Rationale for use of medications for ADHD

Although the research evidence suggests that medication is effective alone, and may be the most effective part of comprehensive multimodal management (Wilens and Biederman 1992; Greenhill 1992), it is the general consensus that educational and behavioral strategies add to the success of management and are essential if medication is ineffective.

There is considerable pressure to treat ADHD with its disruptive symptoms, associated learning, behavioral and emotional problems, family stress, and possible persistence into adolescence and adulthood. For a minority, the outcome is antisocial behavior, criminality and substance abuse.

The multiplicity of aetiology, heterogeneity of presentations, changes over time, and intervention and range of possible treatments, make management complex and confusing.

Approaches to diagnosis and treatment are not equally validated and support is compromised by the lack of, or long waiting lists for, support services. This context emphasizes the use of medication which can have powerful short-term benefits for disrupted behavior and performance.

The need for medication and its effectiveness is relative to the nature and severity of problems and the use of other interventions. A multimodal approach, especially with educational and behavioral supports, should be used if available.

Although medication is the most effective short-term treatment for the disruptive behaviors of ADHD, other approaches may add to the success of medication and be essential if medication is ineffective.

Comprehensive assessment and management is emphasised in managing ADHD. Day-to-day support for the vulnerable individual at home, and in other settings, should be provided. Management is, however, complex and time-consuming and requires collaboration.

Medication, whilst the best validated of the various interventions, is likely to be better accepted when accompanied by advice regarding other supports. Referral to supports should be vigorously pursued though other services may be scanty (Hazell, McDowell, Walton et al 1996).

The prescribing of medication is exclusive to the medical practitioner, but few can provide intensive, prolonged behavioral and emotional management. ADHD usually requires, among other services, psychological or psychiatric support.

North American practice and research dominates pediatric psychopharmacy, particularly in ADHD, reflecting the prevalence of disruptive behaviors responding to medication.

The stimulants are methylphenidate, which is most studied, and dexamphetamine, which is less so, and other medications are used (Werry 1994). A very recent extensive review by Spencer, Biederman, Wilens et al (1996) describes medication in the treatment of ADHD for children and adults.

This refers to 155 controlled studies in over 5700 individuals documenting the efficacy of stimulant medications. Other authors have stated that stimulants are safe and effective drugs (Gadow 1992) and have few side-effects when used in children under proper medical supervision (Werry 1988).

The place of medication may decrease in importance as other vulnerabilities in the child or the child’s environment are dealt with, and as the child moves away from the threshold of the disorder.

Indeed, stimulant medications have been shown to have similar types of effect in children with diagnosed ADHD and individuals regarded as normal controls (Peloquin and Klorman 1986; Rapoport, Buschsbaum and Monte 1980; Rapoport, Buschsbaum and Zahn 1978).

These results emphasise that the diagnosis of ADHD cannot be determined by a positive response to medication.

Stimulant Medication

Pharmacology and stimulant use

Medications for many disorders act on immature or inefficient neurotransmitters, which are localized or distributed throughout the brain, and which have widespread effects on performance and behavior (Gadow 1992; Rogeness, Javors and Pliska 1992; Werry 1988).

There have been reports of mutations in dopamine transporter genes, which may predispose to ADHD (Cook, Stein, Krakowski et al 1995) or in receptor genes (Ebstein, Novick, Umansky et al 1996).

There is, however, no single neurocognitive defect that accounts for ADHD or which can be corrected by a single medication or intervention (Matochik, Liebenauer, King et al 1994).

Dexamphetamine and methylphenidate (Ritalin¨) act on dopaminergic and noradrenergic neurotransmitter pathways and appear to influence mainly prefrontal, frontal and limbic systems with benefits on disruptive behavioral inhibition, impulse control, selective attention, active working memory and executive functioning.

There is no ‘paradoxical effect’ of ‘stimulant’ appearing to ‘sedate’ disruptive behavior. There are no direct effects on consciousness or moral judgement (Greenhill 1992; Wilens and Biederman 1992).

Though similar in clinical effects, the stimulants differ in how they increase neurotransmitter concentrations in the synapse. Dexamphetamine appears to release newly synthesised dopamine and block uptake postsynaptically, while methylphenidate releases stored dopamine (Shenker 1992).

There is wide individual variation in metabolism and effect. Generally, both medications have clinical effects within about 30 minutes and benefits wane after about three hours (though dexamphetamine falls to half its concentration in up to 11 hours compared with the methylphenidate half-life of up to three hours).

Slow release preparations of stimulants used in the United States, or longer-acting agents such as pemoline (Cylert¨), are not generally available in Australia, though they offer the theoretical advantages of less fluctuations in effect than with repeated short doses, and remove the need for dose administration in school.

Their prolonged action, however, may be less intense and their use forfeits the advantages of flexibility and control of titrating that more frequent doses allow (Pelham, Greenslade, Vodde-Hamilton et al 1990).

There is increased experience with combined slow release and fast acting preparations to produce optimum dosage and sustained effects throughout the day.

Extensive research and clinical experience has been robust since early reports (Bradley 1937, 1950; Bradley and Bowen 1941). Many studies involve double-blind placebo-controlled randomised confirmation of stimulant effects.

Reviews of stimulant medication and ADHD include those by Wilens and Biederman (1992), Greenhill (1992), Weiss (1992), Barkley (1994), Shaywitz and Shaywitz (1992), Klein and Wender (1995), and Spencer, Biederman, Wilens et al (1996).

Relevant neurobiological, cognitive and physiological research is reviewed by Ernst and Zametkin (1995), Shenker (1992) and Malone, Kershner and Swanson (1994).

Context of evidence for stimulant effectiveness

Extensive information published over the last twenty years attests to the efficacy and safety of stimulant medication in disruptive behaviors. Older studies claiming only limited efficacy and frequent side-effects have not been replicated with modern understanding and multimodal management of ADHD.

When evaluating this research and extrapolating scientific group data to clinical practice, each of the factors discussed below must be taken into consideration.

Source of research

Most research is from North America, particularly the United States. Early experience was with benzedrine (a racemic mixture of D- and L-amphetamine) and dextroamphetamine, for several hundred patients (Bradley 1950).

The majority of studies evaluate methylphenidate, as dexamphetamine is much less used in the United States where prescribing and manufacturing are more restricted and associations with ‘street speed’ and other substance abuse are strong.

As Australian practice relies much more heavily than does the United States on the use of dexamphetamine rather than methylphenidate, the results of United States research into long-term effects of stimulant medication need to be supplemented by specific studies into the long-term use of dexamphetamine.

Choice of stimulant

The choice of which stimulant to use will depend on individual response.Dexamphetamine and methylphenidate are not identical in pharmacokinetics, clinical benefits and side-effects and one may suit an individual better than the other.

Evaluation must be thorough to determine true non-response in an individual (Arnold 1996). Very few studies directly compare stimulants in the same individual; those reported only involve a total of about 100 children.

Elia, Borcherding, Rapoport et al (1991) attest to a 92 per cent success rate if both are used and dosages thoroughly evaluated.

Some of the same authors compared dexamphetamine, methylphenidate and placebo at different doses in treating ADHD and comorbid Tourette Syndrome, again showing the difference between the stimulants (Castellanos, Giedo, Elia et al 1997).

The relative efficacy and side-effects of dexamphetamine and methylphenidate have been compared in an Australian crossover study which confirms that these vary between individuals (Efron, Jarman and Barker 1997).


The doses used have varied greatly between studies, which have not always been explicit about whether the dose is titrated by tablets, by weight per dose or weight per day.

The lack of any scientific evidence that body weight predicts stimulant response (Rapport, DuPaul and Kelly 1989) adds to this confusion. (Australian statutory prescribing controls limit daily dosage by weight to discourage excessive and toxic doses).

In early research, larger doses were titrated to maximum benefit limited by toxicity whereas, in recent studies, doses have been more conservative in the face of concerns about toxicity and public controversy.

Though linear dose response curves were confirmed in most studies, the preparation only in the form of tablets (dexamphetamine 5 mg and methylphenidate 10 mg) makes it difficult to titrate individual doses for graded responses or side-effects in research or clinical practice.

Doses repeated in the day may have cumulative benefits or adverse effects not evident in studies of single doses.

Anecdotal evidence suggests that adolescents and adults respond to lower doses for their size, though the extent of symptoms, responses and co-morbidity may make prediction and analysis more complex.


The methods used have varied greatly within and between studies. Differences have involved different subject and control populations; study environments; nature, degree and method of defining a ‘response’; dose; and, balancing of benefits with side-effects.

Many double-blind placebo studies employ single doses which evaluate only immediate drug responses, often with single measures which differ greatly. In addition, those who respond to placebo are rarely screened out before the definitive study.

Severity of symptoms on study entry is rarely stated or controlled. The settings are often research classrooms or laboratories, rarely naturalistic ones of home, school, playground or work.

Comparisons of effect with different ages, levels of ability or coexisting neurological problems rarely apply a similar method of dosage and evaluation (Mayes, Crites, Bixler et al 1994). Clinician, teacher or parent may rate response, but not concurrently and perhaps not even while medication is still active.

This may particularly confound comparison between different treatments. Statistical interpretation of correlations or associations between comparisons of different observers, ratings and instruments must be appropriate (Bland and Altman 1986).


Most research has involved primary school-aged students. Wilens and Biederman (1992) record only four studies with preschoolers (a total of 130 children), three of them published around twenty years ago.

They also record, from 1983 to 1991, six studies involving adolescents (113 subjects) and five studies, before 1991, reporting experience with 137 adults.

Mayes, Crites, Bixler et al (1994) claim the first single study with consistent treatment to analyse the relationship between methylphenidate response and age, IQ and other neurodisability.

Type of ADHD studied

There is much less study of ‘nonhyperactive’ children, most recently classified as inattentive type ADHD (DSM-IV).Most research has been carried out on the disruptive behavior end of the ADHD spectrum. Previous studies may have encompassed mixed populations.

There is more subtle language impairment related to learning disabilities and more co-morbid anxiety and depression in the inattentive type. Bauermeister, Alegna, Bird et al (1992) studied a community sample of 614 children whose teachers distinguished different types of ADHD and rated more difficulties in language, memory and reading in the inattentive type.

Further reviews are written by Cantwell and Baker (1992a,b) and Barkley, Grodzinsky and DuPaul (1992).

The recognition of inattentive type ADHD, particularly in girls, and its overlap with traditional designations of ‘learning disability’, pose clinical and conceptual challenges.

Various experts espouse different degrees of efficacy with stimulant treatment of ‘behavioral’ and ‘cognitive’ models of behavioral disinhibition and inattention.

Other factors


Reports may not distinguish between behavioral nonresponse and adverse side-effects, often only recording ‘successful’ treatment. Different symptoms and qualities may respond differently to particular medications and dosage.

Whether ADHD is pervasive at school only or home only, may be significant for mechanisms, management and outcome (Schachar, Tannock, Marriott et al 1995).


Changes in the environment and in the individual may be separate from or closely influenced by medication responses. Neither research studies, nor clinical practice, can control all intrinsic and extrinsic variables which protect from or push towards the threshold of disorder.


Very few medication studies extend beyond a few weeks or include thorough follow-up (Greenhill 1992). Extended treatment studies have been reviewed by Schachar and Tannock (1993). Longer-term observations are anecdotal, with no sustained placebo-controlled studies.

There have been no studies of comprehensive multimodal treatment sustained over several years. Severity of symptoms and co-morbidity will influence mid or long-term outcome.

Current multicentre trials in the United States should give more information about the relative importance of stimulant medication (Greenhill, Abikoff, Arnold et al 1996; Richters, Arnold, Jensen et al 1995).

Ethical conflicts

The need to optimise treatment for an individual conflicts ethically with attempts to give the same multimodal intervention for a fixed time in a controlled, randomised study. The individual may not get optimal treatment for their condition or the study may be compromised.


Reviews of cumulative research are influenced by all the above-mentioned factors. In turn, reviews have major influence on attitudes, knowledge and practice.

The urgency to close the gap between clinical practice, professional and public knowledge and scientific information increases with recognition of ADHD and increased education.

The momentum and direction of this is influenced mainly by a small number of experienced professionals who lead clinical and educational research, practice and teaching and by families and individuals with ADHD, all of whom have varying perceptions, opinions and experience.

Studies on stimulant effectiveness for inattentive versus hyperactive types of ADHD

Semrud-Clikemen, Biederman, Sprich-Buckminster et al (1992) studied 60 children with ADHD, 30 with academic problems, and 36 normal controls and found that, depending on the stringency of academic and psychometric assessment tools for identifying learning disability, reading difficulty varied between 38 per cent and 15 per cent in children with ADHD, 43 per cent to 3 per cent in children with academic problems and 8 per cent to 0 per cent in controls.

The strong interaction between language disorders, learning difficulties, behavior problems and ADHD (McGee, Partridge, Williams et al 1991; Baker and Cantwell 1987) is confirmed by a seven-year follow-up of a Canadian cohort from Ottawa.

Poor linguistic and academic outcome and utilisation of special education were associated with low overall language skills (64 per cent) and poor comprehension (50 per cent), but not isolated poor articulation (Beitchman, Wilson, Brownlie et al 1996a).

Receptive and pervasive speech/language problems convey the greatest risk for later behavioral and emotional problems while poor auditory comprehension is a specific risk factor for hyperactivity and aggression (Beitchman, Wilson, Brownlie et al 1996b).

The Dunedin cohort shows the complex influences on whether symptoms are pervasive or situational for children with reading disability with or without ADHD (Pisecco, Baker, Silva et al 1996).

Famularo and Fenton (1987) showed that methylphenidate improved work output and accuracy in nonhyperactive children. Barkley, DuPaul and McMurray (1991) reported that lower doses may be optimum and there may be more nonresponders among children without hyperactivity (35 per cent) than in hyperactive ADHD (25 per cent).

De Quiros, Kinsbourne, Palmer et al (1994) studied 113 children across the whole spectrum of ADHD and did not support inattentive type ADHD presenting as learning disabilities and responding less to medication.

Further studies on the stimulant response of children with inattentive and hyperactive types of ADHD are described in Section 4.2.4 below.

Early reports on the effects of different doses and cognitive functions showed arithmetic responds better than reading or spelling tasks, perhaps because these are more intrinsically complex tasks (Douglas, Barr, Amin et al 1988; Douglas, Barr, O’Neill et al 1986).

More recent study (Douglas, Barr, Desilets et al 1995) indicates linear dose responses improving speed and accuracy of processing, flexible thinking, problem solving and persistence. Tannock, Ickowicz and Schachar (1995) showed methylphenidate improved working memory.

The presence of learning disability did not affect stimulant response, which increased up to a dose of 0.9 mg/kg/dose and was blocked by coexisting significant anxiety disorder. This may help understand reports of less stimulant effect in inattentive type ADHD, which more commonly coexists with anxiety.

Reports of a child’s learning, performance or behavior while receiving medication, especially stimulants, must state the size and timing of the last dose so that the interaction of a rising, steady or falling drug level on those aspects being assessed can be evaluated.

This is crucial when assessing the benefits of any teaching or therapy program and reports. Dose effects and requirements for inattentive type ADHD may need more thorough and subtle evaluation.

Established benefits of stimulants

The considerations described in Section 4.2.2 also apply to clinical management for the individual with a specified target symptom of ADHD or associated learning, emotional or mood problems.

Double-blind placebo studies, most with randomised selection, demonstrate the following benefits of stimulant medication, usually methylphenidate, which can guide practice.

Improvement in ADHD symptoms is shown in 70-80 per cent of patients who took methylphenidate, the majority moving into the normal range of behavior (Wilens and Biederman 1992). Placebo response averages 18 per cent with reports varying from 2 per cent to 39 per cent (Ullman and Sleator 1986).

DuPaul and Rapport (1993) reported only 20 per cent improvement with placebo on any parameter of classroom functioning compared with methylphenidate inducing improvement in 97 per cent; normalised academic and social functioning in 75 per cent and less day-to-day variability with higher doses.

Routine use of placebo is not necessary but still overemphasised by some nonprescribing professionals.

Placebo use may be helpful to clarify side-effects or equivocal response to stimulant medications and to determine benefits relative to other medications being taken concurrently. It may help resolve concerns, confusion and conflict between patient, parent, and professional regarding perceptions, preferences, philosophies and prejudices.

Truly indistinguishable placebo preparations are not easily available and the techniques of randomised crossover ‘n of 1’ trials in an individual patient should be followed closely (Mahon, Laupacis, Donner et al 1996).

Discussions of detailed procedures for evaluating stimulants include Gadow, Nolan Paolicelli et al (1991), Di Traglia (1991) and McBride (1988).

Disruptive behavior normalizes with medication, as regulatory and inhibitory processes are preferentially stimulated with many possible benefits on function, organization and self-regulation observed as ‘stimulant’ effects.

By improving behavior, emotions, school performance, some learning tasks and peer and family relationships, dexamphetamine or methylphenidate (Ritalin¨) enhance other treatments for ADHD such as behavior modification, support teaching or therapy.

An Australian study by Lewin and Fletcher (1993), involving relatively small numbers collected in 1987, found ‘at 12 months no improvements were evident that could be directly attributed to treatment’.

This supports an outdated meta-analysis (Thurber and Walker 1983) claiming that methylphenidate shows only small short-term, and no long-term, improvement. This does not, however, reflect more recent results of research or current practice and guidelines.

Comparison of methylphenidate and dexamphetamine

Trials involving both stimulants may be conducted separately to determine which is most effective. Direct comparisons are few and are discussed in Elia (1993).

When both stimulants were tried in a range of doses (including up to 3.0 mg/kg/day methylphenidate), 96 per cent of children responded; 94 per cent of teacher ratings showed improved attention and motor restlessness in the children (Elia, Borcherding, Rapoport et al 1991).

Some of the same authors compared dexamphetamine, methylphenidate and placebo at different doses in treating ADHD and comorbid Tourette Syndrome, again showing the difference between the stimulants (Castellanos, Giedo, Elia et al 1997).

Expert opinion provided to the working party in consultation submissions reinforce the conclusion that, while most children respond in a similar way to methylphenidate and dexamphetamine, some children respond better to one than the other, and a small number respond only to one and not the other.

True response or non-response must be thoroughly evaluated for each individual (Arnold 1996). The relative efficacy and side-effects of dexamphetamine and methylphenidate have been compared in an Australian crossover study which confirms that these vary between individuals.

The relative efficacy and predictors of response are still being evaluated in detail and are to be reported separately. However, the side-effect profiles are similar, although negative emotional symptoms are more likely to occur with dexamphetamine (Efron, Jarman and Barker 1997).

In Australia, dexamphetamine is available on the Pharmaceutical Benefits Scheme (PBS) and methylphenidate is not. Administrative processes for the appropriate use of each stimulant will need to be determined.

Effect of dosage

Most studies demonstrate improved benefits with increasing dosage on behavioral symptoms and recently published studies give strong evidence for cognitive improvement with linear dose responses (Douglas, Barr, Desilets et al 1995; Tannock, Ickowicz and Schachar 1995).

The variation between individuals and behaviors requires dosage to be ‘tuned’ to the individual’s most disabling symptoms (limited by maximal effect balanced against toxicity).

Sprague and Sleator (1977) suggested an inverted U-shaped dose response curve on which higher doses up to 1.0 mg/kg/dose improved global behavior ratings but increased error rate on short-term memory tasks with the best correct score on lowest dose of 0.3 mg/kg/dose.

This single unreplicated, uncontrolled study has greatly influenced prescribing practices towards conservative dosage. Other studies summarized close to that time, show group effects of a linear dose response curve in both behavioral and cognitive responses.

The recent work of Tannock, Ickowicz and Schachar (1995) confirms methylphenidate linear dose response on cognitive function of working memory and coexistent learning disabilities, and a less robust response with coexisting anxiety.

Douglas, Barr, Desilets et al (1995) confirm cognitive and behavioral improvement with linear dose responses up to 0.9 mg/kg/dose.

Benefits for hyperactivity increase linearly with the dose, though individual benefits have little relationship to actual levels of the stimulant or their variation in the blood stream during the day.

The more severe the symptom, the greater may be the response, in what has been called a ‘rate dependent response’ (Rapport and DuPaul 1986).

DuPaul and Rapport (1993) report that 97 per cent significantly improve, 75 per cent normalise academic performance and social functioning and there is less day-to-day variability with higher doses of methylphenidate.

The importance of short-term treatment on long-term academic performance has not been conclusively demonstrated.

No studies have examined comprehensive optimum management of ADHD and coexisting learning disability or other emotional disorders with sustained medication, behavioral support, optimum classroom accommodations and academic/vocational planning.

Assessment of improvement

Models of improvement from stimulant treatment of ADHD include vigilance (Perel, Greenhill, Curran et al 1990), improved error rate and delay in responding (Satterfield, Hoppe and Schell 1982); accuracy, flexibility and speed of processing (Douglas, Barr, Desilets et al 1995); improved working memory (Tannock, Ickowicz and Schachar 1995) and improved sense of time (Barkley 1995b).

Specific functional improvements at school include on-task behavior, increased work output, accuracy and neatness (Famularo and Fenton 1987; Douglas, Barr, O’Neill et al 1986). Inefficiency in modulating social insight and interactions may be the most disabling symptom of ADHD.

There is also a close association in individuals and in family patterns with other developmental, behavioral and emotional disorders (Biederman, Faraone, Keenan et al 1992).

Stimulant medication improves social interaction, communication and responsiveness, restores aggression to within the normal range, has a dose dependent effect on antisocial behaviors such as stealing and property damage (Hinshaw, Heller and McHale 1992) and improves but does not normalize peer appraisal (Whalen, Henker, Buhrmester et al 1989).

Medication benefits functioning during sport by improving attention and participation rather than specifically increasing skills (Pelham, Greenslade, Vodde-Hamilton et al 1990).

Family functioning improves when stimulants are taken for ADHD ÷ with enhanced parental warmth and approval, less criticism and less sibling conflict (Schachar, Taylor, Weiselberg, et al 1987; Barkley 1990b).

Adverse effects of stimulants

Adverse effects fall into several groups:

  • physiological effects of noradrenergic activation;
  • those associated with pharmacokinetics of single or repeated doses waxing and waning over three to four hours;
  • different individual responses between doses and between the two stimulants; and
  • other idiosyncratic responses.

In such a chronic condition as ADHD, long-term treatment issues include:

  • whether tolerance develops, requiring increasing dosage;
  • increased side-effects in certain populations;
  • long-term consequences of persisting mild adrenergic activation;
  • effects on physical growth of the child; and
  • potential for misattribution of the action, necessity and power of medication and psychological or pharmacological dependence.

To interpret and identify side-effects accurately requires careful observation. Is the possible symptom present before, unchanged by or begun with medication? Does it relate to a dose wearing off, to a dose increase and is it less if dosage is interrupted?

Does it occur with either stimulant or only one, and at what dose? Is it present in particular situations? Is it affected by other variables ÷ for example, an unreasonable demand on learning disability, change in others’ behavioral responses (eg relief teachers) or any other disruptive event (eg family sickness, disruption or loss)?

The dose required, its benefits and side-effects vary between individuals.

Optimal treatment depends on the balance between the best improvement of the most significant problems (eg disruptive behavior), relatively lesser effect on other problems (eg cognitive improvement) and the existence of any side-effects (eg mild appetite, sleep, mood or tic problems).

Symptoms commonly attributed to stimulant side-effects can occur in children without ADHD, without medication or as the characteristics of ADHD and its differential diagnoses. Side-effects may be confused with symptoms of nonresponse, inadequate dosage, or occur as a dose wears off.

Such confusion may cause vigorous debate and conflict between patient and family members, and between family members and professionals (Jureidini 1996).

Empirical adjustments and observations can establish individual optimum dosage, particularly evaluating doses given later in the day and at weekends and holidays.

Barkley, McMurray, Edelbrock et al (1990) compared side-effects of placebo and methylphenidate in 83 children over seven to ten days at doses of 0.6 to 1.0 mg/kg/day.

Teachers reported over 50 per cent of students receiving placebo with variable or frequent side-effects; some of the reported ‘side-effects’, such as staring, disinterest, sadness and anxiety actually reduced with methylphenidate treatment.

The only symptoms more associated with stimulants were on appetite, sleeplessness, headaches and stomach-ache; only 3.6 per cent had side-effects requiring discontinuation of treatment.

Side-effects are generally mild, can be managed with adjustments of timing or dose and can be confirmed by interrupting medication.

Very few pharmacodynamic studies exist. The mechanisms of short and long-term adverse effects are thoroughly discussed by Greenhill (1992).

Physiological adrenergic activation can produce stimulation of cardiac muscle, contraction of bladder sphincter, bronchodilatation; although a mild rise in blood pressure and tachycardia are poorly correlated with dose and settle over a few weeks in sustained treatment.

Adverse effects of stimulants are far less common and severe in controlled studies than their reputation in popular belief, opinion and prejudice.

Some misconceptions are based on effects of stimulants in severe psychiatric or other co-morbid disorders; in substance abuse of large doses and mixed with other agents and in animal experiments with huge doses.

Methylphenidate has a short half-life and may produce more physiological discomfort or more striking return of symptoms.

Dexamphetamine may theoretically predispose to slightly more noticeable side-effects as confirmed in an Australian crossover study of the two stimulants. The side-effect profiles are similar, but negative emotional symptoms are more likely with dexamphetamine (Efron, Jarman and Barker 1997).

Direct comparisons are few and are well discussed by Elia (1993). If reported, ‘rebound’ deterioration in behavior is usually in the afternoon or early evening.

Possible influences include: the waning of a dose, loss of school day routine and positive behavior management, longstanding negative family interactions and unrealistic comparison with pretreatment behaviors after a more positive period with the medication.

In the first hour or so when a dose of medication begins to act, some children may seem slightly more serious then become more spontaneous and appropriately organized and functional.

Sometimes a balance must be accepted between improved attention, task completion and performance quality and a slight difference in spontaneity. In a report, now twenty years old, Sprague and Sleator (1977) suggested that higher doses may compromise cognitive performance.

This single uncontrolled study has had powerful effects on attitudes to medication. ‘Loss of sparkle’ concerns many teachers alarmed by apocryphal stories of ‘zombie effects’.

Such ‘cognitive constriction’ is not supported by recent studies which confirm cognitive and behavioral improvement with linear dose responses in single methylphenidate doses up to 0.9 mg/kg/dose (Douglas, Barr, Desilets et al 1995; Tannock, Ickowicz and Schachar 1995).

Funk, Chessare, Weaver et al (1993) found no stimulant induced impairment of nonverbal figurative creativity.

Teachers may be tuned in intently to changes in the quality of an individual child, especially in small support classes with structured, clear, individual educational programs and high teacher/student contact (Kasten, Coury and Heron 1992; Barkley, Fischer, Edelbrock et al 1990; Safer and Krager 1988).

Concerns about long-term treatment side-effects centre mainly on growth problems, dose tolerance and chemical or behavioral dependence (Levy 1993).

There is some evidence from control studies that ADHD, and not its treatment, is associated with temporary deficits in growth in height through mid-adolescence that may normalise by late adolescence (Spencer, Biederman, Hardy et al 1996).

In addition to the problems of individual variation of symptoms and response, nearly all group studies are confounded by uncertain compliance; too cautious dosage; short-term or intermittent treatment; poor recording of data eg symptoms, benefits, diets and growth; lack of untreated ADHD controls and practice driven by ignorance and anecdote.

The literature does not yet contain studies of optimum, prolonged, sustained, multimodal treatment or even sustained doses of medication titrated optimally for individuals given for years. Extended treatment studies have been reviewed by Schachar and Tannock (1993).

Only stimulant medication is formally monitored by statutory procedures. Other medications less commonly used for ADHD may have more serious side-effects and are not monitored or restricted to specialist prescribing.

The designation of stimulants as drugs of potential addiction colours the context of ADHD treatment. While anecdotes prevail, misuse of stimulant medication in appropriately treated uncomplicated ADHD is uncommon.

With appropriate therapeutic use, there is no evidence of increased substance abuse and some protection from it compared to the untreated adolescent with poor impulse control. Indeed, there are recent case reports of substance abuse in adults being overcome when ADHD is treated with stimulants.

The Ontario population cohort has followed about 800 Canadian children up to 12 to 16 years. Neither ADHD, presumably treated with stimulants in many children, nor emotional disorder was related to adolescent substance abuse.

Conduct Disorder assessed by teachers predicted use of tobacco and hard drugs; low family support and poor school performance predicted use of tobacco (Boyle, Offord, Racine et al 1993). Lynskey and Fergusson (1995) confirmed these findings in New Zealand for the Christchurch cohort.

Significant side-effects which should always be taken seriously include tics, major mood changes with marked sadness, anxiety, aggression and any bizarre or persecutory thoughts.


  • The use of stimulant medication should be considered as part of the management plan for most children with ADHD. The efficacy and safety of stimulant medication has been established for short-term use.
  • While the response of most children is similar for both methylphenidate and dexamphetamine, the efficacy and side-effects are not identical.
    • Some children may respond better to one than the other. Therefore, children should have equal access to whichever drug is necessary for their optimal treatment.
    • Further research is required to determine the comparative effectiveness and cost-effectiveness of the two medications, and to determine criteria which will predict the optimal therapeutic option for individuals.
  • The routine use of placebo to assess individual response to treatment is not recommended.
  • Further research should examine the efficacy and safety of medications, particularly psychotropic medications, and prolonged or continuous use of stimulant medication.