Proximal humerus fracture_s: A review of current practice

Abstract

The management of proximal humeral fractures (PHF) remains controversial. Its incidence is increasing. Patients should be meticulously assessed clinically for co-morbidities and neuro-vascular injuries. Radiological investigation helps provide information on the fracture configuration and dislocations. Enhanced by 3-dimensional CT scanning, these further help in decision making and operative planning. PHF classifications have been demonstrated to have poor intra-observer and inter-observer reliability. Research has identified some radiographic predictive factors for humeral head ischaemia and likely failure of surgical fixation. The range of management options include non-operative treatment, operative fixation, intramedullary nailing and arthroplasty (hemiarthroplasty, reverse shoulder replacement).

The majority of PHFs are stable injuries and non-operative management is usually successful. Some degree of malunion is readily tolerated especially by elderly patients. Surgical management of significantly displaced, unstable proximal humerus fractures should aim to stabilise the fracture adequately and provide satisfactory function for the long term. Management of the greater tuberosity is pivotal for the eventual outcome. When fixation may appear to be compromised by poor bone quality, likely poor function, age related rotator cuff degeneration or likely humeral head ischaemia clinicians may opt for arthroplasty. Successful hemiarthroplasty outcomes are dependent on sufficient healing of the tuberosity and recovery of the rotator cuff integrity. Reverse shoulder replacement can predictably deliver good functional outcomes for the shoulder in elderly patients, where rotator cuff dysfunction is suspected or as a revision procedure following failure of other surgical interventions. As opposed to hemiarthroplasty, which has shown a downward trend, there has been an increasing trend towards the use of reverse shoulder replacement in proximal humeral fractures. The management of PHFs should be patient specific, fracture specific and meet the functional demands and needs of the individual patient. The surgeon's skill set and clinical experience also plays an important role in the options of management available.

1. Introduction

The incidence of proximal humerus fractures (PHF) is increasing worldwide. PHF typically presents in a bimodal distribution with younger patients associated with higher energy trauma and the elderly (over 65 years), associated with lower-energy mechanisms or fragility fractures.^1,2 There is also a gender discrepancy, with older female patients sustaining PHF two to three times more often than males.³ The association with Osteoporosis has led to an increase in its occurrence as well as complexity of the fracture configuration with increasing age.

The long-term sequelae from PHF can be significant; these can include mal union and non-union, avascular necrosis and traumatic arthritis,. They can precipitate lengthy and costly rehabilitation whether managed surgically or non-operatively, as well as lead to a long term decline in functional ability and persisting residual dysfunction^1,3,4

The treatment of PHF continues to be debated and indeed, there exists a significant geographical variation in the local management of PHF which is indicative of the relative paucity of evidence as to the optimal management for these injuries.^1,3,4 The vast majority of PHFs can be managed non-operatively however for those PHF that do mandate surgical intervention, there remains no consensus as to the optimum treatment strategies. The introduction of locking plates has broadened the possible indications for open reduction with internal fixation, whilst reverse shoulder arthroplasty (RSA) use is on the rise for elderly patients with displaced PHF.

We present an overview of present clinical practice in the management of PHF reinforced on the current literature.

2. Clinical evaluation

The full holistic assessment of a patient with PHF helps guide decision making and therefore an imperative step in formulating a structured management plan. Many elderly patients may suffer from multiple systemic comorbidities and the anaesthetic and surgical risks may outweigh the benefits of surgery. In addition, many elderly patients may opt for non-operative treatment. In this cohort, a focussed history including hand dominance, hobbies/pastimes and functional demands should be obtained. In the relatively younger population occupation, smoking, alcohol history and history of intravenous drug abuse may have a bearing on choice of management.

Open PHFs are uncommon and would require management as per standardised open fracture guidelines.⁵ A neurovascular examination of the distal aspect of the affected limb is the next priority.

Vascular injuries are not common but can be more expected in the setting of a fracture-dislocation of the shoulder. A large expanding haematoma, hypotension with no clear cause, blood tests demonstrating low haemoglobin, pulsatile active bleeding and concomitant nerve trunk or brachial plexus injury ought to raise the awareness of a potential vascular injury. Over 90% of cases reported occur in patients over the age of 50, likely due to the increased stiffness within the lining of vessel walls due to the build-up of atherosclerosis.⁶ If there is any doubt a CT angiogram may aid the diagnosis and a review by the vascular surgeons should follow.

A brachial plexus and axillary nerve examination, in particular, is essential but can prove a challenge following acute trauma, nevertheless assessment of the fingers, wrist, and elbow can still often be attempted. Nerve injuries are, again, more probable in the setting of a shoulder fracture-dislocation.⁷

3. Radiological evaluation

Initial plain radiographs of the glenohumeral joint include a true anteroposterior (AP), lateral, and a mandatory orthogonal view, as a proper axillary view may not always be possible. This should help with the initial determination of glenohumeral joint congruency and PHF pattern.

Computed tomography (CT) scans, with the increasing use of three-dimensional reconstructed (3D-CT) models, enable better visualisation and appreciation of the fracture configuration and facilitate the treatment plan. CT is recommended for complex fractures (with 3D reconstruction), particularly when fracture lines may not be clearly visualised on plain radiographs. Although not commonly used magnetic resonance imaging (MRI) may be valuable for evaluating the integrity of the rotator cuff in the setting of PHF. A prospective study which involved 30 patients demonstrated approximately 40% of the PHFs were in combination with rotator cuff tears.⁸ Fjalestad et al. performed MRI scan on 76 patients whom had PHF, they demonstrated 22 had established rotator cuff tears at the time of sustaining the injury, and 10 went on to develop tears later, at one year. Impaired functional outcome was associated with cuff tears occurring during acute injuries.⁹

Hertel et al.,¹⁰in their seminal paper, suggested that fractures demonstrating a shortened medial calcar (<8 mm), a clear break of the medial hinge, and where there is anatomical neck involvement are most at risk of developing ischaemia. Therefore, patients with any of these signs should be observed closely if being managed non-operatively or indeed should be for higher consideration of surgical fixation. Bone density is another predictor for the quality of surgical fixation possible and the likelihood of potential metal work complications.¹¹

4. Classification

An early classification of proximal humerus fractures was created in 1934 by Ernest Codman, Charles Neer expanded upon Codman's use of the four main anatomic segments in 1970 and it still continues to be commonly used^12,13 The fractures are classified as two part, three part or four part corresponding to the number of displaced parts. The Hertel's binary fracture description approach¹⁰ was based on the analysis of the planes of the fracture and not merely on the number of fragments. It encompasses a total of 12 different basic fracture patterns. Another classification is provided by AO.¹⁴ Studies which have assessed the utility of PHF classification systems show low intraobserver and interobserver reliability.¹⁵ The use of a classification for PHFs in guiding our decision making and aiding within research remains challenging.

5. Non-operative management

Approximately 85% of PHFs can be managed non-operatively.¹² Over the last decade, especially in more elderly patients, PHFs have been increasingly treated non-operatively¹⁴,¹⁶ (Fig. 1). The PROFHER study¹⁷ (a multicentre RCT in the UK, compared non-operative and operative treatment) reported no significant difference in the patient reported outcome measures (PROMs) between surgical and conservative treatment. It is noteworthy, that this pragmatic RCT showed no difference in the outcomes for those fractures where the treating surgeons were genuinely unsure of whether to treat the fracture operatively or non-operatively. However in other fracture configurations where the surgeon felt surgery would be advantageous, they were still treated operatively. Further systematic reviews¹⁸ have also concluded that, in elderly patients, there was no difference in the PROMs for displaced PHF, whether managed operatively or non-operatively.

Fig. 1.

85 years old gentleman treated non-operatively with a displaced 4 part fracture with reasonably good functional out come.

The limitations of the PROFHER study are that most fractures (>80%) were two part with or without greater tuberosity fractures and RSA procedures were not included in the operated group. Additionally only 4.4% of the operated group were Neer four part fractures in the PROFHER study compared to 21% reported in a meta-analysis, demonstrating a discrepancy of their incidence.¹⁸ Another limitation of the PROFHER trial was that different fixation options were all grouped together under the surgical arm and therefore making the generalisability of the results. Discrepancy was further demonstrated with a higher risk of non-union reported in the non-operative group than expected. In non-operative management a medially displaced greater tuberosity fragment significantly overlapping with the humeral articular surface could lead to adverse outcomes.¹⁹

Patients managed non-operatively usually have a short period of immobilisation and progressive supervised physiotherapy. Current evidence supports the early mobilisation, within two weeks, of PHFs, although there is no evidence to suggest its superiority or any demonstrable difference in outcomes.¹⁹ There remains a paucity of high-quality evidence to aid the understanding of optimal rehabilitation timing following PHFs and this would certainly help avoid the detrimental consequences of immobilising patients longer than required.²⁰

6. Operative management

Although most PHFs are managed non-operatively, there are some clear indications for operative management. These include fracture dislocations of the shoulder, open fractures, head splitting fractures, severe varus or valgus displacement of the humeral head and completely off-ended fractures especially if there is significant displacement between the humeral head and shaft. Other indications for surgical treatment can be debated. Before embarking on surgical fixation, one should consider the patients age, their functional level, the fracture personality, the patients' ability to comply with the post-operative rehabilitation, any associated injuries, the patients’ bone quality, the likely disruption to the humeral head vascularity and the experience of the treating surgeon. Most fixation devices known to orthopaedic surgeons have been used in the fixation of PHF, the advantages and disadvantages of which are summarised Table 1. The following are currently the most commonly used surgical techniques.

Table 1.

Summary of surgical management.

	Advantages	Disadvantages
Percutaneous Fixations - K Wires/Screws	-Least soft tissue disruption -Easier for future metal removal -Easier revision for future arthroplasty	-Not rigid fixation -Screws/wires can back out -Relies on established configuration of screws -Unsuitable for very displaced fractures & poor bone quality -May need longer immobilisation period
Intramedullary Nail	-Small incisions -Less soft tissue disruption -Less blood loss -Shorter operating time -Resists varus collapse -Able to mobilise early -Biologically and biomechanically appealing option	-Rotator cuff dysfunction -Proximal migration of the nail -Challenging in 4 part fractures -Can cause impingement -Iatrogenic tuberosity or head fractures intra-operatively
Locking Plate	-Suitable for 4 part fractures -Allows grafting techniques -Ability to deal with all fracture configurations and visually reduce as able	-Increased operating time and blood loss -Extensive Soft tissue dissection -Risk of varus collapse -Screw penetrations -Subacromial Impingement -Difficult revisions or future arthroplasty
Hemiarthroplasty	-Suitable for younger patients with preserved rotator cuff function and increased potential of tuberosity healing -Does not violate the uninjured glenoid	-Outcome is dependent on tuberosity healing -Increased risk of revision to Reverse TSR
Reverse Arthroplasty	-Suitable for unfixable fractures or in failed fixations and failed hemiarthroplasty -Offers predictable pain relief and functional outcome in the elderly -Offers good outcome irrespective of tuberosity healing	-Violates un-injured glenoid -Increased blood loss -Increased duration of operation -Increased cost of the procedure

6.1. Percutaneous fixation OF PHF

Although the indications for this technique are limited they can be used in a range of fracture patterns and even in the setting of a fracture dislocation²¹ (Fig. 2). Percutaneous screw fixation is best used for valgus impacted, three- or four-part fractures and fractures where some improvement in fracture position could lead to a satisfactory position with minimal fixation. The main advantage of this percutaneous technique is minimal soft tissue envelope disturbance, reduced interference of the vascularity to the humeral head and, if required, provides an easier means for future metal work removal.²²

Fig. 2.

PHF Fracture dislocation with a head split in a 38years old patient treated with minimally invasive technique.

Roberts et al.²³ described a technique which relies on an established configuration of screw placement with good outcomes and low complication rates. They highlighted that preservation of the periosteal bridges and soft-tissue attachments of the fractures were important for successful outcomes. Resch et al. further showed low osteonecrosis rates and reduced soft tissue adhesions with this technique.²⁴ Johnson et al. reported low metal work complications and no screw penetrations with the percutaneous technique, unlike with the use of locking plates.²¹

6.2. Locking plate fixation

Locking plate fixation remains the most commonly used fixation modality for displaced PHFs. Locking screws offer stronger angular stability and higher resistance to failure over use of non-locking screws.²² It is vital that the humeral head is not fixed in a varus position and the medial calcar hinge is reduced. In order to achieve this one can use a inferomedial calcar screw, fibular allograft augmentation or bone cement augmentation.²⁵ Augmentation with a fibular strut allograft consistently improves the stiffness of the construct, as well as improving the load to failure and preventing fracture displacement. Given that locking plate use necessitates a degree of soft tissue dissection it can have significant complication rates (especially that of screw penetration). Locking plate fixation requires detailed surgical planning, meticulous surgical technique, soft tissue handling and imperative use of intraoperative imaging. Locking plates are mostly used in two- or three-part PHF in the younger cohort of patients. Although debatable, some studies suggest locking plate use in displaced four-part fractures, even with poor bone quality, demonstrating improved outcomes with low complication rates.²⁶ A Recent systematic review Kavuri et al.²⁷ including more than 3400 proximal humeral fractures treated with locking plates noted the complications of locking plates as follows: Intra-articular crew penetration (9.5%). Primary penetration is due to intra-operative placement. Secondary placement is usually due to the Varus collapse, AVN or failure of fixation. Other complications of locking plates include varus collapse (6.8%), subacromial impingement (5.0%), Avascular Necrosis (4.4%) and the re-operation rate was reported to be 13.8%.

6.3. Intramedullary nailing

Intramedullary (IM) nails are a biologically and biomechanically appealing option. Ideal indications for an IM nail are two-part PHF and selective three-part fractures (Fig. 3). The technique allows for the preservation of the periosteal blood supply and adjacent soft tissue anatomy. However, some of the risks include: Avascular necrosis (AVN) (4%), proximal screw migration, sub-acromial impingement and rotator cuff dysfunction.²⁸

Fig. 3.

IM nailing in a 67 years old lady with displaced 2/3 part fracture.

There have been particular concerns about rotator cuff dysfunction with IM Nails as older curvilinear shaped IM nails had higher complications and re-operation rates.²⁹ Newer generations of IM nails are straighter and offer multiple locking options. The entry point is typically at the apex of the humeral head articular surface, using a supraspinatus muscle split rather than tendon split approach, about 1.5 cm medial to the footprint of rotator cuff attachment and posterior to the biceps tendon thus avoiding the crucial watershed area of vascularity within the rotator cuff, and also any potential iatrogenic tuberosity fracture.³⁰

Muccioli et al. found the incidence of supraspinatus tendon lesions following IM nail to be no higher than its incidence in general population.³¹ Evolving implant designs and surgical techniques have allowed humeral IM nails to become one of the main fixation techniques in managing PHFs.

6.4. Clinical outcomes IM nail VS locking plate

Clinical and radiological outcomes are reported to be similar in PHFs treated with locking plates and locked IM Nails.³² For IM nails used in two and three part PHFs, the re-operation rate was found to be up to 19%, while with locking plates re-operation rates were up to 30%^31,33 However based on the current evidence, it remains a challenge to strongly recommend one implant over the other in the surgical treatment of two and three part PHFs.³⁴ Complications and re-operation rates have been found to be higher with both fixation in four part and head splitting fractures²⁸ but there remains a dearth of evidence in the literature to recommend the ideal method of fixation for four-part PHF.

7. Arthroplasty

Arthroplasty in the context of PHFs is either by way of shoulder hemiarthroplasty (HA) or reverse shoulder arthroplasty (RSA). Patient specific indications for arthroplasty are those patients who are elderly or those with poor bone quality. Fracture specific factors when arthroplasty should be considered are comminuted fractures, or in the setting of a fracture dislocation, head splitting fractures and fracture configurations which are not deemed amenable to an optimal fixation technique.

7.1. Hemiarthroplasty

The role of shoulder HA, whilst diminishing somewhat in its use, still serves as a viable management option in complex un-reconstructable PHF with no evidence of glenohumeral arthritis. It is indicated in the relatively younger patient cohort with three or four part PHFs not amenable to surgical fixation but does mandate a functioning rotator cuff in order to achieve biomechanical success.⁹ The key to the success of HA use is the fixation of the tuberosities and their healing, if the tuberosities fail to heal the outcomes are known to be very poor.³⁵ HA can provide excellent pain relief for patients, however predicting functional outcome is more difficult. The other fundamental factors that should be considered for the successful return of function are restoration of correct height, offset and humeral version.³⁶ When faced with a fracture pattern demonstrating a high degree of comminution of the tuberosity (not amenable to reconstruction) or a high suspicion of a very degenerate rotator cuff, RSA can provide a suitable alterative arthroplasty option.

7.2. Reverse shoulder arthroplasty (RSA)

RSA was primarily used as a salvage procedure in the management of failed HA, but in an increasingly aging population in whom rotator cuff dysfunction is common, and tuberosity fixation doubtful, the indications for RSA have broadened (Fig. 4). The existence of a premorbid rotator cuff tear, glenohumeral joint osteoarthritis and fractures in patients judged to have high risk of surgical fixation failure, form the bulk of the patients for whom RSA has progressively increased in its use. RSA continues to play a role in revision surgery demonstrating excellent outcomes in 79% of patients following failed surgical fixation.³⁷

Fig. 4.

3D-CT helps in surgical planning. This 75 years old man had a reverse shoulder arthroplasty for a 4 part displaced fracture.

Although important, it is not crucial to obtain accurate tuberosity reconstruction for the success of RSA compared with HA, however if well fixed, this has been reported to have improved external rotation.³⁸ However, few authors have reported very good tuberosity healing rates (>80%) nor with improved range of movement, especially that of external rotation.³⁹

A recent meta-analysis⁴⁰ recommended RSA in older aged patients, with three- or four-part fractures as this procedure returned the best patient reported outcome measure (PROMs) ranking with respect to Constant score and re-operation rate. In addition, a few studies⁴¹ have reported predictable improvement in pain and improved functional outcome with RSA when used in elderly patients suffering PHFs. This is likely the reasoning for the increasing use of RSA as the operative treatment of choice for many departments in those more elderly patients whom have displaced three- or four-part PHFs.

7.2.1. RSA vs non-operative management

Lopiz et al.,⁴² conducted a RCT to compare RSA with non-operative management of 59 patients, 80 years and older, with three- and four-part fractures and found minimal benefits with RSA over non-operative management. There was some improvement in pain amongst the RSA cohort over the non-operative cohort. Nonetheless, the study showed conservative management still provided adequate pain control and provides a viable option in more elderly patients or in those whom surgery may be too risky.

8. RSA VS ORIF

In more elderly patients with complex PHF, arthroplasty has several advantages. It can provide a favourable functional outcome with avoidance of further surgery in the future, as would be the case in failed surgical fixation or in the context of AVN. Fraser et al.⁴³ reported improved Constant scores in the RSA group at 2 years follow-up, in their study comparing RSA with operative fixation in elderly patients suffering displaced PHFs.

9. RSA VS HA

A meta-analysis by Shukla et al.⁴⁴ reports improved clinical outcomes with RSA compared to HA for similar indications. A further study⁴⁵ compared HA and RSA in 62 patients 70 years and older, with complex PHFs with a mean follow-up of 28.5 months. The study reports greater DASH scores amongst the HA group but with mean UCLA and Constant scores better in the RSA group. The functional outcome was still satisfactory in the RSA group irrespective of tuberosity healing, while patients with failed tuberosity healing had significantly worse functional outcomes in the HA group. The meta-analysis concluded, that primary RSA overall delivers improved pain relief and function with low revision rate in these fractures. However, conversion from HA to RSA as a salvage procedure did not improve the outcomes.

Whilst the evidence seems to support better functional outcomes following RSA a systematic review⁴⁶ in 2013 reported similar functional outcomes between RSA and HA but with a four times increase in complication rate following RSA (OR 4.0 95% CI 1.9 to 8.5). Another recent meta-analysis⁴¹ found better function in RSA cohorts with improved abduction and external rotation even when the tuberosities were not reattached or failed to heal. The analysis did find equal re-operation rates among both groups but with an increase in the complications in the RSA cohort, these included notching, heterotopic ossification-radiologically and infection, neurologic injury, wound issues, acromion fracture and subluxation/dislocation-clinically.

A recent meta-analysis by Han et al.⁴⁷ evaluated HA and RSA and reported improved post-operative ROM and functional scores in the RSA group, but no real variation in the complication rates. Another analogous study⁴⁸ found statistically significant improvement in pain (mean difference 2.21 95% CI (−0.17,4.58)), PROMS (mean difference 3.47 95% CI (2.25, 4.7)) and range of motion (mean difference 7 95% CI (5.5, 8.51)) in the RSA group and a higher risk of revision surgery in the HA cohort. The conclusion being that RSA appears to show superior short- and medium-term outcomes in comparison to HA.

It is recognised that HA does have a role to play in the management of PHF, that is in the relatively younger patient cohort with preserved rotator cuff function and improved tuberosity healing potential. In view of the current literature however, there is increasing evidence that RSA is the preferred option in elderly patients and indeed, increasing in frequency in younger patients alike, primarily in those patients with PHF deemed unreconstructable.

10. RSA VS ORIF and HA

Rajaee et al.⁴⁹ compared patients that had undergone ORIF, HA and RSA for PHFs. They analysed cost, length of stay, transfusion rates, and patient disposition from 2011 to 2013. They reported an almost 2-fold increase in the rate of RSA being performed, while the other procedures saw a decline. RSA accounted for 32.2% of the procedures performed for PHFs in 2011, and this rose to 53.3% in 2013. The cost involved in RSA was higher compared to the other groups but a greater number of patients were able to be discharged directly home rather than an interim rehabilitation facility. Another study⁵⁰ compared the patient and fracture characteristics and their outcomes with the three different operations. They found that the older patients with high American Society of Anesthesiologists (ASA) scores were treated with arthroplasty while younger, lower ASA patients had ORIF. The range of motion was comparable among all the groups at 10 year follow-up, but the number of re-operations were significantly higher in ORIF and HA groups.

11. Primary VS salvage RSA

Some studies⁵¹ have raised concerns with RSA use for malunion or revision surgery, suggesting a lower survival rate when compared with RSA used for acute fracture management. In a recent meta-analysis,⁵² RSA performed for acute fracture management resulted in a significantly better ROM, lower complications and better PROMs as compared with RSA used as a revision procedure. However, with the development and increasing use of modular platform systems, more versatility is afforded. Furthermore modular systems can allow for a well-positioned and well-fixed humeral stem to be retained if revision arthroplasty is required forgoing the need for violating of the humerus and lengthy operating times thereby reducing the complication rates of revision arthroplasty.⁵³

Overall, there currently exists no clear consensus to assist decision making when deciding between HA and RSA in management of these complex fracture patterns. The general considerations are the age of patient and the integrity of the rotator cuff which will tend to direct a clinician towards a preferred arthroplasty procedure. HA functions well in the relatively younger population in the presence of a reliably, well-functioning rotator cuff. In the more elderly population, where rotator cuff integrity is less reliable, RSA has been shown to provide much more satisfactory functional outcomes. There is, however, a demand for more robust evidence with long term follow-up to establish a coherent agreement on what constitutes the optimal utilisation of these contrasting arthroplasty options. In view of this, a pragmatic multi-centre randomised controlled trial (PROFHER-2), which is underway, sets out to analyse the cost and the clinical effectiveness of RSA versus HA compared to non-surgical treatment and will hopefully go some way in providing some answers to address this.

12. Conclusion

Non-operative management remains an important and viable option in the management of some PHFs. In younger patients with significantly displaced PHF surgical fixation is preferable. Consideration of percutaneous fixation, IM nail or locking plate fixation are commonly used for displaced two- or three-part fracture patterns. Tuberosity management is crucial, and its correct healing is vital to improving the overall outcome.

In unreconstructable PHFs an arthroplasty procedure is the only surgical option. The indications for HA have reduced and currently it is mainly used in younger patients with a well-functioning rotator cuff. A platform system is recommended as it is helpful during the time of revision. RSA has become the mainstay in the management of unreconstructable PHF, especially in the elderly, as well as for salvage situations with reliable outcomes. RSA in fractures is technically demanding and a surgeon's experience is crucial in reducing complications and improving outcomes. The treatment of PHFs should be individualised to the patients and fracture characteristics.

Funding

No outside funding or grants were received for this study by any author.