To define the role of two-portal endoscopic carpal-tunnel release as a method for the treatment of compression of the median nerve at the wrist, a prospective, randomized, multicenter study was performed on 169 hands in 145 patients. Either open or endoscopic carpal-tunnel release was performed in all of the patients who had clinical signs and symptoms consistent with carpal tunnel syndrome, had not responded to or had refused non-operative management, and had had electrodiagnostic studies consistent with carpal tunnel syndrome. Follow-up evaluations were performed at twenty-one, forty-two, and eighty-four days. At the end of the follow-up period, both the open and endoscopic methods had resulted in high levels of achievement of the primary outcomes (relief of pain and paresthesias). The numbness and paresthesias were relieved in eighty (98 per cent) of eighty-two hands in the open-release group compared with seventy-seven (99 per cent) of seventy-eight hands in the endoscopic-release group. This parameter was not recorded for three hands in the open-release group or six hands in the endoscopic-release group. The satisfaction of the patients with the procedure, graded on a scale of 0 to 100 per cent, averaged 84 per cent in the open-release group compared with 89 per cent in the group that had had endoscopic release. We found no significant differences between the two groups with regard to the secondary quantitative-outcome measurements, including two-point discrimination, postoperative interstitial-pressure data for the carpal canal, Semmes-Weinstein monofilament testing, and motor strength. The open technique resulted in more tenderness of the scar than did the endoscopic method. Thirty-two (39 per cent) of eighty-two hands in the open-release group and fifty (64 per cent) of seventy-eight hands in the endoscopic-release group were not tender at eighty-four days. This parameter was not recorded for three hands in the open-release group and six hands in the endoscopic-release group. The open method also resulted in a longer interval until the patient could return to work (median, twenty-eight days, compared with fourteen days for the open-release and endoscopic-release groups). Four complications occurred in the endoscopic carpal-tunnel release group: one partial transection of the superficial palmar arch, one digital-nerve contusion, one ulnar-nerve neuropraxia, and one wound hematoma.(ABSTRACT TRUNCATED AT 400 WORDS)