Background: ADHD medications increase clinical encounters for cardiovascular symptoms. Uncertain are the roles of differences in ADHD medications and restrictive practices by drug programs.
Methods: We conducted two nested case-control studies. The first was nested within a cohort of children de novo users of methylphenidate, amphetamines or atomoxetine and the second case-control study was nested within a subcohort of de novo amphetamine or atomoxetine users with no cardiovascular events prior to the first dispensing of either drug. The outcome for both studies was the composite of physician visits, emergency room visits or hospitalizations for cardiovascular reasons. Cases were matched on sex, age and date of entry within the cohorts, with up to 10 controls. Patients with an active dispensation of ADHD medications at the index date (and up to 90 days previously) were considered exposed. Conditional logistic regression was used to calculate odd ratios (OR).
Results: The full cohort comprised 38,495 patients. Among these patients, 3595 (9.3%) had no prior cardiovascular events (the subcohort). In the full cohort, an association was demonstrated with exposure to amphetamine and atomoxetine (but not methylphenidate) and the cardiovascular encounter outcomes. When the sub-cohort was analyzed the associations with amphetamine or atomoxetine were no longer evident.
Conclusion: Reimbursement policies need to be considered when conducting observational studies. Had the analysis been conducted without consideration of these policies the results would have incorrectly identified amphetamine and atomoxetine as important risk factors for cardiovascular encounters.