Kaplan translate SOCIAL ENVIRONMENT A large proportion of older teenaged suicide completers were drifters neither attending school

nor employed. This process probably started in the mid-teens, and there is a strong relationship between high school dropout and suicide attempts. IMITATION Reading about, seeing, or hearing about suicidal behavior seems to induce imitative behavior in some adolescents. Increased suicidal behavior after exposure to suicide-related material on television, in newspapers, in cinema, and in books has been widely and regularly reported. As would be expected from imitative learning, the effect is proportional to the amount, duration, and salience of the exposure. The individual characteristics of adolescents who are susceptible to imitation are currently unknown. Suicide clusters are striking manifestations of the phenomenon and account for as much as 4 percent of teenage suicides in the United States. The American Foundation for Suicide Prevention has prepared guidelines for journalists on how to minimize risk when reporting on a youth suicide, which include minimizing the graphic display of stories on suicide, avoiding romanticization and precise description of methods, and disseminating treatment resources. Imitation might also play an important role in suicide attempts. There are consistent reports that teenagers who attempt suicide are significantly more likely to have a friend who attempted suicide or to have a family member who attempted suicide. PARENTCHILD RELATIONSHIPS AND ABUSE Suicide victims are more likely than teenagers in the general population to live with only one parent, to have parents with psychiatric illness, and to live in homes with much parentparent discord. Parental psychiatric illness accounts for a significant part of the high rates of physical and sexual abuse found in attempted and completed suicides. However, there also seems to be a specific relationship between abuse and suicidal tendency, because it is found even after accounting for demographic characteristics and parental psychopathology. Parentyouth conflict often precedes an attempt but occurs less often before a death. Conflicts before an attempt usually concern unresolved differences over limit setting. Teenagers frequently view the relationship with their parents as more dysfunctional than the parents acknowledge. It is highly likely that these stressors interact with as-yet unknown individual vulnerability factors to lead to the behavioral response style that makes a suicide attempt more likely. Individual Differences PSYCHOPATHOLOGY

Although few adolescents (or adults) commit suicide in the absence of a psychiatric disorder, few individuals P.3271

with a psychiatric disorder commit suicide. Psychopathology is therefore a necessary but not sufficient cause of suicide. Information about prior psychopathologies is usually obtained through the psychological-autopsy method, which obtains accounts of symptomatic behavior and talk of the suicides from those who had a chance to observe the victims frequently during a period before their deaths. They show that approximately 90 percent of children and teenagers who commit suicide were experiencing a psychiatric disorder at the time of death. The disorder is usually well established and, in approximately one-half of suicides, has been present for 2 or more years. Major depression is the most significant risk factor for suicide in girls, increasing the risk of suicide 20fold. A previous suicide attempt increases the risk for suicide in girls threefold. A prior suicide attempt is the strongest predictor of completed suicide in boys, increasing risk 30-fold. Many young persons who commit suicide with a mood disorder also have a comorbid conduct disorder or substance and alcohol abuse, or both, which, alone or comorbid, is present in as much as two-thirds of older boys who commit suicide. Conduct or oppositional disorder is present in between one-third and one-half of suicides, somewhat more frequently among boys and older teenagers. Although the suicide rate is greatly increased in schizophrenia and bipolar illness, because of their relative rarity, they account for fewer than 5 percent of adolescent suicides each. Many severely depressed patients are not suicidal, whereas others with only mild or transient depression or, indeed, no depression at all can make serious suicide attempts or successfully commit suicide. This has led some to conclude that suicidal tendency should be regarded as an independent condition. In support of this, the biological determinants of suicide appear to be different from those of depression and are found in suicidal patients regardless of their associated diagnoses. The psychiatric profile of suicide attempters is, in many ways, similar to that of suicides. Almost all teenage attempters have a psychiatric illness or have previously experienced one. Most commonly, this is a mood disorder, often comorbid with a conduct or anxiety disorder or, much less commonly, with an eating disorder. Adjustment disorders are also common. Subthreshold symptoms might increase rapidly in response to a stress but return to normal shortly after the attempt. Suicide attempts are significantly more common in heavy smokers. Heavy drinking increases the likelihood of a suicide attempt nearly tenfold. Alcoholism is three times more powerful as a risk factor for attempts than is physical or sexual abuse. PSYCHOLOGICAL FACTORS

Most emphasis has been put on the psychological constructs of hopelessness, impulsivity, aggressiveness, and imitation. Few persons who commit suicide have been studied in this regard before their deaths, so much of the information has been obtained from studies of suicide attempters, and its generalizability to suicide completion is uncertain. Hopelessness The relationship between suicidal tendency and hopelessness remains an area of controversy. Large longitudinal studies have failed to find a relationship between the two after controlling for depression. Those that have found a relationship have noted it to be strongest among girls, who constitute a majority of attempters and a minority of completers. The evidence for impulsivity owes more to the study of suicidal decisions and to the nature of the biological abnormalities in suicides than to standard laboratory measures. In one series of consecutive emergency department admissions, only 10 to 15 percent reported thinking about their attempt for more than a day, and 15 percent of suicide attempters reported no prior planning. However, this might not reflect a trait-like cognitive set, but rather extreme mood volatility. Aggressiveness The relationship between suicidal tendency and aggressiveness has been reported frequently and is in line with the biological findings. The aggressiveness of the suicidal youth might be more than a manifestation of current depression. Aggressive 8-year-olds have been shown to have twice the rate of later ideation and suicidal behavior than matched nonaggressive controls. Sexual Orientation Large community studies have shown that gay, lesbian, and bisexual (GLB) youth are between two and six times more likely than heterosexual adolescents to think about and to attempt suicide. In a nationally representative sample, 9 percent of U.S. GLB adolescents had made an attempt in the last year. In a Massachusetts State sample, as many as 35 percent had made an attempt. The mechanisms of this relationship remain largely unknown. GLB youth are more likely to be victimized at school, but they also have higher rates of drug and alcohol use and higher rates of psychiatric disorders, including major depression, generalized anxiety disorder, and conduct disorder. All or any of these factors might account for some or all of the strikingly increased rates of suicidal tendency among GLB teenagers. BIOLOGY Family History and Genetics Completed teenage suicides and suicide attempters are between two and four times more likely than matched controls to have a first-degree relative who committed suicide. Twin studies demonstrate that this is not simply a consequence of being reared by a psychiatrically ill parent. The increased risk of committing suicide among identical cotwins of suicides is approximately 11 to 1, compared to twice the risk in nonidentical twins.

SEROTONERGIC ABNORMALITIES A great many studies have pointed to abnormalities of serotonin function in suicidal individuals and in impulsive, aggressive individuals. The original investigators pointed to low levels of serotonin metabolites in urine and cerebrospinal fluid (CSF). Subsequent autoradiographic and neuroanatomical studies have shown a reduction in the overall density of serotonin type 1A receptors and serotonin transporter receptors (which regulate serotonin uptake) in the prefrontal cortex and in the dorsal raphe nucleus, the source of serotonin innervation of the prefrontal cortex. Several genetic mechanisms have been proposed to explain the serotoninergic abnormalities associated with suicidal tendency. These include polymorphisms in the serotonin transport gene and the serotonin type 1A receptor gene and polymorphisms on intron 7 of the gene for tryptophan hydroxylase (TPH) (the rate-limiting enzyme for the biosynthesis of serotonin). The suicide phenotype is probably heterogeneous and complex, the genetic effects are probably small, and it is likely that a single genetic variant is less important than patterns of variance. Because this relationship is independent of diagnosis, it is thought to be a marker of an underlying trait. A plausible mechanism would be that these abnormalities or differences modify the response of a mentally ill person to stress, resulting directly in suicidal behavior or in a response that generates adverse consequences and an escalating cascade of stresses and aggression, culminating in suicidal behavior. The proportion of suicides that fall into this category is not known. There has been a single promising report indicating that attempters who have low levels of serotonin metabolites in the CSF are significantly more likely to make a future attempt or to commit suicide, P.3272

or both, but, before these findings can be used for prediction or prevention, replication is needed, along with collection of information on the base rate of these abnormalities in the general nonsuicidal population. PERINATAL MORBIDITY A relationship between perinatal morbidity and subsequent suicide has been reported in several, but not all, studies that have examined this. The mechanism of any relationship is unknown. Treatment of Suicidal Tendency Past and Current Treatment Lifetime contact with mental health professionals ranges from 25 to 60 percent. A small minorityapproximately 15 percentof suicides occurs during the course (within a month) of active psychiatric treatment.

Approximately one-third of suicide attempts result in a clinical evaluation and later treatment. Female attempters are more likely to be treated than male attempters, and Hispanics are more likely to be treated than blacks or whites. Initial Management Suicide attempters are commonly seen in an emergency department, where, after medical stabilization, a decision needs to be made about whether they need continued observation or can return home. ASSESSMENT The prognostic significance of a suicide attempt ranges from benign, with few, if any, negative sequelae, to malignant, with later suicide. These various outcomes are probably dependent on the mental context in which suicidal behavior occurs, and they should be the focus of the clinical evaluation. Suicidal tendency is not, in itself, an adequate diagnosis. Approximately one-fourth of attempters have no evidence of sustained psychopathology. A number of suicide attempters have what can best be described as an adjustment disorder. Between stressors, they might have subthreshold symptoms, but their mental state changes abruptly in response to a stress and then recovers to its former mildly abnormal level shortly after the attempt. The clinician's greatest concern when evaluating attempts is to differentiate between those with a benign prognosis and those with a malignant prognosis. Because of the frequency of ideation and attempts and the rarity of completed suicide, it is statistically reasonable to regard all attempts as benign. However, few clinicians assume this actuarial position. The problem facing clinicians is that a teenager's intent is difficult to gauge before or after the event. Teenagers are poorly informed about the risks that they take. Between 30 and 50 percent of teenagers seen in an emergency department after an attempt say that they wanted to die, even though few took active precautions against being discovered. The teenager and the clinician frequently differ on how dangerous the behavior was. In one study of adolescents who said that they wanted to die, psychiatrists drew the same conclusion in only one-third. This raises the question of if and when it is reasonable to challenge a patient's statement about intent. Intent cannot be easily inferred from the size of an overdose. Normal and disturbed adolescents frequently misjudge the potential lethality of ingestions, in some instances overestimating lethality and, in others, underestimating lethality. In a study that compared attempter and pediatrician estimates of the seriousness of an ingestion, teenagers rated 26 percent of attempts as having lethal potential, but only 2 percent were considered potentially lethal by the pediatricians. Prior planning is often used to assess intent, but suicide attempts and suicides commonly occur shortly after a stressful event, and it is unlikely that they were carefully planned. The Youth Risk Behavior Survey regularly shows that threefourths of teenagers who ideated suicide had made a plan about how to commit suicide. Most of these teenagers do not go on to attempt or commit suicide, and a study of nonplanners shows that they have a similar attempt rate and also show other problems.

Continued observation is indicated for attempters with an abnormal mental state, especially those who are agitated, depressed, irritable, or psychotic, and for the small proportion who declare a persistent wish to die. The features that increase risk for later suicide are listed in Table 45.1-2, and the decision to admit a patient is often made by an imprecise juggling of these risks. The short to medium risk of suicide is small in prepubertal children, and suicidal ideation or behavior should not, by itself, be seen as an indication for hospitalization in that age group. Table 45.1-2 Assessing Child and Adolescent Suicide Attempters (Emergency)

--------------------------------------------------------------------------------

Attempters at greatest risk for suicide Do not discharge without a psychiatric evaluation Mental state Depressed, manic, hypomanic, severely anxious, or a mixture of these states Substance abuse alone or in association with a mood disorder Irritable, agitated, threatening violence to others, delusional, or hallucinating Suicidal history Still thinking of suicide Has made a prior suicide attempt

--------------------------------------------------------------------------------

There are no studies that demonstrate the positive or negative impact of hospitalization, although it has been shown that, within an adolescent ward, attempters tend to seek one another's company, just as they do in other settings, such as college. Deciding whether to admit a currently euthymic recent attempter often depends on essentially nonclinical considerations, such as insurance coverage and bed availability. When an admission can only be brief and can only be to a distant hospital that does not

provide wraparound care, it likely has less value than admission to a site that offers more significant therapeutic intervention. Regardless of mental state, no teenager should be discharged from the emergency department without an interview with a responsible adult, ideally, the caretaker. The youth should be discharged to the care of the adult, who should be told to secure or to remove dangerous medications and firearms. Parents accept and comply with these instructions when they are given clearly but might not think to do so otherwise. It is reasonable to hold the young patient in the emergency department until weapons and medications have been secured or discarded. Patients are often required to sign a contract for safety as a condition for discharge. This contract requires the patient to contact a clinician before engaging in any further suicidal behavior. The impact of these contracts is not known, and, in one study of 135 repeat attempters, 31 percent had previously signed a contract for safety. The emergency department offers an opportunity to obtain a detailed account of the events and feelings that preceded the attempt and its consequences at the first visit. This information can be helpful for planning therapy and is best obtained while memory of the event is fresh. Outpatient Care Only a minority of attempters seen in an emergency department subsequently engage in substantive psychiatric treatment. Unfortunately, noncompliance with offers of further P.3273

treatment is most common in attempters who are symptomatic and who continue to think about suicide. Factors that might improve adherence to planned treatment include (1) teaching the family about the condition and its treatment; (2) dealing with factors that are not directly related to the patient's suicide attempt, such as family strife; (3) offering follow-up appointments from the emergency department before discharge; and (4) establishing procedures to remind the patient of future appointments and proactive follow-up after missed appointments. Psychotherapy Only a small number of psychotherapy studies have been conducted among teenage suicide attempters. Treatments that have failed to reduce the attempt-repetition rate include problem solving, enhanced access to clinical service, and home-based family therapy. CBT, widely tested for depressed adolescents, has not been assessed in suicidal children, but studies of its effect in suicidal adults have not been promising.

Dialectical behavior therapy (DBT) has been shown to reduce repeat attempts in adults, but its effect on teenagers is not known. DBT is a complex and lengthy behavioral intervention given in individual and group contexts. DBT builds on a well-established contemporary model of suicidal tendency, so that its components are worth considering. These include (1) mindfulness training to improve self-acceptance, (2) assertiveness training to reduce interpersonal conflicts, (3) training the patient to avoid situations that trigger negative moods, and (4) increasing tolerance for psychological distress. CBT has been used successfully in adolescent patients with depression, but no studies of individualized CBT with adolescent suicide attempters have been published. Pharmacotherapy Some suicide attempters are extremely volatile and make frequent suicidal attempts. In such cases, small doses of an antipsychotic medication have been shown to be especially useful. SSRI antidepressants have been shown to reduce suicidal ideation in depressed and nondepressed adults with cluster B personality disorders and in individuals who have made a limited number of previous suicide attempts. There is some ecological evidence that links the increased rate of use of SSRIs by teenagers to the decline in the teenage suicide rate. There are anecdotal reports suggesting that at least some antidepressants might have precisely the opposite effect in teenagers. However, they have been shown to be more effective than placebo in treating depressed teenagers, and they reduce the frequency of impulsive and aggressive behaviors, which are common in suicidal teenagers. They are considerably less dangerous in overdose than tricyclic antidepressants. Lithium has been shown to exert a powerful antisuicide effect in bipolar adults. However, when a bipolar patient is withdrawn from lithium suddenly, the risk of suicide increases, even if manic symptoms do not appear. Clozapine (Clozaril) has been found to be effective in reducing suicidal tendency in schizophrenic adults, even when there is no apparent effect or impact on other symptoms of schizophrenia. The antisuicidal effects of lithium and clozapine have not been assessed in children or adolescents. When lithium is being used to treat an adolescent, caution should be used when the drug is withdrawn, and sudden withdrawal should be avoided. Iatrogenic Suicidal Tendency Uncontrolled reports on the induction of suicidal tendency by SSRI antidepressants appeared shortly after their introduction. In some, but not all, cases, suicidal ideation was reversed after treatment was discontinued. However, metaanalyses of many thousands of children and adults suggest that the overwhelming effect of SSRIs is to reduce suicidal ideation. Although it seems likely that instances of suicidal tendency during treatment reflect the fact that suicidal tendency is closely linked to suicide, this might not always be the case. Clinicians should systematically inquire about suicidal ideation before and after treatment is started and should be alert to the possibility of suicidal tendency.

Clinicians should generally be cautious about prescribing medications that can reduce self-control, such as the benzodiazepines, in patients who have made a suicide attempt. Phenobarbital (Barbita) also has a high lethal potential if taken in overdose. Benzodiazepines might disinhibit some individuals, who then exhibit aggression and make suicide attempts. There are indications of similar effects from the antidepressants maprotiline (Ludiomil) and amitriptyline (Elavil), the amphetamines, and phenobarbital. SUGGESTED CROSS-REFERENCES Chapter 13 provides an exhaustive discussion of mood disorders, and suicide is discussed further in Section

45.2: Early-Onset Bipolar Disorders Joseph Biederman M.D. Part of "45 - Mood Disorders in Children and Adolescents" Since the 1980s, the view that bipolar disorder in children is extremely rare or nonexistent has been increasingly challenged by many case reports and series. G. Robert DeLong and G. W. Nieman described a series of children with severe symptoms that highly suggest that mania is responsive to lithium carbonate (Eskalith). Gabrielle A. Carlson suggested that prepubertal mania may be characterized by severe irritability, absence of episodes, and high levels of hyperactivity. Similarly, Hagop S. Akiskal and colleagues reported on the case histories of large groups of adolescent relatives of classic adult bipolar patients. They found that, despite frank symptoms of depression, mania, and frequent mental health contacts, none of these youth had been diagnosed with an affective disorder. Ronald A. Weller and colleagues then reviewed more than 200 articles published between the years of 1809 and 1982. They identified 157 cases that would likely be considered manic by modern standards; however, 48 percent of those subjects retrospectively diagnosed as manic according to the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) criteria were not considered so at the time of referral. Taken together, these reports suggested that pediatric mania may not be rare but is difficult to diagnose. PREVALENCE AND COURSE Classically defined bipolar disorder occurs infrequently in young people. The National Comorbidity Survey (NCS) estimated the lifetime prevalence of the revised DSM-III (DSM-III-R) bipolar I disorder in 15- to 54-year-olds at 0.4 percent and the 1-year prevalence at just under that figure. A study conducted by Peter M. Lewinsohn and colleagues found a lifetime prevalence of bipolar disorders in 1 percent (mainly bipolar II disorder) of adolescents 14 to 18 years of age. Retrospectively recalled age at onset was approximately 12 years of age. A community-based study that has examined recent or current prevalence has found no cases of bipolar I disorder before 13 years of age. Estimates from these studies are uncertain for various reasons. First, few studies possess adequate population samples to estimate prevalence accurately. Second, the definitions of the term bipolar have varied, with some authors referring to the spectrum of bipolar disorders (bipolar I disorder, bipolar II disorder, hypomania, and cyclothymia), whereas others have reported just the rates of bipolar I disorder (full blown, acute manic episodes). Third, controversy remains concerning possible differences between clinical presentations of childhood bipolar disorder, which may present typically as a chronic and unremitting condition, versus adult bipolar disorder, which may often be characterized by clear periods of relapse and recovery. Few studies have followed the same individuals over time to resolve this controversy by providing a careful, developmentally informed bipolar diagnosis. Hence, the actual population prevalence of prepubertal or adolescent bipolar disorder is unknown. The appropriate diagnosis for prepubertal children showing symptoms of mood lability, extreme irritability, or rapid mood cycles remains an unresolved question. Is this an early form of bipolar

disorder, or should it be classified differently? Fueling this controversy, investigators, clinicians, and advocacy groups have increasingly challenged the view that bipolar disorder in children is extremely rare. Part of this controversy may hinge on what is considered to be part of the bipolar spectrum. Although the adult course of pediatric bipolar disorder awaits data from longitudinal studies, the adult literature provides some interesting clues on the subject. The similarity of 12-month and lifetime rates of bipolar disorder in adults from the NCS suggests that, once established, this is a chronic disease. For clinically referred children and adolescents, it also appears that bipolar disorder is a chronic disease. Recently, Barbara Geller and colleagues reported results from a 2-year prospective follow-up of children with prepubertal bipolar disorder. Results show a high rate of chronicity and relapse in this population that is consistent with the chronic course and poor outcome documented in adults with mixed mania, continuous cycling, and psychosis. A review by Susan L. McElroy and colleagues described mixed mania, which affects 20 to 30 percent of adults with mania. Subjects with mixed mania tend to have a chronic course, absence of discrete episodes, onset of the disorder in childhood and adolescence, a high rate of suicide, poor response to treatment, and an early history of neuropsychological deficits that is highly suggestive of attentiondeficit/hyperactivity disorder (ADHD). Thus, McElroy and colleagues identified a manic syndrome in adults that shows the atypical features of pediatric bipolar disorder. These findings suggest that pediatric cases with bipolar disorder may develop into adults with mixed mania. ATYPICALITY OF PEDIATRIC BIPOLAR DISORDER The atypicality (by adult standards) of the clinical picture of childhood mania has long been recognized. The type of irritability observed in manic children is severe, persistent, and often violent. The outbursts often include threatening or attacking behavior toward family members, other children, adults, and teachers. In between outbursts, these children are described as persistently irritable or angry in mood. Thus, it is not surprising that these children frequently P.3275

receive the diagnosis of conduct disorder. Aggressive symptoms may be the primary reason for the high rate of psychiatric hospitalization noted in manic children. In addition to the predominant abnormal mood in pediatric bipolar disorder, its natural course is also atypical compared to the natural course of adult bipolar disorder. The course of pediatric mania tends to be chronic and continuous rather than episodic and acute. For example, in a recent review of the research conducted in the 1990s on pediatric mania, Geller and Joan Luby concluded that childhoodonset mania is a nonepisodic, chronic, rapid-cycling, mixed manic state. Such findings have also been reported by Janet Wozniak and colleagues and were replicated recently by Joseph Biederman and colleagues, who found that the overwhelming majority of 43 children from an outpatient psychopharmacology clinic who met diagnostic criteria for mania on structured diagnostic interview

presented with chronic and mixed presentation. Carlson and colleagues reported that early-onset mania was more likely to have comorbid behavior disorders in childhood and to have fewer episodes of remission in a 2-year period than did adult-onset cases of mania. Thus, pediatric bipolar disorder tends to present with an atypical picture characterized by predominantly irritable mood, mania mixed with symptoms of major depression, and chronic course, as opposed to euphoric, biphasic, and episodic course. Although the atypical features noted in children with manic-like symptoms raise the possibility of misdiagnosis, adolescents with bipolar disorder may provide a valuable touchstone for evaluating the validity of the diagnosis made in children, because bipolar disorder in adolescence has been more readily accepted. Stephen V. Faraone and colleagues compared the features of the mania in children to the diagnosis in adolescents with child-onset mania and in adolescents with adolescent-onset mania. With the exception of more euphoria among adolescents with adolescent-onset mania, the frequencies of other symptoms of mania were strikingly similar between child and adolescent youth with mania. As was the case for children, the clinical presentation of mania among adolescents with childhood onset was rarely biphasic: It was usually chronic and mixed with a simultaneous onset of depression and mania. These results suggested that, despite atypicality, the profile of manic features associated with childhood mania may, in fact, represent a clinically meaningful manic syndrome. COMORBIDITY WITH ADHD A leading source of diagnostic confusion in pediatric bipolar disorder is its symptomatic overlap with ADHD. Systematic studies of children and adolescents show that rates of ADHD range from 60 to 90 percent in pediatric patients with bipolar disorder. Although the rates of ADHD in samples of youth with bipolar disorder are universally high, the age at onset modifies the risk for comorbid ADHD. For example, although Wozniak and colleagues found that 90 percent of children with mania also had ADHD, which is similar to Scheffer's findings that 80 percent of patients had concurrent ADHD, Scott A. West and colleagues reported that only 57 percent of adolescents with mania were comorbid with ADHD. Examining further developmental aspects of pediatric bipolar disorder, Faraone and colleagues found that adolescents with childhood-onset bipolar disorder had the same rates of comorbid ADHD as children with bipolar disorder (90 percent) and that both of these groups had higher rates of ADHD than adolescents with adolescent-onset bipolar disorder (60 percent). Most recently, Gary S. Sachs and colleagues reported that, among adults with bipolar disorder, a history of comorbid ADHD was only evident in those subjects with onset of bipolar disorder before 19 years of age. The mean onset of bipolar disorder in those with a history of childhood ADHD was 12.1 years of age. Although ADHD has a much earlier onset than pediatric mania, the symptomatic and syndromic overlap between pediatric mania and ADHD raises a fundamental question: Do children presenting symptoms suggestive of mania and of ADHD have ADHD or mania, or both? One method to address these uncertainties has been to examine the transmission of comorbid disorders in families. If ADHD and mania are associated owing to shared familial etiological factors, then family studies should find mania in families of ADHD patients and ADHD in families of manic patients.

Studies that examined rates of ADHD (or hyperactivity) among the offspring of adults with bipolar disorder found higher rates of ADHD among these children compared to controls. Although the difference in rates attained statistical significance in only one study, a metaanalysis by Faraone and colleagues documented a statistical and bidirectional significant association between bipolar disorder in parents and ADHD in their offspring, as well as between ADHD in a child proband and mania in relatives. Wozniak and colleagues used familial risk analysis to examine the association between ADHD and bipolar disorder within families of children with bipolar disorder. They found that relatives of children with bipolar disorder were at high risk for ADHD that was indistinguishable from the risk in relatives of children with ADHD and no bipolar disorder. However, bipolar disorder and the comorbid condition of bipolar disorder plus ADHD selectively aggregated among relatives of bipolar disorder youth compared to those with ADHD and comparison children. Almost identical findings were obtained in two independently defined family studies of ADHD probands with and without comorbid bipolar disorder. Taken together, this pattern of transmission in families suggested that bipolar disorder in children might be a familially distinct subtype of bipolar disorder or ADHD. The existence of a familial, developmental subtype of bipolar disorder is consistent with the work of Michael Strober and colleagues and Richard D. Todd and colleagues, who proposed that pediatric bipolar disorder might be a distinct subtype of bipolar disorder with a high familial loading. One problem facing studies of ADHD and bipolar disorder is that these disorders share diagnostic criteria. For example, of seven revised fourth edition of the DSM (DSM-IV-TR) criteria for a manic episode, three are shared with the DSM-IV-TR criteria for ADHD: distractibility, motoric hyperactivity, and talkativeness. To avoid counting symptoms twice toward the diagnosis of ADHD and mania, two different techniques of correcting for overlapping diagnostic criteria have been used to evaluate the association between ADHD and pediatric mania. In the subtraction method, overlapping symptoms are simply not counted when making the diagnosis. In the proportion method, overlapping symptoms are not counted, but the diagnostic threshold is lowered to require the same proportion of symptoms from the reduced set as is required for the original diagnosis. Using these methods, Joseph Biederman and colleagues showed that 48 percent of children with bipolar disorder continued to meet criteria by the subtraction method, and 69 percent met criteria by the proportion method. Eighty-nine percent of children with bipolar disorder maintained a full diagnosis of ADHD by using the subtraction method, and 93 percent maintained the ADHD diagnosis by the proportion method. Taken together, these results suggest that the comorbidity between ADHD and pediatric bipolar disorder is not a methodological artifact due to diagnostic criteria shared by the two disorders. The potential for different rates of comorbidity with bipolar disorder in the combined subtype, inattentive subtype, and hyperactive-impulsive P.3276

subtype of ADHD is in need of further research. Faraone and colleagues studied 301 ADHD children and adolescents consecutively referred to a pediatric psychopharmacology clinic. Among these, 185 (61

percent) were the combined type, 89 (30 percent) were the inattentive type, and 27 (9 percent) were the hyperactive-impulsive type. Bipolar disorder was highest among combined-type youth (26.5 percent) but was also elevated among hyperactive-impulsive (14.3 percent) and inattentive (8.7 percent) youth. COMORBIDITY WITH CONDUCT DISORDER Like ADHD, conduct disorder is also strongly associated with pediatric bipolar disorder. This has been seen separately in studies of children with conduct disorder, ADHD, and bipolar disorder. Wozniak and colleagues reported that preadolescent children satisfying structured interview criteria for bipolar disorder often had comorbid conduct disorder. Maria Kovacs and colleagues reported a 69 percent rate of conduct disorder in a referred sample of bipolar disorder youth and found that the presence of comorbid conduct disorder heralded a more complicated course of bipolar disorder. Similar findings were reported by Stan P. Kutcher and colleagues, who found that 42 percent of hospitalized youths with mania had comorbid conduct disorder. These reports are consistent with the well-documented comorbidity between conduct disorder and major depression, considering that juvenile depression often presages mania. Biederman and colleagues investigated the overlap between mania and conduct disorder in a consecutive sample of referred youth and in a sample of ADHD subjects to clarify its prevalence and correlates. They found a striking similarity in the features of mania, regardless of comorbid conduct disorder. Additionally, the age at onset of mania was similar in subjects with or without comorbid conduct disorder. In both groups, mania presented with a predominantly irritable mood and a chronic course and was mixed with symptoms of major depression. Only two manic symptoms differed between these groups: Physical restlessness and poor judgment were more common in the mania with conduct disorder group compared to the mania only group. Similarly, there were few differences in the frequency of conduct disorder symptoms between conduct disorder youth with and without comorbid bipolar disorder. Although children with conduct disorder and bipolar disorder had a higher rate of vandalizing compared to conduct disorderonly subjects, this difference was not statistically significant. The comorbid and noncomorbid subjects with bipolar disorder had high rates of major depression, anxiety disorders, oppositional disorder, and psychosis compared to conduct disorder and ADHD children. In addition, bipolar disorder comorbid with conduct disorder was associated with poorer functioning and an increased risk for psychiatric hospitalization. Subjects with conduct disorder and bipolar disorder also had a higher familial and personal risk for mood disorders than other conduct disorder subjects who had a higher personal risk for antisocial personality disorder. Wozniak and colleagues compared relatives of four groups of clinically referred probands defined as having (1) conduct disorder and bipolar disorder (N = 26 probands, 92 relatives); (2) bipolar disorder without conduct disorder (N = 19 probands, 53 relatives); (3) conduct disorder without bipolar disorder (N = 16 probands, 58 relatives); and (4) controls without bipolar disorder or conduct disorder (N = 102 probands, 338 relatives). The risk for bipolar disorder in relatives of nonbipolar disorder, nonconduct disorder probands was 4 percent, and their risk for antisocial disorders was 7 percent. In reference to

the control relatives, relatives of both bipolar disorder groups had increased risks for bipolar disorder (14 percent in each group); relatives of both groups of conduct disorder probands had increased risks for antisocial disorder (34 percent and 19 percent). However, the risk of antisocial disorders was not elevated in relatives of bipolar disorderonly probands (8 percent), and the risk for bipolar disorder was not increased in the relatives of conduct disorderonly probands (9 percent). Taken together, these studies suggest that subjects who receive diagnoses of conduct disorder and bipolar disorder may, in fact, have both disorders. Although the resolution of this important issue awaits further research, the mere diagnosis of bipolar disorder in some conduct disorder children offers important therapeutic possibilities, because delinquency and mania require different treatment strategies. COMORBIDITY WITH ANXIETY DISORDERS Although anxiety is frequently overlooked in studies of bipolar disorder, pediatric studies of youth with panic disorder and youth with bipolar disorder document an important and bidirectional overlap between anxiety and bipolar disorder. Wozniak and colleagues reported significantly more panic and other anxiety disorders in children with bipolar disorder. L. Trevor Young and colleagues reported that patients with bipolar disorder and high anxiety scores were more likely to have suicidal behavior, alcohol abuse, cyclothymia, anxiety disorders, and a trend toward lithium nonresponsiveness. Yuan-Who Chen and Steven Dilsaver examined the relationship between panic and bipolar disorders using the Epidemiological Catchment Area database. They reported that the lifetime prevalence of panic disorder was 21 percent among subjects with bipolar disorder compared to only 0.8 percent among subjects who had neither bipolar nor unipolar depression. CORRESPONDENCE BETWEEN CATEGORICAL AND DIMENSIONAL MEASURES Many comorbidity findings in studies of juvenile bipolar disorder have relied on structured diagnostic methodology delivered by trained raters. Although this method is considered state of the art for clinical research, it could be vulnerable to an assessor bias. Thus, it is useful to consider studies using methods that are independent of assessor training or expertise. One such measure is the Child Behavior Checklist (CBCL), one of the best studied psychometric measures of psychopathology in children. The CBCL records, in standardized format, the behavioral problems and competencies of children 4 to 18 years of age, as reported by their parents or parent-surrogates. It is scored on the recently revised social competence and behavior problem scales of the Child Behavior Profile. The scales were originally constructed from analyses of parent ratings of 2,300 clinically referred children and were normed on 1,300 nonreferred children. One study has used the CBCL to characterize children and adolescents with categorical clinical diagnoses of pediatric mania. The results from this report have been strikingly consistent with previous reports, as it found that children satisfying diagnostic criteria for mania have highly abnormal profiles, which is consistent with structured diagnostic interview findings. This study also found significantly elevated tscores on the Aggressive Behavior subscale, the Delinquent Behavior Scale, and the Thought Problems Scale. COMORBIDITY WITH SUBSTANCE USE DISORDERS

An emerging literature suggests an extensive and bidirectional overlap between pediatric bipolar disorder and substance use disorders in P.3277

youth. This literature also suggests that juvenile-onset bipolar disorder may be a risk factor for substance use disorders. Timothy Wilens and colleagues found that bipolar disorder significantly increased the risk for substance use disorders independently of conduct disorder. Furthermore, they reported that the risk for substance use disorders was carried by those subjects with an adolescentonset form of bipolar disorder. Although this may be consistent with the notion that, like adults, adolescents self-medicate bipolar disorder symptoms with substances of abuse, it is also consistent with the hypothesis that child-onset and adolescent-onset bipolar disorder are etiologically distinct forms of the disorder with different risk profiles and natural courses. Biederman and colleagues recently used the family study design to address the putative familial risk of substance use disorder associated with bipolar and conduct disorder. After accounting for conduct disorder in probands, bipolar disorder in probands remained a risk factor for substance use disorders in relatives, including drug and alcohol addiction. Rather than concluding that the effects of conduct disorder or bipolar disorder accounted for the relationship of the other with substance use disorders, they reported results that were consistent with the idea that the effects of conduct disorder and bipolar disorder on the risk for substance use disorder were additive. Given this emerging literature on mood, conduct, and substance use disorders, more work is needed to characterize the developmental relationship between substance use and mood dysregulation in adolescents. It may be that the early identification and treatment of youth with bipolar disorder could prevent the onset or complications of substance use disorders. PEDIATRIC BIPOLAR DISORDER AND TRAUMA Although it has been long suspected that bipolar disorder in children may be the result of trauma and associations between trauma and bipolar disorder that have been reported in adults, there has been relatively limited systematic research of this issue. This report further suggested that behavior problems, including stealing, lying, truancy, vandalism, running away, fighting, misbehavior at school, early sexual experience, substance abuse, school expulsion or suspension, academic underachievement, and delinquency before 15 years of age predicted later posttraumatic stress disorder (PTSD). Not surprisingly, the authors concluded that this association may mean that persons with such behavior in childhood had a greater likelihood of experiencing trauma later on. Because juvenile bipolar disorder is commonly associated with extreme violence and severe behavioral dysregulation, as well as hypersexuality, bipolar disorder in children could be a reaction to or a risk factor for trauma exposure. Using data from a longitudinal sample of boys with and without ADHD, Wozniak and colleagues identified pediatric bipolar disorder as an important antecedent for, rather than a consequence of, traumatic life events. This temporal relationship between bipolar disorder and

traumatic events could have important clinical and therapeutic implications. When traumatized children present with severe irritability and mood lability, there may be a tendency by clinicians to attribute these symptoms to having experienced a trauma. To the contrary, longitudinal research suggests the opposite: Bipolar disorder may be an antecedent risk factor for later trauma and may not represent a reaction to the trauma. TREATMENT RESPONSE In a series of controlled clinical trials, Magda Campbell and colleagues documented the efficacy of mood stabilizers (lithium carbonate and carbamazepine [Tegretol]) in the treatment of aggressive conduct disorder children. However, these psychiatrically hospitalized conduct disorder youth were treated for severe, uncontrollable, and disorganized aggression and not necessarily for delinquency. Thus, it is possible that the therapeutic benefits observed in these children with antimanic treatments could have been due to their antimanic effects in treating aggressive manic children satisfying criteria for conduct disorder. Biederman and colleagues systematically reviewed the clinical records of all pediatrically referred patients who, at initial intake, satisfied diagnostic criteria for mania based on a structured diagnostic interview with the mother. Mood stabilizers were frequently used in these children, and their use was associated with significant improvement of manic-like symptoms that their psychiatrists had recorded in the medical record. In contrast, antidepressants, typical antipsychotics, and stimulants were not associated with improvement of manic-like symptoms. For lithium carbonate and for carbamazepine, higher and more therapeutic doses predicted greater decreases in the manic-like symptoms recorded by the treating clinician in the medical record. Although treatment with mood stabilizers was associated with a statistically significant decrease in manic-like symptoms, this improvement was slow to develop and was associated with frequent relapses. Although somewhat discouraging, these findings are consistent with outcome data from a naturalistic follow-up study of bipolar children and adults. For example, the survival analysis from Biederman and colleagues indicated that 65 percent of the children would improve if treated with lithium carbonate for 2 years. This finding is remarkably consistent with results from DeLong and Ann L. Aldershof, who reported a 66 percent response rate for manic children treated with lithium carbonate over a 10- to 70month treatment period, and with findings reported by Strober and colleagues showing that multiple relapses were most often seen in subjects with mixed mania. More optimistic findings have resulted from investigations of atypical neuroleptics in the treatment of juveniles with bipolar disorder. In a retrospective chart review study of 28 youths with bipolar disorder, 82 percent of subjects showed improvement in manic and aggressive symptoms with risperidone (Risperdal) treatment. In contrast to the duration of treatment required for improvement with mood stabilizers, the average time to optimal response was 1.9 months plus or minus 1.0 months of therapy. Moreover, no serious adverse effects were observed. Mani Pavuluri and colleagues reported improvement in children with bipolar disorder using risperidone in combination with either divalproex sodium (Depakene, Depakote) or lithium. Similarly encouraging results were reported by Jean A. Frazier

and colleagues in an open trial of olanzapine (Zyprexa) monotherapy. They found that treatment with olanzapine was associated with significant improvements in both the Children's Depression Inventory and the Young Mania Rating Scale in 23 manic children after 8 weeks of monotherapy on doses ranging from 2.5 to 20 mg per day. Similarly, Robert L. Findling and colleagues recently reported that risperidone was effective in treating aggression in children with conduct disorder. Although affective disorders were reported to have been excluded, it is unclear if this refers to the rare classic episodes of mania or the atypical cases of pediatric bipolar disorder that are more commonly comorbid with conduct disorder. Thus, the randomized clinical trial by Findling and colleagues may provide replication of the open trial by Frazier and colleagues rather than demonstrating an effect on conduct disorder per se. These initial encouraging results support the need for additional short-term and long-term controlled P.3278

trials of atypical neuroleptics in the treatment of juvenile bipolar disorder, as monotherapy or in combination with mood stabilizers. Although mood stabilizers improved manic symptomatology, they had no demonstrable effect on the symptoms of bipolar depression. Biederman and colleagues reviewed the medical charts of 59 patients with diagnosis of DSM-III-R bipolar disorder from an outpatient pediatric psychopharmacology clinic. Multivariate methods were used to model the probability of improvement and relapse at each visit of clinical follow-up. Serotonin-specific antidepressants were significantly associated with an increased rate of improvement of bipolar depression (relative risk = 6.7 [1.9 to 23.6]; P = .003) and a significantly greater probability of relapse of manic symptomatology (relative risk = 3.0 [1.2 to 7.8]; P = .02). Despite the increased risk of mood destabilization, serotonin-specific antidepressants did not interfere with the antimanic effects of mood stabilizers. Because bipolar youth commonly come to clinical practice with depression, these results underscore the importance of assessing a lifetime history of bipolar disorder in making treatment decisions in depressed youth. David Miklowitz completed a pilot study in 20 adolescents with bipolar disorder, using an adjunctive family-focused psychoeducational treatment (FFT-A). Miklowitz modified the adult version of FFT, which has been shown to be effective in forestalling relapses in two randomized clinical trials involving bipolar adults, and applied it to adolescents. Results from the pilot study in 20 bipolar adolescents (11 boys, 9 girls; mean age, 14.8 1.6) show that the combination of FFT-A and mood-stabilizing medications was associated with improvements in depression symptoms, mania symptoms, and behavior problems over 1 year. Another important therapeutic dilemma in the management of bipolar youth is the approach to comorbid ADHD. To evaluate pharmacological approaches for ADHD in children with bipolar disorder and comorbid ADHD, the medical charts of 38 patients with diagnoses of ADHD and bipolar disorder were reviewed over multiple visits to assess improvement and prescription patterns. The proportion of visits at which ADHD symptoms were rated as improved after initial improvement in manic symptoms was 7.5 times greater than before initial improvement of manic symptoms. The recurrence of manic

symptoms after their initial stabilization significantly inhibited ADHD response to medication. These results support the hypothesis that mood stabilization is a prerequisite for the successful pharmacological treatment of ADHD in children with ADHD and manic symptoms. FUTURE DIRECTIONS The explosive developments in neurosciences, neurobiology, genetics, and neuroimaging will undoubtedly help advance the understanding of this complex and crippling disorder, particularly its relationship to ADHD, conduct disorder, and other psychotic and nonpsychotic neuropsychiatric disorders. It is hoped that such advances can shed light on the etiology and underlying pathophysiology, including the identification of dysfunctional brain circuits that may underlie pediatric mania. For example, an emerging literature on the subject has identified genetic markers associated with bipolar features in children with velocardiofacial syndrome. More imaging research is needed to document the neuroanatomical underpinnings associated with pediatric bipolar disorder. These scientific approaches can also be used in the identification of endophenotypes in unaffected relatives of youth with bipolar disorder. The symptomatic overlap and cooccurrence of bipolar disorder with ADHD have produced debate as to whether these children have ADHD or bipolar disorder, or both. Despite this debate, many clinicians recognize that a substantial minority of children experience an extraordinarily severe form of psychopathology associated with extreme irritability, violence, and incapacitation that is highly suggestive of bipolar disorder. Clarifying the diagnoses of these ill children would have substantial clinical implications. The emerging literature indicates that bipolar disorder can be identified in a substantial number of referred children by using systematic assessment methodology. Thus, this disorder may not be as rare as previously considered. Children with bipolar disorder frequently demonstrate an atypical picture by adult standards, with a chronic course, severely irritable mood, and a mixed picture, with depressive and manic symptoms cooccurring. Most children with childhood-onset bipolar disorder may also have ADHD, which requires additional treatment. Initial clinical evidence suggests that atypical neuroleptics may play a unique therapeutic role in the management of such youth. The high levels of comorbidity with other disorders are common, further requiring the cautious use of a combined pharmacotherapy approach. More research is needed to build a scientific foundation for the notion that pediatric mania is a unique developmental subtype of bipolar disorder. SUGGESTED CROSS-REFERENCES The reader is encouraged to refer to the related mood disorder sections: Section 13.1 on introduction and overview, Section 13.2 on epidemiology, Section 13.3 on genetics, Section 13.4 on neurobiology, Section 13.6 on clinical features, and Section 13.8 on treatment of bipolar disorders. Other related materials include Section 31.8a on carbamazepine and Section 45.1 on depressive disorders and suicide in children and adolescents.