Introduction

The Latarjet procedure,¹ first described in 1954, involves coracoid bone transfer to stabilize the shoulder by the static action of the transferred bone block and its attached coracobrachialis tendon sling. Although coracoid graft transfer often successfully stabilizes anterior shoulder instability for a selected group of patients,² eventual osteolysis of the transferred coracoid bone graft has been reported.³^,⁴

The glenoid bone graft prevents engagement of a humeral head bone lesion by extending the glenoid arch to such a degree that the shoulder cannot externally rotate far enough for the Hill-Sachs lesion to engage the front of the graft.⁵

An important factor contributing to shoulder stability after a Latarjet procedure is the “sling effect” from the conjoined tendon that slings across the anterior-inferior capsule when the shoulder is in 90 degrees abduction and 90 degrees external rotation. This mechanism acts to prevent engagement of the Hill-Sachs lesion even before the anterior capsule is repaired. Stability from this “noncapsular Latarjet procedure” also is contributed from an increased tension in the inferior fibers of the subscapularis muscle due to the transferred coracoid graft with its attached conjoined tendon over the top of the lower subscapularis tendon. Because we trust in the effects of the Latarjet procedure,⁶^,⁷ we do not stress any capsular repair on the glenoid or any capsular reinforcement.

Revision surgery after a failed Latarjet reconstruction is complex, and the success of revision surgery is low. This procedure should therefore be carefully performed.⁴ In Burkhart,⁸ Allain,³ and Hovelius⁹ series, where the capsule is restored on the glenoid margin or in overlap fashion, mild or residual instability is reported at 5%, 12%, and 7%, respectively. At our surgical center, we have reported a rate of around 10%.

The aim of our project is to identify the many possible reasons for the failure of the Latarjet procedure, excluding new and important traumatic events. Few studies exist that cover the possible causes of coracoid graft resorption. We feel coracoid graft nonunion may be due to the incorrect positioning of the graft on the glenoid wall along its longitudinal axis.

Coracoid graft osteolysis concept and proposal of classification system

For there to be proper bone healing, there should be good contact between the coracoid bone graft and the glenoid bone bed, an adequate blood supply from the coracoid bone graft and the glenoid bone surface, and satisfactory mechanotransduction (mechanisms by which mechanical cues modulate bone healing) at the bone healing site. To date, no classification for coracoid bone graft osteolysis has been described that illustrates how these different factors come into play. We present our understanding of coracoid osteolysis and a classification system that divides the coracoid bone graft into eight regions. The two major regions are superficial and deep, which in turn are each divided into superior, inferior, medial, and lateral (Fig. 17-1).

FIGURE 17-1 Coracoid osteolysis classification. A, The coracoid graft is divided into superficial and deep regions. B, Superior, inferior, medial, and lateral regions are shown, for a total of 8 regions.

During the Latarjet procedure, screws for fixation of the coracoid bone graft to the glenoid neck surface should be directed posteromedially because of the steep glenoid neck surface rather than positioned at 90 degrees to the coracoid bone and glenoid neck bone surface. This is done to avoid penetration into the shoulder joint. Because of this, there tends to be a mismatch of the medial aspect of the coracoid bone graft and the corresponding glenoid neck bone surface, which can affect the healing of the medial side of the bone graft. There tends to be good contact between the graft bone surfaces in the lateral regions of the coracoid graft (Fig. 17-2).

FIGURE 17-2 Screw penetration into the shoulder joint at 90-degree insertion of the coracoid bone onto the glenoid neck bone surface. The screw is therefore directed posteriorly, and this can result in a bony mismatch, with a gap formation between the medial aspect of the coracoid bone graft and its corresponding glenoid neck bone surface.

The blood supply of the coracoid bone graft arises from the coracobrachialis attachment, which is on the inferior region of the graft. This promotes healing in the region. The contact of the glenoid bone surface blood supply with the coracoid graft promotes healing in the deep regions of the coracoid graft (Fig. 17-3).

FIGURE 17-3 Blood supply from the conjoined tendon to the inferior aspect of coracoid graft and from the glenoid bone surface to the deep surface of the coracoid graft. A, Insertion of conjoined tendon. B, Glenoid surface contact.

(Redrawn from Abrassart S et al: Arterial supply of the glenoid: an anatomic study. J Shoulder Elbow Surg 15(2):232-238, 2006).

Mechanotransduction factors can modulate the healing of bone graft.¹⁰^,¹¹ The humeral head tends to influence and promote healing of the lateral regions of the coracoid graft. The coracobrachialis attachment sling on the inferior regions of the graft tends to promote healing of the graft’s inferior regions (Fig. 17-4). If each of these factors is given a score of 1, the superficial superomedial region has the lowest score, which corresponds to our findings on computed tomography studies that indicate that osteolysis most often occurs in the superficial superomedial region. The classification system produces an overall score of 6 on the medial side of the graft and a score of 14 on the lateral side (Fig. 17-5 and Table 17-1). The medial region of coracoid graft therefore has less favorable healing potential compared with the lateral regions.