Efficacy and Safety of Lisdexamfetamine Dimesylate and Atomoxetine in the Treatment of Attention-Deficit/Hyperactivity Disorder: a Head-to-Head, Randomized, Double-Blind, Phase IIIb Study

This study enrolled male and female patients (aged 6–17 years) who satisfied Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR) [17] criteria for a primary diagnosis of ADHD of at least moderate severity as shown by a baseline ADHD Rating Scale IV (ADHD-RS-IV) total score of 28 or higher. Inclusion and exclusion criteria related to a patient’s previous exposure and/or response to ADHD medications are outlined in Table 1. Other inclusion criteria included age-appropriate intellectual functioning; ability to swallow a capsule; and blood pressure measurements within the 95th percentile for age, sex and height. Female patients of childbearing potential were required to have a negative urine pregnancy test at baseline and to comply with the contraceptive requirements of the protocol. Other key exclusion criteria included comorbid psychiatric diagnosis with significant symptoms (based on Kiddie-Schedule for Affective Disorders and Schizophrenia for School Age Children—Present and Lifetime diagnostic interview); conduct disorder (excluding oppositional defiant disorder); suicide risk, with a previous suicide attempt or active suicidal ideation; pregnancy or lactation; weight below 22.7 kg; body mass index (BMI, kg/m²) greater than the 97th percentile for age and sex; positive urine drug test (with the exception of a patient’s current ADHD medication); clinically significant electrocardiogram (ECG) results or laboratory abnormalities; known CYP2D6 poor-metabolizer genotype; suspected substance abuse or dependence disorder (excluding nicotine) within the previous 6 months; history of seizures; tics or Tourette’s disorder; pre-existing liver disease or laboratory evidence of liver disease; known cardiac structural abnormality; or any other condition that might increase vulnerability to the sympathomimetic effects of a psychostimulant drug.

Patients were required to discontinue any psychoactive medication for a 7-day washout period prior to baseline (visit 0). At baseline, patients were randomized (1:1) to receive a once-daily, morning dose (at 07:00 ± 2 h) of LDX or ATX for a 9-week, double-blind evaluation period (Fig. 1) with weekly, on-site efficacy, tolerability and safety assessments. Dosing began on the morning after the baseline visit and continued for 9 weeks, starting with a 4-week, stepwise, dose-optimization stage. Randomization of patients was stratified by country, and an automated interactive response system was used to generate the random (concealed) allocation sequence and assign participants to study treatments; patients, caregivers and investigators were blinded to the treatment allocation. All study drugs were over-encapsulated so they appeared identical.

The dose-optimization phase involved adjustment of the dose until an ‘acceptable’ response was achieved [defined as a reduction of at least 30 % from baseline in the ADHD-RS-IV total score and a Clinical Global Impressions-Improvement (CGI-I) score of 1 or 2 with tolerable side effects]. Only one dose reduction was permitted during the optimization phase and, following dose reduction, further increases were not allowed. Dose adjustments were not permitted beyond visit 3, and patients who were unable to tolerate the study drug were withdrawn from the study.

LDX was provided in a single capsule of 30, 50 or 70 mg, with patients initially receiving a 30-mg dose. ATX was available in 10-, 18-, 25-, 40- and 60-mg capsules. All patients in the ATX group who weighed less than 70 kg were started on a daily dose of approximately 0.5 mg/kg body weight, the final target daily dose being 1.2 mg/kg, with a maximum permitted daily dose of 1.4 mg/kg. Patients who weighed 70 kg or more initially received 40 mg and, if required, were titrated to 80 mg and then to 100 mg daily. Some patients treated with ATX would need two capsules to achieve the required dose (e.g. 80 and 100 mg were achieved using two capsules). Therefore, all patients weighing more than 64.5 kg who were titrated to a higher dose were instructed to take two capsules (the second capsule could be either active drug or placebo, as appropriate) to maintain the double-blind study design.

The primary efficacy measure was the time to first clinical response (days) after the initiation of treatment, as assessed by CGI-I scores [18]. The CGI scale provides a global assessment of a patient’s severity of illness; CGI-I scores, which were reported at each post-baseline visit, rate the change in a patient’s condition from baseline on a scale of 1 (very much improved) to 7 (very much worse) [18]. Clinical ‘response’ was defined as a CGI-I score of 1 or 2 (very much improved or much improved). The number of days to first clinical response was calculated as the date of response minus the date of first dose, plus 1 day.

Secondary efficacy outcomes included the proportion of CGI-I responders at each study visit, the proportion of patients who had a decrease of at least one CGI-Severity (CGI-S) category from baseline (assessed at visit 4 and at visit 9), and the change from baseline in ADHD-RS-IV total and subscale scores at each study visit. CGI-S scores, which rate the severity of a patient’s condition on a scale of 1 (normal, not at all ill) to 7 (among the most extremely ill), were assessed at baseline, visit 4 and visit 9 [18]. The ADHD-RS-IV scale [19], designed to reflect current ADHD symptoms, assesses 18 items on a scale of 0 (no symptoms) to 3 (severe symptoms), with a total score ranging from 0 to 54. CGI and ADHD-RS-IV assessments were performed by a medical practitioner or psychologist experienced in the evaluation of children and adolescents with ADHD.

Tolerability and safety assessments included the monitoring of treatment-emergent adverse events (TEAEs), laboratory evaluations, physical examination (including weight), and monitoring of vital signs and ECG parameters. TEAEs were defined as adverse events that first occurred or worsened during the time between the first dose of study drug and the third day following cessation of treatment (inclusive). All TEAEs were coded using Medical Dictionary for Regulatory Activities (MedDRA) (version 14.1) [20]. A serious TEAE was any untoward medical occurrence that resulted in death; was life-threatening; required inpatient hospitalization or prolonged existing hospitalization; resulted in persistent or significant disability/incapacity; was a congenital abnormality/birth defect; or was an Important Medical Event. Important Medical Events may have been considered as serious TEAEs when, based upon medical judgement, they may have jeopardized the patient and may have required medical or surgical intervention to prevent one of the outcomes listed above. The sponsor required any new onset of seizures, syncope or loss of consciousness to be reported as a serious TEAE. Using Centers for Disease Control and Prevention percentile growth charts [21], BMI was categorized into five groups based on BMI percentiles: underweight (<5 %), healthy weight (low; ≥5 to <25 %), healthy weight (high; ≥25 to <85 %), at risk of being overweight (≥85 to <95 %) and overweight (≥95 %). ECG parameters were assessed at screening and visit 4. However, the visit 4 ECG was added as a result of a protocol amendment requested by the French Central Ethics Committee and therefore was not obtained for all patients. The Brief Psychiatric Rating Scale for Children (BPRS-C), the Columbia-Suicide Severity Rating Scale (C-SSRS) and the Udvalg for Kliniske Undersøgelser Side Effect Rating Scale-Clinician (UKU-SERS-Clin) were also used to monitor patient tolerability and safety plus the suitability of individuals to remain in the study [22‐24].

Safety/tolerability assessments were performed using the safety population, defined as all patients who were randomized and received at least one dose of study drug. Efficacy data were analysed using the full analysis set (FAS), also defined as all patients who were randomized and received at least one dose of study drug. One patient was randomized to ATX but received LDX owing to a drug dispensing error. Based on the intent-to-treat principle, this patient was included in the ATX treatment group in the FAS, but was counted in the LDX treatment group in the safety population. Endpoint was defined as the last on-treatment, post-baseline visit with a valid assessment.

Time to first clinical response (days) was calculated using Kaplan–Meier estimates and analysed using a Peto–Peto–Prentice–Wilcoxon (PPPW) test [25, 26], stratified by country and evaluated at a significance level of 0.05 (two-sided). The null hypothesis stated that there was no difference in the time to first clinical response between patients receiving LDX and those taking ATX, with the two-sided alternative of a non-zero difference between the groups. Allowing for a 20 % discontinuation rate, approximately 262 patients (131 in each treatment group) were required to detect a difference in time to first clinical response between the treatment groups with a power of 85 %. Patients who prematurely discontinued from the study without responding, and patients who completed the study up to visit 9 without meeting response criteria, were censored at visit 9 in the primary analysis of time to response, and classified as non-responders in the analysis of responders.

The proportion of responders (CGI-I score of 1 or 2) at each study visit and the proportion of patients who had a decrease of at least one CGI-S category by visit 9 were assessed using the last-observation-carried-forward (LOCF) approach and analysed using a Cochran–Mantel–Haenszel (CMH) test stratified by country. At each study visit, the change from baseline in ADHD-RS-IV scores, using LOCF, was analysed using an analysis of covariance (ANCOVA) model including treatment group (effect of interest), country (blocking factor) and the corresponding baseline score (covariate). Effect sizes were calculated as the difference in least-squares (LS) mean score between the two treatments, divided by the root mean square error obtained from the ANCOVA model. Effect sizes of 0.2, 0.5 and 0.8 correspond to small, medium and large magnitudes of effect, respectively [27].