Sleep quality and anxiety levels were examined using questionnaires and pol
ysomnographic recordings in 22 chronic insomnia patients who regularly used
benzodiazepines to treat their sleeping problems. After abruptly discontin
uing their benzodiazepine medication, patients were randomly allocated to r
eceive either a placebo or zolpidem 10 mg for 1 week, after which they ente
red an open extension phase, receiving zolpidem 10 mg for 3 weeks. Subjecti
vely, sleep quality was considered mediocre during the use of a benzodiazep
ine hypnotic. One week after the discontinuation, an increase in sleep late
ncy was observed in the placebo group, whereas zolpidem induced a significa
nt decrease in sleep latency. Deterioration of other sleep variables (proba
bly rebound) was not suppressed by zolpidem. An explanation for this could
be the selective pharmacological profile of zolpidem. Polysomnographic diff
erences between placebo and benzodiazepine and between placebo and zolpidem
were not reflected by the subjective data on sleep and anxiety. Changes of
sleep structure caused by hypnotics seem not always to be felt as such by
patients. After 3 - 4 weeks of zolpidem treatment, the percentage of non-ra
pid eye movement-4 sleep increased significantly, corresponding with a sign
ificant subjective improvement of sleep quality. This indicates that zolpid
em may restore physiological sleep.