In order to define the role of steroid injection and splinting as a method of treatment of carpal-tunnel syndrome, a prospective study was performed on 50 hands in 41 consecutive patients. All hands were treated with a single injection and 3 wk of splinting. Follow-up ranged from a minimum of 6 mth to a maximum of 26 mth, with a mean of 18 mth. All hands had characteristic symptoms of median-nerve compression at the wrist and increased distal median motor latencies. Eleven (22%) of 50 hands were completely free of symptoms at the end of the follow-up period. Hands that initially had mild symptoms and findings of less than 1 year's duration, normal sensibility, normal thenar strength and mass, and one to two-millisecond prolongations of either distal median motor or sensory latencies had the most satisfactory responses to injections and splinting. Hands with severe symptoms of more than 1 yr's duration and findings of atrophy, weakness, and distal motor latencies of more than 6 milliseconds or absent sensory responses had the poorest response to injections and experienced a high rate of relapse.