16.1 Intro/Preop Workup

Proximal humeral fractures are the third most common fracture in the elderly population (second to hip fractures and fractures of the distal radius), occurring at a frequency of approximately 105 per 100,000 individuals per year [1, 2]. As the general population ages and predicted life spans increase, the desire for procedures that restore joint function and maintain quality of life will become progressively relevant. Following this trend, it is likely the rate of arthroplasty for treatment of proximal humerus fractures—particularly in the elderly population—will likely rise in the decades to come [2]. For complicated fractures of the proximal humerus, both hemiarthroplasty (HA) and the reverse prosthesis have demonstrated promising short- and long-term results [3–10]. In the elderly population and situations involving complex fractures, each approach has exhibited superior outcomes in comparison to open reduction internal fixation (ORIF) and nonsurgical conservative management [11–16]. The reverse prosthesis in particular is selected as a revision procedure for failed HA, ORIF, and nonsurgical conservative management [17, 18].

When deciding between the HA and reverse prosthesis for management, it is crucial to weigh the nature of the injury, analyze the advantages and disadvantages of each surgical approach, and understand the patient’s perspective before moving forward. Additionally, it is important to convey pertinent information to patients regarding the potential complications of surgical intervention. Various studies have found a correlation between increasing patient age and poor outcomes using prosthetic implants [19–21], and the existence of comorbid conditions has also been linked to a multitude of complications following surgery for the treatment of proximal humeral fracture [22–24]. It is imperative that patients understand the expected recovery process, long-term expectations following surgery, as well as possible future sequelae from intervention [25].

Prior to selecting an appropriate course of treatment for patients, a multitude of factors must be assessed. Calculation of patient age, overall health status, comorbid conditions, current medication regimen, and prior injury must precede any discussion of intervention [26]. After determining the severity of the proximal humeral fracture, and confirming the need for surgical mediation, additional factors—quality of bone stock, rotator cuff integrity, and regional anatomy—must be considered. It is also essential to understand the injury from the patient’s perspective. Reflection on desired quality of life and anticipated degree of mobility need to be weighed accordingly in order to minimize potential future morbidity from unnecessary extensive intervention.

16.2 Selection of Surgical Approach

Both the deltopectoral and the superolateral (deltoid-splitting) approach are appropriate for surgical approach. The advantage of the deltopectoral approach is the ability to extend the incision distally to expose the humeral shaft with relative ease. Surgeons choosing the superolateral approach need to be comfortable with exposing the axillary nerve in order to get distal exposure if needed.

16.3 Initial Approach

The long head of the biceps and the bicipital groove are often a reproducible landmark used during surgery to assist with tuberosity mobilization. The author’s preference is to perform a tenodesis or tenotomy of the biceps tendon. The biceps tendon can be followed and released out of the rotator interval. The upper rolled border of the subscapularis can be identified and released medially under the coracoid. This will assist with mobilization of the subscapularis. Once the biceps are adequately mobilized, the specific tuberosity fragments can be identified. A fracture line will occasionally travel through the biceps groove separating the lesser and greater tuberosities. If there is no fracture, the surgeon may choose to osteotomize through the bicipital groove to allow exposure to the joint.

Once the lesser tuberosity and subscapularis are mobilized, a suture can be used at the bone/tendon junction to assist with tagging and traction. Retracting the lesser tuberosity fragment anteriorly and medially can usually expose the fractured articular surface. This can be removed with a clamp. The greater tuberosity can be tagged with one or two sutures to assist with mobilization as well. The cephalad portion of the humeral shaft should be cleaned of callus and the superior insertion of the pectoralis tendon identified.

16.4 Hemiarthroplasty

At this time, if a hemiarthroplasty is planned, the surgeon should inspect the glenoid for any damage or irregularities. If it is determined acceptable to proceed, the next step is to prepare the humerus for stem implantation. There are multiple factors that must be considered when positioning the humeral component including initial tuberosity malposition, tuberosity migration or detachment, as well as prosthetic height and degree of retroversion [19].

The height of the prosthesis is usually the first variable adjusted. There are several landmarks for determining optimal humeral head height intraoperatively. One method involves preoperative planning using a ruler and plain film imaging of both sides. Another method is to look for a fracture key of one of the tuberosity pieces. This will allow the surgeon to estimate the prior position of the anatomic articular surface. Krishnan et al. described restoring the “gothic arch” along the medial edge of the humerus and lateral edge of the scapula. This is analogous to the “Shenton line” of the hip [27]. Another useful intraoperative landmark is identifying the superior edge of the pectoralis major tendon, which is approximately 5.5 cm from the top of the humeral head [28]. The surgeon can also make use of an extramedullary jig [29] or an intramedullary sleeve [30] to assist with positioning of the stem during surgery.

The humeral implant should be adjusted to approximately 20° to 30° of retroversion. Many current implant systems have insertion jigs that allow referencing of the version to the forearm. In addition, the surgeon can verify that the humeral head is pointing directly toward the glenoid fossa with the arm at the side and the arm at 0° of external rotation. Aligning the prosthesis in excessive retroversion risks placing the greater tuberosity in excessive tension with internal rotation of the arm [19]. Conversely, placing a stem with disproportionate anteversion can force tension on the lesser tuberosity and subscapularis during external rotation.

Tuberosity management is usually accomplished with heavy sutures. Prior to final impaction or cementation of the humeral stem, it is useful to place multiple drill holes in the metadiaphyseal bone and place sutures through these holes. These vertical sutures will assist with holding the tuberosities reduced to the shaft of the humerus. The tuberosities themselves are managed with sutures passed through the tendon-bone junction. If there is comminution, it may help to place multiple locking stitches into the tendon [31]. Krishnan et al. recommended using two horizontal and two vertical cerclage sutures to provide stability of the tuberosities [27].

Stem selection is per surgeon preference. There are some lower profile stems which require a graft to achieve appropriate lateralization of the greater tuberosity. Stems with wider proximal bodies can usually restore the appropriate lateral offset without a graft. In addition, having appropriate suture holes in the implant can help with maintaining stability of the tuberosities to allow them to consolidate with the metaphysis of the humerus and to themselves.

Postoperatively a 4- to 6-week period of sling immobilization is generally recommended. Tuberosity stability is evaluated intraoperatively and should be confirmed as adequately stable to allow early passive range of motion. Postoperative x-rays are useful to monitor for any sort of tuberosity migration or resorption. Active range of motion can be started at approximately 6 weeks and strengthening at 12 weeks.

16.5 Reverse Prosthesis

The initial approach to perform a reverse prosthesis is typically similar to hemiarthroplasty. Once the tuberosities are mobilized and the humerus is exposed, the greater tuberosity can be retracted posteriorly behind the glenoid. The lesser tuberosity can be retracted anteriorly by a retractor hooked around the anterior margin of the glenoid. The labrum and soft tissues are released to completely expose the glenoid. Once the glenoid has been identified, it is important to remember the cartilage may need to be removed in these cases. The glenoid base plate is inserted in accordance with the implant manufacturer’s technique, aiming for a 10° inferior tilt. The glenosphere can then be placed accordingly.

Much of the humeral preparation is similar to that described above for hemiarthroplasty. Sutures are placed in the metadiaphyseal region for vertical fixation. The implant is placed at approximately 20° of retroversion. The height of the prosthesis can be estimated by the tension on the conjoint tendon and the deltoid. In addition, temporarily reducing the tuberosities can help with judging appropriate height. The surgeon should take care to avoid placing the stem too high, as this may lead to significant overlengthening and a difficult reduction. In contrast, a stem placed too low in the canal can usually be mitigated somewhat by using a thicker polyethylene insert for final reduction.

Prior to insertion of the final stem, it can be useful to place one or multiple sutures around the medial aspect of the prosthesis. Some implants have a hole or smooth area that will not abrade the suture and allow them to slide. This area is often difficult to access once the joint has been reduced. The posterior limbs of these sutures can be passed through the tendon-bone junction of the greater tuberosity. Once the stem is either cemented or impacted into place, the greater tuberosity can be reduced to the appropriate position. The supraspinatus can be resected or recessed to avoid undue rotator cuff tension. The greater tuberosity can be tied down using the sutures around the medial aspect of the stem. Alternately, some reverse implants have holes and fins that can be used to reduce the tuberosity. The lesser tuberosity can be reduced in a similar fashion. The sutures around the medial aspect of the stem can be tied down as a cerclage around both tuberosities and the stem. The vertical sutures can then be used to optimize stability. Once all the sutures have been placed, the tuberosities are tested for stability. This is important to allow for early postoperative range of motion and rehabilitation.

Postoperatively a 4–6 week period of sling immobilization is generally recommended. Tuberosity stability is evaluated intraoperatively and should be felt to be stable enough to allow early passive range of motion. Postoperative x-rays are useful to monitor for any sort of tuberosity migration or resorption. Active range of motion can be started at approximately 6 weeks and strengthening at 12 weeks.

16.6 Results/Outcomes: Hemiarthroplasty (HA)

Internal fixation is often the surgical treatment of choice for displaced fractures in younger patients with intact bone stock and healthy surrounding soft tissue [22]. In the elderly population (>70 years), and situations where fracture location prevents restoration of proper functional anatomy, more invasive surgical intervention is required [32, 33]. For these complex fractures, especially those in which open reduction internal fixation (ORIF) cannot be achieved, primary hemiarthroplasty (HA) can be a stabilizing treatment. Historically, HA was considered the treatment of choice for such complex fractures and is commonly the suggested approach for 3-part fractures, 4-part fractures, complex avulsion fractures, head-splitting fractures, fracture dislocations, and >50% humeral head involvement in impaction fractures [34–36]. The procedure is also often used for fractures with severe displacement and resultant compromised blood supply to the humeral head [37], as well as a salvage for failed ORIF. Moreover, HA is a viable treatment option when ORIF is contraindicated due to risk of malunion, nonunion, implant failure, and osteonecrosis [33, 37].

HA has demonstrated far superior clinical and subjective outcomes for complex fractures compared to ORIF, and nonsurgical conservative management, particularly when humeral head stabilization is not reasonable [5, 10]. A randomized controlled prospective study demonstrated superior outcomes and reported better quality of life following HA as compared to nonoperative treatment in elderly patients with displaced four-part humeral fractures [38]. HA has superior restoration of functional motion compared to internal fixation, albeit with higher reoperation rates [6]. Timing also plays a critical factor in determining treatment, and the decision to perform HA should not be delayed. Early surgical intervention within 2 weeks of the inciting injury is a major factor contributing to positive short- and long-term postoperative functional outcomes [19, 39–42].

Low bone quality frequently present in the elderly population requires cemented stem components, while younger patients may be amenable to non-cemented stems [43]. Modular prosthesis combined with compression osteosynthesis allows for anatomic restoration of tuberosity alignment with the head and shaft, by variable offset of humeral height and retroversion [44, 45]. Natural humeral head retroversion ranges from 19° to 22°, and this should be properly restored when performing HA [46, 47]. In addition to inter-tuberosity fixation, stability of all components is achieved through tuberosity fixation to the selected prosthesis and to the humeral shaft through drill tunnels. Both the deltopectoral approach and the anterolateral deltoid-splitting approach have exhibited comparable success rates [5, 36]

Numerous studies of HA for complex humeral head fracture have demonstrated ideal long-term pain relief from subjectively reported patient data; however, functional outcome and the restoration of full range of motion are not consistent [19, 39]. Various major and minor factors have been shown to influence the outcomes of shoulder HA. It is a technically demanding surgery, and the restoration of natural humeral length, proper anatomic tuberosity reconstruction, as well as ideal retroversion are difficult to achieve [8, 36]. A functioning rotator cuff, avoidance of superior fixation, and anatomically reduced tuberosities are all essential for long-term outcome satisfaction. The most common complications leading to poor long-term outcomes are nonunion and fixation failure, ultimately leading to decreased function [39, 41]. Nonunion of the tuberosity is a relatively rare complication but is often correlated with severe life dissatisfaction and functional results. A retrospective study of 122 consecutive patients with 3- and 4-part fractures demonstrated significantly reduced (p < 0.001) Constant scores in patients with tuberosity nonunion (n = 53) compared to those with fully healed tuberosities (n = 61) at the 3-year follow-up point [48]. The procedure itself has been reported in certain cases to have relatively high failure rates and long-term dissatisfaction. A retrospective review of analyzing over 800 acute fractures treated with HA resulted in an average reported postoperative Constant score of 56.6, as well as poor functional outcomes with an average forward elevation of 105.7 degrees [40]. Similarly, in a retrospective review of 66 patients who underwent HA for proximal humeral fracture, authors noted a 50% incidence of tuberosity malposition and a patient dissatisfaction rate of 42% [19]. In the same study, average final recorded external rotation was 18°, and forward elevation was 101°. Retrospective reviews have confirmed secondary displacement of the greater tuberosity as the key parameter associated with poor clinical outcomes [42, 49, 50]. The maintenance of tuberosity alignment and proper union may also be influenced by humeral head height and version [41, 42] The upper margin insertion site of the pectoralis major tendon is a viable landmark for restoring proper anatomical humeral height and version; 5.6 cm has been recorded as the estimated average when tracing a line between this upper insertion site and the superior point of the humeral head [28]. Greater than 10 mm of humeral head lengthening has been associated with risk of tuberosity detachment due to the extreme tension placed on the supraspinatus [22, 28]. In addition to proper surgical alignment, poor long-term functional and subjectively reported results have been shown to correlate with increasing age and number of comorbidities, especially those related to poor bone stock and degeneration [22, 23].

While some studies have echoed similarities in demonstrating only moderate functional and range of motion improvement, the subjective outcomes reported by patients often exceed these objective shoulder outcomes. In a retrospective multicenter study of 167 shoulders undergoing HA, 79% of patients were asymptomatic or reported only minimal symptoms after the minimum follow-up of 1 year [51]. Of these same patients, only 35% patients were able to abduct past the horizontal plane at the same 1 year follow-up visit (mean abduction of 85–90°). A retrospective review of 82 consecutive patients undergoing HA for severely displaced proximal humeral fracture demonstrated minimal pain in long-term follow-up despite restricted strength and range of motion [24]. Another retrospective review of 71 shoulders at 2 years follow-up reported 93% of patients “pain-free” and satisfied with their results (average ASES 76.6) with average forward flexion of 128, external rotation 43, and internal rotation to L2 [42].

In acute settings HA is preferred over total shoulder arthroplasty (TSA) due to maintenance of normal anatomical landmarks and the avoidance of glenoid-related complications (e.g., component loosening, polyethylene disease). Using HA as the initial surgical management approach also leaves the option of secondary conversion to TSA in the scenario of mechanical failure [52]. Despite reports of functional limitations following HA, certain studies comparing the procedure to reverse total shoulder arthroplasty (RSA) for the treatment of proximal humeral fractures have demonstrated similarities in outcomes. A systematic review comparing the two procedures revealed similarities in subjective outcomes (ASES and Constant score) as well as a 4.0 times greater rate of postoperative complications in RSA in proximal humeral fractures not amenable to surgical fixation [50]. In contrast, while other studies support similarities in early reported outcomes between the two procedures, RSA has demonstrated superior long-term results [53].

16.7 Results/Outcomes: Reverse Prosthesis

Historically, HA was viewed as the gold standard for complex humeral fractures (specifically 3- and 4-part fracture) in the elderly population, and ORIF was seen as the best approach in younger surgical candidates [20, 21]. However, due to the reliance of these procedures—both HA and ORIF—on proper anatomical tuberosity healing, their outcomes have shown to be unpredictable and often result in the need for revision surgery [19, 54]. Reverse shoulder arthroplasty (RSA) has demonstrated long-term stability and patient satisfaction in comparison to HA and has become the surgical treatment of choice for complex humeral fractures, particularly in the elderly population. The RSA prosthetic component provides an advantage over HA by constructing a more balanced anatomical alignment, thereby stabilizing muscular imbalance arising from imperfect glenohumeral orientation. Moreover, the procedure provides the option of grafting bone to fix potential humeral offset [2, 41, 55].

In patients with preexisting shoulder cuff deficiency, clinically significant osteoarthritis, as well as fracture nonunion of the proximal humerus, RSA has exhibited promising short- and long-term results [16, 36, 38, 56]. In comparison to HA, RSA does not rely on proper tuberosity healing, and as such, patients with poor bone quality and osteoporotic degeneration are ideal candidates for the procedure. Additionally, when compared to both conservative management and alternate surgical approaches, RSA has shown superiority in complex 3- and 4-part humeral fractures and with patients suffering pseudoparalysis as a sequela from the injury [2, 7, 57].

While the effectiveness of RSA in comparison to HA has not reached full consensus, the results have nevertheless been promising. In a study comparing functional outcomes of both procedures, 5-year follow-up of patients demonstrated superior Oxford Shoulder Score (OSS) in the RSA group in comparison to HA; 6-month follow-up did not show any significant differences between the two groups [36]. In a network meta-analysis comparing treatment options for 3- and 4-part proximal humeral fracture, Du et al. selected Constant score reliability as the primary data point and found RSA to be superior in comparison to HA, ORIF, and conservative management [6]. In addition to patient-reported subjective data, RSA has exhibited superior functional outcomes in certain domains. In a randomized prospective study comparing RSA to HA, elderly patients (>70 years) who underwent RSA for acute proximal humeral fracture had better forward elevation (120.3 vs. 79.8), abduction (112.9 vs. 78.7), better pain scores, as well as lower revision rates at a mean follow-up of 28.5 months [18]. Cuff et al. also found superior patient outcomes of RSA in comparison to HA in a prospective randomized trial [58]. At a minimum 2-year follow-up, patients in the RSA group reported higher ASES and SST scores, superior radiographic evidence of healing tuberosities, as well as significant improvement in forward elevation.

RSA for the management of proximal humeral fractures has demonstrated promising results as a primary treatment method and has shown superior outcomes when used as a secondary form of intervention following failure of HA, ORIF, and nonsurgical management.

A key advantage of RSA in comparison to alternative surgical intervention is the significantly lower revision and complication rate. Failure of HA in proximal humeral fracture management is typically due to multiple factors: most commonly rotator cuff tear, general instability, glenoid arthritis, component malpositioning, and infection [23, 41, 42, 59]. Additionally, in contrast to HA and ORIF, the success of RSA is independent of anatomical tuberosity healing and, as such, does not risk the development of tuberosity malunion or nonunion [3, 19, 23]. As such, RSA is frequently selected as revision surgery following failure of primary HA [41, 52]. Holschen et al. followed 35 patients who underwent conversion to RSA following failure of primary HA for proximal humeral fracture [60]. At 61 months follow-up patients reported significant improvements in ASES and Constant scores, as well as increased forward flexion and abduction. Similarly, RSA can be used as a salvage procedure after failed ORIF for proximal humeral fractures [11, 61]. In a prospective study evaluating 53 shoulders with subjectively dissatisfactory outcomes following ORIF, patients who underwent RSA demonstrated a mean relative Constant score improvement of 32% at a minimum 2-year follow-up [11]. In certain situations where tuberosity union cannot be achieved ideally—particularly fractures involving elderly patients—cancellous block autografts have shown positive results as an augmentation to RSA [61].

Although RSA has demonstrated largely positive outcomes following proximal humeral fracture, the procedure itself is associated with certain complications. Scapular notching, glenoid component loosening, and synovitis are all commonly reported complications [62–64] Prosthetic replacement must also be considered in implants set in place for greater than 8–10 years [65, 66] Similar to other surgical procedures, common complications such as infection and severe postoperative pain can occur. Despite the possibility of complications, various studies have demonstrated promising short- and long-term outcomes for RSA, and it is increasingly being selected as the treatment of choice in the management of proximal humeral fracture.