Esketamine in treatment-resistant depression with and without comorbid borderline personality disorder: A real-world longitudinal study of suicidal ideation and self-harm

Treatment-resistant depression (TRD) is described as a severe and disabling condition that is often accompanied by suicidal ideation and self-harm. The paper notes that these risks are especially pronounced when borderline personality disorder (BPD) is also present, because BPD is associated with emotional dysregulation, impulsivity, recurrent self-injury and suicide attempts. The authors also point out that intranasal esketamine has shown antidepressant effects in TRD and rapid anti-suicidal effects in some prior ketamine-based research, but that patients with substantial personality pathology have generally been underrepresented in pivotal trials. As a result, real-world evidence on suicidality and self-harm during esketamine treatment, particularly in people with comorbid BPD, remains limited. The study set out to examine longitudinal changes in suicidal ideation, suicidal behaviour and deliberate self-harm during intranasal esketamine treatment in a naturalistic TRD cohort, with explicit comparison of patients with and without comorbid BPD. A further aim was to explore how trait impulsivity related to suicidality and self-harm at baseline and across follow-up. In doing so, the authors aimed to address a clinically important gap in routine-care evidence for a high-risk subgroup that is often excluded from or underrepresented in clinical trials.

Raffone and colleagues conducted a multicentre prospective observational study in routine outpatient care across four Italian psychiatric services. Patients who consecutively began intranasal esketamine for TRD between 1 September 2024 and 30 September 2025 were followed for 6 months with assessments at baseline, 2 weeks, 1 month, 3 months and 6 months. The sample comprised 90 outpatients aged 18-65 years with a current major depressive episode and TRD, defined as non-response to at least two antidepressant treatments of adequate dose and duration in the ongoing episode. Forty-five participants had comorbid BPD and 45 had no personality disorder. Personality disorders were assessed with the Structured Clinical Interview for DSM-5 Personality Disorders. The BPD subgroup was mainly diagnostically homogeneous: 42 participants had BPD only, and 3 had BPD with comorbid histrionic personality disorder. Current or recent psychotic disorders, substance use disorder within the previous 3 months, and medical conditions preventing reliable assessment were excluded. A lifetime history of hypomanic episodes consistent with bipolar diathesis was recorded in 10 participants, all in the BPD group, but this was not an exclusion criterion. All participants completed the planned follow-up, with no dropouts. Intranasal esketamine (Spravato) was administered on site under supervision according to the product label and local practice. The induction phase involved twice-weekly dosing for weeks 1-4, starting at 56 mg and titrating to 84 mg if needed, followed by weekly dosing during weeks 5-8, then weekly or fortnightly dosing during weeks 9-24 at 56-84 mg according to response and tolerability. Oral antidepressants were continued as clinically indicated, and other psychotropic medication and psychotherapy were permitted as part of standard care. Because session-level administration logs were not retained, the authors reconstructed cumulative 6-month exposure under lower-bound and upper-bound label-concordant assumptions. Suicidal ideation and suicidal behaviour were assessed with the Columbia-Suicide Severity Rating Scale, using binary yes/no indicators at each time point. Deliberate self-harm was measured with the Deliberate Self-Harm Inventory, from which the authors derived the number of self-harm methods endorsed and the frequency of self-harm episodes. Trait impulsivity was assessed with the Barratt Impulsiveness Scale, which yields total and subscale scores for attentional, motor and non-planning impulsivity. Because several outcomes were non-normally distributed, the authors used Friedman tests for within-subject longitudinal change, Wilcoxon signed-rank tests with Bonferroni correction for post hoc comparisons, and Spearman correlations for associations. Exploratory analyses compared recorded doses and reconstructed cumulative exposure between groups with Mann-Whitney U tests and examined relationships between cumulative dose and change in suicidality/self-harm outcomes. The study was approved by the Ethics Committee of Pavia, and all participants gave written informed consent.

At baseline, participants with comorbid BPD were younger, more often female, and had more previous suicide attempts than those without personality disorder. They also had higher impulsivity scores. In terms of co-treatment, the BPD group was more likely to be receiving benzodiazepines, mood stabilisers and psychotherapy; antidepressant and antipsychotic co-prescription rates were similar between groups. Suicidal ideation declined significantly over time, with reductions from baseline becoming evident from month 1 onwards and remaining significant through month 6. Suicidal behaviour also decreased across follow-up (Friedman χ²(4) = 43.46, p < 0.001; Kendall's W = 0.12), with significant reductions versus baseline from month 1 onwards, while the baseline-to-2-week change was not significant after correction. Deliberate self-harm improved robustly. The number of self-harm types decreased across assessments (χ²(4) = 134.76, p < 0.001; W = 0.37), and the frequency of self-harm episodes also declined markedly (χ²(4) = 171.98, p < 0.001; W = 0.51). The baseline-to-6-month effect size for episode frequency was large (r = 0.82). In the BPD subgroup, self-harm episode frequency fell substantially, from a mean of 30.8 at baseline to 2.4 at 6 months, whereas the no-personality-disorder group showed very low baseline levels and a floor effect. Despite these different starting points, both groups showed a downward pattern over time. Baseline trait impulsivity was moderately to strongly correlated with self-harm measures and more modestly correlated with suicidal ideation and behaviour, indicating alignment between impulsivity and risk at the start of treatment. By 6 months these correlations were weaker, which the authors attribute mainly to reduced variability and floor effects as most patients had very low or absent self-harm and suicidal ideation. Safety findings were reassuring in this cohort. Twenty-two patients (24.4%) reported at least one adverse effect, most commonly nausea (7.8%) and dizziness (4.4%). There were no serious adverse events, no treatment discontinuations, and no observed increases in suicidality. No suicide attempts were recorded in either subgroup during the 6-month follow-up. Recorded esketamine doses at the follow-up visits were similar between groups. In the full sample, the mean recorded dose was 72.49 mg at T1, 78.71 mg at T2, 81.20 mg at T3 and 80.58 mg at T4. Reconstructed cumulative 6-month exposure was 1549.02 mg under the lower-bound scenario and 2194.89 mg under the upper-bound scenario. Neither the recorded doses nor the reconstructed cumulative exposure differed significantly by BPD status, and cumulative dose was not significantly associated with change in suicidal ideation, suicidal behaviour, self-harm types or self-harm episode frequency.

The authors interpret the study as showing that, during 6 months of routine adjunctive intranasal esketamine treatment for TRD, suicidal thoughts and self-harm behaviours fell in a sustained way, with significant improvement appearing from month 1 onwards. They emphasise that the pattern was similar in patients with and without comorbid BPD, even though the BPD group had much higher baseline self-harm frequency. They also stress that no increase in suicidality was seen in either subgroup. The paper argues that the time course is clinically informative: changes in suicidal ideation and deliberate self-harm appeared earlier, whereas reductions in suicidal behaviour and self-harm episode frequency were clearer after the induction phase. The authors suggest this may fit frameworks that distinguish suicidal thinking from the transition to action. They also report no clear dose-response relationship between cumulative esketamine exposure and improvement, but note that these analyses were exploratory and limited by the lack of session-level dosing data. In relation to previous research, the authors place their findings within the broader ketamine/esketamine literature, including controlled ketamine studies showing rapid anti-suicidal effects and Phase III esketamine trials in major depressive disorder with acute suicidal ideation or behaviour. They note that their real-world cohort extends earlier work by tracking both suicidal ideation/behaviour and deliberate self-harm across multiple time points, and by including a BPD subgroup that is usually underrepresented in trials. They also state that the results should not be taken to mean esketamine is a stand-alone treatment for BPD; rather, it may have a role as a rapid, monitored adjunct within integrated care when TRD co-occurs with chronic suicidality or self-injury. The authors acknowledge several limitations. The observational design without a comparator group means the improvements cannot be attributed definitively to esketamine, since regression to the mean, symptom fluctuation, and increased clinical contact could also have contributed. Background treatments were not protocolised, which raises confounding concerns. Suicidality was measured in binary form, limiting sensitivity to severity or persistence, and self-harm relied on self-report. The no-personality-disorder group had very low baseline self-harm, creating floor effects. Missing data affected some analyses, the sample was relatively small, the BPD subgroup was largely homogeneous rather than representing the wider Cluster B spectrum, and no objective outcomes such as emergency visits or clinician-rated self-harm severity were included. The authors also note that the age range was limited to 18-65 years by the organisation of the participating services. In terms of implications, the authors say the findings support intranasal esketamine as part of a comprehensive treatment plan for TRD patients with suicidality and self-harm, including those with BPD, alongside structured crisis planning and evidence-based psychotherapy such as dialectical behavioural therapy. They argue that a BPD diagnosis alone should not automatically exclude patients from receiving esketamine within a monitored pathway. They also call for controlled longitudinal studies that can separate the effects of esketamine from concurrent care, include dimensional and objective risk measures, assess suicide attempts and emergency department visits, and examine whether personality traits influence durability of benefit.

The authors conclude that, in this naturalistic TRD cohort, intranasal esketamine was associated with sustained reductions in suicidal ideation, suicidal behaviour and deliberate self-harm over 6 months, with significant improvement evident from month 1 onwards. They highlight that clinically meaningful improvements were also seen in patients with comorbid BPD, that baseline impulsivity became less strongly linked to risk indices at follow-up, and that the exploratory cumulative-dose analyses did not indicate a clear dose-response relationship. They state that these findings support the possible role of esketamine-based treatment within risk-focused management of complex TRD presentations and reinforce the need for controlled, longitudinal studies including patients with BPD.