Real-world effectiveness and safety of intranasal esketamine for treatment-resistant depression: data from the enTRD registry

Treatment-resistant depression (TRD) is described as persistent depression despite at least two adequate antidepressant trials, and it is associated with substantial chronicity, suicidality, and health and productivity burden. The introduction notes that intranasal esketamine offers a different treatment pathway from standard antidepressant strategies, but its true clinical value has remained debated because placebo-controlled trials show statistically significant benefits while effect sizes are modest and the balance against adverse events is still uncertain. The authors also emphasise that real-world evidence is useful for understanding how esketamine performs in routine care, including discontinuation patterns and safety, although it cannot establish causality without a control group. Preiss and colleagues therefore set out to examine real-world induction-phase outcomes for intranasal esketamine in patients with TRD treated in a specialised outpatient clinic. Their aim was to characterise changes in clinician-rated depression severity, patient-reported symptoms, response and remission, suicidality, and tolerability during routine supervised care, with detailed reporting of treatment exposure and concomitant pharmacotherapy.

This was a monocentric, retrospective, observational registry analysis of routine clinical care data from the TRD outpatient clinic at the Department of Psychiatry and Psychotherapy, University Hospital Vienna. Treatment with intranasal esketamine began in January 2021, and the analysis used data cut-off in November 2025. Patients underwent a standardised pre-induction evaluation to confirm the referring diagnosis of TRD, review contraindications, assess psychiatric and somatic comorbidity, and check depressive symptom severity. Adults considered clinically suitable were offered esketamine. For induction-phase analyses, the study included all patients who received at least one esketamine administration; patients who completed all eight planned administrations were classified as completers, and those who stopped earlier were dropouts. Esketamine was given in a low-stimulation clinic setting under direct staff supervision, with self-administration of the nasal spray. During induction, dosing followed the approved product information, titrated up to 84 mg per administration, twice weekly for up to eight administrations over four weeks. Treatment was combined with ongoing antidepressant therapy, typically an SSRI or SNRI, and other psychotropic medication could be continued or adjusted by the treating clinician. Patients were monitored before and after each administration, with blood pressure and other vital signs checked repeatedly. Outcomes were collected across 11 visits: baseline, pre-treatment, eight treatment visits, and end of induction. The primary outcome was change in Montgomery-Åsberg Depression Rating Scale (MADRS) total score from the pre-treatment visit to end of induction. Secondary outcomes included change across a prespecified three-factor MADRS model, trajectory of self-rated depression on the PHQ-9, response and remission rates, prognostic value of baseline hopelessness, and safety and suicidality outcomes. Analyses were prespecified in a statistical analysis plan and conducted in R. The main analysis used a likelihood-based mixed model for repeated measures, with age and sex as covariates. Sensitivity analyses included a covariate-adjusted change-score model, multiple imputation using chained equations, and delta-adjusted pattern-mixture tipping-point analyses to explore missing-not-at-random scenarios. Secondary analyses used mixed models for factor scores, logistic regression for the hopelessness analysis, piecewise mixed modelling for PHQ-9 trajectories, and McNemar's test for paired response/remission comparisons.

From 229 referred patients, 224 formed the pre-induction cohort and 176 initiated intranasal esketamine, constituting the induction analysis population. Of these, 156 completed induction and 20 discontinued before completion. The induction cohort had a mean age of 40.8 years, 54.0% were female, and baseline illness burden was high: mean MADRS was 31.5, mean PHQ-9 was 18.5, and median Clinical Global Impression-Severity was 5. Most patients had recurrent depressive disorder, and 74.9% met criteria for chronic major depressive disorder. Psychiatric comorbidity was common, particularly anxiety disorders, and treatment resistance was substantial, with a median of five prior antidepressant trials. Antidepressant regimens were frequently optimised before or during induction. The median number of esketamine administrations was 8, with a median induction duration of 30 days. In the primary mixed-model analysis, adjusted MADRS scores were 30.3 at the pre-treatment visit and 17.7 at end of induction, corresponding to an adjusted change of -12.5 points (95% CI -14.1 to -11.0; p<0.001). This reflected a large effect size (Hedges' g=1.8). Sensitivity analyses were consistent: the adjusted change was -12.3 points in the ANCOVA model, -12.6 points after multiple imputation, and remained similar in tipping-point analyses. All three MADRS factor scores improved. The affective-anhedonic factor showed the largest reduction, from 3.6 to 2.0, and improvement in this domain was greater than in the anxiety-vegetative and hopelessness factors. The association between baseline hopelessness and non-response did not reach statistical significance. Self-rated PHQ-9 scores also fell steadily across induction, with a total V1-V8 change of -4.9 points (95% CI -5.7 to -4.2; p<0.001), and there was no evidence that the rate of improvement differed between the early and late phases of treatment. At end of induction, remission was 27.2% and response was 57.4% among those with available data. These rates were significantly higher than at the pre-treatment visit. Under a conservative last-observation-carried-forward approach, remission was 24.4% and response 44.3%. Suicidality improved: MADRS item 10 and PHQ-9 item 9 both decreased significantly across treatment. At baseline, 14.8% reported current suicidal ideation, 40.9% had a history of severe suicidality, and 21.7% had at least one prior suicide attempt. Adverse events were recorded in 34.7% of patients. The most common categories were nausea and vomiting, significant blood pressure elevation, dissociation, and sedation. On a per-administration basis, adverse events occurred in 7.8% of administrations. Treatment discontinuation due to adverse events occurred in 3.4% of patients. One patient required hospital admission for worsening depressive symptoms. No suicide or serious self-harm events occurred during induction.

The authors interpret the findings as showing that intranasal esketamine induction in routine outpatient care was associated with a marked and clinically meaningful reduction in depressive symptoms, together with improvement in suicidality and generally acceptable tolerability. They emphasise that MADRS scores were stable between baseline and the pre-treatment visit, which they argue supports the view that the later improvement was linked to esketamine exposure rather than to nonspecific change or regression to the mean. They also note that the robustness of the primary result across multiple sensitivity analyses strengthens confidence in the observed signal. Preiss and colleagues position their results as broadly consistent with earlier real-world esketamine evidence, including a recent meta-analysis of routine-care studies. They state that their cohort was clinically complex, with high rates of chronic depression, psychiatric comorbidity, and prior substance use disorder, which may make the findings relevant to tertiary care settings. They also suggest that their repeated PHQ-9 assessments add detail to previous work by showing continuous improvement across sessions rather than a coarse before-and-after pattern. Their factor-level analyses are presented as extending earlier reports by showing that all MADRS domains improved, with the strongest effect in the affective-anhedonic domain. The authors acknowledge several limitations. The study was retrospective, single-centre, and observational, with no parallel comparator group, so causality cannot be established and improvements cannot be attributed to esketamine alone. They note the possibility of selection and survivor bias because non-initiators and early discontinuers may differ from completers. Adverse events were taken from routine documentation rather than structured rating scales, which may have underestimated frequency. The induction-only focus limits inference about maintenance treatment and durability, and residual confounding from unmeasured clinical factors and concomitant treatment changes remains possible. The authors also note that comparison with prior studies is complicated by heterogeneity in routine-care cohorts and methods. In terms of implications, they argue that real-world data remain important for contextualising benefit-risk balance in routine practice, particularly given ongoing debate about the clinical relevance of average effects in controlled trials. They suggest that longer-term registry follow-up, comparative designs, and more standardised reporting of concomitant medications and adverse events are needed. They also indicate that separating pre-induction non-initiation from induction discontinuation may help services understand access and retention in routine care.