Mania and ADHD: comorbidity or confusion

Abstract

The frequency of occurrence of prepubertal mania is contingent on how much adherence to episodic disorder with separate periods of mania and depression is required. While manic symptoms superimposed on other psychiatric disorders is not uncommon, non-comorbid bipolar disorder is rare. A number of developmental, phenomenological and assessment considerations may complicate simple extrapolation of adult criteria onto young children. Nevertheless, it is clear that a significant number of preadolescents found in outpatient and inpatient samples meet at least symptom criteria for bipolar disorder. Such children have significant comorbidity and impairment. It is likely that some may develop classical bipolar disorder, some will continue to have substantial affective and behavioral comorbidity as do some complicated bipolar adults, and some will continue to have affective lability superimposed on their other, primary psychiatric disorders. Further research and follow-up will be necessary to determine who develops which outcome.

Introduction

It should be apparent from reading the papers in this volume, that in spite of the fact that juvenile mania has been recognized for many years, there is no consensus on the frequency with which it occurs in youth in general, and preadolescents in particular. The divergence of opinion, and inconsistency of data, has to with samples studied (i.e., clinical samples, samples acquired for the purposes of research, samples acquired for treatment studies) and the definition of bipolar disorder. Although the question of juvenile bipolar disorder may be more complicated, it has evolved at the moment into a question about the significance of co-occurring attention deficit disorder and manic symptoms in children and adolescents.

Assuming that the readers of this journal are more versed in the phenomenology of mood disorder than in child psychopathology, it is perhaps worth a quick review to highlight why the study of juvenile mania might be different from the study of bipolar disorder in adults. First, most disorders that primarily begin in adulthood presuppose that there was a significant period of time prior to the onset of the disorder when the subject did not have the disorder. In the case of acute disorders, the demarcation is clearest. In chronic disorders, where onset of symptoms of one disorder evolves and blends into the next disorder, diagnosis becomes more complicated. Thus, deciding whether someone has acute schizophrenia is much easier than deciding whether they have a schizotypal personality disorder that may be gradually evolving into schizophrenia. Defining mania in a 35-year-old pillar of the church is easier than it is in a 23-year-old man who has been abusing substances since the age of 13. One almost never sees juvenile bipolar disorder without an admixture of other serious psychopathology which is usually comorbid, always includes symptoms of attention deficit and hyperactivity, and which always antedates the mania. Thus, attention deficit hyperactivity disorder (ADHD) is most often the point of controversy though, as will be explained shortly, it is by no means the only one. Deciding whether a young person is manifesting symptoms of a new disorder (mania/bipolar disorder), or the exacerbation of another disorder is not as easy as it might appear.

A second complication is deciding how to ascertain the symptoms and from whom. In adult psychiatry, structured interviews and rating scales evolved from a large body of clinical lore that generally described the condition. There was a consensus about what the `template' looked like, and an attempt to operationalize the template to make symptom ascertainment reliable. Developmental age was not a consideration and, for the most part, the expectation was that information was obtained from the subject him/herself.

The template of manic-depressive illness and bipolar disorder has been gradually evolving over the past 30 years. As schizophrenia has gradually retreated from being an acute psychosis with very disorganized positive symptoms to a disorder characterized by chronicity and negative symptoms, manic depressive illness has expanded from being a disorder in which discrete episodes of mania, depression and euthymia with mood-congruent psychotic symptoms prevail, to a disorder where mixed states of affective dysregulation, mood-incongruent psychotic symptoms, and considerable comorbidity occur. Not surprisingly, the less conservatively one defines bipolar disorder, the more frequently one encounters it. The more conservative form of manic depressive illness is almost never seen in prepubertal children.

The assessment instruments and operationalized decisions which are required for systematic research have far-reaching consequences not always intended or anticipated by the researcher. Thus, while great care has been taken in many of the studies reported here to use structured interviews and rating scales, most structured interviews were never designed with children in mind and have been `back adapted' for use. The complications that arise from this are two-fold. First, if one rigidly adheres to adult criteria for certain symptoms, e.g., delusions of grandeur (not just inflated self esteem and bragging) or euphoria (not just silliness), one rarely sees them in children. If one adapts the criteria to be `child-sensitive', it is not entirely clear what is being included. Sexually disinhibited behavior is another good example. There are many reasons why children masturbate, use sexually explicit language, poke the `private parts' of peers, or look at photos of naked people. Sexual abuse studies, for instance, report that such behaviors are signs of sexual abuse, especially early sexual abuse (e.g., McClellan et al., 1996). In the Science section of the New York Times (Gillbert, 1998), a study done by William Friedrich, Ph.D., reported that he found 40% of presumably normal boys and 21% of girls `touch their private parts when at home' and 20% of both try to `look at people when they are nude or undressing'. Thus, the boundaries of `hyper' sexuality in children have yet to be established. In fact, developmentally sensitive epidemiological data do not exist to establish the credibility of the many of the interview/rating scale adaptations made so that clinicians need to be mindful of not simply checking off criteria on an interview without understanding the context of the symptom.

The second complication is what is not asked and what is not counted. The schedule for affective disorders and schizophrenia—childhood version (K-SADS, Orvaschel et al., 1982, Chambers et al., 1985), was based on the adult schedule for affective disorders and schizophrenia (Spitzer and Endicott, 1978), DSM-III (APA, 1980) criteria for children (there were no research diagnostic criteria for children), and the decision that the onset of a disorder did not occur until full criteria were met. Those using it, could not/did not carefully examine their population of prepubertal depressed children for oppositional defiance or incipient conduct disorder. If a child did not meet full criteria for conduct disorder, it was not `counted'. That means that a disorder like depression might appear to be primary because it takes a child more years to accumulate enough sins to meet criteria for full-fledged conduct disorder (Puig-Antich, 1992).

Harrington et al. (1991)have reported that children with conduct disorder and depression more closely resemble children with conduct disorder alone in terms of outcome have similarly reported that children with comorbid depression and conduct disorder have felony rates on adult follow-up equal to those young people with conduct disorder alone. Caspi and Moffitt (1995)have reviewed longitudinal studies of aggression and note that mothers who describe their 3-year-olds as `difficult to manage' and are rated by observers as `restless, impulsive and emotionally labile' distinguish antisocial adolescents from those with no disorders. These findings suggest that over the long haul, behavior disorder problems in childhood are more predictive and `primary' than concurrent depression or mood disorder. In fact, perhaps the treatment resistance that has occurred to psychopharmacological interventions in children is not due to incipient bipolarity (see Geller et al., 1998) but, rather, that depressed prepubertal children, in spite of meeting criteria for MDD, have an externalizing disorder as their primary problem. Their follow-up and treatment response is quite compatible with that outcome. We cannot test the hypothesis because certain information was never obtained on structured interview, or if it was obtained, it was never scored or counted. For instance, given what we know about rates of oppositional defiant/conduct disorder in preadolescent depression (Angold and Costello, 1993), it is difficult to believe that a sample of preadolescent depressed children with bipolar parents does not contain children with co-morbid oppositional defiant/conduct disorder. If only the most severe symptoms are counted, one misses one of the most important predictors of future problems, aggression. The point is that the interviews being used to assess mania in children need to examine subsyndromal comorbidity. Criteria for many disorders are evolving, as they have evolved from DSM-III. Moreover, inpatient units, multiple assessment points, home and school observations, which used to be the backbone of child assessment are often still necessary to understand complex psychopathology in children and families (see case report #2). In other words, a structured interview, even if it takes a day and is done by careful interviewers, may not replace longitudinal assessment with multiple informants. This viewpoint has been similarly articulated by Kutcher et al. in this volume.

The Mania Rating Scale (Young et al., 1978) presents similar complications in children. It was designed for quantifying (not diagnosing) an episode of mania. There was no consideration in its design to come up with ways of discriminating either ADHD or children with temper tantrums/intermittent explosive disorder or language disorders. If one uses it to track manic symptoms in a young person with comorbid ADHD, it will be very difficult to know if what appears to be a partial response represents only modestly treated mania, or residual ADHD—a point that Kafantaris et al. (1998)make in their discussion.

Finally, although full-blown mania is a pretty obvious disorder, and it is very unlikely that one would not get reliable symptoms from an external observer (like a parent), there are a number of other disorders for which examination of the child is very important. Every time I prematurely conclude from the parent history that I know what is wrong with the child (which means that parents have given me apparently lucid histories of well-defined entities), I have regretted it. The most recent experience, and pertinent to this discussion, was a seemingly bright 7-year-old about to be included in a study of valproate for acute mania. A history of fluctuating `manic' symptoms (every 3 weeks or so) was given along with the history of ADHD and oppositional defiant symptoms. However, the child appeared quite aloof on interview. It was explained to me that `he's in his quiet phase'. When a school visit and classroom observation (whole day) was done by our observer, the child appeared autistic-like, and in fact had had no particular cycles in behavior at school—well documented by the teacher who kept daily behavior logs. A more scrupulous assessment of parent history with something besides the K-SADS used for this project revealed a number of telling developmental symptoms. The child's hyperactivity and mood lability have since responded to stimulants (prescribed by someone other than this author), but he clearly continues to have very odd relationships with peers and considerable difficulty with changes in routine. His mood has been stable for 4 months so far. Family history would have been no help since between both parents' first-degree relatives, there were histories of depression, schizophrenia, mental retardation, drug abuse, paranoia and, among second-degree relatives (mother's nephew), autism.

Most of the studies presented in this volume that have tackled the question of mania/ADHD comorbidity in one way or another can be divided into three groups. The first group includes studies of bipolar adolescents. The samples are composed of teens with uncomplicated Bipolar I disorder (a clear episode of depression and an episode of mania requiring hospitalization) as well as those who had preadolescent ADHD (Kafantaris et al., 1998, Kutcher et al., 1998, Strober et al., 1988). The second group of studies examines prepubertal ADHD and mania (Biederman et al., 1998, Geller et al., 1998). An interesting finding in these two studies is the absence of children with mania who do not have ADHD. (We have taken editorial license and included the study by Faraone et al. (1997)who is part of the Massachusetts General Hospital group studying juvenile bipolar disorder.) This, of course, precludes a comparison with non-comorbid bipolar disorder in children. Finally, two studies try to explain the chronicity and comorbidity seen in childhood mania by suggesting that manic symptoms/syndrome in and of itself explains some of the psychopathology being attributed to bipolar disorder.

Table 1 has drawn together information about the adolescent samples studied. Samples are mainly in-patient and are relatively similar with regard to their demographic and phenomenological similarities. Most of the hospitalized bipolar/manic-onset adolescents did not have comorbid ADHD. (I have not included the post pubertal manic teens in Geller's study since they had a childhood onset.) All three studies addressed lithium treatment, either directly or indirectly—the implication being that if there is a response to lithium then the disorder is similar to what is seen in adults. The take home message was that around half of the samples improved considerably with lithium. In Kafantaris's sample, the presence of childhood ADHD did not predict treatment response but the presence of psychosis boded poorly. In Strober's sample, most adolescents did relatively well; those with ADHD had a statistically poorer response, though as Kafantaris pointed out, it is not clear whether what made the response look worse was inadequately treated mania or residual ADHD symptoms. Although the thrust of the Canadian sample was to identify a population of teenagers with conservatively defined manic depressive illness, the description of medication use and the variety of treatments utilized by the sample suggests that, even in a population of relatively uncomplicated bipolar adolescents, different medication strategies are necessary, as is comprehensive treatment planning. Nevertheless, the percentage of teens apparently responding to lithium is in the ball park of what has been reported in recent studies of adults (e.g., Maj et al., 1998).

The Massachusetts General Hospital psychopharmacology clinic sample also included on this chart has been described in detail elsewhere (Faraone et al., 1997). It is somewhat different from the other samples primarily because it is out-patient. Rates of comorbidity with ADHD is considerably higher, and there are virtually no teens with uncomplicated bipolar disorder. This is the case with the post-pubertal manic youth in Geller's sample, also an out-patient sample. Treatment outcomes of these samples have yet to be published (though one expects that may happen in the not too distant future). Given the difference both in the population source and the assessment of bipolar disorder (cross-sectional versus observations for several days in hospital), reasons for what differences exist cannot be satisfactorily resolved.

The most contentious groups, the preadolescents with diagnoses of both ADHD and mania, appear different both from each other and from the adolescent samples. Table 2 outlines some of the areas which both studies address. Geller's sample differs from Biederman's, identifying very high rates of euphoria, and an average of three `cycles' per day in subjects who have been chronically ill for an average of 3 years. Biederman's sample also reports high rates of chronicity (between 60 and 80%), but very low rates of euphoria and, if one extrapolates from Faraone's data, low rates of rapid cycles. It is somewhat difficult to reconcile rates of chronicity in adolescent and adult samples of about 5% (Winokur et al., 1994, Strober et al., 1995), and the high rates of chronicity in prepubertal bipolar subjects. Similarly, given that rates of childhood externalizing disorders have been reported to be about 20% in adult samples (Winokur et al., 1993, Carlson, 1998, Carlson et al., 1998), either the numbers of acute, uncomplicated manic subjects occurring after age 20 must increase enormously, or some of the children with ADHD and mania develop other disorders. Even if the ADHD symptoms subside in adult manics, one would expect that high rates would be present in their childhood histories.

An analogy has been made (Geller and Luby, 1997, Akiskal, 1995b) that childhood mania in the 1990s is analogous to childhood depression in the 1970s. A continuation of that thinking would also aver that manic symptoms are not necessarily the same as bipolar disorder, just as depressive symptoms are not synonymous with major depressive disorder. That does not diminish the importance of the symptoms, but rather suggests that, like much psychopathology, manic symptoms exist as a dimension just as depressive symptoms do. This concept has been articulated by Akiskal (1995a), Akiskal (1995b)and also by Belmaker and van Praag (1980)a number of years ago in a book titled Mania—An Evolving Concept.

Rates of emotional lability, as defined by the presence of `moods go up and down quickly', or a week's worth of elated, expansive or irritable mood (i.e., the `A' symptom of Mania criteria) occurs in about 5–10% of the general population (see Carlson and Kelly, 1998, for review). In addition, people with emotional lability meet criteria for the same comorbidities as do those who meet full syndrome/diagnosis of mania. Conversely, adults who had had a history of severe temper tantrums in childhood, grow up to have personalities which certainly fall within an affective dimension. Attributes like `tends toward under-control of needs and impulses; tends to perceive many different contexts in sexual terms; sees what he can get away with; is self-dramatizing; is irritable; is guileful; has fluctuating moods; is a talkative individual; is self-indulgent; gregarious; initiates humor'. were significantly correlated with the frequency and severity of childhood tantrums (Caspi et al., 1990).

The adult psychiatry literature is replete with case reports of manic symptoms occurring after a stroke or as part of a metabolic disorder. When all criteria for mania are met, this is called secondary mania (see Carlson, 1990Evans et al., 1995, for recent review). This phenomenon is observed in children as well. Swedo et al. (1998)found emotional lability to be present in 66% of children with a post-streptococcal movement disorder (Syndenham's chorea), accompanied by emotional and obsessive–compulsive symptoms. They have coined the term pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections (PANDAS) for this disorder. Another example is the finding that emotional lability was a very common sequela in a population of children systematically studied after a moderate to serious closed head injury, Gerring et al. (1998). It was most serious in children who either developed ADHD after the head injury, or who had their ADHD symptoms exacerbated by the injury. In these cases, three characteristics quadrupled in severity: emotional lability, symptoms of ADHD and aggressive behavior. Some children in fact met symptom criteria for mania on structured interview (Gerring, personal communication, 10 October 1997).

Organic personality syndrome or disorder has been the time-honored designation for behaviors secondary to some sort of brain disease, occurring in clear consciousness. It has been characterized by `lability and shallowness of affect and impairment of judgment' (DSM-II, 1968), or, in greater detail, `affective instability, e.g., marked shifts from normal mood to depression, irritability, or anxiety; recurrent outbursts of aggression or rage out of proportion to stressors, markedly impaired social judgment, e.g., sexual indiscretions, marked apathy and indifference; suspiciousness or paranoid ideation' (DSM-III-R, 1987). Although this has been eliminated from DSM-IV because of its lack of specificity, and it is apparently necessary to be able to define the central nervous system lesion in order to give someone the diagnosis of `mood disorder secondary to' whatever, it is clear there are prenatal influences (poor nutrition, infection, etc.) that are associated in some way with the development of schizophrenia (e.g., Susser et al., in press) and likely affect mood. The ease with which mentally retarded children become behaviorally disinhibited with medications suggests vulnerability to emotional lability possibly because of a subtle `organic' disorder.

Emotional lability accompanies other serious psychopathology in children. Just as there has been a long history of disentangling schizophrenia with affective symptoms from psychotic bipolar disorder in adults, ironically, emotional lability or dysregulation appears to complicate disorders in children which, years ago, had been designated as having `childhood schizophrenia'. Rapoport and colleagues found that 30% of children referred to a study of childhood schizophrenia had `complex developmental disorders and brief psychotic symptoms that did not meet DSM-III-R criteria for schizophrenia' (McKenna et al., 1994, Kumra et al., 1998). Criteria for this syndrome, designated `multidimensionally impaired' (MDI) consist of `nearly daily periods of emotional lability disproportionate to precipitants; poor ability to distinguish fantasy from reality as evidenced by ideas of reference and brief perceptual disturbances during stressful periods or while falling asleep; impaired interpersonal skills despite desire to initiate social interactions with peers; cognitive deficits indicated by multiple deficits in information processing; absence of formal thought disorder'. As was seen in Gerring's description of brain-injured children, a high proportion of children with MDI meet criteria for ADHD (84%). The disorder is also characterized by a very early onset, considerable disability and, though follow-up is ongoing, considerable chronicity to date, not unlike that seen in a follow-up of `borderline' children (Lofgren et al., 1991).

Studies of children with autism and pervasive developmental disorders also found a population of children who did not meet the narrow criteria for infantile autism. The condition is called `multiple complex developmental disorder' (Towbin et al., 1993). Recognizing this condition as possibly encompassing children previously described as `schizophrenic', `borderline', `atypical' or `childhood onset pervasive developmental disorder', the authors admit that they do not know what the ultimate course of this group of children will be. There are three categories of symptoms impacting affect/mood, cognitive functioning, and social interaction. The major affective symptoms include `regulation of affective state and anxiety, impaired beyond that seen in children of comparable age (intense generalized anxiety, diffuse tension or irritability); unusual fears and phobias, recurrent panic episodes, episodes lasting from minutes to days of behavioral disorganization and regression; significant and wide emotional variability with or without environmental precipitants; high frequency of idiosyncratic anxiety reactions such as sustained periods of uncontrollable giggling, laughter, or `silly' affect that is inappropriate in the context of the situation'.

Symptoms sounding vaguely psychotic include impaired cognitive processing (thinking disorder) manifested by irrationality, sudden intrusions on normal thought process, magical thinking, neologisms or nonsense words repeated over and over, desultory thinking, blatantly illogical, bizarre ideas; confusion between reality and fantasy life; perplexity and easy confusability (trouble understanding ongoing social processes and keeping ones thoughts `straight'; delusions, including fantasies of personal omnipotence, paranoid preoccupations overengagement with fantasy figures, grandiose fantasies of special powers, and referential ideation.

This syndrome sounds similar to the MDI disorder described earlier, though there is closer continuity with the pervasive developmental disorders in the category of `consistent impairments in social behavior and sensitivity (e.g., social disinterest, detachment, avoidance or withdrawal in the face of evident competence (at times) of social engagement, particularly with adults. Attachments may appear friendly and cooperative but very superficial, based primarily on receiving material needs; inability to initiate or maintain peer relationships; disturbed attachment displaying high degrees of ambivalence to adults, particularly parents/caregivers as manifested by clinging, overly controlling, needy behavior and/or aggressive, oppositional behavior; profound limitations in capacity to empathize or understand others' affects accurately'.

Wozniak et al. (1997)have reported that 21% of children with pervasive developmental disorder present with clinical features of mania. Examination of their data reveals that the PDD features in this combined group are milder than the PDD symptoms in children without `mania'. This suggests they are perhaps identifying the same population of children.

The point in describing these observations is to emphasize that three completely independent groups, with completely different theoretical and research orientations, have recognized the existence of a group of children with symptoms of ADHD, and unusual behavior characterized by a significant mood component, psychotic symptoms of insufficient intensity or duration to meet schizophrenia criteria, but in whom significant comorbidity and disability exist. Co-morbidity that meets criteria is complicated enough; subsyndromal comorbidity needs to be dealt with as well.

Attention deficit/hyperactivity disorder has become the biggest focus in the controversy about whether hyperactive, inattentive, emotionally labile children are really manic or really `just ADHD'. Again, however, a look back into the history of `the hyperactive child' syndrome reveals that it was initially designated with a good deal of emotional lability, sleep disturbance, and variation in behavior. Although hyperactivity and short attention span were the first two symptoms detailed in Laufer and Denhoff's original description of the hyperkinetic behavior syndrome in children (Laufer and Denhoff, 1957), variability was the third. `Behavior is unpredictable, with wide fluctuations in performance'. Other symptoms included impulsiveness and inability to delay gratification, irritability (`fits of anger are easily provoked'), and explosiveness (`reactions of these children are often almost volcanic in their intensity'). Other relevant observations made by these authors were that in infancy `the baby may have been exceedingly irritable and cried readily', sleep problems are common with early waking and `rampaging through the house in hyperactive, noisy and sleep—disturbing play'. Not surprisingly, in DSM-I terms, before there was a `hyperkinetic reaction of childhood designation', the diagnosis of `passive–aggressive personality', `anxiety reaction' and `emotionally unstable character' were designations used to describe such children (Laufer et al., 1957). Our current comorbidities are oppositional–defiant disorder, various anxiety disorders, and bipolar disorder. In tracing the elimination of impulsivity, low frustration tolerance, variability, etc., from the criteria for ADHD, it would appear that it was the lack of reliability and lack of specificity, not the absence of association that phased these symptoms out of our clinical picture of ADHD (D.P. Cantwell, personal communication, January 1996). The law of unintended consequences may have resulted in a reconfiguring of ADHD, such that children initially described as `hyperkinetic' would be reconceptualized now as `bipolar', at least insofar as onset from birth, variability redefined as `episodes', and irritability/explosiveness/impulsiveness are concerned.

In conclusion, clinicians have been associating emotional lability with attention deficit hyperactivity type of behaviors, and psychotic like behaviors for many years, and labeling them in different ways. This suggests an intrinsic and enduring association between these behaviors. As each disorder has become more stringent with its criteria, a group of children remains that do not quite fit. Insofar as emotional lability is often a common and striking symptom, severe ADHD, schizophrenia spectrum disorders, pervasive developmental disorder spectrums, organic brain disorders, and childhood mania become the fall back diagnoses.

Episodes have always been the defining feature of bipolar disorder and not ADHD. Operationalizing episodes is what is, I feel, at the core of the controversy. How long does a `mood swing' need to last before it becomes an episode? Is that `mood swing' the exclusive defining feature of bipolar disorder? The recent DSMs have not resolved this question. Since there is no consensus, let alone data, on this point, we examined two clinical samples for manic symptoms (i.e., they did not necessarily have the disorder). The first sample comes from the Long Island Follow-Up Study (LIFUS) (Loney et al., submitted), 250 boys ages 6–10, referred from clinics, directly from schools and from a parent support group were evaluated using the Diagnostic Interview for Children and Adolescents, parent and child version (Herjanic and Reich, 1982). Nine percent of the sample was found to have a manic syndrome (Carlson et al., 1998). The second, inpatient sample, used the DSM-III-R version of the Child Symptom Inventory (Gadow and Sprafkin, 1994), which corresponds well to the K-SADS (Grayson and Carlson, 1991). We found 52% of parents endorsed five or more symptoms of mania (Carlson and Kelly, 1998). In both the out-patient and in-patient samples, the children had the same range of diagnoses on structured interview that has been reported with children diagnosed with bipolar disorder. That is, there were high rates of ADHD, oppositional–defiant disorder, conduct disorder, and mood disorder. Responses on the Child Behavior Checklist dimensions had the same pattern as those reported by others, though there was a difference in severity of some factors. Finally, children with manic symptoms were viewed by teachers and in-patient staff as behaviorally more `manic' than children without those symptoms, and their inpatient stays were even longer. The presence of manic symptoms did not preclude a response to stimulants but the response was far from robust.

Given the consistent finding of comorbid externalizing disorder, symptomatology that occurs with manic symptoms, it seemed relevant to ask whether the increased severity observed in children with manic symptoms/syndrome is a function of comorbidity (i.e., having more than one disorder is likely to increase the level of dysfunction), or whether there is a unique morbidity conferred by the manic symptoms themselves. In the LIFUS study, we found that most of the significant differences completely washed out once externalizing comorbidity was controlled for. Only the depression/anxiety factor continued to somewhat differentiate the children with and without manic symptoms—significantly in teachers' and marginally in parents' reports.

In conclusion, manic symptoms, whether they are part of a full-fledged disorder, or are present only as symptoms or in a syndrome state, seem to represent considerable psychopathology. Since explosive, unpredictable aggressive behavior is an important part of emotional lability/manic symptoms, the exacerbation of any disorder in which aggression already occurs (e.g., ADHD) is not surprising. We suspect that some of what is being called `bipolar disorder' with comorbidity, represents manic symptoms/emotional lability occurring as part of the psychopathology of the other disorder. In that regard, manic symptoms are like psychotic symptoms. They presage serious psychopathology, are more associated with certain disorders, but are not, in and of themselves, diagnostic of any particular disorder. In fact, if we have learned anything over the past 100 years in trying to distinguish schizophrenia from mania, it should be that cross-sectional symptoms alone will not do it. The same can be said of attentional symptoms. While inattention is certainly a prominent part of ADHD, its presence is not pathognomonic of ADHD. Nevertheless, its presence certainly means that considerable cognitive impairment will be likely. This conclusion, moreover, does not diminish genetic implications either, since it may well be that dimensions rather than disorders are what is heritable.

That said, there is certainly evidence that manic depressive illness/bipolar disorder can begin some years after an externalizing disorder has clearly begun. The combination is no more benign than either disorder alone and treatment may need to take both disorders into account. Several years ago, we published a small treatment study of children with mixed manic, depressive, and attention deficit disorder symptoms, who had positive family histories of mood disorder. Children were given lithium, methylphenidate and the combination in a randomized, double-blind design. Children's inattention and hyperactivity responded better to combined low dose methylphenidate and lithium than to either high-dose methylphenidate or lithium alone (Carlson et al., 1992).

Whether ADHD and mania represent the co-occurrence of two disorders or a subtype of one or the other awaits studies of comorbidity (including childhood psychopathology) of bipolar disorder in parents, empirically defining the nature, duration and significance of an `episode', the importance of nailing down a truly accurate characterization of other serious childhood psychiatric disorders, and development of assessment instruments unbiased by our currently inadequate knowledge that includes cross-sectional and longitudinal data.Carlson et al., in pressHarrington, in pressHowland and Thase, 1993, Robins et al., 1981