A 5-Year Prospective Assessment of the Risk Associated with Individual Benzodiazepines and Doses in New Elderly Users

Authors: 
Robyn Tamblyn, Michal Abrahamowicz, Roxane du Berger, Peter McLeod, Gillian Bartlett
Research Summary: 

Objectives
To determine the risk of injury associated with the new use of individual benzodiazepines and dosage regimens in the elderly.

Design
Prospective database cohort study with 5 years of follow-up.

Setting
Quebec, Canada.

Participants
Two hundred fifty-three thousand two hundred forty-four persons aged 65 and older who were nonusers of benzodiazepines in the year before follow-up.

Measurements
Population-based hospitalization and prescription and medical services claims databases were used to compare the risk of injury during periods of benzodiazepine use with those of nonuse. Periods of use were measured for 10 insured benzodiazepines by drug and dose as time-dependent covariates. Injury was defined as the first occurrence of a nonvertebral fracture, soft-tissue injury, or accident-related hospital admission. Patient age, sex, previous injury history, concomitant medication use, and comorbidity were measured as fixed and time-dependent confounders. Cox proportional hazards models were used to estimate the risk of injury with benzodiazepine use and to determine the extent to which patient characteristics, differences in dosage, or in the effect of increasing dosage for individual drugs explained differences between drugs.

Results
More than one-quarter (27.6%) of 253,244 elderly were dispensed at least one prescription for a benzodiazepine, and 17.7% of elderly were treated for at least one injury during follow-up, of which fractures were the most common. Patient characteristics, systematic differences in the risk of injury in elderly prescribed different benzodiazepines, and differences in dosage prescribed for individual drugs confounded the risk of injury with benzodiazepine use. The risk of injury with increasing dosage varied by drug from a hazard ratio of 0.92 (95% confidence interval (CI)=0.60, 1.42) for alprazolam to 2.20 (95% CI=1.39, 3.47) for flurazepam per 1 standardized adult dose increase.

Conclusion
The risk of injury varied by benzodiazepine, independent of half-life, as did the risk associated with increasing dosage for individual products. Higher doses of oxazepam, flurazepam, and chlordiazepoxide are associated with the greatest risk of injury in the elderly.