In a multicenter, double-blind trial, 310 patients who had received a diagnosis of generalized anxiety disorder were treated for 6 weeks with either abecarnil, diazepam, or placebo at mean daily doses of 12 mg of abecarnil or 22 mg of diazepam administered three times daily. Patients who were improved at 6 weeks could volunteer to continue double-blind treatment for a total of 24 weeks. The maintenance treatment phase allowed the comparison of taper results for the three treatments at several study periods (0-6 weeks, 7-12 weeks, and more than 12 weeks). Slightly more diazepam (77%) and placebo (75%) patients completed the 6-week study than abecarnil patients (66%). At intake and baseline, after a 1-week placebo washout, the patient was required to have a Hamilton Rating Scale for Anxiety score of ≥20. Major adverse events for both abecarnil and diazepam were drowsiness, dizziness, fatigue, and coordination difficulties. Clinical improvement data showed that both abecarnil and diazepam produced statistically significantly more symptom relief than did placebo after 1 week of treatment. At 6 weeks treatment (using last observation carried forward analysis), however, only diazepam still differed significantly (p < 0.01) from placebo. High placebo response (56% moderate to marked global improvement) at 6 weeks, as well as a slightly lower nonsignificant improvement rate observed with abecarnil, a partial γ-aminobutyric acid (GABA) agonist, when compared with diazepam, a full GABA agonist, most likely contributed to our findings. Finally, taper results showed that only diazepam and not abecarnil caused the presence of temporary discontinuation symptoms, but only in patients who had been treated for at least 12 weeks.