[Effect of the rotator cuff condition on the results of shoulder arthroplasty]

The purpose of this report was to study the repercussions of rotator cuff condition on unconstrained shoulder arthroplasty results.

Between 1986 and 1993, 40 unconstrained shoulder prostheses were performed (24 total and 16 hemiarthroplasties). At the time of the operation, rotator cuff condition was reported to be normal in 15 shoulders, atrophic in 10, scarred in 3, torn in 12. Clinical and radiographical results were analyzed and supplemented with a rotator cuff echography. Postoperative follow-up averaged 42.25 months (12 to 97 months).

When the rotator cuff was intact at the time of operation, clinical results were the best at the time of review. Constant score, overall mobility, forward elevation, external rotation with the elbow along the side were better for intact than for atrophic, scarred or torn rotator cuffs (p < 0.01). Ruptures repaired by local tissue transfer or trapezo deltoidal flap substitution (3 cases) gave satisfactory results, whereas two dacron cuff prostheses failed. In the non repaired ruptures (7 cases), arthroplasty resulted in 3/4 fair or poor results. At the time of review, the total incidence of postoperative rotator cuff tears was high (12 cases), functional repercussions were variable. Among the 15 rotator cuff tears observed at the time of the review, the long head of the biceps was present and in place in 9 cases and ruptured in 6 cases. In these latter cases, Constant score (p < 0.01), the forward elevation (p < 0.01), external rotation with elbow along side (p < 0.05) and overall mobility (p < 0.05) were worse; moreover, humeral head superior migration was greater (p < 0.01).

Our study confirms the superiority of clinical results when the cuff was intact at the time of the operation. The repair of rotator cuff tears seems to be justified. When the rupture has not been repaired, clinical results were fair or poor in 5 cases out of 7, the initial rupture spread to the other tendons in 3 cases out of 7. Secondary ruptures were frequent (12 cases). No predictive factor was identified, but an overlapping of the greater tuberosity appears to be detrimental. Among rotator cuff tears present at the time of review, long head of the biceps rupture jeopardized clinical results and was associated with a proximal migration of the humeral head which was significantly more severe than with an intact non displaced long head of the biceps.

During unconstrained shoulder prosthesis implantation, it is therefore recommend to systematically repair any associated rotator cuff rupture, in order to avoid any greater tuberosity overlap relative to the prosthetic head and to preserve the long head of the biceps tendon which limits upward migration of the humeral head and improves prosthetic kinetics.