Abstract
Keywords
Not only has attention deficit hyperactivity disorder (ADHD) become firmly established in the panoply of psychiatric diagnoses, its treatment with psychostimulants is a growing phenomenon in clinical practice [1]. In Australia, the total number of prescriptions issued for stimulants across all age groups increased more than 10-fold in the years 1990–1998 [McManus P: personal communication, 1999]. Many patients have benefited from this trend, with an increased ability to perform at work or study and an improvement in the quality of their lives [2]. Most psychiatrists, however, approach the prescription of stimulants with some trepidation. Psychostimulants are commonly thought of as drugs of abuse rather than therapy, their long-term effects are not fully known and beneficial effects are hard to quantify objectively. When a patient makes a request for an increased dose of the drug for the treatment of ADHD, it immediately arouses the suspicion of abuse or the fear of adverse consequences. The problem is particularly acute in adolescents and adults because the validity of the diagnosis of ADHD in adults is often questioned [3], the empirical basis for treatment is less well-established than for children [1,3], the co-occurrence of drug abuse is common [4] and many patients have the additional diagnosis of personality disorder [4].
Dosage
If one does accept the legitimacy of stimulant drug treatment in adult ADHD, is there a maximum daily dosage that should not be exceeded? The answer needs to be sought from the efficacy data available, the neuroscientific information that underpins the treatment, the pharmacokinetics and bioavailability data on stimulants in adults in comparison with children and the differential likelihood of long-term side-effects with higher doses. Considerable discrepancy exists between the states and territories of Australia regarding maximum allowable stimulant doses in adults with ADHD. In New South Wales and Tasmania, a prescriber is permitted a maximum of six tablets of dexamphetamine (5 mg each) or methylphenidate (10 mg each) per day, beyond which a second opinion has to be sought. In Western Australia, up to 12 tablets per day can be prescribed. Other Australian states and territories do not set a specific upper limit and allow this to be largely clinician determined.
The efficacy of stimulants, in particular methylphenidate, in adult ADHD has been examined in at least seven studies (n = 193 subjects), five of which were controlled, with an overall report of significant reduction of symptoms of ADHD (for review see [5]). In four of the five controlled studies [6–9], the average daily dose of methylphenidate used was 0.6 mg/kg and the average response rate was about 50%. In the fifth study, a dose of 1 mg/kg was used and a response rate of 70% was reported, the authors arguing for the acceptance of the higher dose in clinical practice [10]. The limitation of such data is that, unless there has been a direct comparison of different doses in the same study, caution is necessary for the acceptance of the differential improvement rates. Furthermore, the maximum dose examined was 1 mg/kg, thereby providing no information on even higher doses, the question that is raised by many clinicians and patients. Higher doses were used in one open study treating predominantly children [11], with the range for dexamphetamine in this study being 1.6–3.6 mg/kg and for methylphenidate 1.4–7.7 mg/kg. The highest dose of dexamphetamine used in this report was 230 mg/day. The open nature of this report is a serious limitation and makes it an unlikely basis to guide therapy [12]. There has been one controlled trial of dexamphetamine in adult ADHD [13] which, together with anecdotal reports [14], supports its being as useful as methylphenidate. The relative dose of dexamphetamine in mg/kg, if usage in children is a guide, is about one-half the methylphenidate dose [15], that is, the standard tablet sizes available for the two drugs are equipotent.
Behavioural pharmacology
An examination of the behavioural pharmacology of stimulants is instructive. Dopaminomimetic drugs, which include stimulants, have a biphasic response in animals, with lower doses reducing locomotor activity and higher doses stimulating locomotion [16–18]. This has been observed in spontaneously active rodents [19], as well as in rats made hyperactive with a neonatal dopaminergic lesion [20]. The dose of amphetamine that reduces activity in mice is in the range of 0.5–1.0 mg/kg, with activity levels increasing progressively with doses from 1.0 to 5.0 mg/kg [18]. The biphasic nature of the response is seen in humans as well, and the commonly used clinical dose of dexamphetamine (0.2–0.6 mg/kg) is in the same range as that in the hypolocomotor animal model. Higher doses in humans lead to excitation, sleeplessness and anorexia, all features of excessive brain stimulation, again like the hyperlocomotor animal.
The pharmacological basis of the biphasic response is incompletely understood, but data from animals have generated some cogent theories. The dopaminergic effect of a drug depends on its influence on the resting level of extracellular dopamine and the increase it produces in the pulsatile dopamine associated with the nerve impulse [21]. In a drug-free condition, a nerve impulse produces a transient 60-fold rise in synaptic dopamine level above the baseline level of about 4 nmol/L [22]. Methylphenidate and dexamphetamine, like cocaine, block the dopamine transporter leading to an increase in dopamine in the synaptic cleft, and dexamphetamine additionally has a direct dopamine-releasing action [23]. These drugs therefore raise the baseline level of extracellular dopamine, as well as the nerve impulse-associated rise. At low doses of stimulants in the rat, the rise in baseline values is about sixfold, and that in the impulse-related levels is twofold [24,25]. The relative rise of dopamine, when expressed as a percentage of the baseline values, is therefore lowered by stimulants at low doses. At higher doses of dexamphetamine (> 1 mg/kg), the rise in baseline is seven-to 35-fold and that in the pulsatile output about sevenfold [26]. This differential effect of low and high doses may account for the biphasic response of stimulants [21]. Since the response in humans is similarly biphasic, and the commonly used dose range for the treatment of ADHD is similar to the hypolocomotor dose in animals, it could be argued that higher dexamphetamine doses (for example greater than 1 mg/kg) will have a stimulating effect and could possibly worsen the features of ADHD.
This argument can be further informed by recent work using positron emission tomography (PET). It has been shown that for cocaine to produce a ‘high', at least 60% of the dopamine transporters must be blocked and this effect should occur and dissipate rapidly [27]. Methylphenidate, which is similar in its action to cocaine except for a more sustained effect, produces a 50% blockade at an oral dose of 0.25 mg/kg, with the peak concentration in the brain being reached at 60 min [28]. The plasma levels of methylphenidate correspond to the dopamine transporter occupancy [28]. Clinical doses of methylphenidate, therefore, can be expected to produce a greater than 50% occupancy which should be pharmacologically meaningful. At a dose of 1 mg/kg, 80% or greater occupancy should be expected [28], arguing against the value of further increases in clinical doses. In fact, adverse effects are likely to outweigh an advantage at very high doses.
Tolerance
If the clinical and experimental evidence does not support the use of high doses of stimulants (e.g. above 1 mg/kg dexamphetamine or 2 mg/kg methylphenidate), is there an argument that the dose would need an escalation, owing to tolerance, beyond the usual recommendation, when the drug is used chronically? Tolerance to some actions of stimulant drugs has been demonstrated in animal studies [29,30]. Long-term administration of dexamphetamine has been shown in monkeys to lead to long-lasting down-regulation of D2 dopamine receptors [31]. Tolerance does not, however, develop to all actions of these drugs [29]. The clinical evidence does not suggest the development of tolerance to the therapeutic effects [15]. In a recent randomised, double-blind, placebo-controlled study of dexamphetamine treatment in children with ADHD, no evidence of tolerance was noted over a period of 15 months [32]. This was a sufficiently long period for tolerance to develop if that was a likely outcome. There is experimental evidence that long-term use of dopamine agonists leads to a reduced density of D2 receptors [33]. Whether this occurs with stimulants as they are used clinically, and what the practical implications are, is not known. It is also possible that chronic stimulant treatment could alter the functional state of both D1 and D2 receptors, but the significance of this is also not known [21]. In conclusion, the empirical basis is weak, but clinical experience argues against the need for an escalation of dose with continuing treatment once the therapeutic effect has been established.
Pharmacokinetics
Examination of the pharmacokinetics of orally administered stimulant medication is an important step in determining a rationale for dosing regimens. After oral administration, both methylphenidate and dexamphetamine are almost completely absorbed [34], with food having little impact on this process [35,36]. In adults, methylphenidate has a pharmaco-kinetic profile similar to that in children [34,35], reaching a peak concentration at 1.5–2.5 h and having an elimination half-life of 2–3.5 h after oral administration of a single dose [34,35,37]. Nonlinear kinetics have been observed with higher doses of methylphenidate [38,39] and have also been proposed with repeated doses [40], probably as a result of saturation of presystemic metabolism. Dose escalation therefore carries the theoretical potential for sudden increase in adverse effects. Dexamphetamine typically takes 2–3 h to reach peak concentration after a single oral dose [34], but its half-life (approximately 7 h) is considerably longer than that of methylphenidate. As is the case in children, peak serum concentrations of both stimulants for the same oral dose may vary by four- to fivefold in adults [35,36,38]. Such variability may arise from differences in presystemic metabolism between individuals [38]. Inter-individual variability in plasma clearance is seen with both drugs [34] and, contrary to earlier suggestions [41], does not appear to be less likely with methylphenidate. As inter-individual differences in pharmacokinetics may be less dramatic when dose is adjusted for body weight [36], it would seem appropriate to use mg/kg as a general guide for determining maximum dose in individual patients. Inter-subject variability might provide an argument for high dosing of individuals failing to respond to conventional dosing strategies. However, as it remains unclear whether clinical response in children and adults is related to peak brain concentration or the rate of increase of the drug in the brain [42], dose escalation purely on the basis of ‘inadequate’ response cannot be recommended. Although there may be an argument for examining the individual pharmacokinetics of stimulants in those patients failing to respond to conventional dosing strategies, this requires ready access to a laboratory with an established assay, and is not feasible in routine clinical practice. Furthermore, data from children suggest that mean serum levels do not differ between drug responders and non-responders [35].
Side-effects
The dose of stimulants is significantly associated with the presence of side-effects [43,44]. As the dose of methylpenidate exceeds 0.5 mg/kg, the rates of insomnia and appetite suppression increase. Within the recommended therapeutic range, this is not generally a problem in adults [2], but at higher doses this could be a major limiting factor. The cardiovascular effects of stimulants (e.g. increases in pulse and blood pressure) have not been well examined, but they cannot be ignored at higher doses, especially in adults with pre-existing cardiovascular disease. The risk of convulsions also goes up with higher doses [43]. In long-term usage, the concerns in adults have been the possibility of drug abuse [45], the precipitation of psychosis [46], the production of tic disorders as has been seen in children [45] and the possibility of as yet unknown permanent brain dysfunction although this seems unlikely. The development of amphetamine psychosis has been associated with frequent high-dose administration of amphetamines, and this has been replicated in an escalating-dose ‘binge’ model of amphetamine psychosis in the rodent [31].
The therapeutic use of drugs of potential abuse raises the important question of whether their prescription encourages abuse or dependence and whether such a problem may be more prevalent at higher therapeutic doses. It is believed that use of stimulants in children does not increase the risk for later abuse or dependence [15]. A recent study demonstrated that treatment of childhood ADHD dramatically decreased the risk of substance-abuse in adolescence [47]. The same cannot be said with confidence when treatment is initiated in adolescence and adulthood [48]. Intranasal and intravenous abuse of prescription stimulants (one's own or from a friend) has been reported [48–50]; oral abuse is also recognised [51]. Since the effect of stimulants on the dopamine transporter is similar to cocaine, and dose is important for the ‘high', large therapeutic doses raises concern. Volkow et al. [28,52] have demonstrated that dopamine transporter blockade by both oral and intravenous methylphenidate is necessary, but not sufficient, to produce self-reports of a ‘high', and postulate that it is the rate of uptake of methylphenidate into the brain that determines its reinforcing effect. This work, and the preference for abuse of stimulants by non-oral routes, suggest that prescribed stimulants themselves, when taken orally in modest doses, are unlikely to encourage abuse or dependence. Clinical practice, however, is complicated by the high prevalence of personality disorders and substance abuse in adults with ADHD [4]. Clinicians hesitate to prescribe stimulants to a patient with a history of drug abuse, even when a diagnosis of ADHD is confidently made. Drug abuse cannot, however, be considered an absolute contraindication [2], but it is advisable to remain within therapeutic doses and monitor patients closely.
Stimulants are recognised as producing a psychosis with persecutory delusions and auditory and visual hallucinations [46]. This is usually associated with intravenous use and high doses, and has not been described with therapeutic use [53]. The possible use of large doses, however, raises this concern. Behaviours such as extreme vigilance, tracking of non-apparent stimuli, constant checking, hyper-responsiveness and grasping in mid-air have been reported in non-human primates treated chronically with high-dose amphetamine [54]. Other behaviours in monkeys include posturing and other catatonic features [55]. Castner and Goldman-Rakic [56] reported that, in monkeys, repeated intermittent but escalating (up to 0.8 mg/kg) low-dose amphetamine treatment led to behavioural sensitisation. While low-dose, chronic treatment may produce similar behaviours at a lesser severity [56], high doses are more likely to cause them. Although both amphetamine and methylphenidate sensitise stereotypic behaviours in rats, only amphetamine induces loco-motor sensitisation [57]. This may reflect the different methods by which the drugs increase synaptic dopamine [57]. There is interest in elucidating alterations in dopamine-stimulated signal transducing mechanisms that may be responsible for different components of behavioural sensitisation [58]. Chronic and high-dose amphetamine exposure in monkeys and rodents has also been reported to produce nigros-triatal toxicity [59,60]. This can be prevented if an intermittent, low-dose schedule is used [61]. Yuan et al. observe [62] that methylphenidate appears to lack the dopaminergic neurotoxic effect shown by dexamphetamine in rats 2 weeks after a high dose. The contrast in neurotoxic potential of these two drugs at high doses suggests that dopaminergic efflux may not be the direct mediator of such damage [62]. While implications of laboratory findings are unclear, they prompt an appraisal of chronic stimulant use, even at low doses, and its adverse consequences. Meanwhile, doses should be restricted to low levels.
Conclusion
Our data justify the use of chronic, low-dose stimulant treatment of ADHD in adults. High-dose treatment is not recommended. There no evidence of greater improvement and any beneficial effect is likely to be compromised by adverse effects. Some short- and long-term side-effects of high doses can be serious. The upper limit is difficult to decide in the absence of data. We recommend limits of 1 mg/kg for methylphenidate and 0.5 mg/kg for dexamphetamine. These doses should be exceeded only after consideration and objective documentation of beneficial and adverse consequences. Monitoring drug levels may be of value if poor compliance or failure of the dose to produce adequate serum levels are being considered, as there is a direct relationship of blood to brain levels as well as to dopamine transporter occupancy. Further research is needed to achieve a sound basis.
Footnotes
Acknowledgements
We thank F. Levy for her help.
