Background. Benzodiazepines (BZs) can impair explicit memory after a single dose and also when taken repeatedly for treatment of anxiety disorders. A previous study with agoraphobia/panic patients found that the BZ alprazolam impaired memory during an 8-week treatment and residual impairments were still manifest several weeks after drug withdrawal (Curran et al. 1994). The present study followed up the same group of patients 3·5 years after treatment to determine whether those memory impairments persisted.

Method. Thirty-one patients, 15 who had originally been treated with alprazolam and 16 with placebo, were assessed on a battery of psychometric tests and self-rating scales.

Results. Ex-alprazolam patients performed at the same levels as ex-placebo patients on the memory task and on other objective tests. Performance levels of both groups were similar to pre-treatment baselines, however there were differences in subjective ratings whereby ex-alprazolam patients rated themselves as less attentive and clear headed and more incompetent and clumsy than ex-placebo patients.

Conclusions. Explicit memory impairments found while patients were taking alprazolam and weeks after drug withdrawal did not persist 3·5 years later. We suggest that the memory impairments observed in our previous study weeks after withdrawal of alprazolam were not residual effects of alprazolam but rather were due to the drug's interference with practice effects on the tests and habituation of anxiety over repeated exposure to the test situation.