Impaired cognitive function in elderly men exposed to benzodiazepines or other anxiolytics

A representative sample of 68 year old men living in Malmo, Sweden, was investigated concerning use of anxiolytic-hypnotic drugs. The subjects were psychometrically tested using the synonyms, Koh's block design, paired associates, visual retention, and digit symbol tests. Subjects with continuous use of anxiolytics, the vast majority of which consisted of benzodiazepines, had an increased risk of impaired cognitive function (RR=3.2 and p=0.008). Multivariate adjustments for differences in education, alcohol habits, exposure to organic solvents and stroke between users and non-users of anxiolytics did not reduce the estimate of risk (RRadj=3.4; 95% confidence interval: 1.3-8.9). Adjusting for differences in the number of subjects with a history of depression reduced the risk estimate slightly (RRadj=2.7; 95% confidence interval: 1.0-7.7) but adjustments for differences in anti-hypertensive and anti-diabetic medication had no influence. Subjects with occasional use of anxiolytics and subjects with occasional or continuous use of hypnotics showed no risk increase. Neither selection bias nor information bias appeared to explain the findings. Accordingly, increased caution is warranted when prescribing benzodiazepines for regular anxiolytic use.