4F-05 - Long-term benzodiazepine treatment among adolescents and young adults in the United States, 2017-2022

Background: Although long-term benzodiazepine (LT-BZD) treatment can be beneficial in certain situations, benzodiazepine (BZD) poses a risk of misuse and dependence, necessitating short-term treatment recommendations. Yet, there remains limited data on the prevalence of LT-BZD treatment in adolescents and young adults, with estimates varying widely, outdated, or focused on older population. Further, such data are lacking for youth with substance use disorders (SUD), who face heightened risks of BZD misuse and dependence.

Objectives: To estimate the frequency of LT-BZD treatment in US adolescents and young adults and determine whether those with SUD have higher risk of LT-BZD use.

Methods: We used MarketScan (2016-2022) commercial claims data and included adolescents (12-17y) and young adults (18-24y) newly initiating BZD treatment with a ≥1 year washout period. We excluded patients with < 180 days of enrollment after BZD initiation to enable assessment of LT-BZD use (≥180 days of BZD use, allowing a 30-day gap). The annual proportion of new BZD initiators becoming LT-BZD users was estimated by age group and SUD diagnosis. Trends were evaluated via linear regressions, and logistic regressions were used to assess the association between baseline SUD diagnosis and LT-BZD use, reporting adjusted odds ratios (aOR) accounting for diagnostic, medication, and healthcare utilization variables.

Results: We identified 130,110 BZD initiators, including 27,028 adolescents (40% male) and 103,082 young adults (33% male). Three percent (3.2%) of adolescents and 9.2% of young adults had a SUD diagnosis. Those with SUD were more often male, and had higher rates of drug poisoning, mental health disorders, suicidal ideation, and hospitalizations. Five percent (4.7%) of adolescent and 6.8% of young adult BZD initiators became LT-BZD users.

From 2017 to 2022, LT-BZD use declined in adolescents (5.4% to 4.1%; p-value for trend: 0.02) and young adults (7.4% to 6.0%; p-value: 0.06). Across the study period, LT-BZD use was more frequent in those with a SUD diagnosis (adolescents=10.6%, young adults=11.3%) than those without SUD (4.5%, 6.4%, respectively). After adjustment, adolescents (aOR=1.32, 95%CI: 1.01-1.73) and young adults (aOR=1.23, 95%CI: 1.13-1.35) with a SUD diagnosis had a modestly elevated odds of LT-BZD use compared to those without SUD.

Conclusions: LT-BZD use was relatively low, and declined from 2017 to 2022, in this privately insured sample of adolescents and young adults newly starting BZD treatment, suggesting more conservative prescribing practices over the study period. However, a higher proportion of adolescents and young adults with a SUD diagnosis became LT-BZD users, underscoring the need for careful prescribing and monitoring in this high-risk population.