Three-Part Fractures: Open Reduction and Internal Fixation, or Arthroplasty?

Maxwell C. Park

Walter G. Stanwood

Louis U. Bigliani

M. C. Park: Center for Shoulder, Elbow and Sports Medicine, New York Orthopaedic Hospital, College of Physicians and Surgeons, Columbia University, New York, New York.

W. G. Stanwood: Center for Shoulder, Elbow and Sports Medicine, New York Orthopaedic Hospital, College of Physicians and Surgeons, Columbia University, New York, New York.

L. U. Bigliani: Center for Shoulder, Elbow and Sports Medicine, New York Orthopaedic Hospital, College of Physicians and Surgeons, Columbia University, New York, New York.

INTRODUCTION

Before the advent of the Neer classification,⁵⁸ Dr. Neer described the treatment of surgical neck fracture-dislocations where “the head fragment remained extruded” from the glenohumeral joint.⁶¹ In this series, the most common treatment was excision of the humeral head. “Replacement prosthesis” was listed as a treatment option being developed at the time. Two years later, Dr. Neer reported on his first series of patients treated with hemiarthroplasty; he outlined the indications for hemiarthroplasty, including “fracture with dislocation of the head fragment.”⁶⁰ He wrote:

Upon exploring the shoulder if it is seen that a reasonable amount of the articular surface is intact and retains soft tissue attachments, [gentle] reduction of the fracture and preservation of the head is the procedure of choice. Rarely, the anatomic head is dislocated and deprived of soft tissue attachments, or is fragmented. In this instance reduction is followed by avascular necrosis and traumatic arthritis. Resection is indicated.

In these few sentences, Dr. Neer touched on the key considerations for treating complex proximal humerus fractures: (1) extent of injury, with dislocation suggesting more trauma; (2) bone quality, as it relates to the size of the fragments; (3) soft tissue and blood supply, as it relates to avascular necrosis (AVN) of the humeral head; and (4) gentle technique during open reduction, suggesting minimal dissection. These salient considerations will be discussed in this chapter as they relate in particular to three-part fractures of the proximal humerus.

Proximal humerus fractures account for 4% to 5% of all fractures and most commonly affect elderly patients with osteoporotic bone.¹,²²,³⁶,⁷¹ Eighty-five percent of proximal humerus fractures are nondisplaced or minimally displaced⁵⁶,⁵⁸ and are adequately treated with early, protected mobilization. Therefore, the incidence of displaced proximal humerus fractures, particularly three-part
patterns requiring surgical intervention, is relatively low.³⁰,⁴¹ Three- and four-part fractures represent 13% to 16% of all proximal humeral fractures.³⁶ When encountered, they can be difficult to treat, even in the most experienced hands.

PATHOLOGY

Risk Factors

Elderly patients are at high risk for proximal humerus fractures,⁵⁰ and females are twice as likely as males to sustain these injuries.⁷¹ Osteoporosis, as it relates to elderly women, is a major consideration in understanding these statistics. Osteoporosis, along with humeral head viability, is one of the most critical considerations when scrutinizing the ultimate treatment decision for complex proximal humerus fractures.

Mechanisms of Injury

The most common mechanism of injury is a fall onto an outstretched arm from a standing height or less⁵⁰,⁶⁴,⁸⁶ or a direct blow to the shoulder itself.⁴ This relatively low-energy trauma is most often associated with the elderly population with osteopenic bone. Higher energy trauma, such as a motor vehicle accident, accounts for more of these fractures in the younger and male populations and presents a greater likelihood of a concomitant dislocation.⁴,²⁹,⁶⁸ Polytrauma is more common in the younger population.⁷¹ Another mechanism of injury, as first described by Codman, involves extreme rotation of an abducted arm, which can cause fracture as the humeral head impinges against the acromion.¹¹ Less commonly, convulsions can cause fracture-dislocations⁴,⁶⁸; the dislocations can be anterior⁷⁴ or posterior.⁵,¹⁸,⁶⁷,⁷⁹ When the mechanism is trivial and a proximal humerus fracture is sustained, consider a pathological etiology.

Concomitant Injuries

Stableforth reported an incidence of 6.1% for brachial plexus injury in proximal humerus fractures.⁸⁸ The brachial plexus is particularly susceptible with fracture dislocations; inspection and palpation should reveal loss of normal shoulder contour. Any displacement at the surgical neck can injure the axillary nerve as it passes underneath the glenohumeral joint; the axillary nerve is the most commonly injured nerve in the setting of proximal humerus fractures.⁴,⁶ When encountering a proximal humerus fracture, the surgeon must palpate the deltoid for voluntary contractions to assess axillary nerve injury.

Given its close proximity to the shoulder joint, the axillary artery is the most commonly injured vascular structure with proximal humerus fractures.⁴ Vascular injuries occur predominantly in the third part of the axillary artery, where it is tethered to the humerus by the anterior and posterior humeral circumflex arteries.⁴,⁸⁸ When vascular damage is present, it is often associated with severe medial shaft displacement through the surgical neck.⁹⁶ A cool extremity with asymmetrical brachial or radial pulses raises suspicion for significant arterial injury requiring emergent intervention.

“Transitory Subluxation”

Occasionally, inferior subluxation occurs after proximal humerus fractures and results from a global atony of the deltoid muscle.⁴ This is generally self-limited and has not been shown to affect outcome.⁵⁹ Support the arm in a sling, and promote recovery through gentle isometric exercises.⁴ If subluxation persists after 4 weeks, then rule out an axillary nerve palsy.

RISK OF AVASCULAR NECROSIS

Avascular necrosis rates of the humeral head are 12% to 25% in three-part fractures.²⁴ The anterior circumflex artery helps supply the anterior rotator cuff, and the posterior circumflex and suprascapular anastomoses supply the posterior cuff.⁷² The anterolateral branch of the anterior circumflex humeral artery provides the primary blood supply to the humeral head.²³,⁴⁷ This artery reaches the bicipital groove at the surgical neck and may be compromised at this level in surgical neck fractures. The anterolateral branch is conceivably more at risk with higher energy trauma (e.g., three-part fracture worse than two-part fracture, dislocation worse than without dislocation). Schai et al. stated that the blood supply of the humeral head comes mainly through muscular insertions at the tuberosities and through vessels entering the bone distal to the anatomical neck, and that the major blood supply of the humeral head is maintained through the intact lesser tuberosity.⁷⁶ The anterolateral branch, before entry into the humeral head (arcuate artery), runs along the lateral aspect of the bicipital groove.²³ Seggl and Wieglein⁷⁸ and Brooks et al.⁸ noted the importance of numerous branches of the anterior and posterior humeral circumflex arteries supplying the humeral head via soft tissue traversing the medial anatomical neck. In a case report by Gerber et al., AVN of the humeral head was absent after post-traumatic rupture of the anterior and posterior humeral circumflex arteries, suggesting the clinical relevance of the intraosseous blood supply by anastomoses of the deep brachial artery.²⁵ Appreciating the significance of vascular anatomy perfusing the humeral head can improve surgical techniques and thus may improve outcomes for the treatment of three-part fractures.

IMAGING AND SPECIAL TESTS

Accurate Radiographs

Radiographic trauma series should be obtained for all patients with proximal humerus fractures.⁵⁸,⁶⁴ This series includes three views: (1) anteroposterior, (2) scapular-lateral, and (3) axillary. When arm abduction necessary for axillary radiographs cannot be tolerated, obtain a Velpeau axillary radiograph.⁷ The trauma series provides the ability to assess fracture displacement and angulation in three planes. Park et al. highlighted the inherent difficulty with obtaining adequate radiographs for making this assessment.⁶⁴ In particular, with respect to the surgical neck fracture, the apex of the humeral head is often not tangent to the radiograph and is either internally or externally oriented depending on the radiograph plate and beam, thus misrepresenting the true head-shaft angle.⁶⁴ However, the trauma series remains the most practical diagnostic tool available to the shoulder surgeon, despite studies questioning the reliability and reproducibility of classifying proximal humerus fractures with radiographs.⁸⁰,⁸¹

Computed Tomography (CT) Scans

Computed tomography (CT) scans can aid in assessing fracture lines involving the articular surface, which would have implications for operative management. Some authors find CT indispensable when the number of fragments cannot be reliably defined.⁷⁶ Three-dimensional reconstructions provide additional information with regard to capital impaction, fracture displacement, and dislocation.⁴³ A head-splitting fracture or large (greater than 40%) humeral head impression fracture are patterns adverse to operative reduction.¹² CT scans, in general, can be used as an adjunct to a proper radiographic trauma series to better appreciate the extent of comminution. When indicated, CT scans with 2-mm cuts should be obtained. However, addition of CT scans to the workup offers little value in improving reproducibility of the Neer classification.²,⁸³,⁸⁴

Magnetic Resonance Imaging (MRI)

Magnetic resonance imaging (MRI) is not routinely used to assess proximal humerus fractures, nor has its significance in this setting been proven in the literature. However, MRI can document the status of the rotator cuff. Codman realized the importance of the rotator cuff tendons in keeping proximal humerus fractures in “mutual apposition.”¹¹ He emphasized “how tenaciously the short rotators with their periosteal prolongations cling to all the fragments and tend to hold them together.” The rotator cuff can be used as a fixation point in open-reduction and internal fixation (ORIF).⁶⁴ In addition, cuff disruption may predispose the humeral head to AVN, if no other remaining perfusion is present after a severe surgical neck fracture. If available to the shoulder surgeon, MRI may facilitate preoperative planning, insofar as rotator cuff status can be reliably assessed. However, in the majority of cases, an MRI is usually not necessary.

Other Studies

As discussed previously, concomitant brachial plexus and axillary artery injuries in the setting of proximal humerus fractures should always be clinically assessed. Angiography is indicated when a cold extremity with asymmetrical pulses presents in the setting of a severely displaced surgical neck fracture, raising the suspicion for arterial compromise. When a brachial plexus injury is present, electromyography (EMG) is appropriate approximately 3 weeks after the time of injury; if persistent and not self-limited, signs of injury are best characterized with EMG during this subacute period.

NONOPERATIVE TREATMENT

Indications

In theory a three-part fracture without soft-tissue interposition might be reduced and treated nonoperatively; in most cases, however, the patho-anatomy poses a challenge to reduction and results in an unstable fracture. The muscles attached to the fragment determine the displacement patterns. For the three-part fracture with greater tuberosity (GT) displacement, the subscapularis internally rotates the humeral head; for the three-part lesser tuberosity fracture (LT), the supraspinatus externally rotates the head. The pectoralis major internally rotates and medializes the shaft. The biceps tendon may be interposed between the fracture fragments. Repeated attempts at closed reduction are not recommended, especially in the elderly population, because the bone quality may be poor and further comminution may occur.⁴ Also, repeated manipulation has been associated with neurological deficit and heterotopic ossification.⁵⁹

If a patient cannot tolerate anesthesia due to medically related comorbid conditions, then operative treatment is precluded and closed management must be pursued. In addition, if a patient cannot participate in postoperative physical therapy, operative management may be unwise. If a pathological fracture secondary to tumor or metastases is present, nonoperative management may be unreliable, and surgical repair may be advantageous.

Results in the Literature

Numerous studies reported acceptable results for the conservative treatment of three- and four-part fractures in elderly patients.³⁸,⁴⁸,⁹⁹, ¹⁰⁰, ¹⁰¹ and ¹⁰² Zyto reported on nine elderly
patients with three-part fractures available for retrospective review at 10 years, the longest follow-up in the literature for this fracture pattern.¹⁰⁰ The mean Constant score¹³ was 59, and disability was reported as none in four patients, mild in three, and moderate in two; no patients had severe disability. All patients could accept their shoulder condition. However, death occurred in 66% of their original cohort (average age 66 years), which consisted of 58 patients. Zyto concluded that nonoperative treatment of displaced three-part fractures in elderly patients should be considered as an acceptable treatment option. In a retrospective review, Ilchmann et al. showed that seven patients (average age 73 years) with three-part fractures treated conservatively had better scores than eight patients (average age 62 years) treated with tension-band osteosynthesis; they used their own clinical outcomes measure which graded pain, function, and motion.³⁸ The conservatively treated group had better average scores for pain and function. They reported one patient with AVN in the conservative group and three in the tension-band group. They concluded that satisfactory results can be achieved with closed treatment of three-part fractures.

Not only have the results shown that closed treatment is acceptable, but several recent studies have proven that the operative treatment is not any better.⁷⁷,¹⁰¹, ¹⁰² and ¹⁰³ Zyto et al. compared 18 patients with three-part fractures treated nonoperatively to 19 patients treated with tension-band constructs.¹⁰¹ Overall, the average age was 74 years. At least 30% contact between the head and shaft were required, and no manipulation was performed. There were no significant differences between conservatively and surgically treated patients after more than 3 years follow-up, and there were no differences in function. Another study by Zyto et al. found no differences between their three-part fractures treated conservatively and surgically with internal fixation; at 3 years follow-up, 96% of their patients, with a mean age of 73 years, accepted their shoulder situation.¹⁰² In two separate studies, Schai et al. showed that patients treated conservatively had similar mean Constant scores¹³ to those treated with ORIF⁷⁶,⁷⁷; in the more recent multicenter study, the patients with three-part fractures treated conservatively and with osteosynthesis had similar Constant scores that were not significantly different—78 and 83, respectively.

Although the previous studies suggest that nonoperative treatment for three-part fractures is acceptable,³⁸,⁴⁸,⁹⁹, ¹⁰⁰, ¹⁰¹, ¹⁰² and ¹⁰³ many studies have reported unsatisfactory results with conservative management associated with pain, malunion, and AVN.²⁸,⁵⁹,⁶¹,⁶⁶,⁸⁹ Rasmussen et al. reviewed 42 patients with displaced two-, three-, and four-part proximal humerus fractures treated conservatively.⁶⁶ The median age was 77 years and the median follow-up was 2 years. There was a statistically significant difference between those with two-part fractures and the rest of the cohort, with the former group doing better. There were 17 three-part fractures, and only 7 patients (41%) had excellent or satisfactory results.⁶⁶ In 1970, Neer reported on 39 three-part fractures in patients with an average age of 55.3 years.⁵⁹ These patients, after closed reduction under anesthesia, were treated with a Velpeau bandage, a hanging cast, or overhead ulnar-pin traction. Only one patient had an excellent result, and two had satisfactory results. Failures were attributed to malunion, nonunion, and AVN of the humeral head. Other complications included neurologic deficit and heterotopic ossification, both thought to have been exacerbated by repeated attempts at reduction. Neer concluded that nonoperative treatment was inadequate and advocated ORIF for displaced three-part fractures.

In general, unless a patient cannot tolerate anesthesia, the authors recommend joint preservation via operative repair for three-part fractures. If severe comminution or osteoporosis exists and limits fixation, consider hemiarthroplasty. In most cases, humeral head replacement in the setting of three-part fractures provides opportunity for shoulder salvage, if ORIF is not possible.

OPEN REDUCTION AND INTERNAL FIXATION

Advantages

The advantages of ORIF in the properly selected patient include joint preservation and the opportunity for salvage via hemiarthroplasty, should the need arise. Other authors raise the concern over prosthesis longevity and favor ORIF for complex fractures of the proximal humerus.²⁴

The relative benefit of ORIF over hemiarthroplasty is emphasized by recalling the potential multiple causes for failure after humeral head replacement. These include: (1) greater tuberosity detachment, (2) prosthetic loosening, (3) nerve injury, (4) glenoid erosion, (5) prosthesis malposition, (6) dislocation, (7) heterotopic bone formation, and (8) deep infection.¹² In addition, multiple studies³¹,⁴⁴,⁵⁷,⁹⁷,¹⁰³ investigating hemiarthroplasty for complex proximal humerus fractures have had less satisfactory outcomes when compared to Neer.⁵⁹ In a recent study, Zyto et al. reviewed 27 patients who had sustained displaced three-or four-part fractures of the proximal humerus and were treated with hemiarthroplasty.¹⁰³ Seventeen patients had three-part fractures, and ten had four-part fractures. The median age was 71 years. The mean follow-up period was 39 months. The median Constant score¹³ was 51 for the three-part fractures and 46 for the four-part fractures. The median range of motion for all the patients in flexion was 70 degrees; internal rotation, 50 degrees; and external rotation, 45 degrees. Nine patients still had moderate or severe pain. Eight patients had moderate or severe disability. They concluded that their results were disappointing and that further studies on open reduction and fixation were justified.

Only gold members can continue reading. Log In or Register to continue