Background. Older adults are the main recipients of repeat prescriptions for benzodiazepine (BZD) hypnotics. BZDs can impair cognitive function and may not aid sleep when taken continuously for years. This study therefore aimed to determine if withdrawing from BZDs leads to changes in patients' cognitive function, quality of life, mood and sleep.

Method. One hundred and ninety-two long-term users of BZD hypnotics, aged [ges ]65 years, were identified in 25 general practices. One hundred and four who wished to withdraw were randomly allocated to one of two groups under double-blind, placebo controlled conditions: group A's BZD dose was tapered from week 1 of the trial; group B were given their usual dose for 12 weeks and then it was tapered. An additional group (C) of 35 patients who did not wish to withdraw from BZDs participated as ‘continuers’. All patients were assessed at 0, 12 and 24 weeks and 50% were re-assessed at 52 weeks.

Results. Sixty per cent of patients had taken BZDs continuously for >10 years; 27% for >20 years. Of all patients beginning the trial, 80% had successfully withdrawn 6 months later. There was little difference between groups A and B, but these groups differed from continuers (C) in that the performance of the withdrawers on several cognitive/psychomotor tasks showed relative improvements at 24 or 52 weeks. Withdrawers and continuers did not differ in sleep or BZD withdrawal symptoms.

Conclusions. These results have clear implications for clinical practice. Withdrawal from BZDs produces some subtle cognitive advantages for older people, yet little in the way of withdrawal symptoms or emergent sleep difficulties. These findings also suggest that, taken long-term, BZDs do not aid sleep.