Background: Good functional results have been reported for arthroscopic repair of rotator cuff tears, but the rate of tendon-to-bone healing is still unknown. Our hypothesis was that arthroscopic repair of full-thickness supraspinatus tears achieves a rate of complete tendon healing equivalent to those reported in the literature with open or mini-open techniques.

Methods: Sixty-five consecutive shoulders with a chronic full-thickness supraspinatus tear were repaired arthroscopically in sixty-five patients with use of a tension-band suture technique. Patients ranged in age from twenty-nine to seventy-nine years. The average duration of follow-up was twenty-nine months. Fifty-one patients (fifty-one shoulders) had a computed tomographic arthrogram, and fourteen had a magnetic resonance imaging scan, performed between six months and three years after surgery. All patients were assessed with regard to function and the strength of the shoulder elevation.

Results: The rotator cuff was completely healed and watertight in forty-six (71%) of the sixty-five patients and was partially healed in three. Although the supraspinatus tendon did not heal to the tuberosity in sixteen shoulders, the size of the persistent defect was smaller than the initial tear in fifteen. Sixty-two of the sixty-five patients were satisfied with the result. The Constant score improved from an average (and standard deviation) of 51.6 ± 10.6 points preoperatively to 83.8 ± 10.3 points at the time of the last follow-up evaluation (p < 0.001), and the average University of California at Los Angeles score improved from 11.5 ± 1.1 to 32.3 ± 1.3 (p < 0.001). The average strength of the shoulder elevation was significantly better (p = 0.001) when the tendon had healed (7.3 ± 2.9 kg) than when it had not (4.7 ± 1.9 kg). Factors that were negatively associated with tendon healing were increasing age and associated delamination of the subscapularis or infraspinatus tendon. Only ten (43%) of twenty-three patients over the age of sixty-five years had completely healed tendons (p < 0.001).

Conclusions: Arthroscopic repair of an isolated supraspinatus detachment commonly leads to complete tendon healing. The absence of healing of the repaired rotator cuff is associated with inferior strength. Patients over the age of sixty-five years (p = 0.001) and patients with associated delamination of the subscapularis and/or the infraspinatus (p = 0.02) have significantly lower rates of healing.

Level of Evidence: Therapeutic Level IV. See Instructions to Authors for a complete description of levels of evidence.