Long-term effectiveness of controlled-release melatonin in 25 children, aged 2.6-9.6 years with autism without other coexistent pathologies was evaluated openly. Sleep patterns were studied using Children's Sleep Habits Questionnaire (CSHQ) and sleep diaries at baseline, after 1-3-6 months melatonin treatment and 1 month after discontinuation. Sleep diary and CSHQ showed a more problematic sleep in autistic children compared with controls. During treatment sleep patterns of all children improved. After discontinuation 16 children returned to pre-treatment score, readministration of melatonin was again effective. Treatment gains were maintained at 12 and 24-month follow-ups. No adverse side effects were reported. In conclusion, controlled-release melatonin may provide an effective and well-tolerated treatment for autistic children with chronic sleep disorders.