Objectives: Subacromial steroid injections are frequently used in the treatment of impingement syndrome. Several studies have pointed out that the localization of the injection makes a significant difference in symptomatic improvement. This study was designed to test the importance of location of the corticosteroid injection in the treatment of impingement syndrome. Patients and methods: The study included 30 patients (18 females, 12 males; mean age 52 years; range 23 to 67 years) who had shoulder pain for at least three months. Involvement was in the right shoulder in 19 cases, and in the left shoulder in 11 cases. The mean duration of symptoms was 11 months (range 3 to 48 months). On examination, all the patients exhibited a positive impingement sign, no instability, and an intact rotator cuff. Magnetic resonance imaging was performed in all the shoulders. The patients were randomized to three groups, equal in number, to receive a single injection of 40 mg methylprednisolone acetate into the subacromial space, deltoid muscle, and gluteal muscle, respectively. Functional and clinical assessments were made with the Constant score and a visual analog scale (VAS), respectively, before and three months after treatment. Results: Subacromial and deltoid injections resulted in significant improvement in both Constant and VAS scores (p<0.003), whereas improvement obtained after gluteal injections was not significant (p>0.05). However, comparison between the three groups yielded no significant difference with respect to the location of injections. Conclusion: Considering possible detrimental effects of repeated subacromial injections on rotator cuff tendons, intradeltoid muscle injections may have a chance in clinical application.