Clearing the confounding confusion: Benzodiazepines and the risk of dementia?

OBJECTIVE

To examine the relationship between duration of benzodiazepine exposure and the risk of dementia.

DESIGN

A retrospective cohort study using administrative health data followed 3 age-based strata (ages 55, 65 and 75) for up to a maximum of 22 years to examine risk of dementia due to benzodiazepine use. Each stratum was analyzed as a whole, then restricted to persons with depression or anxiety, and finally using high dimension propensity scores (HDPS) matched cohort.

SETTING

We used administrative data on subjects receiving standard medical care in Manitoba, Canada.

MEASUREMENTS: Prescription data was used to quantify benzodiazepine using cumulative defined-daily-dose (DDD). Comorbidities and cases of dementia were determined using medical and hospital data.

RESULTS

Dementia risk in high-dose users of the youngest strata was elevated compared to non-users (adjusted HR (aHR) 1.33; 95 % CI 1.05-1.68)). Little to no difference was found in the middle (aHR 1.17; 1.02-1.33) and oldest strata (aHR 1.02; 0.93-1.11). Restriction to persons with depression or anxiety eliminated the association. No association was found using HDPS-matched comparison groups.

CONCLUSIONS

Only a modest increase in dementia risk was seen in the high-dose benzodiazepine users. This association appears to be driven by the confounding due to higher rates of diabetes, cardiovascular disease, depression, and anxiety among users. Using restriction or HDPS to better control for confounding effects eliminates the association. While benzodiazepines do not appear to be a significant risk factor for dementia, tolerance, dependency and adverse effects caution against their long-term use.