11.2.1 Rotator Cuff Arthropathy

Around 2% of people over the age of 80 years suffer from cuff tear arthropathy [8], where glenohumeral arthritis exists in the presence of a massive cuff tear, with the humeral head either remaining concentric (Seebauer 1) or migrating superiorly (Seebauer 2) [7]. Symptomatically, patients may suffer with severe pain, particularly at night, with pseudoparalysis of the arm significantly affecting function. Rotator cuff arthropathy formed the basis for the early studies on RSA, as, prior to this, there were few alternatives [9–11]. This remains the primary indication, for which RSA was approved by the FDA in 2004 in the USA, with studies showing excellent clinical outcomes [12–14].

11.2.2 Immunological Arthritis with Rotator Cuff Tears

While good results for pain relief have been reported with anatomical shoulder arthroplasty for inflammatory arthritides, including rheumatoid arthritis (RA), psoriatic arthritis, inflammatory bowel disease-associated arthritis, scleroderma, and ankylosing spondylitis [15], improvements with range of motion and function have been less satisfactory [16–18]. Furthermore, secondary rotator cuff failure, with proximal migration of the humerus and loosening of the glenoid component, remains a risk [19, 20]. Studies exploring the use of RSA in patients with RA, in the presence of a cuff tear, demonstrate good functional and pain outcomes [21, 22]; however, concerns have been raised over the longevity of the glenoid component, given evidence of glenoid radiographic lucencies on follow-up [22, 23]. Ekelund and Nyberg [24] and Guery et al. [25] have refuted this, demonstrating statistically significant improvements in pain and range of movement for patients with RA undergoing RSA, with the latter showing this to be the case regardless of status of the rotator cuff. As with any joint replacement in RA, increased risk of infection should be borne in mind [19, 20].

11.2.3 Acute Proximal Humerus Fractures

RSA also has a potential place in trauma. Patients who suffer fractures of the proximal humerus that are deemed to require hemiarthroplasty carry the risks associated with the poor quality of the rotator cuff and the reliance on adequate healing of the tuberosities in the elderly population [26, 27].

Several studies have described outcomes for patients undergoing RSA for Neer 3- or 4-part fractures or fracture-dislocations of the glenohumeral joint [27–31]. Results for function and pain, as expected for patients undergoing arthroplasty following trauma, were not as good as for those with rotator cuff arthropathy but equivalent to the alternative of hemiarthroplasty, though without the potential complications of nonunion or malunion of the tuberosities, cuff failure, or erosion of the glenoid [32–34]. It has been shown that RSA is not reliant on healing of the tuberosities [35] and recovery can be quicker than treatment with hemiarthroplasty [36].

11.2.4 Complications of Fracture Healing

Following nonoperative management of proximal humerus fractures, pain and stiffness can be common problems, with altered anatomy, wear of the glenoid, and the possibility of cuff failure [37, 38]. Studies investigating the use of RSA for treatment of fracture sequelae have shown improvement in range of motion and pain comparable to outcomes for patients undergoing RSA for rotator cuff arthropathy [37–39].

Malunion of the proximal humerus is not uncommon following fracture. Asymmetry, in addition to fatty atrophy of the rotator cuff [15], does not create a favorable environment for anatomical shoulder replacement, resulting in uneven forces across the glenoid component [40] and failure because of the “rocking horse” mechanism [41]. Poor results have also been associated with greater tuberosity osteotomy, which may be required during surgery [42]. RSA provides a preferably alternative, because, as described above, it does not require anatomical healing of the tuberosities, removing the requirement for an osteotomy [40]. Comparison of anatomical shoulder replacement and RSA for malunited fractures has shown RSA to produce superior postoperative results [43], although results remain inferior to those achieved in patients who have had RSA for cuff arthropathy and complication rates are higher [44].

11.2.5 Shoulder Dysplasias

It is estimated that 3.5% of patients undergoing shoulder arthroplasty had glenoid dysplasia [45]: this can occur as an isolated congenital condition or in association with a variety of other conditions, including epiphyseal dysplasia, muscular dystrophy, post-traumatic injury, post-infection, arthrogryposis, or obstetric trauma [46]. One case report has described the successful application of RSA in a patient with a dysplastic glenoid secondary to Kniest syndrome [15], but these conditions are rare, so evidence for RSA is limited. However, it has been recognized that it is important to distinguish between a type B2 glenoid and the hypoplastic glenoid, as soft tissue can adapt to glenoid morphology to keep the humeral centered, with implications for orientation of implants during surgery [47, 48].

11.2.6 Revision Surgery

Common reasons for revision surgery to the proximal humerus include failed treatment of fractures (as described previously), ongoing pain, and loss of function. Revision with RSA has also been described for infection (either single- or two-stage revision) [49, 50], after cuff failure following an anatomical shoulder replacement or hemiarthroplasty [44, 51] and after baseplate failure with previous RSA involving a large glenosphere [52].

One retrospective study, comparing revisions for all types of failed arthroplasty, showed significantly lower functional scores with RSA than for patients with primary RSA as well as double the complication rate at 1 month [53]. These results should, though, be taken in the context of the alternative to revision with RSA in the situation of a poor rotator cuff being arthrodesis, which has been shown to have poor results [54]. For this reason, it is recommended that arthrodesis remain a salvage option for when revision with RSA is not feasible [54].

11.2.7 Glenohumeral OA with Severe Glenoid Bone Loss

Severe bone loss involving the glenoid can be seen in a number of shoulder conditions, particularly after failed primary shoulder arthroplasty, inflammatory arthropathies, chronic dislocations, and osteoarthritis with posterior instability [55, 56]. Controversies remain regarding the treatment of this bone loss, although bone grafting is recommended for type B2 and C glenoids [57]. In terms of bone loss in the presence of an intact rotator cuff, the principle of management remains removing enough bone to make a flat surface for the glenoid component and using bone graft to fill the defect [15].

RSA has advantages of anatomical implants in this situation, as the use of a glenoid component which utilizes a central screw for purpose (as seen in a number of RSA systems) negates the requirement for bone grafting, although cases of implantation with a hybrid bone graft (cancellous autograft and femoral neck allograft) have been described [58].

Concerns remain over long-term follow-up in these patients, however, as medializing the glenoid to accommodate the bone loss, while more stable with an implant utilizing a central screw, can alter the mechanics of the prosthesis and potentially lead to loosening [15].

11.2.8 Chronic Glenohumeral Dislocation

Chronic dislocations can be associated with a number of complications: rotator cuff tears, bone loss, soft tissue contractures, osteoporosis of the humeral head, and softening of the articular cartilage [59, 60]. Massive rotator cuff tears are often present, with substantial humeral head or glenoid bone loss [15]. Indications for arthroplasty, in terms of chronic dislocations, include humeral head defects of greater than 50% or significant degenerative changes [59, 61]. While treatment with RSA is not common, several studies have demonstrated improvement in pain and range of motion for patients [62–64].

11.3.1 Deltoid Function

Deltoid function is necessary for active elevation after RSA, and therefore RSA in the presence of impairment will consistently and predictably result in a poor functional outcome [15, 65, 66]. De Wilde et al. particularly highlighted the anterior and middle heads to be crucial to success of RSA [67]. While case reports exist in which patients have been treated with RSA and augmentation of the deltoid (e.g., via latissimus dorsi transfer [68]), the success of these techniques remain in question.

11.3.2 Active Infection

As with any joint replacement, implantation in the presence of active infection is contraindicated.

11.3.3 Age

While improvement in outcomes for patients, particularly those with rotator cuff arthropathy, can be marked following treatment with RSA, concerns remain over longevity. Guery et al. have demonstrated some decrease in function at a mean follow-up of 8 months and a survivorship of 58% at 10 years [25]. Their recommendation that RSA be restricted to patients over 70 years of age is supported by evidence of long-term failures from other studies [24, 27, 69].

11.3.4 Isolated Supraspinatus Tear

An isolated supraspinatus tear does not, in isolation, produce an unbalanced shoulder. While not an absolute contraindication to RSA, Edwards et al. reported, in over 500 cases with 43 months average follow-up, that patients with glenohumeral arthritis and an isolated supraspinatus tear demonstrated the same functional scores, range of motion, and satisfaction ratings as patients without a rotator cuff tear when treated with anatomical shoulder replacement [48].