Suicide in Teenagers: Title and subTitle BreakAssessment, Management, and Prevention

Marie, a 16-year-old white girl, had no academic or social problems until the sixth grade. Her prenatal, birth, and perinatal periods were uncomplicated, and her medical and psychological development were normal, according to her parents and school records. Marie was the oldest of 3 children. Both her parents had successful careers, and there were no acute or chronic stressors in her family. Her mother characterized her as a youngster who wanted to try all sports and performing arts and was academically precocious.

Marie was diagnosed with Osgood-Schlatter disease in the sixth grade. Because of Marie's restricted physical activity, her mother's concern about depression, and Marie's telling a sixth-grade friend that she wished to die, she was seen by a psychologist 5 times. Careful evaluation found no evidence of neglect, domestic violence, abuse (including sexual), or early traumas. No diagnosis was made. The patient developed cold-induced asthma at 12 years of age and occasionally was treated with inhalers. Between the 6th and 12th grades, Marie was not treated by a psychologist or psychiatrist. Between the seventh and ninth grades, she gained excessive weight, but in the months preceding her death, she had lost 9 kg. No evidence of any type of eating disorder was noted by any family member. During the 6 months before her death, her parents were unaware of any thoughts she might have had of hopelessness or worthlessness or any suicidal ideation or suicide attempts. In retrospect, her mother realized that her daughter had almost all the symptoms of major depression ( Article ).¹ Marie's diary entries reflected that her concentration was deteriorating and that she had visited an apartment rooftop on one occasion. There was no evidence of periods of mania, hypomania, thought disorder, or illegal substance abuse, although she may have abused caffeine.

The family history included major depression successfully treated in Marie's biological mother, and her paternal great aunt died in circumstances suggestive of suicide.

Specific Factors
Previous suicide attempt
Mood disorder (major depressive disorder)
Substance abuse disorder (particularly in males)
Aged 16 years or older, male, and living alone
History of physical or sexual abuse
Less Specific Alarming Factors
Recent dramatic personality change
Psychosocial stressor (trouble with family or friends or a disciplinary crisis)
Writing, thinking, or talking about death or dying
Altered mental status (agitation, hearing voices, delusions, violence, intoxication)

During the weeks before the suicide, Marie had significantly impaired sleep. However, her sleep patterns were not characteristic of mania, in which patients sleep only 2 to 4 hours per night. Marie had skipped several days of school because of an upset stomach in the months before her death. She was given a flu shot by her family physician 1 month before her death. According to her parents, the patient was using a nonprescription caffeine product to stay awake and complete more schoolwork.

On the morning of her death, she argued on the telephone with a male friend who lived in another city and with whom Marie was not romantically involved. This friend was well known to the family. By all observers' accounts, she was heterosexually oriented. She e-mailed a suicide note to the male friend, but by the time he received it and called the family, it was too late. Marie's only previous recent verbalization regarding suicidal thinking (6 weeks before her death) was to a former camp counselor, who failed to relay the information to the parents.

On the day of her death, shortly after sending the suicide message, Marie climbed to the roof of a neighboring apartment building and jumped off.

The loss of a talented, bright, and highly motivated 16-year-old highlights the insidious nature of the individual risk factors that affected this adolescent's life. Both the patient's sex and the lethality of the method are noteworthy in this case.

Marie's parents, both professionals and 1 in treatment for depression, were never informed by the camp counselor of suicidal ideation. The adolescent was driven and accomplished but was incapable of sharing her thoughts with her parents or family physician. The immediate stressor, interpersonal loss, is a well-known risk factor but is typical of the average 12th-grade adolescent. The 2 most common familial risk factors, positive family history of suicide and a parent with a psychiatric disorder, were present. Postmortem toxicologic analysis revealed no evidence of drug abuse or pregnancy.

Our review highlights how this case is typical and atypical and why suicide in teenagers may be difficult to predict or prevent. Many risk factors may not be identified in advance. Suicide in teenagers occurs even in the most supportive family environment, such as that of the family described, in which an excellent scholar-athlete was given both parental attention (the family ate dinner together every night) and independence.

The rate of suicide among adolescents has significantly increased in the past 30 years. In 1998, 4153 young people aged 15 to 24 years committed suicide in the United States, an average of 11.3 deaths per day. Suicide is the third leading cause of death in this age group and accounts for 13.5% of all deaths.² Centers for Disease Control and Prevention data show that females contemplate and attempt suicide at much higher rates than males. However, older male adolescents complete suicide 5.5 times more often than female adolescents.³ This incongruence is believed to be due to the tendency of males to use more lethal means for suicide. Children younger than 10 years are less likely to complete suicide, and the risk appears to increase gradually in children between 10 and 12 years of age. However, on average, 170 children 10 years or younger commit suicide each year. Exact figures in the youngest age group are probably underestimates because of lack of recognition of suicide.

A number of facts exacerbate the difficulty clinicians face in identifying suicidal patients: (1) the low prevalence of suicide (0.01%) in the general population; (2) the large percentage (99.9%) of depressed patients who do not commit suicide, despite a 10-fold increase of suicide in adult patients with depression² ; and (3) the high percentage (27%) of suicidal ideation in the general teenaged population.⁴ Rare events are hard to predict. In an attempt to guide clinicians, a large body of evidence has been collected on risk factors, with criteria for assessing risk.

Previous attempt is a major risk factor. Major depressive disorder and other mood disorders, such as bipolar disorder, are most often cited as a risk factor for suicide ( Article ). In one study, Shaffer et al⁵ reported that 52% of subjects who committed suicide met criteria for major depressive disorder, whereas in total, 61% met criteria for a mood disorder (eg, manic depressive disorder and major depression). Substance abuse disorder, conduct disorder (delinquency), and anxiety disorder are also risk factors. Impulsivity and aggressiveness are personality traits frequently seen in persons who attempt suicide. A patient with a history of a suicide attempt, however minor (so-called gestures), is at an increased risk of a repeated attempt. Family history of suicide also increases risk.⁶

Depression
Depressed mood most of the time
Loss of interest or pleasure in activities
Weight loss or gain
Insomnia or hypersomnia
Loss of energy
Feelings of worthlessness
Hopelessness toward future
Lack of concentration
Recurring thoughts of death
Complaints of psychosomatic symptoms
Mania
Elated or irritable mood
Inflated self-esteem, grandiosity
Decreased need for sleep
Pressured speech
Racing thoughts
Distractedness
Excessive participation in multiple activities
Hypersexuality, impulsive spending, uninhibited remarks

*See reference 1.

Finally, one of the distinguishing features of adolescent suicide is that it may be precipitated by a psychosocial stressor, such as a recent loss, rejection, or disciplinary crisis.³ However, these events are common in a normal adolescent's life.

Although suicide may run in families, this observation does not in itself prove a genetic basis for this behavior. Both familial and genetics studies will be reviewed.

Familial Findings. Brent et al⁶ examined 58 adolescent suicide proband subjects and 55 similar control subjects. The rate of suicide attempts in the first-degree relatives of suicide proband subjects was increased compared with that of the control subjects, even after the increased rates of psychiatric disorders in the families of suicide attempters was considered. Tishler et al⁷ reported that 22% of adolescents treated at an emergency department after attempting suicide had at least 1 family member who had exhibited suicidal behavior.

Although some familial factors associated with suicide may be identical to those responsible for psychiatric disorders, such as mood disorder, some are independent of the mood disorder. For example, in their study of depression and suicide in Old Order Amish communities, Egeland and Sussex⁸ documented clusters of suicides in 4 families that also exhibited heavy loading for mood disorders. However, their study also identified families who were heavily loaded for mood disorders yet had no history of suicide. This finding led them to hypothesize that mood disorders and suicidal behavior were inherited independently.

Genetic Factors. In adoption studies in a Danish population,⁹ 12 of 269 biological relatives of adoptees who had committed suicide had themselves committed suicide, whereas only 2 of 269 biological relatives of adoptees who had not committed suicide had killed themselves. In this study, no adopting relatives had died by suicide. In a separate study, Papadimitriov et al¹⁰ interviewed patients with a history of mood disorders and suicidal behavior and their relatives. Using computational analysis, the authors concluded that suicidality is most compatible with polygenic inheritance.

In support of this finding, studies of twins have revealed that monozygotic twins have a 13.2% concordance for suicide, whereas dizygotic twins have only a 0.7% concordance for suicide.¹¹ A separate study¹² demonstrated that monozygotic twins were no more likely than dizygotic twins to attempt suicide when one twin had died of causes other than suicide. The heritability of serious suicide attempts was recently estimated at 55%.¹³ Despite growing knowledge about genetic factors, no specific marker provides clinical usefulness yet.

Studies have linked serotonin with the control of impulsivity, aggression, self-mutilation, and depression. Most studies comparing suicide attempters with nonattempters have found lower levels of 5-hydroxyindoleacetic acid (5-HIAA), a metabolite of serotonin, in the cerebrospinal fluid (CSF) of suicide attempters.¹⁴ Arango et al¹⁵ reported decreased serotonin and 5-HIAA levels in the brain stems of suicide attempters. Genetic studies have suggested that suicidal acts¹⁶ and low CSF 5-HIAA levels¹⁷ are related to a polymorphism in the TPH gene, which codes for tryptophan hydroxylase, the rate-limiting enzyme in the synthesis of serotonin. These studies suggest that low CSF 5-HIAA levels are due to low central serotonin turnover, which is caused by a reduced capacity to hydroxylate tryptophan in the synthesis of serotonin. The authors speculate that the presence of a certain allele of the TPH gene might lead to this reduced capacity.

Although several groups have reported an association between the TPH gene and suicide, the literature disagrees about which allele is associated with this behavior.^{17 - 20} Further, other studies^{21 - 22} did not reveal a significant difference in the TPH polymorphism yet did reveal one in the serotonin transporter gene of suicide completers.

Several researchers have reported physiological changes in the serotoninergic system of suicide completers: increased serotonin 1A¹⁵ and serotonin 2A receptors^{23 - 24} and decreased serotonin transporters²⁵ in the prefrontal cortices of suicide completers. It is unclear, however, whether these changes are caused by a certain genotype or whether they are the result of low levels of serotonin, exposure to psychotropic medications, or another mitigating factor. One such factor is chronic stress or psychological trauma. Attempts to identify clinically useful biological markers are under way, but at present no biological measures are clinically useful.

There are no evidence-based data that psychiatric hospitalization prevents immediate or eventual suicide, despite overwhelming clinical consensus that immediate hospitalization is a critical component in preventing both adult and teenaged suicidal patients from completing suicide. In one study that investigated the efficacy of hospitalization,²⁶ parasuicidal adult patients were randomized to home or a hospital. No significant difference was found in outcome as measured by subsequent suicide or general functioning. In a separate controlled trial of adult suicide attempters who were randomized to admission or discharged home, no significant difference in repetition rate was noticed after 16 weeks of follow-up.²⁷ This study has not been done in teenagers.

No treatment program has reduced subsequent attempts in adolescent suicide attempters.²⁶ In addition, no controlled studies have demonstrated that treating conduct disorder or substance abuse reduces the number of future suicide attempts, which is due to the lack of established treatments with proved long-term efficacy for these disorders.²⁶

Adolescents are generally not compliant with psychiatric treatment. One study²⁸ found that 40% of suicidal teenagers are removed from treatment for not attending therapy. A separate 3-month follow-up study²⁹ of 62 adolescents discharged from an emergency department after a suicide attempt found that 52% of them failed to attend more than 2 sessions of therapy, and 16% of patients never attended outpatient therapy. Suicidal and nonsuicidal children dropped out of treatment, but the suicide attempters ended therapy earlier than nonsuicidal children.³⁰

In a review of 4 novel approaches to intensive outpatient care, Greenhill and Waslick²⁶ determined that none of the follow-up or treatment programs was superior to the others. The studies involved adults, and there are no comparable controlled trials in adolescent suicide attempters. Only one study³¹ of cognitive behavior therapy included adolescent suicide attempters, and this study demonstrated modest effects compared with those of a brief, problem-oriented therapy.^{3 ,26} For multiple reasons, it is highly problematic to assume that the cognitive behavior therapy performed on adult study subjects can be transferred to adolescents, given the critical role of family factors in adolescent life and the fact that, developmentally, this age group might be incapable of using this approach. Nevertheless, without a clear-cut solution, a prudent approach appeared to be conscientious clinical follow-up of teenagers to ensure that they were engaged in treatment.²⁶

Unless the general practitioner, internist, or pediatrician has particular interest, training, or expertise in suicide prevention, teenagers with suicidal ideation or multiple risk factors or who have attempted suicide should be referred for a complete mental health evaluation and careful treatment. In areas with few psychiatrists available, careful communication between physicians and psychologists or psychiatrists is essential.

Both older and newer selective serotonin reuptake inhibitor (SSRI) antidepressants unequivocally reduce symptoms of major depression and generalized anxiety in adults. There is also clear-cut and overwhelming evidence that the descriptive phenomenology of unipolar and bipolar disorders in adolescents is essentially identical to that in adults.³² However, historically, experts in adolescent depression have had difficulty demonstrating that antidepressants reduce symptoms of major depression and generalized anxiety in this age group. They cite the high placebo response rate in teenagers as a methodologic hurdle to demonstrating efficacy in this subpopulation. Recent studies^{33 - 34} have emerged with more definitive positive results of the efficacy of antidepressants in a teenaged population.

Despite the evidence base for the efficacy of antidepressants in depression, little evidence exists that antidepressants significantly lower suicide rates in adult patients treated with antidepressants.⁴ Tondo et al³⁵ state, "Despite broad clinical use and intensive study of antidepressants for four decades, evidence that they significantly alter suicidal behavior, or reduce long term suicidal risk, remains meager and inconclusive." The introduction of SSRIs and other modern antidepressants that are much less toxic in acute overdose than older drugs appears not to have been associated with a decrease in suicide rates. It is therefore safe to say that there are no evidence-based data to suggest that tricyclic antidepressants or SSRIs prevent suicide in teenagers. However, suicide rates began to plateau in the 1990s, starting with the widespread introduction of SSRIs in the United States.³⁶

Benefits of lithium augmentation for depression have been reported in 2 studies^{37 - 38} in adolescents, although less than half the subjects seemed to respond to this additional treatment. There is strong and conclusive evidence that in adults with bipolar disorder, lithium as part of long-term treatment provides "substantial protection against suicide attempts and fatalities" (relative risk reduction ratio of 7.18 in a meta-analysis of 22 studies).³⁵ In addition, "[a] recent international collaborative study found that discontinuation of lithium maintenance was associated with sharp increases in suicidal risk in a large sample of Bipolar I and Bipolar II patients."³⁵ This study found that rates of suicide completions and attempts increased 20-fold after lithium use was discontinued.

Bipolar illness commonly presents in adolescence with classic adult features (grandiosity, pressured speech, decreased sleep, agitation, intense irritability). Much academic controversy exists about the identification of bipolar illness in very young adolescents and prepubertal children. Lithium has not been subjected to large-scale studies in adolescents with childhood depression, and where it has been studied, clear-cut effects have been noted on aggression but not mood. There is no evidence on lithium's effectiveness in reducing suicide in teenagers, but the evidence in adults is strong.

The possibility that pharmacotherapy, especially SSRIs, may increase suicidal thinking or behavior is reviewed in a meta-analysis of 17 studies (of adults) by Beasley et al.³⁹ They could not show any evidence that suicide rates were different between patients treated with SSRIs (fluoxetine), tricyclic antidepressants, or a placebo. No evidence exists that the results would be different in teenagers.

As summarized by Goodwin and Ghaemi,⁴⁰ only 2 studies of adults address the impact of anticonvulsants (mood stabilizers) on suicide. Despite the effectiveness of carbamazepine and valproate in controlling acute manic episodes, one study showed that carbamazepine was less effective than lithium in preventing adult suicide. No data exist for valproate.⁴⁰ Psychosis as part of bipolar illness, psychotic depression, or adolescent-onset schizophrenia clearly is associated with increased suicidal behavior in adults, and the use of antipsychotic drugs in adolescent psychosis is clearly efficacious for psychotic symptoms.

The practice parameters of the American Academy of Child and Adolescent Psychiatry³ summarized the limited empirical data on the measurable effects of school-based programs and hot-line crisis services on suicidal behavior. "Early studies . . . failed to show that hotlines reduce the incidence of suicide . . . and their value remains untested. . . . Suicide awareness programs in schools . . . have not been shown to be effective either in reducing suicidal behavior or in increasing help seeking behavior."³

The prevalence of suicide has leveled off in the United States during the past 10 years,² a trend that correlates with significant increases in the use of antidepressant medications in adolescents and young adults.³⁶ However, the effects of widespread use of these newer antidepressants on suicide rates of teenagers and whether the leveling-off trend will continue have yet to be ascertained. Further, there are no clear criteria for hospitalizing and discharging a patient at moderate risk for suicide, and community-based suicide prevention programs have not yet been clinically proved to have an effect on suicide rates. It also is unclear what type of psychotherapy efficaciously prevents suicide attempts. Surprisingly, there are no data to indicate that the treatment of a patient's underlying condition (eg, conduct disorder or substance abuse) prevents suicidal behavior. Anecdotal reports have not shown that the penetration of managed care has increased the number of teenaged suicides, yet it is uncertain whether closer scrutiny of patient admissions and pressure for early discharge will affect the rate of suicide attempts and completions. Finally, the biological basis of suicidal behavior remains unclear despite the large body of literature discussing it. The surgeon general noted⁴¹ that increased support for evidence-based treatment and prevention was needed. The Public Health Service⁴² further identifies ways in which investment in clinical research, professional training, and evidence-based community prevention should lead to reductions in suicide rates in the United States.

Although much clinical practice is based on limited research, incomplete knowledge, and extrapolations from adult studies, several recommendations can be made to physicians faced with the task of assessing and treating suicidal teenagers and preventing suicide. Physicians should know the symptoms of depression and the risk factors for suicide in teenagers and inquire about these symptoms during an office visit. Physicians should also ask about firearms and other lethal means that might be found in the home and should always screen for substance abuse, conduct disorder, and poor school performance. Physicians should pay particular attention to patients who were recently hospitalized or those with multiple hospitalizations for suicidal behavior³ and should obtain a family history of depression and suicide.

Physicians who choose to treat suicidal teenagers should not rely on contracts (ie, a verbal agreement between patient and physician that the patient will not attempt suicide). They should also be aware that suicide risk seems highest at the beginning of depressive episodes, so expeditious treatment or referral is crucial. Hospitalization may be required for suicidal teenagers with altered mental status such as psychosis or thought disorder, who have actively abused substances or have attempted suicide, who experience hopelessness or impulsivity, or who have a lack of adequate parental supervision, among many other factors. Physicians should prescribe SSRIs instead of tricyclic antidepressants for safety reasons, since SSRIs are much less likely to be lethal in an overdose. In general, adult doses may be used in older teenagers, but teenagers and family should be counseled about adverse effects, including sexual dysfunction. In addition, parents should be warned about manic activation when antidepressant medication is used. Although the data regarding the precipitation of suicidal ideation by SSRIs in adolescents are not at all clear, "the wisest course of action for the practitioner during the early stages of . . . treatment . . . is to systematically inquire about suicidal ideation before and after treatment is started, especially . . . if SSRI treatment is associated with the onset of akathisia [motor restlessness, active or subjective]."³ Physicians should also not instruct their patients to discontinue medication use when target symptoms subside. They should provide careful follow-up after medication use is discontinued. Some patients will require many years of continuous pharmacotherapy and follow-up, since depression is a chronic and recurring illness. If faced with a denial for treatment by an insurer, physicians should emphasize to the managed care organization that the patient is suicidal. For further recommendations, physicians should review the American Academy of Child and Adolescent Psychiatry practice parameters.³

If a patient commits suicide, the physician should counsel the victim's family, friends, and community. In addition, physicians should be aware of the risk of related suicides and how to guide the media to prevent "glorification" of suicide.

Finally, if suicidal thinking and behavior are signs of psychiatric disorder, then symptomatic relief or amelioration of common risk factors and underlying psychiatric illness should reduce suicide in teenagers. If health care professionals vigilantly screen for the constellation of factors that lead teenagers to commit suicide, then this rare but tragic behavior will be reduced.