Rapidity of Symptom Improvement With Intranasal Esketamine for Major Depressive Disorder: A Systematic Review and Meta-Analysis

Major depressive disorder (MDD) is a leading cause of disability worldwide and clinicians lack rapid-acting pharmacological options because conventional oral antidepressants typically require weeks to months for full effect. Over the past decade, intravenous (IV) ketamine and its S-enantiomer, esketamine, have shown fast-acting antidepressant effects in controlled trials, and intranasal (IN) esketamine is now approved in the US and EU as an adjunctive treatment for treatment-resistant depression (TRD). However, individual Phase III IN esketamine trials failed to demonstrate a statistically significant excess of sustained responders at day 2 (24 hours) for esketamine versus control, a key secondary endpoint agreed with regulators for demonstrating rapidity of action. Hock and colleagues therefore undertook an updated systematic review and meta-analysis to quantify the magnitude of the treatment effect of adjunctive IN esketamine versus intranasal placebo at 24 hours after the first dose, and as a secondary objective to estimate the magnitude of effect at study endpoint. The analysis also planned exploratory subgroup and dose-related analyses to examine potential moderators of response.

The researchers conducted systematic searches up to November 2020 of PubMed, abstracts from major psychiatric meetings, and ClinicalTrials.gov, with no language restrictions. They sought randomized, double-blind clinical trials that compared adjunctive IN esketamine plus a standard antidepressant against adjunctive IN placebo (saline) plus the same antidepressant in adults with MDD. Excluded reports were those focusing on bipolar disorder, dysthymia, psychotic MDD, minor depression, seasonal affective disorder, depression secondary to medical conditions, active substance or alcohol abuse, non-original data, or non-acute-phase designs. Included trials used either the Hamilton Depression Rating Scale (HDRS-14) or the Montgomery–Åsberg Depression Rating Scale (MADRS) as the primary outcome; in practice, all included trials used MADRS. Data extraction used a pre-coded form and was performed independently by two authors (G.I.P. and N.I.), with discrepancies resolved by joint review. Extracted items included randomized arm sizes, population type (TRD versus MDD with suicidal ideation, SI), setting, region, trial duration, dosing/frequency of esketamine, type of control, baseline and mean change scores with standard deviations for the primary outcome at 24 hours and at endpoint, and remission rates. A small number of missing data points were obtained from the Yale University Data Access (YODA) project under an agreement with the trial sponsor. The primary quantitative outcome was the pooled standardized mean difference (SMD) in change scores on the primary outcome between adjunctive esketamine and placebo; SMD standardises the treatment effect relative to pooled standard deviation. Homogeneity was assessed via the DerSimonian and Laird test statistic, and the investigators presented results from a random effects model, which incorporates both within- and between-study variance. Exploratory subgroup and meta-regression analyses evaluated dose groups (28 mg, 56 mg, 84 mg) and study population (TRD versus SI). Remission outcomes were summarised with risk ratios (RRs) and numbers needed to treat (NNT). All meta-analyses used the meta package implemented in Stata 15. The study was unfunded and the senior author had full access to the data and final publication responsibility.

The search identified 26 PubMed abstracts, of which 7 described eligible randomized, double-blind IN esketamine trials; one additional completed trial was identified via ClinicalTrials.gov, yielding 8 trials included in the meta-analysis. Because two trials used multiple doses and one used a sequential-parallel design, the dataset yielded 16 esketamine-versus-placebo comparisons. Across trials, a total of 1,437 patients with MDD were randomised, with 802 allocated to adjunctive IN esketamine and 635 to adjunctive IN placebo. The dataset included studies of patients with TRD and studies of patients with MDD presenting with acute suicidal ideation; the extracted text reports 509 patients with MDD with SI but does not clearly state the exact number classified as TRD. All trials used MADRS as the primary outcome measure. At 24 hours after the first dose, adjunctive IN esketamine produced greater MADRS score reduction than adjunctive IN placebo: pooled SMD = 0.34 (95% CI = 0.11 to 0.46, P < .0001). In raw-score terms this corresponded to a 2.9-point greater mean reduction in MADRS for esketamine versus control. At study endpoint the pooled SMD was 0.26 (95% CI = 0.16 to 0.37, P = .004), corresponding to a 3-point greater mean MADRS reduction for esketamine; endpoint was 4 weeks in all studies except one, which had a 1-week endpoint. Remission outcomes favoured esketamine. The pooled RR for remission at 24 hours was 2.31 (P < .0001) with no statistically significant heterogeneity (heterogeneity test P = .7). At endpoint the RR for remission was 1.37 (P < .0001) with no statistically significant heterogeneity (P = .7). Mean remission rates were 16.7% for esketamine versus 7.3% for placebo at 24 hours (NNT = 10), and 27.9% versus 22.9% at endpoint (NNT = 20). Exploratory subgroup analyses by dose showed that all three studied IN doses (28 mg, 56 mg, 84 mg) were significantly more effective than placebo at 24 hours. At endpoint, only the 84 mg dose remained statistically superior to placebo; the 28 mg and 56 mg doses did not differ significantly from placebo at endpoint, although fewer comparisons were available for the lower doses (n = 3 for 28 mg, n = 4 for 56 mg, n = 7 for 84 mg). Meta-regressions did not identify dose as a predictor of SMD at 24 hours (P = .8) or at endpoint (P = .7), nor as a predictor of RR for remission at 24 hours (P = .4) or endpoint (P = .7). Likewise, study population (TRD versus SI) did not predict SMD at 24 hours (P = .8) or at endpoint (P = .9), nor RR of remission at 24 hours (P = .5) or endpoint (P = .7). Baseline severity was associated with a lower RR of remission at 24 hours (coefficient = -0.2; P = .04) but not at endpoint (coefficient = -0.6; P = .09). Baseline severity did not robustly predict SMD at 24 hours (P = .05) or at endpoint (P = .08). The authors report that all available efficacy data (SMDs of change) were obtained for the included trials, making the meta-analysis inclusive of existing published and identified completed studies.

Hock and colleagues interpret their pooled findings as evidence that adjunctive intranasal esketamine produces a statistically significant and clinically meaningful antidepressant effect that emerges rapidly, with a moderate pooled SMD of 0.34 on MADRS at 24 hours and a smaller SMD of 0.26 at endpoint. They note that an SMD of 0.34 corresponds to roughly one third of a pooled standard deviation in change scores (pooled standard deviations across studies were approximately 7 to 12 MADRS points). The smaller SMD at endpoint is attributed to larger measurement error at later timepoints across studies, despite similar absolute mean MADRS differences at 24 hours and endpoint. The authors place the IN esketamine effect in clinical context by pointing out that the derived effect sizes (SMD, NNT) at 24 hours are comparable to effect sizes seen for conventional antidepressants measured at endpoint in acute-phase trials, while emphasising that the present dataset involved patients with more difficult-to-treat presentations (TRD or imminent suicidal ideation) and used intranasal placebo plus active care as the comparator rather than an inert oral pill. These aspects, the investigators argue, support the clinical meaningfulness of a rapid antidepressant effect with IN esketamine. Exploratory and dose-stratified analyses provided limited evidence for a dose-response relationship: all three studied doses (28 mg, 56 mg, 84 mg) were superior to placebo at 24 hours, but only 84 mg remained superior at endpoint; meta-regressions did not identify dose as a significant predictor of pooled effect sizes. The authors acknowledge that sample size limits the strength of dose-response conclusions and cite mixed findings from IV ketamine studies as context for uncertainty about optimal dosing. Key limitations acknowledged include the relatively small overall sample size for robust dose-response assessment, the fact that all trials administered esketamine as an adjunct to an antidepressant (so monotherapy efficacy is untested), and restriction of trial populations to adults with MDD, limiting generalisability to adolescents or other mood disorders. The investigators conclude that adjunctive IN esketamine produced rapid and significant improvement in depressive symptoms at 24 hours and at study endpoint, but they call for further research to clarify dose-response relationships and broader clinical applicability.