A clinical and electrophysiological study evaluated the usefulness of local steroid therapy for carpal tunnel syndrome (CTS). To evaluate the efficacy of local steroid therapy 32 patients (53 nerves) were randomly assigned to one of two groups: one (27 nerves) received 15 mg methylprednisolone acetate injected locally and the other (26 nerves) received the same amount of saline solution. The injections were repeated after a week. Clinical and electrophysiological findings were evaluated, double blind, at regular intervals. A clear-cut efficacy of steroid treatment was found. Only 8% of nerves were not benefitted while a marked early improvement was observed in most of the nerves. In order to appraise the long-term effect of local steroid treatment on CTS, 53 patients (91 nerves) were studied and followed up by means of clinical and electrophysiological examinations performed every 2 months for 2 years. The benefit of steroid treatment was transient. About 50% of the nerves became worse within 6 months and 90% within 18 months. Only a small percentage (8%) of the nerves remained improved at the 2-years follow-up. The clinical features were not useful in foretelling the duration of the improvement, which appeared to be related to the antidromic SAP latency.