We have performed a double-blind placebo-controlled trial of moderate doses of extracorporeal shock-wave therapy (ESWT) for non-calcific tendonitis of the rotator cuff. Adults (74) with chronic tendonitis of the rotator cuff were randomised to receive either active (1500 pulses ESWT at 0.12 mJ/mm2) or sham treatment, monthly for three months. All were assessed before each treatment, and at one and three months after the completion of treatment. The outcome was measured with regard to pain in the shoulder, including a visual analogue score for night pain, and a disability index. There were no significant differences between the two groups before treatment. The mean duration of symptoms in both groups was 23.3 months. Both showed significant and sustained improvements from two months onwards. There was no significant difference between them with respect to change in the Shoulder Pain and Disability Index (SPADI) scores or night pain over the six-month period. A mean (+/-SD; range) change in SPADI of 16.1 +/- 27.2 (0 to 82) in the treatment group and 24.3 +/- 24.8 (-11 to 83) in the sham group was noted at three months. At six months the mean changes were 28.4 +/- 25.9 (-24 to 69) and 30.4 +/- 31.2 (-12 to 88), respectively. Similar results were noted for night pain. We conclude that there is a significant and sustained placebo effect after moderate doses of ESWT in patients with non-calcific tendonitis of the rotator cuff, but there is no evidence of added benefit when compared with sham treatment.