Proximal humerus fractures account for 5% of all fractures and are increasing in frequency,¹ particularly among the elderly. It remains the second most common fracture of the upper extremity and third most common fracture in patients older than 65 years. Current research suggests an approximate threefold increase in the incidence of proximal humerus fractures, concomitant with the aging population.² While a majority of proximal humerus fractures are minimally displaced and have acceptable clinical results with nonoperative treatment,³ for fractures with significant displacement, operative intervention is often indicated. Displaced three- and four-part fractures, treated nonoperatively, can result in chronic pain and loss of function. Unsurprisingly, due to the bone morphological changes associated with aging, elderly patients sustaining lower energy injuries often present with more complicated fracture patterns than the younger population¹ and, as a result, can present a formidable challenge. Locked plating for proximal humerus fractures has been associated with high rates of osteonecrosis, intra-articular screw penetration, and loss of fracture reduction, resulting in inconsistent results. Shoulder hemiarthroplasty (HA) has shown reliable outcomes with respect to pain, but has been associated with unpredictable results for range of motion and function, likely due to tuberosity displacement, nonunion or resorption, and rotator cuff dysfunction.⁴

Traditionally, reverse total shoulder arthroplasty (RTSA) has been used to manage cuff tear arthropathy and other glenohumeral disorders in the setting of irrepairable rotator cuff tears or insufficiency. However, with increasing experience, the indications have been expanding. Due to the unpredictable results of operative management of proximal humerus fractures with open reduction and internal fixation (ORIF) and HA, RTSA has become a reliable and predictable alternative in appropriately selected patients (VIDEO 29.1).

There have been numerous classification systems developed for proximal humeral fractures due, in part, to the difficulty of defining the fracture patterns. Codman first attempted to classify these fractures in the 1930s by dividing the proximal humerus into four parts: the greater tuberosity, the lesser tuberosity, the head, and the shaft based upon the epiphyseal segments. Neer built upon this scheme in an effort to better define fracture displacement, which he defined as angulation >45° or greater than 1 cm of separation. Despite being the most commonly used classification system, studies have found the Neer classification to have relatively poor interobserver reliability (κ = 0.52).⁵ The AO classification of proximal humerus fractures places emphasis on the blood supply to the articular surface and divides these fractures into three groups, with increasing rates of osteonecrosis: A (extra-articular, unifocal), B (extra-articular, bifocal), and C (extra-articular with compromise to the vascular supply).⁶ Hertel devised a system for reading proximal humerus fractures where classifications are classified using a binary description system used to classify the four fracture parts and five basic planes of fracture.⁷ Despite these various classification systems (FIGURE 29.1), no one system has shown accuracy in predicting the development of osteonecrosis in the setting of acute proximal humerus fractures.

A special consideration for proximal humerus fractures of the three- and four-part variety in the elderly are that these fractures are often associated with a high incidence of osteonecrosis and poor bone healing.¹,² Multiple factors have been associated with increased osteonecrosis and complications including loss of the medial hinge, four-part fracture patterns, posteromedial metaphyseal head extension of less than 8 mm, angular displacement of the head greater than 45°, and fracture displacement greater than 1 cm.⁷ Treated nonoperatively, these fractures can result in debilitating malunion and poor function.³ A recent review article by Iyengar et al found that three- and four-part proximal humerus fractures reached a 98% rate of radiographic union; however, they were associated with highest overall complication rate at 48% (23% rate of varus malunion and 14% rate of osteonecrosis).³ ORIF with locked plating was initially a promising treatment modality for proximal humerus fractures. However, multiple studies have demonstrated high complication rates due to loss of reduction, screw cutout of up to 67%, postoperative stiffness, and osteonecrosis
in up to 55% of cases.⁴,⁷ Studies have shown that poor bone quality is associated with a significantly increased failure rate with ORIF.⁶,⁸ HA alleviates the threat of osteonecrosis and screw cutout; however, the functional outcomes are completely dependent on tuberosity union, and multiple studies have demonstrated postoperative tuberosity osteolysis and displacement with subsequent loss of motion and function and poor patient satisfaction.⁵,⁹,¹⁰,¹¹,¹²,¹³ Additionally, fractures with non reconstructable tuberosities have been associated with poor outcomes following ORIF and HA. Concern for humeral head osteonecrosis, poor bone quality, compromised healing
capacity, and poor functional outcomes following treatment of three- and four-part fractures has made RTSA a favorable treatment option.

When evaluating any patient with an acute proximal humerus fracture, it is important to appropriately assess the patient’s general health status. This includes an assessment of the patients’ medical comorbidities, cognitive status, functional demands, and expectations. A patient unable to comply with a postoperative rehabilitation program or restrictions will compromise the outcome of the surgery. A history of previous shoulder injury or surgery should be identified since it may factor into the selected treatment plan. Neurological status must be accurately documented, particularly axillary nerve function, which has been reported to be injured in up to 67% of patients with low energy proximal humerus fractures.⁹ Although deltoid function is considered essential for a successful RTSA, Ladermann et al in a retrospective study of 49 patients undergoing RTSA with deltoid dysfunction reported a 98% patient satisfaction rate based upon increases in forward elevation and Constant scores.¹⁰ Nonetheless, careful examination of deltoid function is imperative and is performed by assessing for contraction of all three portions in an isometric assessment of elevation, abduction, and extension. Assessment of deltoid function in a patient with an acute proximal humerus fracture can be a challenge but is essential. If contraction is confirmed, the function of the muscle is thought to be adequate. If a contraction is not confirmed, then an electromyography and further workup is required.

Radiographs should include a standard shoulder trauma series. This includes a scapular anterior-posterior or Grashey, axillary, and scapular-Y view. Images of the contralateral shoulder may be useful for templating purposes. A CT scan is useful for more precisely defining bone and fracture morphology including tuberosity comminution and displacement, articular segment involvement, degree of osteopenia, and possible glenoid fracture. If preoperative planning software is utilized, then thin slice CT scans of less than 1 mm are necessary.

Due to the unpredictable results of ORIF and HA and the successful results of RTSA for an expanding spectrum glenohumeral pathologies, RTSA has become an attractive option for the treatment of complex proximal humerus fractures. Our indications include displaced four-part fractures and fracture-dislocations; displaced three-part fractures and fracture-dislocations in selected patients with poor bone quality that is not amenable to ORIF; fractures involving a head split in the elderly; proximal humerus fractures with associated glenoid fractures compromising glenohumeral joint stability; fractures with significant tuberosity comminution not amenable to ORIF; and elderly patients requiring early mobility and function of injured arm such as wheelchair dependence or contralateral paralysis.

Contraindications to RTSA for fracture include an associated brachial plexopathy and open fractures. Isolated axillary nerve compromise is a not an absolute contraindication since the vast majority resolve over time. However, it should be closely monitored both preoperatively and postoperatively. Associated injuries to the glenoid and acromion, although very uncommon, require careful assessment to determine if it will impact the outcome of RTSA. Since the proximal humerus is a common site for primary and metastatic oncologic lesions, if there is suspicion for a pathologic fracture, further preoperative evaluation is required including consultation with an orthopedic oncologist before proceeding with RTSA.¹¹,¹²

The surgeon’s preference is to use general anesthesia without a regional anesthetic. This allows for an immediate postoperative neurological examination. A periarticular infiltration of a pain cocktail is also used for postoperative pain control at the conclusion of the procedure, which will be described. Patients are positioned in the beach chair position with the operative arm secured to an arm positioner (FIGURE 29.2). Ensure the endotracheal tube is secured on the opposite side of the operative arm. A standard deltopectoral approach is utilized, with care taken not to compromise the deltoid muscle because of its importance in postoperative stability, range of motion, and overall function.

Once the deltopectoral interval is identified, the cephalic vein is most commonly mobilized lateral with the deltoid. There are several soft-tissue considerations to be mindful of which drastically improves surgical exposure. If necessary, the pectoralis major tendon insertion can be released approximately 1 cm from its insertion onto the humerus to aid in exposure. Subacromial and subdeltoid adhesions, hematoma, and any early callus must also be released to allow adequate visualization. At this time, the biceps tendon is usually identified at the bicipital groove and tenotomized with a 2-0 Vicryl suture.

This gives the surgeon exposure to the greater and lesser tuberosities. In three-part fractures, an osteotome is often necessary to separate the greater and lesser tuberosity at the bicipital groove. Once the tuberosities are isolated, they are secured using two braided, nonabsorbable #5 heavy sutures at the tendon-bone interface evenly spaced to allow control of the tuberosity fragments. They are then mobilized to allow the fracture morphology to be assessed (FIGURE 29.3). If there is comminution of the tuberosities present, it is suggested that the comminution remain attached to the rotator cuff. At this point, the humeral head should be extracted and saved to be used for bone grafting (FIGURE 29.4). The status of the humeral neck/calcar, the quality of cancellous and cortical bone, and additionally comminution should be assessed at this time.

Once the humeral head has been extracted, it is the surgeon’s preference to first prepare the glenoid. In the context of proximal humerus fracture, glenoid exposure is generally more easily achieved due to capsular injury and loss of proximal humeral bone from the trauma. The proximal biceps stump as well as the labrum must be completely excised from the glenoid face. Care should be taken during the inferior capsulolabral resection to avoid injury to the axillary nerve (FIGURE 29.5). It is recommended to detach the capsule directly off the glenoid bone to avoid neurologic injury. Once glenoid exposure is achieved, placement of a central pin perpendicular to the glenoid allows for reaming, which is particularly important in setting of acute fracture due to the abundant articular cartilage that may be present (FIGURE 29.6). If preoperative templating identifies significant glenoid deformity or version, glenoid reaming can be adjusted to allow either bone or metal augments to be utilized. Once the glenoid is reamed, either a central screw or post is prepared, and the glenoid baseplate is inserted and should sit flush with the bone. Peripheral screws are used to further secure the baseplate to the scapula. A glenosphere is then applied onto the baseplate in the standard fashion (FIGURE 29.7). Depending on the amount of proximal humeral comminution, consideration for a lateralized or larger glenosphere should be given to improve postoperative stability.

The humeral canal is then exposed (FIGURE 29.8) and prepared using hand reaming to prevent perioperative fracture in osteoporotic bone. Using trial components, the humeral prosthesis height and version of the implant is measured relative to the normally intact humeral calcar. If significant comminution of the humeral canal is
present, the superior border of the pectoralis tendon is utilized, and a distance of 5.0 cm is used to identify the most superior aspect of the humeral prosthesis.¹⁴ Alternatively, radiographs of the contralateral unaffected shoulder can be obtained to better approximate humeral height.¹⁵ The proper humeral height is critical to retain appropriate soft-tissue tensioning, with instability being the major consequence of inappropriate tensioning. A version rod is generally used to guide the amount of planned version for the humeral stem. It is our preference to place the humeral stem in 20° to 30° of retroversion, as these values have been found to provide optimal stability and minimal scapular notching.¹⁶,¹⁷ Often, a surgical sponge (raytec) is used to help insert humeral trial implants to control rotation, and a trial reduction is performed and stability assessed (FIGURE 29.9). The reduction of the tuberosity fragments must be assessed at this time as well as tuberosity healing is critical for optimal postoperative function.