Safety and effectiveness of intranasal esketamine for treatment-resistant depression: a real-world retrospective study

Major depressive disorder (MDD) is common and a leading cause of disability. A substantial subgroup—treatment-resistant depression (TRD)—is conventionally defined as failure to respond to at least two adequate antidepressant trials and accounts for about one third of MDD cases, contributing disproportionately to the disease burden. Existing pharmacological and somatic treatments (switching or combining antidepressants, augmentation strategies, electroconvulsive therapy, transcranial magnetic stimulation, vagal nerve stimulation, deep brain stimulation) have limitations including delayed onset, high cost, procedural risk or limited evidence for TRD, leaving an unmet need for rapid-acting, tolerable therapies. Ketamine and its S-enantiomer esketamine have mechanistic plausibility for rapid antidepressant effects and esketamine nasal spray received FDA approval in 2019 on the basis of Phase III trials demonstrating faster onset and relapse-prevention effects under a Risk Evaluation and Mitigation Strategies (REMS) programme that mandates clinic-based administration and post-dose monitoring. Brendle and colleagues note that patients treated in routine clinical practice may differ from clinical trial samples in symptom severity, comorbidity and treatment schedules, so real-world safety and effectiveness data are limited. The present study set out to characterise the demographic and clinical profile, treatment patterns, clinical outcomes (depression, anxiety, suicidal ideation) and adverse events among adults with TRD receiving intranasal esketamine at a private outpatient integrative psychiatric clinic between July 2019 and June 2021, using retrospective electronic health record (EHR) data.

This was a retrospective observational cohort study using EHR data from a private outpatient integrative psychiatric clinic that specialises in ketamine and esketamine treatments. The analysis followed International Society for Pharmacoepidemiology good practice guidance and received institutional human subjects review board approval. Inclusion criteria were adults aged 18 years or older with a diagnosis of recurrent major depressive disorder without psychotic features (ICD-10 F33.2) who received intranasal esketamine between July 2019 and June 2021. Patients who received other ketamine formulations (intravenous, oral, sublingual, intramuscular) were excluded. Data were extracted from structured EHR fields and by chart review when necessary. Extracted variables included demographics, completed history questionnaires (medical, psychiatric, social, family), diagnoses, concomitant medications, treatment dates, adverse events and REMS-related monitoring data such as blood pressure and observation time. The primary clinical outcomes were change in depression symptoms measured by the Patient Health Questionnaire (PHQ-9), change in anxiety symptoms measured by the Generalized Anxiety Disorder scale (GAD-7) and suicidal ideation measured by PHQ-9 item 9. The authors noted established interpretive thresholds for the instruments (PHQ-9 range 0–27; a 3-point reduction is often considered clinically meaningful and a 6-point reduction clinically substantial; GAD-7 range 0–21). Statistical analyses were descriptive and comparative. The cohort was stratified by number of esketamine treatments; paired two-sample t-tests compared baseline to last available treatment scores. Correlation analyses (Spearman) assessed associations between number of sessions and outcome changes. Multiple linear regression models were built using forward selection (entry p = 0.15) to identify predictors of change in PHQ-9 and GAD-7 scores; covariates with missing data were excluded. Regression diagnostics (outlier and influence checks, collinearity, residual plots, normality) were applied. Analyses used STATA 16.1 and statistical significance was defined as p < 0.05.

A total of 171 patients met inclusion criteria. The sample had a mean age of 36 (±12) years, 60% were female and 92% were recorded as white. Common psychiatric comorbidities included generalized anxiety disorder (39%; n = 67), attention-deficit hyperactivity disorder (25%; n = 43), insomnia (25%; n = 43) and post-traumatic stress disorder (12%; n = 20). Most patients (98%; n = 168) had recorded use of other psychiatric medications, with an average of 5.8 (SD 4.0) psychiatric medications per patient; 94% had an oral antidepressant prescription and these patients were taking a mean of 2.3 (SD 1.4) oral antidepressants. Treatment exposure and symptom trajectory: the cohort median number of esketamine sessions was 11 (IQR 10.5). For sessions 1–8 mean PHQ-9 and GAD-7 scores decreased steadily; scores showed a transient increase around session 9 after which trends were variable but did not exceed levels observed at session 14. Median days between sessions increased from 3–6 days during sessions 1–9 to 7 days thereafter. Change from baseline to last observed treatment: 148 patients had baseline and follow-up PHQ-9 scores taken more than two weeks after the first treatment; baseline PHQ-9 mean was 16.7 (SD 5.8) and last-treatment mean was 12.0 (SD 6.4), a significant reduction (p < 0.001). Subgroup analyses by number of treatments showed significant baseline-to-last reductions for most strata except the 2–5 treatment group; baseline PHQ-9 was higher among patients who went on to receive more treatments. For suicidal ideation, 140 patients had paired data: mean PHQ-9 item 9 dropped from 1.09 (SD 1.06) at baseline to 0.79 (SD 0.94) at last treatment (p < 0.001), with significant subgroup reductions in the 6–10, 11–15 and 31–71 treatment categories. For anxiety, 120 patients had paired GAD-7 scores: baseline mean 12.0 (SD 5.8) to last-treatment mean 8.7 (SD 5.6), a significant reduction (p < 0.001); all treatment-number subgroups showed significant reductions. Associations and multivariable models: change in PHQ-9 correlated negatively with number of treatments (Spearman rho = -0.2645; p = 0.001), indicating greater symptom reduction with more sessions. Change in GAD-7 also correlated negatively with session count (rho = -0.1871; p = 0.04). In the final linear model predicting PHQ-9 change (n = 137) higher age (p = 0.021), higher baseline PHQ-9 (p < 0.001) and longer treatment duration (p = 0.034) were associated with greater decreases; female sex was associated with an increase in PHQ-9 from baseline to last treatment (p = 0.021). The mean PHQ-9 change in that subsample was -4.7 (SD 6.2) and the model R2 was 29%. For GAD-7 (n = 120), higher baseline GAD-7 predicted larger decreases (p < 0.001) while a diagnosis of GAD predicted an increase in GAD-7 from baseline to last treatment (p = 0.026); the mean change was -3.32 (SD 4.68) and R2 was 27%. Safety and tolerability: dissociation was reported in 73% of patients (n = 125) and sedation in 22% (n = 38). When examined per treatment, sedation occurred on average in 8% (SD 22%) of treatments and dissociation in 48% (SD 41%) of treatments. Most sedation onsets occurred within the first 30–60 minutes and resolved within 2 hours in 97% of cases; dissociation onset occurred within 30 minutes for most treatments and resolved within 2 hours in 99% of cases. There was one serious adverse event (0.58%) of prolonged dissociation/sedation with nausea/vomiting after an 84 mg dose; the episode resolved the same day. Blood pressure measures showed no clinically notable sustained elevations in the extracted text and most patients were ready for discharge within approximately 90 minutes post-administration under REMS monitoring.

Brendle and colleagues interpret the findings as showing that intranasal esketamine was associated with clinically meaningful reductions in depressive and anxiety symptoms in this real-world TRD cohort, without major safety concerns. The authors highlight that patients receiving 6–10 treatments had an average PHQ-9 reduction of 4.1 points, while those receiving 11–71 treatments showed reductions in the range of 5.5–6.9 points; these changes exceed the typical 3-point threshold for clinical meaningfulness and are larger than the treatment effects reported in short-term randomized controlled trials, which the authors cite as approximately -2.2 to -2.8. They note the temporal pattern in symptom scores, with improvement during the induction phase (sessions 1–8) followed by increased variability after treatment frequency was reduced around session 9. The median interval between sessions rose to 7 days after session 9, and the authors propose that less frequent dosing may explain the plateau in improvement because esketamine has rapid, relatively transient effects. Continued administration and frequent clinical contact are presented as factors likely contributing to sustained symptom improvement; the discussion aligns these observations with SUSTAIN-1 trial data, which showed that continuation of esketamine reduced relapse risk compared with switching to placebo. However, the authors caution that these results do not demonstrate lower relapse risk compared with oral antidepressants and situate their findings in the context of historical relapse rates from other trials. On population characteristics, the cohort was relatively young, predominantly female and white, with higher rates of comorbidities (for example, GAD) and greater exposure to psychiatric medications than some claims-based TRD samples. The authors suggest that future prospective studies should track medication management to determine whether oral antidepressants or other psychotropic medications can be reduced as depressive symptoms improve with esketamine. Regarding safety, the investigators report that most acute adverse effects (dissociation, sedation, transient blood pressure elevations) were short-lived and resolved within the REMS-mandated monitoring period, supporting the sufficiency of the two-hour observation window; only one serious adverse event occurred and resolved the same day. They note that the observed monitoring and typical readiness for discharge within 90 minutes may inform considerations about time commitments and access barriers for esketamine therapy. The authors acknowledge variability in response and the limitations inherent to a retrospective, single-clinic EHR analysis—such as potential selection bias, missing questionnaire data for some patients, and lack of a control group—although these specific limitations are not detailed exhaustively in the extracted text. They frame the study as providing complementary real-world evidence on effectiveness, tolerability and treatment patterns for esketamine in routine outpatient care.