Incidence of and Characteristics Associated With Long-term Benzodiazepine Use in Finland.

Importance: The proportion of patients who develop long-term benzodiazepine use remains controversial, as do the length of time before long-term use develops and the factors associated with long-term use. Objective: To investigate the incidence of long-term benzodiazepine and related drug (BZDR) use and factors associated with the development of long-term use implementing a follow-up design with new BZDR users. Design, Setting, and Participants: This population-based cohort study used a nationwide cohort of 129 732 new BZDR users in Finland. New users of BZDRs aged 18 years or older were identified from the prescription register maintained by the Social Insurance Institution of Finland as individuals who initiated BZDR use during 2006 and had not used BZDRs from 2004 to 2005. The follow-up continued until death, long-term hospitalization, a gap of 2 years in BZDR use, or December 31, 2015. The population was analyzed according to age at treatment initiation, categorized into younger (<65 years) and older (≥65 years) subcohorts. Analyses were conducted from May 2019 to February 2020. Exposures: Use of BZDRs, modeled from register-based data using the PRE2DUP (from prescriptions to drug use periods) method. Main Outcomes and Measures: Long-term BZDR use, defined as continuous use of 180 days or longer, and factors associated with long-term vs short-term use, compared using Cox proportional hazards models. Results: Among the 129 732 incident BZDR users, the mean (SD) age was 52.6 (17.7) years, and 78 017 (60.1%) individuals were women. During the follow-up period, 51 099 BZDR users (39.4%) became long-term users. Long-term treatment was more common in the older subcohort (19 103 individuals [54.5%]) than the younger subcohort (31 996 individuals [33.8%]). At 6 months, 28 586 individuals (22.0%) had become long-term users: 11 805 (33.7%) in the older subcohort and 16 781 (17.7%) in the younger subcohort. The largest proportions of initiators who became long-term users were those persons who initiated treatment with nitrazepam (76.4%; 95% CI, 73.6%-79.1%), temazepam (63.9%; 95% CI, 62.9%-65.0%), lorazepam (62.4%; 95% CI, 59.7%-65.1%), or clonazepam (57.5%; 95% CI, 55.9%-59.2%). Factors associated with the development of long-term use included male sex, older age, receipt of social benefits, psychiatric comorbidities, and substance abuse. Conclusions and Relevance: The findings of this population-based cohort study conducted in Finland suggest that the incidence of subsequent long-term BZDR use in individuals who initiate use of BZDRs is high, especially among older persons, and that the specific BZDR used initially is associated with the development of long-term BZDR use and should be carefully considered when prescribing BZDRs. The observed factors that appear to be associated with development of long-term BZDR use also should be considered in clinical decision-making when starting and monitoring BZDR treatment.