Results from a new retrospective study published in the British Medical Journal suggest that the risk of all cause mortality associated with benzodiazepine initiation may be smaller than previously stated.

Researchers from Harvard Medical School analyzed the U.S. healthcare database, Optum Clinformatics Datamart, selecting patients who started taking a benzodiazepine between July 2004 to December 2013 and who had not previously filled a prescription for a benzodiazepine in the past six months (benzodiazepine initiators). 

Benzodiazepine initiators were matched using a 1:1 in high dimensional propensity score methodology to randomly selected benzodiazepine non-initiators. The final sample used for analysis included 1,252,988 patients who initiated a benzodiazepine and the same number of individuals who had a medical visit but did not initiate benzodiazepine treatment.

The researchers found that approximately 75% of prescriptions were for short-acting benzodiazepines. Alprazolam was the most often prescribed short-acting benzodiazepine (47.2%), while diazepam was the most often prescribed long-acting benzodiazepine (87.7%). The mean age of initiators was 46.0 and non-initiators was 45.2.

Without propensity score matching the number of deaths among benzodiazepine initiators was significantly greater than non-initiators, however once 1:1 propensity score matching was applied the hazard ratio was consistent with no increased risk of death associated with benzodiazepine initiation compared with non-initiation (hazard ratio (HR) 0.89, CI=0.85–0.93).  

Related Articles

A total of 5,061 and 4,691 deaths occurred in the high dimensional propensity score matched benzodiazepine initiators compared to non-initiators, which equated to 9.3 vs. 9.4 events per 1,000 person years, respectively (HR 1.00, 95% CI, 0.96–1.04).

The researchers identified a small but statistically significant increased risk of mortality among initiators vs. non-initiators with 12 and 48-month follow-up (HR 1.04, 95% CI, 1.01–1.08; HR: 1.05, 95% CI, 1.02–1.07, respectively), in younger patients (<65 years old; HR 1.09, 95% CI, 1.02–1.15), and those taking short-acting benzodiazepines (HR 1.06, 95% CI 1.02–1.10). 

The authors write that this large population study demonstrated, “either no increase or a small increase in the risk of all cause mortality associated with benzodiazepine initiation.” They go on to assert that if a mortality risk does exist it is lower than what has been stated in previous large scale studies. 

For more information visit