To the Editor: Atomoxetine has been recently introduced for the management of attention deficit hyperactivity disorder (ADHD) (1), and a vigorous campaign is ongoing to encourage physicians to write prescriptions for this drug. A media blitz is being directed to consumers, encouraging them to seek this medication. Before this expensive norepinephrine enhancer is used as a first-line medication to treat ADHD, its advantages relative to the generically prescribed stimulants need to be established. Ideally, a placebo-controlled blinded study model such as the one previously used by us to study another norepinephrine enhancer, imipramine (2), should be used. Because the costs of administering atomoxetine are about $90 per month and generic stimulants cost, on average, about $25 per month, atomoxetine’s role as a first-line therapy should be supported by research.

With this in mind, we evaluated this drug effectiveness in our clinical program by employing measures used routinely to gather data in our program among children who were nonresponders to clinical trials of stimulants.

Seven patients were selected from our clinic (which was previously described [3]). Their average age was 10.5 years, and their IQ was 75.6. Their IQ is deemed average by the New York City Board of Education in its special education program, in which most children have an artificially deflated performance that is most likely consequent to comorbid learning disabilities. All patients were diagnosed with ADHD by using standard DSM-IV criteria. In accordance with the company’s recommendations, we used doses of atomoxetine starting with 0.5 mg/kg/day for 3 days and then increased them up to 1.4 mg/kg/day. Parents of the children consented to treatment in accordance with routine hospital procedure.

We measured behavioral changes at baseline (without drug) and at either 1.2 mg/kg/day or when behavioral exacerbation obligated discontinuation by using the 10-item hyperactivity index derived from the Conners Teacher’s Rating Scale (4).

In this open-label clinical observation of children taking atomoxetine, no change was seen. Tests performed between subjects at baseline (mean=2.66, SD=0.49) and after treatment (mean=2.56, SD=0.51) resulted in t=0.37, df=12, p=0.72, with a 95% confidence interval of –0.48 to 0.68.

Our current clinical observation cautions us that atomoxetine may also be of limited value in children who do not respond to treatment with standard stimulant therapy. The role of atomoxetine needs to be firmly established with a cost-effective analysis if it is to be considered as first-line therapy, and its effectiveness in nonresponders should be demonstrated if it is to be considered for an expensive trial among nonresponders to first-line stimulants.