Anxiety Drug Use May Boost Dementia Risk
Older patients who used benzodiazepine anti-anxiety drugs were at substantially higher risk of developing dementia than nonusers, a French study found.
Among 1,063 randomly selected individuals 65 and older in southwest France who agreed to participate in a long-term observational study, the risk of new-onset dementia during follow-up was 60% greater (adjusted odds ratio 1.60, 95% CO 1.08 to 2.38) for those who had used benzodiazepines relative to never-users.
"The results remained robust after control for potential confounders, in pooled analysis across the follow-up time, and in a nested case-control study," reported PhD student Sophie Billioti de Gage, of the University of Bordeaux Segalen in Bordeaux, France, and colleagues online in BMJ.
"Considering the extent to which benzodiazepines are now prescribed, physicians and regulatory agencies should consider the increasing evidence of the potential adverse effects of this drug class for the general population," they wrote.
The possibility that benzodiazepines impair cognition over the long term is not new, the researchers noted, given that short-term effects "are well recognized."
But earlier studies seeking to examine the question have yielded conflicting results: some have found an increased risk of cognitive impairment or dementia, whereas others detected no relationship or even a protective effect.
In the current study, Billioti de Gage and colleagues analyzed data from the PAQUID cohort study, in which a total of 3,777 people in southwestern France were followed for up to 20 years starting in the late 1980s, with clinic visits every 2 to 3 years.
The researchers excluded those failing to remain in the study for at least 5 years, those with prevalent dementia at the 5-year follow-up, and those with a history or current use of benzodiazepines at year 3 or with missing data on use of these drugs.
That left 1,063 participants for analysis, with a mean age at enrollment of 78.
After adjusting for age, sex, education, wine intake, use of anti-diabetic and cardiovascular medications, and trajectory of Mini-Mental State Examination scores during the study's first 3 years, Billioti de Gage and colleagues came up with the odds ratio of 1.60 for incident dementia after study year 5 associated with new benzodiazepine use.
Using the larger PAQUID cohort, the researchers also performed a nested case-control evaluation, comparing each of 467 individuals who developed dementia after being in the study for 8 years with up to four other participants who were dementia-free at the time when the case's dementia was diagnosed. (Under this procedure, a participant might serve as a control at one point and separately as a case if dementia developed later on.)
Those who had ever used benzodiazepines in this analysis were more likely to develop dementia after the 8-year mark, with an odds ratio of 1.55 compared with never-users (95% CI 1.24 to 1.95).
This value was virtually identical for those whose benzodiazepine use was recent (first reported at their last follow-up visit before the 8-year mark) and for those whose use was earlier in the past.
However, there were not very many recent benzodiazepine users -- 17 among cases and 52 among controls -- and the increase in dementia risk did not reach statistical significance.
An important limitation of the study was that the PAQUID protocol did not collect data on diagnoses of anxiety or sleep disorders, so Billioti de Gage and colleagues could not adjust for them in their analyses.
Also, the exclusion of participants with missing data -- who might have had higher rates of incident dementia than other participants -- could have skewed the results.
The researchers suggested that future studies examine "possible correlations between [benzodiazepine] dosage or cumulative length of exposure and dementia."
The study was funded by agencies of the French government and internal university sources.
Some co-authors reported relationships with Ipsen, Merck, Serono, Novartis, Allergan, MAP Pharmaceuticals, Sanofi, and BMJ.