Transdiagnostic suicidality in depression: More similar than different?

While extensive work has been conducted regarding risk factors for suicide attempts and deaths, clinical prediction of suicide-related behaviour at an individual level remains an unmet challenge.1 Given this reality, some have recommended a shift from risk prediction to risk management and a focus on formulation and therapeutic engagement.2 There is a paucity of prospective data directly comparing suicidal behaviour across commonly encountered psychiatric conditions and their symptom severity. Major depressive disorder (MDD), bipolar disorder (BD) and borderline personality disorder (BPD) are known to be associated with higher risk for suicidal ideation, suicide attempt and death by suicide than the general population, especially during the course of major depressive episodes (MDEs).3 Mood disorders and BPD also frequently co-occur, with about 20% of adults with BD having comorbid BPD and an even higher frequency of BPD traits.4, 5 The dynamic interplay between depressive symptom fluctuation, baseline BPD traits and emergence of suicidal behaviour is of great clinical relevance. In this edition of the journal, Isometsä et al. examine suicide attempts and suicidal ideation prospectively in a cohort of outpatients with an MDE within the context of MDD, BD or comorbid with BPD.6 Findings include higher rates of suicide attempts in those with BPD, a correlation between suicide attempts/suicidal ideation and BPD symptom severity as measured by the borderline personality disorder severity index (BPDSI) at baseline, and a correlation between severity of suicidal ideation and depressive symptom severity. Although the rates of suicidal ideation and attempts are lower over the period of follow up in the BD group than in the other two groups, the link between depression severity changes and suicidality is greater in this group. Hopelessness is found to be a transdiagnostic predictor of suicidal ideation. Interestingly, non-suicidal self-injury (NSSI) is also high in the MDD group, and not only in the BPD group. NSSI is lower in the BD group. This work provides an important contribution to the field based on the prospective design and direct comparison of uniform and relevant outcomes between these three commonly encountered clinical groups. The sample is comprised of treatment-seeking subjects in an outpatient clinic, with on average moderate depression severity at baseline. While the strictly outpatient setting may reduce generalizability to other settings, the outpatient setting is particularly salient as past work has demonstrated the most common mental health contact prior to suicide completion is an ambulatory care visit, rather than emergency room visit or psychiatric hospitalisation.7 Furthermore, the rate of lifetime suicide attempts in this group is 15.6% for the MDD group, 30% for the BD group and 60% for the BPD group, which suggests generalizability to most clinical populations.8 The authors also highlight several limitations including the amalgamation of suicidal thoughts and thoughts of self-harm within the suicidal ideation measure. The lethality degree of the suicide attempts reported is also unknown and is obtained by self-report. The Isometsä et al. study supports findings from prior studies which demonstrated that BPD features are correlated with higher suicidal ideation in those with mood disorders, and that depressive symptom severity and hopelessness are associated with suicidal ideation in those with BPD.9, 10 The dynamic mapping of suicidal ideation and depressive symptom severity over time, particularly hopelessness, is of great interest. The authors note that the choice of a 6-month prospective period links with the mean duration and remission of an MDE. A prospective study as part of the Systematic Treatment Enhancement Program for Bipolar Disorder trial (STEP-BD) also found an increase in suicidal ideation and depressive symptoms, mainly depressed mood and anhedonia, prior to a suicide attempt.11 A prospective study in MDD showed a decline in depressive symptoms preceding a decline in SI, with a median period of around 2 months for SI resolution.12 The exact duration and trajectory of the acute period of crisis and increased risk may be difficult to ascertain at an individual level given the known fluctuation of suicidality. The correlation between changes in specific depressive symptoms and suicidality versus absolute severity of depressive symptoms would also warrant further study, especially as Isometsä et al. report that the magnitude of this change in severity of overall depressive symptoms may be more closely related to suicidal ideation shifts within BD. This may relate to a closer link between depression and suicidality in BD due to hopelessness in the context of a chronic illness. Alternatively, other factors may more strongly mitigate the relationship between depression and suicidality in those with non-BD MDE. Given that suicide is a rare event when measured over relatively short durations, suicide attempts and measures of suicidal ideation and self-harm may be reasonable proxies to understand death by suicide, although the relationship is not a direct one. Transdiagnostic studies have identified past suicide attempts as among the most important risk factors for future attempts and eventual suicide death.13 This may vary by disorder and as a function of comorbidities. Chronic suicidal ideation and recurrent self-harm are core features of BPD. Self-harm may also predispose to further suicidal behaviour by enhancing readiness for suicide due to factors such as tolerance for physical pain.14 Most patients with BPD have, on average, three lifetime suicide attempts, and the lifetime risk of death by suicide is estimated at 5%–10%.15 Previous BPD studies did not find a correlation between number of past suicide attempts and death by suicide, and identified that other factors such as impulsivity, hostility and comorbidities play prominent roles.16 It is also plausible that the combination of BPD and depression may alter the relationship between suicide attempts and completion. In BD, comorbid BPD was associated with suicide attempts, but not with suicide completion.17 The cumulative effect of longer periods of elevated suicidal ideation and suicidal behaviour in those with mood disorders with comorbid BPD or BPD traits may also contribute to hopelessness about the future and a sense of entrapment in the current pain. This course of illness contrasts significantly with that of individuals with intermittent mood episodes and inter-episode periods of symptomatic and perhaps even functional recovery. In addition, it is hypothesised that patients with BPD are at higher risk of death by suicide later in the course of illness while being at higher risk for suicide attempts and frequent self-harm earlier in the course of illness.18 This speaks to the importance of hopelessness as a driver of suicidal behaviour, which is further highlighted by the findings from the Isometsä et al. study. What are salient points for clinicians to consider from this extant literature? First, among patients presenting with an MDE, clinicians should consider screening for BPD and rating BPD symptom severity, even in those with only borderline personality traits. Exploring symptoms more typically associated with BPD such as impulsivity, self-harm and interpersonal instability in those with mood disorders may further help in understanding the drivers of a patient's suicidal behaviour. Given that the present study also found increased rates of NSSI in MDE without BPD, screening for self-harm in this population may be warranted. Furthermore, patients with mood disorders may benefit from interventions aimed at BPD symptoms such as enhancing coping skills, impulsivity management and work towards stable connections and relatedness to others. These may in turn modify risk for suicidal behaviour. Conversely, clinicians should focus on assessing closely those with BPD for the presence of a depressive episode. Indeed, clinicians may often focus on BPD symptoms in those with a history of the diagnosis given the inherent externalising nature of BPD symptoms. Patients with BPD often present with recurrent suicide attempts or ideation, and identification of increased risk for suicide beyond the chronic elevated risk may prove particularly daunting. In addition, it can often be challenging to identify an MDE cross-sectionally in patients with severe BPD due overlapping symptoms of psychological pain, distress, and suicidal ideation. Longitudinal assessments using depression rating scales and prospective mood rating in the outpatient clinical, as done by Isometsä et al. could be of great benefit. Ultimately, the goals of clinical encounters with suicidal patients are to understand the patient's unique vulnerabilities to suicidal behaviour, instil a sense of hope and connection, and co-create a management plan. By prospectively linking suicidal ideation and attempts during depressive episodes to BPD symptom severity, fluctuation of depression severity and hopelessness, Isometsä et al. eloquently confirm the importance of a multidimensional and transdiagnostic assessment of our patients. Dr. Toma reported salary support from the Sunnybrook Foundation. Dr. Mitchell reported receiving grants from the American Foundation for Suicide Prevention, the TD Pooler Fund, and the Sunnybrook Foundation, and an Academic Scholar Award from the Department of Psychiatry at Sunnybrook Health Sciences Centre and the University of Toronto, and an honourarium from Medscape outside of the submitted work. Dr. Sinyor reported salary support through an Academic Scholar Award from the Departments of Psychiatry at Sunnybrook Health Sciences Centre and the University of Toronto. Dr. Schaffer reported salary support through an Academic Scholar Award from the Departments of Psychiatry at Sunnybrook Health Sciences Centre and the University of Toronto. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/acps.13600. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.