Second-Generation Antipsychotics and Real-World Outcomes in Youth With On- and Off-Label Indications

Objective Second-generation antipsychotic (SGA) treatment has increased in youth globally. However, the effects of these agents on real-world outcomes are largely unknown, especially for non-approved indications. Method This study included a national cohort of incident SGA recipients 7 to 17 years of age in Sweden from 2007 to 2020. Youth were followed from 1 year before SGA initiation through up to 1 year afterward, and in recurrent treatment analyses, up to 13 years after initiation. Within-individual comparisons were used to examine the risk of psychiatric hospitalization, self-harm, accidental injury, and violent crime before and after treatment. Results The study included 21,306 SGA initiators (53.1% male; median age 14.8 years at baseline). Indications included psychosis-related disorders (1,966 [9.2%]), autism spectrum disorders (5,140 [24.1%]), depression/other mood disorders (3,669 [17.2%]), intellectual disability (987 [4.6%]), and attention-deficit/hyperactivity disorder (2,992 [14.0%]). Within a year of initiation, 5,604 (26.3%) experienced psychiatric hospitalization, 1,749 (8.2%) self-harm, 954 (13.6% of those ≥16 years of age) violent crime, and 3,134 (14.7%) accidental injury. Risks were generally elevated immediately before SGA initiation and decreased across subsequent treatment without fully returning to baseline, although there was variation across indications. For example, psychiatric hospitalization and self-harm displayed larger pre-initiation elevations and subsequent decreases for youth with psychosis-related disorders (eg, psychiatric hospitalization: pre-initiation odds ratio [OR pre ] = 11.30 [95% CI, 9.84-12.97]; OR post = 2.59 [95% CI=2.21-3.02]). Conversely, youth with neurodevelopmental disorders demonstrated greater decreases in violent crime (eg, attention-deficit/hyperactivity disorder: OR pre = 1.66 [95% CI, 1.23-2.26]; OR post = 1.26 [95% CI, 0.94-1.70]). Accidental injury risk did not statistically significantly differ (eg, ASD: OR post = 0.90 [95% CI, 0.77-1.07]). Conclusion After having escalated before initiation, risk of psychiatric hospitalization, self-harm, and violent crime decreased during SGA treatment without fully returning to baseline. There was little evidence of associations with accidental injury.