Cardiovascular Safety of Stimulants in Children with Attention-Deficit/Hyperactivity Disorder

To The Editor:

In their analysis of data from a longitudinal cohort study in Denmark, Dalsgaard et al. (2014) recently reported that treatment with stimulants doubles the risk of an adverse cardiovascular event in children and adolescents with attention-deficit/hyperactivity disorder (ADHD). Although the large sample size and the study design are major strengths, the findings of this study as reported may generate more concern than is warranted.

To begin with, the article's title – with its focus on cardiovascular safety – is somewhat misleading. The outcome measure of this study was not safety. Rather, this study focused on “cardiovascular events,” defined very broadly. The majority of patients with cardiovascular events had a relatively nonspecific diagnosis of either “heart disease not otherwise specified (NOS)” (14%) or “cardiovascular disease NOS” (40%); these cases were all classified by the authors as “nonserious.” Only five children with ADHD experienced a serious cardiac event (e.g., cardiac arrest, uncompensated heart disease, and ischemic heart disease) and none had a history of stimulant use. Likewise, of the six children with a history of stimulant use who died during the study period, the authors note that none died from cardiovascular causes.

Second, it is important to note that although there was a statistically significant twofold increase in the risk of cardiovascular events with stimulant treatment, the incidence in both the ADHD group and the general population was extremely low. The authors themselves acknowledge that cardiovascular events were “rare.” The total rate of cardiovascular events among children with ADHD in the study was 170 events per 100,000 person years.

Most importantly, we are concerned that referral bias may have significantly skewed the results of this study. Although the authors themselves acknowledged the possible influence of “systematic information bias” in their discussion, this likely confound cannot be overemphasized, and may explain some or all of the observed differences. Given that cardiovascular safety of stimulants has received considerable attention in both the medical literature and popular media, it is likely that children being treated with stimulants are either screened more carefully for cardiovascular problems or referred more frequently for evaluation.

Apart from this major concern about referral bias, the article raises many other concerns. For example, it is unclear who in the general population other than youth with ADHD were taking stimulant medications. Also, the cohort included youth treated with amphetamines; however, there is no mention in the results whatsoever about this subset of stimulant users. Lastly, we believe that there is an error in Table S4. It states that the highest daily dose ever for child #3 was 30 mg; however, this child is classified as having been on a high dose (i.e., >30 mg) at time of death.

In short, although clinicians should be vigilant in screening for cardiac risk factors and symptoms in children for whom stimulant treatment is being considered, the findings in this study are in many ways very reassuring. Referral bias may explain some or all of the findings, and contrary to the article's title, it does not suggest that stimulants pose a significant safety risk to children. Cardiovascular events in children on stimulants were rare, serious cardiac events were noted only in children not on stimulants, and no deaths were attributed to stimulant therapy.