Expert group and authors

Charles cohen-salmon, Vulnerability, Adaptation and Psychopathology, CNRS UMR 7593, CHU (University Hospital Centre) Pitié-Salpêtrière, Paris

Sylvana côté, School of Psychoeducation, University of Montreal, Canada

Pierre fourneret, Institute of Cognitive Sciences, UMR 5015, CNRS, UCBL1, Bron

Isabelle gasquet, Eating Disorders in Adolescents, INSERM U 669 and Department of Medical Policy, AP-HP (Paris Hospitals and Public Aid), Paris

Antoine guedeney, Child and Adolescent Psychiatry, AP-HP, , CHU Bichat-Claude Bernard, Université Paris VII, Paris

Michel hamon, Neuropsychopharmacology, INSERM U 677, CHU Pitié-Salpêtrière, Paris

Béatrice lamboy, National Institute for Prevention and Health Education, Saint-Denis

Marie-France le heuzey, Psychopathology in Children and Adolescents, Robert Debré Hospital, Paris

Grégory michel, Department of Psychology, François Rabelais University, Tours and Psychopathology in Children and Adolescents, Robert Debré Hospital, Paris

Jean-Philippe reneric, University Department of Child and Adolescent Psychiatry, CH (Hospital Centre) Charles Perrens, Bordeaux

Richard E. tremblay, Chairman of Canadian Research into Child Development, University of Montreal, Canada and Faculty of Social Sciences, University of Utrecht, Holland

Mathias wohl, Phenotypic, Developmental and Genetic Analysis in Addictive Behaviour, INSERM U 675 and Louis Mourier Hospital, AP-HP Paris VII, Colombes

Other contributors

Catherine blatier, Clinical and Pathological Psychology, University of Grenoble

Michel born, Psychology of Delinquency and Psychosocial Development, University of Liège, Belgium

Marie choquet, Eating Disorders in Adolescents, INSERM U 669, Maison des Adolescents, Cochin Hospital, Paris

Jacques fortin, Department of Epidemiology and Public Health, Calmette Hospital, Lille

Laurent mucchielli, UMR 2190, Sociological Research Centre on Law and Penal Institutions, Guyancourt

Nadine neulat-billard, Department of School Teaching, Ministry of National Education, Higher Education and Research, Paris

Bernard ribiolet, Inter-Ministerial Delegation for the City and Social Urban Development, Saint-Denis La Plaine

Scientific and editorial coordination

Fabienne bonnin, Scientific assistant, INSERM collective expertise center, Faculty of Medicine Xavier-Bichat, Paris

Catherine chenu, Scientific assistant, INSERM collective expertise center, Faculty of Medicine Xavier-Bichat, Paris

Jean-Luc daval, Head of Expertise, INSERM collective expertise center, Faculty of Medicine Xavier-Bichat, Paris

Jeanne étiemble, Director, INSERM collective expertise center, Faculty of Medicine Xavier-Bichat, Paris

Anne-Laure pellier, Scientific Assistant, INSERM collective expertise center, Faculty of Medicine Xavier-Bichat, Paris

Bibliographic Assistance

Chantal rondet-grellier, Documentalist, Centre d'Expertise Collective de l'Inserm, Faculty of Medicine Xavier-Bichat, Paris

Conduct disorder manifests as various symptoms defined in the classification systems

Conduct is defined as a way of acting and behaving, a definition that includes a degree of moral judgement. The phrase "conduct disorder" implies a behaviour pattern that contravenes accepted social rules. Such a disorder is therefore at a junction between psychiatric, social, and legal concepts.

Historically, since Lombroso's idea of the "born criminal" ¹ , paediatric conduct disorders have raised questions regarding the frontier between responsibility and blame as well as between law and medicine. Some experts have proposed constitutional explanations based on ideas such as a "congenital abnormality of instinct", whereas others (influenced by ideas from psychoanalysis) point to the importance of environment. Different terminologies reflect different attitudes, often coloured with some form of moral judgement, e.g., perverse child, guilty child, psychopath, etc.

In 1968, the term conduct disorder was first used in the Diagnostic and Statistical Manual of Mental Disorder (DSM-II) drawn up under the auspices of the American Psychiatric Association (APA). In 1977–1978, the term was taken up by the World Health Organisation (WHO) in its updated International Classification of Diseases (ICD-9).

In DSM-IV and its updated version (DSM-IV-TR), one chapter is entitled "Attention-Deficit Disorder and Disruptive Behaviour", in which a distinction is made between four entities, namely attention-deficit/hyperactivity disorder (ADHD), conduct disorder, oppositional defiant disorder (ODD), and non-specific disruptive behaviour disorder.

Conduct disorder covers three criteria (A, B, and C). Criterion A defines the disorder as "A repetitive and persistent pattern of behaviour in which the basic rights of others or major age-appropriate societal norms or rules are violated, as manifested by the presence of three (or more) of the following criteria in the past 12 months, with at least one criterion present in the past 6 months." Symptoms are broken down into four broad categories:

• aggression to people or animals
• destruction of property (without physical aggression
• deceitfulness or theft
• serious violations of rules

Criterion B addresses the degree of impairment in social, academic, or occupational functioning. If the individual is 18 years of age or older, a diagnosis of conduct disorder can only be made if the criteria for antisocial personality disorder (criterion C) are not met.

Two parameters have to be taken into account, namely the age of onset and the severity of the disorder. With respect to age, the distinction is made between childhood-onset (prior to the age of 10) and adolescent-onset (after the age of 10). These two types differ in terms of symptoms, development, relative prevalence in boys and girls, and co-morbidity. The severity of the problem is graded as mild, moderate, or severe.

Similarly, ICD-10 defines conduct disorder as a repetitive and persistent pattern of behaviour in which the basic rights of others or societal norms or rules are violated, again with the idea that the behaviour has persisted for at least 6 months. This system gives a list of 23 identifiable symptoms, and the type of conduct disorder depends on which of these symptoms are found together in the patient:

• socialised or un-socialised, depending on whether the child has normal relationships with his/her peers
• conduct disorder confined to the family environment
• oppositional defiant disorder (ODD), which is included in conduct disorder

DSM-IV and ICD-10 agree on the global definition of conduct disorder and on the value of differentiating childhood-onset and adolescent-onset. However, there are differences: ICD-10 emphasises the socialisation aspect and classifies ODD as a subtype of conduct disorder, whereas in DSM-IV, ODD is a distinct, strictly personalised diagnostic category.

In ICD-10, attention-deficit/hyperactivity disorder is called hyperkinetic disorder and there is a special mixed category for those in whom hyperkinetic disorder and conduct disorder co-occur.

Although the first version of the French classification system for mental disorders was quite different from the international systems, the version published in 2000 was more similar, including a "conduct and behaviour disorder" category.

The different classification systems have made no distinction between symptoms in boys and girls. In girls, the data on age of onset are variable, and the prognosis for late-onset is considered as being the same as for early-onset, unlike observations in boys. Some experts propose that, to distinguish between early and late onset in girls, the threshold of puberty is more relevant than the age of 10. Adolescent-onset in girls is characterized by sexual behaviour problems, premature pregnancy, and depression. In general, girls commit fewer aggressive acts and more frequently manifest manipulative behaviours.

The virtue of category-based classification systems such as ICD and DSM is that they specify groups of symptoms and make it possible to constitute uniform groups of patients for studies and trials to evaluate treatment modalities. However, such systems do not characterise relationships between the different types of disorder or the probability of one developing into another. As mentioned above, ODD is classified as a conduct disorder in ICD-10, whereas it is a quite separate entity in DSM-IV. It is considered as a precursor for conduct disorder in some cases and as co-morbidity in others. Should ADHD, which is classified as a separate entity, also be considered a conduct disorder or a co-morbid condition?

Finally, it is important to note the definition of antisocial personality as being one who exhibits "a pervasive pattern of disregard for and violations of the right of others occurring after the age of 15". To meet these diagnosic criteria, the subject must be at least 18 and must have met the diagnostic criteria for conduct disorder since at least the age of 15. Thus, antisocial personality disorder is both a differential diagnosis and a possible sequel to conduct disorder. Some experts have questioned the notion of antisocial personality as a diagnostic entity and its inclusion in classification systems for mental disorder, given the absence of any effective treatment coupled with the possibility that the diagnosis might be used to avoid penal sanctions.

The prevalence of conduct disorder in boys of 15 years old is between 5% and 9% in the general population

Many studies, including some multiple cohort studies with long follow-up times, have yielded high-quality data on the prevalence of conduct disorder in children and adolescents in the general population. Key factors affecting the pertinence of such studies are how representative the sample is, the use of validated instruments for evaluation, and high follow-up rates (for prospective studies). Fewer studies have focused on young offenders or those in special educational establishments.

On the basis of the classification systems used to establish the psychiatric diagnoses (ICD or DSM), a distinction is usually made between ODD and conduct disorder (which itself is divided into two types, namely aggressive and non-aggressive).

In studies focusing on children, the parents are not the ideal source of information, and the diagnosis is often established after talking to the child and his/her teachers. In adolescents, it is the subjects themselves that are the best source of information.

In the population as a whole, the prevalence of ODD peaks at 8–10 years of age (reaching 3–4%) in both sexes, thereafter dropping (to 1–3%). The prevalence of conduct disorder increases up until 15 years, after which it stabilises in boys (at 5–9%) and drops in girls. Aggressive-type conduct disorder is very rare in girls.

Conduct disorder tends to persist, with two-thirds of children thus diagnosed still being diagnosed as having the disorder in adolescence.

Studies in populations of delinquents show prevalence of 30–60% for conduct disorder and of 8–30% for ODD. The scarce data available on delinquent girls tend to suggest that both problems are equally as common in this group as in delinquent boys. The prognosis is usually considered as poorer if the disorder appeared before the age of 10 (childhood-onset) as opposed to later (adolescent-onset). However, among girls in prison, later onset seems to be associated with more severe conduct disorder than in boys.

The results of studies show that the great majority of adults with antisocial personality disorder have a history of conduct disorder, and that about one-half of all young people with conduct disorder will go on to have antisocial personality disorder in adulthood. The risk of developing antisocial personality disorder seems to be higher in young people with a conduct disorder who also abuse psychoactive substances in their teenage years and adulthood.

In France, a single study conducted in 6–11-year-olds in 18 primary schools in Chartres reported a prevalence of 6.5% (9% in boys, 3% in girls) with 2.8% with a severe form. The level was significantly higher in special classes (17%) than in normal classes. No difference emerged between private and public schools.

Conduct disorder is often associated with other mental problems

Conduct disorder is rarely found in isolation. The international literature reports a high level of inter-current morbidity with a great diversity of co-morbid conditions. One of the most commonly associated psychiatric disorders is ADHD, with several epidemiological studies pointing to a continuity between paediatric ADHD and conduct disorder in adolescence. ADHD during childhood seems to be all the more predictive of subsequent conduct disorder if it is combined with ODD. Other more recent studies have suggested that the symptoms of ODD and ADHD are often associated with physical aggression (a symptom of conduct disorder) from very early childhood.

The relationship between ODD and physical aggression in small children could lead to specific problems, both clinical and developmental, in adolescence, thus constituting a prodrome for adult antisocial personality disorder. According to other experts, the clinical picture in which hyperactivity, impulsiveness, and attention deficit are associated with conduct disorder would be characterised by high rates of aggression and criminal violence.

Co-morbid depression is also common. Depression could be involved in the development and persistence of conduct disorder and vice versa, and such a synergy could increase the likelihood of suicide. Studies conducted in clinical populations have shown co-morbidity with bipolar disorder in 17–42% of cases. Some experts believe that bipolar disease in childhood could predispose to the development of conduct disorder, whereas others believe that the simultaneous onset of both problems reveals the existence of a specific subgroup. Co-occurring conduct disorder and depression would increase the likelihood of developing antisocial personality disorder in adulthood as well as increase the chance of dependence on alcohol or other psychoactive substances.

General population studies indicate a fairly high level of co-occurring/co-morbid anxiety disorders. Most studies, of both general populations and patients, indicate that anxiety problems have a moderating effect on the severity of both conduct disorder and antisocial behaviour. However, this effect seems to disappear in severe forms of conduct disorder, especially in prison inmates. One anxiety problem that is commonly associated with conduct disorder is post-traumatic stress disorder. This co-morbidity is most often observed in adolescents who have experienced sexual violence or those with severe conduct disorder (e.g., those in prison). Moreover, adolescents with conduct disorder are at higher risk of post-traumatic stress disorder because they tend to expose themselves to situations in which trauma is likely. Finally, the existence of post-traumatic stress disorder increases the risk of conduct disorder.

Psychoactive substance abuse is commonly associated with conduct disorder. The age of onset and severity of the conduct disorder are key factors in the initiation and maintenance of abusive behaviour, and reciprocally, early substance abuse is associated with more severe forms of conduct disorder. It seems, therefore, that the synergy might act in both directions insofar as the one problem exacerbates the other. In addition, this strong co-morbidity could be the expression of a cumulative continuity: drug abuse promotes delinquent behaviour, and delinquent behaviour involves the use of illicit substances. Substance-abusing teenage girls with co-occurring conduct disorder are more commonly depressed or anxious, whereas boys with a similar clinical profile tend to commit violent or criminal acts.

In the opinion of some experts, the association of conduct disorder and ADHD seems to promote psychoactive substance abuse, whereas others emphasise the association between conduct disorder and ODD. With respect to cannabis, longitudinal studies have shown that conduct disorder is predictive of early use, a link that is stronger in girls. Finally, some studies have shown that the superimposition of depression or anxiety on a developing conduct disorder might increase the likelihood of substance abuse.

Apart from co-morbid psychiatric disease, it has been shown that certain high-risk activities are associated with conduct disorder, including risky games, extreme sports, dangerous driving, bulimia, sexual risk-taking, harassing behaviour, and violent suicidal behaviour. These high-risk behaviour patterns reveal impaired self-control and behavioural inhibition mechanisms, as well as some degree of difficulty or even inability to foresee possible adverse consequences of high-risk behaviour on the part of children and adolescents suffering from conduct disorder.

Studying populations of children to define how the symptoms of conduct disorder develop

To identify factors that predispose to and protect against conduct disorder, it is important to understand how the problem develops. This entails formulating working definitions of the relevant symptoms at different ages and following cohorts longitudinally from birth to the age of 18 in representative samples. Few studies to date have adopted this approach.

Conduct disorder must be considered from a developmental point of view. The developmental aspect is highlighted by the fact that it is a prerequisite to the diagnosis of antisocial personality disorder in adulthood. Moreover, DSM-IV-TR specifies two different types of conduct disorder: that beginning before 10 years of age and that beginning after.

Individuals who develop early-onset conduct disorder (before 10) are usually physically aggressive throughout childhood, and boys are more likely to develop a personality disorder in adulthood. Those who develop conduct disorder later (after 10) are not particularly aggressive and are at low risk of antisocial personality disorder. Thus, the prognoses associated with the two different types are quite different. This underlines the importance of studying how the various symptoms of conduct disorder develop.

Aggression against people and animals includes physical aggression, such as "fighting", "physical cruelty", and "bullying, threatening or intimidating others". Most children begin to commit acts of physical aggression at a very young age with the frequency of such acts increasing around the age of 4. Subsequently, this type of behaviour becomes rarer in most children, but in a minority (3–11%), the level of aggression is sustained into adolescence; this minority includes those children who are identified as suffering from childhood-onset conduct disorder. In these children, the frequency of acts of physical aggression tends to rise again in adolescence, although not back up to the rate seen in young childhood; however, the consequences for the victims are more serious because of the aggressor's greater strength and the possible use of weapons. At the beginning of adulthood, the frequency of acts of physical aggression seems to drop sharply, even in those with a chronic history of conduct disorder with physical aggression. The available data suggest that it is extremely rare that a child who has never manifested any evidence of physically aggressive behaviour before the age of 10 will exhibit such behaviour at a later stage.

Behaviour patterns involving the "destruction of property" have not been studied longitudinally to the same extent as physical aggression. Such acts begin in early childhood and are observed in the majority of children. The frequency of such acts also seems to drop with age although a relatively high level of this type of behaviour is sustained in a minority of children (including those with conduct disorder). With increasing age, the property destroyed tends to be of increasing monetary value and usefulness to the community (e.g., serious vandalism such as the burning of motor cars or school premises).

Theft, either with or without confrontation of the victim, begins in young childhood, and lying is also common in young children. How stealing, lying, and deceitfulness develop in relation to the diagnosis of conduct disorder remains poorly understood. It is likely that those who are identified as suffering from early-onset conduct disorder often lie or steal. Deceitfulness is probably a type of behaviour that appears later in development, although this of course depends on the definition used.

Symptoms corresponding to "serious violations of rules", such as "staying out at night before the age of 13", "running away and staying away from home overnight" and "playing truant from school before the age of 13", can begin in preadolescent children with childhood-onset conduct disorder. In these children, such serious violations of established rules represent an extension of a disorder that first appeared in young childhood. For the others, such violations might be symptoms of adolescent-onset conduct disorder or the expression of other adaptive problems.

Co-morbidity studies show that a diagnosis of conduct disorder is commonly associated with a diagnosis of ODD or ADHD. It has often been suggested that hyperactivity and oppositional problems induce conduct disorder. Longitudinal studies following young people from the beginning of primary school up to the end of adolescence show that early physical aggression is more predictive of violence in adolescence than are either hyperactivity or oppositional behaviour. These findings, together with the association between hyperactivity, physical aggression, and oppositional behaviour at the age of 18 months, suggest that these types of problems often appear simultaneously and interact with one another from early childhood through adolescence. As a result, hyperactivity might be a less reliable predictor of violence during adolescence than physical aggression in young childhood. Physical aggression, antisocial behaviour, and delinquency are not often associated with anxiety, whereas reactive aggression is.

Given that chronic physical aggression begins in early childhood and persists into adulthood in the most serious cases, it is not surprising to see that similar risk factors for the trait are observed in young children, older children, and adolescents. One of the risk factors is the child's gender: although small boys tend to commit more acts of physical aggression than girls, the difference is actually small until primary school, by which time the difference has become more significant; by adolescence, there is a substantial difference between the sexes with respect to rates of arrest for acts of physical violence.

The risk factors for high levels of physical aggression resemble those for conduct disorder, antisocial behaviour, and delinquency—whether evaluated at a single time point or at a series of time points. The main factors are a parental history of antisocial behaviour during adolescence, the mother being young at the birth of her first child, the mother's educational level being low, the mother being a smoker, family conflict, poverty, and coercive behaviour toward the child by the parents.

In the absence of data on the developmental pathways of behaviour patterns associated with "deceitfulness, lying, stealing, destruction of property and serious violations of the rules", the relevant risk factors cannot be identified. However, it seems likely that the majority of risk factors for physical aggression will be good predictors for other symptoms of conduct disorder because pathologically aggressive children tend to exhibit the other symptoms. Nevertheless, some children who are diagnosed with conduct disorder do not show a high level of physical aggression, e.g., some are thus diagnosed because they steal, lie, commit fraud, or violate rules without any aggressive component. The risk factors in these cases are as yet unknown, and they may differ from those implicated in children with a high level of physical aggression. Proportionally speaking, more girls than boys are diagnosed as having conduct disorder with no manifestation of physical aggression.

Longitudinal studies that have addressed developmental pathways during the primary school years with a view to predicting medium- and long-term social adaptation show that, in boys, the path toward chronic physical aggression is a better predictor of delinquency—both with and without violence—in late adolescence than it is in girls. These studies also show that children who develop multiple symptoms of conduct disorder are at high risk of failing to adapt socially, possibly leading to academic failure, rejection by their peers, precocious sexual activity, promiscuity, smoking, alcohol or substance abuse, membership in delinquent gangs, depression, suicidal thoughts, child pregnancy, unemployment, and poor physical health.

To our knowledge, there have not been any studies of pathways to physical aggression as a risk factor for antisocial personality disorder in adulthood, although most children who exhibit a high level of physical aggression continue to do so through adolescence and also present most of the other symptoms of conduct disorder. However, from the available information, it seems fairly clear that the older the individual, the less they resort to the types of behaviour that correspond to DSM-IV criteria for conduct disorder or antisocial personality disorder.

Genetic determinance of conduct disorder is around 50 %

As with any other multifactorial disease, the question of genetic determinism revolves around the idea of predisposition. Genetic factors increase the risk, modify how the disorder is expressed, and have to be appreciated in the context of a set of dynamically interacting etiologic factors. Epidemiological data give indications of the relative weight of genetic and environmental factors that specifically predispose to conduct disorder, ODD or ADHD. The same studies have yielded clues as to genetic factors that are common to all three problems.

The first clue that points to a genetic component in these disorders is the fact that relatives of an affected subject are more likely to suffer from the same problem. Family studies find a strong clustering of ADHD in families. Although there are no specific studies on conduct disorder, being related to a hyperactive child is known to increase the risk of conduct disorder and ODD. This clustering in families indicates a common etiologic basis. It actually corresponds to cosegregation of ADHD and conduct disorder. In practice, only those forms of ADHD that are inter-current with conduct disorder increase the familial risk of the latter—and in the relatives, the conduct disorder is itself often associated with ADHD. This cosegregation of conduct disorder and ADHD suggests that there is a familial form of ADHD that is associated with antisocial behaviour; at least some of the etiologic mechanisms of this form would be distinct from those involved in isolated forms of ADHD. This familial impact on the development of conduct disorder cannot be extrapolated to independent forms of ADHD and, moreover, familial clustering does not necessarily imply that genetic factors are involved, which can only be confirmed by quantifying the heritability of the condition.

The genetic component corresponds to the part of the phenotypic variance that can be attributed to genetic factors. This is equivalent to the percentage of the disease that can be accounted for in terms of genomic differences between different individuals, estimated using methods designed to separate the influence of heredity from that of the environment, notably adoption and twin studies.

The type of adoption study that has yielded the most information on paediatric problems is based on comparisons of unrelated siblings who have been adopted with brothers and sisters who have been adopted by the same family. Such studies have resulted in estimates of 40–70% genetic heritability in the case of ADHD and 55% for externalizing problems.

A similar degree of genetic heritability has been found in twin studies, of the order of 50–60% for externalizing symptoms. Such studies also suggest reasons for this average weighting of genetic factors as well as the extreme diversity of the results. In the first place, permanent, non-contextual disruptive behaviour patterns are the most heritable (h²=80%). Secondly, the various different diagnostic categories that cover disruptive behaviour do not share the same degree of genetic heritability: ADHD is strongly heritable (70–80%) with a greater genetic component than in the dimension of motor hyperactivity or in comorbid forms. The minor contribution of environmental factors observed in twin studies does not mean that such factors are not important in the etiology of ADHD because the heritability component includes interactions between genetic factors and the environment. Conduct disorder appears to be less heritable than ADHD, about 50%. This seems to be independent of the child's gender but may vary with age, heritability being higher for antisocial behaviour that begins at school age than for that which begins either earlier or in adulthood, although these findings remain to be confirmed in longitudinal studies. Finally, among the symptoms of conduct disorder, physical aggression is the most heritable (h²=60–70%), with other symptoms mainly depending on shared environmental factors.

Twin studies also reveal the relative genetic and environmental components in etiologic factors common to the different problems. The phenotypic covariance between exteriorised and internalizing problems is predominantly environmental. In contrast, conduct disorder, ODD, and hyperactivity share genetic susceptibility, with up to 87% of the covariance attributable to genetic factors. In addition to these common global factors, there are on the one hand, supplementary genetic contributions in ADHD and in the temperamental trait of the seeking of novel experiences, and on the other hand, specific environmental components in conduct disorder and substance abuse.

These consistent findings justify looking for vulnerability genes. The functional approach involves selecting candidate genes and testing their links with the various disorders. Many such molecular genetics studies have been undertaken, giving highly disparate results. Of the mono-amine-based neurotransmitter systems, attention has focused on the dopamine system, which is known to be involved in motor excitability and concentration. Also, psychostimulants that target dopaminergic neurotransmission have been used to manage ADHD. Recent meta-analyses show a link between ADHD and the genes that encode dopamine receptors D4 and D5, although the correlations are weak and do not by any means fully account for the epidemiological picture with the presence of each of the alleles associated with ADHD increasing the risk by a factor of just 1.25. However, most of these studies have been category-based and have not been conducted on enough subjects to yield information about comorbid forms, although these are the most heritable. A few studies have pinpointed a specific link between the gene for the D4 dopamine receptor and intercurrent ADHD/conduct disorder, and between the D5 receptor and inter-current ODD/substance abuse. Finally, the most recent studies have sought to define cognitive profiles and drug response patterns in children with ADHD in order to identify phenotypes that are particularly tightly linked to certain candidate genes.

Figure

Meta-analyses of studies addressing the link between ADHD and intra-familial polymorphisms in candidate genes (Odds Ratios and 95% confidence intervals) .

An alternative way of identifying genes that predispose to conduct disorder involves the holistic investigation of multiple risk factors. Genetic factors are all the more difficult to define because their frequency varies in parallel to the variability of environmental factors. Exposure to a particular type of environment might exacerbate the situation in a genetically predisposed child to a greater extent than would result from the simple addition of risk factors. Moreover, certain genetic or environmental factors might be protective. There are many reasons for studying interactions between genes and the environment. Firstly, that such interactions do indeed occur is illustrated by the fact that the same genetic factors are associated with conduct disorder, ADHD, and ODD, whereas the relevant environmental factors are specific to each of these conditions. Secondly, adoption studies based on a parent-child model show that physiological susceptibility can be revealed by life events and interacts with the parents' child-raising practices: the child-raising strategies of adoptive parents tend to be less successful with children from a biological family with a history of antisocial behaviour. Such studies also point to a link between disruptive behaviour and biological history emerging only if the child is removed from the adoptive home. It is therefore logical that interactions between genes and environment be taken into account when we are looking at the genetics of conduct disorder—and such approaches are yielding valuable information. The allele that is associated with reduced monoamine oxidase A activity appears to be associated with the development of antisocial behaviour, but only in subjects who have been mistreated during childhood. This does not correspond to a simple superimposition of risk factors but rather a synergistic effect between different vulnerability factors. Taking interactions between life events and genes into account could help shed light on why the results of different studies are so divergent and could reveal the type of susceptibility conferred by genes in multifactorial problems such as disruptive behaviour, making it a potentially useful approach to the genetics of behaviour.

Temperament and personality may be factors that predispose to conduct disorder

For many years, the impact of personality and temperament were neglected in favour of sociological explanations for conduct disorder and delinquency. However, many studies have shown that personal variables such as temperament and personality are key factors in the development, maintenance, and severity of conduct disorder.

In children, the existence of a set of traits that collectively constitute a difficult temperament (e.g., negative mood, lack of perseverance, inability to adapt to situations, readiness to be distracted, intense emotional reactions, hyperactivity and social withdrawal) is strongly predictive of impaired psychosocial adaptation in adolescence and adulthood. A difficult temperament is especially predictive of conduct disorder if the child comes from a dysfunctional family. However, it is not specific to conduct disorder because it is also found in ADHD and internalised problems such as anxiety and depression. In fact, a difficult temperament seems to be a common precursor for mental problems in general.

On the other hand, of the full range of temperament-related behavioural characteristics that can be ascertained at an early age, aggression against others, weak emotional self-control, and disobedience are most predictive of adolescent conduct disorder, independently of both the family environment and the nature of the child's peer group. Impulsiveness is also predictive of this type of problem, as well as delinquency insofar as this paediatric temperamental dimension may inhibit socialisation processes, e.g. by entailing difficulties with relationships or poor social adaptation. Although impulsiveness is far from specific to conduct disorder (it is also associated with ADHD and ODD), it is particularly marked in the more severe forms (e.g., among prison inmates).

Impaired behavioural control has been shown to be involved in conduct disorder and antisocial behaviour in both boys and girls. But this type of temperamental trait is involved in a non-specific way and is broadly predictive of exteriorised problems such as ADHD. However, it seems that it is mostly the coupling of a strong need for behavioural activation with weak behavioural inhibition that is predictive of conduct disorder. The involvement of both these systems (activation/inhibition) could promote the expression of certain personality dimensions (e.g., psychoticism).

In terms of emotional dimensions, it has been shown that a lack of guilt feelings in a child is strongly predictive of physical aggression, delinquency, and conduct disorder. The individuals concerned are less empathetic, less able to recognise emotions such as anger and sadness in others, and have a lower level of affective morality. Self-esteem does not seem to be a significant risk factor for conduct disorder insofar as it is not specific to a type of trait. On the other hand, egocentrism (the tendency to relate everything to oneself) and narcissistic personality could be typical of certain forms of conduct disorder with physical aggression.

Some experts believe that all these emotional characteristics express a personality trait defined by affective coldness, insensitivity, and a tendency to charm. Adolescents with conduct disorder coupled with this personality trait tend to be relatively aggressive and violent and exhibit more symptoms of ODD (92% of cases) and ADHD (48%). This trait associated with conduct disorder might correspond to a specific group which constitutes a predictor of psychopathy in adulthood, especially if the subject exhibits high impulsiveness.

The profile—combining both temperament and character—that would be specifically characteristic of conduct disorder would be one of strong "novelty seeking" (exploratory excitability, impulsiveness, extravagance) coupled with relatively low levels of "pain avoidance" dimensions (i.e., low levels of fatigability, timidity, fear of the unknown, and anticipatory anxiety), "dependence on compensation" (lack of empathy, little sensitivity to social reinforcement), and "determination" (low level of individual maturity and sense of responsibility).

Psychoticism (tough-mindedness, interpersonal hostility, self-centredness, affective coldness) seems to be predictive of conduct disorder and antisocial behaviour in both boys and girls insofar as this dimension reflects a tendency toward antisocial attitudes and impulsive behaviour. Extraversion and neuroticism seem to be particularly important in delinquent girls and less serious delinquent behaviour, whereas people who commit extremely violent acts are usually relatively introverted.

A longitudinal study has shown that character traits such as affective coldness, a tendency to manipulate, cynicism, and aggression could significantly influence the age at which aggressive behaviour is manifested, the stability of conduct disorder through adolescence, and antisocial personality disorder in adulthood.

A number of studies have set out to analyse links that might exist between the temperament of the child and parental attitudes. To a greater or lesser extent, all findings point to the centrality of the "goodness of fit" phenomenon between parents and child in the development of conduct disorder. According to this hypothesis, conduct disorder is the consequence of incompatibility between the child's temperament (e.g., one who is impulsive without great inhibitory capacities) and the demands of those close to him (e.g., child-raising style). Thus, a boy with a "resistant" temperament (who has trouble concentrating, a tendency to oppose, and little capacity for inhibition) would only be susceptible to conduct disorder if his parents adopt a broadly "permissive" child-raising style, and the outcome would be quite different if they were to exert adequate control. In addition, according to some studies, mother-child interactions that are typified by reciprocal anger and poor emotional adjustment (laughing/anger) or, in contrast, emotional insensitivity could lead to conduct disorder but only in boys.

The involvement of personality and temperamental traits in the occurrence and maintenance of conduct disorder should be situated in a dynamic developmental perspective in which physiological and environmental factors interact with one another.

Perinatal events might be involved in the pathogenesis of conduct disorder

Various antenatal and perinatal events have been proposed as possibly exacerbating the risk of conduct disorder. Nevertheless, it seems unlikely that such events would have any very specific impact; rather, interactions with other risk factors (notably genetic) would determine the form of the disorder.

It is generally accepted that the mother's life style, especially if psychoactive substances were used during a critical period of the pregnancy, can affect cerebral development in the foetus with long-term repercussions upon the child's neuro-behavioural future. In this context, a large number of studies have consistently shown that the offspring of mothers who smoke are more likely to develop conduct disorder; a significant correlation has emerged in several longitudinal studies between maternal smoking and the occurrence of conduct disorder, especially in boys.

With respect to drinking, the worst manifestations are seen when the mother's alcohol consumption during pregnancy is very high, leading to a set of physical and neurological problems collectively referred to as foetal alcohol syndrome. However, a number of longitudinal studies have suggested that prenatal exposure to even moderate levels of alcohol can have adverse effects on a child's cognitive and behavioural development.

In longitudinal studies, a correlation has been observed between cannabis smoking during pregnancy and behavioural problems in the child, notably impulsiveness and impaired concentration. An increased level of exteriorised problems has also been reported, especially in boys whose mothers took cocaine during pregnancy: these boys scored two times higher in tests for aggressive behaviour and delinquency than did those in a control group.

Premature birth and low birth weight have both been suggested as possible risk factors for conduct disorder, and a relationship has been identified between health problems in premature babies and the incidence of exteriorised problems at the age of 5, notably oppositional behaviour and hyperactivity. Similarly, low birth weight correlates with ADHD and antisocial behaviour.

Complications during delivery can also lead to neurological deficit which itself is associated with neuro-behavioural problems. Recent studies conducted on large populations have shown that there is a correlation between obstetric complications and antisocial behaviour in childhood or adolescence. Among the various problems that may arise during delivery, experts have singled out foetal asphyxia as a specific risk factor for conduct disorder.

There is now clear evidence of an association between the mother's age at pregnancy and the likelihood of conduct disorder in the child. In a study conducted at a special centre, it was shown that pregnancy at a young age (before 20 and to an even greater extent before 18) directly correlated with the number of symptoms of conduct disorder in sons between the ages of 6 and 13. If the mother had a history of conduct disorder, the likelihood of her becoming pregnant at a young age was higher, as was the incidence of the diagnosis of conduct disorder in her children.

In addition, it emerged that head injury, even mild, during early childhood may be associated with an increased rate of ADHD and conduct disorder between 10 and 13 years of age.

Other risk factors, such as maternal smoking and complications during delivery have been identified in methodological robust studies however they seem to be relatively non-specific, and the correlations with pathology in later life could involve other parameters, such as disturbance of the mother-child relationship or parameters related to the family environment in general.

The pathogenesis and persistence of conduct disorder are affected by familial and environmental factors

Various contextual factors, essentially related to family or psychosocial environment, are associated with conduct disorder in adolescence and adulthood. Adverse environmental conditions are all the more important in that they are commonly systematic, long term, and associated with other risk factors. Poverty in itself is not a risk factor, but extreme poverty is associated with multiple risk factors, which have impact on parenting behavior.

Quite a lot of studies have addressed the impact of parental problems, be it behavioural problems in the father or mother, dysfunction in the couple, maternal post-natal depression, parental alcoholism, or other forms of dependence. These studies have consistently identified a link between conduct disorder in the offspring and antisocial personality disorder in one of the parents. The children of fathers who are substance dependent, particularly to alcohol, and also have an antisocial personality are at higher risk of both conduct disorder and ADHD; and the sons are at greater risk of themselves becoming addicted. Similarly, if the mother has an antisocial personality, the likelihood of the children developing conduct disorder is enhanced, a fact that was underestimated for a long time.

Maternal depression is known to have an effect on the development of the child although the magnitude of such effects in conduct disorder remains controversial. However, it can be readily appreciated that depression-related disturbances of the relationship between mother and baby could have adverse long-term consequences on the ability of the child to concentrate and control his/her emotions. The children of mothers who experience post-natal depression may constitute a specific high-risk group.

Adolescent pregnancy is associated with conduct disorder in the offspring. The risk of conduct disorder with aggressive behaviour is particularly high among boys whose mothers were very young and whose educational level is low.

In several longitudinal studies of a high risk population, insecure attachment and disturbances or disruption of relationships within the family significantly exacerbated the risk of behavioural problems among children. Insecure attachment between child and both parents seemed to increase the risk of conduct disorder substantially, specifically if associated with temperamental difficulties, adverse family events, and disturbances in parent-children relationships. A child whose relationship to principal attachment figures has been disrupted at a very young age tends to exhibit aggressive and directive behaviour toward their parents, particularly if the child has been subjected to violence or sexual abuse. Disorders of attachment, compounded with other dimensions (e.g., temperament, unsuccessful parental strategies, a dysfunctional family environment, etc.) exacerbate the future risk of conduct disorder. Attachment could be involved in gene/environment interactions as a resilience factor when the attachment is secure and as a risk factor when it is insecure, especially if attachment is disorganized. Many experts have drawn attention to a link between disruption of the family structure, through divorce and subsequent aggressive behaviour, conduct disorder, and delinquency. However, it has been shown in longitudinal studies that this factor acts long before the actual separation of the parents, and it may be more closely related to the conflict between the parents rather than to the divorce itself. Thus, the long-term impact of a divorce on subsequent problems in the child may be more due to conjugal discord than to the eventual separation. Parental conflict disrupts children's capacities to regulate their emotions and, as a result, inhibits their ability to learn how to control aggressive behaviour. If one of the parents has an antisocial personality or is substance dependent, or if there is conjugal violence, the separation may help protect the child.

Various studies have shown that deleterious parental attitudes and inappropriate child-raising strategies emerge as a family characteristic linked to delinquency. Parenting style seems to be more important when it comes to predicting aggressive and oppositional behaviour patterns than for ADHD. Results agree that a failure on the part of the parents to provide adequate control of their child's behaviour is primordial, as is also inconsistency or overly strict discipline. Parental indifference to truancy from school or to the absence of the child from the family home is predictive of antisocial behaviour and delinquency, particularly in underprivileged settings.

The influence of siblings cannot be ignored. Various studies have shown that a child with a brother or sister (especially if he or she is older) who has been convicted for an act of delinquency is significantly more likely to come before the courts. It may be that, in a family in which there is already a child with conduct disorder, his or her siblings may be directly influenced, independently of any other risk factors. This association is significantly amplified in underprivileged settings. Similarly, the fact of having older brothers or sisters who perform well at school may be protective.

Many experts agree that the type of aggressive and violent behaviour patterns that are common in underprivileged settings depend on the concurrence of diverse risk factors. Concentrations of gangs of children and adolescents with conduct disorder in certain neighbourhoods tend to attract other young people from the same neighbourhood, and mixing with delinquents doubles the risk of aggressive or delinquent behaviour patterns being sustained through adolescence. Young misfits reinforce each others' antisocial behaviour patterns, sometimes in response to rejection by their better-adjusted peers.

The available data suggest that schooling experiences need to be taken into account when evaluating risk factors for conduct disorder—truancy, rudeness at school, and academic failure have all been related to conduct disorder and delinquency. When children are presenting symptoms of both ADHD and early-onset conduct disorder, they are very likely to fail at school.

Numerous studies have revealed a significant link between exposure to media violence and aggressive behaviour in young people. Recent studies have confirmed that 8-year-old children who watch violent shows on television are far more likely to act aggressively in the long term (11 to 22 years later). This correlation is independent of intelligence quotient (IQ) and socio-economic status. Televised violence may not only lead to real violence but it also leads to desensitisation, entailing trivialisation and habituation, resulting in a passive reaction to and tolerance of violent acts. In certain children, playing violent games enhances physiological excitation, exacerbates aggressive attitudes, and inhibits positive social behaviour patterns. The inundation of children—who might already be vulnerable by virtue of their family or social environment, or who might already be presenting signs of nascent conduct disorder—with violent stimuli such as are so ubiquitously and continuously dispensed via television and video games, enhances the attraction of violence, all the more so because the violent behaviour is at best trivialised or exempted of any feeling of guilt, and at worst glamorised and encouraged. It is nevertheless important to remember that it is children who already have problems with real violence who are the most affected by exposure to virtual violence in the media.

Neurocognitive deficits are associed with conduct disorder

Two types of neurocognitive deficit seem to be associed with conduct disorder in children and adolescents, namely impaired verbal skills and impaired executive inhibition mechanisms.

It has long been recognised that children and adolescents with conduct disorder perform poorly on the verbal components of psychometric tests compared with their performance as a whole. Adolescents with severe conduct disorder, particularly characterised by violence, tend to have a lower Verbal Intelligence Quotient (VIQ) than those without any significant conduct problem. Furthermore, poor verbal skills represent one of the most important risk factors for delinquent behaviour in adulthood. The fact that this deficit persists over time from childhood through adulthood is probably evidence that language functions play a key role in the development of so-called disruptive behaviour patterns. For example, retrospective analysis often identifies delayed language acquisition in children with ADHD. These children also exhibit many verbal difficulties, including poor comprehension, limited vocabulary, and low verbal fluency. Similarly, particularly aggressive children tend to have greater difficulty with both oral expression and written language than their non-aggressive peers. Both of these types of difficulty are also seen in delinquent adolescents and are good predictors of criminal behaviour in adulthood.

Three explanations could account for the functional importance of verbal deficit in the cognitive mechanisms that underlie conduct disorder. The first concerns the regulatory function of language in human behavioural control processes. It is now generally accepted that the steady improvement in self-control that is seen in the course of development is associated with parallel improvement in children's capacities for understanding and verbal expression. The progressive internalisation of language during development, in the form of an "internal dialogue", allows the child not only to plan and guide his behaviour more effectively but also, and more importantly, to acquire an understanding of what he should do and what he should not do according to aim and context—before the action is executed. Verbal deficits can inhibit the development of the child's or adolescent's ability to formulate symbolic and abstract representations, resulting in a lesser understanding of situations in which social understanding is called for. The idea of "rules" is also integrated into an individual's psychological functioning in the form of mental representations constructed around language-based concepts such as "You can do this but not that". Adequate language development is therefore essential to establishing prosocial behaviour patterns, and this necessitates anticipating the consequences of actions and referring to accepted norms of conduct in a given socio-cultural environment.

The second explanation pertains to the capacity of language to express the subject's emotions in an accurate fashion and also to decode the emotions being expressed by others (i.e. empathy). Verbal deficit can therefore restrict a child's ability to express his own emotional experiences as well as his ability to process and express the emotions felt by those around him. A low level of empathy has been documented in children and adolescents with conduct disorder: any such deficit will compromise the quality of communication and trigger defensive or aggressive reactions in the child.

The third explanation emphasises the high correlation known to exist between verbal skills and academic abilities. Impaired verbal skills are associated with poor academic performance, and the subsequent school career of children arriving at kindergarten with poor verbal skills is characterised by difficulty in learning to read and learning disorders in general.

In parallel to poor verbal skills, many empirical observations also suggest that children and adolescents with conduct disorder have impaired executive functions. The idea of executive function covers a whole set of complex, disparate cognitive processes that are necessary for the performance of goal-driven tasks, including concentration, the ability to select and refine a course of action, the inhibition of automatic responses, the monitoring of an action under way, and checking the efficacy of the response in view of the goal envisaged and the context. In accordance with this definition, impaired executive skills in a child or adolescent would compromise his capacities for analysis, for abstract reasoning, for formulating appropriate responses in the context of the relevant demands, and motor control (impulsiveness).

A number of studies have suggested that children and adolescents with conduct disorder exhibit poor executive control, particularly those with a tendency to violent behaviour. The correlation between executive deficit and conduct disorder is stronger if the latter begins at a relatively early stage of development. Executive deficit could also account for the severity of the conduct disorder and its persistence into adulthood. However, many experts minimise the strength of the link between executive dysfunction and conduct disorder because ADHD was not taken into account in most of the relevant studies. Further studies are needed in which executive performance is evaluated in children and adolescents with isolated conduct disorder, conduct disorder plus ADHD, and controls.

These two types of cognitive deficit, verbal and executive, are as common in boys as in girls, but in the opinion of some experts, their impact may not be identical in the two sexes in the course of development, as a result in particular of the modulation of their effects by environmental factors that may be gender dependent. There are presently, however, insufficient data for age- and gender-based comparison, so it is difficult to draw any conclusions on this point.

In terms of brain anatomy, poor verbal skills in children and adolescents with conduct disorder are probably due to impaired function in the temporal regions of the left hemisphere. Executive deficit is associated with the frontal lobes, and more broadly speaking with fronto-striato-thalamic circuits. However, precise morphological magnetic resonance imaging (MRI) data or functional magnetic resonance imaging (fMRI) data in children and adolescents with conduct disorder are presently lacking. Recording of cerebral activity (e.g., by evoked potential) should also shed light in coming years on the neuro-physiological processes underlying conduct disorder. By way of illustration, some interesting experimental results have shown a significant reduction in the amplitude of P300 waves in adolescents with conduct disorder. The degree of attenuation is closely related to the severity of the disorder and appears to be more specifically related to cases in which conduct disorder is associated with rule violation. Before the age of 16, this electrophysiological abnormality is focused in the posterior regions of the cerebral cortex but later shifts to the anterior regions. It is observed in both affected boys and girls and might possibly reflect delayed maturation of the frontal regions responsible for executive functions, as opposed to a frank defect. Although such perturbation of the P300 wave is not specific to conduct disorder (being also observed in ADHD and schizophrenia), it supports the idea that neurodevelopmental factors are involved in the pathogenesis of the condition in children and adolescents.

A diagnosis of conduct disorder requires rigorous multimodal evaluation

Before a child or adolescent can be diagnosed with conduct disorder, the diverse nature of the symptoms and the frequency of inter-current problems require that a multimodal investigation be undertaken by a multidisciplinary team based on diverse sources of information, and taking into account the history of the family and its functioning.

Categorical and dimensional measures have been developed in English to review the patient's history and to identify and grade the clinical manifestations of conduct disorder. The value and relevance of these instruments—be it a standardised interview, a behavioural scale, or a questionnaire to be completed by the patient—varies according to their psychometric validity and how ease of use in daily practice. Most of the scales designed to evaluate behaviour or aggression are available in versions for both parents and teachers. Diverse instruments should be used in parallel to ensure as global an overview of the symptoms as possible.

Results should be interpreted in the light not only of the symptoms of the child or adolescent with suspected conduct disorder but also of the parents' tendency to minimise the gravity of the problem. It is thus important to compare the information given by the child and his/her parents with information from other sources, such as teachers or peers. In the same way, it is important to take into account the child's level of cognitive development when he is asked to evaluate his own behaviour (especially his verbal understanding, which will affect his perception of the problem as well as how he expresses his experiences). In addition, given the variability over time of behavioural manifestations in developing children, it is important to perform regular assessments in a child at risk of conduct disorder, e.g., one who is exhibiting high levels of physically aggressive activity and/or showing the symptoms of ADHD.

Most of these instruments have yet to be translated into French and validated in populations of French-speaking children and adolescents—a task requiring a significant amount of work.

Preventive measures will depend on vulnerability factors

A distinction may be made between three classes of preventive measure dependent on the degree of vulnerability of the subjects being targeted. "Universal" preventive measures concern the population as a whole or those in the population who present no particular risk of conduct disorder. "Selective" preventive measures target individuals or groups who are at high risk by virtue of environmental, social, or family factors: in the case of conduct disorder the targets might be people living in underprivileged urban neighbourhoods, poor children and parents, the children of young, single mothers, children growing up in a conflict-ridden family environment, and the children of alcoholics or drug addicts. "Indicated" preventive measures target individuals or groups with personal risk factors or who are already exhibiting early symptoms of conduct disorder. Indicated prevention programmes may target young people with learning disorders, children manifesting impulsive and/or aggressive behaviour, and children who have already been diagnosed with ADHD or ODD.

There are few specific programmes aimed at conduct disorder. The majority of programmes are designed to combat delinquency, aggressive behaviour, or violence in general, and neither DSM-IV nor ICD criteria are ever used in evaluations. The efficacy of programmes is judged on the basis of measurements of aggressive and/or antisocial behaviour patterns derived from standardised questionnaires and official reports from the judicial system or the police.

Hundreds of programmes designed to combat violence are already in use, especially in Britain and America, although only about 20 have been validated. For a programme to be validated and qualified as evidence based, certain criteria must be fulfilled: intervention modalities must be based on a reference theory; the intervention programme must be stipulated in a protocol, and there should be monitoring of subject adherence to the programme; interventions must be demonstrated to have been effective in an empirical context (a rigorous, quasi-experimental plan, significant positive effects, results duplicated in at least two different sites, available scientific data, etc.).

Validated conduct disorder prevention programmes require special intervention modalities. These modalities can be classified according to the age of the subject (very young children, children, and adolescents) and the target (affected subject, parents, teachers or the subject's environment). One or more preventive measures may be planned in a given programme, which is thus classified as either unimodal or multimodal. Such measures are integrated into a preventive programme that might be universal, selective, or indicated; although in the prevention of conduct disorder, a distinction is rarely made between the last two, the same programmes targeting both groups.

Validated preventive measures for small children (0–3 years of age) mainly focus on the parents and on the children's capacities to control themselves and their own behaviour. This involves home visits and support to the parents in the form of health education and counselling on parenting skills. At crèches and kindergartens, the emphasis is on developing social, cognitive, and emotional skills. The Elmira Home Visitation Study and the Perry Preschool/Preschool Curriculum Comparison Study have proven effective in this field. Both are selective/indicated programmes, the first aimed at children up to 2 years of age, and the second focusing on 3–4-year-olds.

Other validated methods are integrated into preventive programmes aimed at children and adolescents between 3 and 16 years of age. Interventions focused on these young people are designed to develop their social, cognitive and emotional skills through such modalities as conflict-resolution, anger-management, emotional understanding, and the fostering of prosocial attitudes and reasoning ability. Most of the validated programmes are based on this type of intervention strategy, whether combined with another type of modality or not. Some of these programmes (Linking the Interests of Families and Teachers, Seattle Social Development Project, Montreal Prevention Experiment, and Fast Track) include interventions aimed at the parents as well as their offspring. In most cases, this involves training to help develop emotional control, the establishment of a positive disciplinary routine, and improved communication. Training programmes can also be formulated for teachers, and programmes like the Seattle Social Development Project and The Incredible Years Series are designed to promote co-operative teaching, proactive management, positive reinforcement, and conflict resolution, among other goals.

A few preventive programmes have focused on the environment. Two such methods have been validated, one based on interventions at school (e.g., the School Transitional Environmental Project and the Olweus Bullying Programme) and the other on a mentoring system (the Big Brother/Big Sister of America programme).

Apart from these validated methods, scientific data built up over the last 20 years in this field make it possible to identify the characteristics of successful preventive programmes. Interventions which have been proven to be effective have integrated interactions with experienced counsellors that go well beyond the simple transmission of information. The successful programmes also tend to be long term and intensive, running for many years with an annual investment of at least 20 hours. Programmes involving different methods and targeting several different factors (the young person, parents, teachers, etc.) have proven particularly effective. The same is true of those that match the intervention method to the target population. The extent of presence on the ground is also crucial: the idea of an "effective programme" only has meaning if it is closely associated with tight control of the intervention processes. Finally, the latest information about the development of aggression and associated factors highlights the importance of the early (post-natal and pre-school) implementation of selective and indicated interventions focusing on promoting the mother's parenting skills.

In conclusion, it seems that at the international level, many different measures have been developed for the assessment of conduct disorder, and numerous validated preventive programmes are currently in use. However, the situation in France remains problematic. No results from any French programme designed to prevent either conduct disorder or violence in the broader sense have been published in the scientific literature to date. It might be that interventions are currently under way in France in the framework of the prevention of violent behaviour, but if this is so, they remain isolated, one-off instances, the results of which have not been published. Moreover, such studies are generally not scientifically validated, and outcomes are not rigorously evaluated—evaluation in France seems to be confined to descriptive parameters, which give no real information about the effects of a given measure or its success.

Psychotherapy incorporates various treatment approaches

The management of conduct disorder should involve different, complementary treatment approaches whenever possible. So-called psychosocial management strategies include interventions focusing on the parents, aimed at the child him- or herself and, in some cases, his/her teachers. The global goals of management can be summarized under four objectives:

• establishing a system to provide back-up and social support for the entire family
• promoting the child's contacts with "prosocial" peers
• reducing the child's contacts with "antisocial" peers
• stepping up support in school and contacts with teachers

The treatment approaches that have been shown to be the most successful in controlled studies in Great Britain and America have been based on providing support for the whole family of the young person with the conduct disorder. The Functional Family Therapy (FFT) programme is supervised by trained, experienced therapists and involves three phases spread out over a dozen sessions over a period of 3 months, often conducted at home in the most difficult cases. The purpose of the "commitment and motivation" phase is to establish a therapeutic bond with every family member by minimising distancing factors (cultural and/or social). In this phase, availability and respect are essential to fostering confidence and promoting collaboration. The second "parent training" phase aims to modify how family members interact with one another by facilitating communication and working on ways of resolving conflicts. The third "generalisation" phase aims at extending or broadening any positive interactional changes that have occurred in the family to their relationships with other community systems, including mental health and legal authorities. Taken together, the results show that this programme has significantly reduced the rate of relapse of serious antisocial behaviour patterns in young people with conduct disorder. Even among those with persistent antisocial behaviour, the programme reduces the severity of the acts. Improvement was sustained, with results stable over 5 years.

Books (e.g. "Living with Children") and video tape recordings can help to reinforce the parents' daily efforts and therefore increase confidence and compliance with the programme. Success rates are enhanced if the video aid is used in a discussion group supervised by a therapist. Daily, structured exercises help parents generalise and apply at home the principles and methods that they have learned at the therapeutic unit, relating to positive reinforcement of the child's behaviour, the setting of limits, a non-violent disciplinary approach, the respect of rules, conflict resolution…

Although the programme is aimed at the child, it is designed to foster empathy, anger management, friendship, communication, and positive relationships with the school and teachers. Incorporation of a teacher-centred component into the child-centred intervention enhances the outcome in personal programmes in terms of capacities of adaptation, behaviour, and conflict resolution.

Multisystemic therapy (MST) involves a precise evaluation of the overall environment in which the young person lives and grows up (family, school, friends, neighbours, and neighbourhood) and his/her interactions with all these systems with the aim of identifying the main factors underlying the problems. The goals of management are to: try to keep the child away from unsuitable friends (who set a poor example); reinforce his/her relationships with conventional institutions such as family or school; enhance the family's ability to monitor the child and enforce discipline; and enhance the adolescent's social and academic skills. The therapists may use various methods, including family therapy and interventions at school as well as cognitive and/or behavioural techniques. The programme is conducted by a team of three or four trained therapists who are closely supervised and are available to intervene around the clock, 7 days a week. The programme lasts 3 to 5 months on average, representing a total of about 60 hours of intervention time, mostly with the young person and the family but including the social network. It is crucial that a collaborative relationship be established between parents and teachers. Multisystemic therapy has been shown in controlled studies to be effective in the long-term at reducing criminal activity, violent acts, arrests for substance abuse, assignments to special centres, and the incarceration of subjects with conduct disorder. Relapse is also substantially reduced.

Special schools and centres, often located in out-of-the-way places, have been used for decades as a social, "therapeutic" response for young delinquents and aggressive adolescents. Apart from the high cost of this type of strategy (which requires a high ratio of supervisory and security personnel), far from attenuating behavioural problems, bringing troubled young people together seems to broaden the "misfit's" repertoire of aggressive behaviour patterns. Disappointing outcomes led to a switch toward foster care programmes for young people with conduct disorder. Foster care treatment programmes use methods based on parental guidance, both for the biological family and the foster family. Such programmes are run by therapists and a full-time experienced supervisor. The supervisor contacts the foster family every day to review the subject's behaviour that day and, if necessary, adjust the treatment regimen. A set of procedures aims at improving the subject's behaviour in a sustained way. The foster families are also taught how to work together with the programme personnel and with the school. Whenever possible, they are asked to participate in family therapy sessions together with the biological (or adoptive) family because the ultimate goal is for the child to go back to his/her own home. Outcomes have been far better with this sort of programme than with placement in a special centre, with better compliance on the part of the troubled young people who run away less often and more often complete the treatment programme. Over the follow-up period, fewer days of incarceration are recorded, and the rate of relapse is reduced. As for other treatment models, changes in the way the family manages the child's behaviour (following parental training) and distancing the subject from "misfit" peers seem to be the key determinants of general improvement.

Finally, the purpose of individually focused interventions is to enhance the child's social skills, i.e. his sociability. Working in collaboration with a therapist, the child learns conflict resolution strategies through role playing and situational scenarios. In a more general way, cognitive/behavioural interventions focusing on the child have a slight-to-moderate effect on aggressive and antisocial behaviour patterns, particularly in children over 10. It seems that combining this type of treatment (which has proven successful in other problems like hyperactivity) with a parental training component helps normalise the child's behaviour in a sustainable way.

Drugs can be useful as second line strategies

Drugs may have a place in a global, priority-ranked, multimodal treatment strategy. They are usually prescribed following the failure of other strategies—apart from emergency situations (e.g., violence or aggression against others or self).

No specific drugs exist for conduct disorder, but certain drugs can moderate aggressive behaviour. Various pharmacological classes have been tested in this context, often on the basis of knowledge about the neurobiology of impulsiveness and aggression, both of which are known to involve dopamine, serotonin, and GABA pathways (among others). Most of the information on efficacy and safety was first established in adults, in most cases in the context of problems other than conduct disorder. As is the case for psychopharmacology in children and adolescents in general, there is little specific data from controlled studies in this age group. A review of the last 20 years of publications yields about 30 relevant articles, but only 10 describing studies based on rigorous methodology (with at least 500 subjects of between 6 and 18). Fewer than 10 studies report any efficacy with regard to conduct disorder.

Three major classes of drugs have been evaluated, namely antipsychotic drugs, central nervous system stimulants, and thymoregulatory drugs.

Although antipsychotic drugs are in widespread use for aggressive behaviour, only four controlled studies have been carried out focusing on conduct disorder, two with neuroleptics (haloperidol and molindone) and two with new-generation drugs (risperidone and olanzapine). The last two have a promising pharmacological profile in that they inhibit both serotonin and dopamine receptors, both systems being closely associated with aggressive behaviour. They are also better tolerated than the classic neuroleptics, especially with respect to neurological side effects such as tardive dyskinesia. All reports indicate efficacy against aggressive behaviour that is supported by the results of open studies and case histories. Classic side effects are observed, notably weight gain, drowsiness, and dystonia. Numerous medium-term studies conducted with risperidone for other indications (such as invasive developmental problems) indicate broadly satisfactory tolerance. In contrast to lithium and anticonvulsive drugs, antipsychotic drugs act immediately on the dimensions of aggression and agitation and are particularly useful in acute or emergency situations.

Unlike the antipsychotic drugs, central nervous system stimulants promote executive cognitive functions. They attenuate impulsiveness by inducing inhibition mechanisms and enhancing self-control. They mainly act by stimulating dopamine-based transmission in the brain but also have some stimulatory activity in noradrenaline and serotonin pathways. They are prescribed to treat ADHD, which is known to be the most common co-morbidity of conduct disorder. Therapeutic trials (seven controlled studies) have been conducted to evaluate the efficacy of methylphenidate against conduct disorder (with inter-current ADHD in six of the seven studies). Methylphenidate was found to have some degree of efficacy against aggression in moderately severe conduct disorder, and the treatment proved quite safe (with side effects confined to negative effects on appetite and sleep).

Seven controlled trials conducted to evaluate the efficacy of thymoregulatory drugs, notably lithium, have yielded mixed results in that in three cases, no efficacy was detected. Data on safety were more consistent with a high incidence of adverse reactions (weight gain, acne, gastrointestinal upset and polyuria), some of them serious (kidney failure and hypothyroidism). Such reactions entailed compliance problems in a population which is in any case resistant to restrictions. This was compounded by the fact that lithium's low therapeutic index means that regular blood testing is necessary to monitor the concentration of the drug in the blood throughout the course of treatment. Nevertheless, lithium is a preferred drug in conduct disorder associated with bipolar disease.

Other thymoregulatory drugs belonging to the anticonvulsive class have also been evaluated. Recently, very promising outcomes were obtained in two controlled studies of sodium valproate. Regular blood testing is required with this drug as well but not so often as lithium because of its higher therapeutic index. As with lithium, this drug is indicated in subjects with inter-current bipolar disease or who experience mood swings.

Because the serotonin system is known to be intimately involved in the pathogenesis of aggression and impulsiveness, trials have been carried out to investigate the efficacy of selective serotonin reuptake inhibitors (SSRIs) and non-selective serotonin receptor stimulants like buspirone. Only two clinical trials have been conducted in children and adolescents, but both these reported efficacy against aggression: an open study of trazodone which addressed several symptoms of conduct disorder; and a study of citalopram in aggressive children and adolescents with conduct disorder. Side effects were moderate—mainly headache, drowsiness and nightmares.

Anecdotally, small, uncontrolled trials have been reported on alpha 2-adrenergic receptor stimulants (clonidine), beta-adrenergic inhibitors, benzodiazepines, and opioid inhibitors such as naltrexone. Clonidine gave promising results in an open, pilot study, with reduced aggression correlating with decreased GABA levels in the blood after treatment.

In a confirmed case of conduct disorder, the treatment strategy should be formalised and matched to the subject, based on an accurate evaluation of the disorder itself as well as any inter-current problems.

Animal models can be of use in investigating behaviour patterns associated with conduct disorder

Of all the types of behaviour that interest ethologists and behavioural scientists, it is aggression that has generated the greatest number of publications—which is to be expected because this type of behaviour represents one of the key factors driving evolution and the establishment of territory. However, in the laboratory, choices have to be made about which parameters are to be studied because the experimenter must simplify the model to be able to make meaningful measurements reflecting the influence of any given factor.

Studies of agonistic conduct patterns in animals have revealed numerous behavioural correlations that are helping us to understand human conduct disorder, especially in the context of motor hyperactivity, attention-deficit and anxiety, all questions that have been studied in depth in laboratory animals.

Ethologists, and more recently biologists and geneticists, have traditionally divided agonistic conduct patterns into two broad classes, making a distinction between "defensive" aggression in response to attack and "offensive" aggression, which is typical of the interactions between males of a species. The latter includes predatory aggression in which the victim is killed and eaten, not necessarily for reasons of pure hunger. Other types of aggression exist, e.g., prepubertal playful aggression and the aggression of the breast-feeding mother—the purpose of which is to protect her offspring against male aggression. These various forms of aggressive conduct, which are easily differentiated by specific associated behaviour patterns, have been carefully characterised and described.

Investigation of exogenous etiologic factors that lead to aggressive conduct in laboratory animals, both prenatal and postnatal factors, has focused on the paradigm of stress. Physical or social stress can be used to modulate an animal's reactions. It is important to highlight the importance of the temporal dimension in the effects of stress on agonistic conduct: the literature shows that chronically stressed animals are more aggressive than controls, whereas animals who experience a single, highly stressful event tend to become more passive. The fact that the consequences of stress differ according to whether it is acute or chronic provides reference points when it comes to dissecting the pathogenic mechanisms underlying conduct disorder, notably with respect to the temporal dimensions—duration, repetition—of relevant life events.

The duration and the intensity of stress also seem to be important parameters in the pathogenesis of motor hyperactivity, attention deficit, and impaired behavioural inhibition which all commonly co-occur with conduct disorder in humans.

In rodents, 3 to 8 weeks of chronic exposure to repeated, unpredictable, moderately intense stressors stimulates activity and exploratory behaviour. In contrast, exposure to varied, unpredictable stressors of much greater intensity leads to reduced physical activity. Acute exposure to a severe stressor over a much shorter time frame increases locomotion without any effect on exploratory behaviour, but if this type of stress is sustained for a longer period, the rats' global level of locomotor activity drops within 24 hours.

There appears therefore to be a link between stress and motor activity. Depending on the intensity of the stressor, the victim's motor activity either increases or decreases. The effect of chronic stress seems to have a converse and more complex effect, pushing the victim toward the "impulsive" end of the "psychomotor" dimension; this is compatible with the idea of a drop in cerebral serotonin levels.

Unawareness of the environment entails indifference to danger. This is clearly observed in chronically stressed animals exposed to high-risk situations (e.g., being placed on an unstable surface where there is a risk of falling off). This type of behaviour associated with "unawareness of the danger" is typical of impaired inhibition mechanisms, and it resembles in some ways the effects of certain anxiolytic drugs. In the same way, in stressed animals performing free exploratory tasks, hyperactivity manifests as impaired habituation, in other words as a failure to ignore irrelevant information from the environment. Thus, relatively mild but chronic stress (modelling chronically stressful life events) has a major effect on selective attention and leads to serious impairment of behavioural inhibition mechanisms. The enhanced psychomotor reactivity of these animals can be addressed, as in the hyperactive child, from the perspective of attention-related processes and their consequences on inhibitory mechanisms.

Moreover, the aggressive responses to such environmental effects (i.e., stress) depend on the state of the subject (age, endocrine parameters, etc.). Experiments in which animals are confronted with a violent environment (e.g., living with aggressive peers) at different ages reveal the existence of a relatively short sensitive period around puberty during which exposure to violence can lead to vulnerability. Living with violent subjects around the age of puberty leads to the transformation of perfectly well-characterised, playful-type aggression into adult-type aggression which is itself perfectly well-characterised. The same is true for social isolation, which seems to have opposite effects according to whether it occurs before this critical period (in which case aggression is exacerbated) or after (in which case passivity is increased).

Comparison of the findings on aggression in small laboratory animals and epidemiological and neurobiological data on conduct disorder in children and adolescents underlines the inverse relationship between anxiety and aggressive behaviour. Psychobiological correlations with aggressive behaviour, together with the results obtained in rats and mice, are providing explanations that are of great heuristic value because they make it possible to isolate the effects of two variables that are fundamental to aggression, namely fear (which is difficult to address in human experiments) and anxiety. The fact that these two variables can be separated in animals has shown aggression and anxiety to be independent of one another (as are aggression and the effects of anxiolytic drugs), whereas aggression and fear are directly related, the latter being independent of anxious reactions. This line of research may lead to new pharmacological options for the treatment of conduct disorder.

Research in biology, neurobiology, and brain imaging is enhancing our understanding of conduct disorder

In biology, the study of aggression and violence is a relatively recent endeavour currently indicating parameters that might be relevant to conduct disorder.

Abnormally low cholesterol levels have been observed in the blood of men who have committed violent acts and in people who have attempted suicide by violent means. A possible causal relationship between low blood cholesterol levels and aggressive or violent behaviour is suggested by the fact that monkeys on a low-cholesterol diet tend to be more aggressive. Moreover, people taking cholesterol-lowering drugs (statins) have reported episodes of severe irritability, aggression, and even impulses to torture and murder. However, other studies have failed to confirm these results, showing that the relationship between blood cholesterol and aggressive or violent behaviour is a complex one, probably under the control of many as yet undefined factors that vary substantially from one individual to another.

Episodes of behavioural perturbation associated with low levels of blood cholesterol could result from changes induced in neurotransmission in serotonin, dopamine, and GABA circuits in the brain, although solid data are scarce.

It is generally accepted that peripheral sympathetic hypoactivity could be associated with social indifference, low emotional reactivity (e.g., in response to punishment), attention deficit, and impulsiveness. Children and adolescents of 6–16 years of age with conduct disorder have been shown to have abnormally low levels of dopamine beta-hydroxylase activity in their blood. This enzyme is released along with noradrenaline in response to sympathetic stimulation. Consistent with this finding, the average heart rate (which is under the direct control of the sympathetic system) in young boys (11–12) with conduct disorder and aggressive/impulsive behaviour patterns has been shown to be lower than in their non-aggressive peers. However, this relationship between heart rate and antisocial or aggressive behaviour seems to disappear in the course of development and is no longer observed in adulthood.

Other observations consistent with those concerning the sympathetic system suggest that conduct disorder could be related to hypoactivity in stress systems, probably in relation to the emotional deficit and social indifference observed in affected subjects. On the basis of results reported in animals, there could also be a causal relationship between low blood cortisol levels and conduct disorder.

On the other hand, early studies pointed out the existence of a positive link between elevated blood testosterone levels and aggressive or violent behaviour but no systematic trend has really been observed. It might be that this profile may be relevant in a sub-population of subjects, but no such group has yet been characterised in terms of genetics, temperament, or biological characteristics.

According to extensive work in the field of clinical biology, an important factor in conduct disorder and the commission of violent acts could be serotonin (5-hydroxytryptamine or 5-HT), the levels of which are determined by fluctuations of its precursor tryptophan at the periphery. As a general rule, a reduction in the concentration of this essential amino acid (which is exclusively derived from food in humans and rodents) exacerbates aggressive behaviour; inversely, an increase in tryptophan intake can moderate aggressive conduct. However, such effects remain relatively modest and are not systematically reproducible.

It has been established however that specific biological patterns are related to certain neurobiological patterns, notably affecting neurotransmitter systems in the central nervous system.

Findings from this type of research have been compared with data on how neuroactive and psychotropic substances work to attenuate this type of behaviour, especially neuroleptic drugs, antidepressants, and anxiolytics. Consideration of the molecular targets of various classes of drug have led to the conclusion that serotonin, the catecholamines (especially dopamine), and GABA are involved in impulsiveness, aggression, and possibly the other traits of antisocial personality disorder. Studies performed in humans (including school children) and laboratory animals tend to confirm this conclusion. However, other neuroactive compounds—in particular neuropeptides like substance P, arginine-vasopressin, and corticotropin releasing hormone (CRH)—also play important roles in the neurobiological mechanisms that underlie behaviour, including impulsive behaviour and aggression.

In parallel, the development of relevant animal models, especially in rodents (rats and mice, including genetically modified lineages) but also in primates, has refined our understanding of how the various neuromediators (e.g., neurotransmitters such as the monoamines and neuropeptides) might be involved in regulating the expression of impulsive and aggressive behaviour and violent manifestations. Impulsiveness and aggression have been observed to be exacerbated in mice in which the 5-HT1B serotonin receptor has been knocked out. This has triggered a search for agonists of this receptor that might attenuate such behaviour patterns. Effective agonists have been generated (called "serenics"), but prolonged administration of these candidate drugs leads to tolerance and reduced responsiveness, so no clinically useful psychoactive compounds have as yet been identified.

Finally, recent advances in brain imaging techniques are paving the way toward a new era in research into neurobiological mechanisms through identification of the key neural circuits involved, including those associated with impulsive, aggressive, and violent behaviour.

Limbic structures (the hippocampus, the hypothalamus, the septum, the amygdala, and the bed nucleus of the stria terminalis) and the anterior cingulate and orbito-frontal areas of the pre-frontal cortex have been pinpointed as being the parts of the brain in which activities are altered in association with impulsive, aggressive, and violent manifestations. In consequence, it is now known that limbic structures play a key role in controlling emotions and self-control mechanisms, whereas motivation and passage to violent action is associated with activation of the anterior cingulate and orbito-frontal areas of the pre-frontal cortex. In patients, these are the very areas in which changes are observed in the activity of neural networks implicating monoamines and other neuropeptides, which clearly play a central role in the expression and regulation of these types of behaviour. The deficit in serotonin that has been associated with impulsiveness and aggression might well affect the anterior cingulate and orbito-frontal areas of the pre-frontal cortex, where certain serotonin receptors are known to be particularly dense. This deficit could result in a decrease in the cortical efferent capacity to inhibit the subcortical structures that are responsible for aggression and passage to violent action. Consolidation of serotonin-based regulatory systems in particular, and monoamine-based systems in general, through diverse environmental and/or pharmacological interventions could therefore correct a functional deficit in the frontal cortex. At this time, the most likely hypothesis is that this functional deficit appears in childhood or adolescence. It is known that the prefrontal cortex is the last part of the brain to be differentiated in humans, with definitive connections only ultimately becoming established by the end of adolescence, under the control of multiple genetic and environmental influences. Childhood and adolescence are therefore critical periods in the construction of personality via the progressive maturation of different neurotransmitter systems, especially in this particular part of the brain. In other words, conduct disorder could be related, at least in part, to impaired development of the cortico-subcortical circuits involved in the control of impulsiveness, aggression, and social interactions. It is reasonable to believe that progress in functional brain imaging techniques—coupled with the elucidation of the mechanisms of action of those neurobiologically active and psychotropic agents which are known to have some degree of efficacy in the treatment of conduct disorder, ADHD or ODD—will help advance our understanding of the etiology of these conditions and may open the way toward new therapeutic options.

Training and Information

Screening

Diagnosing and treating conduct disorder

Prevention

Promote research

Promote research