A positive total-score comparison does not mean that every component improved more with one medicine. A nine-week lisdexamfetamine-versus-atomoxetine trial reported a more selective pattern of parent-rated functional differences.
A particular prior-treatment population
Nagy and colleagues studied children and adolescents aged 6–17 with ADHD and an investigator-judged inadequate response to methylphenidate. This was not an unselected first-treatment population. Oral lisdexamfetamine was titrated from 30 to 50 or 70 mg/day as needed. Atomoxetine used weight-dependent titration: 40–100 mg/day at 70 kg or above; below 70 kg, approximately 0.5 mg/kg/day initially toward 1.2 mg/kg/day, with a permitted maximum of 1.4 mg/kg/day.
The double-blind, active-controlled trial randomized 267 participants, 133 to lisdexamfetamine and 134 to atomoxetine. Two hundred completed the study: 99 and 101, respectively. Functional outcomes were secondary analyses using the parent-report WFIRS-P, with valid-data counts varying by domain and visit.
What did the ratings show?
Both groups improved from baseline across the six functional domains. The least-squares mean total-score changes were −0.35 with lisdexamfetamine and −0.27 with atomoxetine.
The between-treatment difference favored lisdexamfetamine for total score, with an effect size of 0.27, and for Learning and School and Social Activities, with effect sizes of 0.43 and 0.34. Differences in the other four domains did not reach the reported significance threshold.
Thus within-group improvement in every domain is distinct from a between-treatment difference in every domain. The first does not imply the second.
How certain are the subgrouped outcomes?
Reported p values were nominal and not adjusted for multiple comparisons. The many domain comparisons increase the importance of treating the findings with appropriate caution.
The paper also noted the absence of an established minimum clinically important difference between two active medicines on these functional scales. Statistical significance does not alone establish a large or noticeable benefit to an individual family.
What does the endpoint represent?
Parents completed ratings at baseline and week nine or early termination. The endpoint used the last valid on-treatment assessment. It is not equivalent to complete nine-week follow-up for everyone.
There was no placebo arm in this comparison. Randomization supports the active-treatment contrast, while improvement from baseline can include influences other than a specific treatment effect. Parent perceptions also do not directly measure independently verified grades or social behavior.
The defensible conclusion
The historical trial supports a modest, domain-specific functional advantage under its protocol in youth with inadequate prior methylphenidate response. It does not establish universal superiority across daily life, efficacy in all ADHD populations or an individual driving-safety finding.
Any summary should retain the selected population, short duration, secondary nature of the outcomes and uncorrected statistical testing.
Scientific source
Nagy P, Häge A, Coghill DR, et al. Functional outcomes from a head-to-head, randomized, double-blind trial of lisdexamfetamine dimesylate and atomoxetine in children and adolescents with attention-deficit/hyperactivity disorder and an inadequate response to methylphenidate. European Child & Adolescent Psychiatry. 2016;25:141–149. doi:10.1007/s00787-015-0718-0. Original commentary on the supplied paper.
Leave a Reply