Intranasal esketamine in treatment-resistant depression with and without comorbid borderline personality disorder: A multicenter real-world longitudinal study

Treatment-resistant depression (TRD) remains a severe condition with persistent symptoms, functional impairment, and elevated suicide risk despite adequate antidepressant treatment. Intranasal esketamine has shown rapid antidepressant effects in trials and meta-analyses, but there is still uncertainty about how it performs in routine care for patients with complex comorbidity. In particular, borderline personality disorder (BPD) frequently co-occurs with TRD, is associated with affective dysregulation, impulsivity, self-harm risk, and poorer antidepressant outcomes in some earlier studies, yet its influence on esketamine response in real-world settings had not been clearly established. Beyond mood symptoms, anxiety, impulsivity, and cognitive difficulties may also affect prognosis, but their change during esketamine treatment had not been well characterised in this population. Mazzoni and colleagues set out to compare six-month depressive symptom trajectories in outpatients with TRD receiving intranasal esketamine, contrasting those with comorbid BPD against those with no personality disorder. They also aimed to examine response and remission rates, changes in anxiety and impulsivity, and exploratory change in global cognitive functioning. The study further explored whether baseline impulsivity moderated antidepressant response, with the broader goal of informing the use of esketamine in routine care for clinically complex patients.

The researchers conducted a multicentre, prospective, real-world observational study across four Italian psychiatric centres. Consecutive outpatients aged 18-65 years who initiated intranasal esketamine between 1 September 2024 and 30 September 2025 were considered for inclusion. TRD was defined clinically as failure to achieve a meaningful response to at least two adequate antidepressant trials in the current episode. Participants were allocated to a BPD group or a no personality disorder group, with 45 patients in each. BPD diagnoses were established using the Structured Clinical Interview for DSM-5 Personality Disorders (SCID-5-PD). Three patients in the BPD group also had comorbid histrionic personality disorder; no other personality disorder diagnoses were identified. Patients with current or recent psychosis, substance use disorder within the prior three months, or medical/cognitive conditions preventing reliable assessment were excluded. Esketamine nasal spray was given under medical supervision in routine practice. Treatment began with a four-week induction phase of twice-weekly 56 mg dosing, with the option to increase to 84 mg, followed by a maintenance phase lasting weeks 5-24 in which 56-84 mg was administered weekly or fortnightly depending on response and tolerability. Oral antidepressants were continued without washout, other psychotropics were allowed, and baseline psychotherapy was recorded as present or absent rather than standardised. Patients were monitored for at least two hours after each administration, and vital signs and adverse events were recorded throughout follow-up. Assessments took place at baseline, 2 weeks, 1 month, 3 months, and 6 months. The primary outcome was depressive severity measured by the Montgomery-Åsberg Depression Rating Scale (MADRS). A response was defined as a 50% or greater reduction in MADRS score, and remission as a MADRS score of 10 or less at 6 months. Secondary outcomes included anxiety on the Hamilton Anxiety Rating Scale (HAM-A), impulsivity on the Barratt Impulsiveness Scale-11 (BIS-11), and global cognition on the Montreal Cognitive Assessment (MoCA), which was treated as exploratory because it is a brief screening tool. Analyses used mixed-design repeated-measures models, non-parametric tests for within-group change, chi-squared and odds ratio comparisons for categorical outcomes, and sensitivity analyses with generalised estimating equations adjusting for age, sex, baseline psychotherapy, benzodiazepine use, mood stabiliser use, and lifetime hypomanic episodes. No formal sample size calculation was performed; the sample was determined by consecutive eligible patients during the study period.

The sample included 90 outpatients with TRD, evenly split between the BPD and no personality disorder groups. Baseline characteristics differed between groups: the BPD group was younger, more often female, had higher baseline MADRS and BIS-11 scores, more lifetime suicide attempts, more frequent lifetime hypomanic episodes, and greater use of benzodiazepines, mood stabilisers, and psychotherapy. The no personality disorder group had a higher prevalence of family history of mood disorders. Depressive symptoms improved substantially over time in both groups. MADRS scores fell significantly from baseline to 6 months, and the mixed-model analysis showed a large time effect with a significant time × BPD interaction, indicating faster early improvement in the BPD group. The BPD group had higher baseline depression severity and a lower MADRS score at 1 month, but no between-group differences were seen at later time points. Sensitivity analyses adjusting for baseline psychotherapy and concomitant medications, and analyses excluding the 10 participants with lifetime hypomanic episodes, gave similar results. The BPD group also showed greater overall improvement from baseline to 6 months. At 6 months, remission was achieved by 48 of 90 participants (53.3%). Remission did not differ significantly by BPD status: 57.8% in the no personality disorder group versus 48.9% in the BPD group. Response rates, defined as at least a 50% reduction in MADRS, were higher in the BPD group from 1 month onwards and remained higher at 6 months, when 97.8% of the BPD group responded compared with 77.8% of the no personality disorder group. This difference remained after excluding participants with lifetime hypomanic episodes and after stratifying by baseline psychotherapy. In adjusted logistic regression, BPD status was not associated with remission at 6 months. Exploratory secondary outcomes showed that anxiety decreased significantly during follow-up, with a greater reduction in the BPD group despite their higher baseline anxiety. Impulsivity decreased across all BIS-11 subscales over time. MoCA scores were largely in the normative range at baseline and did not worsen over 6 months; the mean score increased modestly from 27.34 to 28.88, which the authors treated as exploratory. No serious adverse events occurred, no participants discontinued treatment, and 22 participants (24.4%) reported at least one adverse event. The most common were nausea (7.8%) and dizziness/vertigo (5.6%); dissociative symptoms and paresthesia each occurred in 3.3% of participants, while vomiting, transient blood pressure increase, depersonalisation, mild euphoria, and incontinence were each reported in one participant.

The authors interpret the findings as showing that intranasal esketamine produced a rapid and sustained reduction in depressive symptoms over six months in routine care, with a large overall time effect. They argue that comorbid BPD did not predict worse outcomes, because depressive trajectories converged between groups and remission rates were similar at 6 months. The higher response rate in the BPD group from month 1 onwards is presented as evidence that response kinetics may be faster in this subgroup, while the similar remission rates are attributed to differences in baseline severity and the fact that response is defined proportionally whereas remission depends on an absolute symptom threshold. The authors place these findings in the context of earlier work suggesting that personality disorder comorbidity can be associated with poorer antidepressant outcomes, but note that the present results do not support excluding patients with BPD from esketamine treatment. They say the findings were robust in sensitivity analyses, including analyses excluding participants with lifetime hypomanic episodes and models adjusting for psychotherapy and medication use. They also suggest, cautiously, that BPD-related affective dysregulation and impulsivity might relate to glutamatergic and synaptic plasticity pathways targeted by ketamine derivatives, although they describe this as speculative. The authors emphasise several limitations. The observational design and lack of a control group mean the study cannot establish causality or fully separate treatment effects from regression to the mean or concurrent care. The study did not measure borderline symptom severity dimensionally, so subthreshold personality difficulties in the comparison group may have gone undetected, and the BIS-11 does not specifically assess emotion-driven impulsivity such as negative urgency. All participants with lifetime hypomanic episodes were in the BPD group, which may confound interpretation despite consistent sensitivity analyses. Concomitant pharmacotherapies, psychotherapy, and service organisation were not standardised, and selection bias is possible because there were no dropouts and no formal log of eligible non-initiators. The sample size also limited power for small differences in categorical outcomes. For practice, the authors suggest a pragmatic approach in which BPD status should inform monitoring and integrated care rather than restricting access to esketamine. They also note that future studies should include borderline-specific dimensional measures such as the BSL-23, along with more comprehensive neurocognitive testing and functional outcomes, to clarify whether borderline symptomatology itself changes with treatment.