To assess the effects of gabapentin on sensory and motor symptoms in patients with restless legs syndrome (RLS). Patients with RLS (22 idiopathic, 2 secondary to iron deficiency) were randomized and treated for 6 weeks with either gabapentin or placebo. After a 1-week washout they crossed over to the alternative treatment for 6 weeks. Patients were rated at baseline and at scheduled intervals by the RLS Rating Scale, Clinical Global Impression, pain analogue scale, and Pittsburgh Sleep Quality Index. At the end of each treatment period, all-night polysomnography was performed. Compared to placebo, gabapentin was associated with reduced symptoms on all rating scales. In addition, sleep studies showed a significantly reduced periodic leg movements during sleep (PLMS) index and improved sleep architecture (increased total sleep time, sleep efficiency, and slow wave sleep, and decreased stage 1 sleep). Patients whose symptoms included pain benefited most from gabapentin. The mean effective dosage at the end of the 6-week treatment period was 1,855 mg, although therapeutic effects were already observed at the end of week 4 (1,391 mg). Gabapentin improves sensory and motor symptoms in RLS and also improves sleep architecture and PLMS.