We treated 24 patients with narcolepsy for 4 weeks with gammahydroxybutyrate (GHB), 60 mg/kg/night, in a randomized double-blind placebo-controlled cross-over trial. Both clinical and polysomnographic criteria were used to assess the results. Compared to placebo, GHB reduced the daily number of hypnagogic hallucinations (from 0.87 to 0.28; p = 0.008), daytime sleep attacks (from 2.27 to 1.40; p = 0.001) and the severity of subjective daytime sleepiness (from 1.57 to 1.24 on a 0-4 scale; p = 0.028). The number of daily cataplexy attacks was reduced from 1.26 at baseline to 0.56 after 4 weeks of GHB intake. This reduction, however, was not statistically significantly different from the difference between baseline and placebo. GHB stabilized nocturnal rapid eye movement (REM) sleep, i.e. it reduced the percentage of wakefulness during REM sleep (p = 0.007) and the number of awakenings out of REM sleep (p = 0.016), and tended to increase slow wave sleep (p = 0.053). Adverse events were few and mild. We conclude that GHB is an effective and well-tolerated treatment for narcolepsy.