Meta-analysis of aggression or hostility events in randomized, controlled clinical trials of atomoxetine for ADHD

Background We systematically examined potential aggression/hostility-related events in a meta-analysis of acute clinical trials of atomoxetine for attention-deficit/hyperactivity disorder (ADHD). Methods Pediatric patients from 14 trials of atomoxetine were subdivided into a placebo-controlled (atomoxetine n = 1308, placebo n = 806) or active comparator databases (atomoxetine n = 566, methylphenidate n = 472). A third database comprised adult patients from placebo-controlled studies (atomoxetine n = 541, placebo n = 405). A computerized search of adverse events and comments identified patients with potential aggression/hostility events. Mantel–Haenszel incidence differences (MHID) were calculated. Results In the placebo-controlled database, we observed 21 atomoxetine and 9 placebo patients with reported aggression/hostility events, MHID of .6% (95% confidence interval [CI]: –.4, 1.7). In the active comparator database, there were seven events in atomoxetine and four in methylphenidate patients, MHID = .2% (95% CI: –1.0,1.3). In the adult database, there were no events in 0 atomoxetine and one placebo patient, MHID = –.3% (95% CI: –.8, .2). Conclusions Aggression/hostility-related events occurred in less than 2% of patients and were more frequent in pediatric patients treated with atomoxetine versus placebo (risk ratio of 1.33; not statistically significant). The risk of aggression/hostility events was similar in patients treated with atomoxetine or methylphenidate.