Long-term Use of Benzodiazepines for Insomnia in Primary Healthcare Facilities in Beijing, China

Background: Insomnia affects 25% of the primary care outpatients in Beijing and regular use of benzodiazepines (BZD) for insomnia is common, yet little is known about long-term use, which has been associated with dependence and adverse effects. Using BZD for more than 4 weeks is not recommended by clinical guidelines.

Objectives: To assess long-term use of BZD for insomnia among primary care outpatients in Beijing.

Methods: We included adults aged ≥18 years with a diagnosis of insomnia and who were prescribed at least one oral BZD from all 63 primary care facilities in Dongcheng district in Beijing between July 2019 and June 2022. The following BZDs were included: estazolam, lorazepam, clonazepam, diazepam, alprazolam, nitrazepama and midazolam. We excluded adults with any other mental health disorder and prevalent users prescribed BZD between Jan and June 2019. Long-term use was defined as a continuous BZD course (total days between the first and last prescription plus the days’ supply of the last prescription) >30 days. Discontinuation of a course was defined as not having a benzodiazepine prescription ≥120 days (4-month extension after observation was set to prevent misclassification). Outcomes were proportion of patients having long-term BZD course, or having more than one BZD courses, and number of days between first and second courses. We applied generalized linear models to examine the adjusted associations between long-term BZD use and patient age (at their first dose during observation), controlling for patient gender and intra-cluster correlation among each clinician.

Results: A total of 89,459 patients were diagnosed with insomnia between July 2019 and June, 2022. Of 34,905 prevalent BZD users for insomnia between Jul 2019 and Jun, 2022, 59.1% were female and 62.1% were aged ≥65 years. Overall, 49.3% of incident BZD users received at least one long-term course, 35.7% a >90-day course and 20.8% received a >180-day course. One-third of the incident BZD users reinitiated BZD treatment after a median of 211 (IQR 154-329) days. Higher long-term rates were found in older adults ≥85 years (56.0%, P< 0.001, OR 2.25 [1.95-2.59]), 75-84 years (53.2%, P< 0.001, OR 2.02 [1.75-2.33]), and 65-74 years (49.8%, P< 0.001, OR 1.73 [1.52-1.96]) when compared with adults 18-44 years (35.2%). Males were more likely to have long-term courses than females (50.6% vs 48.6%, P< 0.001, OR 1.09 [1.04-1.14]).

Conclusions: Despite the concerns of potentially harmful effects and limited evidence of clinical benefits, half of the primary care outpatients that were newly prescribed BZD for insomnia received long-term courses, especially among oldest adults.