Psychotic spectrum symptoms across the lifespan are related to lifetime suicidality among 147 patients with bipolar I or major depressive disorder

Suicide is one of the leading causes of preventable death, with more than 800,000 people dying by self-inflicted injuries every year [1].

In high-income countries, up to 90 % of people who die by suicide suffer from a mental disorder [2]. In particular, mood disorders are among the most important risk factors for suicidality [3‐6]. The lifetime risk for suicide is estimated to be around 4 % in patients with any mood disorder [7] and 8 % in people with bipolar disorder [1, 8, 9]. Importantly, the risk of suicidal behavior increases with comorbidity. Significantly higher risk of suicide has been in fact reported among individuals with more than one mental disorder, with odds of serious suicide attempts 89.7 times the odds of those with no psychiatric disorders [2, 10]. Moreover, recent data suggest that also subthreshold comorbidity might increase the risk of suicidal ideation and attempts [11‐15].

Several investigations in clinical and non-clinical samples have evaluated the relationship between psychotic symptoms and the risk of developing suicidal thoughts or behaviors. Recent epidemiological data have shown subthreshold psychotic experiences to be predictive of suicidal thoughts and attempts in non-clinical populations [16‐23]. Moreover, a number of studies have identified a correlation between positive psychotic symptoms, or positive psychotic symptoms-related distress, and suicide in patients with schizophrenia [24‐26]. As for mood disorders, some authors suggested that suicidal behaviors might be more frequent in patients with mood disorders with psychotic features with respect to those without [27], despite this has not been confirmed in all studies [28, 29]. Ran and colleagues [30] found higher levels of delusions in suicide attempters than in those without suicide attempts among 166 patients with affective disorders. Park et al. [31] reported the same result among 1183 subjects with depression. Suicidal ideation was found to be associated with psychotic features among 644 subjects with unipolar or bipolar II major depression [32]. Song and colleagues [33] found auditory hallucinations and the age of first psychotic symptoms to be significant predictors of suicide attempts among 212 subjects with bipolar disorder. Conversely, Pawlak and colleagues [34] failed to find an association between psychotic symptoms and suicide attempts among 597 subjects with unipolar and bipolar affective disorders. Consistently, in a recent register-based, nationwide, historical prospective cohort study, evaluating 34,672 severely depressed subjects followed at Danish psychiatric hospitals between 1994 and 2010, psychotic depression was not identified as an independent risk factor for suicide [35].

However, the extensive literature on suicide risk factors among subjects with mood disorders has mostly been focused on positive psychotic symptoms, while more subtle dimensions of the psychotic spectrum have not been hitherto carefully investigated.

The data were collected in 11 Italian Departments of Psychiatry located at nine sites: Pisa, Bari, Cagliari, Firenze, Milano, Sassari, Siena, Torino, and Udine.

To the best of our knowledge, this is the first study investigating the relationship between psychotic symptoms over the lifespan and lifetime suicidality among individuals with BD I or MDD by means of a spectrum approach. The majority of studies, in fact, explored the prevalence of suicidal thoughts or attempts comparing patients with the presence of psychotic symptoms such as delusions and hallucinations, retrieving conflicting evidences [32‐35].

Our results show no differences in lifetime suicidality between subjects with MDD and those with BD I. Moreover, we did not find a significant correlation between suicidality and the presence of psychotic features. According to DSM-IV criteria, these latter were assessed on the presence of delusion and/or hallucinations. However, evaluating the study sample by means of our spectrum instrument encompassing not only typical positive symptoms, such as hallucinations and delusions, but also atypical, subthreshold and negative manifestations, our results suggest that further dimensions of the psychotic spectrum occurring across the lifespan may have a significant relationship with lifetime suicidality. In particular, linear regression analyses indicated lifetime suicidality to be significantly related to the Paranoid, the Schizoid and the Interpersonal Sensitivity domain score of the SCI-PSY. The latter domain remained significantly correlated to lifetime suicidality even when inserted in a regression model that included the other SCI-PSY domains correlated to suicidality and the total depressive lifetime score of the MOODS-SR. While the final regression model showed to account only for a moderate percentage of variance in suicidality (about 30 %), the results showed that for one unit increase in Interpersonal Sensitivity, suicidality is expected to increase by .164 units.

Some previous studies reported a link between suicidality and interpersonal sensitivity. Interpersonal sensitivity has been found to discriminate between individuals engaging in self-harm and those who do not engage in such behaviors among 1986 Air Force recruits [40]. Engin et al. [41] found interpersonal sensitivity among the risk factors for suicidal thoughts in a sample of 1992 first-year university students. Consistently, Gupta et al. [42] reported that interpersonal sensitivity might represent the mediating factor between cutaneous body image dissatisfaction and suicidal ideation in a non-clinical sample of 312 subjects. As for clinical samples, Stepp et al. [43] evaluated the relationship between attachment styles and self-injury among 406 patients with a broad range of mental disorders, including both Axis I and Axis II disorders, highlighting that this relationship is arguably mediated by interpersonal sensitivity. We may argue that this latter may render individuals prone to feel easily hurt or rejected, and thus to avoid approaching others, having pessimistic expectancies about the outcome of their interpersonal interactions; as a consequence, they may experience feelings of personal inadequacy and negative emotional states, such as depression or anxiety. In addition, such avoidance may worsen social isolation and social exclusion that have been shown, in turn, to be associated with increased suicidal risk [44‐46].

Our data also showed a relationship between the Paranoid domain of the SCI-PSY and lifetime suicidality. An increase in paranoid behaviors has been indicated as a signal of risk of suicide among subjects with schizophrenia [47]. Consistently, the paranoid schizophrenia subtype has been repeatedly associated with an elevated risk of suicide among other types of schizophrenia [48, 49]. As far as we know, however, the relationship between paranoid symptoms and suicidality has not been specifically evaluated among subjects with BD I or MDD.

Further, we found a relationship between lifetime suicidality and the Schizoid domain of the SCI-PSY. This domain explores the lifetime occurrence of symptoms that are typical of autism, schizoidism and schizotypy. The comorbidity with cluster A personality disorders has been repeatedly shown to increase the risk for suicide in patients with different mental disorders, including mood disorders [50‐52]. On the other hand, growing data indicate that autistic traits may both represent an independent risk factor for suicide and increase the likelihood of engaging in more lethal suicidal behaviors in depressed patients [53, 54]. Such literature also appears consistent with recent reports indicating that subjects committing almost fatal suicide attempts show significantly lower self-disclosure, more schizoid tendencies and more feelings of loneliness and communication difficulties than those who did not report medically serious consequences [55‐57]. We might speculate that subjects with autism spectrum symptoms and those with schizoid or schizotypal traits share a common experience of interpersonal problems, lack of intimacy and social isolation [12]. Such symptoms are likely to lead to difficulties in accessing social support and to reduce the individual’s resilience to stressful life events [58], increasing feelings of helplessness and hopelessness and, ultimately, giving rise to suicidal thoughts and behaviors.

Several limitations of this study have to be acknowledged while interpreting our findings. First, we grouped together patients from both in- and outpatient treatment settings, which may have biased reports of lifetime suicidality. Second, the lack of information about age of onset of mood disorder, number of episodes, current treatment and the lack of assessment of Axis II comorbidity impaired us from analyzing the possible contribution of these factors to psychotic spectrum symptoms. Further, suicidality was assessed solely by selecting six questions of a 161-item instrument and grouping together very different aspects, from ideas that life is not worth living up to attempted suicide, preventing us from elucidating whether some dimensions of the psychotic spectrum specifically correlate with suicidal behaviors rather than with transient suicidal thoughts. Moreover, the use of a lifetime assessment did not allow us to establish whether the psychotic symptoms preceded, co-occurred or followed suicidal ideation or attempts. Last, the exclusion of individuals with severe medical and neurological disorders may have resulted in an underestimation of lifetime suicidality in patients with BD I or MDD.

In the context of the limitations listed above, we are reporting a significant association between lifetime suicidality and some dimensions of the psychotic spectrum occurring over the lifespan among patients with BD I or MDD. Our findings suggest the potential usefulness of a spectrum approach in the assessment of suicidality among patients with mood disorders. On the other hand, they highlight the importance of a dimensional approach to rate severity for the core symptoms of schizophrenia not only among patients with schizophrenia spectrum and other psychotic disorders [59], but also among those with mood disorders. Further research is needed to corroborate our results and to elucidate what implications, if any, the findings have on the assessment and treatment of patients with BD I or MDD.