measures across treatment weeks and interindividual variability: In this model, MADRS score was the dependent variable; week, benzodiazepine, lamotrigine and lithium exposure were included as fixed effects; and participant was included as a random-effect. In the mixed-effects model, week was significantly associated with lower MADRS scores over time (β = -2.82, p < 0.001),Among the concomitant medications, benzodiazepines were associated with higher MADRS scores across treatment weeks (β = 4.46, p = 0.026) Neither lamotrigine (p = 0.777) nor lithium (p = 0.343) showed significant main effects, and no significant interactions were observed between medications or between medications and time. Residual analyses confirmed model adequacy, and all tests were two-tailed with a significance threshold of p < 0.05. Dose variability was not included as a covariate given the pragmatic nature of the study and the absence of a predefined dose-response hypothesis.
A total of 178 patients were included in the analysis, of whom 109 (61%) were female. The mean age was 40.0 ± 13.6 years (Table). Regarding diagnosis, most patients had unipolar depression (68%), while 32% were diagnosed with bipolar disorder. The mean baseline MADRS score was 32.1 ± 6.5, indicating severe depressive symptomatology. The level of treatment resistance, as measured by the Maudsley Staging Method, was 10.7 ± 1.8, consistent with moderate to high treatment resistance (Table). -INSERT TABLE 2 ABOUT HERE -Baseline demographic and clinical characteristics stratified by medication exposure are presented in TableBaseline MADRS scores were similar across medication exposure groups..INSERT TABLE 3 ABOUT HERE --
Concomitant psychotropic medications were frequently used in this cohort. Lithium was prescribed in 80 patients (45.0%), benzodiazepines in 109 patients (67.3%), and lamotrigine in 30 patients (18.5%). Twenty-two patients (13.6%) received esketamine without concomitant use of lithium, benzodiazepines, or lamotrigine. Because patients could receive more than one medication, these exposure categories were not mutually exclusive. In the primary statistical analyses, medication exposures were modeled as independente binary variables (presence vs absence of each medication class). A progressive reduction in MADRS scores was observed across all groups with successive esketamine administrations. In descriptive analyses, patients treated with esketamine monotherapy showed a more homogeneous pattern of symptom improvement over time, whereas groups receiving concomitant lamotrigine or multiple medications displayed greater dispersion in MADRS trajectories. However, in the mixed-effects model, no significant medication-by-time interactions were observed, indicating that the overall trajectory of symptom improvement did not differ significantly between subgroups.
No differences in MADRS scores change were observed between subgroups when analyzed separately (Wilcoxon test, p > 0.05). In linear mixed-effects models assessing the effects of concomitant medications and their interactions on MADRS trajectories over time, no significant medication-by-time interactions were observed (all p > 0.05). A trend toward a negative interaction between benzodiazepine and lithium use was observed (p = 0.068), although this interaction did not reach statistical significance.
In the mixed-effects model including treatment week as a time variable and individual patient as a random effect, a significant main effect of week was observed (β = -2.82, 95% CI -3.24 to -2.39, p < 0.001), indicating a consistent reduction in depressive symptoms over time. Benzodiazepine use was associated with higher MADRS scores across treatment weeks (β = 4.46, 95% CI 0.53 to 8.39, p = 0.026). Neither lamotrigine (β = 1.04, 95% CI -6.14 to 8.21, p = 0.777) nor lithium (β = 1.70, 95% CI -1.82 to 5.22, p = 0.343) showed significant main effects on MADRS scores. No significant medication-by-time interactions were observed. Regarding concomitant medications, benzodiazepine use was associated with higher MADRS scores across all treatment weeks compared with patients not using benzodiazepines (β = 4.46, p = 0.026). Importantly, no significant benzodiazepine x time interaction was observed, indicating that the rate of symptom improvement over time did not differ between groups. Neither lamotrigine (β = 1.04, p = 0.777) nor lithium (β = 1.70, p = 0.343) showed significant main effects on MADRS scores, and no medication × time interactions reached statistical significance. -INSERT FIGUREClinical response and remission rates were also examined at treatment endpoint (week 6). Overall, 66 patients achieved clinical response, defined as a ≥50% reduction in MADRS score from baseline, and 34 achieved remission (MADRS ≤10). Response and remission analyses were restricted to patients with available baseline and endpoint MADRS scores, and denominators therefore differ slightly from those used in medication exposure analyses. Among benzodiazepine users, 43 of 101 patients (42.6%) achieved response and 19 (18.8%) achieved remission, compared with 23 of 52 (44.2%) and 15 (28.8%), respectively, among patients not receiving benzodiazepines. Similar patterns were observed for lithium and lamotrigine exposure (Table). -INSERT TABLEABOUT HERE.
The present study investigated the impact of commonly prescribed psychotropic medications-lamotrigine, lithium, and benzodiazepines-on the antidepressant efficacy of subcutaneous esketamine in a real-world cohort of patients with treatment-resistant depressive episodes. Although the present study focused on subcutaneous esketamine administration, most clinical trials and regulatory approvals have involved the intranasal formulation (Spravato®), which is currently approved by the U.S. Food and Drug Administration and other regulatory agencies for treatment-resistant depression. Differences in pharmacokinetics and bioavailability across routes of administration-including intravenous racemic ketamine, intranasal esketamine, and subcutaneous administration-may influence plasma exposure and onset of action. Nevertheless, available evidence suggests broadly comparable antidepressant effects across administration routes when adequate dosing strategies are usedPrimary analyses indicated that concomitant benzodiazepine use was associated with higher MADRS scores during treatment as evidenced by a smaller reduction in MADRS scores compared to the esketamine-only group. Importantly, no benzodiazepine x time interaction was observed, suggesting that the trajectory of symptom improvement over time did not differ between groups. Our findings are broadly consistent with earlier observations suggesting that benzodiazepine exposure may be associated with reduced antidepressant response to ketamine. For example,reported that concomitant benzodiazepine use was associated with attenuated response to intravenous ketamine in patients with treatment-resistant depression. However, the clinical interpretation of such findings remains complex in naturalistic settings, where benzodiazepines are frequently prescribed to patients with greater anxiety, insomnia, or illness severity. As highlighted in recent clinical perspectives, the use of ketamine and esketamine in routine practice often occurs in the context of ongoing polypharmacy, and understanding potential medication interactions remains an important area for future investigation. In contrast, both lithium and lamotrigine co-administration did not appear to interfere in the therapeutic trajectory. These findings provide critical insights into the pharmacological optimization of esketamine treatment in complex clinical settings. The association between benzodiazepine use and higher MADRS scores observed in this study aligns with emerging evidence from both preclinical and clinical research.first suggested that benzodiazepines might "blunt" the antidepressant response to racemic ketamine, a finding later supported by, who reported delayed and reduced response rates in patients receiving concomitant GABAergic agents. The present results extend these observations to the subcutaneous route of esketamine administration and are consistent with hypothesis suggesting that GABAergic potentiation may interfere with the glutamatergic mechanisms essential for rapid symptom relief. Benzodiazepines act as positive allosteric modulators at the GABA-A receptor, enhancing inhibitory tone and dampening the glutamatergic surge that mediates ketamine's synaptic potentiation. This GABAergic modulation may interfere with downstream AMPA-BDNF signaling, thereby reducing antidepressant effects. In addition, benzodiazepine exposure in our cohort was heterogeneous, with a wide range of diazepam-equivalent doses. Although doses were standardized using diazepam equivalents, the present study was not designed to evaluate dose-response relationships. Furthermore, the timing of benzodiazepine administration relative to esketamine sessions could not be evaluated, which may influence acute antidepressant responses. An alternative explanation for this finding is confounding by indication, a common limitation in naturalistic observational studies. In routine clinical practice, benzodiazepines are frequently prescribed to patients presenting with greater anxiety symptoms, insomnia, agitation, or overall illness complexity. Therefore, benzodiazepine exposure may act as a proxy marker of a more severe clinical phenotype rather than a direct pharmacodynamic interaction with esketamine. The absence of a benzodiazepine × time interaction in the mixedeffects model further supports the interpretation that symptom trajectories over time were broadly similar across groups. In contrast, neither lamotrigine nor lithium was associated with significant differences in antidepressant response These findings are noteworthy in light of preclinical and small clinical studies proposing potential pharmacodynamic interactions between these agents and ketaminerelated neuroplastic mechanisms. Theoretical models proposed that lithium might potentiate ketamine antidepressant action by inhibiting glycogen synthase kinase-3 (GSK-3), a mechanism linked to increased synaptic resilience and neurotrophic signaling. Similarly, lamotrigine could theoretically attenuate glutamate release and modulate excitatoryinhibitory balance. However, the absence of measurable interaction in this cohort suggests that, at typical clinical doses, neither lithium nor lamotrigine meaningfully alters the antidepressant trajectory of esketamine. These results are consistent with small-scale clinical studies and preclinical data showing that the net pharmacodynamic effects of these agents may be dose-, timing-, or formulation-dependent. From a clinical perspective, these findings have potentially relevant implications for the management of treatment-resistance depressive episodes in routine practice. Given the widespread use of benzodiazepines among patients with TRD-often for managing anxiety, insomnia, or agitation-understanding their impact on ketamine efficacy is crucial. While short-term anxiolytic use may remain necessary in selected cases, present findings suggest that clinicians should consider minimizing chronic or high-dose benzodiazepine co-administration when initiating treatment with esketamine, especially during the induction phase, to optimize antidepressant effects. Conversely, the absence of significant effects of lithium and lamotrigine provides reassurance regarding the concomitant use of mood stabilizers in treatment-resistant depressive episodes, supporting treatment continuity when esketamine is introduced. This study also highlights the value of real-world data in complementing randomized clinical trials. The sample drawn from a university clinic reflects the complexity of treatmentresistance depressive episodes in clinical practice-characterized by long illness duration, polypharmacy, and varying degrees of treatment resistance. Despite this heterogeneity, the consistent week-to-week improvement observed across all strata underscores the antidepressant effectiveness of the subcutaneous route in clinical settings, extending findings from randomized clinical trials using other routes of administration. Although the present study focused on subcutaneous esketamine administration, most of the available evidence on ketamine-based treatments for depression derives from studies using intravenous ketamine or intranasal esketamine. Therefore, the present findings should be interpreted within the broader context of ketamine-related antidepressant treatments rather than being restricted to a specific route of administration. Taken together with prior randomized and real-world studies, these results contribute to the growing literature suggesting that ketaminebased therapies can produce consistent symptom improvement in treatment-resistant depressive episodes across different clinical settings and administration routes. The antidepressant effectiveness observed in this naturalistic cohort is further supported by emerging randomized evidence favoring flexible, response-guided dosing strategies for subcutaneous ketamine. A recent multicenter phase 3 trial demonstrated that flexible doses up to 0.9 mg/kg were associated with superior antidepressant outcomes compared to fixed-dose regimens. Together, these findings suggest that dose optimization may represent a clinically relevantyet still underexploredmoderator of ketamine-related antidepressant response in real-world populations. A key strength of this study lies in its high ecological validity, as it draws on a large realworld cohort of patients with severe treatment-resistant depressive episodes treated in a tertiary, specialized mood disorders service, thereby reflecting the clinical complexity and polypharmacy that typically characterize treatment-resistant population in routine practice. The use of prospectively collected, repeated MADRS assessments across the treatment enabled a longitudinal characterization of symptom trajectories. In addition, by focusing on subcutaneous esketamine, the study contributes data on a route of administration that remains comparatively underrepresented in the literature dominated by intravenous ketamine and intranasal esketamine trials, offering pragmatic evidence relevant to services that employ cheaper, less complex delivery strategies. Finally, the explicit examination of three commonly co-prescribed medication classesbenzodiazepines, lamotrigine, and lithiumaddresses a clinically salient knowledge gap regarding potential pharmacodynamic interactions that may meaningfully influence outcomes and inform treatment optimization in treatment-resistant depressive episodes. Present findings should be interpreted in light of several limitations. First, the retrospective and naturalistic design precludes causal inference and is vulnerable to selection bias and confounding by indication, as patients receiving benzodiazepines, lamotrigine, or lithium may differ systematically in illness severity, comorbidity burden, anxiety/insomnia symptoms, or prior treatment history. Although baseline characteristics were described according to medication exposure groups, the mixed-effects models were not adjusted for potential confounders such as age, sex, diagnosis subtype, or baseline symptom severity. Consequently, residual confounding by indication therefore cannot be excluded. Moreover, most patients continued background pharmacotherapy, including antidepressants and antipsychotics, as part of routine clinical care. Although medication regimens were generally stable during the induction phase, the potential influence of concomitant antidepressant treatment on symptom trajectories cannot be excluded. Second, although repeated MADRS assessments enabled modeling of symptom trajectories across treatment weeks, the observational nature of the dataset limits the ability to fully characterize clinical factors that may influence treatment response. MADRS subitems were not analyzed separately, and therefore it was not possible to explore whether specific symptom domains such as anxiety, inner tension, agitation, or sleep disturbance differentially influenced response trajectories during esketamine treatment., In addition, although sex distribution was described in the sample, the present analyses did not formally evaluate sex as a moderator of treatment response. Emerging evidence suggests that biological sex may influence response trajectories during esketamine treatment, and future studies should explore potential sex-bytreatment interactions in larger samplesThird, although esketamine dosing was flexibly adjusted according to clinical response and tolerability, mean dose levels were relatively consistent across the cohort. Dose was therefore not modeled as a time-varying covariate in the mixed-effects analyses, and future studies should explore potential dose-response relationships. Medication exposure was operationalized categorically, without granular characterization of benzodiazepine type, dose (e.g., diazepam equivalents), timing relative to esketamine administration, or treatment duration; likewise, lithium exposure was not systematically anchored to serum concentrations at each assessment point, and lamotrigine dose escalation schedules were not uniformly available. Fourth, the sample sizes of some subgroupsparticularly the lamotrigine group (n = 30) limited statistical power to detect potential differences involving this subgroup and increased the risk of type II error. Fifth, the study was conducted in a single tertiary academic clinic, which may limit generalizability to other healthcare settings, routes of administration (e.g., intranasal), or less severely ill populations. Finally, missing data inherent to real-world follow-up and the chosen handling strategy may have introduced bias if data were not missing at random. Subcutaneous esketamine treatment was associated with a clinically meaningful reduction in depressive symptom severity in this real-world cohort of patients with treatment-resistant depressive episodes. These findings apply to the acute six-week induction phase and should not be interpreted as evidence regarding longer-term treatment outcomes. Benzodiazepine use was associated with higher MADRS scores during treatment; however, no benzodiazepine × time interaction was observed, indicating that the trajectory of symptom improvement over time did not differ between groups. Thus, benzodiazepine exposure in this cohort was associated with higher overall symptom severity rather than a slower rate of improvement during treatment. In contrast, lamotrigine and lithium use were not associated with differences in depressive symptom trajectories, supporting the clinical feasibility of maintaining these agents during esketamine treatment when clinically indicated. Given the retrospective design, potential confounding by indication, and limited statistical power for some subgroups (particularly lamotrigine), these findings should be interpreted cautiously. Prospective, adequately powered studies are needed to clarify medication-specific interactions-particularly potential dose-and timing-dependent benzodiazepine effects-and to provide clearer guidance for the management of polypharmacy in patients receiving esketamine for treatment-resistant depressive episodes.
Create a free account to open full-text PDFs.