The heterogeneity of causes and courses of attention-deficit/hyperactivity disorder


  • H-C. Steinhausen

    1. Aalborg Psychiatric Hospital, Aarhus University Hospital, Denmark
    2. Clinical Child and Adolescent Psychology, University of Basel, Switzerland
    3. Department of Child and Adolescent Psychiatry, University of Zurich, Switzerland
    Search for more papers by this author

  • Invited paper

Hans-Christoph Steinhausen, Aalborg Psychiatric Hospital, Aarhus University Hospital, Mølleparkvej 10, DK-9000 Aalborg.


Objective:  Attention-deficit / Hyperactivity Disorder (ADHD) is a frequent mental disorder with onset in childhood and persistence into adulthood in a sizeable number of people. Despite a rather simple clinical definition, ADHD has many facets because of frequent co-morbid disorders and varying impact on psychosocial functioning. Thus, there is considerable heterogeneity in various domains.

Method:  A review of recent research findings in: i) selected domains of aetiology reflecting the role of genes, brain structures and functioning and the interplay of causal factors and ii) clinical heterogeneity in terms of co-morbidities, gender effects, courses and outcomes.

Results:  Molecular genetic studies have identified a number of candidate genes which have a small effect on behavioural variation in ADHD. In the most recent Genome Scan Meta Analysis of seven ADHD linkage studies, genome-wide significant linkage was identified on chromosome 16. The volume of both the total brain and various regions including the prefrontal cortex, the caudate nucleus and the vermis of the cerebellum is smaller in ADHD. Functional MRI has documented a specific deficit of frontostriatal networks in ADHD. Integrative aetiological models have to take the interaction of gene and environment on various dysfunctions into account. Clinical heterogeneity results from frequent associations with various co-morbidities, the impact of the disorder on psychosocial functioning, and gender effects. Partly, these effects are evident also in the course and outcome of ADHD.

Conclusion:  ADHD is a chronic mental disorder with a complex aetiology. So far, various neurobiological factors have been identified that need to be studied further to better understand their interaction with environmental factors. The clinical presentation and the long-term course of ADHD are manifold.

Clinical recommendations

  •  Neurobiological research findings on the complex aetiology of ADHD do not yet have clear clinical implications.
  •  The manifold combinations of the core ADHD symptoms with co-morbidities and the negative impact on psychosocial functioning at all developmental stages asks for detailed clinical assessment and multi-dimensional treatment.

Additional comments

  •  The persistent character of the disorder poses large burdens on the affected individuals, their social environment and the society at large.
  •  The upcoming revisions of the international systems of classification will have to consider the heterogeneity of ADHD.


Attention-Deficit/Hyperactivity Disorder (ADHD) is one of the most frequent disorders with onset in childhood with some 3 to 8 per cent of children affected by the disorder worldwide. Both in the ICD-10 and in the DSM-IV, the cardinal features are inattentiveness, hyperactivity and impulsiveness. In terms of subtypes, the two systems of classification differ with ICD-10 considering a combined (C) type and the co-morbid condition of hyperkinetic conduct disorder, whereas DSM-IV makes a distinction between the C subtype, the inattentive (IA) subtype and the hyperactive-impulsive (HI) subtype. Despite its relatively simple definition, ADHD aetiologically and clinically is more a rather heterogeneous than a homogenous syndrome as will be shown in the following.

Aims of the study

The purpose of this contribution is to provide an overview of the heterogeneity of ADHD in terms of the current understanding of aetiological factors and the different courses that the disorder may take in the affected individuals.

Material and methods

This review sought to highlight some of the emerging major lines of research dealing with the multitude of factors that contribute both to the causes and the courses of ADHD. Given the very dynamic developments in the field of ADHD research, the following review is not exhaustive and has to be selective in reporting to some extent.


Heterogeneity of causes

There is no doubt that ADHD is caused by a multitude of factors with a large set of neuro-biological and environmental factors operating in a highly complex fashion. The present overview will deal only with a selected number of findings on factors that have been emphasized strongly in the recent past. These factors include genes, brain structure and functioning, and some environmental conditions.

Genes.  Research into the genetics of ADHD started with family studies, adoption studies, and twin studies demonstrating that ADHD clusters in families and that the disorder is highly heritable. These early studies have shown that genetic factors cause about 75% of the ADHD-related behavioural variance. In addition, behavioural genetic studies have shown that individual (non-shared) environmental factors contribute to the behavioural variance, whereas familial (shared) environmental factors seem to play at least a minor role (1).

With the recent advances in molecular genetic research there have been an increasing number of studies looking for various candidate genes with both positive and negative results. However, based on replicated findings of risk alleles for ADHD with pooled odds ratios across at least three studies from 1.1 to 1.5 there are some convergent findings. Thus, polymorphisms of the dopamine D4 and the dopamine D5 receptor gene, the dopamine transporter gene, the serotonin receptor gene (HTR1B), and the synaptosomal-associated protein 25 gene are among the most robust findings (2). However, these associations are small and consistent with the idea that many genes of small effect contribute to genetic vulnerability in ADHD.

Furthermore, genetic linkage scan methods allow for the identification of relatively large chromosomal regions which may contain susceptibility genes for ADHD, based on the sharing of chromosomal segments among affected members of a family. In the most recent Genome Scan Meta Analysis of seven ADHD linkage studies, genome-wide significant linkage was identified on chromosome 16 between 64 Mb and 83 Mb (3).

So far, there have been two genome-wide association studies (GWAS) of ADHD. The International Multi-centre of ADHD Gene (IMAGE) study examined trio families (two parents and one ADHD child) using a large number of tagging single nucleotide polymorphisms (SNPs) designed to be in high linkage disequilibrium with all untyped SNPs in the genome. No finding achieved genome-wide significance in the primary analysis of the ADHD diagnosis (3, 4).

There are a number of limitations that have to be kept in mind when it comes to the interpretation of these genetic findings. Even a strong genetic distribution does not imply determinism because genetic influences are expressed through interaction with the environment. Furthermore, significant associations of certain candidate genes account only for rather small proportions (less than five per cent) of explained behavioural variance.

The small amount of explained variance and the inconsistencies of findings across studies may be due to genetic heterogeneity, i.e. different combinations of gene variants in affected individuals. In addition, the expression of certain gene variants may be affected by moderator variables like sex and age (2), age at onset (5), or gene-gene interactions (6).

Gene-environment interactions may also be operant. Two recent studies focussing on the DAT1 gene are of particular interest. One study showed that family adversity moderates the impact of the DAT1 genotype on the expression of ADHD symptoms (7). In another study it has been shown that only those individuals carrying the DAT1 risk alleles whose mothers used alcohol during pregnancy showed an increased risk for ADHD (8). However, even in this study the question remained open whether there were direct toxic effects as one would expect from the teratogenic effects of alcohol seen in the foetal alcohol spectrum disorders (FASD) which lead to ADHD in a sizeable proportion of the affected individuals (9). However, the observed interaction may also have been due to a correlation of drinking with deficient parenting or maternal psychopathology, or due to genes causing changes both in parents and offspring by gene–gene interaction (epistasis).

The clinical implications of many of the genetic findings are even further limited, because most children with DNA variants do not have ADHD whereas most children with ADHD do not have the known DNA variants. In addition, an association of the disorder with certain variant alleles does not yet prove a cause. The genetic associations may be part of only certain aspects of the ADHD phenotype, the so-called endophenotypes, which are intermediate between clinical diagnosis and the biological variables that are considered the cause of the disorder. Currently, a large number of biochemical, neuropsychological, electrophysiological and other endophenotypes are studied to find out whether or not they may serve as specific genetic vulnerability markers.

Genetic associations may also relate to co-morbid conditions rather than the pure ADHD behavioural phenotype. Accordingly, latent class analyses of twin data have indicated that there are a number of genetically independent forms of ADHD that differ in severity of the symptom domains and vary in co-morbidity patterns (10). Further behavioural genetic studies assumed that it is likely that the co-morbidity between ADHD and oppositional-defiant disorder (ODD) / conduct disorder (CD) is due to a shared genetic liability either operating directly or indirectly through gene-environment correlations or interactions. It seems that the co-variation is governed by a common set of genes with ODD / CD being influenced also by additional genetic factors (11). In another study it was shown that ADHD shares most of its genetic liability with CD / ODD and executive functions (EF) deficits and there was preliminary support for additional genetic influences underlying CD, ODD, and EF that are independent of ADHD(12). The heritability of reading disorders (RD) with ADHD is stronger in the inattentive (h2 = 0.42–0.47) rather than in the hyperactive-impulsive sub-type (h2 = 0.20–0.22) (13). Australian twin data showed a strong additive genetic component between most subtypes of ADHD and developmental coordination disorder (DCD) to the subtypes of the other disorder (14). Finally, molecular genetic studies have also shown that the same locus / candidate gene can affect several disorders and functions (so-called pleiotropic effect). This is true for ADHD and RD (on chromosome 6p22) (15), and for ADHD and Autism (16p3) (16).

Brain structure and functioning.  Anatomical MRI studies have shown that children with ADHD compared with typically developing children show a subtle reduction in total brain volume. Besides this global reduction in volume, individuals with ADHD shows: i) smaller prefrontal volumes, ii) smaller volumes of the caudate nucleus and iii) differences in the cerebellum, in particular, a smaller vermis (17–19). In addition, various functional MRI studies revealed that differences in cognitive control between subjects with and without ADHD are associated with differences in brain activation patterns. These studies have demonstrated reduced activation in prefrontal and striatal regions during neuropsychological test paradigms that require subjects to suppress prepotent tendencies as part of the task. A dysfunction in temporal and parietal areas also plays a role, mostly during tasks that examine attentional processes. The most consistent evidence points to the importance of the frontostriatal networks in ADHD. However, very little is known how changes in frontostriatal functioning and associated cognitive deficits are related to anatomical changes (20).

There is a long standing question whether ADHD is because of a developmental delay or persistent deficit. Two recent studies have addressed this question with different methods and came to quite different conclusions. The developmental trajectories of cortical grey matter were investigated by use of MRI in a large ADHD sample and controls and it was found that cortical grey matter thickness in ADHD follows the same trajectory as in typically developing children. However, it is lagging behind by an average 3 years. This structural finding constitutes evidence that ADHD is associated with a developmental lag in cortical maturation (21).

In the other study by Döhnert et al. (2009), an EEG-based examination of the development of neurophysiological markers of attention (Cue P300, CNV) and inhibition (NoGo P300) in ADHD and control groups was performed repeatedly from childhood to adolescence. Event-related potential (ERP) maps were recorded during a cued Continuous Performance Test (CPT) at all assessments, and analyzed using scalp and source (sLORETA) measures. CPT performance showed common effects of ADHD and younger age, consistent with (but not specific to) developmental lag. The NoGo P300 developed earlier and became stronger in controls than in the ADHD group, which is also consistent with an initial developmental lag. In contrast, the attenuation of the Cue P300 and the CNV with ADHD at all assessments was opposite to the enhancement with younger age, and thus inconsistent with developmental lag. The sLORETA source localization also differed between ADHD and developmental effects.

These functional results provide strong evidence for multiple and persistent neural processing deficits in ADHD. They do not support the developmental lag hypothesis for attentional dysfunction in ADHD despite partial evidence that developmental lag contributes to inhibitory brain dysfunction during early adolescence (22).

Recent neuroimaging and genetic studies have used convergent methods and resulted in some integrated findings (20). In these studies, brain changes have also been found in family members of affected individuals. For instance, reductions in cortical grey matter have been found in siblings with and without ADHD. Even more interestingly, two ADHD risk genes have been shown to directly impact frontostriatal grey matter volumes. In these studies, the DRD4 genotype was associated with a reduction in prefrontal grey matter volume, whereas the DAT1 genotype was associated with reduced caudate nucleus volume. Thus, it has become possible to map the effects of ADHD risk genes on neuroanatomical measures. This was shown both indirectly in terms of effects shared by family members, who may be carriers of unknown risk genes, and directly by showing that ADHD risk genes may affect brain structure. More recent studies have even shown differential effects of the DAT1 genotype on brain activation patterns between individuals at genetic risk for ADHD and controls. In these studies it was found that the DAT1 genotype interacted with familial risk for ADHD in the striatum, but not in the vermis of the cerebellum.

Integrated causal models.  The various causal influences need quite a complex model that takes into account both neurobiological and psychological constructs as outlined in Fig. 1 (23). The model starts from the contribution by genes and environment that have an impact on various sorts of dysfunctions comprising, e.g. theoretical constructs like inhibition deficits, delay aversion, or other neurophysiological or neurochemical dysfunctions which are currently considered to be central to the understanding of the key deficits in ADHD. These dysfunctions lead to various ADHD phenotypes representing the heterogeneous clinical manifestations of ADHD. The model also takes account of the various bidirectional interactions of both neurobiological and environmental risk and protective factors on each level and may serve as an aid to better understand the causes of ADHD.

Figure 1.

 A model of the complex aetiology of attention-deficit/hyperactivity disorder (ADHD).

Heterogeneity of clinical course

Comorbidities and effects on psychosocial functioning.  The clinical picture of ADHD is not only marked by different severity of the defining symptoms but also by rather frequent co-morbid disorders with oppositional defiant disorder, conduct disorder, anxiety disorders, and mood disorders being the most frequent conditions and multiple combinations being quite frequent. Evidence for these associations has been provided both in epidemiological surveys (24, 25) and in clinical samples (26, 27). A recent large Pan-European project studied the impact of co-existing problems with ADHD on behavioural features, psychosocial functioning, and quality of life. It was found that having multiple co-existing psychiatric problems increases the severity of ADHD in all domains (behavioural features, psychosocial impairment, quality of life) and that a similar though less consistent pattern applies also to co-existent ODD/CD problems, whereas the effect of other co-existing problems like anxiety/depression, tics, and developmental coordination disorder is not different from ADHD without co-existing problems (28).

Many affected children suffer from serious impairments in cognitive and social functioning (29). School performance may be affected not only because of deficits of attention but also because of subtle impairment of intelligence. It is unclear, whether the latter are because of inattention or as a result of shared genetic causes (30). Furthermore, ADHD may have serious impact on education, employment, and social functioning. In comparison with controls, adolescents with ADHD have been shown to develop higher rates of attempted suicide, intentional injury, incarceration and substance abuse whereas adults with ADHD suffer from higher rates of unemployment or lower job status, relationship/ marital difficulties, chaotic household organization, risk taking behaviour, accidents and legal violations (31, 32).

Figure 2 summarizes the complexity of the clinical phenomena by illustrating how the core features of ADHD in combination with co-morbid disorders and symptoms impact on psychosocial functioning in various domains.

Figure 2.

 The association of core symptoms, co-morbidities and psychosocial functioning in ADHD.

Gender effects.  Females with ADHD are underrepresented in both epidemiological and clinical samples. There was a 1 : 4 female to male ratio in the British Mental Health Survey 1999 which differed for the subtypes with a more pronounced difference in the HI subtype (1 : 7) rather than in the Combined (1 : 5) and in the IA (1 : 3) subtype. The rate of referral of girls in clinical settings is even higher (1 : 9). In clinic-based studies girls show less disruptive behaviour, more anxiety, depression and peer rejection, and have more psychiatric admissions in adulthood (33). A meta-analysis of findings (34) showed that there were no gender differences in impulsiveness, social and peer functioning and academic performance though girls were rated lower on hyperactivity and externalizing problems.

When all ADHD types were taken together among non-referred children, there were no differences on core symptoms, co-morbidity or impairment, but more somatic complaints and better school functioning in girls. The impairment was equal or higher in the IA group, but less in the COMB and HI type in girls in one Australian study, there were no differences in subtypes, co-morbidity, treatment history, and functioning in a study from the US, and there was the lowest gender (1 : 2.3) ratio, a higher prevalence of co-morbid anxiety disorders in girls with IA, and more co-morbid mood disorders in boys with the C subtype in an epidemiological study from Puerto Rico (33).

The recent Pan-European study of a large sample of children with ADHD found gender ratios from 1 : 3 to 1 : 16 (f:m), few gender effects in core symptoms and correlates, but more parent-rated emotional symptoms and prosocial behaviour in girls. They were also more likely to be the victim of bullying and less likely to be the bully. Both genders had similar levels of co-existing psychiatric and physical health problems, and received the same type of treatment (33).

Course and outcomes.  In many affected individuals hyperactivity and, less strongly, inattention decline from childhood to adolescence (35). Stability of diagnostic criteria is relatively high at ages 4–6 (36) whereas there is a considerable decline at the turn of late childhood to early adolescence (37). However, a sizeable proportion of these adolescents not fulfilling diagnostic criteria still have quite abnormal behavioural features that may need intervention. Various outcome studies have followed-up children with ADHD into adolescence and found varying rates of persistence (43–72 per cent), high rates of co-morbid CD or ODD, and increased rates of substance abuse most frequently with alcohol, tobacco, or marijuana, rather, than cocaine or opiates and other street drugs (38).

The outcome in adolescents may be predicted by a number of features each having relatively limited prognostic power when considered in isolation: i) low socioeconomic status is associated with higher severity and poor prognosis of ADHD, and lower intelligence has a negative impact on scholastic career, ii) the severity of early peer problems is a predictor of personal relationship problems in adolescence and adulthood, iii) co-existent CD predicts various indicators of poor psychosocial functioning including poor school career, relationship problems and substance abuse, iv) parental ADHD, particularly in combination with antisocial personality disorder and substance use disorder (SUD), increases the likelihood of mental disorders in the adolescent, v) early hostile parent-child-interactions are associated with later conflicts between parents and adolescents and contribute to adolescent aggressive behaviour and vi) severity of ADHD relates to achievement at school only (38).

Various controlled long-term adult outcome studies have documented that ADHD may persist completely or may develop into a residual syndrome and may constitute a risk factor for various other mental disorders. The rate of persistence of ADHD is strongly dependent on the kind of criteria used for the definition of outcome. If very strict criteria are used, only 15 per cent of former children with ADHD still have a persistent disorder in adulthood whereas the figure rises to 65 per cent if partial remissions are counted (35). Frequent co-morbid conditions include SUD which in the majority of cases had been proceded by CD. Other frequent co-morbid disorders include cluster B (histrionic, borderline, antisocial) personality disorders, mood disorders and anxiety disorders, and there is also an increased rate of delinquency and criminality (31, 39, 40). In addition, findings of a large meta-analysis indicate that in comparison with children also adults with ADHD display the same neuropsychological deficits in terms of lacking impulse control, behavioural inhibition, attention and verbal short-term memory (41). Findings on prognostic indicators are either controversial or of limited significance with the exception that severity of ADHD and treatment status may predict persistence of ADHD in adulthood (38).


There is clear indication both from research and clinical experience that despite a very simple definition by three key behavioural features and a few additional diagnostic criteria, ADHD is a rather complex and heterogeneous disorder. The final aetio-pathological pathways are not yet fully understood. However, it is clear that a very complex interaction of various neuro-biological factors including genetic abnormalities with environmental factors have to be considered in any model that tries to integrate the multitude of diverse findings.

As a consequence, the developmental trajectories of the affected individuals are manifold and vary with the life-span. Impaired school careers and behavioural problems carry the greatest risk on the development of children and adolescents with ADHD. The second largest risk is derived from co-morbid ODD and CD that may lead to serious problems like delinquency, antisocial personality and SUD in adulthood. A third major risk starts with pronounced relationship problems in children with ADHD and may continue with similar problems including risky sexual contacts and partnership problems in adult life. Compared with normal controls, adults with ADHD show lower educational and vocational status, frequent co-morbid disorders, and higher instability in their relationship to others.

From a clinical point of view, it is mandatory that this large group of patients with a heterogeneous and chronic mental condition gets continuous professional counselling and treatment by highly qualified specialists. Scientifically, a better understanding of ADHD has to be based on the notion of heterogeneity of the disorder which has to be taken into consideration also in the upcoming revisions of the major international systems of classification of mental disorders.

Declaration of interests

The author does not have to declare any competing interests with the content of the present contribution.