Intranasal esketamine effectively treats treatment-resistant depression in adults regardless of baseline irritability

Irritability—an elevated tendency to anger or aggression in response to frustration or perceived threat—is a frequent but often under-recognised feature of major depressive disorder (MDD). The extracted text notes that persistent irritability is reported by roughly 30%–50% of adults with MDD, is linked to earlier illness onset and comorbid anxiety, and is associated with worse outcomes including longer and more severe episodes, reduced functioning, and greater suicidal ideation. Despite these associations, the relationship between irritability and treatment outcomes in treatment-resistant depression (TRD) has not been well characterised. Jha and colleagues used pooled data from two Phase III trials of intranasal esketamine (ESK) plus a newly initiated oral antidepressant (ESK+AD) versus placebo nasal spray plus a newly initiated oral antidepressant (AD+PBO) to examine whether baseline irritability influenced antidepressant response in adults with TRD. The investigators also explored change in irritability symptoms over the 28-day double‑blind treatment period as a secondary interest.

This report is a post hoc analysis of pooled data from two similarly designed, multinational, randomised, double-blind, active-controlled Phase III trials, TRANSFORM-1 and TRANSFORM-2. Eligible participants had moderate-to-severe depression (MADRS total score ≥28) and documentation of nonresponse to 1–5 oral antidepressants in the current episode at screening; at randomisation all participants met a TRD definition of nonresponse to ≥2 adequate oral antidepressant trials in the current major depressive episode. Key exclusions included recent suicidal intent or behaviour, psychotic or bipolar disorders, recent moderate-to-severe substance use disorder, and positive drug screens for specified substances. The primary clinician-rated outcome was the Montgomery–Åsberg Depression Rating Scale (MADRS), administered at baseline and on days 2, 8, 15, 22 and 28 by independent blinded raters using a structured interview guide. Irritability was operationalised using item 6 of the GAD-7 ("Becoming easily annoyed or irritated"); patients were classified as high (item score ≥2 at both screening and baseline), varying (item score ≥2 at either visit), or low (item score <2 at both visits). Safety was assessed through treatment-emergent adverse events (TEAEs) coded with MedDRA v20.0. Efficacy and safety analyses used the full analysis set (all randomised patients who received ≥1 dose of study nasal medication and ≥1 dose of oral antidepressant during the 4-week induction). Changes in MADRS total score at day 28 were analysed with ANCOVA models including fixed effects for treatment, irritability level, study identifier, class of oral antidepressant (SSRI or SNRI) and baseline value as covariate; logistic regression models examined response (≥50% MADRS reduction) and remission (MADRS ≤12) with treatment, irritability level and study identifier as predictors and with a treatment-by-irritability interaction term tested. Number needed to treat (NNT) was estimated from response and remission rates. Changes in irritability (GAD-7 item 6) were analysed with Wilcoxon signed-rank and rank-sum tests. The extracted text states no multiplicity adjustment was applied.

The pooled full analysis set included 560 patients (ESK+AD n = 339; AD+PBO n = 221). At baseline, 296 patients (52.9%) met criteria for high irritability, 134 (23.9%) for varying irritability, and 130 (23.2%) for low irritability. Patients with high irritability were reported to be younger at MDD diagnosis, to have higher BMI, longer current episode duration, more prior oral antidepressant nonresponses in the current episode, greater lifetime suicidal ideation/behaviour, higher anxiety symptoms and more comorbid anxiety disorders than those with low irritability; baseline MADRS scores were similar across irritability subgroups. A total of 518 patients (ESK+AD n = 310; AD+PBO n = 208) had MADRS recorded at day 28 and were included in the post hoc endpoint analyses. In the pooled population, ESK+AD produced a greater mean improvement in MADRS total score at day 28 than AD+PBO (least squares mean change −19.9 vs −15.6, P < 0.001). This superiority reached statistical significance within the low (P = 0.011) and varying (P = 0.008) irritability subgroups; the high irritability subgroup showed a numerically greater improvement with ESK+AD but did not meet significance (P = 0.122). There was no statistically significant interaction between baseline irritability and treatment effect on MADRS change (interaction P = 0.267). Response and remission rates also favoured ESK+AD overall. At day 28, response occurred in 58.7% of ESK+AD patients versus 45.2% of AD+PBO patients (P < 0.001), and remission in 42.3% versus 30.8% (P = 0.004). NNTs for response or remission were <10 for most subgroup comparisons, although the NNT to achieve remission in the high irritability group was reported as 16, higher than in the low and varying groups (both <10). Logistic regression estimated that ESK+AD nearly doubled the odds of response overall (OR 1.92, 95% CI 1.33–2.76). Subgroup ORs for response ranged from 1.79 (95% CI 1.09–2.93) in the high irritability group to 2.21 (95% CI 1.03–4.73) in the low irritability group; some subgroup CIs crossed 1. For remission the overall OR was 1.75 (95% CI 1.20–2.56), with subgroup ORs reported as 1.39 (95% CI 0.84–2.32) for high irritability, 2.35 (95% CI 1.01–5.47) for varying irritability, and 2.31 (95% CI 1.07–4.97) for low irritability. Irritability measured by GAD-7 item 6 improved significantly from baseline to day 28 within all treatment subgroups (within-group P ≤ 0.005). Between‑group differences in change were not statistically significant, though trends favouring greater improvement with ESK+AD were noted in the high and varying irritability groups (P = 0.087 and P = 0.084, respectively); the largest median improvement (2 points) occurred in ESK+AD patients with high baseline irritability. Safety findings were described as consistent with prior reports and generally tolerable across irritability subgroups. The extracted text indicates four patients experienced at least one serious adverse event (SAE) distributed across treatment arms, but the extraction is fragmented and does not clearly report denominators for all events. The most common TEAEs in ESK+AD recipients during the double-blind phase included nausea (range 16.7%–33.9%), dissociation (25.6%–30.9%), dizziness (18.0%–28.9%), dysgeusia/bitter taste (11.1%–26.9%) and vertigo (16.7%–35.9%).

Jha and colleagues interpret these post hoc pooled-trial analyses to indicate that intranasal esketamine, when added to a newly initiated oral antidepressant, is associated with greater improvement in depressive symptoms and higher response and remission rates than oral antidepressant plus placebo in adults with TRD, irrespective of baseline irritability. The investigators highlight that more than half of the sampled TRD patients reported high irritability and that high irritability was associated with earlier age at MDD diagnosis, greater anxiety, higher BMI, more prior antidepressant nonresponses and more frequent lifetime suicidal ideation/behaviour; baseline MADRS scores did not differ by irritability level, suggesting that MADRS alone may not capture the broader syndromic burden of irritability. The authors note that although MADRS improvements were numerically larger with ESK+AD across irritability strata, statistically significant differences in MADRS change were clearer in the low and varying irritability groups than in the high irritability group, whereas response rates showed a statistically significant advantage for ESK+AD even in the high irritability subgroup. They further discuss NNT findings, pointing out that NNTs <10 for response across groups support potential clinical benefit, while the higher NNT for remission in the high irritability group suggests remitting these patients may be more challenging. Several limitations are acknowledged. The analysis was post hoc and the trials were not prospectively designed to test irritability-specific hypotheses. Subgroup sample sizes were limited, no adjustment for multiplicity was made, and irritability was assessed using a single GAD-7 item that likely does not capture the full construct or provide informant perspectives. The authors also recognise potential confounding because irritability was associated with other clinical features (for example, prior nonresponse and comorbid anxiety), and they note the sample was predominantly White, which limits generalisability. Finally, the investigators call for prospective studies that use more comprehensive irritability measures and include more diverse samples to replicate and extend these findings. They conclude that the current analyses provide supportive evidence of esketamine's efficacy in TRD regardless of baseline irritability and offer preliminary indications that esketamine may reduce irritability symptoms, while emphasising the need for further targeted research.