Patterns of tendon retraction in full-thickness rotator cuff tear: comparison of delaminated and nondelaminated tendons

Objective To analyze full-thickness rotator cuff tears, compare retraction patterns in delaminated and nondelaminated tendons, and correlate retraction distances with anteroposterior tear lengths. Materials and methods In 483 MR examinations reported as showing full-thickness cuff tear, two musculoskeletal radiologists independently characterized tendons as delaminated or nondelaminated. Tendon delamination was defined as either horizontal intra-substance splitting of bursal and articular layers by an intervening plane of fluid, or differential retraction of bursal and articular layers. In a subset of 144 shoulders with surgically proven full-thickness cuff tears (45 delaminated, 99 nondelaminated tendons), matched cohorts ( n = 45) were further analyzed to compare tendon retraction distance, anteroposterior tear length and retraction ratios (retraction distance/anteroposterior length). Results Delamination was present in 13% of 483 total tears, and 31% of 144 operated tears ( p = 0.001). In nondelamination and delamination cohorts, mean anteroposterior tear length measured 30.0 and 31.5 mm respectively ( p = 0.6). Although nondelaminated tendons showed mean retraction 31.5 mm, articular and bursal layers of delaminated tendons showed mean retractions 36.3 mm and 21 mm respectively ( p