In Review

Paul S Links, MD1, Brent Gould, MD2, Ruwan Ratnayake3

 

 

Objective: This paper has 3 objectives. First, we review the epidemiologic evidence for the association between suicidal behaviour and suicide in individuals diagnosed with anti- social, borderline, or narcissistic personality disorder. Second, we examine whether any potentially modifiable risk factors are associated with these diagnoses, based on existing empirical evidence. Last, we discuss clinical approaches to assessing youth with antisocial, borderline, or narcissistic personality disorder presenting at risk for suicide.

Method: We reviewed the English-language literature for the last 12 years (from January 1, 1991, to December 31, 2002), using as search terms the names of the 3 disorders, as well as the key words suicide, suicidal behaviour, youth, and adolescents.

Results: Patients with antisocial or borderline personality disorder are likely to be at in- creased risk for suicidal behaviour when they demonstrate such comorbid disorders as major depressive episodes or substance abuse disorders, when they experience recent nega- tive life events, or when they have a history of childhood sexual abuse.

Conclusions: For patients with antisocial personality disorder, the risk of violence has to be judged in addition to the risk of suicide or self-harm. For patients with borderline per- sonality disorder, interventions are determined by differentiating acute-on-chronic from chronic risk of suicidal behaviour. Finally, patients with narcissistic personality disorder can be at high risk for suicide during periods when they are not suffering from clinical de- pression. These episodes can seem to be unpredictable.

(Can J Psychiatry 2003;48:301–310)

Information on author affiliations appears at the end of the article.

 

Clinical Implications

• Youth diagnosed with antisocial, borderline, or narcissistic personality disorders are at in- creased risk for suicide and suicidal behaviour.
• Patients with antisocial or borderline personality disorder are likely to be at increased risk for suicidal behaviour when they demonstrate comorbid disorders, such as major depressive epi- sodes or substance abuse disorders, when they experience recent negative life events, or when they have a history of childhood sexual abuse.
• Clinical approaches to assessing youth with antisocial, borderline, or narcissistic personality disorders are provided.

 

Limitations

• The review was restricted to English-language publications from the last 12 years.
• The clinical assessment approaches are not yet supported by empirical evidence.

 

Suicide continues to be a leading cause of death among youth in Canada. Statistics Canada has reported that, in 1999, a total of 558 suicides were recorded among youth aged 15 to 24 years (468 male youths and 90 female youths) (1). The male-to-female ratio was approximately 5 to 1. Robins and colleagues have already demonstrated the strong associa- tion between mental disorders and suicide (2), but the relation between suicidal behaviour and personality (Axis II) disor- ders has only been systematically studied within the last 2 de- cades. This paper has 3 objectives. First, we review the epidemiologic evidence for the association between suicidal behaviour and suicide in individuals diagnosed with anti- social, borderline, or narcissistic personality disorder. Second, based on existing empirical evidence, we examine whether any potentially modifiable risk factors are associated with these diagnoses. Last, we discuss clinical approaches to assessing youths with antisocial, borderline, or narcissistic personality disorder presenting at risk for suicide.

For this review, we define suicidal behaviour as having 3 com- ponents: suicide, suicide attempts, or self-injurious behav- iours. We have adopted the definitions of O’Carroll and others, who define suicide as self-injurious behaviour with a fatal outcome for which there is evidence (either explicit or implicit) that the individual intended at some (nonzero) level to kill himself or herself (3). A suicide attempt is defined as self-injurious behaviour with a nonfatal outcome for which there is evidence (either explicit or implicit) that the individ- ual intended at some (nonzero) level to kill himself or herself. We have also employed a definition of self-injurious behav- iour not intended to be fatal: Simeon and Favazza define self-injurious behaviour as all behaviours that involve delib- erate infliction of direct physical harm to one’s body with zero intent to die as a consequence of this behaviour (4). Our re- view reports on estimates of the risk of suicide and suicide at- tempts, based on existing literature. We reviewed the English-language literature for the last 12 years (from 1991 to 2003), using as search terms the names of the 3 disorders, as well as the key words suicide, suicidal behaviour, youth, and adolescents. In particular, we focused on research that exam- ined potential risk factors in youth with these diagnoses, com- pared with other psychiatric disorders. The paper’s final section describes clinical approaches to patients with anti- social, borderline, or narcissistic personality disorder who are at risk for suicidal behaviour. These observations are based on clinical experience and not on empirical evidence. They do not replace the need for a comprehensive suicide risk assess- ment based on the formats described in the following papers:

• “An Evidence-Based Approach to Evaluating and Managing Suicidal Emergencies” (5)
• “Suicide Assessment: An Overview and Recommended Protocol” (6)

• “The Emergency Interview” (7)
• “The Assessment, Management, and Treatment of Suicidality: Toward Clinically Informed and Balanced Standards of Care” (8)
• “Assessment of Suicide Risk” (9)
• “A Clinical Model for Suicide Risk Assessment” (10)

For this review, we selected the age range 15 to 24 years. This coincides with the age range for which nations report mortal- ity rates to the World Health Organization (11). For a mean- ingful discussion of the suicide risk associated with personality disorders, one must by definition focus on youth in late adolescence and early adulthood (12): while there is ev- idence that younger adolescents can demonstrate clinical fea- tures of a personality disorder, there is debate about whether these features remain stable over time (13,14).

 

Epidemiologic Evidence

The rates of personality disorders among adolescents who died by suicide have been studied. In the Comprehensive Psy- chological Autopsy Study in Finland (15), Marttunen and oth- ers estimated that 17% of the adolescents aged 13 to19 years who died by suicide met criteria for conduct disorder (CD) or antisocial personality disorder (APD). When Marttunen and others examined adolescents with nonfatal suicidal behav- iour, approximately 45% of male adolescents and 33% of fe- male adolescents were characterized by antisocial behaviour (16). In other research, suicidal behaviour was found to be higher among adolescents with CD than in the comparison groups, even after controlling for major depression (17,18). Beautrais and others studied individuals who had made medi- cally serious suicide attempts and compared them with sub- jects from the community (19). After controlling the data for the intercorrelations among mental disorders, these research- ers found that the risk of a serious suicide attempt was 3.7 times higher for individuals with APD than for those without the disorder. When they examined men under age 30 years, the risk of a serious suicide attempt was almost 9 times higher among individuals with APD than in those without the disor- der. For women, the risk of a serious suicide attempt was 2.3 times higher in individuals with APD than in those without.

A few studies have documented the lifetime risk of suicide in samples of individuals with APD. In a 5-year follow-up of a small sample of persons with APD (n = 59), Maddocks esti- mated a 5% lifetime risk of suicide (20). Laub and Vaillant ex- amined causes of death of 1000 delinquent and nondelinquent boys followed from ages 14 to 65 years (21). Deaths due to vi- olent causes (that is, accident, suicide, or homicide) were sig- nificantly more common in delinquent, compared with nondelinquent, boys; however, equal proportions of both groups died by suicide.

Patients with borderline personality disorder (BPD) represent 9% to 33% of all suicides (22,23). Bongar and others studied patients with chronic suicidality who made 4 or more visits in a year to a psychiatric emergency room; most often, these pa- tients met criteria for BPD (24). These patients accounted for over 12% of all psychiatric emergency room visits during the year studied. Crumley has shown a high incidence of BPD in the adolescents and young adults aged 15 to 24 years who en- gage in suicidal behaviour (25). Paris and colleagues indi- cated that this diagnosis significantly increases the risk of eventual suicide (26). Depending on the study, the lifetime risk of suicide among patients with BPD is between 3% and 10% (27–32). Those at highest risk appeared to be young, ranging from adolescence into the third decade (33–35), which likely reflects a decrease in symptom severity later in adulthood for most patients (25,35–38). The high rates of sui- cidal behaviour in patients with BPD are reflected by the in- clusion of recurrent suicidal behaviour, gestures, threats, or self-mutilating behaviour as diagnostic criteria in the DSM-IV (12)—although some would say they are a result of this inclusion. A history of suicidal behaviour is found in 55% to 70% of individuals with a personality disorder (39–41) and in 60% to 78% of individuals with BPD (42–45). Suicides by individuals with BPD may carry an extra burden for survivors: Runeson, Beskow, and Waern found that 44% of suicides by patients with BPD were witnessed, compared with 17% of suicides by patients with other diagnoses (46).

Narcissistic personality disorder is an uncommon diagnosis in community samples, compared with APD and BPD. In a sur- vey of a community sample in Baltimore, Maryland, approxi- mately 0.1% of the sample met criteria for the disorder (47). For the most part, individuals with narcissistic personality are also uncommon in clinical settings. They probably make up less than 1% of outpatient samples, and most of the literature regarding patients suffering with narcissistic personality dis- order is based on clinical experience and theoretical formula- tions, rather than on empirical evidence (48). Few data exist regarding the risk of suicide in individuals with narcissistic personality disorder. It is infrequently identified in samples of suicide victims studied according to the psychological au- topsy method. However, Apter and others studied 43 consecu- tive suicides that occurred among Israeli males aged 18 to 21 years, during compulsory military service (49). They carried out psychological autopsies, using preinduction assessment information, service records, and extensive post-mortem interviews. Based on this methodology, the most common Axis II personality disorders were schizoid personality disor- der in 16/43 (37.2%) and narcissistic personality disorder in 10/43 (23.3%). Stone’s extensive follow-up study of 550 pa- tients admitted to the general clinical service of the New York State Psychiatric Institute provided some information on this outcome for individuals hospitalized with the diagnosis of narcissistic personality disorder (50). According to the 15-year follow-up, patients with narcissistic personality dis- order or traits were significantly more likely to have died by suicide, compared with patients who did not have narcissistic personality disorder or traits (14% vs 5%; P < 0.02).

 

Risk Factors Related to Antisocial, Borderline, and Narcissistic Personality Disorder

Many studies have identified factors at a population level that alone or in combination increase the risk of suicide. Although extrapolating these risk factors to an individual allows for risk categorization, it does little to predict which individual will commit suicide, and when. “The goal of a suicide assessment is not to predict suicide, but rather to place a person along a pu- tative risk continuum, to appreciate the bases of suicidality, and to allow for a more informed intervention” (6, p 4). Many risk factors are fixed (for example, age, race, and sex), provid- ing little opportunity to intervene. However, several of the most significant risk factors are modifiable. Allen captured the concept eloquently in his statement “Suicide is almost al- ways the catastrophic result of inadequately treated psychiat- ric illness” (51) and suggested concentrating on modifiable risk factors. Therefore, this review discusses risk factors that place youth with antisocial, borderline, or narcissistic person- ality disorders at a higher risk, relative to other individuals with like disorders, or places them at higher risk, relative to other times in the course of their illness. In addition, we dis- cuss purportedly modifiable risk factors as they may present opportunities for interventions.

Tables 1 and 2 summarize the published studies from the Eng- lish literature in the last 12 years that discuss risk factors for suicidal behaviour in youth and adults suffering from APD or BPD, compared with other psychiatric disorders. By far, most of the studies have focused on subjects with BPD, with little or no research on youth with narcissistic personality disorder. We review each of the major risk factors and discuss in some detail findings from key studies—those employing carefully characterized comparison groups and controlling for potential confounding factors. We limit discussion to a review of the risk factors that are felt to place patients at higher risk when present, relative to points in the course of their disorder when these factors are not present and relative to patients without these factors. In addition, these factors are felt to be poten- tially modifiable, although the discussing the interventions is beyond the scope of this paper.

Comorbid Disorders Comorbidity is found to be higher in adolescents than in adults, and the presence of 2 or more psychiatric disorders ap- pears to substantially increase the suicide-attempt rate, com- pared with individuals having a single disorder (52). Most of the research has been done with regard to BPD; the following studies indicate that certain specific comorbidities may in- crease the risk for suicidal behaviour in youths with BPD.

Major Depressive Episode (MDE). Several studies have doc- umented that the existence of depression plus BPD may con- fer an increased risk for suicidal behaviour. In adolescents, coexistence of disruptive behaviour and depression is felt to be a particularly dangerous combination (53). Soloff and oth- ers completed the most careful study of this combination in adults (54). They examined a well-characterized group of pa- tients with BPD comorbid with MDEs and compared them with subjects having BPD without an MDE and with subjects having current MDE only. Soloff and others found suicidal behaviours more frequently in the total sample of patients with BPD, compared with patients having MDEs without comorbidity. This finding is anticipated of course, because suicidal behaviour is a DSM criterion for the diagnosis of BPD. The number of lifetime attempts significantly differen- tiated the comorbid patients from the other 2 comparison groups: subjects with comorbid BPD had a mean of 3.0 life- time attempts, compared with a mean of 1.9 lifetime attempts for subjects with BPD only and 0.8 lifetime attempts for sub- jects with MDEs only. When the researchers examined the most recent and the most serious lifetime attempt, they found no differences in their lethal intent as rated by the Suicide In- tent Scale (SIS). The total sample of patients with BPD dem- onstrated a greater lifetime level of medically lethal suicidal behaviours, compared with the patients having MDE without comorbidity. Patients with comorbid disorders reported sig- nificantly higher levels of objective planning, based on the most serious lifetime attempt, than did the other comparison groups. Using regression analysis, the researchers demon- strated that BPD diagnosis and comorbidity, history of ag- gression, and level of hopelessness predicted the number of lifetime attempts. Level of hopelessness alone predicted the level of intent for the most serious lifetime attempt. Finally, the lifetime number of suicide attempts predicted the level of medical damage for the most serious lifetime suicide attempt. Overall, the patients with comorbidity demonstrated an in- creased risk for suicidal behaviour, particularly with a higher number of lifetime attempts and evidence of more objective planning. The authors concluded that suicidal behaviour in in- patients with BPD should not be considered “less serious” than the suicidal behaviour of inpatients with an MDE. Other researchers, in earlier studies (25,55–57), found that comorbid depressive disorder increased the risk of suicidal behaviour in patients with BPD. An earlier study by Soloff and others (58), which did not include the above 3 comparison groups, did not find that comorbid major depressive disorder predicted suicidal behaviour in patients with BPD. However, the level of depression was found to be a risk factor for sui- cidal behaviour.

Substance Abuse Disorder. A relation has also been found be- tween increased suicidal behaviour and comorbidity of sub- stance abuse disorder with BPD (58,59). Links and others examined the prognostic significance of comorbid substance abuse in patients with BPD (60). These patients were followed prospectively over a 7-year period. The researchers found that patients with comorbid substance abuse and BPD perceived themselves to be at significantly more risk for suicide than did the comparison groups of patients having BPD without comorbidity, patients having substance abuse without BPD, and patients having borderline traits only. The patients with comorbidity also demonstrated a more frequent pattern of self-mutilating behaviour and reported a more frequent pat- tern of suicide threats and attempts than did the patients with- out comorbidity. Evidence also indicates that comorbidity of substance abuse disorder and CD increases the risk for sui- cidal behaviour in youth (16,61) (see Table 1).

Antisocial Personality Disorder. The presence of comorbid BPD and APD identifies a group at particularly high risk for suicide, as demonstrated in Stone’s long-term follow-up study of the PI 500 (62). Similarly, Soloff and others identified the fact that patients with comorbid BPD and APD had attempted suicide more frequently (58). Impulsivity as a trait character- istic, also found in patients with BPD and bulimia nervosa, seems to be a risk factor for increased suicidal behaviour in pa- tients with BPD (58,63,64).

 

 

 

 

Table 1 Risk factors for suicide or suicidal behaviour in youth with borderline or antisocial personality disorder vs other psychiatric disorders

Studies                                             Findings

Suicide Marttunen and others (16)        In suicide victims with APD, higher rates of the following: ·     Separation from a parent ·     Separation from father (before age 12 years) ·     Parental alcohol abuse ·     Parental violence ·     Alcohol abuse or dependence ·     Comorbid mental disorders Runeson and Beskow (59)       In suicide victims with BPD, the following risk factors found: ·     Substance abuse ·     Parental abuse ·     Unstable employment (more than 2 job changes) ·     Financial problems (existing debt at time of death) ·     Lack of permanent residence ·     Sentence by a court of law ·     Parental substance abuse Stone (62)                                   Suicide victims with BPD experienced: ·     More life events per patient (loss, incest, parental brutality, or parental humiliation); a higher frequency of  incest than all other diagnoses who had committed suicide Suicidal behaviours Kelly and others (61)                  ·    Higher rates of comorbid alcohol use disorder and CD in suicide attempters ·     Higher rates of comorbid cocaine use disorder and CD in suicide attempters ·     Higher rates of comorbid hallucinogen use disorder and CD in suicide attempters Stone (62)                                   Suicide victims with BPD experienced: ·     More life events per patient (loss, incest, parental brutality, or parental humiliation); a higher frequency of  incest than all other diagnoses who had committed suicide Verkes and others (63)               ·    Bulemia nervosa with BPD subtype may be predisposed to suicidal-type behaviour (that is, poor impulse control, anger, and depression). Young and Gunderson (68)       ·    More highly suicidal  adolescents rate themselves higher on family alienation and social isolation.  At the same time, parents do not see this attribute. ·     More highly suicidal BPD adolescents rate themselves as more impaired in overall function- ing.  At the same time, parents do not see this attribute. ·     More highly self-destructive (for example, burn self or run away) see themselves as more socially isolated.  Parents see that adolescent experiences social isolation.

APD = Antisocial Personality Disorder; BPD = Borderline Personality Disorder; CD = conduct disorder; MDD = Major Depressive Disorder; MDE = Major Depressive Episode; SA = Substance Abuse

 

Table 2 Risk factors for suicidal behaviour in adults with borderline or antisocial personality disorder vs other psychiatric disorders

Suicidal behaviours studies                  Findings

Brodsky and others (64)                        BPD patients who had trait impulsivity had a higher number of past suicide attempts than those who did not meet the criteria for trait impulsivity. Corbitt and others (57)                            ·    MDD + BPD patients were more likely to have made a high number of previous suicide attempts (greater than or equal to 3), compared with those with no personality disorder or another person- ality disorder. ·     MDD + BPD patients were younger at time of first attempt (mean 21.9, SD 7.0 years) compared with those with no personality disorder (mean 33.5, SD 14.4 years) or another personality disor- der (mean 34.7, SD 10.6 years) or another Cluster B personality disorder (mean 24.2, SD 11.7 years). ·     MDD + BPD patients had increased severity of suicidal ideation. Dubo and others (70)                              ·    BPD patients with incidence of parental sexual abuse had more incidences of self-mutilative be- haviour in the past. ·     BPD patients with incidence of parental sexual abuse had longer duration of self-mutilative be- haviour in the past. ·     BPD patients with parental sexual abuse had longer duration of suicidal behaviour. Links and others (60)                              ·    BPD + SA patients perceived themselves as significantly more likely than comparison groups to kill  themselves in the future. ·     BPD + SA patients had higher scores on the impulse-action measure. ·     BPD + SA patients had more frequent pattern of self-mutilative behaviour. ·     BPD + SA patients had more frequent pattern of suicide threats and attempts. Silk and others (71)                                With reference to the duration of sexual abuse, on-going sexual abuse predicted parasuicidal behaviour in BPD patients. Soloff and others (58)                            In BPD patients who attempted suicide, the following risk factors were found: ·     Older age ·     More impulsive actions ·     APD ·     Depressive state ·     Number of prior attempts Soloff and others (54)                             ·    Patients with comorbid BPD + MDE had a higher number of previous suicide attempts. ·     The diagnosis of BPD patients, hopelessness, and history of aggression predicted the lifetime number of suicide attempts. Soloff and others (72)                            BPD patients who have experienced childhood sexual abuse are 10 times more likely to attempt than those without childhood sexual abuse. Rietdijk and others (69)                           ·    Female BPD patients who had lower scores on the Reasons for Living Inventory’s Survival and Coping Beliefs subscale were 6.8 times more likely to show suicidal behaviour during the 6 month follow-up period, compared with high scorers. ·     Female BPD patients who had a comorbid diagnosis of depressive personality disorder were 6.7 times more likely to show suicidal behaviour during the 6 month follow-up period, compared with those without the diagnosis. ·     Female BPD patients who had lower scores on the Utrechtse Coping Lijsts’ reassuring thoughts were 11 times more likely to evidence suicidal behaviour Kelly and others (65)                               ·    Suicide attempters had significantly more life events. ·     BPD patients low on social adjustment were more likely to be classified as suicide attempters, compared with those with depression.

APD = Antisocial Personality Disorder; BPD = Borderline Personality Disorder; CD = conduct disorder; MDD = Major Depressive Disorder; MDE = Major Depressive Episode; SA = Substance Abuse

 

Recent Life Events

Adverse life events may push high-risk patients into actual suicidal crises. Kelly and others studied the impact of recent life events and the level of social adjustment in patients with major depression, patients with BPD, and patients with comorbid major depression and BPD (65). They found that the suicide attempters within this sample had experienced more adverse life events recently, particularly in the area of stressful events at home, with the family, or financially (65). In addition, the total number of life events was related to in- creased risk of suicidal behaviour. When the authors did a re- gression analysis to predict suicide attempter status, the diagnosis of BPD was predictive, as was the level of social adjustment in the family unit. In particular, an interaction was found between the presence of low social adjustment and the borderline diagnosis. It indicated that these patients were 16 times more likely to be classified as suicide attempters than were the patients with MDEs. Runeson and Beskow also found that the number of stressful life situations was related to death by suicide for adolescents with BPD, compared with others not having BPD (59). These stressful situations in- cluded unstable employment, financial problems, lack of a permanent residence, and a sentence by a court of law. Discharge from hospital should also be considered a stressful event. Kullgren found that, when discharge from hospital was imminent, patients with BPD were at somewhat increased risk for suicide and that such suicides occurred during the period of inpatient care and in the weeks following discharge (66). Kjelsberg and others noted that, compared with surviving BPD patients, patients with BPD who died by suicide during or following hospitalization were more frequently discharged after violating an inhospital contract (67). Young and Gunderson reported that adolescents with BPD were more highly suicidal if they saw themselves as alienated from their family and if they perceive their functioning to be impaired (68). Often, parents did not see these attributes and were thus not as aware of the patients’ perception of the risk for suicide. Maladaptive coping strategies appear to partly explain how patients with BPD are at increased risk for suicidal behaviours when facing stressful events (69).

The history of childhood abuse needs to be mentioned, al- though whether this is a modifiable risk factor is debatable. In his study of adolescents who died by suicide, Stone found that such patients with BPD were more likely to have experienced parental brutality and parental humiliation in their past (62). Dubois and others found a relation between a history of paren- tal sexual abuse and the risk of self-mutilative behaviours (70). This association has also been documented by several other investigators (59,71). Soloff and others completed a key study examining the relation between childhood abuse and suicidal behaviour in a sample of patients with BPD (72). They found that, in terms of the number of attempts, suicidal behaviour was predicted by a history of childhood sexual abuse, by the severity of BPD, and by the level of hopeless- ness. In fact, childhood sexual abuse continued to predict the number of attempts independently, even after the researchers entered several other selected risk factors into the analysis. The severity of childhood sexual abuse was associated with the severity of comorbid depression in these patients, with the presence of antisocial traits, and with a trend toward greater hopelessness. These factors may mediate childhood sexual abuse as a risk factor for suicidal behaviour. Soloff and others indicated that, in persons with BPD, the history of childhood sexual abuse increases tenfold the risk of suicidal behaviour, compared with patients without such a history (72).

In summary, patients with APD or BPD are likely to be at in- creased risk when they demonstrate the above risk factors. The presence of comorbidity, particularly when it is acutely evident, may lead high-risk patients into episodes of acute sui- cidal behaviour. The accumulation of recent life events or the lack of intimate or family support also indicate times of high risk for these patients. If factors such as a history of childhood sexual abuse and its associated psychopathologic deficits or the level of impulsivity are modifiable, based on clinical inter- ventions, then these interventions can potentially reduce the ongoing risk in youth with BPD or APD (73).

 

Clinical Approach to Patients With Antisocial, Borderline, and Narcissistic Personality Disorders

Patients with APD or CD present a unique challenge to clini- cians. When these patients present in crisis, the clinician is faced with the risk of assessing the potential for violence in addition to the risk of suicide or suicidal behaviour. For exam- ple, Marttunen and others reported that 10/23 patients with antisocial behaviour had a history of violence against others (16). In fact, all the patients that met criteria for CD or APD had a history of violence. Clinicians need to consider carefully their interventions (such as hospitalization), based on the po- tential risk to the patient vs the risk this individual might rep- resent to copatients.

Apart from the usual factors involved in a risk assessment, the psychopathy concept can be clinically valuable when clini- cians are trying to balance risks to the patient vs risks to others. Cleckley classically describes patients with psychopathic per- sonality disorder as having a “disinclination” toward suicide (74). Clinically, he observed that, among ward patients, sui- cidal behaviour was much rarer in psychopathic patients than in other patients. Cleckley wrote, “instead of a predilection for ending their own lives, psychopaths, on the contrary, show much more evidence of a specific and characteristic immunity for such an act” (74, p 359).

Cleckley’s clinical observation has had some support from empirical research. Hare developed the Psychopathic Check- list-Revised to capture the aspects of Cleckley’s psychopathic concept (75). Research has shown that the concept comprises 2 underlying dimensions. The first dimension, called Factor 1 or “emotional detachment,” includes the affective component of psychopathy: the glibness, superficial charm, grandiose sense of self-worth, pathologic lying, cunning and manipulativeness, lack of remorse or guilt, shallow affect, cal- lousness, lack of empathy, and failure to accept responsibility. Factor 2 relates to antisocial behaviour. In this Factor, items such as a proneness to boredom, poor behavioural controls,

early problematic behaviour, lack of realistic long-term goals, impulsivity, irresponsibility, juvenile delinquency, and revo- cation of conditional release were found (75). In a direct ex- amination of the relation between the factors of psychopathy, APD, and suicide risk, Verona and others attempted to look at whether a suicidal history was differently related to Factors 1 and 2 from the Hare Psychopathic Checklist-Revised (76). The authors found that suicidal history was significantly re- lated to Factor 2. A history of suicidal behaviour was nega- tively related to Factor 1; however, this relation was not statistically significant.

The mechanism by which the Factor 1, “emotional detach- ment,” might work to lessen the risk of suicide seems related to the emotional deficit found in some psychopaths. Physiologic studies have demonstrated that psychopathic in- dividuals have reduced startle response when processing ad- verse stimuli (77,78). This deficit may indicate a temperamental difficulty in their capacity to reflect negative affect or to experiencing depression or dysphoric states (79).

Therefore, assessing the psychopathic Factor 1 might assist clinicians to decide on the relative risk patients present for themselves vs others. It appears that patients demonstrating elements of emotional detachment are more likely to be a risk to others than to themselves. In making particular decisions to admit such patients to an inpatient psychiatric environment, one must carefully weigh the risks to vulnerable others vs the risk to the patient.

The clinical assessment of the patient with BPD in crisis is complicated. Often these patients have made multiple suicide attempts, and it is unclear whether a short-term admission will have any impact on the ongoing risk of suicidal behaviour. Figure 1 demonstrates a way of assessing and communicating the suicidal risk of patients with BPD and a history of repeated suicide attempts. These patients typically are at a chronically elevated risk of suicide, one that is much above that of the gen- eral population. This risk exists because of their history of multiple attempts. In addition to the history of multiple at- tempts, a history of self-injurious behaviour also increases the risk for suicide (80,81). Stanley and others found that patients with self-injurious behaviour were at risk for suicide attempts because of their high level of depression, hopelessness, and impulsivity and also because they misperceive and under- estimate the lethality of their suicidal behaviours (81). The pa- tient’s level of chronic risk can be estimated by taking a care- ful history of previous suicidal behaviour and by focusing on the times when the patient may have demonstrated attempts with the greatest intent and medical lethality. By documenting the patient’s most serious suicide attempt, one can estimate the severity of the patient’s ongoing chronic risk for suicide.

 

Figure 1 Assessing suicide risk in patients with borderline personality disorder

In patients with BPD, the acute-on-chronic level of risk is re- lated  to  several  factors  (Figure  1,  shaded  arrow). An understand the psychological meaning of their suicidal thoughts and behaviours and to learn new methods of communicating these emotional issues.

Assessing risk in patients with narcissistic personality disorder presents a unique clin- ical challenge. Ronningstam and Maltsberger’s description is thought pro- voking: they note how narcissistic patients can be at risk for suicide at times when they are not suffering from depression (83). Certainly, patients with narcissistic per- sonality disorder will be at increased risk during episodes of comorbid depression (84), but they can also become acutely sui- cidal outside episodes of clinical depres- sion. Suicide attempts in narcissistic patients can arise because of their very fragile self-esteem and in response to perceived narcissistic injury (85). Ronningstam and Maltsberger observe that acute-on-chronic risk will be present if the patient is suffering from comorbid major depression or if the patient is demon- strating high levels of hopelessness or depressive symptoms, as reviewed above. In addition, patients with BPD are known to be at risk for suicide during times of hospitalization and dis- charge. The clinical scenario of a patient presenting in crisis shortly after discharge from an inpatient setting illustrates a time when the risk assessment must be very carefully com- pleted to ensure that a proper disposition is made. This patient is potentially at an acute-on-chronic risk and the assessment cannot be truncated because of the recent discharge from hos- pital. Proximal substance abuse can increase the suicide risk in a patient with BPD. Of course, the existence of a diagnosis of substance abuse increases the chronic risk for suicidal be- haviour. This risk is acutely elevated when patients have less immediate family support or when they have lost or believe they have lost an important relationship.

Gunderson makes the distinction that patients with BPD who are attempting to manipulate the environment are at less risk than patients with BPD who present in a highly regressed dissociative state (42). At these times, acute interventions fre- quently have to be put in place to reduce the risk of suicide at- tempts or self-harm. The acute-on-chronic model can effectively communicate decisions regarding interventions. For example, if a patient is felt to be at a chronic, but not acute-on-chronic, risk for suicide, one can document and communicate that a short-term hospital admission will have little or no impact on a chronic risk present for months and years. However, a patient demonstrating an acute-on-chronic risk (Figure 1, the shaded arrow) maywell require an inpatient admission. In this circumstance, a short-term admission may allow the level of risk to return to chronic preadmission levels. Managing the chronic level of suicide risk in patients with BPD often involves strategic outpatient management. Dialectical behaviour therapy has been shown to effectively reduce suicidal behaviour (80,82). In working with out- patients with BPD, clinicians have to be able to tolerate the chronic risk of suicide. Over time, patients can be helped to

suicidal behaviour can have several meanings in these pa- tients, including an attempt to raise self-esteem through a sense of mastery; an attempt to protect themselves against an- ticipated narcissistic threats—“death before dishonour”; a vengeful act against a narcissistic trauma; the false belief of indestructibility; and a wish to destroy or attack an imperfect self. Narcissistic individuals, therefore, can demonstrate the so-called “Richard Corey suicide”; that is, they can take their lives despite seeming to have every happiness and good for- tune. Apter and others noted this in the recruits who took their lives in their sample (49). The recruits with narcissistic disor- der tended to be seen as promising young soldiers who, based on their preinduction interviews, seemed well equipped to handle service-related stress. Apter and others felt that these individuals might have been at risk because of their isolative and perfectionist traits.

Clinicians can take 4 steps to monitor the risk of suicide and suicidal behaviour in patients with narcissistic personality disorder. First, such patients should be routinely monitored for evidence of coexisting major depression or for an acute ep- isode of lowered self-esteem resulting from a felt narcissistic injury. Because suicide attempts in narcissistic individuals tend to arise abruptly, the risk can be lessened by denying pa- tients access to a means of suicide. Therefore, attention should be paid to ensuring that patients have no access to highly lethal means of suicide, such as guns or large quantities of pills. The patients’ families and other significant supports should be aware of the potential for an acute onset of suicidal feelings and the need to avoid access to lethal means. Finally, Kohut suggested that narcissistic patients may be at less risk of acting out suicidal behaviour once they have established a stable transference within a therapeutic relationship and once the therapist has established some empathic closeness to the pa- tient’s fragmented self (86). Creating a stable therapeutic rela- tionship seems to be an important factor that can lessen the risk  of  suicide  in  patients  with  narcissistic  personality disorder; it should be a consideration in their ongoing outpa- tient management.

 

Summary

Youths with antisocial, borderline, and narcissistic personal- ity disorder present unique challenges to clinicians. First, these diagnoses clearly identify individuals at increased risk for suicidal behaviour and death by suicide. Second, the as- sessment of their risk of suicide is problematic. For patients with APD, the risk of violence has to be judged, in addition to the risk of suicide or self-harm. For patients with BPD, one has to differentiate an acute from a chronic risk and determine interventions based on this differentiation. Finally, patients with narcissistic personality disorder can be at high risk for suicide outside times of clinical depression. These episodes can seem to be unpredictable. This paper offers some clinical approaches to assessing these patients. However, we hope that future empirical evidence will provide a more sound footing for assessing and managing youth with antisocial, borderline, and narcissistic personality disorder.

 

References

 

1. Statistics Canada. Causes of Death, 1999. Ottawa (ON): Statistics Canada; 2002. Catalogue no 84F0208XPB.
2. Robins E, Murphy GE, Wilkinson RH, Gassner S, Kayes J. Some clinical consid- erations in the prevention of suicide based on a study of 134 successful suicides. Am J Public Health 1959;49:888–99.
3. O’Carroll PW, Berman AL, Maris RW, Moscicki EK, Tanney BL, Silverman MM. Beyond the Tower of Babel: a nomenclature for suicidology. Suicide Life Threat Behav 1996;26:237–52.
4. Simeon D, Favazza AR. Self-injurious behaviors: phenomenology and assess- ment. In: Simeon D, Hollander E, editors. Self-injurious behaviors. assessment and treatment. Washington (DC): American Psychiatric Publishing, Inc; 2001. p 1–28.
5. Kleespies PM, Dettmer EL. An evidence-based approach to evaluating and man- aging suicidal emergencies. J Clin Psychol 2000;56:1109–30.
6. Jacobs DG, Brewer M, Klein-Benheim, M. Suicide assessment: an overview and recommended protocol. In: Jacobs DG, editor. The Harvard Medical School guide to suicide assessment and intervention. San Francisco (CA): Jossey-Bass Publishers; 1999. p 3–39.
7. Kleespies P, Deleppo J, Mori D, Niles B. The emergency interview. In: Kleespies P, editor. Emergencies in mental health practice: evaluation and management. New York: Guilford; 1998. p 41–72.
8. Rudd MD, Joiner T. The assessment, management, and treatment of suicidality: toward clinically informed and balanced standards of care. Clinical Psychology: Science and Practice 1998;5:135–50.
9. Thienhaus OJ, Piasecki MP. Assessment of suicide risk. Psychiatr Serv 1997;48:293–
10. Kral MJ, Sakinofsky I. A clinical model for suicide risk assessment. In: Leenaars AA, Maltsberger JT, Neimeyer RA, editors. Treatment of suicidal people. Wash- ington (DC): Taylor and Francis; 1994. p 19–31.
11. Lester D. Adolescent suicide risk today: a paradox. J Adolesc 1998;21:499–503.
12. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington (DC): American Psychiatric Association; 1994.
13. Levy KN, Becker DF, Grilo CM, Mattanah JJ, Garnet KE, Quinlan DM, and oth- ers. Concurrent and predictive validity of the personality diagnosis in adolescent inpatients. Am J Psychiatry 1999;156:1522–8.
14. Grilo CM, Becker DF, Edell WS, McGlashan TH. Stability and change in DSM-III-R personality disorder dimensions in adolescents followed up 2 years after psychiatric hospitalization. Compr Psychiatry 2001;42:364–8.
15. Marttunen MJ, Aro HM, Henriksson MM, Lonnqvist JK. Mental disorders in ad- olescent suicides, DSM-III-R Axes I and II among 13 to 19 year olds. Arch Gen Psychiatry 1991;48:834–9.
16. Marttunen MJ, Aro HM, Henriksson MM, Lonnqvist JK. Antisocial behaviour in adolescent suicide. Acta Psychiatr Scand 1994;89:167–73.
17. Apter A, Bleich A, Plutchik R, Mendelsohn S, Tyano S. Suicidal behavior, de- pression, and conduct disorder in hospitalized adolescents. J Am Acad Child Adolesc Psychiatry 1988;27:696–9.
18. Brent DA, Perper JA, Moritz G, Allman C, Friend A, Roth C, and others. Psychiatric risk factors for adolescent suicide: a case-control study. J Am Acad Child Adolesc Psychiatry 1993;32:521–9.

19. Beautrais AL, Joyce PR, Mulder RT, Fergusson DM, Deavoll BJ, Nightingale SK. Prevalence and comorbidity of mental disorders in persons making serious suicide attempts: a case-control study. Am J Psychiatry 1996;153:1009–14.
20. Maddocks PD. A five-year follow-up of untreated psychopaths. Br J Psychiatry 1970;116:511–5.
21. Laub JH, Vaillant GE. Delinquency and mortality: a 50-year follow-up study of 1000 delinquent and nondelinquent boys. Am J Psychiatry 2000;157:96–102.
22. Runesson B, Beskow J. Borderline personality disorder in young Swedish sui- cides. J Nerv Ment Dis 1991;179:153–6.
23. Kullgren G, Renberg E, Jacobsson L. An empirical study of borderline personal- ity disorder and psychiatric suicides. J Nerv Ment Dis 1986;174:328–31.
24. Bongar B, Peterson, LG, Golann S, Hardiman JJ. Self-mutilation and the chroni- cally “suicidal” emergency room patient. Ann Clin Psychiatry 1990;2:217–22.
25. Crumley F. Adolescent suicide attempts. JAMA 1979;241:2404–7.
26. Paris J, Zweig-Frank H. A 27-year follow-up of patients with borderline person- ality disorder. Compr Psychiatry 2001;42:482–7.
27. Perry J. Longitudinal studies of personality disorders. J Personal Disord Suppl 1993;1:63–85.
28. McGlashan TH. The Chestnut Lodge Follow-Up Study III: long-term outcome of borderline personalities. Arch Gen Psychiatry 1986;43:2–30.
29. Stone MH, Stone DK, Hurt S. The natural history of borderline patients. Psychiatr Clin North Am 1987;10:185–205.
30. Akiskal HS, Chen SE, Davis GC. Borderline: an adjective in search of a noun. J Clin Psychiatry 1985;46:41–8.
31. Pope HG Jr, Jonas JM, Hudson JI. The validity of DSM-III borderline personality disorder: a phenomenologic, family history, treatment response, and long-term follow-up study. Arch Gen Psychiatry 1983;40:23–30.
32. Fyer MR, Frances AJ, Sullivan T, Hurt SW, Clarkin J. Suicide attempts in pa- tients with borderline personality disorder. Am J Psychiatry 1988;145:737–9.
33. Berman AL. The teenager at risk for suicide. Med Aspects Hum Sex 1985;19:123–9.
34. Friedman RC, Corn R. Suicide in the borderline depressed adolescent and young adults. J Am Acad Psychoanal 1987;15:429–88.
35. Stone MH, Hurt S. W., Stone DKA. The PI 500: long-term follow-up of border- line inpatients meeting DSM-III criteria: I. Global outcome. J Personal Disord 1987;1:291–8.
36. Plakun E, Burkhardt PE, Muller JP. Fourteen year follow-up of borderline and schizotypal personality disorders. Compr Psychiatry 1985;26:448–55.
37. Paris J, Brown R, Nowlis D. Long-term follow-up of borderline patients in the general hospital. Compr Psychiatry 1987;28:530–5.
38. Stone MH. The fate of borderline patients. New York: Guilford Press; 1990.
39. Clarkin JF, Friedman RC, Hurt SW, Cora R, Aronoff, M. Affective and character pathology of suicidal adolescents and young adults in patients. J Clin Psychiatry 1984;45:19–22.
40. Casey PR. Personality disorder and suicide intent. Acta Psychiatr Scand 1989;79:290–5.
41. Gomez A, Lolas F, Martin M, Barrera A, Jaar E, Suarez I. The influence of per- sonality on suicidal behavior. Actas Luso Esp Neurol Psiquiatr Cienc Annes 1992;20:250–6.
42. Gunderson JG. Borderline personality disorder. Washington (DC): American Psychiatric Press; 1984.
43. Shearer SL, Peters CP, Quaytman MS, Wadman BE. Intense lethality of suicide attempts among female borderline patients. Am J Psychiatry 1988;145:1424–7.
44. Zanarini MC, Gunderson JG, Frankenburg FR, Chauncey DL. Discriminating borderline personality from other axis II disorders. Am J Psychiatry 1990;147:161–7.
45. Kjellander C, Bonger B, King A. Suicidality in borderline personality disorder. Crisis 1998;19:125–34.
46. Runeson BS, Beskow J, Waern M. The suicidal process in suicides among young people. Acta Psychiatr Scand 1996;93:35–42.
47. Samuels J, Eaton WW, Bienvenu III J, Brown CH, Costa PT, Nestadt G. Preva- lence and correlates of personality disorders in a community sample. Br J Psychi- atry 2002;180:536–42.
48. Links PS, Stockwell M. The role of couple therapy in the treatment of narcissis- tic personality disorder. Am J Psychotherapy 2002;56:522–38.
49. Apter A, Bleich A, King RA, Kron S, Fluch A, Kotler M, and others. Death without warning? A clinical postmortem study of suicide in 43 Israeli adolescent males. Arch Gen Psychiatry 1993;50:138–42.
50. Stone M. Long-term follow-up of narcissistic personality disorder. Psychiatr Clin North Am 1989;12:621–41.
51. Allen M. Emergency psychiatry. Review of psychiatry. Volume 21. Washington (DC): American Psychiatric Press; 2002.
52. Lewinsohn PM, Rohde P, Seeley JR. Adolescent psychopathology: III. The clini- cal consequences of comorbidity. J Am Acad Child Adolesc Psychiatry 1995;34:510–9.
53. McCracken JT, Cantwell DP, Hanna GL. Conduct disorder and depression. In: Koplewicz HS, Klass E, editors. Monographs in clinical pediatrics. Volume 6. Depression in children and adolescents. New York: Harwood Academic Press; 1993. p 121–32.
54. Soloff PH, Lynch KG, Kelly TM, Malone KM, Mann JJ. Characteristics of sui- cide attempts of patients with major depressive episode and borderline personal- ity disorder: a comparative study. Am J Psychiatry 2000;157:601–8.

55. Fyer MR, Frances AJ, Sullivan T, Hurt SW, Clarkin J. Suicide attempts in pa- tients with borderline personality disorder. Am J Psychiatry 1988;145:737–9.
56. Friedman RC, Aronoff MS, Clarkin JF, Corn R, Hurt SW. History of suicidal be- havior in depressed borderline inpatients. Am J Psychiatry 1983;140:1023–6.
57. Corbitt EM, Malone KM, Haas GL, Mann JJ. Suicidal behavior in patients with major depression and comorbid personality disorders. J Affect Disord 1996;39:61–72.
58. Soloff PH, Lis JA, Kelly T, Cornelius J, Ulrich R. Risk factors for suicidal be- havior in borderline personality disorder. Am J Psychiatry 1994;151:1316–23.
59. Runeson B, Beskow J. Borderline personality disorder in young Swedish sui- cides. J Nerv Ment Disease 1991;179:153–6.
60. Links PS, Heslegrave RJ, Mitton JE, van Reekum R, Patrick J. Borderline per- sonality disorder and substance abuse: consequences of comorbidity. Can J Psy- chiatry 1995;40:9–14.
61. Kelly TM, Cornelius JR, Lynch KG. Psychiatric and substance abuse disorders as risk factors for attempted suicide among adolescents. Suicide Life Threat Behav 2002;32:301–2.
62. Stone MH. Suicide in borderline and other adolescents. Adolesc Psychiatry 1992;18:289–305.
63. Verkes RJ, Pijl H, Meinders E, Kempen GMJ. BPD, impulsiveness, platelet monoamine measures in bulimia nervosa and recurrent suicidal behaviour. Biol Psychiatry 1996;40:173–80.
64. Brodsky BS, Malone KM, Ellis SP, Dulit RA, Mann JJ. Characteristics of bor- derline personality disorder associated with suicidal behavior. Am J Psychiatry 1997;154:1715–9.
65. Kelly TM, Soloff PH, Lynch KG, Haas GL, Mann JJ. Recent life events, social adjustment, and suicide attempts in patients with major depression and borderline personality disorder. J Personal Disord 2000;14:316–26.
66. Kullgren G. Factors associated with completed suicide in borderline personality disorder. J Nerv Ment Dis 1988;176:40–4.
67. Kjelsberg E, Eikeseth PH, Dahl AA. Suicide in borderline patients — predictive factors. Acta Psychiatr Scand 1991;84:283–7.
68. Young DW, Gunderson JG. Family images of borderline adolescents. Psychiatry 1995;58:164–72.
69. Rietdijk EA, van den Bosch LMC, Verheul R, Koeter MWJ, van den Brink W. Predicting self-damaging and suicidal behaviors in female borderline patients: reasons for living, coping, and depressive personality disorder. J Personal Disord 2001;15:512–20.
70. Dubo ED, Zanarini MC, Lewis RE, Williams AA. Childhood antecedents of self-destructiveness in borderline personality disorder. Can J Psychiatry 1997;42:63–9.
71. Silk KR, Lee S, Hill EM, Lohr NE. Borderline personality disorder symptoms and severity of sexual abuse. Am J Psychiatry 1995;152:1059–64.
72. Soloff PH, Lynch KG, Kelly TM. Childhood abuse as a risk factor for suicidal behavior in borderline personality disorder. J Personal Disord 2002;16:201–14.

73. Bergmans Y, Links PS. A description of a psychosocial/psychoeducational inter- vention for persons with recurrent suicide attempts. Crisis 2002;23:156–60.
74. Cleckley H. The mask of sanity. 5th ed. St. Louis (MO): CV Mosby; 1976.
75. Hare RD The Hare Psychopathy Checklist — Revised. Toronto: Multi-Health Systems; 1991.
76. Verona E, Patrick CJ, Joiner TE. Psychopathy, antisocial personality and suicide risk. J Abnorm Psychol 2001;110:462–70.
77. Patrick CJ. Emotion and psychopathy: startling new insights. Psychophysiology 1994;31:319–30.
78. Herpertz SC, Werth U, Lukas G, Qunaibi M, Schuerkens A, Kunert HJ, and oth- ers. Emotion in criminal offenders with psychopathy and borderline personality disorder. Arch Gen Psychiatry 2001;58:737–45.
79. Lovelace L, Gannon L. Psychopathy and depression: mutually exclusive con- structs? J Behav Ther Exp Psychiatry 1999;30:169–76.
80. Linehan M. Cognitive behavioral treatment of borderline personality disorder. New York: Guilford; 1993.
81. Stanley B, Gameroff MJ, Michalsen V, Mann JJ. Are suicide attempters who self-mutilate a unique population? Am J Psychiatry 2001;158:427–32.
82. Koener K, Linehan MM. Research on dialectical behavior therapy for patients with borderline personality disorder. Psychiatr Clin North Am 2000;23:151–67.
83. Ronningstam EF, Maltsberger JT. Pathological narcissism and sudden suicide- related collapse. Suicide Life Threat Behav 1998;28:261–71.
84. Maltsberger JT. Ecstatic suicide. Archives of Suicide Research 1997;3:283–301.
85. Perry CJ. Personality disorders, suicide and self-destructive behavior. In: Jacobs D, Brown H, editors. Suicide: understanding and responding. Madison (CT): International Universities Press; 1990.
86. Kohut H. Thoughts on narcissism and narcissistic rage. Psychoanal Study Child 1972;27:360–400.

 

Manuscript received and accepted March 2003.

1Arthur Sommer Rotenberg Chair in Suicide Studies, Professor of Psychiatry, Department of Psychiatry, St Michael’s Hospital, University of Toronto, Toronto, Ontario.

2Fellow in Emergency Psychiatry, Department of Psychiatry, St Michael’s Hospital, University of Toronto, Toronto, Ontario.

3Undergraduate Research Program, Suicide Studies Unit, Arthur Sommer Rotenberg Chair in Suicide Studies, St Michael’s Hospital, University of Toronto, Toronto, Ontario.

Address for correspondence: Dr PS Links, 30 Bond Street, Suite 2010, Shuter Wing, St Michael’s Hospital, Toronto, ON  M5B 1W8