Evidence-based guidelines for
management of attention-deficit/
hyperactivity disorder in adolescents in
transition to adult services and in
adults: recommendations from the
British Association for Psychopharmacology
D. J. Nutt Psychopharmacology Unit, University of Bristol, Bristol, UK.
K. Fone University of Nottingham, Nottingham UK.
P. Asherson MRC Social Genetic Developmental Psychiatry, Institute of Psychiatry, King’s College London, UK.
D. Bramble Telford & Wrekin PCT, Shrewsbury, UK.
P. Hill London, UK.
K. Matthews University of Dundee, Dundee UK.
K. A. Morris c/o Psychopharmacology Unit, University of Bristol, Bristol, UK.
P. Santosh Institute of Psychiatry, London, UK.
E. Sonuga-Barke University of Southampton, Southampton, UK.
E. Taylor Institute of Psychiatry, London, UK.
M. Weiss University of British Columbia, Vancouver, Canada.
S. Young Bethlem Royal Hospital, Kent, UK.
For the Consensus Group (other invited participants at the consensus meeting ‘ADHD in transition
from child to adult’ are listed in the Acknowledgements).
21(1) (2007) 10–41
©2007 British Association
SAGE Publications Ltd,
London, Thousand Oaks,
CA and New Delhi
Journal of Psychopharmacology
Attention-deficit/hyperactivity disorder (ADHD) is an established
diagnosis in children, associated with a large body of evidence on the
benefits of treatment. Adolescents with ADHD are now leaving children’s
services often with no readily identifiable adult service to support them,
which presents problems as local pharmacy regulations often preclude
the prescription of stimulant drugs by general practitioners (GPs). In
addition, adults with ADHD symptoms are now starting to present to
primary care and psychiatry services requesting assessment and
treatment. For these reasons, the British Association for
Psychopharmacology (BAP) thought it timely to hold a consensus
conference to review the body of evidence on childhood ADHD and the
growing literature on ADHD in older age groups. Much of this initial
guidance on managing ADHD in adolescents in transition and in adults is
based on expert opinion derived from childhood evidence. We hope that,
by the time these guidelines are updated, much evidence will be
available to address the many directions for future research that are
ADHD, hyperkinetic disorders, hyperactivity, impulsivity,
psychostimulants, psychotherapy, co-morbidities
Corresponding author: Prof. David J. Nutt, Psychopharmacology Unit, Dorothy Hodgkin Building, Whitson St, Bristol BS1 3NY, UK.
ADHD management in adolescents to adults
In recent years, UK health services have proven inadequate in
meeting the needs not only of children with ADHD in transition
from children’s services, but also of the growing number of adults
newly presenting with ADHD symptoms, often after their child
has been given the diagnosis. Specialist services are now being
established to manage adolescents in transition and adults with
ADHD. The BAP thought it helpful to act as the vehicle to prepare
a consensus document on best practice in an area that is both con-
troversial and rapidly changing.
The BAP is an association of psychiatrists, psychopharmacolo-
gists and pre-clinical scientists who are interested in the broad
field of drugs and the brain. BAP is the largest national organ-
ization of its kind worldwide, and publishes the Journal of Psy-
chopharmacology. The association started publishing consensus
statements more than a decade ago, and the first BAP guidelines
on depression were considered a landmark publication when pub-
lished in 1993 (Montgomery et al., 1993). That document, which
was updated in 2000 (Anderson et al., 2000), has become the stan-
dard of care in many countries because it is considered an accessi-
ble consensus to guide practising psychiatrists. The BAP now has
a target of publishing one consensus statement per year in the
Journal of Psychopharmacology. Recent guidelines have covered
management of bipolar disorder (Goodwin, 2003) and drug treat-
ments for addiction (Lingford-Hughes et al., 2004), with guide-
lines on anxiety published late last year (Baldwin et al., 2005).
Forthcoming consensus conferences are planned on child psy-
chopharmacology, old age psychopharmacology and schizo-
phrenia, all of which will utilize a similar style and process.
All guidelines are available
(http://www.bap.org.uk) and the intention is to update each guide-
line every 5 years.
via the BAP website
A consensus conference was held at the Novartis Foundation
building, London on 24 February 2005. The meeting was spon-
sored by unrestricted educational grants from Cephalon, Janssen,
Lilly, Shire UK and Shire US pharmaceutical companies, which
had no input into and no responsibility for the meeting and its
content. Those invited to attend the meeting included BAP
members, representative clinicians from services with a strong
interest in ADHD, and other recognized experts and advocates in
the field, including a representative from Canada who had recently
participated in that country’s consensus statement on adult ADHD
(www.caddra.ca). Observers from the sponsoring companies were
invited to attend but not to participate in the proceedings or in
drafting the guidelines. All attendees completed conflict of interest
statements that are held at the BAP office as per usual BAP policy.
Speakers were asked to present a review of the literature and
identification of the standard of evidence in their area, with an
emphasis on meta-analyses, systematic reviews and randomized
controlled trials (RCTs) where available. Each presentation was
followed by a lengthy discussion that aimed to reach consensus
where the evidence and/or clinical experience was considered ade-
quate, or otherwise to flag the area as a direction for future
research. A draft of the taped proceedings was drawn up by an
attending clinician writer and circulated to all speakers and other
participants for comment. Key subsequent publications were
added by the writer and speakers at draft stage. All comments
were incorporated as far as possible in the final document, which
represents the views of all participants, although the authors take
final responsibility for the document.
Categories of evidence for causal relationships, observational
relationships and strength of recommendations are given in Table 1
and are taken from Shekelle et al. (1999). The strength of
recommendation reflects not only the quality of the evidence, but
also the importance of the area under study. For example, it is pos-
sible to have methodologically sound (category I) evidence about
an area of practice that is clinically irrelevant, or has such a small
effect that it is of little practical importance and therefore attracts a
lower strength of recommendation. However, more commonly, it
has been necessary to extrapolate from the available evidence (e.g.
in childhood) leading to weaker levels of recommendation (B, C
or D) based upon category I evidence statements.
Scope of the guidelines
Our intention is to present a comprehensive statement to guide cli-
nicians, who are managing adolescents with ADHD in transition
from children’s services, and adults newly presenting with ADHD
symptoms. Although rigorous evidence exists in a few areas,
overall there is a dearth of evidence that pertains to adolescents
and especially adults. Many of the statements in this document are
therefore based on the experience of experts who are currently
treating this adolescent or adult group, with extrapolation from the
greater body of rigorous data in children. We anticipate that the
situation with regard to research on adult ADHD is likely to
change markedly in the future. Until that time, these guidelines
must be considered a first attempt to reach consensus in an area
that continues to attract controversy, and in which we cannot be
certain that childhood findings will be confirmed in older age
groups. We hope that the body of evidence may be considered far
more definitive when these guidelines are due for updating.
Our consensus on adolescents in transition and adults with
ADHD is based on the premise that a clinical entity, ADHD,
exists in childhood and, moreover, that the condition may persist
in some form in older age groups. The meeting recognized the fact
that even ADHD in childhood is a controversial diagnosis to make
in some quarters of society; thus, a diagnosis in older age groups
might be more controversial in these quarters. The crux of the con-
troversy seems to be whether the symptoms of ADHD need to be
understood as a medical condition or whether the behaviours iden-
tified as ADHD are an extreme of the normal spectrum of human
behaviours (e.g. ‘very naughty little boys’). A further dimension is
whether such behaviours necessitate pharmacological treatment or
whether non-pharmacological modifications are sufficient to
The consensus of the meeting is that the extreme behaviours
characterized under current diagnostic systems represent a clinical
condition in so much as the behaviours cause problems for the
ADHD management in adolescents to adults
affected individual, are identified as problematic by others and
may respond to various forms of clinical treatment or other forms
of management, such as environmental restructuring. A diagnosis
of ADHD is thus warranted to allow affected individuals to access
appropriate forms of support from health care and other systems.
We hope the controversy will be lessened over ADHD in older
age groups, because in many cases, the affected individual will be
the one to initiate clinical contact. Indeed, it is partly in response
to the growing number of such requests that this consensus
meeting was arranged.
With current science, it is not possible to determine whether
ADHD is an extreme variant of normal behaviour or in a distinct
category. However, current science does support the improvement
of ADHD symptoms with treatments, particularly but not exclu-
sively pharmacological. It is in this spirit that this consensus docu-
ment is presented. The question of who should provide and pay for
service provision is a matter of health policy and thus beyond the
scope of our consensus, although we hope that this consensus will
prompt this debate.
The exact cause(s) of ADHD are unknown but the growing liter-
ature on genetics, neuroimaging and neuropsychology suggests
features consistent with a neurodevelopmental condition in which
multiple environmental and biological factors, of individually
small effect, interact to produce an abnormal brain condition that
manifests as cognitive and behavioural deficits (Sonuga-Barke
et al., 2005). Subsequent bidirectional interactions then occur
between these deficits and environmental and biological factors to
modify the phenotype further, resulting in various diagnostic sub-
groups under the umbrella term ADHD. Overall, the literature on
genetic, neurobiological and neuropsychological deficits shows
marked disparities between subgroups. This may indicate that cau-
sation is likely to show marked heterogeneity, and/or that tradi-
tional theories on cause and effect in ADHD, that regard ADHD
as a homogeneous condition characterized by a single causal
pathway, are inadequate (Sonuga-Barke, 2003, 2005; Coghill
et al., 2005; Castellanos et al., 2006).
Genetics and neuroscience
ADHD is highly heritable, with heritability estimates from twin
studies in the range of 65–90% (Thapar et al., 2001) and average
across 20 studies of 76% (Faraone et al., 2005b). Family studies
report that both parents and siblings of a child with ADHD are
around four to five times more likely to have ADHD than the
general population (Faraone et al., 2000). Although ADHD is
diagnosed using operational diagnostic criteria, measures of
ADHD symptoms are continuously distributed in the general
population. Mathematical modelling using De Fries and Fulker
(DF) analysis supports the hypothesis that ADHD is the extreme
of behaviours that vary genetically throughout the entire popu-
lation (Gjone et al., 1996; Levy et al., 1997; Willcutt et al., 2000;
Price et al., 2001). These data suggest that genetic influences on
ADHD are distributed throughout the population and correlate
with quantitative trait measures of ADHD symptoms. This is
important since this implies that ADHD is best perceived as a
quantitative trait, or series of quantitative traits, rather than a cate-
gorical disorder. The implication for clinical practice is that,
similar to other common psychiatric disorders such as anxiety and
depression, appropriate clinical cut-offs need to be established that
link ADHD symptoms to significant clinical impairments; given
that ADHD is a developmental disorder, this will require estab-
lishment of age- and gender-specific norms. A particular issue for
the diagnosis of ADHD in adults is that the age-appropriate
expression of clinical symptoms has yet to be firmly established.
Categories of evidence and strength of recommendations
Categories of evidence for causal relationships and treatment
Ia:Evidence from meta-analysis of randomized controlled trials
Ib: Evidence from at least one randomized controlled trial
IIa:Evidence from at least one controlled study without randomization
IIb:Evidence from at least one other type of quasi-experimental study
III:Evidence from non-experimental descriptive studies, such as comparative studies, correlation studies and case-control studies
IV:Evidence from expert committee reports or opinions and/or clinical experience of respected authorities
Categories of evidence for observational relationships
I:Evidence from large, representative population samples
II: Evidence from small, well-designed, but not necessarily representative samples
III:Evidence from non-representative surveys, case reports
IV: Evidence from expert committee reports or opinions and/or clinical experience of respected authorities
Strength of recommendation
ADirectly based on category I evidence
B Directly based on category II evidence or extrapolated from category I evidence
CDirectly based on category III evidence or extrapolated from category II evidence
DDirectly based on category IV evidence or extrapolated from category III evidence
SStandard of clinical care
ADHD management in adolescents to adults
Genetic influences on ADHD are likely to be the result of mul-
tiple genes of small effect size and are expected to interact with
environmental risk factors (Asherson et al., 2005). Four linkage
scans have been completed and highlight a number of potential
chromosomal regions containing genes that increase risk for
ADHD, although there is as yet no clear consensus across the
various data sets, and no genes have been identified that account
for linkage signals (Fisher et al., 2002; Smalley et al., 2002;
Bakker et al., 2003; Ogdie et al., 2003; Arcos-Burgos et al., 2004;
Ogdie et al., 2004; Hebebrand et al., 2005). Genetic association
studies have focused on the analysis of monoamine system genes,
due to the marked and rapid response of ADHD symptoms to stim-
ulants that block the reuptake of dopamine and norepinephrine (see
later). Several genes have been reported to be associated with
ADHD in multiple studies, with small but significant effects con-
firmed by meta-analysis (Asherson et al., 2005; Faraone et al.,
2005b, Thapar et al., 2005a). The best evidence for association is
with DNA variants of the dopamine D4 (DRD4) and D5 (DRD5)
receptors and the dopamine transporter (DAT1), with strong addi-
tional evidence for associations with the serotonin IB receptor
(5-HT1B), serotonin transporter (SERT) and synaptosomal associ-
ated protein-25 (SNAP-25). Gene–environment interactions have
been reported between DAT1 and maternal use of tobacco (Kahn
et al., 2003) and alcohol (Brookes et al., 2006) during pregnancy
on the risk of ADHD, and between catechol O-methyltransferase
and low birth weight on risk for antisocial outcomes among ADHD
cases (Thapar et al., 2005b). Although most studies have focused
on children with ADHD, it is interesting to note that the only two
studies to investigate the DRD4 association in adults with ADHD
were both positive (Muglia et al., 2000; Lynn et al., 2005).
Neuroimaging studies (Bush et al., 2005) have documented
significant overall and regional reductions in white matter volumes
between unmedicated ADHD patients, including those never
exposed to relevant medication, and healthy controls (Castellanos
et al., 2002). Medicated patients differ from unmedicated patients,
but not from healthy controls, in overall and regional (frontal,
parietal and temporal) white matter volumes. These studies have
clearly demonstrated differences in the anatomy and processing of
patients with ADHD compared with controls, although these dif-
ferences do not have the sensitivity or specificity in individuals to
contribute to differential diagnosis (Seidman et al., 2005).
Several neuroimaging studies, though not all (van Dyck et al.,
2002), have indicated that striatal DAT binding is increased com-
pared with controls in ADHD patients (Krause et al., 2000;
Madras et al., 2002; Spencer et al., 2005a; Volkow et al., 2005)
including studies in adults (Dougherty et al., 1999; Dresel et al.,
2000; Krause et al., 2000), suggesting increased density of the
transporter protein in the striatum. In addition, there is evidence
from two out of four studies that the DAT1 risk genotype for
ADHD is associated with increased striatal DAT binding (Heinz et
al., 2000; Jacobsen et al., 2000; Martinez et al., 2001; Cheon et
al., 2003). Moreover, other groups have reported that
methylphenidate treatment, which
dopamine in the human striatum (Volkow et al., 2001), is accom-
panied by reductions in DAT binding to control values (Dresel
et al., 2000; Spencer et al., 2005a; Volkow et al., 2005).
Detailed coverage of the diverse literature on the influence of
environmental factors was beyond the scope of the meeting and
thus is briefly reviewed in this document. It is noted that environ-
mental factors of significance include pregnancy and birth factors,
prematurity, diet, institutional care and other aspects of the
psychosocial environment (see Biederman, 2005 for a recent
review). Several studies also suggest that childhood environmental
factors shape the phenotypic expression of ADHD, and co-
morbidity with other behavioural disorders, by acting on genetic
influences (e.g. Hudziak et al., 2005; Burt et al., 2005; Dick et al.,
2005; Jester et al., 2005 I–II). Furthermore, predictors of persist-
ence into adulthood include family history of ADHD, psychiatric
co-morbidity and psychosocial adversity (Biederman, 2005). In
terms of treatment, various dietary manipulations can improve
ADHD symptoms in at least some children (Egger et al., 1985;
Carter et al., 1993; Boris and Mandel, 1994; Rowe and Rowe,
1994; Schmidt et al., 1997; Dykman and Dykman, 1998; Richard-
son and Puri, 2002 Ib), although the mechanism of these improve-
ments is unknown. Activities in natural ‘green’ settings have been
linked with symptom improvement in a national sample (Kuo and
Taylor, 2004 I). These and other findings (e.g. that brain abnor-
malities may not determine phenotype (Castellanos et al., 2002
IIa)) could suggest a greater importance of environmental factors
in some individuals, both in childhood and in older age groups.
Environmental restructuring might be particularly important as
part of the management of adults with ADHD, in our experience
(see ‘Psychotherapeutic approaches’ below).
Neuropsychological research, mainly in children, has documented
various deficits in executive function (especially inhibition), moti-
vation and reinforcement, attention, timing, memory and energet-
ics (covered in detail in ‘Neuropsychological assessment’ below
(Doyle et al., 2005)). This research has clarified that although
about a third of patients have difficulties with one or another neu-
ropsychological test of executive functioning, this is neither uni-
versal nor diagnostic. Functional neuroimaging studies have often
suggested that brain activation is abnormal during affected tasks
compared with controls, and specifically is more diffuse than in
controls with failure of inhibitory functions. For example, Tamm
et al. (2004) have linked a key deficit – abnormal response inhibi-
tion on a Go/No Go task – with abnormal activation patterns in
specific temporal regions compared with controls. This has led to
the hypothesis that individuals with ADHD may have to use other
brain regions, when faced with attention demanding tasks, that are
perhaps less efficient (Bush et al., 2005). Recent evidence also
suggests that this is reversible with medication.
Various studies are aimed at linking the specific neuropsycho-
logical traits (known as endophenotypes – e.g. delay aversion,
executive motor inhibition, deficit in arousal/activation regulation)
that may underpin symptoms and behaviours, with neurobiological
changes (e.g. dysfunction in frontostriatal circuitry, reward
systems and catecholamine function (Waldman, 2005; Willcutt et
ADHD management in adolescents to adults
al., 2005)). The effects of the identified genetic variants on brain
functioning and resultant phenotypic traits remain unclear, so the
possibility remains that such variants could contribute to several
phenotypes of which one is currently classified as ADHD.
Finally, neuropsychological deficits neither consistently predict
behaviours nor outcome of treatment (e.g. Rhodes et al., 2004,
2005). This could reflect a lack of specificity of the deficit tested
to the diagnostic group under study or the fact that treatment effi-
cacy is not mediated via those structures and functions involved in
the causative pathway(s) of ADHD.
Several major prospective follow-up studies have documented the
natural history of ADHD, especially as patients transit through
adolescence into adulthood (Hechtman and Weiss, 1983; Barkley
et al., 1990; Weiss and Hechtman, 1993; Biederman et al., 1996a;
Biederman et al., 1996b; Faraone et al., 1996; Biederman et al.,
1998; Mannuzza et al., 1998; Mick et al., 2004; Faraone et al.,
2006). These studies suggest that a large proportion of child
patients show ADHD symptoms that persist to adolescence (esti-
mated variously at 50–60%) or adulthood (10–66%; Hill and
Schoener, 1996; Spencer et al., 1996; Faraone et al., 1996; Bieder-
man et al., 2000; Barkley, 2002; Faraone et al., 2006). A recent
meta-analysis of studies that followed children identified with
ADHD and matched controls into adulthood has been completed
(Faraone et al., 2006). When the definition of ADHD included
only those that met full diagnostic criteria for ADHD, the rate of
persistence was approximately 15% at age 25 years. However the
rate was far higher, approximately 65%, when individuals fulfill-
ing the DSM-IV definition of ADHD in partial remission were
included, referring to the persistence of some symptoms associ-
ated with significant clinical impairments. Prevalence rates of
adult ADHD estimated at around 4% have been found in several
different studies, using DSM-IV criteria, but will vary depending
on clinical measures used and the precise way in which the diag-
nostic criteria are applied (Hill and Schoener, 1996; Gallagher and
Blader, 2001; Bloom and Dey, 2004; Faraone and Biederman,
2005; Kooij et al., 2005; Kessler et al., 2006; Faraone et al.,
While some symptoms (e.g. hyperactivity) tend to diminish or
present differently with age, perhaps due to self-medication, adap-
tation and neurodevelopment, other symptoms, especially inatten-
tion, (Millstein et al., 1998) are more likely to persist and may
have a greater impact on adults. Community follow-up of young
people with untreated ‘hyperactivity’ suggests that impulsiveness
declines in absolute terms, but remains deviant relative to that of
age-matched peers (Taylor et al., 1996). These studies have inves-
tigated the frequency of these symptoms in adults but no study to
date has evaluated the relative impairment caused by continuation
of the various symptoms that do remain in the adult population.
Attentional difficulties are likely to persist at least in their impact
on functioning. Symptoms diminish in frequency and severity as a
consequence of natural developmental processes seen in all chil-
dren and may further diminish, at least in part, due to learned
skills, coping strategies and environmental restructuring.
However, the overall impact may worsen due to the increased
demands of an adult environment. While the full complement of
diagnostic deficits can often disappear in adulthood, at an age-
related rate dependent on criteria, impairment persists and may
Consensus: ADHD as a neurodevelopmental condition
ADHD is a neurodevelopmental psychiatric condition that most likely results from the interaction of multiple genetic and
environmental factors, each of small effect, with different patterns creating multiple pathways to symptoms each marked and
mediated by different deficit profiles (B).
ADHD is currently understood to be a life-long condition and currently a diagnosis of adult ADHD needs to include childhood
impairment (either prospectively or retrospectively) (C). The status of disorder of later onset needs to be established.
In the absence of specific markers common to the entire group of ADHD patients, assessment and treatment are guided by phe-
Co-morbidity is common in both childhood and adulthood, and may determine outcomes (D). Clinical assessment of ADHD
needs to include careful evaluation for other disorders.
Expression of ADHD and co-morbidities is highly heterogeneous, thus management needs to be individualized (C).
A better aetiological theory is needed that accounts for the causal heterogeneity in the condition (Coghill et al., 2005; Sonuga-
Is ADHD a categorical difference or the end of spectrum of a population trait?
How does phenotype match to genetic, neuropsychological and neuroimaging markers?
What are the best ways to subdivide ADHD into subtypes?
What underpins the relation between ADHD and environmental factors, such as diet and sleep?
How many adults with adult ADHD currently receive alternative diagnoses and treatments within adult psychiatric and
primary care services?
Should screening of parents and children of referred patients be considered, and what would be the resource implications of
ADHD management in adolescents to adults
worsen in a subset. In such patients, impairment is usually pervasive
or affects more than one domain of activities and may manifest as:
educational, organizational or occupational failures; substance use
disorders and other dependent, risky, antisocial or forensic behav-
iours; emotional and relationship difficulties and increased medical
morbidity (Millstein et al., 1998; Swensen et al., 2000).
Co-morbidity is widely considered to be a common finding in
adolescent and adult ADHD patients, and will affect outcomes
(see ‘Comorbid disorders and special situations’). Epidemiological
and clinical studies have shown that more than 80% of adults with
ADHD will have another disorder, and the finding that co-morbid-
ity in adults is similar to that in children, but increases with time,
raises the question of whether some of this co-morbidity may be
driven by untreated ADHD. Common co-morbid disorders in
adulthood include anxiety, depression and antisocial behaviour
disorders, and persisting neurodevelopmental disorders such as
dyslexia (Heiligenstein and Keeling, 1995; Biederman et al.,
1996b; Marks et al., 2001; Biederman, 2004; McGough et al.,
2005; Kessler et al., 2006). The frequency of self-medication and
substance use disorders is thought to increase with age but no
studies have confirmed this supposition.
Diagnosis and assessment
Several issues arise when considering the diagnostic process for
adult ADHD. Two major sets of criteria are used worldwide to
diagnose conditions characterized by inattention and/or hyperac-
tivity – the Diagnostic and Statistical Manual of Mental
Disorders, fourth edition text revision (DSM-IV-TR, APA, 2000),
which recognizes ADHD, and the International Classification of
Diseases, tenth edition (ICD-10, WHO, 1992), which encom-
passes hyperkinetic disorders. Due to differences in criteria, each
system identifies different clinical entities with different names –
ADHD (APA, 2000) or hyperkinetic disorders (WHO, 1992); the
latter generally describes a more restricted and severe subset of
DSM-IV combined-type ADHD. We have chosen to use the term
ADHD as shorthand for ADHD and hyperkinetic disorders, except
where the two are contrasted.
In the absence of any definitive and objective test (e.g. genetic
or neuropsychological), the purpose of the diagnostic process
needs consideration. Currently, diagnosis is a process of identify-
ing extreme behaviour that requires and is amenable to profes-
sional help. However, this process can be used to validate
observers’ views of extreme behaviour rather than to identify
markers of an underlying pathological process, with manifesta-
tions that map to the label of ADHD. As in most areas of medi-
cine, ADHD is diagnosed clinically. The only valid ‘test’ for
ADHD is the use of rating scales that have been normed on large
populations, and can identify whether the child is affected by this
disorder, other disorders and functional impairment compared
with age- and gender-matched peers. The use of rating scales com-
bined with a developmental history, observation, family and other
risk factors, and impairments consistent with the disorder allows a
high level of diagnostic certainty between clinicians and predict-
ing response to treatment (Barkley, 2006). Treatment response is
not considered sufficient to make a diagnosis of ADHD; although
the criteria for hyperkinetic disorders better predicts treatment
response than the criteria for ADHD, treatment response varies
between individuals, while other disorders can respond to current
The two diagnostic systems have various limitations, not least
the internal tautology that the definition of ADHD or hyperkinetic
disorders rests solely on the criteria that have been established.
For example, the possibility exists that boys are diagnosed with
ADHD more often than girls because the diagnostic criteria relate
to disruptive behaviours as markers that are more prevalent in
male children. Importantly, these systems do not include associ-
ated symptoms, reported by patients which could arguably be the
main target of treatments.
Although DSM-IV-TR and ICD-10 have criteria to diagnose
childhood ADHD, these have not been adapted to be appropriate
to adults – e.g. there is no age adjustment to the criteria appropri-
ate to adults. The associated symptoms are similar in the two clas-
sifications and full criteria are given in the Appendix.
following criteria for a diagnosis of ADHD:
The DSM-IV-TR (APA, 2000) requires all the
Either at least six of nine symptoms of inattention or at least
six of nine symptoms of hyperactivity/impulsivity persisting
for at least 6 months to a maladaptive degree, inconsistent
with developmental level. DSM-IV-TR allows for ADHD ‘in
partial remission’ that can usefully be applied to adults who
had ADHD and have persistence of some of the symptoms
associated with continued clinical impairments, but who no
longer fulfil the full criteria.
Some symptoms that caused impairment were present before
age 7 years.
Some impairment from symptoms is present in at least two
settings (e.g. at home and at school/work).
Clear evidence of clinically significant impairment in social,
academic or occupational functioning.
Symptoms do not occur exclusively during a pervasive
developmental disorder or psychotic disorder and are not
better accounted for by another mental health disorder.
DSM-IV-TR distinguishes three subtypes of ADHD on the basis
of clinical symptoms: (a) predominantly inattentive type, (b) pre-
dominantly hyperactive/impulsive type or (c) combined type
(presence of inattention and hyperactivity/impulsivity). These sub-
types are defined by the absence of a six of nine symptom cut-off
in any one of the domains, although the patient may still have
significant symptoms (five of nine), impairment, or be markedly
discrepant for age and gender norms in the area in which they fail
to meet this cut-off criterion.
general criteria and the diagnostic criteria for research (DCR),
Two types of WHO criteria exist (WHO, 1992) – the
ADHD management in adolescents to adults
which are more specific than the general criteria and closer to the
DSM-IV-TR criteria. The diagnostic basis of ICD-10 hyperkinetic
disorders is the existence of both impaired attention and overactiv-
ity evident in more than one situation (e.g. home, educational set-
tings, clinic). The presence or absence of conduct disorder
constitutes the basis for the main subdivision of hyperkinetic dis-
orders – disturbance of activity and attention versus hyperkinetic
For the general ICD-10 diagnosis, behavioural problems
should be longstanding and have started before age 6 years. The
caveat is made that due to wide normal variation in activity, only
extreme degrees of hyperactivity should lead to a diagnosis in pre-
school children. Associated features are not necessary or sufficient
for the diagnosis but help sustain the disorder, including; disinhi-
bition in social relationships, recklessness in situations involving
some danger, and impulsive flouting of social rules (as shown by
intruding on or interrupting others’ activities, prematurely answer-
ing questions before they have been completed or difficulty in
The ICD-10 DCR recognizes seven criteria:
Demonstrable abnormalities at home: at least three of five
problems of inattention, at least three of five problems of
hyperactivity, at least one of three problems of hyperactivity.
Demonstrable abnormalities at school or nursery: at least two
of four attentional problems and at least three of five activity
Directly observed abnormality of attention or inactivity.
Does not meet criteria for pervasive developmental disorders,
mania, depressive or anxiety disorder.
Onset before the age of 7 years.
Duraction of at least six months.
IQ above 50.
Thus the DSM-IV-TR criteria identify a broader group of patients
than the ICD-10 (Tripp et al., 1999). No definitive system exists to
diagnose adult ADHD, while use of these childhood diagnostic
systems in adults raises various difficulties, not least the alteration of
symptoms in adulthood. Thus, the consensus group thought it would
be helpful to present an extended checklist of adult symptoms in this
document, based on childhood diagnostic symptoms plus additional
adult symptoms (APA, 2000; WHO, 1992; Wender, 1995).
diagnose adult ADHD, including three self-report scales. The 61-
item Wender Utah Rating Scale focuses on retrospective symp-
toms in childhood plus current hyperactivity, inattention and other
symptoms. It relies on retrospective recall by the individual but
has been validated against parent report and found to be reliable
(Ward et al., 1993). Others include the Adult Self Report Scale
(Adler et al., 2004; Kessler et al., 2005), Conners Adult ADHD
Rating Scale (Conners et al., 1998), the 40-item Brown Adult
Attention Deficit Disorder Scale (Brown, 1996) and the Barkley
Self, Other and Past ADHD symptom checklists (Barkley and
Murphy, 2006). None of these scales is sufficient for diagnostic
Various rating scales have been developed to help
Proposed BAP extended adult symptom checklist
1 Lack of attention to detail or carelessness
2 Inattention in tasks or activities the patient finds tedious
4 Failure to follow instructions
5Starting many tasks while having difficulty finishing
6 Poor organizational skills
7Avoidance of, dislike of, or inability to expend sus-
tained mental effort
8Losing or misplacing things
9 Ready distractibility
12 Restlessness or an inability to sit still in low-stimulation
13Inappropriate or excessive activity or an internal feeling
of restlessness or edginess
14Difficulty keeping quiet; talking out of turn
15Unfocused mental activity; difficulty turning thoughts
16 Blurting out responses; poor social timing in dialogue
17 Trouble waiting if there is nothing to do
18Interrupting or intruding on others
19Irritability, impatience or frustration
20Affective lability or hot temper
22 Impulsivity or risk-taking in activities
purposes but may be useful in conjunction with a formal clinical
evaluation, or serially to evaluate symptom changes. Several dif-
ferent types of scales may be helpful: screens for adult ADHD
symptoms, screen for co-morbidity (Gadow et al., 1999), and
examination of impact of ADHD on functional impairment (Weiss
and Weiss, 2004). Generic functioning scales have shown poor
correlation with ADHD symptoms, making it clear that ADHD
specific functional scales are needed (Gordon et al., 2006).
Two shorter scales have been developed to allow a rapid initial
screen in various clinical settings, and are based on the 18 symp-
toms listed in both the DSM-IV-TR and the ICD-10 DCR criteria.
The six-item Adult ADHD Self-Report Scale screener (ASRS-
v1.1, Adler et al., 2004; Kessler et al., 2005) is patient-rated while
the ADHD Rating Scale (ADHDRS-IV-Inv, Adler et al., 2005b) is
clinician administered with good reliability and moderate
agreement between raters. The ASRS (available at www.med.
nyu.edu/psych/assets/adhdscreen18.pdf) and the Canadian Con-
sensus screening checklist (www.caddra.ca) are given in the
Appendix. Other scales can be downloaded from www.caddra.ca
Many studies in children but few in adolescents have documented
various deficits of neuropsychological testing. The limited liter-
ADHD management in adolescents to adults
ature in adults reports similar patterns, although marked hetero-
geneity exists (Schoeclin and Engel, 2005; Frazier et al., 2004;
Harvey et al., 2004; Seidman et al., 2004; Nigg, 2005).
Commonly reported impairments are found in tasks involving
executive functions, selective and sustained attention, response
inhibition, working memory and reward-related motivation. Find-
ings on specific tests in various domains are shown in Table 2.
However as suggested above, the traditional neuropsychologi-
cal model of ADHD, which holds that symptoms are the product
of a common core deficit expressed by all affected individuals, is
now contentious (Sonuga-Barke, 2002, 2003, 2005; Castellanos et
al., 2006). The emerging consensus is that ADHD is a neuropsy-
chologically heterogeneous disorder, with different patterns of
impairments seen in different individuals (Sonuga-Barke, 2002,
2003, 2005; Coghill et al., 2005). Alternate theories include a
failure of specific response-inhibition mechanisms (Willcutt et al.,
2005), altered reward or motivational pathways (Luman et al.,
2005), abnormal processing of time-related cues (Toplak and
Tannock, in submission), non-working memory (Rhodes et al.,
2004, 2005) and diffuse abnormalities in energetics (Seargant,
2005). Current thinking is that these impairments may reflect dys-
regulation in distinct underlying neural circuitries, that is: frontal,
(dorsolateral prefrontal cortex), dorsal and medial striatal function
reflecting executive function deficits; cerebellum and its outputs
affecting timing (with modulation by norepinephrine); ventro-stri-
atal-prefrontal (orbito-frontal) circuitry affecting reward and pun-
ishment processing, motivation and delay aversion (with
modulation by dopamine); temporal and amygdalo-hippocampal
Summary of evidence on neuropsychological assessment
Child (approx 6 to 12) Adolescence (approx 12 to 18)Adulthood
Test QualityQuantity Strength
Quality Quantity Strength
Inhibition-based executive functions
Reward and punishment processing
Quality: quantitative review (A); qualitative review (B); no review (C); no studies (D)
Quantity: more than ten studies (A); five to ten studies (B); two to four studies (C); one study (D); no studies (E)
Strength: I=no effect d<0.2; II=small effect d 0.2–0.4; III=moderate d 0.4–0.7; IV=large d 0.7–1.0; V=very large d>1.0
Tests are: Continuous Performance Test commission errors; Continuous Performance Test ommission errors; Stop Signal Reaction Times; Spatial working
memory; Verbal working memory; Tower of London/Hanoi; Trials-B; Stroop interference; Wisconsin card sorting task; Time discrimination; Time
reproduction; Interstimulus interval effects; responses to reward, punishment; Choice for delayed rewards
ADHD management in adolescents to adults
circuitry affecting non-working memory; and distributed circuitry
implicated in abnormal cognitive energetics.
While little is known about the value of neuropsychological
assessment in differential diagnosis in adults and adolescents
(Lovejoy et al., 1999) most studies in children reveal that tests of
executive function and attention show good positive predictive
power but poor negative predictive power (Barkley et al., 1992;
Sharma et al., 1991; Grodzinsky and Barkley, 1999; Rielly et al.,
1999; Doyle et al., 2000; Dickerson Mayes et al., 2001; Berlin et
al., 2004). Thus, use of such tests alone will lead to false-negative
diagnoses (children without neuropsychological executive impair-
ment will still have the condition), which is unsurprising given the
heterogeneity of neuropsychological findings.
The effect of co-morbidities on such tests is not well
researched. General intelligence testing can be useful to relate IQ
to the level of academic and occupational achievement and to
investigate or exclude learning disabilities. However, it raises the
issue of labelling based on IQ testing alone. This is particularly
important for individuals with ADHD since it is possible that IQ
testing results will show increased variation in this group and are
likely to be affected by the attention and motivation of individuals
and/or by underachievement in education. For example, a recent
study of the WISC-IV in 118 children showed the working
memory tests to be the lowest score in all 118 children, thus low-
ering the overall IQ score secondary to a deficit considered spe-
cific to ADHD (Mayes and Calhoun, 2006).
Until the heterogeneity of ADHD is better characterized and an
optimal battery of tests in multiple domains is compiled, the prac-
tical value of experimental neuropsychological assessment may lie
in profiling areas of particular deficits in individual cases, which
may then help to indicate specific solutions to task-related impair-
ments, and then to monitor outcomes. Further, neuropsychological
subtyping of patterns of deficits might shed more light on the
underlying pathological processes and potentially help tailor treat-
ments to phenotype.
Assessment and differential diagnosis
The importance of a full clinical assessment needs to be emphas-
ized, including neurodevelopmental history (Weiss and Murray,
2003, 2004; Biederman, 2005; Asherson, 2005). It may be worth
setting aside any preconceptions of an individual’s diagnosis from
previous contacts with health services and starting from first prin-
ciples, using standard assessment techniques available in psychi-
atric services. The presence of ADHD symptoms needs to be
elicited plus symptoms of differential and co-morbid disorders
(see also ‘Co-morbid disorders and special situations’), and their
impact on various domains of functioning needs to be assessed.
The differential diagnosis of ADHD includes neurodevelopmental
disorders (learning disability, autistic spectrum disorders, commu-
nication difficulties, etc.), anxiety, depression, bipolar disorder,
substance use disorders and personality disorders. Some clinicians
also advocate a full examination including neurology, since subtle
deficits like gaze instability and word-finding defects have been
reported (Weiss and Murray, 2004 IV). The clinical assessment
may generate hypotheses about specific disorders or deficits in an
individual, which can then be explored further with rating scales,
and where indicated neuropsychology. Treatment response is not
thought to be sufficiently specific to form part of the diagnostic
Diagnostic difficulties in adults
may be made under several circumstances. The most straight-
forward scenario is the persistence of symptoms or problem
behaviours in individuals in transition from child and adolescent
mental health services. Individuals that present in adulthood may
self-report symptoms of ADHD or observations of others. ADHD
might also be identified through presentation with a co-morbid
diagnosis such as substance use disorders or a difficulty in a
domain of functioning (e.g. referral from occupational health serv-
ices). The availability of management strategies suggests the pos-
sibility of screening – either in the parents or siblings of those
identified with ADHD or in children and adults with problem
behaviours. This latter strategy might present a potential source of
gender bias since a focus on conduct disorder and related behav-
iours might be one explanation of the high male: female ratio
among children diagnosed with ADHD (Pineda et al., 1999;
Jackson and King, 2004). Some evidence suggests that equal
numbers of women and men are affected with ADHD as adults
(Murphy and Barkley, 1996; Rowland et al., 2002), which raises
the possibility that girls with ADHD are underdiagnosed.
Diagnosis of ADHD in adults raises other issues in addition to
those already identified. The importance of making a diagnosis in
adults is the identification of impairments that are amenable to
treatment. Moreover, being given a diagnosis of adult ADHD may
help individuals understand why their attainment has failed to cor-
relate with that expected of them, but can also reduce self-esteem,
which may itself have an adverse impact on functioning. The diag-
nostic and assessment process is further complicated by differ-
ences in co-morbidities, which may be a more important focus for
management than core ADHD symptoms.
To make a diagnosis of ADHD in adults, current diagnostic
systems require the presence of ADHD symptoms from childhood.
Whilst this is unlikely to be a problem for individuals in transition
from children’s services, presentation for the first time in adult-
hood raises the question of whether ADHD is always a continuum
from childhood deficits or whether adult-onset ADHD can occur.
A view shared by many clinicians is that whilst symptoms are
likely to have been present throughout development, there are
groups for whom these symptoms only become impairing at
particular stages of development, often around significant trans-
itions, e.g. from primary to secondary school or from school to
college or to work. Whilst the current diagnostic frameworks
would not formally recognize these cases as late-onset ADHD,
they will meet all criteria for diagnosis excepting that for onset of
Another requirement for diagnosis is the number of symptoms
reported but, given the natural history of the disorder and other
sources of heterogeneity, many adults may not reach full diagnos-
tic criteria. Furthermore, adults have a greater ability to adapt to
their environment and the demands that this places on them, which
may alter the ways in which symptoms are expressed. However,
A diagnosis of adult ADHD
ADHD management in adolescents to adults
some will still have substantial impairments amenable to treat-
ment, which will not be highlighted by use of diagnostic systems
but rather through a detailed assessment of associated impairments
(Weiss and Murray, 2004). Epidemiological data support the
validity of applying a lower threshold in adults (e.g. four or five
out of nine criteria), as this lower threshold correlates significantly
with impairment (Kooij et al., 2005). The emphasis on functional
impairment has several effects. First, it permits recognition of the
need for treatment in those patients who are impaired but subsyn-
dromal. Second, it raises questions about the value of treatments in
patients with symptoms and no evidence of current impairment,
thus narrowing diagnostic criteria to those who are actually having
Use of the current childhood diagnostic systems (APA, 2000;
WHO, 1992) in adults requires validation of ADHD symptomatol-
ogy from observers since self-reported symptoms are not included.
However, individuals who present in adulthood with ADHD
report their symptoms but may also lack insight into their current
symptoms (Magnusson et al., 2006). In addition, the need to
collect data on possible impairments as a child raises the possibil-
ity of recall bias in the individual, especially for symptoms such as
impulsivity and their impact. Use of informants, if available, to
Recommendations – diagnosis and assessment
Diagnosis in adulthood requires specialist skills – i.e. those used by psychiatrists – but should involve primary care practitioners,
who should be trained to be aware of the diagnosis (D). Assessment includes symptoms (past and present), impairments in different
contexts, influence of changing demands through life, exclusion of differentials, and application of clinical examination, rating
scales and other tools as indicated.
Diagnosis has a major impacts, so the purpose of diagnosis is to identify those who are likely to benefit from treatment options
where these are available (D).
Preferable diagnostic and assessment criteria are an extended checklist based on DSM-IV-TR and ICD-10 and adult symptoms
(D) (see panel and Appendix).
Assessment needs to confirm impairment in different scenarios: (a) a history of childhood ADHD or suggestive symptoms
(impairments/failure of attainment/demands and co-morbidity), (b) current evidence of symptoms leading to pervasive impair-
ment in more than one domain (given context demands and skills) (D).
Multiple informants need to be used where possible (at least one contemporary and one developmental) with consent, espe-
cially for younger patients (D).
Current neuropsychological tests based solely on executive function are likely to be of limited diagnostic value, though test
batteries that assess multiple domains of neuropsychological performance may be useful to determine individual deficits and to
suggest tailored management strategies (D).
Sensitive use of general intelligence tests can be useful to ascertain potential attainment, and to diagnose co-morbid learning
disabilities (D). Interest, reward and educational achievement are important complicating factors (D).
Clinicians need to be aware of the impact of diagnosis and testing on the individual, particularly on self-esteem (D).
Treatment response cannot be used to make a diagnosis of ADHD (D).
What criteria and checklists are required to incorporate better subjective adult symptomatology, to highlight possible neuropsy-
chological deficits and to map treatment responses?
What will be the impact of extending the symptom checklists on sensitivity and specificity of diagnosis?
How can the diagnostic process avoid gender and other biases and match to an underlying pathological process reflected in
symptoms rather than to external complaints?
What is the practical place of neuropsychology (who to test, with what, and when with what value)?
What can neuropsychological subtyping tell us about different neuropsychological traits (endophenotypes) within ADHD and
their relationship with causation, behaviours, co-morbidity and treatment responses?
What are the natural history or neurodevelopmental milestones in different domains of neuropsychological functioning espe-
cially in relation to transitions between childhood/adolescence and adulthood?
Can a diagnostically useful battery of tests that measures across domains of functioning, be compiled to improve diagnosis?
What are the useful thresholds regarding diagnosis to determine whether symptoms will or will not be amenable to treatment?
Since the current subtypes are based on male child norms that are problematic, what further research is needed on subtypes? Of
note, the inattentive subtype includes patients who have never had problems with hyperactivity, but also those who have sub-
threshold hyperactivity and impulsive symptoms that are representative of the combined type. There are very few patients who
are just hyperactive and no research to indicate that these people are impaired or simply not reporting problems with attention.
ADHD management in adolescents to adults
report on current and childhood behaviour requires consent and
could be unreliable depending on the nature of the informant and
their relationship to the individual (e.g. partner, parent, employer,
etc.) but is nevertheless useful.
Physicians identify and treat patients based on symptoms.
However, patients come to clinical practice with the hope of func-
tional change, which is the key goal of management. Change in
psychological functioning and functioning in other domains is
essential since ‘pills don’t build skills’ and treatment of inattention
and other core symptoms is not beneficial if patients have nothing
to do. A package of care needs to be developed from available
options on an individual basis after a thorough assessment (Weiss
and Murray, 2004; Asherson, 2005, IV).
Drug treatment in childhood
Drug treatment for children with ADHD has substantially
increased in the past decade, in part due to increased recognition
of the disorder and perhaps also due to changing views on the
impairment thresholds for drug treatment. In the UK, GP prescrib-
ing of methylphenidate has increased sixfold (PPA data www.
rates remain about 20 times lower than in the USA for both pre-
scribing and treatment. Therapeutic agents licensed in the UK and
comparative costs are listed in Table 3, and are all used in primary
care. These agents plus mixed amfetamine salts (Adderall) are
generally available elsewhere. Safety alerts have been issued over
atomoxetine use in childhood, concerning rare hepatotoxicity (one
possible case, one probable case in 3.8 million) and, recently,
increased suicidal thoughts (five cases out of 1357 cases) or
behaviour (one case). All of these rare events are below the preva-
lence of similar events in the general population.
An appraisal for the National Institute for Clinical Excellence
(NICE, 2006 Ia) has identified 65 acceptable trial reports on the
first-line agents, although most are of low quality, and the hetero-
geneity of measures, methods and participants has precluded a
Agents licensed in children (data from the British National
GenericProprietary Annual cost
40mg once daily=£712
40mg twice daily=£1424
Note – prices may change and the dosages given are not necessarily
meta-analysis. Qualitative assessment suggests that all agents are
more effective than placebo and have similar efficacy, although
there have been few head-to-head comparisons. Cost-effectiveness
modelling indicates that use of all three agents sequentially is ben-
eficial, and although the order of sequential prescribing is not clear
on a clinical basis, dexamfetamine would be first line on the basis
of cost alone. However, these agents are not equivalent in terms of
side effects (for example, dexamfetamine is considered by many
to have greater abuse potential than methylphenidate), and the
response to different agents varies both between individuals and at
The effects of drug treatment require careful monitoring and
dose adjustment. Scales such as the Clinical Global Adjustment
Scale (Shaffer et al., 1983), the ADHD checklists (Adler et al.,
2005; Barkley, 2006) and Conners’ scales (Conners et al., 1998)
are useful tools for disease monitoring and service audit. Patients
and parents should be questioned for concordance over reported
effects, as well as compliance. Any history of tolerance or diurnal
changes in symptom control should be specifically elicited. Ado-
lescents need to be advised of the potential interaction with recre-
ational drugs, including possibly cannabis, although strong
warnings against substance use might need to be balanced against
the need to engage the patient with treatment. Substance use or
dependence might need to be addressed prior to being able to
intervene effectively in ADHD.
Various questions remain open over prescribing in childhood,
and would benefit from future research. Comparable data on cost
effectiveness of different drugs are lacking, as are data on long-
term effects, including adverse events and effects on different
symptom domains, including broader areas of functioning such as
social adjustment. Data are also lacking regarding the effects of
longer acting preparations on phenomena such as tolerance and
sensitization. The mechanisms of diurnal and long-term changes
in symptoms per se are not well understood nor is the basis of
variation in effective medication dosages, though increasing dose
with increasing body size is well recognized. Better outcome
measures are required to assess broader effects, including quality
of life measures because the QALY is inadequate in this scenario.
Drug treatment of adults
Drug treatment of adults with ADHD is relatively new, having
been prompted by the entry into adult services of people previ-
ously treated in children’s services. However, clinical experience
in specialist adult ADHD clinics suggests that the majority of
adult patients are diagnosed as adults, while most adolescent
patients stop or are taken off medication. While it is not known
whether all findings in children can be extrapolated to adults, psy-
chostimulants have comparable effects in adults as in children and
adolescents (Faraone et al., 2004; Kooij et al., 2004; Spencer
et al., 2005b 1a). Global functioning and adult symptoms such as
unstable mood and ceaseless mental activity are thought to
respond as well to psychostimulants as do other core symptoms
Prescribing and monitoring strategies are no more complex
than those used in the management of children, and can be based
ADHD management in adolescents to adults
on usual approaches utilized in adult services. However, a rigor-
ous approach to diagnosis is warranted since adult patients are per-
ceived to present commonly with and be treated for anxiety and
depression either as the primary disorder or as part of a personality
disorder. It is important that such patients are correctly diagnosed
and treated within adult psychiatric services, since symptoms can
respond to treatment. Moreover, specialist services are able to
monitor treated patients in a similar manner to those with other
conditions, for compliance, response and potential adverse effects,
including physical side effects like hypertension and weight loss,
and effects on co-morbid symptoms. The best outcomes for moni-
toring response within trials and for individuals are not known but
may include ratings of core and co-morbid symptoms and effects
on various domains of functioning. In general, compliance may be
better in adults than in children, although the reverse may also be
true. Patients may discontinue treatment due to disorganization,
difficulty with persistence, negative and uninformed media
(leading for example to employer prejudice), fears over
dependency, mistaking the treatment as the cause of the stigma of
the disorder, and lack of knowledge on other long-term effects.
Atomoxetine is licensed in the USA for the treatment of adult
ADHD and in the UK for the treatment of adults who have previ-
ously been treated for ADHD as children. No agent is currently
licensed in the UK for adults newly diagnosed with ADHD so pre-
scribing is off-label. The drugs of first choice for the treatment of
adult ADHD are classified as either psychostimulants (e.g.
methylphenidate, amfetamines (Maidment, 2003a)) or non-stimu-
lant (e.g. atomoxetine (Thomason and Michelson, 2004)) (see Table
3). Other non-stimulant agents reported to have some efficacy
include alpha2 adrenoceptor agonists (clonidine and guanfacine),
tricyclic antidepressants, bupropion, modafinil and venlafaxine
(Maidment, 2003b). No efficacy has been found for selective sero-
tonin-reuptake inhibitors, which may be because effective agents
generally act via dopamine and/or norepinephrine as discussed next.
In general, core symptoms are thought to respond better to psychos-
timulants and atomoxetine than to antidepressants, although no
head-to-head studies have been done (Biederman and Spencer,
2001; Maidment, 2003a; Maidment, 2003b).
Mechanisms of drug action
fied act on dopamine and/or norepinephrine neurotransmission,
either as agonists or as reuptake inhibitors with the exception of
modafinil, whose mechanism of action remains unclear (Fone and
Nutt, 2005). Methylphenidate is a potent inhibitor of dopamine
reuptake (Andersen, 1989; Thomason and Michelson, 2004), by
binding to the cocaine-binding site on the DAT, but in vitro data
on reuptake inhibition suggest that it also has a very high affinity
for the norepinephrine transporter (NET). The primary action of
dexamfetamine is inhibition of dopamine reuptake but in addition,
amfetamines can cross the cell membrane by a mechanism
independent of the transporter, and interact with the vesicular
monoamine transporter 2 (VMAT2), thereby displacing vesicular
dopamine and causing the release of newly synthesized intraneu-
ronal monoamine (Ferris and Tang, 1979; Fleckenstein and
Hanson, 2003). Atomoxetine is a relatively selective NET
inhibitor, having approximately a 300-fold higher affinity for NET
All effective agents thus far identi-
than DAT, (Thomason and Michelson, 2004; Gehlert et al., 1995).
Neuroimaging studies in humans have failed to clarify the relative
therapeutic benefit resulting from targeting DAT or NET inde-
pendently. In particular, no high-affinity ligands exist to visualize
the norepinephrine transporter in the human brain making it diffi-
cult to establish the therapeutic importance of NET inhibition.
Nonetheless, imaging studies indicate that methylphenidate binds
strongly to DAT and indirect evidence suggests that this elevates
dopamine levels within an hour when given in therapeutic doses
by oral administration (Volkow et al., 1998; Krause et al., 2000).
These changes are accompanied by an increase in synaptic
dopamine that is dependent on cell-firing rates. Thus, following
methylphenidate administration, a salient stimulus is found to
elicit greater synaptic dopamine levels than a neutral stimulus
(Volkow et al., 2005).
Animal studies show that psychostimulants induce an increase
in cFos-like immunoreactivity consistent with it causing neuronal
activation in the striatum, including the caudate, and in the
mediofrontal cortex (Lin et al., 1996). However, microdialysis in
rats shows that intraperitoneal methylphenidate increases the
synaptic overflow of hippocampal norepinephrine and striatal
dopamine efflux to a similar magnitude (Kuczenski and Segal,
2001). Furthermore, with oral methylphenidate, hippocampal nor-
epinephrine efflux is evoked by lower doses than are required to
elicit nucleus accumbens dopamine efflux (Kuczenski and Segal,
2002). In contrast, atomoxetine causes a selective increase in cFos
within the prefrontal cortex without increasing expression within
the nucleus accumbens or striatum, consistent with it being a
selective inhibitor of NET as indicated from in vitro data (Bymas-
ter et al., 2002). Yet, systemic atomoxetine administration
produced an equivalent threefold increase in prefrontal norepi-
nephrine and dopamine efflux measured by microdialysis (Bymas-
ter et al., 2002), whereas extracellular norepinephrine but not
dopamine is increased in other brain regions, including the hip-
pocampus (Swanson et al., 2006). The prefrontal cortex contains
very low levels of DAT but dopamine and norepinephrine have
very similar affinity for NET, so it is possible that dopamine reup-
take may occur via NET in this brain area. Furthermore, adreno-
ceptor agonists (guanfacine and clonidine), which are thought to
activate prefrontal cortex postsynaptic alpha2 adrenoceptors, have
been shown to improve performance of non-human primates in a
spatial working memory task (Avery et al., 2000).
the lack of head-to-head studies with adequate and unbiased
methodology. In general, all agents with evidence of efficacy in
adults can reduce core symptoms, although effects on different
symptoms and global functioning may vary between agents and
individuals. In the absence of any obvious prescribing hierarchy,
choice of agent may depend on pharmacological factors other than
efficacy (particularly abuse potential, side-effect profile and toxic-
ity in overdose), as well as individual factors (such as patient
choice and co-morbidity). The effect of genotype on treatment
response is unclear, although a poor response to methylphenidate
has been reported in children homozygous for the less common
nine-repeat DAT1 genotype (Stein et al., 2005).
Discussions on clinical efficacy are limited by
ADHD management in adolescents to adults
In children and adolescents, over 200 trial reports indicate
around a 70% short-term response rate with methylphenidate treat-
ment (Smith et al., 2000 Ib; Schachter et al., 2001). A recent
meta-analysis of several adult trials of methylphenidate indicates
that response rates and effect size of treatment in adults is compa-
rable with, although somewhat lower than, those in children
(Faraone et al., 2004 Ia). Notably, greater response rates were
identified by physician ratings compared with self-report, and in
global functioning rather than core symptomatology. Long-acting
preparations are reported to have similar response rates and effect
sizes to standard methylphenidate (Santosh and Taylor, 2000 Ib).
It should be noted, however, that the long-acting preparations may
improve compliance, minimise abuse, diminish rebound symp-
toms and address impairment later in the day, which has been
shown to lead to preferential effectiveness in children (Steele
et al., 2006).
The meta-analysis found response rates varied between trials
(25–78%); this report and a subsequent trial (76%) indicate that
greater responses are found with higher dosing regimes (1mg/kg)
that are comparable with those effective in children (Faraone et
al., 2004; Spencer et al., 2005b Ia). Other factors found to corre-
late with response include reduced level of functioning (psychi-
atric outpatients versus high functioning academic underachievers)
and psychiatric co-morbidity, although the latter was not con-
firmed in the latest trial (Spencer et al., 2005b).
Similar but more limited findings are reported for dexamfeta-
mine (dextro- or D-amfetamine) (Arnold et al., 1989; Paterson et
al., 1999; Taylor and Russo, 2000, 2001 Ib), Adderall mixed
amfetamine salts (Spencer et al., 2001; Biederman et al., 2005 Ib),
the psychostimulant pemoline (Wilens et al., 1999b Ib), and the
dopaminergic agent bupropion (Wender and Reimherr, 1990;
Wilens et al., 2001 Ib). However, pemoline is associated with
hepatotoxicity (Marotta and Roberts, 1998 III) and has been with-
drawn from use.
Several randomized, double-blind placebo-controlled trials
have confirmed the efficacy of atomoxetine in children and adoles-
cents (Thomason and Michelson, 2004 Ib), and similar findings
with an effect size of 0.35–0.4 have been reported from two trials
in adults (Michelson et al., 2003; Faraone et al., 2005a Ib).
However, effects on overall functioning are debated, while no
long-term efficacy trials have yet been reported, although an
interim analysis in adults has been reported (Adler et al., 2005a).
Thus, its place in clinical practice is not yet fully defined. Other
norepinephrine uptake blocking agents such as desipramine
(Wilens et al., 1996 Ib) have similar effects in trials while open
studies on venlafaxine report comparable responses (Adler et al.,
1995; Hedges et al., 1995; Findling et al., 1996 IIb). It is not
known whether other noradrenergic agents like reboxetine, dulox-
etine and lofepramine have any efficacy in ADHD. Guanfacine is
reported in one small trial to have similar efficacy to dexamfeta-
mine (Taylor and Russo, 2001 Ib), as is the non-stimulant wake-
fulness-promoting agent modafinil (Taylor and Russo, 2001 Ib)
Monitoring and adverse effects
ments, patients should be specifically questioned about efficacy on
core symptoms and in various domains of functioning, as well as
effects on co-morbid symptoms and side effects noted. The find-
ings of the MTA study strongly suggest that active and fairly
intensive monitoring of drug treatments improves effectiveness
(MTA Cooperative Group, 1999). Patients do not always report
psychiatric effects, which they may not realize can be induced by
medication. Therefore, direct questioning about changes in affect,
anger or personality should be part of the follow-up interview. The
physician also needs to ask directly about difficulty with com-
pliance, since this is often problematic in ADHD and may not be
reported spontaneously. Diurnal changes in effects should also be
sought, since these may be ameliorated by adjustment of dose
timing or with longer-acting preparations. In addition, long-term
monitoring of blood pressure, pulse and weight is indicated. All
the treatments for ADHD are associated with mild statistical
increases in blood pressure and pulse which may not be problem-
atic in children, but could be problematic in those with antecedent
cardiac disease, hypertension or those who engage in extreme
sports. Treatment adjustment should be guided by report of symp-
toms and functioning in various domains as well as rating scores.
Optimal outcome is the dose which leads to best functional
During review of drug treat-
Guidance on prescribing
Off-label. Prescribing for adult ADHD is necessarily off-label since no agent is licensed for this indication – although atomoxetine is
licensed for use in adults but only when ADHD treatment was initiated in childhood. The BNF (Joint Formulary Committee, 2005)
states: ‘Unlicensed use of medicines becomes necessary if the clinical need cannot be met by licensed medicines; such use should be
supported by appropriate evidence and experience.’ However, the BNF also states that prescribing medicines outside the recommenda-
tions of marketing authorization alters (and probably increases) the doctor’s professional responsibility. This latter statement might
explain a reluctance to prescribe beyond marketing authorization by clinicians, particularly in primary care. Although controlled evid-
ence for prescribing in adults is not extensive, this consensus statement can be considered to meet the criteria for adequate evidence and
experience in prescribing standard medications to adults with ADHD, when done in the context or with support of specialist psychiatric
services. However, it is noted that supplementary prescribers are not contracted for unlicensed prescribing.
Controlled drugs. Prescriptions for psychostimulants (CD) are subject to prescription requirements, which might be altered in the
future, notably for computer prescribing. Currently in the UK, prescriptions must be indelible, signed and dated by the prescriber
with their address, and must always state in the prescriber’s own handwriting: name and address of patient; the form and strength of
a preparation (e.g. 10mg tablets); total quantity or number of dose units in words and figures (e.g. 900mg=nine hundred milligrams
OR ?90=ninety doses); the dose (e.g. 10mg tds).
ADHD management in adolescents to adults
outcome, balancing adverse effects against benefits for that indi-
Various concerns have been raised over long-term treatment
effects, including the potential for tolerance over time, and
adverse effects of psychostimulants such as psychosis, sensitiza-
tion, dependency and withdrawal reactions (Ashton et al., 2006).
Few data exist to guide clinicians since no long-term treatment
trials have been conducted in adults with ADHD. The long-term
effects were reviewed in an analysis of existing short- and long-
term follow-up studies of stimulants in children with ADHD
(Hechtman and Greenfield, 2003). The conclusions were that chil-
dren with ADHD treated with stimulants for as long as 2 years
continue to benefit from the treatment, with improvements
observed in ADHD symptoms, co-morbid oppositional defiant dis-
order and academic and social functioning, with no significant
problems of tolerance or adverse effects. Long-term, prospective
follow-up studies into adulthood show that stimulant treatment
in childhood has slight benefits regarding social skills and
self-esteem. Long-term adverse effects from stimulant treatment in
childhood regarding adult height or future substance abuse have
not been supported by existing studies.
Substance use is a particular concern, both surrounding the
abuse potential of psychostimulants and the effect of treatment on
co-morbid substance use disorders. Abuse potential is related to
the route of administration and the rate of absorption/bioavailabil-
ity of the drug, and may relate to the resultant rate of dopamine
release, which is higher for dopamine releasers than pure reuptake
inhibitors (Kollins, 2003). Slow-release preparations have less
abuse potential and are less amenable to use through abused routes
such as snorting. They are to be preferred for patients with a
history of, or risk factors for, drug misuse.
Diversion of psychostimulants is reported, in particular for
short-acting preparations. Very few clinical studies have evaluated
the risk/benefit ratio of long-term stimulant medication at thera-
peutic doses on human brain and behaviour (Volkow and Insel,
2003) – imaging studies indicate that cocaine induces a similar
blockade of striatal DAT to that seen with methylphenidate
(Volkow et al., 1999). Circumstantial evidence exists for sensiti-
zation and tolerance, which can be precursors to dependency; in
particular, from baseline data of COMACs (Swanson et al., 2004;
Sonuga-Barke et al., 2005) and from the titration phase of the
MTA (Galanter et al., 2003), in which possible rebound was seen
that might be due to tolerance (Sonuga-Barke et al., 2004).
Co-morbid substance use is common in untreated ADHD (see
also ‘Substance use disorders’ below). Treatment in children has
been repeatedly shown to be linked to a reduction in acquiring
substance misuse, but the effect of treatment at other ages is not
known. A meta-analysis of six trials in children found a 1.9-fold
reduction in incidence of substance use in medicated compared
with unmedicated patients (Wilens et al., 2003 Ia). Notably,
studies of methylphenidate-treated rats have found that treatment
during the equivalent age to human childhood reduces the reward-
related effects of cocaine (Carlezon et al., 2003), while treatment
during the equivalent age to adulthood enhances cocaine reward
(Brandon et al., 2001). However, clinical experience suggests
probable lower use of recreational drugs in medicated adults,
although adolescence might carry a differing vulnerability. Thus,
rigorous studies are needed to determine the long-term effects of
pharmacotherapy in different age groups, particularly on co-
morbid recreational substance use. Research is also needed to
determine whether treatment of ADHD in patients currently using
cannabis and/or alcohol on a daily basis is effective.
Recommendations – drug treatments
Proven drug treatments in children include psychostimulants (A) and atomoxetine (A) as first-line treatments, with imipramine,
which is metabolized to desipramine, (B) and bupropion (B) as second-line treatments and clonidine and guanfacine as pos-
sible adjunctive treatment (D).
Adult patients with ADHD are most likely to present to primary care, but drug treatment is best initiated and optimized by sec-
ondary/specialist services (D).
Drug treatment needs to be chosen and adapted to best fit the individual, including the patient’s preferences and concerns (D).
Meta-analysis of methlyphenidate in adults demonstrates similar drug response effect sizes to that seen in children (A).
The limited evidence in adults suggests that agents that enhance synaptic dopamine have far better efficacy than other treat-
ments for core symptoms – amfetamines, methylphenidate and atomoxetine are all effective but not equivalent, since they have
different actions and hazards (A).
Drug prescribing in adults is usually off-label but clinicians are supported in prescribing by these BAP guidelines (D).
Drug treatment requires regular, preferably structured, monitoring and review (e.g. for dose adjustment). For uncomplicated
cases this should be every 6–12 months.
Clinicians need regularly to assess patients on medication for ADHD and other symptoms, global and specific functioning,
adverse effects, concordance of effects (e.g. between patient, doctor, informants), psychiatric side effects, cardiovascular
effects, compliance and tolerance (daily and long term) (D, S).
Drug treatment should NOT be initiated if the diagnosis is uncertain or benefit is unlikely (D).
Abuse potential is related to drug action and formulation – abuse by patients seems low, but diversion can occur with stimu-
lants for performance enhancement or weight loss (D). Slow-release preparations of these agents or atomoxetine are to be pre-
ferred for patients with a history or who are at risk of drug misuse (D). Controlled studies are required to confirm these
ADHD management in adolescents to adults
The Multimodal Treatment Study in ADHD (MTA Cooperative
Group, 1999) has provided definitive evidence on the role of psy-
chotherapy in childhood, with the conclusion that psychotherapy
provides a small additional benefit when added to drug treatment.
Compared with children, adults are mostly self-referred and thus
are self-selected and generally more motivated. However, adults
have usually left the education system where skills are taught and
may also be affected by lifetime failures and effects of co-morbidity.
Good evidence of the effects of psychotherapy in adulthood is
sparse. Moreover, the usefulness of published studies is limited
due to small numbers, lack of evidence on long-term unmedicated
patients, failure to consistently rate effects beyond core and co-
morbid symptoms, exclusion of patients with extreme symptoms
or co-morbidity, and analysis of completers only.
Clinical experience indicates that general psychotherapeutic
support and psycho-education around the time of adult diagnosis,
treatment initiation and review seems helpful. Aims of psycho-
education are to inform on the condition, natural history and prog-
nosis, to prevent further negative effects on self-esteem or
unrealistic expectations of treatment and to give perspective to
individual neurodevelopmental history. Advice to and screening of
family members may also be initiated at this time. A review of the
psychosocial environment is essential to assess specific limitations
in various domains, the level of executive demands and coping
skills already acquired. Assistance with improving coping skills
and efforts at environmental restructuring may have a powerful
impact on the functioning of adults with ADHD. Group therapy,
where available, may help with social isolation, especially around
the critical period where treatment initiation can enhance
experience of failure and lead to loss of self-esteem, depression
and substance use (Safren et al., 2005a).
In terms of psychotherapy, an initial retrospective notes review
conducted by Ratey et al. (1992 III) found that a dynamic psy-
chotherapeutic approach led to decreased self-esteem and
increased frustration. In a review of 36 patients improved but not
remitted on medication, 85% of those who also had co-morbidity
found benefits of adapted cognitive–behavioural therapy on
anxiety, depression and functioning (Wilens et al., 1999a III).
Two controlled, non-randomized studies are reported. Seventeen
patients given psycho-education and organizational skills teaching
versus no treatment found improvements in organization, attention
and emotional stability but a reduction in self-esteem (Wiggins
et al., 1999 IIa). In 15 patients, dialectical behaviour therapy
showed improvements in ADHD symptoms, depression and func-
tioning compared to control treatment (Hesslinger et al., 2002 IIa).
Three RCTs are published, with some good effect sizes
reported in terms of core symptoms, and providing some insights
into the interaction of psychotherapy with medication. A cognitive
remediation programme using a brain injury model in 44 patients
versus waiting list controls showed moderate effects on core
symptoms and organization with minimal improvement of self-
esteem and anger (Stevenson et al., 2002 Ib). Notably, effects on
core symptoms and organizational skills were maintained at 1 year
while effects on anger were not. A retrospective examination of
the effects of cognitive remediation found no difference between
medicated and unmedicated patients but numbers were small
(Stevenson et al., 2002 Ib).
Weiss and the ADHD Research Group have examined the
effects of individualized problem-focused therapy in 33 patients
on dexamfetamine versus those on placebo (unpublished data Ib).
Problem-focused therapy was found to be more effective in redu-
cing core symptoms and improving functioning in medicated
patients than controls. Modified cognitive–behavioural therapy
plus medication demonstrated greater benefits than psychotherapy
or medication alone particularly on ADHD symptoms, with lesser
effects on anxiety and depression in 32 patients (Safren et al.,
Can we extrapolate from treatment studies in children?
Can drug treatment be continued and monitored in primary care with training and support? With such an arrangement, what
would the workload and resources implications be for both primary and secondary care?
How is individual outcome of drug treatment best assessed?
How do classes of drugs compare with each other in head-to-head comparisons and with psychotherapeutic approaches in
What factors underlie individual differences in drug response, tolerance and other side effects and how is treatment best
matched to individuals and types of impairments?
What is the effectiveness of various drug treatments beyond clinical ADHD trials?
Do different drug treatments have differential effects on different underlying deficits (endophenotypes)?
What are the effects of different drug treatments on symptoms beyond ADHD core (e.g. functioning/impairment), risk–benefit
analysis and tolerance?
What are effects of long-term drug treatment and how feasible are long-term follow-up studies given likely high dropout rates?
What are better measures of effectiveness and cost effectiveness for drug trials?
What are the important interactions with other prescribed psychoactive medications (e.g. SSRIs) and with recreational drugs?
When are drug treatments best avoided (e.g. women of child-bearing age, certain age groups, patients with nothing to do)?
What to offer drug non-responders?