From May 1994 to May 1995, we performed a prospective controlled double-blind clinical study of 85 patients(85 shoulders). For inclusion in the study all patients met the following criteria: 1) a diagnosis of impingement syndrome by lidocaine injection test; 2) symptoms for at least three months; 3) no previous subacromial injection; 4) no evidence of rotator cuff tear. We excluded the patients from the study who showed a instability or other clear primary cause. The patients were randomized to receive either 5_cc 2% lidocaine with 4ee of 20mg/cc Depomedrol(Group 1) in 45 cases or 5_cc 2% lidocaine solely(Group 2) in 40 cases. At three month follow-up, all results were better in steroid group than control group. But at one year, improvement of pain and daily activity showed no difference between two groups. So we conclude that the use of subacromial steroid injection is effective for short-term therapy in the treatment of subacromial impingement syndrome, but the long-term effectiveness is doubtful.