Antipsychotic Use May Hike Diabetes Risk in Children

Initiating antipsychotics may raise a child's risk for type 2 diabetes by nearly 50%, according to data.

Initiating antipsychotics may increase a child’s risk for type 2 diabetes by nearly 50%, with the risk being even higher among children who are also receiving antidepressants, new data published in JAMA Pediatrics indicate.

Recently, second-generation antipsychotics have been prescribed more frequently to children enrolled in Medicaid. Overall, more than 25% of Medicaid-enrolled children receiving prescription medications for behavioral problems were prescribed antipsychotics by 2008, usually for less severe disorders like attention-deficit/hyperactivity (ADHD) disorder, according to a press release.

“With such vast numbers of children being exposed to these medications, the implications for potential long-lasting harm can be jarring,” lead study author David Rubin, MD, MSCE, co-director of PolicyLab at the Children’s Hospital of Philadelphia, said in the release.

Although second-generation antipsychotics have been associated with adverse metabolic effects, delineating the connection between these medications and type 2 diabetes has proven difficult, according to background information in the study.

To learn more, Rubin and colleagues conducted a retrospective national cohort study of children enrolled in Medicaid between 2003 and 2007.

Those aged 10 to 18 years with a mental health diagnosis and enrolled in a Medicaid fee-for-service arrangement were included in the analysis, while those with chronic steroid exposure, a diabetes diagnosis or use of second-generation antipsychotic use during 1-year look-back period were excluded. The study included 107,551 children who initiated second-generation antipsychotic use and 1,221,434 who did not.

Mean follow-up was 17.2 months, according to the data.

Results revealed an increased risk for incident diabetes among children who initiated second-generation antipsychotic use, as compared with those who did not (OR=1.51; 95% CI, 1.35-1.69).

Further, diabetes risk was higher among children who were also taking antidepressants at the time of initiating second-generation antipsychotic use (OR=1.54; 95% CI, 1.17-2.03) than those who were initiating second-generation antipsychotic use only. This was not the case, however, for children who were simultaneously using stimulants.

The researchers also found that, compared with those initiating risperidone, risk for diabetes was higher among children initiating treatment with ziprasidone (OR=1.61; 95% CI, 0.99-2.64) and aripiprazole (OR=1.58; 95% CI, 1.21-2.07), but not quetiapine fumarate or olanzapine.

However, the researchers noted that caution should be exercised when interpreting their findings. For example, baseline risk for diabetes among children not exposed to antipsychotics in the study was only 1 in 400, rising to 1 in 260 among those initiating antipsychotics to 1 in 200 among those who initiated antipsychotics while also receiving antidepressants, according to the release.

“Although these findings should certainly give us pause, we should not reflexively overreact to them. Rather, we need to incorporate these new revelations about the risk for diabetes into a more thoughtful consideration of the true risks and benefits of prescribing an antipsychotic to a child,” Rubin said, emphasizing that in some cases, the advantages of treatment with antipsychotics may outweigh the possible harms.

The researchers concluded that the trend toward prescribing antipsychotics to children and teens likely won’t change as solutions are sought for challenging behavior in children.

“At the end of the day, the approach to the individual child who is in crisis is still a case-by-case decision between a family and the treating provider,” Rubin said. “We can only hope that those decisions are made in full recognition of our findings, and that for some children, alternatives to these powerful medications — such as counseling or supportive services — will be considered first.”


  1. Rubin DM et al. JAMA Pediatr. 2015;169(4):e150285.