Abstract
Attention deficit hyperactivity disorder (ADHD) in adults is an increasingly recognised clinical entity but remains controversial as regards diagnostic criteria and treatment. Editorials from the American Journal of Psychiatry [1] and the British Journal of Psychiatry [2] highlight its international emergence as a childhood onset disorder with variable persistence into adulthood. In Australia, there has been a National Health and Medical Research Council report on ADHD in children and adults [3]. Evidence from both childhood [4] and adulthood [5–8] suggests that stimulant medication is the cornerstone of management for ADHD in both children and adults.
The consumption figures for stimulant medication (dexamphetamine and methylphenidate) in Australia, for children and adults combined, demonstrate the following level of clinical usage [9]. In 1997, about 25 million dexamphetamine tablets and about 14.5 million methylphenidate tablets were consumed. In addition, the rate of consumption has been increasing: from 1994 to 1997 the consumption rate for methylphenidate doubled, and for dexamphetamine it tripled [9].
The common clinical usage of dexamphetamine contrasts with the paucity of research on its efficacy in adults. There have been no clinical trials of dexamphetamine in the treatment of adult ADHD. The authors of this paper could find only one reference to the usage of dexamphetamine in adults and that was in a non-controlled study of five prisoners in Norway [10].
Most pharmacotherapy evidence concerns the efficacy of other stimulant medication, especially methylphenidate, and is summarised in a recent review [8]. This review highlighted that in contrast to the consistently robust responses to stimulant medication seen in children and adolescents (approximately 70%), controlled studies in adults have shown more equivocal responses to stimulant medication (25–78% response rate). Variability in the response rate appears to be related to several factors such as the diagnostic criteria used to determine ADHD, stimulant medication dosages, levels of comorbidity and methods of assessing overall response. Higher dosages appeared to result in more robust outcomes. Long-term data are generally lacking, which is important given the chronicity of the condition [5].
The diagnostic criteria used to determine ADHD in adults is a subject of major discussion [5],[11]. It is argued that the DSM-IV criteria are too restrictive as the symptom wording reflects childhood behaviour, symptom patterns change over time, and self-report rates (more common with adult studies) are lower than relatives' reports [11],[12].
Murphy and Barkley [11] suggest that the diagnostic threshold for ADHD in adults is lower than in children and decreases with increasing age. The standard DSM-IV ADHD symptom checklist consists of 18 symptoms which are included as present if they are ‘often’ observed [13]. The DSM-IV threshold is the presence of six, out of a possible nine, symptoms for inattention and/or hyperactivity-impulsivity, shortened to hyperactivity hereafter. As for children, ADHD in adults has three subtypes: the inattentive ADHD subtype qualifies just on the inattentive criteria; the hyperactive type qualifies just on the hyperactive criteria; and the ADHD combined subtype is present if a patient reaches the threshold on both the inattentive and the hyperactive criteria [13].
A study by Murphy and Barkley [11] used a self-report questionnaire to determine normative data for the DSM-IV symptom thresholds in adults. The results suggest that for ages 17–29 the threshold should be the presence of at least four inattentive and/or five hyperactive symptoms, for ages 30–49 at least three inattentive and/or four hyperactive symptoms, and for those older than 50 at least two inattentive and/or three hyperactive symptoms.
The aim of this study is to examine the short-term efficacy of dexamphetamine in adult ADHD by conducting a randomised, double-blind, placebo-controlled trial in a naturalistic setting.
Method
Two psychiatrists, working in private practice, screened consecutive patients for a research trial into adult ADHD using a questionnaire based on the DSM-IV symptoms. These patients had been referred in the usual way by their general practitioners because of concerns that the patients may have ADHD. In Western Australia, only psychiatrists or specialist physicians can prescribe stimulant medication for adult (i.e. aged 18 years or older) ADHD.
Patients were asked to fill out a DSM-IV ADHD symptom checklist. Patients were eligible for inclusion in the trial if they reported the presence of at least four inattentive and/or five hyperactive symptoms during the previous 6 months. These thresholds were based on the observation by Murphy and Barkley [11], as discussed in the Introduction, that adults have lower thresholds than children. However, given the controversy in the area it was thought best to try and find some mid-ground between the restrictive DSM-IV thresholds and the more lenient thresholds they proposed. The thresholds we used were those suggested by Murphy and Barkley [11] for adults aged 17–29 years, and thus will provide a more conservative threshold for patients in older age groups.
Patients were excluded from the study on the grounds of either having an insufficient ADHD score, or comorbidity for other major psychiatric disorders, including a history of current substance abuse. Typically, in clinical practice in Western Australia, and probably elsewhere, other major psychiatric disorders would be treated before embarking on treatment of ADHD [5–7]. Patients were also screened for organic disorders that would contraindicate the use of dexamphetamine. Finally, all patients eligible for the trial had a sample of urine tested to screen for illicit substance abuse. Patients included in the trial signed written consent forms acknowledging their involvement in the research. Ethical approval was gained from the local State Ethics Committee of the Royal Australian and New Zealand College of Psychiatrists.
Patients were randomly assigned to a 6-week, double-blind trial of either active dexamphetamine treatment or placebo treatment. Dosage level was set in a manner which approximated the practice of both psychiatrists: that is, begin at a low dose, gradually increase and take before early afternoon to avoid the side effect of insomnia.
Specifically, patients for the first week took one tablet each morning after breakfast. For the second week, they took one tablet after breakfast and one tablet after lunch. For the third week, they took two tablets after breakfast and one tablet after lunch. Their progress was reviewed at the end of this 3-week period. For weeks 4–6, patients were instructed that they were now allowed to take up to six tablets per day, but that incremental increases should be by no more than one tablet per day, with 2 days between increases. They were then reviewed at the end of week 6.
Patients were instructed to record both the number of tablets taken each day, and any side effects noticed.
Measures
The DSM-IV ADHD criterion list with the modified thresholds outlined above was used at initial assessment, 3 weeks and 6 weeks [13]. In addition, patients' relatives were asked to fill out these questionnaires and the results of these were compared with the patients' self-reports.
The Brief Symptom Inventory (BSI) [14] is considered as a general measure of psychological ill-health. It is a 53-item, self-report symptom inventory designed to reflect the psychological symptom patterns of psychiatric and medical patients, as well as community non-patient respondents. According to the BSI manual, the BSI may be used as a single point-in-time assessment of clinical status, or it may be used sequentially to document trends over time, or in pre-post evaluations. It produces several global indices and dimensions, including anxiety and depression scores and a Global Severity Index (GSI). The GSI is the most sensitive single indicator of a respondent's level of distress.
The BSI raw scores were translated into T-scores based on normative data from psychiatric outpatients [14]. The BSI provided an ongoing measure of whether patients' possible improvement in ADHD was just part of a general improvement in psychological health. This measure was repeated at the 3-week and 6-week assessments.
Three Clinical Global Impressions subscales were used: Severity of Illness, Global Improvement and Efficacy Index [15]. First, the clinicians rated the patients' Severity of Illness after the initial assessment. Second, the clinicians rated the patients' Global Improvement after 6 weeks. Third, an Efficacy Index was calculated by using a ratio of the benefits of treatment against the side effects of that treatment. A score of ‘1’ indicates that the benefits are equal to the side effects; this might mean no improvement and no side effects, or a situation where there was vast improvement with vast side effects. The possible range of the Efficacy Index is from 0 to 4.
At the end of the trial a five-point Likert Scale [16], measuring patient satisfaction was filled in by the patients. Of the adverse reactions to dexamphetamine described [17], weight loss and elevation of blood pressure were specifically assessed at 3-weeks and 6-weeks. At 6 weeks, urinalysis was again undertaken to ensure compliance, and to exclude drug abuse.
Statistical analysis
Statistical calculations were carried out using SPSS (v 7.0) for Windows (SPSS Inc. Chicago, IL, USA). Analyses were carried out on an intention to treat basis. That is, when there were missing data on week 6, the patient was assumed to be unchanged and scores of the previous assessment (week 0) were substituted. Ratings from relatives were missing for 13 patients. The Wilcoxon sum rank test was the chosen statistic for between-group comparisons of the ADHD rating scale scores, since item scores are not normally distributed. For continuous outcome variables, t-tests were carried out. Paired tests were used to compare scores between two time points (paired t-test and paired Wilcoxon sign test). Two-tailed tests of significance were used throughout.
Methods
Sixty-eight consecutive referrals were screened but 17 patients were found not to be suitable for the following reasons: 10 patients had ADHD scores below the cut-off levels; one patient was diagnosed with a mood disorder; and six patients were excluded because of current drug usage (one for alcohol and five for illicit drugs). This left 51 patients to enter the trial.
The urine drug screening performed on all of the 51 patients showed that six had been taking drugs despite denials about this. Of the six positive drug tests, four were for tetrahydrocannibol (THC), one for amphetamines, and one was for a combination of THC and amphetamines. These six patients were excluded from the trial a few days post assessment.
One more patient left the trial because of the onset of depression about a week after the trial began. The code was broken and it was revealed that the patient was taking dexamphetamine. The patient was withdrawn from the trial and was considered as a treatment failure. This patient has been included in the analysis on an intention to treat basis as described above. In fact, the patient's depression was treated and the patient subsequently did well on dexamphetamine and an antidepresssant. Thus, the results for 45 patients have been analysed, with 24 being allocated to the dexamphetamine group and 21 allocated to the placebo group.
Demographics
There were 27 males and 18 females, the mean age was 35.5 years (range = 19–57). There was no significant difference between the groups with regard to age and gender distribution (Pearson Chi-squared p = 0.143). Most subjects had completed school up to the age of 15, with eight completing tertiary education.
Attention deficit hyperactivity disorder subtypes
The two treatment groups were similar, with just over half of the cases being of the inattentive type. Nearly all of the remainder were of the combined type (that is, they met the criteria for both inattentive and hyperactive types as outlined in the previous section), with only one case (in the dexamphetamine group) being of the hyperactive type.
Attention deficit hyperactivity disorder symptom scores change over time
Self-rating
As displayed in Table 1, at 6 weeks, there was a generalised, significant reduction in total mean symptom scores for both the placebo treatment group (p = 0.042) and the dexamphetamine treatment group (p < 0.001). The dexamphetamine treatment group had significantly lower symptom scores at 6 weeks compared with the placebo treatment group (p = 0.045). There was no significant difference in the response to dexamphetamine between the genders and across the age range of the patients.
Mean attention deficit hyperactivity disorder (ADHD) symptom scores and mean (SE) Brief Symptom Inventory (BSI) ‘T’ scores at 0 and 6 weeks, by treatment group
The hyperactivity scores decreased over the trial period in both groups, but this was only significant in the dexamphetamine group (p = 0.004). Likewise, the inattentive scores decreased in both groups and only in the dexamphetamine group was this significant (p < 0.001).
Assessment by relatives
There were 32 patients (18 placebo/14 dexamphetamine) for whom relatives filled out a questionnaire, both on entry to the trial and after 6 weeks. Table 2 shows the mean ADHD symptom scores as rated by these patients and by their relatives, divided by treatment group. The only significant difference between patients' and relatives' scores occurred in the placebo group on entry to the trial, namely the hyperactive scores (p = 0.045).
Mean attention deficit hyperactivity disorder scores by patients and relatives at entry and 6 weeks, by treatment group
That is, the trial shows that in the patient/relative pairs studied, both the treatment effect of dexamphetamine, and the lack of effect of placebo on these symptoms, were detected by the relatives as well as by the patients themselves.
Brief Symptom Inventory (BSI)
Table 1 shows the mean ‘T’ scores for anxiety, depression and the Global Severity Index (GSI) on entry and at the end of 6 weeks. The scores are based on norms for adult psychiatric outpatients, calculated separately for males and females. Most of the scores are below the average for the norm, which are based on cross-sectional sampling of new and continuing psychiatric outpatients.
There were no significant differences between the placebo and dexamphetamine groups at the beginning of the trial. Overall, females on entry had a significantly higher anxiety score than males (p = 0.023), but there was no difference in the depression or GSI scores.
At the end of the trial there were no differences between the placebo and the dexamphetamine groups. In both groups the scores had been reduced: in the placebo group all the reductions were significant, but in the dexamphetamine group only the reduction in the anxiety score was significant.
Clinical Global Impressions
Severity of illness
At entry point, the clinicians rated both the dexamphetamine and placebo groups as scoring a mean of 4.05 (and median of 4.0) which corresponds with a rating of ‘moderately ill’.
Global Improvement Scale (GIS)
Of the patients taking placebo, 19 out of 21 (90.5%) were rated as unchanged or minimally worse, and none of them very much improved at 6 weeks. In contrast, at the end of 6 weeks 14 out of 24 patients (58%) taking dexamphetamine were assessed as being much improved or very much improved. The difference in the GIS between the groups was significant (p < 0.001). This confirms, clinically, the findings from the ADHD symptom scores.
Efficacy Index
At 6 weeks, 20 out of 21 (95%) placebo patients were rated by the clinicians as experiencing equivalent levels of both benefits and side-effects. The dexamphetamine patients were significantly different (p < 0.001) in that 18 out of 24 (75%) of these patients were rated as experiencing a level of benefits greater than the level of side-effects. This difference is significant (p < 0.001).
Dosage levels
Dexamphetamine treatment group
Of the 15 patients who significantly responded to dexamphetamine (i.e. had 6 week symptom scores which fell below the threshold levels), the average dose being taken at the end of 6 weeks was about five tablets per day (mean = 4.77, range = 1–7 tablets per day). Of the eight patients in this group who did not respond to dexamphetamine, four patients were on six tablets per day, two patients were on four and one patient was on two (no record kept in one case). The mean dose range for these eight patients was 4.25 tablets per day.
Placebo treatment group
Four patients did not record their dosage accurately. Of the remaining 17 patients in this group, 14 ended up on six tablets per day, one on two, one on four and one on seven tablets per day (mean = 5.7).
Side effects
General
The patients kept a diary of side effects. Records were missing or incomplete for 14 patients (eight dexamphetamine/six placebo). Four patients on dexamphetamine and 11 on placebo reported no side effects at all. Of the remaining 15 patients, the most common side effects reported for dexamphetamine versus placebo were sleep disturbance (nine vs one), headache (six vs three), dry mouth (seven vs zero), and thirst (three vs zero). A number of other side effects occurred occasionally. No recorded side effects were severe enough to cause the patient to discontinue dexamphetamine. As was seen in the results of the Efficacy Index, the benefits of dexamphetamine outweighed the side effects.
Specific
Table 3 shows the mean weight of patients on entry to the trial and at 6 weeks. The data for the patient who was withdrawn from the trial are not included. In the dexamphetamine group, 21 out of 23 (91%) patients lost weight, with the mean weight loss being 3.6 kg (SE = 0.525). This was significant (p < 0.001). In the placebo group, there was an overall mean weight loss of 0.286 kg (SE = 0.403) with nine patients losing weight and 12 remaining unchanged or increased. The weight loss for the placebo group was not significant.
Table 3 also shows the mean systolic and diastolic blood pressures for the patients on entry to the trial and at 6 weeks. In both groups there was a tendency for the blood pressure to be lowered, but the differences were not significant.
Mean (SE) weight, systolic and diastolic blood pressure at entry and 6 weeks, by treatment group
Patient satisfaction
The number of patients satisfied with the treatment was significantly higher in the dexamphetamine group compared with the placebo group both at 3 weeks and 6 weeks (p < 0.001).
Urinalysis
Urinalysis was performed at 6 weeks on all patients to screen for drugs. The results showed that all patients in the dexamphetamine group were complying with the treatment and that no patient in the placebo group was taking dexamphetamine. There was no evidence of concurrent drug abuse in any of the patients.
Conclusions
This study shows that daily usage of dexamphetamine is efficacious in adults with ADHD when compared with placebo after 6 weeks' usage. The response was similar in both genders and across the age ranges. The reduction in the number of ADHD symptoms was noticed by both the patient and their relatives on the self-administered ADHD questionnaires. It was also detected by looking at the patient satisfaction questionnaires and the clinicians' independent ratings.
This study provides evidence about the efficacy of dexamphetamine in adult ADHD as regards inattentive symptoms, but is less conclusive about hyperactive symptoms. For instance, there was only one patient in the ADHD hyperactive subtype in the sample at the entry point, and the mean hyperactive score was below the diagnostic threshold at the entry point. It would appear from this sample that adult ADHD patients do not have prominent hyperactive symptoms and may explain why ADHD was previously thought to often cease at puberty (i.e. the more obvious hyperactive symptoms decrease and the less readily observable inattentive symptoms continue) [18].
Not all patients who took dexamphetamine for 6 weeks felt satisfied with the treatment (five were neutral and two were dissatisfied) and not all reduced their ADHD symptom scores. Two patients showed a partial response in that their inattentive symptoms declined but not their hyperactive symptoms. One patient responded to dexamphetamine in the first 3 weeks of the trial, but then relapsed with the onset of influenza in the second 3 weeks of the trial. Afterwards, the patient resumed the medication and responded well. Five patients did not respond at all to dexamphetamine. Their progress after the trial was noted. One changed to methylphenidate and did well, two persisted with dexamphetamine and seemed to improve after another month, one responded to a higher dose of dexamphetamine (10 tablets per day), and one (the only pure hyperactive case) had to be admitted 1 month later with an ‘adjustment disorder’. It was thought that excessive dexamphetamine use may have contributed to the decompensation.
An obvious point about this trial is that it only lasted 6 weeks and yet dexamphetamine for adult ADHD appears to be a long-term treatment, lasting years, if not indefinitely [7]. It remains to be shown as to whether dexamphetamine has a long-term benefit in adult ADHD and whether long-term use will produce significant side effects.
The response to placebo was significant, although to a lower degree than for the dexamphetamine. The placebo effect was demonstrated most noticeably in the reduction of the BSI scores. The BSI is a general measure of wellbeing rather than the more specific ADHD symptom scores. This significant placebo effect is not unusual in psychoactive drug trials [19],[20], although it has not been a consistent finding in ADHD research so far [6],[8].
When compared with placebo, dexamphetamine was superior in reducing ADHD symptom scores but was equipotent in reducing the BSI scores. That is, this trial appears to show that dexamphetamine has a greater impact on ADHD symptom scores, specifically, compared with its effect on a patient's general wellbeing.
The differential effect of dexamphetamine on the anxiety scale (significant decrease) and the depressive scale (nonsignificant decrease) is interesting given the literature on ADHD comorbidity in children and adults. As to the possible relationship between ADHD and mood disorders, there is some support for mood disorders being comorbid with ADHD [6], and some support for them to be simply disorders which affect attention, but not commonly comorbid with ADHD [21]. Patients with significant mood disorders were excluded from this trial. The fact that dexamphetamine went on to have a more significant impact on anxiety symptoms rather than depressive symptoms during the trial may suggest some type of ADHD/anxiety overlap in adults which may warrant further investigation.
The dose of dexamphetamine which the patients decided had greatest benefit was about five tablets per day, but did vary from one tablet to seven tablets per day: that is, one patient took more than was suggested. The therapeutic dosage range is yet to be determined for adults but, compared with children, ‘anecdotal evidence suggests that adolescents and adults respond to lower doses for their size’ [3]. As expected, the patients in the placebo treatment group finished the trial on about six tablets per day, the maximum amount allowed. It appeared that the patients on placebo had gone to maximum doses in search of therapeutic benefit, but the dexamphetamine treated patients had settled on a relatively lower dose when given some dosage flexibility in the second 3 weeks of the trial.
The side-effect profile of dexamphetamine appears to be mild in the short term, with the only significant side effect being weight loss. One patient discontinued with the trial after 1 week, possibly related to taking dexamphetamine, although in the longer term the patient responded clinically to combined therapy with dexamphetamine and antidepressant medication.
The comorbidity with substance abuse has been noted with ADHD [1]. In this study, it is worth noting that six possible ADHD patients were excluded from the trial because of current drug usage, and then five more were excluded after inclusion in the trial because of positive urine drug screening (despite denials of current drug usage).
It is probably important to note that this research was not carried out in a research or tertiary centre. The patients assessed for admission into the trial were routine, consecutive referrals from general practitioners to two private psychiatric practitioners experienced in the management of ADHD. In Western Australia, most adult ADHD patients are treated in the private sector. The two psychiatrists endeavoured to assess and treat as they would routinely do, with the aim of testing a hypothesis about dexamphetamine that would be useful in general clinical practice. Socioeconomic status, age and gender distribution indicated that a broad range of patients were assessed and, after randomisation into the two treatment groups, there were no significant differences between the two groups as regards the latter two factors.
Furthermore, in order for this research group to be typical of usual patients seen in clinical practice, modified ADHD scores were used for inclusion/ exclusion criteria as previously discussed. Because the modified scores were marginally lower than the more strict DSM-IV criteria scores, the study is open to possible criticism. However, much thought went into choosing the criteria for caseness, with one of the problems being that ADHD in adults is an emerging entity whose boundaries are not yet clearly defined. In this study, slightly more conservative criteria than that suggested by Murphy and Barkley [11] have been used and the authors believe that this is a reasonable compromise, given the rapidly evolving nature of adult ADHD practice and research.
In summary, dexamphetamine has been shown to be therapeutically useful in the short-term management of adult ADHD. As this is the first trial in this area, replication is needed to validate the findings. In addition, adult ADHD appears to be a long-term condition which will require long-term treatment, and it remains to be shown as to whether dexamphetamine has long-term efficacy.
Footnotes
Acknowledgements
The authors wish to thank: the Health Department of Western Australia for providing a research grant; Sigma Pharmaceuticals for the provision of dexamphetamine placebo tablets; Murray Patterson (Western Australia Health Department) for medication prescription and dispensation; Doctors Tim Threllfall and Dick Allen (Western Australian Health Department) for help with research planning; Sandra Dowds for assistance with manuscript preparation and administration; Nina Keane for administrative assistance; and research staff at the Centre for Clinical Research into Neuropsychiatry, Linda Bradley and Wayne Hill, for data entry and statistical advice. Finally, special thanks go to all the patients and their relatives who participated in the research trial.