Linda M. Nicholas, MD, MS, Assistant Professor of Psychiatry, and Robert N. Golden, MD, Professor and Chair of Psychiatry, University of North Carolina School of Medicine Chapel Hill, North Carolina
[Clinical Cornerstone 3(3):47-57, 2001. © 2001 Excerpta Medica, Inc.]
Seventy percent of persons who commit suicide see their primary care physician within 6 weeks of death. Of those who attempt suicide, 35% seek medical attention the week prior to the attempt. It is thus imperative that all physicians learn to recognize the risk factors for suicide, actively assess a patient's current risk, understand the differential diagnoses, and develop appropriate treatment approaches.
The evaluation of the suicidal patient requires familiarity with both demographic and individual risk factors. Although demographic or population risk factors may give us clues to which groups of people are at risk, it is the individual factors that help us understand what may drive a person to such hopelessness and despair that he or she would attempt to end his or her life.
In the United States, suicide rates among adolescents have risen 3-fold in the past 30 years. In 15-to 19-year-olds, suicide is the third leading cause of death after accidents and homicides. Several factors put the adolescent at risk for self-inflicted death, including impulsive and high-risk behaviors, untreated mood disorders (especially major depression and bipolar disorder), access to lethal means (eg, firearms), and substance abuse. The use of firearms is the most common method of completed suicide in children and adolescents and accounts for about two thirds of suicides in boys and almost one half of suicides in girls.
Gender. Men of all ages commit suicide 2.5 to 4 times more frequently than do women, whereas women attempt suicide 2 to 3 times more often than do men. However, attempts by men may be masked as risk-taking behaviors or car accidents. The predominance of males in completed suicide is accounted for in part by their more lethal and violent means as well as an increased rate of substance abuse.
Ethnicity and community. In the United States, suicide rates are highest among whites, followed by American Indians, African Americans, Hispanics, and Asian Americans. Suicide rates have been increasing in young African American men, however. Suicide rates among inner-city youths exceed national rates. Immigrants are also at risk, as their suicide rates exceed those of both the population of their country of origin and their adopted country.
Marital status. Married persons have the lowest risk of suicide. Rates for single persons are twice those of married persons, whereas rates for divorced, separated, or widowed persons are 4 to 5 times higher.
Employment. Work appears to be a protective factor for suicide, probably due to increased feelings of usefulness as well as increased social interaction in the workplace. Suicide rates increase in times of economic depression and high unemployment. In the United States, the highest suicide rates in the 20th century were during the Great Depression.
Neurobiologic correlates. Biologic factors may play a role in setting the "threshold" for impulsive, violent behavior and suicide. Measures of reduced serotonin function have been consistently associated with suicidal behavior in various diagnostic groups. Low concentrations of the serotonin metabolite 5-hydroxy-indoleacetic acid (5-HIAA) in cerebral spinal fluid (CSF) as well as blunted prolactin response to fenfluramine challenge have been found in patients who have attempted suicide, and postmortem studies have found presynaptic and postsynaptic changes in serotonin receptor sites in suicide completers compared with controls who died of other causes. Furthermore, reduced CSF 5-HIAA has been shown to predict future suicidal behavior.
Low serum levels of total cholesterol have been associated with suicide and suicide attempts in multiple studies. Primary prevention trials designed to lower cholesterol in patients with coronary artery disease resulted in an increase in deaths by suicide.
Current mental status. A thorough mental status examination is of utmost importance in the assessment of every patient with suicide potential. Special attention should be given to thought content, mood, affect, and factors that predispose to impulsive behavior.
Suicidal thoughts, intent, and plans. When assessing a patient's risk for self-harm or suicide, it is of utmost importance to directly inquire about thoughts of death or suicide. Suicidal thoughts may range from fleeting to persistent and intrusive. When suicidal thoughts are present, the clinician should question the patient's family and other associates with regard to communication of intent.
Intent may be expressed as putting one's affairs in order or saying good-bye to important people in one's life. When a patient admits to suicidal thoughts or intent, or both, the next step is to ask directly about the presence of a plan. The specificity and lethality of the plan must be evaluated, and a discussion regarding the patient's intent to act on it must be conducted. An important question to ask is, "Is the method of possible self-injury available?" Interestingly, the method used influences the success of the suicide attempt independent of the severity of the intent to die. Plans that are lethal and well thought out pose a particularly acute danger.
Affect/mood. Sadness, hopelessness, and withdrawal predispose a patient to suicide more so than anger. The degree of a patient's hopelessness must be assessed. Hopelessness is the symptom most associated with long-term suicide risk. Those patients who seem to have given up and can see no "light at the end of the tunnel" pose a significant risk. The degree of communication between physician and patient is also important. A patient who is withdrawn and noncommunicative should raise red flags for the clinician, more so than the patient who is responsive and warm. Also worrisome is the patient who expresses a desire to meet again with deceased loved ones.
The presence of severe anxiety, agitation, or insomnia greatly increases short-term risk of suicide. In one study, the presence of anxiety predicted 93% of suicides occurring within a year of diagnosis of major affective disorder, with the risk proportional to the severity of the anxiety. Sleep disturbance is a symptom of several psychiatric illnesses (particularly the mood and anxiety disorders) associated with increased suicide risk. Agitation is another state that appears to increase the risk of self-harm. Akathisia, a common side effect of certain psychotropic drugs, including conventional antipsychotics and some antidepressants, is another important form of agitation.
Impulsive behavior. Impulsive behavior is often seen in psychiatric disorders, particularly in bipolar disorder, some personality disorders, and substance use disorders. Impulsive behavior increases the risk for suicide. Some patients, especially women, may be hesitant to admit to such behaviors.
Inquiry as to how a patient handles stress or the breakup of a relationship and whether there is a history of domestic violence or abuse may give clues to potential for impulsive behavior. For example, an individual who responds to bad news or rejection by head banging, driving recklessly, binge drinking, or destroying property would be at high risk. Impulsive behavior is especially dangerous when combined with substance abuse or alcohol abuse.
Psychosis. The presence of psychotic symptoms is an important risk factor for suicide. Some patients with psychosis may have persecutory delusions from which they may wish to escape, or they may experience command hallucinations telling them to kill themselves.
Organic disease. Patients with a medical cause for a mood or psychotic disorder may be at increased risk for self-harm. In an acute delirium, insight and judgment are characteristically impaired. Brain injury due to illness (eg, stroke) or an accident may predispose to depression and at the same time may damage those parts of the brain needed for resourcefulness, judgment, and impulse control.
History. Evaluation of suicide risk must involve a careful patient history, including psychiatric or medical conditions, prior suicide attempts, as well as family and social histories.
Psychiatric and medical history. The greatest risk factor for suicide is the presence of a psychiatric disorder. The mental disorders most highly associated with suicide are the mood disorders (depression and bipolar disorder), schizophrenia, borderline and antisocial personality disorders, alcoholism, and drug abuse.
Mood disorders are the most common psychiatric disorders associated with suicide. In general population studies, 30% to 64% of persons who commit suicide have a primary depressive dis-order. Fifteen percent of depressed patients die by suicide. Comorbid psychiatric disorders (particularly alcohol or substance abuse) and medical illness further increase the risk of suicide in depressed persons. Depressed patients with delusions (eg, delusions of guilt, thought insertion, mind reading, or paranoia) are 5 times more likely to kill themselves than nondelusional depressed patients. Depressed patients are at especially increased risk of suicide in the 6 to 12 months following discharge from the hospital.
At some point in their lives, 25% to 50% of patients with bipolar disorder attempt suicide. Bipolar patients are more at risk during the rapid fluctuation in mood states at the beginning or end of a depressive episode and when there are simultaneous depressive and manic symptoms (mixed mood states). Irritability, poor sleep, recklessness, medication noncompliance, and comorbid sub-stance abuse increase the likelihood of suicide during hypomanic, manic, and mixed states.
Anxiety and panic are important short-term risk factors for suicide. National Institutes of Health Epidemiologic Catchment Area data indicate that 20% of persons with panic disorder have made a suicide attempt. Particularly worrisome is the combination of panic and hopelessness. In one study, patients with simple panic disorder without comorbidity had a lifetime rate of suicide attempts of 7%. When panic patients had comorbid depression, the risk rose dramatically to 19%. Assessment of suicide risk in patients with anxiety disorders therefore should include determination of superimposed depression. In patients with post-traumatic stress disorder, survivor guilt is associated with suicide.
Suicide is the leading cause of premature death (approximately 10%) in patients with schizophrenia. Suicide tends to occur in the first years of the illness, and patients with schizophrenia who commit suicide tend to be younger than other cohorts who commit suicide. Approximately one third of suicides in schizophrenic patients occur during the first few weeks following discharge from the hospital, and another approximately one third occur during inpatient hospitalization. Approximately 75% of schizophrenic persons who commit suicide are unmarried males. Although a small percentage may be actively psychotic with command hallucinations or delusions, risk is highest when symptoms are under control, possibly reflecting the painful recognition of their illness. Depressive symptoms are closely associated with suicide in schizophrenic patients.
The presence of either alcohol abuse or substance abuse increases the risk for suicide. The majority of alcohol abusers and substance abusers have suicidal thoughts, most commonly at the end of a binge. Alcohol abuse or dependence is associated with 25% to 50% of suicides and is second only to depression in its association with suicide. Comorbidity of alcohol or substance abuse with other psychiatric illnesses greatly increases the suicide risk.
Impulsive behavior, parasuicidality, and suicide characterize borderline and antisocial personality disorders. Many of these individuals use self-harm as a response to stress. Although persons with antisocial disorder or borderline disorder are more likely to be repeat attempters than completers, the death rates by suicide for these 2 groups are 5% and 8%, respectively.
Multiple studies have confirmed that the presence of physical illness increases the risk of a suicide attempt. In suicide completers aged >50 years, medical illness plays a role in the motivation to commit suicide in more than one half. Patients with AIDS have suicide rates up to 20% greater than those in the general population. Factors that may increase risk in patients with medical illness include any comorbid psychiatric disorder or sub-stance use disorder, chronic pain, threat of financial loss, and feelings of being a burden. Suicidal patients with a serious or terminal illness common-ly suffer from a treatable depressive disorder.
Often, particularly for adolescents, suicide occurs after a stressful event. There may have been a recent disapproval or rejection, disciplinary crisis,
or academic failure. However, healthy individuals adjust to these stresses and disappointments without thoughts of self-harm. Persons who react to these events with suicidal behavior need to be carefully evaluated for the presence of an underlying, treatable psychiatric disorder.
History of suicide attempts. A history of suicide attempt(s) significantly increases the lifetime risk for suicide completion. Sixty percent of completed suicides are preceded by attempts, and one third of patients who make a suicide attempt will repeat an attempt within 2 years. Eight times as many individuals attempt suicide as complete it. The methods previously used in attempts may give clues to the patient's intent and help to predict the lethality of future attempts.
Family history. Family history of suicide increases the risk of suicide attempts and completed suicide across diagnostic groups. Twin and adoption studies suggest a genetic transmission to suicidality that is independent of its relationship to mental illness. It may be that the genetic predisposition toward suicide and suicide attempts is related to an inability to control impulsive behavior.
Social history. Social factors that predispose to suicide include early death of a parent, childhood physical and social abuse, and social isolation. In a retrospective study, 30% of persons who completed suicide were found to have experienced a recent loss. Early parental loss is associated with chronic suicidality. The children of parents who are abusive, rejecting, or depressed demonstrate higher rates of suicidal behavior. Support systems always need to be assessed and maximized in any patient with suicidal thoughts because social withdrawal or isolation is a factor in half of suicides. A fall in socioeconomic status is sometimes a predisposing factor.
When deciding whether to hospitalize the suicidal patient, consider not only individual factors, but also support systems, living arrangements, and the feasibility and safety of other alternatives such as intensive outpatient treatment (Table II). When possible, work with other mature, involved adults (eg, family, friends, roommates, clergy, other health professionals). However, always remember that the bottom line -- regardless of the patient's stated preference, the family's wishes, or managed care restrictions -- is that safety comes first. If necessary and indicated, consider involuntary hospitalization. Explore and modify the safety of the environment, restricting access to lethal means. Make arrangements to remove dangerous drugs or weapons from the home and other environments. Ensure that the patient does not have access to firearms. "No-suicide contracts" -- formal agreements between patient and caregiver that no attempt at suicide will be made and that the clinician will be contacted in the event that the patient has a suicidal intent or plan -- may be useful in some situations, particularly with patients with whom there is an established alliance, but do not overly rely on them. Often such contracts mainly serve to make the clinician feel better rather than being truly meaningful for the patient. Remember that suicide is most often an irrational act.
Mood stabilizers. There is evidence that lithium may be prophylactic against suicide in bipolar depression and unipolar depression and in schizoaffective disorder. Lithium's antisuicidal effect may be secondary to its serotonergic activity and the stabilization of agitated and aggressive states. Lithium treatment must be monitored closely due to its low therapeutic margin and issues of compliance. Although anticonvulsant medications also have mood stabilizing and antiagitation effects, to date, they have not been shown to lower the risk of suicide.
Antidepressants. The selective serotonin reuptake inhibitors (SSRIs) are effective in the treatment of depression and certain anxiety disorders, and there is evidence that they may also improve impulsiveness and irritability. The SSRIs have an excellent safety profile, unlike the tricyclic antidepressants and monoamine oxidase inhibitors. However, bear in mind that initial side effects may include increased agitation and insomnia; therefore it may be wise to titrate slowly and to be prepared to add a temporary agent (eg, trazodone) to improve sleep at the onset of treatment. Remember that an antidepressant used alone may precipitate a switch into mania or the induction of rapid cycling in depressed patients with bipolar disorder.
In treating patients who are suicidally depressed, bear in mind that they may be at increased risk for self-harm during the transitional period of treatment, when neurovegetative symptoms such as low energy and lack of motivation start to improve but feelings of sadness, hopelessness, and low self-esteem are still severe.
Antipsychotics. Antipsychotic drugs clearly provide substantial benefit for patients requiring them, but their potential side effects should also be considered. Control of psychotic symptoms reduces suicide risk. Some of these agents, however, especially the older "typical" antipsychotics, may induce akathisia. The new-generation atypical antipsychotic medications such as clozapine, olanzapine, and quetiapine are preferable in this regard.
Psychologic and psychosocial interventions. Sometimes cognitive behavioral psychotherapy or crisis intervention therapy can be a useful adjunct to pharmacologic approaches. Providing psychosocial support such as housing and vocational rehabilitation is important for the patient with schizophrenia or other chronic psychiatric illness. These interventions may enable patients to avoid the despair associated with the chronicity and progression of their illness.
In addition to familiarity with the risk factors for suicide, the public and the medical community must understand that thoughts of self-harm and suicidal behaviors are intimately linked to major psychiatric illnesses. Physicians should recognize and promptly implement treatment for specific psychiatric disorders. Remember that the majority of persons who commit suicide have seen their primary care physician during the weeks prior to death. It is important to recognize the most serious danger periods, including: (a) when depression begins to improve; (b) when chronically medically ill patients receive discouraging news regarding prognosis; (c) when financial stresses secondary to psychiatric or medical illness begin to mount on the family; (d) when the patient enters the posthospitalization period; (e) when the family disperses after rallying around an acute problem; and (f) after any major loss or adverse life event.
Depression
Alcoholism~70% of elderly suicide victims
~20% of elderly suicide victimsPhysical losses Medical problems, pain, increased vulnerability, helplessness Financial losses Retirement, medical expenses, loss of security Social losses Loss of home, death of spouse or friends, loss of work role, social isolation, loneliness Emotional losses Hopelessness, dependence on others, despair Cognitive losses Decreased self-concept and self-esteem
*Remember, it is impossible to predict with certainty whether an individual will commit suicide.
Strongly Consider Hospitalization May Consider Outpatient Management Prior attempt of high lethality No history of potentially lethal suicide attempts Well-thought-out plan Lack of plan/intent; cooperative family member or other adult Access to lethal means Removal or lack of availability of lethal means Uncommunicative Communicative Recent major loss Availability of intensive outpatient care Social isolation Good social support Hopelessness Hopefulness History of impulsive, high-risk behavior Active substance abuse or dependence Untreated mood, psychotic, or personality disorder
GOLDEN
When we talk about a major psychiatric illness, we mean disorders that are diagnosable in the DSM-IV. The other 10% include a combination of situations; for example, a person who impulsively reacts to bad news with a suicidal act: he comes home, finds out his spouse has left him, becomes totally distraught, uncharacteristically gets drunk, and gets into a car accident. One might argue that such a patient has a psychiatric condition, but there's no clear diagnosis.
GOLDEN
For those with a psychiatric illness, the risk remains as long as the illness, for example, depression or mania, goes untreated. For individuals in the small group without a defined psychiatric illness, once they have settled down from the acute trauma, in the short term, things are probably going to be fine. But in the long term, if they experience another major traumatic event, their reactions to past traumas will likely be predictive of their reactions to future traumas.
GOLDEN
Any kind of violent or destructive impulsive behavior; the act doesn't have to be directed toward oneself. Among those at risk would be someone who has a history of pounding a hole in the drywall when angry, someone who has a history of getting upset and driving off in the car without any purpose, someone who has a history of suddenly screaming, running out of the room, and then later saying, "My God, I don't know what happened, I just lost control." Any kind of impulsive, dramatic, especially aggressive or violent, out-of-control act is of concern.
GOLDEN
Drug-induced akathisia is a side effect that needs to be taken very seriously. At first glance, drug-induced restlessness may not seem to be a serious symptom. However, in patients with a serious psychiatric illness such as schizophrenia, the development of akathisia is a major cause of noncompliance with antipsychotic medication. Akathisia can also make these patients susceptible to command hallucinations and can drive them to impulsive acts out of frustration, including suicide.
GOLDEN
No. In fact, in a prospective cohort study of people who made nonserious suicide attempts, which was conducted by a major suicide-study center, a direct relationship between the number of prior superficial attempts and the risk of ultimately dying of suicide was demonstrated. In other words, this study found that people who had made a single suicidal gesture were at less risk than those who had made 3, who, in turn, were at less risk than those who had made 7. Thus, the number of attempts in the past, in a stepwise fashion, increases prospectively the risk of ultimately dying of suicide. The interpretation of these data is difficult, but one possibility is that if a person makes 1 superficial gesture, the psychiatric diagnosis might be adjustment disorder, whereas if a person makes 7 or 8 attempts, he may have a chronic persistent or a chronic recurrent psychiatric illness like manic-depressive illness or schizophrenia or alcohol and substance abuse, which is much more dangerous. So it could be that a history of multiple attempts is a kind of indirect marker for a serious suicide-risk diagnosis.
GOLDEN
In the studies from which these figures were derived, family members were queried only about the last time the patient saw a doctor, not about the exact nature of the visit. I would suspect, however, that many of these patients were struggling with a psychiatric illness like major depression or bipolar disorder. Unfortunately, even today in our enlightened society, people remain embarrassed talking about psychiatric illness, more embarrassed than talking about other types of disease. This would lead me to suspect that after these patients walk into the office, it is tough for them to bring up the subject, so instead they present nonspecific physical complaints like fatigue, low energy, or maybe the flu, and the true reason for the visit doesn't come to light unless the physician suspects depression and queries them about suicidal thoughts.
In light of this, I want to underscore that there is absolutely no truth to the dangerous misconception that making an inquiry about suicide may put the idea into a patient's head, precipitating a suicidal attempt or act. I emphasize to all of our trainees to have a very low threshold for asking about suicide whenever they see someone who might have a mood disorder, an anxiety disorder, a substance abuse disorder, or a psychotic disorder. There is no evidence that any suicidal patient got the idea from a doctor, but there are many cases where a person had the idea, wasn't asked about it, and thus there was no intervention to save a life.
GOLDEN
Although I don't have hard data to back this up, my belief is that the risk of suicide occurs, not in the midst of the panic attack, but in the larger period of time between panic attacks. In these interim periods, the patient is thinking, "I have this horrible, incapacitating illness. I am afraid to go out of my house and to live my life. I am going to turn down that promotion at work because I can't bear these panic attacks coming on when I am talking to colleagues. Do I want to live a life like this?"
I believe that this is when the suicidal thinking and suicidal behavior occurs, between attacks, when patients feel their lives are being destroyed and there is no hope for overcoming the illness.
GOLDEN
It is hard to come up with a "one-size-fits-all" strategy because many factors depend on social rather than strictly clinical circumstances: for example, consider a mildly depressed college freshman, away from home for the first time, living in a large, impersonal dormitory with a roommate whom he doesn't know at all. Our follow-up for him will be more aggressive than that for a woman who is in a good, solid marriage to a spouse who is going to watch her very, very carefully, has a lot of support systems, and is in a safe environment with more monitoring. The latter patient may have a more severe depression than the former, but the former is the patient whom you would want to see frequently. How frequently? It depends. When it comes to suicide prevention, brief, frequent contact is a reasonably good safety net. You don't need to spend an hour a day, 7 days a week, with an outpatient who has some suicidal ideation. It might be that you or a nurse need to touch base over the phone every day to see how he is doing. If he seems to be doing fine, and a relative gets on the phone and confirms that things are heading in the right direction, then a brief phone call might be enough. If any red flags come out of that brief contact, then, of course, we would have the patient come in for a more extensive face-to-face evaluation. I think that frequent, brief follow-up is oftentimes a safer plan than less frequent, more extensively involved follow-up in the suicide prevention component.
GOLDEN
There are several factors: underlying psychiatric dis-order and its severity; history; and current emotional state. The latter can cut across psychiatric diagnoses -- a patient with depression or schizophrenia can also feel anxious and agitated. Comorbidities, both medical and psychiatric, are very important. Someone with depression plus alcoholism is at greater risk than someone with depression and no alcoholism. Someone with depression and chronic pain syndrome is at great risk. Someone with depression and any chronic, debilitating, incapacitating medical syndrome is at great risk. Someone with underlying dementia or any form of organic impairment is also at risk. And another very important factor, just to reemphasize, is the patient's social circumstances. Is he alone and isolated? Is he living with an irresponsible housemate, someone who has problems of her own? Or is he in a caring environment, both at home and in the community, where people will be looking out for him? All these things have to be taken into account.
GOLDEN
There clearly is a familial association, and that is why a very important part of the assessment is asking whether there is a family history of suicide. It is not clear whether the association is genetic or environmental. If it is genetic, it might be because there are certain diseases such as bipolar disorder, schizophrenia, depression, or alcoholism that have a genetic component and a risk for suicide. Or it could be that there is simply a genetic-component for self-destructive, impulsive, aggressive acts, due to, perhaps, serotonin dysregulation. On the other hand, some studies suggest that persons with a nonbiologic relative -- a spouse or adoptive parent -- who commits suicide are at increased risk. Like so many things, it is probably a combination of nature and nurture, but there clearly is a familial risk factor, which may include both a genetic component and a learned or environmental or psychologic component.