Zolpidem (10 mg/day) and zopiclone (7.5 mg/day), administered at night, were compared in a 14-day, double-blind, equivalence trial on 479 chronic primary insomniacs (zolpidem, 231; zopiclone, 248) throughout Japan, with a 1-week follow-up to assess rebound. The primary endpoint was the investigators' rating of global improvement of sleep disorders. A total of 32 patients in the zolpidem group (13.9%) and 45 patients in the zopiclone group (18.1%) withdrew from the study before the end of the treatment. In the zolpidem group, 67.9% (142/209) of patients were rated at least 'moderately improved' versus 61.6% (135/219) with zopiclone, zolpidem being at least as effective as zopiclone (90% confidence interval: -1.7, 14.3). With zolpidem, sleep onset latency improved in significantly more patients (85.8% versus 77.5%) and significantly fewer patients showed aggravated sleep onset latency relative to baseline at follow-up (4.5% versus 15.4%). Significantly fewer patients receiving zolpidem experienced drug-related adverse events (31.3% versus 45.3%), with bitter taste representing 5.8% (six of 104) of such complaints with zolpidem compared with 39.9% (69/173) with zopiclone. In conclusion, zolpidem was at least as effective as zopiclone, showed significantly less rebound on discontinuation and was better tolerated.