Atomoxetine, a Novel Treatment for Attention-Deficit-Hyperactivity Disorder

Alisa K. Christman, Pharm.D.; Joli D. Fermo, Pharm.D.; John S. Markowitz, Pharm.D.


Pharmacotherapy. 2004;24(8) 

In This Article

Clinical Trials

The safety and efficacy of atomoxetine were established in six pivotal, randomized, double-blind, placebo-controlled trials and in one open-label comparative trial against methylphenidate.[23,24,25,26,27] These are the largest studies to date evaluating the treatment of ADHD in children and adults. The clinical trials are summarized in Table 2 .[23,24,25,26,27,28,29,30,31,32,33]

In five trials, atomoxetine was evaluated in children and adolescents with ADHD.[23,24,25,26] The data from two of the trials were presented together, as the trials were identically designed.[25] One trial was an open-label trial comparing atomoxetine with methylphenidate.[26]

The investigators in the first trial evaluated atomoxetine twice/day in children and adolescents (aged 8-18 yrs) with ADHD.[23] In this multicenter study, the investigators also evaluated the effects of atomoxetine in poor metabolizers by performing phenotypic testing to analyze for the CYP2D6 genotype in all enrolled subjects. Two hundred ninety-seven participants were enrolled, of which 71% were boys and 29% were girls. In approximately 67% of patients, the diagnosis was for the mixed subtype ADHD (both inattentive and hyperactivity-impulsivity types); 38% had the psychiatric comorbid opposition defiance disorder (ODD). Participants were eligible if they met the DSM-IV-TR criteria for ADHD by clinical assessment and confirmed by a structured interview using the behavioral module of the Kiddie Schedule for Affective Disorders and Schizophrenia for School-Aged Children-Present and Lifetime versions (KSADS-PL) and by a symptom severity score at least 1.5 standard deviations above the age and sex norms on the ADHD Rating Scale-IV-Parent Version: Investigator Administered and Scored (ADHD Rating Scale) for the total score or for either of the inattention or hyperactivity-impulsivity subscales. Exclusion criteria were an IQ less than 80, a serious medical illness, comorbid bipolar disorder or any history of psychosis, history of a seizure disorder, ongoing use of any psychoactive drug other than the study drug, and a history of substance abuse within the previous 3 months.

The primary outcome was an improvement in the symptoms of ADHD assessed with the ADHD Rating Scale and defined by a mean change in the total score from baseline to end point. The hyperactivity-impulsivity and inattention subscales of the ADHD Rating Scale, the Conners' Parent Rating Scale-Revised (CPRS-R) short form, and the Clinical Global Impressions of Severity (CGI-S) assessed secondary outcomes. The Children's Depression Rating Scale-Revised (CDRS-R) was used to assess affective symptoms, whereas the Child Health Questionnaire assessed the change in the subject's social and family functioning. Safety and tolerability were assessed through open-ended questions concerning adverse effects that occurred, as well as regular monitoring of vital signs and laboratory data.

Atomoxetine dosages were 0.5, 1.2, and 1.8 mg/kg/day. After being assessed for depression and anxiety using the KSADS-PL depression and anxiety models, patients were randomly assigned to receive placebo or one of the three dosages of atomoxetine for approximately 8 weeks. All patients in the atomoxetine arm began therapy at 0.5 mg/kg/day, and the dosages of those assigned to the higher dosage arms were later titrated with intermittent steps of 0.8 and 1.2 mg/kg/day at 1-week intervals.

Atomoxetine was determined to be superior to placebo on the primary outcome measure of an improvement in ADHD symptoms in those patients assigned to receive 1.2 and 1.8 mg/kg/day (p<0.05). No difference was noted between the 1.2- and 1.8-mg/kg/day dosages as indicated by the change in ADHD Rating Scale scores. Similar outcomes were seen in the secondary end points of a reduction in the scores of the inattention and hyperactivity-impulsivity subscales, the CGI-S, and the CPRS-R scores. Symptom reduction was the same in children compared with adolescents as seen in ADHD Rating Scale scores. However, older children and adolescents had a significant response to 0.5 mg/kg/day compared with placebo as seen by a mean reduction in ADHD Rating Scale scores (p<0.05). Reduction of affective symptoms, measured by the CDRS-R, was greater between the two higher dosages compared with placebo, as indicated by a change in score (1.2 mg/kg/day group -1.5, 1.8 mg/kg/day group -2.0, placebo group +1.1, p<0.05). Also, improvements in social and family functioning were superior for all atomoxetine dose groups compared with those in the placebo group.

Seventeen participants classified as CYP2D6 poor metabolizers were randomized to treatment. The mean change in the improvement of ADHD Rating Scale scores was the same between the poor and extensive metabolizers. This indicates that dosing adjustments are not necessary in patients classified as poor metabolizers.

The safety and tolerability of atomoxetine were favorable for all dosages. No statistically significant differences were seen in adverse events between the 1.2- and 1.8-mg/kg/day groups. A dose-response effect was suggested with somnolence and anorexia but did not prove to be statistically significant.

The time to onset of effect with atomoxetine at the varying dosages was not assessed. The 0.5-mg/kg/day arm included about half of the patients, as in the other two higher dose treatment arms the dosages were titrated upward, in an attempt by the investigators to provide evidence of a dose-response effect and a threshold dosage for drug effect rather than for efficacy of atomoxetine. This, however, may affect the internal validity of the study as a comparison of dosages in the determination of the primary outcome, as a decrease in ADHD symptoms, may be skewed to reflect that the lower dosage of atomoxetine may not be as effective as the higher dosages. As such, the study data suggest that there is an effect on the symptoms of ADHD with atomoxetine 0.5 mg/kg/day, but that there is a graded response as the dosage is titrated upward. Based on the results of this study, there does not appear to be a greater improvement in symptoms beyond the 1.2-mg/kg/day dosage, although the 1.8-mg/kg/day dosage was well tolerated.

A second trial[24] performed by the same group of investigators was a dose-ranging study to evaluate the efficacy of once-daily administration of atomoxetine in participants aged 6-16 years who met the same criteria for diagnosis and assessment of ADHD as those in the previous study. The primary objective was to provide evidence that atomoxetine was effective for the treatment of ADHD when given once/day, as measured by a change in the total score of ADHD Rating Scale from baseline to end of study (response > 25% reduction from baseline in total score on the ADHD Rating Scale). Secondary end points were a reduction in the ADHD Rating Scale subscales of inattention and hyperactivity-impulsivity, and a reduction in CPRS-R, Conners' Teacher Rating Scale-Revised (CTRS-R), CGI-S, and a change in family and social behavior as assessed by a parent-rated diary developed specifically for this study. Patients were excluded if they had a history of substance abuse (> 3 mo), a serious medical illness, comorbid bipolar disorder or any history of psychosis, history of a seizure disorder, or ongoing use of a psychoactive drug other than the study drug.

One hundred seventy-one patients were randomly assigned to receive either atomoxetine or placebo; of these 171 patients, 70.6% were boys and 29.4% were girls. One assigned patient did not receive any drug and therefore was excluded from all analyses. Fifty-five percent of patients in the atomoxetine arm had ADHD of the mixed subtype. The most common comorbid psychiatric disorder was ODD, occurring in 18.8% of the patients in the atomoxetine arm. Fifty-five percent of the patients in the total population had been treated with a stimulant. The treatment period was 6 weeks.

Patients in the atomoxetine arm were started at 0.5 mg/kg/day for 3 days, then the dosage was increased to 0.75 mg/kg/day for the remainder of the week. After the first week, the dosage was increased to 1.0 mg/kg/day for 4 weeks, when efficacy was assessed by using the CGI-S scale to determine severity of symptoms. Those who had a score greater than two, indicating more than minimal symptoms, had a further dosage increase to 1.5 mg/kg/day, which remained as the maximum dosage throughout the study. In addition to using the CGI-S score, efficacy was measured by using the ADHD Rating Scale, the CPRS-R, the CTRS-R, and a 13-item parent-rated diary developed by the investigators to assess the efficacy of atomoxetine during the evening and early morning periods. The safety and tolerability of the two dosages were assessed through the use of open-ended questions and frequent monitoring of vital signs.

Overall improvement seen in the ADHD Rating Scale total score indicated that treatment with atomoxetine was superior to placebo, with a 59.5% response in the atomoxetine group versus 31.3% in the placebo group when comparing scores from baseline to end point. Secondary efficacy was achieved by the reduction in the mean changes of the CGI-S score from baseline and in the mean change in score of the CPRS-R (p<0.001 for both measures), indicating that atomoxetine was superior to placebo (p=0.003), with 28.6% of atomoxetine-treated patients demonstrating a reduction of symptoms versus 9.6% of patients taking placebo. Atomoxetine significantly reduced the CTRS-R scores as well (p=0.016). No significant differences were seen in the parent ratings of offspring behavior, suggesting interuser variability when assessing symptoms and questioning objectivity of the results. Statistical significance was shown for atomoxetine in only two items -- inattention and distractibility in the evening (p=0.003), and difficulty settling at bedtime (p=0.03) -- suggesting that the drug may decrease evening symptoms.

The most important finding of the study was the clinically significant effects of atomoxetine on symptom control throughout the day. Despite its relatively short plasma half-life (approximately 4 hrs in extensive metabolizers), the duration of effect of atomoxetine persisted into the evening after a once-daily dose was given in the morning. This effect was seen in the individual analysis of the symptom of inattention from the daily diary, suggesting a drug-specific benefit occurring late in the day and early evening. Unfortunately, a comparative twice-daily dosing arm was not included to determine the relative efficacy of once-daily versus twice-daily dosing. Evaluation of these data might suggest that once-daily dosing may be as effective as twice-daily dosing in producing symptom reduction; however, adequately sized direct comparisons are needed before definitive conclusions can be drawn.

Another group of authors presents evidence that once-daily atomoxetine therapy provides continuous symptom relief throughout the day when given as a single daily dose in the morning.[28] Their study, performed in children aged 6-12 years, demonstrated that atomoxetine 1.3 mg/kg/day was significantly more effective than placebo in reducing the core symptoms of ADHD by about 40% in atomoxetine-treated patients versus 17% in placebo-treated patients (total score ADHD Rating Scale, p<0.05). Likewise, continued efficacy in reducing ADHD symptoms into the evening hours and through the night was determined through evaluation of Daily Parent Ratings of Evening and Morning Behavior-Revised (DPREMB-R) scores (a reduction in the total score of 44% for atomoxetine-treated patients vs 29% reduction in placebo-treated patients). The DPREMB-R total score decrease for the atomoxetine group during the first week of treatment was significantly different from that of the placebo group after 1 day of treatment, indicating a rapid onset of effect at a dosage of 0.8 mg/kg/day (p<0.001).

The third and fourth trials to evaluate atomoxetine in children and adolescents were conducted by the same group.[25] The two identical randomized, double-blind, placebo-controlled, proof-of-concept trials were run in parallel and evaluated the efficacy of atomoxetine at a maximum dosage of 2.0 mg/kg/day, admin-istered twice/day. A smaller methylphenidate treatment arm was used in the event that atomoxetine showed no difference from placebo. The study period was 12 weeks, which included an additional 2-week drug washout period and a 1-week drug discontinuation period. The primary objective was to determine the possibility of atomoxetine as an alternative to stimulant therapy.

Two hundred ninety-one patients aged 7-12 years were randomly assigned to receive atomoxetine (65 patients in trial 1, 64 in trial 2), methylphenidate (20 in trial 1, 18 in trial 2), or placebo (62 in trial 1, 62 in trial 2). In both trials, 81% of the patients were boys and 19% were girls. Inclusion criteria were slightly different from those of previous studies. Patients needed to meet the DSM-IV-TR criteria for ADHD of the inattentive type confirmed by a clinical interview using the KSADS-Episodic:ADHD, and a symptom severity score greater than 1.5 standard deviations above age and sex norms on the ADHD Rating Scale. There was one major difference from enrollment of other studies. Whereas previous studies included patients who were CYP2D6 poor metabolizers, this study excluded any patient who had a CYP2D6 poor metabolizer genotype. Other exclusion criteria were a history of substance abuse, a serious medical illness, comorbid bipolar disorder or any history of psychosis, history of seizure disorder, and weight less than 25 kg.

Participants were stratified into two arms: one arm comprised patients with prior treatment with a psychostimulant, the other arm comprised patients with no prior psychostimulant treatment. Those who had been stratified to the previous psychostimulant arm received either atomoxetine or placebo; the patients in the other arm, without previous psychostimulant treatment, were randomly assigned to receive atomoxetine, methylphenidate, or placebo.

Primary efficacy was achieved as a reduction in the total ADHD Rating Scale score for all patients in the comparison of atomoxetine with placebo (trials 1 and 2, p<0.001). No significant differences in efficacy were noted between methylphenidate and atomoxetine when given to stimulant-naïve patients. In addition, atomoxetine achieved significance in reduction of ADHD Rating Scale total score for patients previously treated with a stimulant (trial 1, p<0.001; trial 2, p=0.048) and in reduction of inattention (trials 1 and 2, p<0.001) and hyperactivity-impulsivity (trial 1, p<0.001; trial 2, p=0.002) subscales of the ADHD Rating Scale, compared with placebo. Secondary efficacy was achieved for atomoxetine in the reduction of the CGI-S score for both studies (trial 1, p=0.003;trial 2, p=0.001). All patients treated with atomoxetine were classified as responders, defined by a 25% or greater reduction in ADHD Rating Scale scores (trial I, atomoxetine 61.4% vs placebo 24.6%, p<0.001; trial 2, atomoxetine 58.7% vs placebo 40.0%, p=0.048). This indicated that selective inhibition of norepi-nephrine transport would provide a reduction in ADHD symptoms in school-aged children and a nonstimulant option for ADHD treatment.

These studies demonstrate that atomoxetine is effective in children and adolescents, when compared with placebo, for the treatment of ADHD. Interpreting the possible comparative efficacy of atomoxetine with that of methylphenidate is difficult, as the use of methylphenidate in these studies was intended to validate the study design in the event that atomoxetine showed no difference from placebo. Also, whether the frequency of adverse effects was increased at the higher dosages of atomoxetine was not reported.

A number of post hoc analyses from the two proof-of-concept studies[25] have examined the efficacy of atomoxetine in children with comorbid ODD, children who had failed psychostimulant therapy, children with ADHD of the inattentive type, and girls with ADHD.[29,30,31,32] All analyses were favorable for atomoxetine in reducing symptoms of ADHD. In the analysis involving children with comorbid ODD, the reduction in symptoms associated with ODD with atomoxetine was not statistically significant. The results of the studies are summarized in Table 2 .[23,24,25,26,27,28,29,30,31,32,33]

A prospective, randomized, open-label trial assessed the comparability of atomoxetine and methylphenidate for 10 weeks in children and adolescents.[26] The data reported are from a relapse-prevention study in which previous responders to stimulant ADHD therapy were enrolled to determine how a nonstimulant therapy would compare with the traditional stimulant therapy. This study enrolled boys aged 7-15 years and girls aged 7-9 years who met the DSM-IV-TR diagnostic criteria for ADHD. After an evaluation and washout period, patients were randomly assigned to receive open-label treatment with either atomoxetine or methylphenidate for 10 weeks. The study enrolled 228 patients. Randomization was based on a 3:1 ratio (atomoxetine:methylphenidate) with a block size of four for the first four patients. For the remaining patients, randomization was based on a 5:1 ratio with a block size of six. Block randomization was used to provide a balance for the early treatment phase and provide room for internal decision making.

At study entry, patients were tested to determine whether they were poor or extensive metabolizers through the analysis of DNA taken from whole-blood samples for CYP2D6 poor metabolizer alleles. Patients determined to be poor metabolizers were started at a lower dosage of 0.2 mg/kg/day, with the dosage titrated to a maximum of 1.0 mg/kg/day. The dosage for the extensive metabolizers was titrated to a maximum of 2.0 mg/kg/day. Patients assigned to the methylphenidate arm were started at 5 mg 1-3 times/day, with dosage titration based on the investigators' clinical assessment of patient response and tolerability. The maximum daily dosage of methylphenidate was 60 mg/day.

No differences were noted between the groups for the study's primary and secondary end points. Primary efficacy was defined by a total change in ADHD Rating Scale score from baseline to the end of the study period. Secondary efficacy was defined as a reduction of the total ADHD Rating Scale (parent), CPRS-R, CTRS-R, and CGI-S (ADHD subscale).

No girls were assigned to the methylphenidate arm and only 17 were assigned to the atomoxetine arm. The small sample (44 patients) treated with methylphenidate limited the comparability of effects between atomoxetine and methylphenidate, although the lack of significant differences noted in the outcomes suggest that the two are not different in terms of efficacy. The smaller size of the methylphenidate arm will inflate any effects seen, thus allowing the data to appear comparative to atomoxetine for effect. Larger studies with comparable sample sizes between the atomoxetine and methylphenidate arms may clearly show the superiority of one agent over the other in efficacy. In addition, a gradual dose-titration design as well as the varying schedule of methylphenidate dosing at once/day to 3 times/day confound the time and dose effects of the two therapies, therefore making it impossible to assess for onset of effect for both groups. Effects of methylphenidate may be seen earlier in treatment based on its pharmacologic effect. In patients with severe ADHD, use of a pharmacologic agent that has a proved rapid time to effect may be more favorable to an agent that may take longer before a clinical effect is seen. The final mean dosage of methylphenidate was 18.7 mg/day, which for most patients may be an average dosage for maintenance of ADHD.

By study end, the average dosage of atomoxetine in the poor metabolizer group was one third of the dosage in the extensive metabolizer group (approximately 0.5 vs 1.5 mg/kg/day). The rate of adverse effects was not larger in the poor metabolizer group than in the extensive metabolizer group; however, the small sample of poor metabolizers limits the interpretability of these results. Previous studies with larger populations of poor metabolizers evaluated similar dosages of atomoxetine and determined that the safety and tolerability of these dosages are similar for both groups.[23]

The study was open-label; therefore, bias may have been introduced by the investigator and parent assessments of symptoms based on their expectations of the treatment. In addition, the groups were not well matched in terms of sex, in that there were significantly more boys than girls for each group and no girls were randomly assigned to the methylphenidate group. Bias also may have been introduced into the results for atomoxetine as it is thought that girls with ADHD do not exhibit as aggressive of symptoms as do boys and therefore may appear to respond more favorably to treatment. Comparison with a stimulant may show greater reduction in symptoms based on previous results evaluating the efficacy of methylphenidate. Finally, the effects of either treatment on school behavior in comparison to home behavior are questionable, owing to the lack of direct teacher assessments. Conclusions drawn from improvements in school functioning are based on parent reports, the validity of which was not confirmed.

Two identical trials conducted by the same group evaluated the efficacy of atomoxetine in adults older than 18 years.[27] These were double-blind, placebo-controlled, phase III trials to assess the efficacy of atomoxetine 60-120 mg/day given in two doses for 10 weeks. Two hundred eighty participants were enrolled and randomly assigned to receive atomoxetine or placebo in trial 1. Two hundred fifty-six patients were enrolled in trial 2. In trial 1, 64% of patients were men and 36% were women; in trial 2 66% were men and 34% were women. Patients were eligible if they had moderate-to-severe symptoms of ADHD based on the Conners' Adult ADHD Diagnostic Interview (CAARS-INV) for the DSM-IV-TR. Exclusion criteria were a history of substance abuse (previous 3 mo), a serious medical illness, comorbid depression or bipolar disorder or any history of psychosis or anxiety, ongoing use of psychoactive drugs other than the study drug, hypo- or hyperthyroidism, and a history of a seizure disorder.

Primary efficacy was defined as changes in the total and subscale scores of the CAARS-INV from baseline to end of study. A significant reduction was achieved in the total CAARS-INV score at study end for both trials (trial 1, p=0.062; trial 2, p=0.002) in patients treated with atomoxetine. For the inattention (trial 1, p=0.010; trial 2, p=0.001) and hyperactivity-impulsivity (trial 1, p=0.017; trial 2, p=0.012) subscales of the CAARS-INV, a significant reduction in scores indicated the superiority of atomoxetine over placebo for an improvement of these symptoms in both trials. Secondary efficacy was defined and achieved in both trials as a change in the total and subscale scores of the patient-rated tests: the CAARS-Self (trial 1, p=0.003; trial 2, p=0.008), CGI-S (trial 1, p=0.011; trial 2, p=0.002), and the Wender-Reimherr Adult Attention Deficit Disorder Scale (trial 1, p=0.001; trial 2, p=0.041) for atomoxetine over placebo. Significant differences were not found in trial 1 concerning improvements in social, family, and work functions as measured by the Sheehan Disability Scale total and domain scores. In trial 2, however, a significant difference was noted in the improvements of social, family, and work functioning for the total (p=0.022) and the domain (p=0.007) scores. This indicates that treatment with atomoxetine in adults with ADHD may improve quality of life by improving social and work functioning; however, future studies will be required to truly determine this effect. Limitations to the study include the lack of a comparative arm with a stimulant, such as methylphenidate, to evaluate treatment efficacy against the standard therapy in this population.