BACKGROUND: The purposes of the study were to examine rotator cuff tendon degeneration with respect to harvesting location, to determine a rationale for debridement of the torn end, and thus, to determine adequate debridement extent. METHODS: Twenty-four patients with a full-thickness rotator cuff tear were included in the study. Tendon specimens were harvested during arthroscopic rotator cuff repair from three locations; from torn ends after minimal regularization of fraying (native end group, NE group), from torn ends after complete freshening of the frayed end (freshened end group, FE group), and from the macroscopically intact portion just distal to the musculotendinous junction (musculotendinous junction group, MTJ group). Control samples were harvested from patients admitted for surgery for proximal humerus fracture. Harvested samples were evaluated using a semi-quantitative grading scale. RESULTS: Mean total degeneration scores in the NE group (13.3 +/- 3.21), the FE group (12.5 +/- 2.30), and in the MTJ group (10.8 +/- 3.10) were significantly higher than those in the normal control group (5.0 +/- 2.87; all P>0.001). Mean total degeneration score in the NE group was significantly higher than that in the MTJ group (p=0.012), but was not from that of the FE group. Mean total degeneration score in the FE group was not significantly different from that of the MTJ group. CONCLUSIONS: Tendon degeneration exists throughout the entire tendon to the macroscopically intact portion of full-thickness rotator cuff tear. Therefore, aggressive debridement to grossly normal appearing, bleeding tendon is unnecessary for enhancing healing after repair.