Text

Objectives and questions: Rotator cuff re-tears (RCRT) present diagnostic and therapeutic challenges with multifactorial causes. This study categorizes re-rupture etiology in patients undergoing arthroscopic revision and evaluates failure characteristics using preoperative MRI and standardized arthroscopic imaging.

Material and methods: Institutional registry data from 56 patients with posterosuperior RCRT undergoing arthroscopic revision were analyzed. Re-tears were classified as traumatic or degenerative based on history. Preoperative MRI was assessed using the Sugaya classification by two independent reviewers and compared with intraoperative findings. Suture-anchor construct failure and tendon healing were evaluated. Four failure types were defined: Type I: Footprint failure with suture-anchor construct failure, Type II: Footprint failure with intact suture material, Type III: Medial failure with intact footprint and suture material, Type IV: Medial failure with intact footprint and non-intact suture material. Interrater reliability (IRR) was analyzed using Cohen’s Kappa.

Results: The cohort had a mean age of 57±10 years, 70% were male. 66% of re-ruptures were atraumatic, 30% trauma-related, 4% caused by postoperative infection. The mean interval between the primary surgery and revision was 26±35 months, 38% of cases were classified as Type II failure, 29% were Type I, 20% Type III, 7% Type IV, 6% of cases could not be classified due to insufficient intraoperative documentation. A complete re-reconstruction was achieved in 92% of cases, with double-row (DR) techniques used most frequently (68%). 35% of DR cases used 2 medial and 2 lateral anchors. 33% used 2 medial and 1 lateral anchor. Patch augmentation (xenograft or autograft) was performed in 11%. Side-to-side sutures were sufficient in 13%, while single-row reconstruction was used in 11%. Partial re-reconstruction with superior capsular reconstruction (SCR) was performed in 4%. Latissimus dorsi tendon transfer was performed in 2%, and reconstruction was not possible in another 2%. MRI-based evaluation of RCRT patterns showed: 4% Sugaya Type I, 11% Sugaya Type II, 22% Sugaya Type III, 63% Sugaya Type IV (IRR: kappa = 0.181). The MRI findings often underestimated intraoperative severity. The newly defined failure types demonstrated an IRR of kappa = 0.266.

Discussion and conclusions: Preoperative MRI-based Sugaya classification often underestimates the severity of RCRT. Footprint failures (Types I–II) were the most common failure mechanisms. Re-reconstruction was achieved in nearly all cases. The differentiated classification of failure mechanisms (Types I–IV) showed improved IRR compared to the Sugaya classification and provides a systematic framework for understanding RCRT etiology and characteristics, enabling more targeted surgical strategies.