Carpal tunnel syndrome is the clinical condition resulting from the entrapment of the median nerve in the wrist. It has been accepted as the most frequent entrapment neuropathy. The most common symptoms are tingling, numbness, and pain in the hand that may radiate to the forearm or shoulder. There may be weakness and atrophy of the thenar muscles associated with sensory loss in the affected fingers. There is no universally accepted therapy for carpal tunnel syndrome. Surgical treatment is widely preferred to non-surgical or conservative therapies for overtly symptomatic patients, while mild cases are usually not treated. The objective of this review is to compare the efficacy of surgical treatment of carpal tunnel syndrome with non-surgical treatment in improving clinical outcome. We searched the Cochrane Neuromuscular Disease Group register for randomised or quasi-randomised trials as well as MEDLINE, EMBASE and LILACS (to July 2001). We checked the bibliographies in relevant papers and contacted the authors to obtain information about other published or unpublished studies. All randomised and quasi-randomised controlled trials comparing any surgical and any non-surgical therapies. Two reviewers independently assessed eligibility of the trials. Only one randomised controlled trial was found. It included 22 female patients, 11 allocated to surgical section of the anterior carpal ligament and 11 to splinting for one month. The trial was not blinded and it is not clear if the allocation was properly concealed. Data reported allowed an intention-to-treat analysis on two secondary outcomes. The results favour surgery for both of them. There was a significant clinical improvement at one year follow-up in 10 out of 11 patients allocated to surgery and two out of 11 allocated to splinting (relative risk 5.00, 95% confidence interval 1.41, 17.76). Eight out of 11 patients allocated to splinting required surgery during follow-up, compared with apparently no re operation in the surgical group (relative risk 0.06, 95% confidence interval 0.00, 0.91). Surgical treatment of carpal tunnel syndrome seems to be better than splinting. There is a need for randomised controlled trials comparing surgical and non-surgical therapies for carpal tunnel syndrome, particularly in patients with mild symptoms in whom there is greater uncertainty concerning the balance of risks versus benefit of surgical therapy.