Proximal humerus fractures can be classified in different ways [5] - [10]. Arguably the most widely used is the Neer classification [5]. Neer divided the proximal humerus into 4 distinct parts: the greater tuberosity, the lesser tuberosity, the humeral head and the proximal shaft. They then classified fractures based on the displacement and angulation of these 4 parts (see Figure 1). According to the Neer classification, displacement of greater than one centimetre or angulation

of more than 45˚ is required for classification as a part [5] [6]. The significance of these 4 parts is each has a different deforming force meaning that reduction of 3 and 4 part fractures is challenging.

Hertel (2004) devised a similar classification and described 5 basic fracture planes with 12 distinct fracture patterns (see Figure 2). Hertel proposed that it was the fracture pattern rather than the amount of individual fragments that was most important. Hertel was interested in finding factors that could predict the likelihood of avascular necrosis of the humeral head. Factors they found to be important were; Length of metaphyseal head extension and Integrity of the medial hinge [7]. Importantly, they also correlated the fracture pattern with the incidence of humeral head ischemia with a positive predictive value of 0.7 [7] [8].

Although the Neer classification is widely used it is not correlated to humeral head ischemia. This is important because humeral head ischemia is a major contributing factor when deciding whether a proximal humerus fracture needs operative intervention with arthroplasty [11] [12] [13]. Ultimately, regardless of what scoring system one prefers, clinical judgment will be required when devising a management plan for a proximal humerus fracture. This case series hopes to provide information to both patients and clinicians when deciding if a reverse shoulder arthroplasty will be their best treatment option.

Up to 85% of Proximal humerus fractures are minimally displaced and can be treated conservatively [3] [14]. There is no consensus on what is the optimal treatment strategy for more complex fractures e.g 3 and 4 part Neer fractures or Hertel type 12 [5] [6] [7] [8]. Various techniques have been described for fixation of proximal humerus fractures. Some of which include; Intramedullary nailing [15] , Tension band wiring [16] , Closed reduction and pining [17] and open reduction with a various combinations of plates and screws [12] [13] [18].

Recommendations for most displaced two- and three-part fractures include internal fixation, with possible prosthetic replacement in three and four-part fractures that are not amenable to reconstruction [6] [11] [12] [13]. It has been well described that more complex fractures are more likely to disrupt the blood supply to the humeral head leading to AVN [7] [8]. Studies have suggested that AVN of the humerus head can be as high as 90% in some four?part proximal humerus fractures [6] [19]. Considering this data it is easy to see why a prosthetic shoulder replacement is an attractive treatment option for complex proximal humeral fractures.

To date there has been few randomised control trials looking at conservative management compared with arthroplasty for three and four part proximal humerus fractures [20]. The poorer outcomes associated with displaced, multipart fractures have resulted in surgical treatment options being explored. Neither the use of plates or replacement of the fractured humeral head (hemiarthroplasty) has been shown to be clearly superior to non operative fixation [21] [22] [23] [24]. Despite the trend of increased surgery for patients with displaced fractures of the proximal humerus the evidence available would seem to suggest that conservative management is as effective as some of the more traditional surgical treatments for 2 and some 3 part proximal humerus fractures [25].

Currently there is limited level 1 evidence regarding the optimal treatment of proximal humerus fractures with the most well know study being the PROFHER Trial [25]. As a result of this trial which looked at conservative versus surgical management of proximal humerus fractures in 250 patients there was a general move towards a more conservative approach for proximal humerus fractures [26]. However, it should be noted, that a reverse total shoulder replacement is only indicated in the most complicated of proximal humerus fractures. The PROPHER Trial included all patients with 2, 3 and 4 pat proximal humerus fractures [25]. Therefore although the general trend is now moving towards a more conservative approach the reverse total shoulder replacement is still considered a viable option for operative replacement.

Studies comparing reverse total shoulder replacement to hemiarthroplasty for the treatment of proximal humeral fractures have shown improved pain scores and functional outcomes [27] [28]. This would suggest that a reverse total shoulder replacement is a superior form of surgical treatment compared to the more traditional methods of treating these fractures such as hemiarthroplasty. If we consider that a hemiarthroplasty is not superior to conservative management for proximal humerus fractures in terms of outcomes and a reverse total shoulder replacement is superior to a hemiarthroplasty in terms of pain scores and functional outcomes then, it seems logical, that a reverse total shoulder replacement is superior to conservative management for complex proximal humerus fractures.

The shoulder is a complex joint that relies on many opposing forces and tensions to provide a large range of motion [29]. As previously stated there are a number of deforming forces concentrated on a relatively small surface area [14].

The design of the reverse total shoulder arthroplasty medialises the center of rotation and brings the humerus to a more inferior position. This allows tension to be restored in the deltoid muscle belly and the functional lever arm which in turn improves the ability of the deltoid to move the humerus in space [29]. It has been reported that reversed total shoulder replacement provided excellent symptomatic relief and a pain free functional range of movement [30]. This is in stark contrast to some of the earlier studies on reverse total shoulder replacements which was reported to increase patient pain and reduce functional mobility [31]. New prosthetic design characteristics and surgical technique improvements are potentially factors that have dramatically improved patient outcomes.

Recent evidence would suggest that a reverse shoulder arthroplasty would be the preferred treatment method for surgeons in complex proximal humerus fractures [30]. However, to our knowledge, there is no single centre case series in Ireland that focuses on patient outcomes and quality of life following a reverse shoulder arthroplasty in the setting of trauma. Data from the United States suggests that we will see an exponential increase in proximal humerus fractures by the year 2030 and with this the increase in the demand for shoulder arthroplasties is inevitable [2]. This case series aims to build the evidence for a reverse shoulder arthroplasty as a treatment option in the setting of trauma for complex proximal humerus fractures.

As previously stated a reverse total shoulder replacement is a superior form of surgical treatment compared to the more traditional methods of surgical replacement such as hemiarthroplasty in terms of pain and functional outcome [27] [28]. In order to assess these variables it is important to utilise numerous scoring systems so a general trend between patients can be observed. The ones choosen for this study focus on patient related outcome measures nameslyrhe DASH and the OSS.

The DASH outcome measure is a 30-item, self-report questionnaire designed to measure physical function and symptoms in patients with musculoskeletal disorders of the upper limbs [32]. The DASH score is calculated by scoring patients on their ability to complete everyday tasks such as opening a jar or changing an overhead light bulb. The DASH score was shown to be reliable, valid, and responsive for patients with shoulder disabilities [33] [34] [35]. A higher DASH score (total 100 points) indicates a greater disability. Hunsaker (2002) [36] reported that the general population would score 10.1 on the DASH with a standard deviation of 14.68. This indicates that even in the uninjured shoulder the DASH accounts for reduced function that may be present from conditions such as osteoarthristis.

The Oxford Shoulder Score is also a personal subjective assessment of shoulder pain and function assessing everyday activities such as ability to dress and pain in bed at night. The OSS has been described as a reliable, easy to use outcome measure which is sensitive to surgical outcome and clinical changes [37] [38]. In total it is scored out of 48 with a higher score indicating less pain and better shoulder function (see Table 1).

As the OSS and DASH are patient self-assessment measurement outcomes they could be expected to be interrelated to each other but may sometimes be used combined. Studies have reported that OSS is better to assess the surgical outcome in patients with rheumatoid arthritis as compared to the DASH [39] [40]. Neither has been shown to be superior to the other for trauma patients but both were used in this study as they focus on patient related outcomes measures and should have, in theory, a good correlation.

The aim of this retrospective case series is to evaluate the clinical outcome and the impact on quality of life of the reverse shoulder arthroplasty in the setting of trauma in a single centre case series in Ireland.

Extrapolation of theatre records to identify patients who have undergone a reverse total shoulder replacement in the setting of trauma between January 2014 and December 2017 in Tullamore Regional Hospital (n = 12). 2 patients were not willing to participate. While 2 further patients were lost to follow-up. In total 8 patients were included in the case series.

There were 8 patients included in the case series; 1 male and 7 females (see Table 2). The average age of the participants was 73.3 years +/− 4.9 years. The surgical site was evenly distributed between left and right. The indication for surgery was a comminuted 3 or 4 part proximal humeral head and neck fracture in all cases. All Fracture paterns were either a Neer classification 3 or 4. There were

three Hertel class 12’s, three Hertel class 9’s, one Hertel class 8 and one Hertel class 7. The average time to surgery from time of the fracture was 8.6 days +/− 6.2 days. There were 5 non smokers, 2 ex smokers and 1 current smoker in the study. The average time of data collection post op was 23 months +/− 10 months. The mechanism of injury was a fall from standing height in all cases.

All procedures were performed by 2 consultant surgeons who specialise in upper limb injuries (n = 2 and n = 6). The implants used were the Aequalis II Reversed Shoulder System^TM. The approach used was the delto-pectoral approach in all cases.

Participants were contacted by telephone and asked a series of questions relating to their quality of life. 2 separate scoring systems were used?The Disabilities of the Arm, Shoulder and Hand (DASH) Score and the Oxford Shoulder Score. These scores broadly categorised patients into good, average and poor function groups.

All patients who participated in the study were invited into outpatient’s clinic where flexion, extension, internal and external rotation of the affected limb were measured and compared with the uninjured limb. All measurements were ascertained by a consultant orthopaedic surgeon and 2 orthopaedic trainees using a standard goniometer. In total 6 measurements were taken of each limb.

Patient records were reviewed and any intra operative and post operative complications were recorded.

X-rays taken at 1 day, 2 weeks and 3 months post operatively were reviewed independently by 2 orthopaedic trainees and 1orthopaedic consultant and any complications noted.

All Statistical analysis was carried out using SPSS version 24. An Independent t-test was used to assess difference between groups for all shoulder movements. A linear regression analysis was used to determine the relationship between the DASH and OSS.

The average Disabilities of the Arm, Shoulder and Hand (DASH) Score was 27.5 +/− 24.4 while the average Oxford Shoulder score was 37.9 +/− 9.2 see Table 3. There was a high negative linear correlation between the 2 scoring systems r = −0.95.

Flexion, extension and abduction movements were similar in all patients between the injured and the uninjured limb with statistically significant differences noted in all movements (see Table 4 and Figure 3).

Flexion

There was a statistically significant difference between flexion of the injured limb (130.5˚ +/− 16.85˚) and the uninjured limb (161˚ +/− 6.55˚) p < 0.001.

Extension

There was a statistically significant difference between extension of the injured limb (51.25˚ +/− 4.27˚) and the uninjured limb (57.62˚ +/− 3.58˚) p = 0.002.

Abduction

There was a statistically significant difference between abduction of the injured limb (131.36˚ +/− 12.56˚) and the uninjured limb (144.13˚ +/− 5.06˚) p = 0.004.

Adduction and external rotation were similar for all patients between the uninjured and the injured limb with no statistically significant difference observed between the 2 groups. There was a significant difference between the uninjured and injured side for internal rotation (p < 0.001) with the injured limb having a much less degree of movement compared to the uninjured side. Of note internal rotation is the only movement which showed clinically significant differences between the 2 groups. See Table 5 and Figure 4.

Adduction

There was no statistically significant difference between adduction of the injured limb (43.37˚ +/− 2.92˚) and the uninjured limb (43.37˚ +/− 1.77˚) p =1.

Internal Rotation

There was a statistically significant difference between internal rotation of the injured limb (23.75˚ +/− 6.36˚) and the uninjured limb (83.75˚ +/− 4.33˚) p < 0.001.

External Rotation

There was no statistically significant difference between external rotation of the injured limb (41.75˚ +/− 4.86˚) and the uninjured limb (43.13˚ +/− 2.3˚) p = 0.528.

There were no reported Intraoperative or post operative complications. Of note one patient experienced altered sensation along the distribution of the radial nerve that was present before their operation. This altered sensation originated from the time of their fall and had not fully resolved at 1 year follow-up.

There were no radio-graphical complications recorded on plain films of the shoulder joints taken at 1 day, 2 weeks and 3 months post operatively.