Externalizing Disorders of Childhood (Attention-deficit/Hyperactivity Disorder, Conduct Disorder, Oppositional-Defiant Disorder, Juvenile Bipolar Disorder) (February 14-16, 2007)

Prepared by Michael B. First, M.D., DSM Consultant to the American Psychiatric Institute for Research and Education (APIRE), a subsidiary of the American Psychiatric Association

The APA, in collaboration with the WHO and NIH, convened a diagnosis-related research planning conference focusing on externalizing disorders of childhood in Mexico City on February 14-16, 2007. The conference was the ninth in a series of 12 NIH-funded conferences on "The Future of Psychiatric Diagnosis: Refining the Research Agenda" that is administered by APA's American Psychiatric Institute for Research and Education (APIRE). The Externalizing Disorders conference co-chairs were David Shaffer, MD, FRCP, New York State Psychiatric Institute/Columbia University (New York, NY), Luis Augusto Rohde, MD, ScD, Hospital del Clínicas de Porto Alegre (Porto Alegre, Brazil), and Ellen Leibenluft, MD, National Institute of Mental Health (Bethesda, MD). Thirty invited scientists from around the world participated.

The conference began with three presentations presenting recent findings from longitudinal studies and their implications for nosology. Terrie Moffitt, PhD (London, UK) focused on findings relevant to the diagnoses of Conduct Disorder (CD), Oppositional Defiant Disorder (ODD) and Antisocial Personality Disorder (ASPD). She reviewed a number of issues pertaining to possible changes in CD. The age-of-onset subtypes (i.e., childhood, adolescent, unspecified) were examined using trajectory analyses in longitudinal cohorts and were validated (e.g.., individuals with childhood-onset type continue to meet criteria up to their late 20's). Furthermore, a family history of psychiatric disorders, especially externalizing disorders, discriminates among the three subtypes (e.g., the childhood-onset group has the strongest family history). Although the trait of being callous and unemotional was excluded from previous DSMs because of concerns about poor reliability of assessment, it has been shown that this trait can be assessed with good reliability and that it offers incremental validity in terms of better prediction of outcome. With respect to the diagnostic relevance of physiological biomarkers (i.e., stress hormones, neurotransmitters, perinatal complications, and heart rate), all four have a moderate to strong relationship with conduct disorder (with heart rate the strongest) but their predictive validity in terms of prognosis and treatment response is not yet well known. Using genotypes as biomarkers, Dr. Moffitt reviewed five candidate genes and whole genomic scans; the finding that is best replicated is the finding that boys with the low MAOA genotype who experience severe maltreatment are more likely to develop CD. Furthermore, manifestations of low MAOA genotype using functional and structural MRI in healthy subjects can be discerned in terms of reduced volume and altered function in the amygdala fear regulating circuit. Very few studies have examined neuroimaging biomarkers and CD. Although one study did show less amygdala activation to emotional stimuli in boys with callous unemotional traits, the small samples and lack of follow-up render such findings of dubious use in clinical practice. Considering the relationship between ODD, CD, and ASPD, many children meet criteria for ODD; only a small subset progress to CD and yet a smaller subset progress to ASP. Although making a diagnosis of CD in the 2- to 5-year-old range may have advantages in terms of helping families get access to affordable therapy, there are dangers of false-positive diagnoses with harmful labeling, and setting the child up to embody a self-fulfilling prophesy. Although it has been suggested that the CD threshold be lowered in girls from 3 to 2 symptoms, all of the reviewed literature suggest that dimensional models work the same for males and females so that any rationale for lowering the cut-off would apply equally well to boys. Similarly, although adding female-specific criteria such as relational aggression and precocious sexuality would increase prevalence rate in girls, whether this addition offers any incremental validity is unknown. Early substance use was one of the criteria for CD in DSM-III but was dropped because of very high base rates of experimentation. Given that early substance use has repeatedly been shown to be an integral part of CD and that it also has predictive validity, early substance use should be reconsidered as a criterion for DSM-V. In considering whether CD is best represented as a category or a continuum, there are hundreds of papers that support the use of dimensional measures for CD. Given the clinical need for a CD category, it would make sense to use both categorical and dimensional approaches in tandem. Dr. Moffitt concluded that most people are generally satisfied with the DSM-IV approach. Proposals for DSM-V to add subtypes and criteria should be evaluated in terms of whether they offer improved specificity, sensitivity, or incremental prediction, and whether can they be translated into clinical practice.

Jane Costello, PhD (Durham, NC) presented findings from longitudinal studies (e.g., the Smoky Mountain longitudinal study) relevant to the diagnosis of Attention-Deficit/Hyperactivity Disorder (ADHD) and focused on five questions. The first question was whether adolescents and adults, as well as children, have ADHD. She concluded that the prevalence of ADHD is similar across the age span from age 2 through childhood (around 2-5%) and that while adolescent and adult prevalence is lower (around 1-4%) it nonetheless is significant. Reviewing recent epidemiological studies to answer the second question (i.e., Does childhood ADHD continue into adulthood?), Dr. Costello noted that about one-third of individuals with childhood ADHD go on to have adult ADHD and that persistence was more probable in girls than in boys. The third question concerned the age of onset requirement (i.e., "some hyperactive-impulsive or inattentive symptoms that caused impairment were present before age 7 years"). Epidemiological data suggests that this criterion makes no difference with respect to the actual age of onset in a sample of individuals with ADHD; age of onset is around 5.3 years regardless of whether or not this criterion is required. The fourth question addressed whether comorbidity between ADHD and ODD is simply the result of overlapping items in their criteria sets. Dr. Costello noted that although that the comorbidity between every ADHD and ODD item is higher than you would expect by chance, the absence of a pattern suggests that they are tapping different constructs. The final question is whether longitudinal data provide support for having a combined ADHD/ODD diagnosis in DSM-V. She concluded that although the same children are likely to have both sets of symptoms, the onset of ODD is 4 years later than any of the sub types of ADHD and that comorbid ADHD/ODD cases are only a small proportion of all cases of ODD, and only half of all cases of ADHD. Furthermore, although childhood ADHD predicts adolescent ODD in both sexes, childhood ODD does not significantly predict adolescent ADHD. Therefore there is no evidence to support a combined ADHD/ODD diagnosis.

Boris Birmaher, MD (Pittsburgh, PA) examined results from the COBY study (Course and Outcome of Bipolar Youth) to help clarify the phenomenology and course of pediatric Bipolar Disorder (BP). In this study, which included children and adolescents with DSM-IV bipolar I (58%), bipolar II (7%) and bipolar NOS (35%), the participants are followed every month for 10 years. The vast majority of the bipolar NOS cases (86%) met symptom severity criteria for either bipolar I or bipolar II but the episodes fell short of the duration requirement. Overall, roughly 66% recovered in 2 years and 48% had at least one recurrence. Predictors of poorer outcome at 2 years included early-onset, psychosis, being diagnosed with bipolar NOS, having had mixed episodes, lower socio-economic status and the presence of comorbid anxiety and ADHD. Compared with a 2002 sample of adults with bipolar I, children with bipolar I disorder are more likely to have mixed episodes, rapid cycling, fewer periods of wellness, and more fluctuations in polarity. Based on these findings, Dr. Birmaher recommended that BD criteria continue to be applied to children but that elation, grandiosity, irritability, psychosis, etc. be defined with developmentally appropriate examples. Furthermore, he suggested several additional potential changes, including possibly lowering the number of days required for a manic or hypomanic episode, requiring periodicity of mood, requiring that symptoms be present in more than one setting, and requiring that those symptoms overlapping with other disorders (e.g., hyperactivity) need to worsen in order to be counted towards the diagnosis of BP.

The next three presentations focused on the impact of cultural and environmental influences on the nosology of externalizing disorders. Alan Flisher, MD, PhD (Capetown, South Africa) began his presentation by reviewing the various ways that DSM-IV deals with culture, which include having entries in the descriptive text for disorders describing cultural variations in their clinical presentation, an appendix containing an outline of cultural formulation, and a description of culture-bound syndromes (i.e., recurrent locally-specific pattern of aberrant behavior). According to the DSM-IV text, ADHD is noted to occur in various cultures, and variations in reported prevalence among countries probably arise more from different diagnostic practices than from differences in clinical presentation; the text for CD cautions that the "diagnosis should be applied where the disorder … is symptomatic of an underlying dysfunction … and not … a reaction to the immediate social context." DSM-IV is silent regarding cultural variations in ODD presentations. Although it has been proposed that ADHD is a culture-bound disorder, studies comparing rates of ADHD in the US vs. rates in the rest of the world have shown that prevalence rates and correlates are comparable across settings, suggesting that it is not simply an American disorder. A study of disruptive behaviour in Limpopo, South Africa also supported the cross-cultural validity of ADHD, noting that the small differences in prevalence rates are likely due to methodological factors and lack of equivalence in terms of scales, norms, and language. Dr. Flisher recommended that the existing features of DSM-IV continue to be used while keeping normative equivalence issues in mind. Regarding DSM-V, he recommended increased prominence of cultural issues in the introduction; including more material on cultural issues in the text; and developing both a separate cultural axis and separate chapter on culture-bound syndromes. Urgent research needs include studies of adult ADHD in non-Western countries, cross-cultural studies of CD and ODD, studies of "emic" aspects (e.g., the role of cultural explanations, environment, help-seeking, and course), and intervention studies.

Enrico Mezzacappa, MD (Boston, MA) in his presentation examined ethnic-racial differences in response thresholds for externalizing behavior problems in a diverse sample of over 1,800 9 to 15 year-old children from the Project on Human Development in Chicago Neighborhoods in order to assess measurement equivalence for aggression and delinquency as defined by the Child Behavior Checklist (CBCL). Given the socially contextualized nature of CD, it is crucial that assessment tools take into account variations in the appraisal of symptoms arising across social class, race, ethnicity, or immigration/acculturation status that are due to differences in the way these behaviors are understood and reported across comparison groups. By applying item response theory to parent ratings of aggression and delinquency from the CBCL, differences in rates of endorsement due to non-equivalencies in response thresholds could be determined across the three different ethnic-racial groups studied: namely Whites, Blacks, and Hispanics. Overall, item endorsement probabilities were consistent with expected severities across all groups. However, non-equivalencies in item response thresholds were identified for both scales, and across all group comparisons, indicating the absence of measurement equivalence for CBCL-defined aggression and delinquency across ethnic-racial groups in this sample. Furthermore the pattern of non-equivalencies suggested that cultural factors tied to ethnicity may have been more influential than those related to race, although both clearly contributed to the differences in response thresholds for both sets of externalizing behavior problems.

Dr. Francisco de la Peña Olvera (Mexico City, Mexico) began his presentation by noting that there is evidence that culture defines and generates stress, models the form and quality of illness experience, influences the expression of symptomatology, determines symptom interpretation, guides the search for help, and rules the social response to illness and impairment. Although there is conceptual equivalence of externalizing disorders among cultures (e.g., there are no significant differences in the prevalence of ADHD in developing countries as compared to developed countries), manifestations of externalizing disorders are modulated by cultural factors (e.g., according to teacher reports, boys are rated as having higher externalizing scores than girls in all cultures, except for Iran and Jamaica). Even though there is great heterogeneity of the Externalizing Disorders across cultures, ADHD is consistently integrated with two factors (i.e., inattention and hyperactivity/impulsivity), ODD can be recognized as both an independent factor and also commonly comorbid with ADHD, and two main factors hold up for CD: non aggressive (non predatory) and aggressive (predatory). Methodological limitations in doing cross-cultural research include limited data on how a cultural threshold is recognized and established and a lack of culturally validated instruments. Regarding the task of modifying criteria in future classifications to be more culturally accurate, ethnicity, environmental factors, and the cultural background of both the interviewee and the interviewer need to be considered.

The next three presentations considered how current neurobiological knowledge might influence the nosology of externalizing disorders. Joseph Sergeant, PhD (Amsterdam, Netherlands), in his presentation, examined the evidence regarding whether abnormalities in executive function (EF) are either necessary or sufficient for a diagnosis of ADHD and therefore worthy of inclusion in the criteria. Tasks measuring executive function include inhibition, planning, vigilance, and working memory. Dr. Sergeant reported that although meta-analyses of studies have demonstrated that measures of EF are able to discriminate children and adults with ADHD from controls, non-executive function measures (e.g., the Trails A task) discriminate adults with ADHD from controls as well. Tests that measure response time variability, independent of the type of task, can also discriminate ADHD from controls. EF tests are less able to discriminate among the ADHD subtypes, however. They fail to discriminate between individuals with ADHD and those with other disorders like CD, ODD, high functioning autism, and those with reading disabilities. Since a relatively high proportion of non-affected siblings of ADHD children have EF problems, EF tests cannot currently be used for clinical diagnostic purposes. Future research endeavors recommended by Dr. Sergeant include greater use of non-EF measures to determine measurement specificity, use of IQ measures to determine differences in fluid and structural intelligence, and studying children with a wide age range to assess developmental effects in EF research. Studies are urgently needed with multiple group comparisons to establish clinical specificity, and large-scale family studies to determine both the false positives and false negative rates of EF measures in assigning children to an ADHD classification.

Edmund Sonuga-Barke, PhD (Southampton, UK), in his presentation, considered what role should markers of motivational dysfunction play in the diagnosis of ADHD. Renewed interest in motivational dysfunction as a cause of ADHD is a reaction to a growing recognition that executive dysfunction is not a necessary or sufficient component of ADHD. Dr. Sonuga-Barke postulated that given the heterogeneity suggested by the moderate strength of association with executive dysfunction, ADHD is likely not a single pathophysiologic entity but rather a heterogeneous condition consisting of multiple more or less discrete groupings of patients each with distinctive elements to its neuropsychological profile. He proposed that sensitivity to reward-related delay (i.e., preference for smaller immediate rewards over large delayed rewards) may be the most robust indicator of motivational dysfunction in terms of differentiating those with ADHD from controls. Like executive dysfunction, delay aversion is moderately associated with ADHD but is neither a necessary nor a sufficient condition with only a substantial minority of cases affected. Therefore, like executive dysfunction measures, delay aversion markers add little value to current approaches to the diagnosis of ADHD as currently conceptualized. However, delay aversion and executive function markers may be crucial in future refinements of ADHD definitions in as much as they help partition heterogeneity (e.g., executive dysfunction and delay aversion could be seen as separate pathways to ADHD grounded in functionally segregated brain circuits). Dr. Sonuga-Barke called for research into whether subgroups of children within the ADHD umbrella can be identified with a predominantly "motivational type" or "cognitive type" and whether this categorization helps predict long-term outcome and treatment response.

Ian H. Gotlib, PhD (Stanford, CA) discussed how the neurobiology of emotional dysregulation and heightened reactivity to stress might inform the diagnostic criteria for mood disorders. Although emotion dysregulation and a heightened reactivity to stress (i.e., regulation of sustained affective responses when confronted with stressors or negative events) occurs in depression, its precise relation to depression (is it cause or effect?) is unclear. Examining these constructs prior to the onset of depression may elucidate their role in affecting vulnerability to depression. To investigate this, Dr. Gotlib compared a group of high-risk individuals (never-disordered daughters of mothers who had experienced recurrent episodes of depression during the daughters' lifetime) with low-risk individuals (never-disordered daughters of never-disordered mothers) to determine whether daughters of mothers with recurrent depression are characterized by a specific pattern of stress reactivity and recovery, and whether they exhibit specific patterns of neural activation in response to a sad mood induction and when regulating a negative mood state. Both groups received functional MRI scans during the induction of a sad mood and while subsequently being asked to recall positive memories to repair the sad mood. Dr. Gotlib and his colleagues found that the high-risk girls exhibited increased amygdala activity both during the development of sad mood and during its regulation; low-risk girls, in contrast, did not increase amygdala activation during these tasks. Dr. Gotlib found further that, compared with their low-risk counterparts, the high-risk girls exhibited greater HPA-axis reactivity and a slower recovery in response to a laboratory stressor. Interestingly, this pattern was most pronounced in those participants who had a short-short polymorphism of the serotonin transporter gene (which has previously been shown to be a risk factor for depression when accompanied by exposure to stress). Because both girls at risk for depression and currently depressed people are characterized by increased amygdala activation, anomalies in cortisol production, selective attention to negative stimuli, and a polymorphism of the serotonin transporter gene, these characteristics are clearly not specific to the disordered state. Importantly, Dr. Gotlib noted that they are also not specific to depression, nor are they specific to a particular developmental period or stage. Dr. Gotlib concluded that these findings raise a number of challenging questions, including how to understand the lack of specificity of this neural functioning with regard to the diagnosis of affective disorders as well as the comorbidity of different disorders (or whether comorbidity actually might be a single disorder) and whether there are finer-grained patterns of interaction of these factors and domains that may discriminate disorders. Finally, Dr. Gotlib raised concerns about the viability of making DSM diagnoses based on findings from the neuroscience literature, and suggested that it will be difficult to "carve nature at its neurobiological joints."

F. Xavier Castellanos, MD (New York, NY), in his presentation discussing how current neurobiological knowledge might influence psychiatric nosology, considered the "best case" example of cystic fibrosis (CF). Although a CF diagnosis is established in 71% of individuals by age 1, in 8% it is not established until age 10 or older. Even though a reliable and valid test for CF exists (i.e., the sweat test which measured chloride concentration), because of milder variants borderline or even near-borderline negative results may be used to diagnose CF. Thus clinical presentation, family history, and patient age still need to be taken into account. In developing a diagnostic test, even if the new system performed perfectly (i.e., sensitivity of 100%), it would still incorrectly classify as negative or borderline up to 5 of every 100 individuals with CF, given the 95% confidence intervals of 83% to 100%. CF is caused by a single locus mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. However, given that over 1,000 mutations in the CFTR gene have been found (most not associated with CF), most commercial tests look for 32 or fewer different mutations, with the result that a negative screen in not confirmatory. In psychiatry, there are currently only two conditions with known genetic etiologies (Rett's, Fragile X). Although the paucity of data might engender nosological nihilism, Dr. Castellanos argued that growing optimism is in fact justified. We now have increased awareness of the complexity of disorders and their risk factors, and there have been recent advances towards understanding the pathogenesis of psychiatric disorders (e.g., the role of the insula in supporting persistence of addiction and decreased GABA receptor subtype expression in DLPFC related to working memory deficits in schizophrenia).

The next three presentations focused on the validity and implications for classification of recent descriptions of bipolar disorder in children and adolescents. In his introductory presentation, David Shaffer, F.R.C.P., F.R.C. Psych. (New York, NY) pointed to the significant increase in the number of children and adolescents now being diagnosed as having bipolar disorder. He stated that the most significant variations from the currently accepted (DSM-IV and ICD-10) diagnosis of bipolar disorder is the very early onset, the absence of episodicity, and the frequency of ultradian mood changes. The problem of diagnosis is complicated by the difficulty of differentiating normal childhood exaggerations and exuberance from pathological grandiosity and expansiveness. Comorbidity with CD and ADHD and social phobia are common, and, unlike adults with bipolar disorder, greater "goal directedness" is unusual. The field of child psychiatry in the U.S. has accepted that there are young children who do meet DSM-IV criteria without modifying the criteria. There is controversy, howver, over whether early-onset, non-episodic cases represent: 1) an age-specific bipolar variant; 2) cases of severe ADHD; or 3) children suffering from an impairing, discrete, and hitherto unlabeled diagnosis. The high prevalence of such children is noted even when grandiosity is required instead of irritability and when criteria that overlap with ADHD are not considered for the diagnosis. Findings from several family studies of juvenile bipolar disorder differ from studies of later-onset disorder by showing relatively lower heritability, higher shared and unique environment components, and family risk for a broader range of psychiatric diagnoses. Dr. Shaffer indicated that, from a public health perspective, the most important thing is that, along with the increasing diagnosis of bipolar disorder among children and adolescents, there has been a great increase in the prescription of mood stabilizers for children and teens. While there are over 200 double-blind, randomized, controlled studies of stimulants in children and youth, there have been a very small number of such studies for those medications, such as mood stabilizers and antipsychotics, that are currently being prescribed for this condition.

Barbara Geller, MD (St. Louis, MO), focused on how the diagnostic criteria for bipolar I disorder in children can be formulated to avoid incorrectly diagnosing mania because of the overlap of many of the criteria with those for ADHD. This is important because ADHD occurs in over 80% of children with mania and is thus the main differential diagnostic problem for the pediatric age group. In this formulation, euphoric mood and/or grandiose behaviors would be necessary criteria, analogous to the way depressed mood or diminished interest is required for a diagnosis of a major depressive episode. Dr. Geller proposed that although there are differences between pediatric and adult bipolar disorder (e.g., much higher frequency of rapid cycling in children as compared to adults), these differences do not justify having separate criteria sets for children vs. adults. Furthermore, she noted that although aggression/irritability (A/I) is a highly sensitive symptom of mania across the age span, basing the diagnosis of mania on this non-specific symptom would likely result in overdiagnosis given that 95% of irritable children do not have a mood disorder. To assess pediatric bipolar disorder, the WASH-U-KSADS was developed to incorporate age specific manifestations of mania. Summarizing the evidence that child and adult bipolar are the same disorder, Dr. Geller noted that the symptom distribution is similar, the course resembled that seen in the most severely ill adults (i.e., long episode duration, daily mood switches), the predictors of outcome are similar (e.g., lack of maternal warmth, psychosis), and adults who report childhood onset of their bipolar I disorder also had comorbid ADHD. There is also evidence for overlapping genetic susceptibilities (e.g., pediatric and adult BP occur within the same families) and family based linkage disequilibrium of the BDNF Val66 allele is similar in pediatric and adult bipolar probands.

Ellen Leibenluft, MD (Bethesda, MD) presented on the clinical and pathophysiological differences between non-episodic irritability and childhood bipolar disorder. Specifically, she studied two possible clinical phenotypes of pediatric bipolar disorder: a narrow BP phenotype, with one or more full-duration episodes of mania or hypomania with elation and a broad phenotype that has been called "severe mood dysregulation" (SMD) , defined as having non-episodic symptoms, irritability (i.e., reactivity to negative emotional stimuli at least three times per week, baseline anger or sadness, at least three ADHD-like symptoms that resemble "B" mania criteria (e.g., distractibility), and impairment in at least two settings. Data from longitudinal studies suggest that chronic irritability in childhood (i.e., SMD) predicts MDD in early adulthood but not bipolar disorder. Looking at clinical variables, compared to the BP group, the SMD group was more heavily male and had a much earlier age of onset (5.1 vs.10.1 years). However, the two patient groups were similarly impaired in terms of hospital stays, number of medications, and global functioning. Furthermore, children with BP and SMD differ with respect to family history, the pathophysiology of frustration, and response to punishment. Although children with BP disorder and SMD, unlike other patient groups, have deficits in their ability to label face emotion, the brain regions engaged when processing faces differ. Based on these findings, Dr. Leibenluft concluded that 1) it would be unwise to "jump the gun" and assume that SMD is a phenotype of BD; and 2) it is important to have a population of children assessed for BD using techniques parallel to those used in adults, albeit with developmentally appropriate symptom thresholds.

Eric Taylor, MD (London, UK), in his discussion, gave an overview of suggested research issues for conditions of emotional dysregulation. He noted that an important priority should be to develop a better glossary for defining what the symptoms are by having clinician panels rate vignettes and videos. Regarding the issue of classification and weighting of symptoms, research should address the differentiating characteristics of bipolar I in children and explore the relationship of the broader spectrum of emotional dysregulation to childhood bipolar I disorder by pooling large data sets. With respect to neurobiology, research needs to address the similarities and differences between BP and ADHD, paying attention to direct comparisons and discrepant cases and using genetically informative samples (e.g., distinguishing from a trait in one twin vs. the other). Finally, it is very important to study longitudinal course by examining existing cohorts, given that many children with BPNOS eventually turn into BP (suggesting they are continuous) and that irritability and ADHD predict the future development of emotional disorders.

The next five presentations focused on the nosological issues surrounding ADHD. Luis Augusto Rohde, MD, ScD (Porto Alegre, Brazil) gave an introduction to some of the problems that have been identified with the DSM-IV criteria for ADHD. Dr. Rohde began with the results of a meta-analysis of the worldwide prevalence of ADHD. Although the overall prevalence of ADHD/hyperkinetic disorder was 5.29%, there was huge heterogeneity across the studies based on age, gender, and geographical location. The main reasons for the heterogeneity included differences in information source (parents vs. teachers), impairment criteria, and diagnostic criteria (e.g., DSM-IV vs. ICD-10). He identified a number of key research questions, including whether we are ready to incorporate neurobiological markers into the classification of ADHD, whether there is validity to differentiating the current ADHD subtypes, whether the age of onset of impairment requirement is needed, whether cut-off scores need to be adjusted according to developmental stage-age or gender (e.g., 6 symptom cut off may be too high for females), how to address problems in the construct validity of core symptoms (e.g., what does "often" mean), how to integrate the different information sources to accomplish the requirement of establishing the presence of symptoms causing impairment in different settings, and whether to continue to exclude the diagnosis of ADHD in pervasive developmental disorders but not in mental retardation.

Richard Todd, PhD, MD (St. Louis, MO) discussed the validity and usefulness of ADHD subtypes based on evidence from population- and clinic-based samples. Although the current ADHD subtypes are based on symptom clusters (i.e., predominantly inattentive, predominantly hyperactive, combined), other subtyping strategies could instead be comorbidity based (i.e., comorbid mania, ODD, substance abuse), etiology-based (e.g., genes and environment), treatment-based (e.g., stimulant responder), or some mixture of these. There are a number of problems with the current ADHD subtypes, including poor stability across development (e.g., for those diagnosed with the hyperactivity subtype, on follow-up half became combined and half became asymptomatic), no sex, age or IQ specific criteria, different reporters describing the same child differently (i.e., self, parent, teacher), limited or no family specificity, and unwarranted exclusionary criteria (i.e., ADHD excluded if criteria met for PDD). Looking at the association between candidate genes and population-based ADHD subtypes, DRD4, DAT, and DRD5 were associated with the combined subtypes whereas CHRNA4 was only associated with the inattentive subtype. Prenatal exposure to tobacco smoke and genotype at both the DRD4 and DAT gene loci dramatically increased the risk of having the combined subtype of ADHD, suggesting the presence of subtype-specific gene-environment interactions in the etiology of ADHD.

James McGough, MD (Los Angeles, CA), in his presentation, explored two issues: whether genetic studies provide any guidance as to the validity of the ADHD diagnosis and its subtypes, and whether there are new data that would support the maintenance of, or a change in, the age of onset criterion (i.e., symptoms prior to age 7 in ICD-10 and symptoms and impairment prior to age 7 in DSM-IV). Genetic studies reveal that ADHD runs in families (20-25% have affected family members) and, according to twin studies, the heritability of ADHD (0.8) is similar to that of height. Although candidate gene associations from meta-analysis do not reveal a huge impact of individual genes, it is likely that genome scan studies will provide more conclusive evidence. There is also evidence of a gene-by-environment interaction in the development of some cases of ADHD; children of mothers who smoke during pregnancy who also have two DAT+ alleles have the highest risk of developing ADHD. Summarizing the genetic findings for ADHD, Dr. McGough noted that ADHD is among the most heritable of psychiatric disorders and is likely to arise from multiple genes of small to moderate effect in interaction with environmental variables. While family studies provide validation for ADHD as a distinct biological syndrome, molecular genetics studies are preliminary and findings nonspecific. There is currently no role for genetics as a basis for diagnosis, and genetic studies have provided little, if any, support for categorical subtypes as defined by DSM-IV. Regarding the age of onset criterion, the DSM-IV requirement is not supported by its own field trial in which some individuals had an age of onset later than age 7. In another study comparing adults with "full-onset" ADHD with those who had late-onset, both groups had similar patterns of psychiatric comorbidity, functional impairment, and familial transmission. Dr. McGough concluded that that age of onset requirement was added primarily for research purposes, that few other childhood and no adult disorders are so precisely defined, that there is no empirically-based scientific or clinical value, and that raising the age of onset requirement to age 16 years identifies 98% adults with ADHD.

Jan Buitelaar, MD, PhD (Nijmegen, Netherlands) presented on the diagnosis of ADHD in adults. He argued that the ADHD criteria are not developmentally appropriate for adults, and that the childhood criteria should be rephrased or modified (e.g., "climbing and running around" is not applicable) and that new criteria should be added for adults with ADHD. Studying ADHD symptoms in adult primary care patients, he examined the distribution of symptoms in the general population, properties of individual items, internal validity, and the relationship between ADHD symptoms and impairment in functioning. In his sample, subjects with four or more symptoms were significantly more impaired, suggesting that a lower threshold might make more sense for adults. Alternatively, there is evidence that some symptoms could be left out without affecting validity; for example, with respect to inattentive symptoms, difficulty in sustaining attention and avoiding mental efforts predict the most significant impairment; and on the hyperactive side, running and climbing and being unable to relax during leisure time predict the most impairment. He concluded that the criteria should be rephrased and modified for adults, that new criteria should be introduced for both adults and children, that the diagnostic algorithm be modified to require 4 out of 9 symptoms, and that the age of onset should be set at 16 years. Additional research is needed to determine new types of impairment more applicable to adults, and to determine the validity of subtypes in adults.

Russell Barkley, PhD (Syracuse, NY) concluded the presentation section on ADHD by providing his perspective on the issues in revising the ADHD criteria for adults. His goal was to look for impairment in a more comprehensive way and to develop symptoms that best distinguish between adults with ADHD and controls. He argued that the one symptom that best rules out normal individuals (and applies to children as well as adults) is "often easily distracted by extraneous stimuli." More important, however, are items that discriminate adults with ADHD from adults with other mental disorders. These include items such as "fails to give close attention to details," "difficulty sustaining attention to tasks," "fails to follow through on instructions," and "difficulty engaging in leisure activities quietly." Dr. Barkley also proposed creating a new symptom pool based on concepts of executive functioning (inhibition, working memory, organization to time and future events, emotional self-control, self-motivation and planning), ultimately ending up with a list of nine items which performed better than the 18 DSM-IV items. Dr. Barkley also covered a number of technical issues, including whether "developmentally inappropriate" should be specified (as in mental retardation or learning disorder) and whether "often" refers to absolute frequency or frequency relative to a peer group. He argued against specifying fixed thresholds and symptoms across developmentally changing constructs, and also questioned the utility of requiring symptom onset before age 7 because of a lack of qualitative or quantitative differences between those with symptom onset before age 7 and those between age 7 and 16, and because of problems with recall.

The next four presentations considered the relationship between CD and ODD and their continuity with antisocial behaviors. Reviewing evidence from clinical longitudinal studies, Jeffrey D. Burke, PhD (Pittsburgh, PA) reported that although ODD and CD are related disorders, they are distinct. For example, there may be two types of ODD, both of which are comorbid with ADHD; one that is prodromal to CD and one that is subsyndromal, with little likelihood of progression to CD. Furthermore, although it is common for CD to occur almost exclusively with ODD in clinical samples, in community samples there are typically relatively high proportions of CD cases without ODD (from 38% up to 74%). With respect to the relationship between ODD, CD, and depression, ODD but not CD symptoms predict depression. Negative life events (e.g. removal from the home, grade retention, peer rejection) explains any association between CD and depression. One possible explanation for the link from ODD to depression stems from factor analyses which suggest that ODD may yield two factors: a negative affect factors (i.e., touchy, spiteful and angry) which predicts depressive symptoms and an oppositional behavior factor (i.e., argues, defies, and loses temper) which predicts later CD symptoms. The relationship between negative affect and later depression may explain the relationship between SMD and depression, given that children with SMD are characterized by chronic negative affect. In fact, ODD may represent a point of divergence between an antisocial behavioral pathway and an affective disorder pathway. Negative affective symptoms of ODD predict affective disorder symptoms and suicidal ideation, and oppositional behaviors and interpersonal callousness predict antisocial behavior symptoms. Therefore, assigning a diagnosis of ODD, even when CD is also diagnosed, may provide additional information regarding the risk for affective disorders. Regarding the relationship between ODD, CD, and antisocial personality disorder (ASPD), ODD, and perhaps an ODD behavioral component in particular, predicts CD. CD, however, measured in childhood or late adolescence, predicts ASPD. Regarding the question of whether the criteria for ODD, CD, or ASPD can be revised to permit improved prediction of transitions between the disorders, including items covering interpersonal callousness with symptoms of ODD may improve the prediction of CD, and adding interpersonal callousness with symptoms of CD may improve the prediction of ASPD and psychopathy. Dr. Burke concluded by noting that if ODD is intended to reflect the earliest step in a hierarchical pathway of antisocial behavior, the symptoms should be revised to be more specific to later antisocial outcomes. However, if ODD is instead to be useful as an indicator of general dysregulation, perhaps CD should be regarded as one possible outcome.

Frank Verhulst, MD (Rotterdam, Netherlands) examined the relationship between ODD, CD, and ASP using epidemiological longitudinal data. Although it is established that conduct disorder is strong predictor of antisocial personality in adulthood, very little is known about risks of ODD for adult functioning, except for the inclusion of ODD behaviors in ADHD. Using longitudinal follow-up data from a study on a cohort of 2000 children and adolescents from the general population of the Netherlands to examine the relationship between ODD and CD symptoms and the future development of ASPD, Dr. Verhulst determined that CD predicts ASPD in the general population, that ODD is a less strong predictor of ASPD, that ODD and ADHD do not add to the prediction of ASPD when comorbid CD is accounted for, and that adolescent-onset CD related problems did not predict ASPD. He concluded that the findings support the distinction between ODD and CD, and that an onset distinction with "life course" problems predicts a worse outcome than adolescence onset.

Hector Bird, MD (New York, NY) presented on gender, age, and cultural variations in the presentation of ODD and CD. He noted that it is critical to apply criteria uniformly for cross-cultural comparability in research and that one needs to assume that sets of core symptoms of psychiatric disorders are universally applicable across cultures. This assumption does have empirical support, given that studies in a broad variety of cultural settings have yielded relatively small differences in syndrome scales across studies and that there is considerable similarity across cultures in the associations of syndromes with risk factors. In order to examine the relationship between ethnicity and environmental context in the presentation of ODD/CD symptoms, Dr. Bird presented data from a study that looked at the development of antisocial behaviors in children of Puerto Rican background living in two contrasting contexts (i.e., San Juan, PR and South Bronx, NY). Given that these children from the same ethnic group ended up showing dissimilar patterns of disruptive behavior disorders in the two settings, he suggested that differences across studies may be more a function of context than ethnicity. Furthermore, focus groups indicated that, at least in this Hispanic group, there were no unique features of antisocial behavior or disruptive behavior disorders that were culturally specific. As expected, both cross-sectionally and longitudinally, the risk for boys is greater than girls and older children have greater risk than younger. Finally, despite the low prevalence of CD in this study, children with ODD but no CD at baseline were more likely to develop CD in within the next two years than those with no ODD at baseline.

Kang-E Michael Hong, MD (Seoul, South Korea), in his presentation on the relationship between CD and ODD and antisocial behaviors, noted that a number of neurobiological predisposing factors have been suggested by investigators, including genetic predisposition, prenatal and perinatal problems, prematurity, neurological disorders, malnutrition, brain injury, temperament, neurobiological susceptibility, intelligence, and other psychopathology. These non-specific factors become etiological for a specific disorder when combined with specific psychosocial factors Psychosocial factors that contribute to the development of ODD include socioeconomic status, family dysfunctions, parental psychopathology, and distinct mother-child interaction problems characterized by conflicts, emotional outbursts and power struggles. Psychosocial factors contributing to the development of CD include family problems, parenting problems (i.e., negative, rejecting, lacking maternal warmth), parental psychopathology, and peer rejection. Factors contributing to the progression of ODD to CD include earlier onset, family variables, atypical family structure in boys, and physical fights. Factors contributing to the progression of CD to antisocial behavior include younger age, chronic delinquent behavior, low constraint, academic failure, peer rejection and social bond, comorbidity, family, parenting, and partner problems, neurobiological factors, and a callous unemotional trait. Factors contributing to the progression from ADHD to CD include low scores on social phobia symptoms, and unique familial underpinnings, with some cases of ADHD becoming earlier onset CD. Dr. Hong concluded that ODD and CD are clearly differentiated both clinically and empirically and made several suggestions including 1) conducting a pathway analysis (e.g., LISREL) to clarify the relationships among the externalizing disorders, 2) recognition of a callous-unemotional trait as a core factor in a subgroup of CD and ASPD individuals with consideration of setting up a separate subtype; 3) recognition of "reactive" disruptive behavior disorders associated with rapid, massive social changes in many developing countries (e.g., the rapid modernization of Korea in the past 40 years has brought about dramatic social and family changes that lead to changes in value orientation, confusion, and crisis in child rearing, all of which create many reactive developmental, emotional and behavior problems); and 4) a consideration of subtyping schemes for ODD that might include ODD with ADHD vs ODD with CD vs ODD only; subsyndromal CD vs. ODD prodromal to CD; and underdisciplined OD behaviors (i.e., spoiled child) vs ODD; and a consideration of a subtyping scheme for CD (CD with callous unemotional trait vs. CD without).

The final two presentations focused on the relationship between externalizing disorders and substance-related disorders. Wilson M. Compton, MD (Bethesda, MD) began his presentation by asking whether addictive disorders might be a late manifestation of CD and ASPD. Given that antisocial behaviors have a strong association with drug addiction, this might explain the "common genetic vulnerability" across substances. Furthermore, continuous antisocial behavior is strongly associated with drug use and substance disorders, especially among women, whereas childhood limited antisocial behavior is less associated with drug use and substance disorders. Given that persistent conduct syndromes are remarkably strongly associated with drug use and substance disorders, genetic and/or environmental exposures might predispose to both conduct problems and substance use disorders. A second theme presented by Dr. Compton is that addiction can be considered a developmental disorder with onset during adolescence. Recent studies have shown that maturation of the brain's gray matter moves from back to front and that during adolescence, cognition-emotion connections are still undergoing development. In adults, emotional self-regulation is normally implemented by neural circuits linking various prefrontal regions and subcortical limbic structures, raising the question of whether variations in frontal lobe development pose a general risk for externalizing disorders, including addictions and whether frontal lobe functioning can be strengthened. Dr. Compton concluded by considering some of the possible implications for DSM-V, including how diagnosis can account for the high comorbidity of addiction and conduct disorders. One possible suggestion from the audience was to add early substance abuse as a criterion for conduct disorder, as it was in DSM-III.

Guilherme Borges, DrSc (Mexico City, Mexico) reported on the comorbidity between psychiatric disorders and substance dependence in Mexican adolescents, using results from the Mexican Adolescent Mental Health Survey. He reported that a large amount of comorbidity was found among Mexican adolescents (25% with two or more disorders) and that temporally, primary mental disorders were strong predictors of first onset of substance dependence (with odds ratios (OR) ranging from 1.7-4.7). In general, all classes of disorders were associated with the new onset of substance use disorders and among externalizing disorders of childhood, risk estimates seem very similar (OR ranging from 2.3-6.5), with the exception of ADHD (no increase in risk) and a larger OR for CD. Further analyses at the symptom level are to be done next.

Upon conclusion of the presentations, participants reconvened in three breakout groups (ADHD, pediatric bipolar, ODD/CD) to formulate recommendations for research and suggestions for future versions of the classifications of mental disorders (e.g. in the DSM-V and ICD-11). Each workgroup was assigned (in advance) a series of questions/issues to address.

The ADHD group considered 21 questions.

  1. Are the high levels of comorbidity that have been noted between ADHD and ODD in most studies a function of overlapping criteria? The group concluded that there is enough research data from clinical and community samples to support the validity of the two disorders as independent yet overlapping constructs.
  2. Should comorbid ADHD/ODD be categorized separately from either ADHD or ODD? The group concluded that there is no evidence supporting ADHD/ODD as a specific category or distinct disorder from ADHD.
  3. Has research emerged to support the validity of the three main subtypes of ADHD? The group decided that the DSM-IV subtypes lack validity. The workgroup suggested the need for more research to assess the validity of having new ADHD subtypes, taking advantage of modern statistical approaches. More research is needed to assess the "sluggish cognitive tempo" construct and the subdivision of ADHD according to exo- and endo-phenotypes or comorbid conditions (e.g., ADHD + CD). Efforts should be made to identify more clinically useful and valid subtypes than currently exist in DSM-IV.
  4. There are age of onset differences between ICD-10 (symptoms) and DSM (impairment): It has also been suggested that age of onset is usually "age of recognition," which, in turn, is dependant on observations made in school. Are there new data that would support the maintenance of, or a change in, this criterion? The group suggested that the age of onset/age of onset of impairment criterion - 7 years - should be abandoned in favor of a "childhood to adolescent onset". They suggested that existing datasets be explored for further guidance as to specific age and whether it should be expressed in terms of onset of symptoms or of symptoms producing impairment.
  5. Some behaviors are represented by multiple DSM-IV criteria, e.g., A1b. difficulty sustaining attention in tasks and play, A1d. fails to finish schoolwork, chores, A1h. is easily distracted (which all address the same feature), 2B. often leaves seat, 2C. runs about or climbs excessively in situation in which inappropriate, and 2E. is often on the go or often acts as if driven by a motor. This overrepresentation of some features probably weights caseness towards those clinical features. A single clinical feature represented by three criteria allows that single behavior to provide 50 percent of the "necessary" requirements. Should the overlap between criteria be reexamined, either by grouping alternate descriptions of a single behavior or by weighting overlapping and unique criteria separately. In other words, can the existing item sets be reduced without loss of utility or accuracy of diagnosis? The group concurred that existing datasets need to be explored to address this issue.
  6. Should more numerically specific definitions of frequency/intensity for criterion "A" items be used instead of "often"? The group recommended that it should be made explicit in the text that "often" refers to a symptom occurring more frequently than is expected for typical individuals at the same age.
  7. Is there evidence from longitudinal studies that show that individual ADHD symptoms wax and wane with age without the diagnosis losing its integrity i.e. retaining its ability to predict future course? The group noted that there is evidence that developmental inappropriateness may be more persistent than meeting full diagnostic criteria in DSM-IV. Existing datasets need to be further explored to evaluate developmental sensitivity of items and item sets across the lifespan.
  8. Research findings suggest that teachers' observations are more reliable and relate better to a clinical diagnosis than do mothers'. Is this information robust, and how should clinicians integrate information from disparate sources with respect to both the presentation of symptoms and impairment in different settings? The group suggested that it be made explicit in the text that the best clinical approach is to use multiple informants and multiple sources reporting on symptoms and impairments from different settings and examined over the history of the individual and not just the current time or context.
  9. Should the diagnosis of autism or PDD exclude a diagnosis for ADHD? The group asserted that there is no evidence for this approach.
  10. Is there sufficient evidence from cognitive, anatomical, or neuroimaging studies to regard basic cognitive or structural abnormalities as either necessary or sufficient in ADHD and therefore worthy of inclusion in the criteria? The group agreed that although this is not yet the case, these studies can improve our knowledge on the heterogeneity of the disorder. The group suggested exploring existing and new datasets for further guidance.
  11. Do genetic studies provide any guidance as to the validity of the diagnosis and its subtypes? The group noted that behavioral genetic studies have provided evidence of validity for the disorder but not for its subtypes. Current molecular genetic studies do not contribute to specific diagnosis. Future behavioral and molecular genetic studies may increase our knowledge of the etiology and the heterogeneity of the disorder.
  12. Should different weights be given to individual ADHD symptoms? The group decided that more work needs to be done to explore existing datasets concerning the weighting or importance that should be given to some symptoms over others.
  13. Is the current symptom list inadequate for the identification of ADHD in adults? The group suggested that existing datasets be mined further to determine if DSM-IV items sets can be modified for use with adults, if new thresholds need to be used, and/or if new symptoms can be identified.
  14. Should the term "developmentally inappropriate" be more explicitly defined in the text? The group stated that it should be more explicitly defined to refer to typical individuals of the same age. The use of well-standardized rating scales or other measures may assist the clinician in making this determination
  15. Should the window of observation for symptoms of 6 months be broadened to 12 months for the preschool age group? The group recommended that existing datasets be examined to further address this question.
  16. Is there a need to retain two separate thresholds on two symptom lists or could a single threshold representing total symptoms across the two lists be more appropriate? The group declared that the issue is important and should be explored using existing datasets and various statistical techniques.
  17. Should different thresholds be applied based on developmental stage? (For example, some evidence shows that if DSM-IV lists were retained, thresholds of 4+ on either list would identify adults with ADHD who are impaired). The group noted that the answer will depend on answers to earlier questions concerning the need for new item sets for different ages.
  18. Should the inattention list be retitled to reflect its broader representation of working memory and executive function deficits (e.g., inattention/metacognition, inattention-working memory, inattention-executive dysfunction, etc.)? The group proposed that this can be addressed through text clarification and that the item sets capture a broader array of cognitive deficits than the term inattention implies.
  19. Should items reflecting poor behavioral inhibition or impulsivity be added and hyperactive symptoms be reduced to reflect current conceptualizations of ADHD in which inhibition is a core construct? The group recommended that existing datasets be explored to address this issue.
  20. Should the term impairment be clarified to indicate that it is defined relative to the general population average or norm, what is known in law as "the average person standard?" [This is to avoid current clinical confusion and even abuse of the diagnosis by clinicians who diagnose people whose impairment is defined relative to high functioning local peer groups or relative to high general intellectual ability (intra-person discrepancy). It would also make the definition consistent with legal rulings related to the Americans with Disabilities Act.] The group decided that further efforts to operationally define the term impairment is needed at higher-level DSM committees given that the term appears across many other disorders besides ADHD.
  21. Should more domains of major life activity be added to the impairment criterion (home, school, and occupation) to reflect the more numerous and important domains in which teens and adults must function (i.e., social, marital/cohabiting, child-rearing, community activities, money management, driving, etc.)? The group agreed that the text should be revised to clarify this.

The pediatric bipolar group began their presentation with some basic recommendations, namely that bipolar disorder is a mood disorder and that it is important not to leave behind non-episodically irritable, hyper-aroused children who do not meet the "narrow" bipolar disorder phenotype since they are more common than those with DSM-IV bipolar disorder and are as severely impaired. In order to learn more about the role of irritability in the definition of mania, the group recommended that researchers tap into large longitudinal data sets to look at the outcome and phenomenology of irritability, that more detailed questions be developed so that a typology of irritability can be constructed, that there be observational studies of irritability across diagnoses, and that a scale be developed to measure severity of irritability. With respect to euphoria, the group recommended that the normative boundaries of happiness be examined to determine how those boundaries vary developmentally. They suggested that videotapes of very happy children be shown to clinicians to see if they would be diagnosed as euphoric and that it would be helpful to include peer complaints to help differentiate normative from deviant. They also recommended studies comparing different assessment methods for pediatric bipolar disorder in which children are diagnosed according to the different techniques with the focus on the discordant cases. Finally, they recommended more studies contrasting pre-pupertal vs. adolescent onset bipolar disorder in terms of prognosis, familiality, heritability, and treatment response.

The group then addressed the following eight questions:

  1. Should the same criteria be used to diagnose BD in adults and children and are there children who meet adult criteria? The group asserted that children who meet strict adult criteria can be identified and that they are an important comparison group for children whose diagnostic status vis-à-vis bipolar disorder is unclear. However, the diagnostic system also needs to accommodate children who do not meet DSM criteria for bipolar disorder; these include children with BP-NOS (episodes of too short duration) and children with SMD-like pictures (i.e., non-episodic, impairing irritability and ADHD-like symptoms).
  2. Is there evidence to justify a BD diagnosis for children who don't meet adult criteria? Currently, no such evidence exists. However, it would be justified to assign a bipolar diagnosis to children who do not meet adult criteria if children are identified who convert to adult-like bipolar, have a family history of bipolar disorder, similar comorbidity patterns, similar pathophysiology and genetics, and perhaps similar treatment response.
  3. Should severe mood dysregulation (SMD)-like children be diagnosed with BD or should a new category be created (and would it be a mood disorder)? The group suggested that although these children should not be diagnosed with bipolar disorder, it is important not to leave these severely impaired children behind as they are much more common than DSM-IV BD. Options for these children include creating a research definition that might go in the research appendix, creating a subtype of ADHD or ODD, or putting the condition in the bipolar spectrum.
  4. Should those B criteria in mania that overlap with ADHD simply be restated or should they be given a special adjusted weight? The group recommended that there be comparative studies of symptoms in ADHD with and without negative affect, and Bipolar Disorder with and without ADHD.
  5. Should there be different definitions for episodes and cycling in adults vs. children? The group suggests employing the same external validators as would be used to answer question #2.
  6. How should manic irritability, grandiosity, etc be operationalized? At the current time, it is important to maintain consistency between adults and children in these definitions in order to facilitate comparative developmental studies. To ascertain whether, ultimately, different definitions should be used for children vs. adults, the group suggested that a glossary be developed at the start (based in part on systematic study of clinicians ratings of standardized clinical information such as video) which would then be modified as data is gathered and that these definitions be applied across diagnoses (not just limited to bipolar) and cut across the age span. It was also suggested that the pediatric bipolar group work with the adult bipolar group as those who study adult bipolar have already worked on definitions of episode, change in polarity, etc.
  7. Can pathophysiological research inform these questions? The group responded by noting that a lot of work on phenomenology is needed before it makes sense to focus on pathophysiology.
  8. What are the most important future genetic and pathophysiological studies? The group suggested genetic behavioral studies (especially twin studies), studies of irritability in controls and in those with ADHD, and agreeing on definitions that will help to accumulate large samples of prepubertal and adolescent BD subjects for inclusion in large genetic studies.

The ODD/CD group considered seven questions:

  1. Has further evidence emerged to justify the current division of ODD and CD into two different disorders? ICD maintains them as a single entity. Can we learn whether this is a useful distinction for understanding course, natural history, or psychopathology? The group concluded that ODD and CD should continue to be kept separate, in part based on their differential associations with other disorders. It still needs to be determined whether ODD without comorbidity but with impairment, is persistent over time and is related to poor outcomes. The group also suggested it needs to be demonstrated whether relationships which have been found using continuous symptom counts are consistent when categorical diagnoses are examined. Relevant to addressing this question is the associated issue of how potential subsets of ODD symptoms might relate to factors within CD, such as overt and covert distinctions. It may also be helpful to include a requirement that ODD be present across more than one context in order to meet criteria for the disorder. Furthermore the effect of informant differences needs to be more fully understood and accounted for. Finally, the relational nature of ODD suggests that transactions with others are essential to the disorder. The question remains as to whether and how this might need to be more explicitly reflected among the symptoms.
  2. Some DSM CD requirements would be more appropriate in a combined diagnosis. For example, ODD has more high-frequency ("often") items, which would make the requirement that behaviors have a "repetitive and persistent pattern" easier to apply. The group knows of no evidence to support decisions regarding the application of frequency criteria and the group agrees that for some symptoms, such a modification may be helpful, but for others, the occurrence one time seems to be sufficiently distinctive of disorder. Furthermore, there is controversy regarding the statement that a diagnosis of CD should not be made for children in certain circumstances or contexts. The group is not aware of any evidence to say whether or not full criteria for CD are ever met based solely on "acceptable" adaptations to context.
  3. Can the criteria for CD, ASPD, or both be revised to permit a smoother transition between the two? Is the notion of ASPD reasonable, or should it be an Axis I disorder? What constitutes "evidence of CD in the past?" Some consideration of callousness in regards to CD may improve the prediction of transition to APD. The group's consensus is that APD should be moved to Axis I, and regarded as "antisocial disorder." The description of the criteria must be changed to be consistent with social changes and social context (some common operationalizations of symptoms are anachronistic). Symptoms of antisocial behavior in adulthood must also be devised or operationalized from a developmental standpoint.
  4. Have data emerged to justify subtyping CD by age of onset? The group agrees that trajectory analyses tend to reveal early onset and adolescent onset antisocial behavior groups, although they disagree whether there should also be other subgroups. The results hold for both boys and girls. Some evidence suggests that callous and unemotional traits are associated with early onset trajectories, and may be associated with poorer outcomes.
  5. ICD combines CD with other Axis I diagnoses such as MDD into a single compound diagnoses (e.g., depressive conduct disorder). This assumes that the comorbid entity has a distinctive set of antecedents, course, and response to treatment. Is there evidence for this? The group reported that there is no support for this. Available evidence suggests that it would be best to allow the diagnosis of separate disorders if both are present. It may be useful to add a modifier of "with depressive features" if full criteria for depression are not met. ODD appears to explain the relationship between CD and depression. Further evidence must be collected regarding the influence of negative affective symptoms of ODD and their association with affective disorders.
  6. Should a separate category of psychopathy (perhaps a subcategory of CD) be defined, since pathophysiological research in the area of callousness has allowed the measurement of that descriptor? Ample evidence implicates callousness in the course of antisocial behavior. It remains to be determined whether it would be best incorporated as a modifier for CD, within the symptoms for CD, or somehow separate from CD. It will be particularly important to determine whether it predicts poor response to treatment.
  7. Can we make more progress in measuring the extent of age and gender bias in CD, and can these be corrected? The group commented that question should read, "Is there age and gender bias?" New technologies may be used to determine whether differences in item endorsement exist between boys and girls, and over the course of development, and what implications these have for the course and outcomes of ODD and CD.

All papers will appear in a monograph published by American Psychiatric Press, Inc.

From: 
Email:  
To: 
Email:  
Subject: 
Message: