We have carried out a prospective double-blind randomised controlled trial to compare the efficacy of a single subacromial injection of the non-steroidal anti-inflammatory drug, tenoxicam, with a single injection of methylprednisolone in patients with subacromial impingement. A total of 58 patients were randomly allocated into two groups. Group A received 40 mg of methylprednisolone and group B 20 mg of tenoxicam as a subacromial injection along with lignocaine. The Constant-Murley shoulder score was used as the primary outcome measure and the Disability of Arm, Shoulder and Hand (DASH) and the Oxford Shoulder Score (OSS) as secondary measures. Six weeks after injection the improvement in the Constant-Murley score was significantly greater in the methylprednisolone group (p = 0.003) than in the tenoxicam group. The improvement in the DASH score was greater in the steroid group and the difference was statistically significant and consistent two (p < 0.01), four (p < 0.01) and six weeks (p < 0.020) after the injection. The improvement in the OSS was consistently greater in the steroid group than in the tenoxicam group. Although the difference was statistically significant at two (p < 0.001) and four (p = 0.003) weeks after the injection, it was not at six weeks (p = 0.055). Subacromial injection of tenoxicam does not offer an equivalent outcome to subacromial injection of corticosteroid at six weeks. Corticosteroid is significantly better than tenoxicam for improving shoulder function in tendonitis of the rotator cuff after six weeks.