To evaluate the long-term efficacy of non-surgical treatment methods for mild and moderate carpal tunnel syndrome, 120 patients with clinical symptoms and electrophysiologic evidence were included in a prospective, randomized and blinded trial: 60 patients were instructed to wear splints every night, 30 received injections of betamethasone 4 cm proximal to the carpal tunnel, and 30 received injections distal to the carpal tunnel. After approximately 1 year (mean, 11 months; range, 9-14), 108 patients were available for final evaluation. We assessed clinical symptom severity and performed detailed electrophysiologic examinations before and after treatment. Splinting provided symptomatic relief and improved sensory and motor nerve conduction velocities at the long-term follow-up when the splints were worn almost every night. Proximal and distal injections of steroids were ineffective on the basis of both clinical symptoms and electrophysiologic findings.