Reverse Shoulder Arthroplasty: A Role Defined

Results of Hemiarthroplasty 

Because of the previously mentioned failures of both nonconstrained and constrained total shoulder arthroplasty designs for cuff tear arthropathy, the current recommendations for this condition consist of either hemiarthroplasty or the reverse prosthesis. Humeral head replacement in this situation has been the standard of care based on several well done retrospective studies with mid- to long-term follow-up and acceptable results (Fig. 1).2, 3, 6, 7, 8, 9, 15

View Large Image Download to PowerPoint

Figure 1. (A) Anteroposterior (AP) radiograph demonstrating superior migration of the humeral head, acetabularization of the coracoacromial arch, including the acromion, distal clavicle, and coracoid process. Also note femoralization of the humeral head (rounding off of the greater tuberosity), formation of a superior “biconcave” glenoid, and joint space narrowing. These changes are characteristic of cuff tear arthropathy. This patient is a 78-year-old female with preoperative active elevation of 95°. (B) Coronal magnetic resonance image (MRI) showing these bone changes in addition to a massive rotator cuff tear retracted to the level of the glenoid. (C) Intraoperative photograph of the hemiarthroplasty cemented into position. (D) Postoperative radiograph showing a well-positioned cemented hemiarthroplasty. Postoperative active elevation was 110°. (Color version of figure is available online.)

Pollock and coworkers6 were awarded the Charles S. Neer Award for their study showing improved outcome of humeral head replacement versus total shoulder arthroplasty in patients with rotator cuff deficiency. In their study, shoulders with hemiarthroplasty similar satisfaction, pain relief, and functional improvement compared with those with total shoulder arthroplasty, but those with hemiarthroplasty gained significantly greater active elevation after surgery (+52° versus +2°) as well as decreased operative time and decreased blood loss. Persistent postoperative pain remained in 2 of 19 patients at average 3 year follow-up (range 1-9 years). Arntz and coworkers9 showed, in a series of 18 shoulders, that active forward elevation improved from an average of 66° preoperatively to 109° after surgery, with 3 of 18 patients with persistent postoperative pain at average 3 year follow-up (range 2-10 years).

Williams and Rockwood8 retrospectively reviewed 21 cases of hemiarthroplasty for glenohumeral arthritis with rotator cuff deficiency at average 4 year follow-up (range 2-6.6 years). Using Neer’s “Limited Goals Criteria,” they reported 86% satisfactory results, with 3 of 21 patients reporting persistent moderate pain and active elevation improving from 70° before surgery to 120° after surgery. Similar midterm results, indicating that pain generally improved, although it was unpredictable, and active elevation also improved, but rarely to the overhead level, have more recently been reported by Field and coworkers,3 Sanchez-Sotelo and coworkers,7 and Zuckerman and coworkers.15

Goldberg and coworkers2 recently reported on a series of 34 shoulders at an average of 8.5 years follow-up (range 2-16 years) that had undergone hemiarthroplasty for glenohumeral arthritis with massive rotator cuff tears. It was found that, even in the long term, 76% of shoulders were satisfied according to Neer’s limited goals criteria, and that the mean improvement in active forward elevation was 78° before surgery to 115° at early follow-up, 119° at midterm follow-up, and 108° at long-term follow-up. Interestingly, in this study, it was found that, when patients are grouped into those with preoperative active elevation of less that 90° and those with greater than 90°, the results were significantly different. The group with greater than 90° of preoperative motion had 88% satisfactory results and had significantly better function, pain relief, and higher total American Shoulder and Elbow Surgeons (ASES) scores than patients who were unable to elevate the arm 90° before surgery.

Results of Reverse Arthroplasty 

Because the reverse prosthesis based on the Grammont design is a more recent innovation, the length of follow-up is necessarily shorter than for hemiarthroplasty on average (Fig. 2). However, there are some well done retrospective studies with midterm follow-up that are available for comparison. Studies examining the reverse have made it clear that the results are highly dependent on the indication for surgery. For example, studies in which the indication for surgery was primary cuff tear arthropathy without previous arthroplasty, the results are better and the complications fewer than in those studies that included revision of existing arthroplasty to reverse. In the largest study focusing exclusively on cuff tear arthropathy, Sirveaux and coworkers16 reported on 80 shoulders at a mean follow-up of 44 months (range 24-97) after insertion of the Grammont prosthesis. Persistent pain remained in 4% and mean active forward elevation improved from 73° to 138°, and the constant score improved from 22.6 before surgery to 65.6 after surgery. They found that the presence of the teres minor preoperatively had a significant influence on the postoperative Constant score and particularly on postoperative external rotation. The reoperation rate was 5% in this study. Other studies focusing on reverse for cuff tear arthropathy have reported reoperation rates ranging from 0 to 13%.17, 18

View Large Image Download to PowerPoint

Figure 2. (A) AP radiograph of an 84-year-old female with 30° of preoperative active elevation. Note characteristic findings of cuff tear arthropathy. (B) Coronal MRI demonstrating a massive rotator cuff tear. (C) Intraoperative photograph of humeral stem and prepared canal. (D) Postoperative radiograph of well-positioned reverse prosthesis. Postoperative active elevation was 160°. (Color version of figure is available online.)

In the past 2 years, major shoulder centers have reported their 3-year results with the reverse prosthesis used for multiple indications. Frankle and coworkers19 reported results of 60 shoulders at average 33 month follow-up. Indications in this study included cuff tear arthropathy, fracture, and rheumatoid arthritis, but not revision arthroplasty. Active elevation increased from 55° to 105°, ASES scores increased from 34 to 68, and the visual analog pain scale improved from 6.8 to 2.2. There was a 17% complication rate and 12% revision rate in this series. Werner and coworkers20 reported results on 58 patients followed for an average of 38 months (minimum 24 months), with indications including cuff tear arthropathy and revision arthroplasty. Active elevation improved from an average of 42° to 100°. The overall complication rate was 50% (29/58), with a 33% (19/58) reoperation rate. Twenty-one patients underwent reverse arthroplasty as a revision of a previous arthroplasty, with a higher complication rate (57%) and a higher reoperation rate (38%). Complications included hematoma formation, infection, dislocation, prosthetic loosening, nerve damage, and acromion fracture (Fig. 3). In addition, the problem of scapular notching is well known, but the long-term sequelae of this phenomenon remain unknown and could be severe in longer-term follow-up.

View Large Image Download to PowerPoint

Figure 3. Complications seen at our institution: (A) Anterior dislocation of reverse performed as a revision arthroplasty. (B) Dissociation of the glenosphere. (C) Glenoid baseplate loosening with screw breakage.

Indications: “A Role Defined” 

Multiple studies have demonstrated that hemiarthroplasty for glenohumeral arthritis with rotator cuff deficiency provides acceptable levels of pain relief and patient satisfaction with long-term durability. However, in every series there is a small percentage of patients with persistent pain, and, especially in patients with preoperative pseudoparalysis, functional improvement can be minimal. The reverse, on the other hand, often results in greater pain relief and dramatically improved active elevation, but there is a high rate of early complications and reoperations. The key to success with both of these operations is in selecting the appropriate indications.

General requirements for the reverse prosthesis include a low-demand patient, typically over the age of 70, with adequate glenoid bone stock for cementless fixation, a functional deltoid muscle, and the ability to comply with postoperative instructions. Because hemiarthroplasy patients with preoperative active elevation of greater than 90° have significantly better results in terms of postoperative elevation, ASES scores, and pain relief than those with pseudoparalytic shoulders elevating less than 90°, and the complication rate for hemiarthroplasty is dramatically lower than for the reverse, it is clear that these who are able to raise their arms before surgery should undergo hemiarthroplasty. In patients with pseudoparalysis, however, a hemiarthroplasty may relieve some pain, but they may have persistent pain and will be unlikely to enjoy any significant functional gains. These patients should be counseled on the relative advantages of the reverse prosthesis but should also be aware of the risk of complications. This role for the reverse is fairly straightforward.

As the surgeons comfort level grows, however, indications are often expanded. The use of a reverse arthroplasty for revision of a failed shoulder arthroplasty with a nonfunctional rotator cuff is fraught with complications. In this situation, the probability of a good outcome is poor with all currently available treatments. Hemiarthroplasty is the solution least likely to result in future reoperations and serious complications but is unlikely to improve the patient’s function significantly. A reverse arthroplasty may produce dramatic results in terms of improving function if all goes well and complications are avoided, but the risk of serious complications in this patient group is up to 50% as previously discussed. A good way of assessing the advisability of a reverse prosthesis in these situations is to ask whether the patient would be better off with a resection arthroplasty than in their current condition. If the answer is “yes,” then perhaps it is worth the significant risk to implant a reverse prosthesis, because, if it does fail, the patient may be forced to ultimately accept a resection arthroplasty. If the answer is “no,” then perhaps it would be better not to gamble on a good outcome with the reverse prosthesis and “play it safe” with a hemiarthroplasty. New potential indications are beginning to emerge, including primary treatment for badly displaced four-part proximal humerus fractures and pseudoparalysis due to massive rotator cuff tears in the absence of glenohumeral arthritis. Further study is warranted before such indications can be addressed.

Conclusion 

It is clear that the reverse prosthesis is an exciting new technology, with the promise of addressing the shortcomings of the gold-standard hemiarthroplasty, namely, unpredictable pain relief and failure to improve active motion in the pseudoparalytic shoulder. While enthusiasm is clearly warranted since the prosthesis offers the possibility of a good result in clinical situations in which previously there was none, expansion of indications should be done with caution.

Indications are clearly the most important factor affecting results at this point. Some indications, such as pseudoparalysis combined with cuff tear arthropathy or osteoarthritis with rotator cuff deficiency, are supported by the literature as having improved clinical results with a relatively low risk of complications. Other indications, such as revision of failed arthroplasty, may be reasonable despite high complication rates as long as the patient understands the risks. Other indications will require more research before they can be recommended. Perhaps Gilles Walch stated it best during the reverse symposium at the 2006 AAOS Annual Meeting, “The best indication is when there is nothing else satisfactory to propose.”