Conservative Treatment of Proximal Humerus Fractures: When, How, and What to Expect

Abstract

Purpose of Review

The key question to answer during the decision-making process for proximal humerus fractures (PHF) is whether the amount of displacement of a specific fracture pattern will be acceptable taking into account the anticipated demands on the patient. The aim of this review article was to provide some clarity regarding the features that contribute to poor clinical outcomes when PHF are treated non-operatively and to review the reported outcomes of conservative treatment.

Recent Findings

Conservative treatment for non-displaced or minimally displaced fractures leads to good outcomes in 80% to 90% of patients. However, with increasing fracture complexity and displacement, functional outcomes tend to diminish. In active patients with significant functional demands, the challenge is to predict which fractures will do poorly when treated non-operatively.

Summary

A better understanding of fracture patterns and fragment displacement may improve treatment indications. To avoid complications related to conservative treatment, surgery should be considered (1) in fractures in which the humeral head is severely compromised (due to fracture-dislocation, severe impaction, or a split of the head itself), (2) in non-impacted fractures with gross instability between the humeral shaft and humeral head, and (3) in those cases in which displacement of the tuberosities or the final shape of the proximal humerus after healing will lead to symptomatic malunion.

Introduction

Deciding when to offer a given patient with a proximal humerus fracture (PHF) surgery or non-operative treatment can be a very straight-forward decision in the two ends of the spectrum of severity: for example, nondisplaced fractures are treated non-operatively, whereas an anterior dislocation of the fractured humeral head with both tuberosities fractured in a young patient requires surgery. The problem is that there are many fractures that are somewhat displaced, but not severely displaced. The answer to this question depends not only on fracture pattern and displacement, but also on patient factors, including age, comorbidities, and anticipated demands by the patient on the fractured shoulder. This review article intends to provide some clarity regarding the features that contribute to poor clinical outcomes when PHF are treated non-operatively and reviews the reported outcomes of conservative treatment.

When Should Conservative Treatment vs Surgery Be Considered?

Decision-making in proximal humeral fractures depends on fracture characteristics (pattern, displacement, complexity, stability) and patient-related factors (age, functional expectations, comorbidities).

Fracture Characteristics

The proximal humerus may fracture in many different ways but there are a few patterns more common than others [1, 2, 3, 4, 5, 6, 7•, 8••, 9, 10]. In each pattern of fracture, it is important to analyze displacement of the fragments, stability of the fracture, and expected outcomes.

Our preference is to use the Mayo-FJD classification [8••] that contemplates seven common fracture patterns which correlate with different outcomes when treated conservatively: isolated fracture of the greater or lesser tuberosity (GT, LT), fractures involving the surgical neck without head deformity (SN), fractures impacted in varus (VPM) or in valgus (VL), and fractures involving the humeral head where the head is dislocated (head dislocation, HD), split (head-splitting, HS), or depressed (head impaction, HI).

Isolated Fracture of the Tuberosities

The isolated fracture of the greater tuberosity (GT) can be the result of an avulsion in the setting of an isolated injury [Fig. 1a] or an anterior dislocation [Fig. 1b]. The isolated fracture of the lesser tuberosity (LT) may occur as an avulsion of the subscapularis through the bone, or in the setting of a posterior dislocation [Fig. 1c]. When fracture of the greater or lesser tuberosity occurs in the setting of shoulder dislocation (anterior or posterior, respectively), residual displacement should be judged after reduction of the dislocation. Nondisplaced tuberosity fractures can be treated conservatively.

Fig. 1.

Tuberosities fracture. a Minimally displaced GT. b Antero-posterior radiograph of anterior humeral head dislocation and associated GT fracture. c Axial view of CT scan of posterior humeral head dislocation and associated LT fracture

The main potential adverse outcome with non-operative treatment of these fractures when displaced is impingement of the tuberosities on the glenoid rim or subacromial space or dysfunction of the rotator cuff secondary to changes in length and line of pull [Fig. 2].

Fig. 2.

Greater tuberosity fracture. a Posteromedially displaced GT fracture (red arrow). b Supraespinatus tear retracted to the glenoid and absence of GT on the coronal view of MRI. c Axial view of MRI showing complete infraspinatus tear. d Sagital view of MIR showing edema of supraspinatus and infraspiantus. e Secondary glenohumeral osteoarthritis

Fractures of the Surgical Neck

In this fracture pattern, the fracture plane separates the head and both tuberosities from the diaphysis. The anatomic neck is preserved. Surgical neck fractures with contact and no gross instability can be managed satisfactorily with conservative treatment (Fig. 3a, b). Fracture instability is the main driving factor to consider surgery in surgical neck fractures (Fig. 3c, d). The shaft is typically displaced medially by the pectoralis major. When treated non-operatively, nonunion at the level of the surgical neck has been associated with displacement of the humeral shaft between 33 and 100% [4, 5]. When there is no head-shaft contact and the neck is grossly unstable, these fractures can be considered disengaged neck (DN) fractures. In this situation, there is a high risk of fracture nonunion that may handicap patients in their routine daily activities [5].

Fig. 3.

Surgical neck (SN) fracture. a Minimally displaced SN fracture with associated GT fracture. b SN + GT fracture consolidation. c–d SN fracture with anteriorly displaced humeral shaft. e SN fracture nonunion

Varus Posteromedial Fractures (VPM)

Varus posteromedial (VPM) fractures are characterized by comminution of the posteromedial neck-head junction. The humeral head articular cartilage faces posteriorly and inferiorly. These fractures are considered stable when the humeral shaft is impacted into the humeral head and there is no major displacement on the lateral side (Fig. 4a–d). However, the more comminution at the posteromedial level, the more unstable the fracture becomes (Fig. 4e–h). When there is complete loss of contact between the head and the shaft, this extreme displacement converts the fracture into a disengaged fracture and carries a high risk of nonunion.

Fig. 4.

Varus posteromedial (VPM)–impacted fractures. a–d Coronal view of CT scan of impacted VPM fracture with progressive displacement. e–h 3D reconstruction of VPM fracture with progressive displacement

Compared to other fracture patterns, varus fractures continue to settle during the initial treatment phase [6, 9]. However, varus deformity can be relatively well tolerated. Court-Brown et al. reported 80% of satisfactory results in these fractures independent of the fracture angulation (Fig. 5a, b), with only 20% of unsatisfactory results using Neer criteria [11] (less than 90 degrees of elevation, 20 degrees of ER, unsatisfied with the results and/or substantial pain). Nonetheless, VPM fractures present worse results when associated with a fracture of greater tuberosity or both tuberosities [7•]. Greater tuberosity fractures typically occur posterior to the bicipital groove and displace posterosuperiorly due to traction of the posterosuperior rotator cuff (Fig. 5c). Malunion of the GT (Fig. 5d) can lead to loss of external rotation strength.

Fig. 5.

VPM fracture and associated GT fracture. a VPM fracture. b Consolidated VPM fracture. c VPM fracture with associated GT fracture that occurs posterior to the bicipital groove and displaced posteromedially. d Malunion of the GT

Valgus-Impacted Fractures (VI)

Due to comminution on the lateral side of the humerus, in valgus fractures, the articular surface is facing superiorly or superolaterally. Court-Brown et al. [2] reported that valgus fractures can be impacted stable fractures (Fig. 6a) or non-impacted fractures with displaced humeral shaft (Fig. 6b–d).

Fig. 6.

Valgus-impacted fractures (VL). a 3D reconstruction of valgus-impacted fracture with intact hinge and associated GT fracture. b–c 3D reconstruction showing valgus-impacted fracture with disrupted medial hinge and medial displacement of the humeral shaft. d 3D reconstruction showing valgus fracture with completed disruption of the humeral head that faces laterally under the GT

Stable impacted fractures lead to satisfactory results after conservative treatment and early mobilization [12] (Fig. 7a, b). However, marked valgus displacement leads to a higher chance of disruption of the vascularity of the humeral head and a higher risk of avascular necrosis [7•] (Fig. 7c, d). Scores deteriorate with increasing displacement of the fracture, with worse results reported in the presence of displaced humeral shaft [2] (Fig. 7e).

Fig. 7.

Valgus-impacted fractures (VL). VL fracture + associated GT fracture minimally displaced (a) and satisfactory fracture healing (b). displaced VL + GT fracture (c) and posterior avascular necrosis (d). VL fracture with associated displaced shaft fracture (d)

Valgus fractures with associated tuberosity fracture experience tuberosity displacement due to a combination of the position of the humeral head and the pull of the rotator cuff muscles. The greater tuberosity may fracture posterior to the bicipital groove (Fig. 8a, b) and displace posteromedially by the infraspinatus and teres minor muscles. When fracture occurs through the bicipital groove (Fig. 8c, d), both tuberosities displace in the transverse plane, widening the gap created by the fracture plane; the GT is pulled posteromedially by the infraspinatus and teres minor muscles, while the LT is pulled anteromedially by the subscapularis muscle. In other cases, LT fractured with the bicipital groove and the anterior part of the GT (Fig. 8e, f) and displaced anteromedially. The main potential complications of non-operative treatment are avascular necrosis of the humeral head, malunion or nonunion of the tuberosities leading to impingement on the glenoid rim or subacromial space and weakness in external or internal rotation due to rotator cuff deficiency on their attached GT or LT, respectively.

Fig. 8.

Tuberosities fracture associated with VL fracture. 3D reconstruction of valgus fracture with tuberosities fracture occurring posterior to the bicipital groove (a–b), through the bicipital groove (c–d), and fracture of both tuberosities together (e–f)

Fractures Involving the Humeral Head: Fracture-Dislocation, Head-Splitting, and Head Impaction Fractures

In these fractures, the humeral head itself is either fractured or dislocated. Conservative treatment in these pattern groups has been associated with maltracking, instability, osteoarthritis, avascular necrosis, nonunion, and malunion. For these reasons, except for those patients in which surgical treatment is contraindicated, non-operative treatment is not recommended when the humeral head is severely compromised due to fracture-dislocation (Fig. 9a), severe impaction (Fig. 9b), or a division (split) of the head itself (Fig. 9c).

Fig. 9.

Fractures involving the humeral head. a Antero-posterior radiograph of anterior proximal humerus fracture-dislocation. b Sagittal view of magnetic resonance imaging of proximal humerus head impaction after dislocation. c 3D reconstruction of proximal humerus head split

Patients Factors

Patient’s age, anticipated demands, and comorbidities play a major role in the decision-making process regarding treatment for PHF [13]. One study reported worse functional outcomes for more complex and displaced fractures [14]. In addition, socioeconomic factors have also been reported to negatively affect the outcome of PHF when treated conservatively [15].

Older patients seem to adapt to shoulder functional limitations better. In addition, failure of internal fixation seems to also correlate with older age [16]. However, severely displaced proximal humerus fractures treated non-operatively can lead to loss of independence for those elderly patients who were able to live independently prior to the fracture [7•].

How? Non-operative Treatment Protocol

Type of Immobilization

Non-operative treatment of PHF usually involves a period of immobilization followed by physiotherapy. Immobilization provides support and pain relief during healing, while physiotherapy aims to restore the function and mobility of the injured arm. In our opinion, PHF should be immobilized in some external rotation. This is particularly important for fractures of the surgical neck and for any fracture that includes a greater tuberosity fragment. For fractures of the humeral neck, if the shoulder is immobilized with the arm resting on the patient’s belly, the fracture will heal with an internal rotation malunion, and patients may have difficulty regaining functional external rotation once their fracture is healed. Additionally, to counteract the pull of the pectoralis major on the proximal aspect of the diaphysis, consideration may be given to placing a small bump or pillow in the axilla. Regarding fractures involving the greater tuberosity, since the posterosuperior rotator cuff is under more tension with internal rotation, placing the shoulder in some external rotation will decrease displacement of the greater tuberosity.

Time of Immobilization

Shoulder immobilization is typically maintained for 3 to 4 weeks, depending on fracture stability and healing progression. Unfortunately, 1 month of complete shoulder immobilization can substantially impact the independence of elderly patients. Moreover, there is a risk of the shoulder becoming stiff and painful with substantial reduction of function. Thus, different strategies for early mobilization with the expectation of a faster recovery in conservative treatment have been used [17, 18].

Early mobilization in PHF has been shown to lead to a faster and better outcome, with less pain and no major redisplacement or other complications in stable fractures, especially those with impaction [18]. Hodgson et al. [18] studied the timing of physical therapy for two-parts PHF and found that at 16 weeks after injury, patients who started physical therapy within 1 week achieved greater function and reported less pain than those immobilized for a period of 3 weeks. Lefevre-Colau et al. [12] performed a randomized trial comparing 72 h with 3 weeks of immobilization for impacted PHF treated non-operatively. Early mobilization turned out to be more effective in restoring the function of the injured shoulder, as measured using Constant scores, than the conventional 3-week period of immobilization. However, the follow-up was limited to 6 months. Kristiansen et al. found that 1 week of immobilization resulted in less pain and better function at 3 months when compared with a 3-week immobilization period [18] in minimally displaced PHF. These differences disappeared at 12- and 24-month follow-ups. More recently, Torrens et al. [19•] in a prospective randomized controlled trial of 111 proximal humeral fractures treated non-operatively demonstrated similar functional results in those managed with a short immobilization period (1 week) compared to those immobilized during 3 weeks, independent of the fracture pattern. However, almost 82% of these fractures were relatively simple fractures.

Rehabilitation Protocol

In our rapid rehabilitation protocol, at 1 week, patients are allowed to perform activities of daily living for self-care and hygiene, and the immobilizer is removed when patients are resting as well as for physical therapy exercises [17]. Exercises are initiated as soon as pain allows, which for most patients occurs between 1 and 2 weeks. Codman pendulum exercises can be performed for passive range of motion exercises of the shoulder from the beginning. These should be performed four to six times per day. Passive forward elevation exercises are allowed from the third to fourth week and are best tolerated in the supine position (“the prayer”). As the patient better adapts to these exercises, they can be continued in the sitting or standing position. As bonne union progresses, active assisted range of motion exercises may be added at 6 weeks, with strengthening starting 3 months after the injury. Patients are encouraged to perform the exercise program at least 2 times each day for 10 to 15 min each time, with 10 to 15 repetitions for each exercise. At any evaluation point, home therapy exercise may be supplemented with sessions with a physical therapist if pain is worse than expected for the average patient or loss of recovery of motion is observed.

Aguado et al. [17] reported good clinical outcomes in one-, two-, three-, and four-parts PHF with a home-based self-exercise program although fractures involving greater tuberosity presented a risk of cranial tuberosity displacement. Carbone et al. [20] in a randomized controlled trial of 120 patients with stable impacted osteoporotic PHF showed no advantage with aggressive rehabilitation regimen compared to immediate conventional mobilization protocol.

Radiological Evaluation

Close follow-up is required to detect any possible secondary displacement that would potentially change the indication of treatment from non-operative to surgical. The older the patient and the greater the initial displacement, the more the chances of progression of displacement. Surgical neck fractures with potential for instability should be evaluated with radiographs on a weekly basis for the first 3 to 4 weeks, since they have the highest likelihood of displacement leading to nonunion. For non-displaced or minimally displaced PHF, serial radiological evaluation seems to be necessary only in those presenting with comminution, while three of four patients may not need subsequent radiological evaluation [21]. Our preference is to obtain radiographs at weeks 3, 6, and 12 after the injury.

What to Expect? Outcomes

Most of the studies on the outcome of PHF treated non-operatively have been reported using Neer’s classification as a frame of reference. In general, conservative treatment in non-displaced or minimally displaced impacted fractures leads to good outcomes in 80% to 90% of patients [3, 22]. Spross et al. [23] reported 97% of satisfactory results in 89 PHF treated conservatively according to an evidence-based algorithm. Fractures selected for conservative treatment included 1-part fractures with <5 mm of GT displacement (74%), 2-part valgus-/varus-impacted fractures (25%), and only one 3-part fracture (1%). Only three patients (3%) needed unplanned surgery after conservative treatment due to secondary displacement with pain.

However, more complex or displaced fractures are associated with worse outcomes after conservative treatment [7•, 9, 14, 24••]. Torrens et al. [14] prospectively analyzed functional and radiologic outcomes of 70 PHF in patients with a mean age of 72 years treated non-operatively. According to Neer’s classification, there were 20% one-part fractures, 21% two-part GT, 25% two-part SN, 14% 3-part GT, 9% four-part, and 9% two-part GT fracture-dislocation. One-part fractures achieved better outcomes in terms of pain, ADL, forward elevation, abduction, internal rotation, strength, and total Constant score (CS), with most patients pain-free and mean anterior elevation of almost 120°. When analyzing displaced fractures, four-part fractures obtained the worst functional results (CS 34) followed by three-part fractures (CS 55) and two-part fractures (CS 66). Mean active anterior elevation was 90° for 3-part fracture and between 60 and 90° for 4-part fractures with limited internal and external rotation in both groups. However, despite limited functional outcomes obtained in displaced PHF fractures, no significant differences were observed in subjective perception of quality of life between displaced and non-displaced fractures. They suggested that this was probably because the older the population selected, the less the range of motion influenced the patient’s quality of life. Similar outcomes were reported by Lopiz et al. [24••]. In a prospective study comparing outcomes in PHF between conservative treatment and reverse shoulder arthroplasty (RSA), in the group of 29 PHF treated non-operatively (15% 3-part and 83% 4-part), mean CS at 12 months was 55, mean DASH score was 29, and mean elevation ranged between 61 and 90° in 45% of cases. They reported 100% of malunion, 59% of avascular necrosis, and 3.4% of nonunion. Despite these limited functional results, 93% of the patients treated with conservative treatment would undergo the same treatment again. They emphasized that this population presented a high comorbidity index (Charlson Comorbidity Index of 6.1) with a mean age of 85 years, and that might influence results of perceived quality of life. This means that, in particularly elderly patients with limited functional expectations, non-operative treatment could be considered appropriate for any type of fracture because good pain relief is to be expected although with limited function. However, these data cannot be extrapolated to patients with significant functional demands [13].

The results of conservative treatment in PHF are influenced not only by the fracture displacement and patient age, but also by the fracture pattern. Foruria et al. emphasized the importance of fracture pattern in PHF when they are treated conservatively [7•]. After conservative treatment of 111 PHF (52% varus, 14% valgus, 17% isolated GT fracture), they reported higher rate of unsatisfactory results in valgus fractures followed by varus fractures. The same author observed that progression of the deformity was influenced by initial fracture displacement and patient age, especially in varus posteromedial impacted fractures while GT fractures remained stable in most cases [9]. Nonunion at the surgical level was observed in two patients with a healing rate of 98% and 6 patients developed avascular necrosis (higher rate in valgus fractures), but 5 of these patients remained with minimal symptoms.

In active patients, surgical treatment may be considered to avoid complications related to conservative treatment that could lead to poor functional outcomes. The main complications associated with non-operative treatment are nonunion, malunion, and avascular necrosis. Nonunion of the surgical neck rises with increasing translation of the humeral shaft [4, 5]. Nonunion of the GT may lead to retraction and atrophy of the external rotator muscles with subsequent loss of external rotation [25]. Malunion of the humeral head leads to changes in the orientation of the articular surface. In the case of varus fractures, this translates into less elevation with worse varus, but increased external rotation due to head retroversion. Regarding the tuberosities, malunion of the tuberosities may lead to impingement with the glenoid in external or internal rotation. Finally, avascular necrosis of the humeral head has been reported to range between 1 and 10% [26]. Several specific fracture situations have been associated with an increased risk of AVN such as head-splitting fractures, fracture-dislocations, medial hinge comminution, and very displaced valgus fractures [7•]. Regardless of the incidence and risk factors of AVN, patients with proximal humerus AVN often are minimally symptomatic. Gerber et al. [27] demonstrated that, in patients with AVN but with anatomic fracture healing, shoulder elevation averaged 125° and the Constant score was 65% of the contralateral shoulder, similar to hemiarthroplasty or complex open reduction and internal fixation. Conversely, in the setting of malunion [27, 28] or in the presence of glenoid erosion due to the prominence of surgical implants, AVN led to significantly worse outcomes [25, 26].

Summary

In conclusion, conservative treatment leads to satisfactory results in non-displaced PHF but several displaced fractures may be treated operatively according to patient age and functional expectations and the risk of nounion, malunion, and avascular necrosis. The key question to answer during the decision-making process for proximal humerus fractures is whether the amount of displacement of a specific fracture pattern will be acceptable taking into account the anticipated demands on the patient. In our opinion, in active patients with significant functional demands, to avoid complications related to conservative treatment, surgery should be considered (1) in fractures in which the humeral head is severely compromised (due to fracture-dislocation, severe impaction, or a split of the head itself) and (2) in non-impacted fractures with gross instability between the humeral shaft and humeral head (disengaged neck fractures). If there is no head involvement and the shaft and the head are close enough to heal, the last question to answer is if the final shape of the proximal humerus after healing going to allow a good function or will lead to symptomatic malunion. At this time, we consider the grade of displacement of the rest of fracture patterns (varus fractures, valgus fractures, surgical neck) and the displacement of associated tuberosities fractures.

Declarations

Conflict of Interest

The author declares that she has no competing interest.

Human and Animal Rights and Informed Consent

This article does not contain any studies with human or animal subjects performed by any of the authors.