Background: Previous randomized trials on cuff repair have included mainly degenerative tears, but studies on acute traumatic tears are lacking. We aimed to compare early surgical repair with nonoperative treatment for traumatic supraspinatus tears. Methods: We did a 2-center randomized controlled trial of patients with small rotator cuff tears mainly involving supraspinatus, comparing surgical repair (n = 32) and physiotherapy (n = 26). The primary outcome was a group difference in the Constant-Murley score at 12-month follow-up. Secondary outcomes were differences in the Western Ontario Rotator Cuff index, pain (Numerical Rating Scale 0-10), and Euro quality-of-life-visual analog scale. We used magnetic resonance imaging to assess retear rate, tear progression, fatty infiltration, and atrophy. Results: The mean age was 59.7 years (range, 44-77 years), median sagittal tear size was 9.7 mm (range, 421 mm), and baseline characteristics were well balanced between the 2 groups. The repair group had a median Constant-Murley of 83 (25 quartile range [QR)) and the physiotherapy group 78 (QR, 22) at 12 months, with the between-group difference in medians of 4.5 (-5 to 9, 95% confidence interval; P = .68). The corresponding values for the Western Ontario Rotator Cuff index were 91% (QR, 24) vs. 86% (QR, 24), with the between-group difference of 5.0 (-4 to 9, 95% confidence interval; P- .62). There was no difference in Numerical Rating Scale or in Euro quality-of-life-visual analog scale. Retear was found in 63% of repaired patients and tear progression amp;gt;5 mm in 29.2% of unrepaired patients. Conclusions: We found no significant differences in clinical outcomes between cuff repair and nonoperative treatment at 12-month follow-up. Approximately one third of unrepaired patients had a tear enlargement of more than 5 mm. (C) 2019 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.