A 12-month, multicenter, open-label study was conducted to evaluate the long-term safety and efficacy of the new nonbenzodiazepine anxiolytic buspirone. The study population consisted of 700 patients with DSM-III generalized anxiety disorder who ranged in age from 16 to 84 years. Ninety-two percent had symptoms of anxiety that had persisted for 3 months or longer. According to the Hamilton Rating Scale for Anxiety, the Physician Global Opinion of Improvement, and the Patient Opinion of Improvement, the use of buspirone resulted in a significant reduction in anxiety. The degree of this improvement correlated with the duration of drug use. After the third month of therapy, there was an extremely low incidence of patient discontinuation due to side effects or insufficient improvement. Adverse effects were typical of those observed in other double-blind, placebo-controlled studies that compared buspirone with benzodiazepines. Overall, buspirone was well tolerated for maintenance therapy.