Failure with continuity in rotator cuff repair "healing".

BACKGROUND Ten to seventy percent of rotator cuff repairs form a recurrent defect after surgery. The relationship between retraction of the repaired tendon and formation of a recurrent defect is not well defined. PURPOSE/ HYPOTHESES: To measure the prevalence, timing, and magnitude of tendon retraction after rotator cuff repair and correlate these outcomes with formation of a full-thickness recurrent tendon defect on magnetic resonance imaging, as well as clinical outcomes. We hypothesized that (1) tendon retraction is a common phenomenon, although not always associated with a recurrent defect; (2) formation of a recurrent tendon defect correlates with the timing of tendon retraction; and (3) clinical outcome correlates with the magnitude of tendon retraction at 52 weeks and the formation of a recurrent tendon defect. STUDY DESIGN Case series; Level of evidence, 4. METHODS Fourteen patients underwent arthroscopic rotator cuff repair. Tantalum markers placed within the repaired tendons were used to assess tendon retraction by computed tomography scan at 6, 12, 26, and 52 weeks after operation. Magnetic resonance imaging was performed to assess for recurrent tendon defects. Shoulder function was evaluated using the Penn score, visual analog scale (VAS) score for pain, and isometric scapular-plane abduction strength. RESULTS All rotator cuff repairs retracted away from their position of initial fixation during the first year after surgery (mean [standard deviation], 16.1 [5.3] mm; range, 5.7-23.2 mm), yet only 30% of patients formed a recurrent defect. Patients who formed a recurrent defect tended to have more tendon retraction during the first 6 weeks after surgery (9.7 [6.0] mm) than those who did not form a defect (4.1 [2.2] mm) (P = .08), but the total magnitude of tendon retraction was not significantly different between patient groups at 52 weeks. There was no significant correlation between the magnitude of tendon retraction and the Penn score (r = 0.01, P = .97) or normalized scapular abduction strength (r = -0.21, P = .58). However, patients who formed a recurrent defect tended to have lower Penn scores at 52 weeks (P = .1). CONCLUSION Early tendon retraction, but not the total magnitude, correlates with formation of a recurrent tendon defect and worse clinical outcomes. "Failure with continuity" (tendon retraction without a recurrent defect) appears to be a common phenomenon after rotator cuff repair. These data suggest that repairs should be protected in the early postoperative period and repair strategies should endeavor to mechanically and biologically augment the repair during this critical early period.