Chapter 13 Reverse Shoulder Arthroplasty for Massive Anterosuperior Rotator Cuff Failure
Reverse shoulder arthroplasty (RSA) was developed for the treatment of cuff tear arthropathy and irreparable anterosuperior cuff tears in the elderly. Over the course of the last decade, the success of the procedure has led to the expansion of its indications. RSA is now performed to treat proximal humeral fractures in the elderly, significant glenoid defects (in osteoarthritis of the glenohumeral joint with an intact rotator cuff), and as a revision procedure for a failed hemiarthroplasty and total shoulder arthroplasty (TSA). The deltopectoral approach has been the classical approach to perform an RSA.
In this chapter, we detail the anterosuperior approach for performing an RSA. This approach provides an excellent end-on view of the glenoid, the option of subscapularis preservation while enabling accurate placement of implants, earlier return to function, and a lower rate of postoperative instability. We have detailed the sequence of steps that enable the surgeon to perform this procedure in a safe and reproducible manner. As the RSA enters into its third decade of use, better understanding of the modes of failure and survival rates of the prosthesis is understood. Ten-year survival rates of over 90% have been reported. Instability, loosening, infection, and fracture remain the most common causes for revision.
Shoulder motion is a summation of synergistic action of the muscles surrounding the joint. As the joint itself has minimal intrinsic bony stability, its complex movement is primarily dependent on the rotator cuff and the deltoid. The rotator cuff stabilizes the humeral head and depresses it, acting as a fulcrum to the strong abductor moment arm of the deltoid. Rupture of the rotator cuff creates loss of this fulcrum and results in an unbalanced shoulder. This leads to proximal (superior) migration of the humeral head with deltoid contraction. The humeral head with time impinges upon the undersurface of the acromion, which now becomes the new fulcrum for abduction. Neer in 19831 described this process in detail and described the term “cuff tear arthropathy.” He proposed the “preferred method appears to be a resurfacing total shoulder replacement with rotator-cuff reconstruction and special rehabilitation.”1
Repair of the cuff in conjunction with a total shoulder replacement, though a rational idea, was often not possible. Poor quality or irreparable tendon led to early failures of the total shoulder replacement.2 Neer explored constrained shoulder designs, as he believed that constraint would obviate the need for a cuff repair. In an effort to create a fixed fulcrum prosthesis, Neer developed the Mark I, Mark II, and Mark III prostheses. These prostheses lateralized the center of rotation, which often made repair of the residual cuff difficult. Failure of these designs convinced Neer that rotator cuff repair, not constraint, was critical for improving shoulder function.3
The concept of reversing the ball to the glenoid and socket to the humerus saw the development of various prosthesis designs.4 The first and most successful out of these was the design proposed by Grammont in 1987. Grammont5,6 popularized his novel idea, which was based upon four underlying principles: (1) the prosthesis must be inherently stable; (2) the weight-bearing part must be convex and the supporting part concave, (3) the center of the sphere must be at or within the glenoid neck, and (4) the center of rotation must be medialized and distalized.5 This led to the development of the Delta I, II, and III prosthesis. This fundamental concept still holds true and is the foundation of all the current reverse shoulder arthroplasty (RSA) designs.
The Delta prosthesis has been implanted in Europe since the early 1990s and received approval of the Food and Drug Administration (FDA) in the United States in 2004. The RSA was initially developed to treat cuff tear arthropathy. Due to its widespread success, RSA has been increasingly performed for younger patients and for various etiologies beyond cuff tear arthropathy. The present indications for RSA now include cuff tear arthropathy, irreparable rotator cuff tears with or without glenohumeral arthritis, glenoid deformities, acute and chronic trauma, tumor, systemic/inflammatory arthritis, and revision arthroplasty.
13.1.1 Clinical Results of RSA
In a survivorship study of 527 RSA performed between 1985 and 2003 by Favard et al with a minimum of 2-year follow-up, the revision-free survival rate was 89% at 10 years with a breakpoint occurring at 2 and 9 years. Survivorship to a Constant-Murley score of less than 30 was 72% at 10 years with a marked break observed at 8 years. Progressive radiographic changes were observed after 5 years. Increasing frequency of large notches was noted with long-term follow-up. The authors advised caution in performing RSA in a young patient.7 In another study, Walch et al analyzed the 10-year survival of 135 Delta (DePuy) RSA implanted between 1992 and 1999. They divided the analysis into three subgroups: group A (92 cases)—cuff tear arthroplasty and massive cuff tear with pseudoparalysis; group B (39 cases)—failed hemiarthroplasty (HA) and failed TSA; and group C (14 cases)—varied. The survival curve to prosthetic removal showed an overall survivorship of 92% at 10 years. Segmentation according to etiology showed a 97% survivorship for group A and 88% for group B. This difference was not significant. The survival curve for a Constant score of less than 30 showed an overall survivorship of 90% at 10 years. Segmentation according to etiology showed a significant difference at 10 years in favor of group A (92%) compared to group B (86%) with a break of the curve after 9 years for group B.8
The Australian Joint Replacement Registry in its analysis of 9,682 RSA with a cumulative follow-up to 8 years reports no difference in the revision rate based on primary diagnosis. Instability, loosening, infection, and fracture were the top four causes for revision, contributing 42.2, 19.4, 15.2, and 11.4%, respectively, of all revisions. The registry states a cumulative revision rate of 1.35 (1.21–1.50) revisions/100 revision observed years (95% confidence interval [CI]).9
In the aforementioned systematic review, the authors concluded that the complications for an RSA varied based on the prosthesis utilized and the indication for the RSA. The study classified the prosthesis based on lateralized designs and medialized designs. Lateralized prosthesis had a higher arc of rotation and lower rate of scapular notching but a higher rate of glenoid loosening.10 However, the authors concluded that no conclusion could be drawn if the new generation lateralized prostheses increase the risk of glenoid loosening compared to medialized prostheses.
Generally revision surgery (changing an anatomic or reverse prosthesis to an RSA) had a higher complication rate compared to nonrevision procedures.11 Wall et al13 in their study reported 36.7% complication rate in a revision setting compared to 13.3% (p value < 0.001) in nonrevision procedure.
13.1.2 Cuff Tear as an Indication for RSA
A recent systematic review of 35 studies of 2,049 patients by Samitier et al10 and Alentorn-Geli et al11 demonstrated that RSA is an excellent surgical solution with great improvements in clinical outcomes. The authors concluded that primary RSA performed for cuff tear arthropathy or massive anterosuperior cuff rupture had significantly better outcome than an RSA performed as a revision procedure (after hemiarthroplasty or total shoulder replacement). This result has been echoed in other studies. Boileau 12 found that patients with cuff tear arthropathy had significantly higher improvements in Constant score compared with patients undergoing revision of their prosthesis. In a study based on the etiology of RSA, Wall et al13 noted that patients with primary rotator cuff tear arthropathy, primary osteoarthritis with a rotator cuff tear, and a massive rotator cuff tear had better outcomes, on average, than patients who had posttraumatic arthritis and those managed with revision arthroplasty.
13.1.3 Anterosuperior Approach/Transdeltoid Approach for Reverse Shoulder Arthroplasty
The anterosuperior approach was initially described by Neviaser14 and Mackenzie.15 The approach initially described by Grammont and used in the text is a variation of the original description. The anterosuperior approach has been widely utilized for repair of rotator cuff tears and fixation of proximal humeral fractures. The simplicity of the approach and excellent exposure of the glenoid enable the approach to be utilized for an RSA. It has the added advantage of providing the ability to preserve the subscapularis tendon.
The preference of deltopectoral approach to anterosuperior approach is widely debated. In the deltopectoral approach, the main advantage is that it offers an extensile approach, providing excellent access to the humeral diaphysis; however, detachment of the subscapularis and the required shoulder dislocation to gain access to the joint are its main disadvantages. Posterior retraction of the head to obtain visualization of posterior glenoid is often challenging in some glenoid deformities such as Walch B2–3 and Walch C glenoids.
Excessive traction on the posterior retractor (retract the humeral head away, for approaching the glenoid) may result in damaging the anterior deltoid as well as traction injury to the plexus by retropulsion of the scapula.
The main drawback of the deltopectoral compared to the superior approach is the increased risk of instability of the implant.
The anterosuperior approach has the following advantages: direct exposure of the glenoid, deltoid muscle preservation, no need for glenohumeral dislocation, no need to release subscapularis tendon in most cases, better assessment of the infraspinatus or great tuberosity reattachment in case of fracture, and lower risk of postoperative instability.
The disadvantages include the following: risk of weakening the attachment of the anterior deltoid, injury to the anterior branch of the axillary nerve, difficulties to access the humeral inferior osteophyte, and risk of malpositioning (superior tilt) the glenoid implant.
The anterosuperior approach has been widely utilized to perform an RSA. In a multicentric trial of 527 RSA, Molé et al16 compared the outcome of deltopectoral (DP, hereafter) approach with anterosuperior (AS, hereafter) approach (227 AS vs. 300 DP) with a mean follow-up of 2 years. No difference in the Constant–Murley scores or range of motion was noted between the two approaches. The AS approach demonstrated a lower rate of instability (0.8 AS vs. 5.1% DP) and a lower rate of humeral and acromial fracture. Paradoxically, the rate of axial nerve injury was higher in the DP group (0.8 AS vs. 1.3% in DP). The rate of scapular notching was noted to be higher with the AS approach (74 AS vs. 63% DP), as was the rate of glenoid loosening (6.6 AS vs. 2.3%). Loosening was often associated with superior tilt of the glenoid implant.