Key clinical point: The risk of psychosis appeared higher among patients with ADHD who were prescribed amphetamine rather than methylphenidate.
Major finding: There were 106 episodes of psychosis among patients receiving methylphenidate (0.10%) and 237 new cases among patients receiving amphetamine (0.21%).
Study details: Data from two commercial insurance claims databases on 221,846 patients aged 13-25 years with ADHD and received methylphenidate or amphetamine.
Disclosures: Dr. Moran reported receiving grants from National Institute of Mental Health (NIMH). The other authors reported grants, personal fees, and other relationships with several entities, including Boehringer Ingelheim, the Food and Drug Administration, the NIMH, and Takeda.
Moran LV et al. N Engl J Med. 2019. doi: 10.1056/NEJMoa1813751.
The findings by Moran et al. are consistent with other randomized controlled trials that suggest a better safety profile for methylphenidate over amphetamine. But the data cannot determine causality in this patient population, Samuele Cortese, MD, PhD, wrote in a related editorial.
“The findings of the current study should not be considered definitive. Observational studies such as this one can provide information on uncommon adverse events in real-world clinical practice that are challenging to assess in randomized trials performed over brief periods,” he said. “However, even sophisticated approaches, such as the ones used in this study to address possible biases, do not have the advantages of randomized trials in excluding confounding factors.”
It is still unclear why some patients developed psychosis, such as in cases of patients with stimulant use and had a “low” or “high” vulnerability to developing psychosis after exposure. The lack of association between psychosis and prescribing amphetamines among psychiatrists also might indicate that those clinicians identified risk factors in patients that predicted the development of psychosis and thus avoided prescribing amphetamines to these patients, he said.
“Currently, it is not possible to predict which patients will have psychotic episodes after stimulant treatment,” Dr. Cortese concluded. “Perhaps techniques such as machine learning applied to large data sets from randomized trials, combined with observational data, will provide predictors at the individual patient level.”
Dr. Cortese is affiliated with the Center for Innovation in Mental Health at the University of Southampton (England). These comments summarize his accompanying editorial (N Engl J Med. 2019. doi: 10.1056/NEJMe1900887 ). He reported nonfinancial relationships with the Association for Child and Adolescent Central Health and the Healthcare Convention & Exhibitors Association.