Purpose: The risk of recurrence would be greater after arthroscopic treatment of shoulder instability compared with open repair. We wanted to find what effect bone lesions and more particularly posterosuperior notch in the humerus have on outcome after arthroscopic management of shoulder instability.

Material and methods: We made a retrospective analysis of 60 patients who underwent an arthroscopic procedure between February 1993 and December 1998. Mean age was 28 years three months. All patients in this series had a bony notch measured on the preoperative scan. Six measurements were made to determine the volume of the notch approximated to the half volume of a revolution ellipse. The ratio of this volume to total volume of the humeral head (approximated as a sphere) was also calculated. The glenoid cavities were classed in four groups: normal, abrasion, fracture, amputation. Operative result was assessed with the Duplay score at a mean follow-up of 42.7 months.

Results: Outcome was excellent in 24 patients, good in 14, fair in seven, poor in five including four related to recurrent instability, and a failure in ten requiring reoperation. The rate of recurrence with reoperation was 16.6%. Notch volumes varied from 0 to 4792 mm³ (mean 1019±1253). The notch/ head ratio ranged from 0 to 10.5% (mean = 2.28±2.63). Only 12 patients (20%) did not have a bony notch; 11 of them had an excellent or good result. Mean volume of the notch in patients with excellent, good, or fair outcome (group A) was 639.72 mm³. Mean volume of the notch in the failure cases (group B) was 2158.11 mm³. Glenoid cavity classes in group A were 17 normal, 20 abrasion, 8 fracture, 0 amputation. In group B the classes were 2 normal, 9 abrasion, 2 fracture, 2 amputation. The volume of the notch was significantly correlated with outcome (Spierman test). There was no significant correlation for the glenoid cavity.

Discussion: Despite the absence of a significant correlation with the glenoid cavity classes, the presence of a glenoid lesions in cases with a bony notch smaller than 500 mm³ had an important impact in three out of the four failures (two amputations and one fracture). Considering a threshold value of 1000 mm³, the rate of failure was 7.7% irrespective of the glenoid lesions and 52.5% if glenoid amputations and fractures were excluded.

Conclusion: It would appear that arthroscopic treatment should be reserved for patients with a small humeral notch (< 1000 mm³). Glenoid lesions should also be quantified in patients with small notches.