We performed a prospective arthroscopic study to explore the variability of the posterior labrum glenoid insertion. We aimed to classify the insertions and to explore whether these insertions can be identified by pre-operative arthro-CT scan.

From January to December 2011, 86 patients were prospectively included in the current study. During arthroscopy, anterior labrum was evaluated and posterior labrum was assessed in 3 different locations: superior, medial, and inferior. For each segment, the labrum was considered normally inserted (directly to the glenoid cartilage), medialized (inserted at the posterior part of the glenoid bone, without direct contact with the cartilage), torn (macroscopic degenerative changes, tears, fragments) or absent (agenesis). Imaging was analyzed segment by segment by an experienced osteoarticular radiologist, using the same classification.

Four types of posterior labrum insertions were identified. Type 1, 60% of the cases, corresponded to a posterior labrum totally inserted in the glenoid, with direct contact with the cartilage. Type 2, 20% of the cases, represented medialized insertion of the superior segment. Type 3, 15% of the cases, represented an associated medialization of the superior and medial segment of the posterior labrum. Type 4 is a medialized insertion of the all-posterior labrum. Fifty-six shoulders were used for arthro-CT and arthroscopy correlation: for the superior segment (n=22/56), the sensitivity of arthro-CT to identify an abnormal insertion when the labrum is medialized was 68.18%, specificity 70.59%, positive predictive value (PPV) 60%, and negative predictive value (NPV) 77.42%. For the medial segment (n=16/56), the sensitivity of arthro-CT to identify an abnormal insertion when the labrum is medialized was 81.25%, specificity 57.50%, PPV 43.33% and NPV 88.46%. For the inferior segment (n=5/56), the sensitivity was 100%, specificity 47.60%, PPV 15.63% and NPV 100%.

The current study points out the high variability of shoulder posterior labrum glenoid insertion, and thus the risk of misdiagnosis with posterior labral tears, especially in posterior instability and also the risk of considering as labral lesions some non-pathological aspects.

Level III. Anatomic prospective study.