Two hundred thirty patients with generalized anxiety and Hamilton Rating Scale for Anxiety (HAM-A) scores greater than or equal to 18 were subdivided at random, according to a double-blind design, into one group treated with 5-10 mg of oral buspirone t.i.d. or one group treated with 10-20 mg of oral oxazepam t.i.d. for 6 weeks. No anxiolytic treatment was allowed 3 months prior to trial entry. Analysis of demographic variables revealed no significant imbalance between the two treatment groups. Twenty patients were excluded from efficacy analysis because of treatment withdrawal before the first efficacy evaluation on Day 7. Another 4 patients were excluded because they were taking concomitant psychotropic medication. The remaining 206 patients displayed a decrease in HAM-A scores (mean +/- SD) from 23.9 +/- 4.1 to 10.6 +/- 7.7 in the buspirone group and from 23.9 +/- 4.2 to 11.5 +/- 8.0 in the oxazepam group. The two treatment groups were also found to be virtually identical in an "intent to treat" analysis of all 230 patients as well as in other ratings (Hamilton Rating Scale for Depression, Raskin Depression Scale, Covi Anxiety Scale, Physicians Questionnaire, global ratings, and Hopkins Symptom Checklist [HSCL]-56). However, oxazepam was never superior to buspirone in any of the efficacy analyses. Of the 230 patients, 127 spontaneously reported adverse events, including drowsiness, dizziness, headache, nausea, and nervousness. Adverse events were relatively similar in the two groups. In conclusion, buspirone and oxazepam appear to be equally effective in the treatment of generalized anxiety encountered by general practitioners. This outcome, in addition to a previously documented absence of any dependency liability, makes buspirone a clinically important anxiolytic drug.