Patients with shoulder pain or injuries not infrequently have concomitant neurologic conditions, and orthopedic surgeons caring for such patients must be aware of them. In addition, the practice of reconstructive shoulder surgery carries an inherent risk of iatrogenic injury to neighboring neurologic structures. Knowledge of the common nerve lesions about the shoulder allows surgeons to recognize these entities when they see them, and familiarity with the relevant neural anatomy will help surgeons avoid potential neural injuries when they operate. Surgeons must adopt a systematic approach to the challenges of evaluating and treating patients with nerve-related disorders about the shoulder region.
Patients with nerve injuries are often seen in situations involving significant trauma. Commonly, the patient is confused, incoherent, sedated, or even unconscious, and it can be difficult to perform a satisfactory neurologic examination before initiating surgical care. Nonetheless, a good neurologic examination should be attempted in the emergency department. If for any reason an adequate examination cannot be performed, this should be noted in the patient’s medical record. Specifically, if the function of a particular nerve cannot be properly assessed preoperatively, the record should include such information. Too often in the emergency setting, the patient receives only a cursory evaluation, and perhaps the most junior person on the orthopedic surgical team writes in the clinical record that the extremity was “neurovascularly intact.” These two words, if inaccurate, are sometimes the origin of unnecessary litigation. Such general terms should not be used when recording a patient’s examination; instead, individual muscle strength, sensory examination, and deep tendon reflexes should be carefully documented in the patient’s record.
A detailed neurologic evaluation of the upper extremity can be performed on a coherent patient in a relatively short time, even if the patient has a shoulder dislocation or proximal humeral fracture. It is often easiest to start at the hand and progress proximally during the examination. Radial, median, and ulnar nerve function can all be assessed by a thorough evaluation of the hand and wrist, which should take less than 1 minute. Elbow flexion and extension strength are relatively simple to determine. The examiner should be aware, however, that it is possible to flex the elbow strongly with the action of the brachioradialis without having any function of the biceps. Loss of motor or sensory function in the distal end of the extremity can help direct examination of the more proximal musculature. For example, loss of radial nerve function should make the examiner look closely at axillary nerve function because both nerves are derived from the posterior cord. Likewise, loss of median nerve function might also affect the musculocutaneous nerve if the lesion is in the lateral cord.
Progressing up the arm, the condition of the medial and lateral pectoral nerves can be assessed by individually testing the strength of each of the major portions of the pectoralis major. The deltoid and rotator cuff muscles are then examined. The deltoid can be assessed even in the case of a painful proximal humeral fracture or glenohumeral dislocation. With the arm at the side, the patient is instructed to push out or “elbow” the examiner’s hand (which is placed at the lateral aspect of the elbow, with the other hand over the deltoid region to feel for contracture). If the patient is in a great deal of pain and the examiner cannot adequately determine the condition of the axillary nerve, this information should be recorded in the clinical record. Do not assume that it “might be okay.”
Examining shoulder abduction is an important part of the examination both to record muscle strength and to visualize shoulder kinesis through the arc of motion. Two important points are relevant here. First, some patients are able to abduct the shoulder through a full arc of motion by using either just the supraspinatus or the deltoid in the face of complete paralysis of either of them. Ensuring muscle contraction is a critical element in this part of the examination. Second, visualizing and palpating the scapula are a necessary part of the examination, especially when there is possible dysfunction. For example, patients with winged scapulae due to weakness of the serratus anterior or trapezius may have difficulty abducting the arm fully without the scapula stabilized and may compensate with trick motions. The examiner should recognize the clinical appearances and know the techniques to examine winging of the scapula, particularly with respect to distinguishing serratus anterior, trapezius, and rhomboid muscle dysfunction. In addition, a useful test for serratus anterior function that can be applied even in a patient with a complete brachial plexus lesion is to stabilize the inferior pole of the scapula while the patient pushes the arm forward; patients unable to perform this test would not be able to perform the more standard push-off test with the arms extended against a wall.
The examiner needs to evaluate the patient for other conditions besides a neurologic etiology that can cause lack of movement. In particular, it should be considered whether the inability of a patient to externally rotate the arm or perform the lift-off test might represent a neurologic lesion affecting the infraspinatus or subscapularis, a rotator cuff tear, or both. A patient who is feigning paralysis in the upper limb for secondary gain issues cannot voluntarily stop the latissimus dorsi from contracting while coughing.
Examination of the shoulder should include examination well above the shoulder and even above the neck, as well as the distal portion of the limb. Proximal and distal lesions should always be considered when examining a patient with shoulder pain or weakness and establishing a differential diagnosis. Cervical radiculopathy is a common cause of pain in the shoulder accompanied by motor weakness and sensory loss in the upper extremity. In this situation flexion and extension of the cervical spine or Spurling’s maneuver might reproduce or exacerbate the patient’s symptoms. Upper motor neuron lesions can also result in shoulder weakness. In these cases the deep tendon reflexes may be hyperreflexic, pathologic reflexes may be present, and tone may be increased. The clinical examination should exclude referred pain as a possibility, because cardiac and other intrathoracic as well as intraabdominal complaints may manifest as shoulder pain.
Finally, examination of the shoulder should always be performed with both shoulders exposed. Visualization is the first component of a physical examination, but this step is often neglected because of time constraints or modesty issues. If the bare scapula is not examined, it is easy to miss atrophy of the spinatus muscles. Bilateral atrophy or weakness would certainly change the differential diagnosis and force the examiner to consider an underlying neurogenic or myopathic condition.
Musculocutaneous Nerve Injury
Musculocutaneous nerve injury is most commonly associated with severe brachial plexus trauma. Although the nerve can be injured in glenohumeral dislocation, it is unusual to diagnose such injury as an isolated neuropathy. When seen as an isolated nerve injury, it is most often associated with a form of penetrating trauma, open surgical reconstruction, or a direct blow to the chest (near the coracoid). Occasionally, musculocutaneous neuropathy can occur after strenuous physical activity, such as rowing.
The musculocutaneous nerve travels obliquely below the coracoid process and enters the coracobrachialis. The anatomy of this juncture has been investigated in several studies. Small branches of the nerve can be found inserting into the coracobrachialis as close as 1.7 cm below the coracoid. The main trunk of the musculocutaneous nerve enters the coracobrachialis approximately 5 cm from the coracoid and exits at 7 cm. The nerve then enters the biceps, typically more than 10 cm from the coracoid. The nerve is at risk during anterior shoulder procedures that result in significant retraction medially, and during medial surgical dissection. The Bristow procedure has been thought to be associated with injury to this nerve, although such injury is probably related less to the transfer and more to manipulation of the nerve.
It is important for the operating surgeon to recognize the protective value of maintaining the origins of the coracobrachialis and biceps. When these muscles are allowed to remain on the coracoid during surgery, they act as a tether to overzealous medial retraction. In fact, exposure of the posterior cord or the axillary nerve can lead to musculocutaneous nerve palsy as a result of retraction. Care must be taken to avoid excessive traction on the musculocutaneous nerve during dissection. As an alternative, detachment of the conjoined tendon allows excellent exposure as well as potentially decreased pressure in the entire brachial plexus. Actual detachment of these muscles from the coracoid and reattachment to the anterior glenoid (the Bristow procedure) relaxes the musculocutaneous nerve. Nonetheless, the surgical manipulation required can result in damage to the nerve.
The nerve can also be damaged during arthroscopic surgery, although such injury is very uncommon. Anterior portals that stray medial to the coracoid put the musculocutaneous nerve and other branches of the brachial plexus at potential risk for injury. Low anterior portals, such as the 5-o’clock portal, can bring instruments to within 10 mm of the nerve. Carofino and colleagues reported on iatrogenic nerve injuries during shoulder surgery. In their report out of 26 patients with iatrogenic nerve injuries, 4 had undergone arthroscopic surgery and 10 had undergone combined arthroscopic and open shoulder surgery.
A patient with a musculocutaneous nerve lesion typically has a mixed sensory and motor lesion. Less commonly, a pure sensory lesion of the lateral antebrachial cutaneous nerve, the distal sensory termination of the musculocutaneous nerve, can occur. Lesions of the lateral antebrachial cutaneous nerve at the level of the elbow often have an atraumatic etiology and should be distinguished from more proximally occurring (incomplete) musculocutaneous lesions manifesting with sensory loss. These patients might have numbness or paresthesia along the lateral elbow crease that extends distally along the anterolateral aspect of the forearm. Treatment involves splinting or corticosteroid injection and, possibly, surgical exploration. Occasionally, surgical exploration reveals a thickened aponeurosis that is compressing the lateral antebrachial cutaneous nerve as it transits between the biceps and brachioradialis muscles. Patients might respond to surgical decompression of the nerve in this area.
Patients seen after trauma or surgery with an injury to the musculocutaneous nerve should be observed for 3 to 4 weeks. If no improvement in function is noted in that period, electromyography (EMG) and nerve conduction study (NCS) may be performed to assess the extent of nerve damage. Most postoperative musculocutaneous neuropathies are traction injuries that resolve over a period of weeks to months, depending on the extent of the injury.
If no improvement in biceps and brachialis function is seen on clinical examination or electrophysiologic studies, surgical exploration should be undertaken, ideally within 6 months from the injury. Surgical treatment options vary for persistent musculocutaneous neuropathy. If at surgical exploration the nerve appears intact but is compressed by scar tissue and demonstrates electrical conduction across the lesion, neurolysis may be indicated. If surgical exploration reveals a neuroma in continuity that does not conduct a nerve action potential (NAP) or reveals a rupture or transection of the nerve, additional treatment options should be considered.
In patients who have an isolated musculocutaneous nerve injury, a standard approach is to perform interpositional nerve grafting across the lesion. However, nerve transfers can also be used to shorten the distance (and time) for reinnervation or to bypass a scarred or avascular segment. A possible nerve transfer is the Oberlin transfer, a technique in which one or two fascicles of the ulnar nerve are transferred directly to the motor branch to the biceps. The distance to achieve reinnervation is extremely short because the site of repair is in the proximal part of the arm (several centimeters from the biceps end-organ) rather than a more lengthy repair from the neck or shoulder region. This technique can be used in patients with upper plexus lesions. In two large series each with more than 30 patients, grade 3 or 4 Medical Research Council (MRC) function was achieved in more than 90% of the patients treated with this technique. Reinnervation in the biceps was noted approximately 3 months after the procedure. Importantly, no patient suffered loss of distal ulnar nerve function or sensation. Recent modifications of this procedure have been reported with double reinnervation of elbow flexion: an ulnar nerve fascicle may be transferred to the biceps motor branch and a median nerve fascicle to the brachialis branch.
If patients are seen for more than 1 year after musculocutaneous nerve injury, nerve repair or reconstruction is significantly less likely to be effective. Nevertheless, nerve transfer techniques can be considered in select cases.
Some patients with little or no biceps or brachialis strength still function extremely well by using their brachioradialis alone for elbow flexion. The majority of patients however need augmentation. A number of tendon or muscle transfer procedures that achieve good results can be used. Popular options include Steindler flexorplasty (proximal advancement of the flexor-pronator muscle group) and triceps, pectoralis major, pectoralis minor, and latissimus dorsi transfer. Free muscle transfer is also a possibility in patients in whom these other potential donor muscles are not available.
The axillary nerve is one of the more commonly injured nerves about the shoulder. It is a terminal branch of the posterior cord and is derived from the fifth and sixth cranial nerves. The axillary nerve lies lateral to the radial nerve, posterior to the axillary artery, and anterior to the subscapularis muscle. It enters the quadrilateral space accompanied by the posterior humeral circumflex artery and is in close contact with the inferior shoulder capsule.
The axillary nerve is easy to locate at surgery during an anterior exposure by sweeping an index finger inferiorly over the anterior subscapularis and gently hooking the nerve while simultaneously palpating it on the underside of the deltoid with the other index finger. As it exits the space, the axillary nerve continues to the posterior aspect of the humeral neck and divides into anterior and posterior branches. The position of the anterior branch is commonly reported as lying 4 to 7 cm inferior to the anterolateral corner of the acromion. The posterior branch innervates both the teres minor and the posterior portion of the deltoid. The branch to the teres minor usually arises within or just distal to the quadrilateral space and enters the posteroinferior aspect of the teres minor muscle.
The internal topography of the axillary nerve has been studied by Aszmann and Dellon. As the nerve leaves the posterior cord, it is monofascicular, but as it enters the quadrilateral space, it has three distinct groups of fascicles: motor groups to the deltoid and teres minor and the sensory group of the superior lateral cutaneous nerve. The deltoid motor fascicles are found in a superolateral position; those of the teres minor and superior lateral cutaneous nerve are located inferomedially.
Etiology and Clinical Manifestation
Most axillary nerve injuries occur as part of a combined brachial plexus injury; isolated axillary nerve injury occurs in only 0.3% to 6% of brachial plexus injuries. Injury to the axillary nerve most often follows closed trauma involving traction on the shoulder. Axillary nerve paralysis is the most common neurologic complication of shoulder dislocations. Some patients with a proximal humeral fracture or shoulder dislocation have a subclinical axillary nerve lesion that is evident by EMG/NCS but which is not apparent clinically because of the associated discomfort. The vast majority of these patients recover from the nerve injury as they rehabilitate from the dislocation or fracture. Blunt trauma to the anterolateral aspect of the shoulder has also been noted to cause injury to the axillary nerve as it travels on the deep surface of the deltoid muscle.
Open reconstructive surgery ( Fig. 18-1 ) and some newer arthroscopic techniques can put the axillary nerve at risk. For example, capsular shrinkage procedures can create a local increase in temperature in the inferior capsule that can lead to nerve injury. Injury to the nerve has been reported in 1% to 2% of thermal capsular shrinkage procedures, but fortunately, the vast majority of these injuries seem to be only temporary. The axillary nerve is also at risk during capsular resection for adhesive capsulitis. Because the nerve is in close proximity to the anteroinferior capsule, great care should be taken when resecting in this area. A safer method of inferior capsular resection in this area is to visualize the axillary nerve with the arthroscope during the procedure.
Young patients may be able to compensate for complete deltoid paralysis and can often perform activities of daily living with only partial disability. The shoulder can easily maintain a full range of motion with an intact rotator cuff. However, most patients experience early fatigue in the involved side if asked to perform repetitive activities. Although deltoid atrophy is quite evident in a fit person, it can sometimes be difficult for the examiner to detect deltoid atrophy in a less fit patient. Injury to the superior lateral cutaneous nerve of the arm can lead to sensory loss over the lateral aspect of the shoulder. It is possible for patients with a complete deltoid motor deficit to experience only a mild loss of sensation over the lateral part of the shoulder. The diagnosis of axillary neuropathy should not be determined by the presence or absence of lateral shoulder sensation. It is unclear whether the sensory branch is spared from injury or whether the sensory zone is supplied by overlapping innervation from other cutaneous branches.
Another potential cause of posterior shoulder pain that has been described is the quadrilateral space syndrome, which presumably results from compression of the axillary nerve within the quadrilateral space. This syndrome is a controversial clinical entity; it might simply be a manifestation of Parsonage-Turner syndrome (brachial neuritis). Tenderness may be noted posteriorly along the shoulder joint; otherwise, the clinical examination is often normal. Deltoid atrophy or lateral sensory changes are uncommon, and EMG examination is usually normal. Magnetic resonance imaging (MRI) may demonstrate signal change indicating denervation in the deltoid and teres minor muscles. Observation is the usual treatment for quadrilateral space syndrome, and the vast majority of patients improve with time. Surgical exploration of the quadrilateral space and release of scar or fibrous bands to achieve decompression of the axillary nerve are rarely needed.
Patients with a history of blunt trauma to the axillary nerve should be observed over at least a 3-month period before operative treatment is considered. At 3 to 4 weeks, a baseline EMG/NCS examination should be obtained. Physical therapy, including active and passive exercises, should be initiated to preserve the maximal range of motion and prevent joint contracture while awaiting return of function. Electrical stimulation of the deltoid has been used in an attempt to preserve muscle viability, although it is unclear whether this approach has any genuine effect.
The results of nonoperative treatment of a blunt traumatic lesion have generally been good. Leffert reported that axillary nerve injury after fracture or dislocation is more common than is usually appreciated, yet the majority of patients progress to full recovery. In a study of 73 patients with proximal humeral fracture or dislocation, 33% were noted on EMG to have an axillary nerve injury, with 9 complete and 15 partial lesions. All of the patients, including those with complete nerve lesions, recovered with no objective loss of function. In a series of 108 elderly patients with anterior shoulder dislocation, 9.3% were found to have an axillary nerve injury, but all went on to full recovery by 12 months. However, some patients do not make the expected recovery. In these patients surgical exploration with neurolysis or possibly nerve grafting can be undertaken if no clinical or EMG recovery is evident by 3 to 4 months. If the patient has experienced a sharp penetrating wound or a surgical injury, surgical exploration should be performed at an earlier date.
The proximal monofascicular structure of the axillary nerve with primarily motor fibers and its relatively short length from the posterior cord to the deltoid motor end plate are characteristics that lend themselves to surgical intervention. Alnot and Valenti reported on 37 axillary nerve surgeries, including 33 cases of sural nerve grafting, 3 neurolysis procedures, and 1 direct repair. In 23 of the 25 isolated axillary nerve lesions M4 or M5 strength was achieved. The fact that 33 of the 37 patients required sural nerve grafting illustrates the difficulty in adequately mobilizing the nerve for direct repair. The small number of patients undergoing neurolysis (3 of 37) is an indication that mild nerve compression by scar or fibrous bands is not common. Repair of the axillary nerve with a short interposed cabled sural nerve graft has been the most common method and has demonstrated the most consistent results.
We have found the use of intraoperative NAPs and EMG techniques to be invaluable in evaluating neuromas and helping us decide whether to perform neurolysis alone or to resect the neuroma and perform a graft. In all cases we are also prepared to expose the axillary nerve posteriorly. In selected cases such exposure is necessary to identify normal nerve more distally for grafting.
Other techniques for repair of the axillary nerve include nerve transfer. Direct neurotization with a donor nerve, such as the medial pectoral, thoracodorsal, or radial, has yielded satisfactory results in cases in which direct repair or short cable grafting of the axillary nerve itself is not possible or not preferable. Nerve transfers using the spinal accessory nerve or upper intercostal nerves have been described, but these require an interpositional sural nerve graft and have demonstrated less optimal results. A nerve transfer using a triceps branch to the anterior division of the axillary nerve has been described in cases of upper trunk brachial plexopathy ; because of promising results, it is now being used by some surgeons instead of nerve grafting techniques in cases of isolated axillary paralysis.
Patients seen longer than 15 to 18 months after trauma usually do not benefit from surgical repair of the nerve because of the poor condition of the deltoid muscle and its motor end plates. In patients with poor shoulder function that limits their activities of daily living but a normal rotator cuff, muscle transfer procedures can be considered. Potentially, if the posterior deltoid and middle deltoid are innervated and the anterior deltoid is not functioning, the posterior-middle deltoid can be rotated anteriorly on the clavicle. This procedure however has the potential to harm the remaining deltoid and is not strongly recommended.
Pedicled pectoralis major transfer can be performed to reconstruct the anterior deltoid ( Fig. 18-2 ). The clavicular head of the pectoralis major alone or combined with the upper two thirds of the sternal head can be rotated on its pedicle to reconstruct mostly the anterior deltoid. Another option is the transfer of pedicled latissimus dorsi based on the thoracodorsal neurovascular bundle to reconstruct the anterior deltoid ( Fig. 18-3 ). One of the authors of this chapter (B.T.E.) has performed the first of these procedures on 17 patients and the second on 23 patients, with very promising results. Alternatively, to reconstruct middle deltoid function, the upper/middle trapezius can be transferred to the proximal humerus. Different techniques have been described for this transfer, including transfer of the muscle on top of the acromion, prolonged with Achilles tendon allograft, or transfer of the acromial bony insertion to the proximal humerus. The outcome of this transfer has been very marginal, especially in patients with no rotator cuff function. We favor pedicled muscle transfers, as described above, when these are available because of the more predictable improvement of shoulder function.
Spinal Accessory Nerve
Injury to the spinal accessory nerve can occur after penetrating trauma to the shoulder. Blunt trauma can also cause loss of trapezius function. Most commonly, surgical dissection in the posterior triangle of the neck, such as for lymph node biopsy, can expose the nerve to possible damage ( Fig. 18-4 ).
The spinal accessory nerve passes through the upper portion of the sternocleidomastoid muscle, which it innervates. It then crosses the posterior cervical triangle, which is bordered anteriorly by the sternocleidomastoid muscle, posteriorly by the trapezius, and inferiorly by the clavicle. The nerve lies on the floor of the posterior triangle with only the overlying fascia as protection against injury. It abuts the posterior cervical lymph nodes. The nerve enters the anterior surface of the trapezius and travels inferiorly, parallel to the medial border of the scapula. The trapezius has a broad origin from the ligamentum nuchae to the 12th thoracic vertebra and inserts over the lateral part of the clavicle, acromion, and spine of the scapula. The upper trapezius is the prime elevator of the scapula, acting to upwardly rotate the lateral aspect of the bone ( Fig. 18-5 ).
A diagnosis of injury to the spinal accessory nerve is often missed, and appropriate treatment may be delayed. Patients usually have vague shoulder pain as their primary complaint. Loss of motion may be a secondary concern. Unless trapezius function is specifically tested, the diagnosis may not be recognized. Because of the common occurrence of anterior shoulder pain in these patients, with occasionally only slight visible wasting of the trapezius muscle, the pain may be assumed to represent postoperative pain, or it may be misinterpreted as resulting from another condition associated with shoulder pain, such as rotator cuff pathology. The trapezius receives some innervation from the upper cervical nerve roots; thus complete atrophy of the muscle might not occur.
The spinal accessory nerve supplies the sole innervation to the lateral portion of the muscle, which is critical in supporting abduction of the shoulder. If the patient is carefully observed, the scapula appears to be rotated forward at the shoulder. The shoulder might hang lower or droop in comparison to the contralateral side; this difference can even cause symptoms of thoracic outlet syndrome. Winging of the scapula often occurs but is not as dramatic as that associated with injury to the long thoracic nerve. Usually, a shoulder shrug produces the deformity; however, a strong levator scapula can compensate well, and only minimal deformity may be noted. The patient might not be able to abduct the arm fully with the forearm either pronated or supinated.
Some of the pain that patients experience might come from strain of the other parascapular muscles as they attempt to compensate for the lack of trapezius function. Additionally, because the scapula cannot properly rotate the acromion away from the humerus as the arm is elevated, impingement of the rotator cuff can cause secondary rotator cuff tendinopathy. If attention is directed only at the rotator cuff, the underlying nerve pathology will be missed. In addition, the spinal accessory nerve, although mainly a motor nerve, still has sensory fibers. Injury to the nerve can also produce neuropathic pain. Finally, examination should also cover possible concomitant injury to other neighboring nerves, such as the cervical plexus or great auricular nerve.
Whenever possible, surgical exploration of the nerve should be performed within 6 months. Although surgical reconstruction of the nerve may be considered up to 1 year after injury, better results occur with treatment as soon after injury as possible. We favor early exploration of these nerve injuries when they occur immediately after surgery, especially when the nerve was not identified and protected as part of the operation.
Surgical options include neurolysis, direct repair, or nerve grafting, depending on intraoperative observations and electrophysiologic testing. During surgical reconstruction, it is important to consider the acromion-mastoid distance in the anesthetized patient. If direct repair of the nerve is performed with the head tilted toward the operated shoulder, significant traction can occur when the patient is awakened after surgery and transported to the recovery room, and this can disrupt the repair. If the nerve ends are found to have retracted at surgery, it is best to use an intervening graft, such as a sural nerve or the great auricular nerve, to reduce tension on the repair ( Fig. 18-6 ).
Nerve transfer using a pectoral branch can be performed in cases of proximal injury to the accessory nerve, inability to identify proximal stump, or late referral. If more than 12 to 18 months have passed since injury to the spinal accessory nerve, nonoperative treatment may be considered if the patient has compensated reasonably well. The degree of disability varies from patient to patient, even despite aggressive physical therapy. Some patients experience only a persistent ache in the shoulder, whereas others feel and act completely disabled with respect to the upper extremity. Braces have been advocated as adjunctive treatment, but they are bulky and not used consistently by the patients. A patient symptomatic enough to attempt to use a brace is potentially a candidate for surgical reconstruction.
Modern surgical procedures currently involve dynamic muscle transfer techniques. Earlier historical procedures however initially involved mostly static repairs. Henry and others advocated static stabilization of the medial aspect of the scapula to the vertebral spine with strips of fascia lata. Dewar and Harris described lateral transfer of the levator scapulae to the lateral part of the scapula, combined with a static fascial sling from the vertebral spine to the medial part of the scapula. However, static repairs with fascia, tendon, or artificial materials tend to stretch out or rupture over time.
Dynamic transfer of the levator scapulae along with the rhomboid major and rhomboid minor was described in Germany by Eden and Lange. Bigliani reported good results with this technique.
Elhassan and Wagner proposed a modification of the Eden-Lange (EL) procedure, called the “triple transfer,” with good to excellent results in 22 patients. The authors reported that the line of pull of the rhomboid major and minor when transferred to the body of the scapula worked in almost the opposite direction to the lower trapezius, and for this reason these muscles should independently be transferred more proximally for better replication of the line of pull of the middle/lower trapezius.
The surgical approaches for the EL procedure and the Elhassan-Wagner (EW) modified transfer are similar and involve an inverted L -incision with a vertical incision placed midway between the vertebral spine and the medial edge of the scapula and a horizontal incision extended laterally to the mid-aspect of the supraspinatus fossa. The atrophied trapezius is detached from the spine of the scapula and acromion and can be used to cover the muscles’ transfer at the end of the case. The levator scapulae and rhomboidei are exposed and dissected. If the EL procedure is performed then the levator scapulae is detached with a small bony insertion, and the rhomboid minor and major are detached together with small bony insertions. This release can be accomplished with a small saw or an osteotome. At this time the spine of the scapula is debrided to prepare it for the bony insertion of the muscle transfer. With the EL procedure ( Fig. 18-7 ), the levator scapulae is attached laterally on the spine of the scapula, posterior to the acromion. The levator should be placed as far laterally as possible on the spine of the scapula, usually 5 to 7 cm medial to the posterolateral edge of the acromion. The rhomboid major and minor can be elevated individually or together from the medial edge of the scapula. The dorsal scapular nerve is potentially at risk during this part of the procedure. The infraspinatus is partially elevated from the scapula in a medial-to-lateral direction. The rhomboids are then placed as far laterally as possible—at least 4 cm—on the posterior aspect of the scapula and secured in place via suture and drill holes through the scapula. Alternatively, the rhomboid minor can be transferred cephalad to the spine of the scapula into the supraspinatus fossa. The infraspinatus is then sutured back in position over the transferred rhomboids.
With the EW transfer, the levator scapula is placed as lateral as possible on the spine of the scapula (as described above), the rhomboid minor is attached on the spine of the scapula just medial to the levator scapulae, and the rhomboid major is attached on the spine of the scapula just medial to the rhomboid minor. Because of the wider insertion of the rhomboid major, its attachment on the medial spine of the scapula is better performed by breaking the bony attachment in its middle part and then closing the two edges of the bone (like closing a fish’s mouth) around the medial spine of the scapula ( Fig. 18-8 ).
Postoperatively, patients are placed in a shoulder abduction brace that holds the arm at approximately 70 degrees of abduction for 6 to 8 weeks, depending on the preference of the treating surgeon. Because of the smaller size of the levator scapulae and rhomboidei, we prefer 8 weeks of immobilization to allow full healing of the transfers and to reduce the risk of stretch injuries from early mobilization. At 8 weeks, patients are started on progressive range of motion for 6 weeks followed by gentle strengthening and aquatherapy for 8 weeks; they are allowed unrestricted activities at 6 months.
Good outcomes have been reported with both techniques, but the EW procedure has been reported to have a better outcome and less risk of failure. If an EL procedure fails, the only possible remaining muscle transfer option available is to transfer the lower contralateral trapezius origin to the lateral spine of the scapula to restore scapula retraction. One of the authors (B.T.E.) has performed this procedure on five patients with a failed prior EL procedure and the results were promising in all five cases. The main salvage procedure to stabilize the scapula in symptomatic patients with failed EL procedure is scapulothoracic fusion. Different techniques have been used to perform scapulothoracic fusion, but most involve passing wires through the scapula and around several ribs, with a broad iliac crest bone graft or metallic plate for support. The complication rate can be high, with the potential for pneumothorax or hardware failure.
Long Thoracic Nerve
The serratus anterior originates from the upper nine ribs and inserts on the anteromedial border of the scapula. This insertion is only a few millimeters wide at the midportion of the scapula, but it becomes more substantial at the inferior pole of the scapula ( Fig. 18-9 ). It is the inferior portion of the muscle that is important in maintaining protraction and upward rotation of the scapula during forward elevation of the shoulder. The long thoracic nerve has a relatively long course after taking origin from the C5, C6, and C7 nerve roots. After crossing over the first rib, it travels 10 to 20 cm to its motor end plate in the serratus anterior. It is vulnerable to blunt trauma over the first rib along the lateral chest wall and can be crushed by forceful displacement of the scapula.
The nerve is rarely injured as a result of penetrating trauma, but injury can occur during thoracic outlet surgery in the region of the first rib, breast surgery, or lateral chest wall procedures, such as axillary node dissection. Spontaneous cases of entrapment at the scalenus medius have been described. The most common cause of serratus anterior dysfunction is probably Parsonage-Turner syndrome; indeed, this condition is most likely the underlying cause of long thoracic nerve dysfunction attributed to overexertion, including athletic activities.
Isolated injury to the long thoracic nerve is usually manifested as winging of the scapula. Velpeau first described injury to the long thoracic nerve causing paralysis of the serratus anterior in 1837. However, winging of the scapula has many possible causes, with multidirectional instability probably being the most common cause of mild winging. Spinal accessory nerve injury can also cause winging, but this injury tends to be milder and results in more of a rotational deformity of the scapula. Some patients have volitional control over the scapula and can demonstrate significant winging at will.
In our opinion, observation should be the standard therapy in idiopathic or nonpenetrating trauma cases. Other than continued use of the shoulder as tolerated, no specific physical therapy protocol has been found to be especially helpful. Braces have been advocated to help hold the scapula against the chest wall ; however, although these can be somewhat effective, they are usually found to be awkward and are not well tolerated by patients. Most patients with a nontraumatic or idiopathic cause tend to recover from the paralysis and regain serratus anterior function within 6 months to 1 year.
There are potential surgical options available for treating injury to the long thoracic nerve in the early stages. Some surgeons have favored neurolysis of the nerve with decompression at the level of the scalenus medius. However, because it is difficult to be certain where the lesion resides, or if compression (rather than inflammation, for example) is responsible for the dysfunction, we do not recommend this approach for most patients. In cases where there has been no spontaneous recovery, another strategy is to perform neurotization (or nerve transfer) using one or two intercostal nerves or the thoracodorsal nerve (M.B. Wood, personal communication, 2002). Because it is not usually clear where the damaged section is, the nerve can be connected with the donor nerve close to the motor end plate. This technique has been helpful in a small number of reported cases. Many patients with atraumatic lesions recover spontaneously, so we do not routinely recommend surgery for most patients before 6 or 9 months after the nerve deficit develops.
If a patient does not recover serratus anterior function and use of the shoulder is compromised, a number of reconstructive options are available ( Fig. 18-10 ). Scapulothoracic fusion has been performed when the scapula is fixed to the underlying ribs. Although this technique can eliminate winging of the scapula, it will reduce shoulder girdle motion by at least 30%, with mostly forward elevation and extension affected. Pneumothorax is a risk with this technique, and the pseudarthrosis rate is not insignificant. For these reasons, scapulothoracic fusion should be reserved for the salvage situation or for patients with symptomatic facioscapulohumeral dystrophy.
Tendon transfers provide dynamic control of a winging scapula and are now our preferred method in patients with a neurologic deficit that has lasted for 1 year or more. Tubby described transfer of the pectoralis major to the serratus anterior in 1904. Although this transfer might offer initial relief of the winging, the paralyzed serratus anterior tends to stretch out with time, and this procedure is not recommended today. Other techniques described include transfer of the pectoralis minor, rhomboids, or levator scapulae.
Pectoralis Major Transfer
The technique that seems to give the most consistent result is transfer of the pectoralis major to the scapula with tendon graft augmentation. Durman, Ober, and Marmor each described successful transfer of the pectoralis major with a fascial extension graft in a few cases. More recent studies have demonstrated the excellent ability of this tendon transfer procedure to control winging of the scapula. The pectoralis major is ideally suited as a transfer to substitute for the paralyzed serratus anterior. The direction of pull of the pectoralis major is similar to the path of the serratus anterior, and the bulk of the pectoralis major provides enough strength to resist winging of the scapula. The technique has been used with transfer of the entire pectoralis major or with only the sternal head ( Fig. 18-11 ); equally good results have been reported with both procedures. However, the sternal head of the pectoralis major is much more in line with the direction of pull of the serratus anterior than is the clavicular head, and it also provides for a less bulky transfer. Furthermore, the sternal head is more substantial than the clavicular head and provides a stronger tendon transfer. Potential concerns of cosmesis or visible deformity of the anterior chest wall should be minimal. In most instances regardless of whether the entire pectoralis major or only the sternal head is transferred, there is little change in the normal contour of the anterior chest.
This procedure can be performed with a single large incision across the axilla or through two separate incisions. The two-incision technique is not technically harder than a single incision and is somewhat more cosmetic and is the preferred procedure. The choice of tissue for augmentation of the transferred pectoralis major tendon depends on the surgeon’s preference. The most common choice has been a large portion of fascia lata rolled into a tube. Other graft options include semitendinosus or gracilis autograft or allograft. Fascia lata can be rolled into a spiral tube and draped around the pectoralis major muscle and tendon to provide very strong proximal fixation.
More recently, Elhassan and Wagner reported on the outcome of direct transfer of the sternal head of the pectoralis major with its bony insertion to the scapula for patients with symptomatic scapula winging as result of chronic long thoracic nerve palsy.
Pectoralis transfer to the scapula augmented with fascia lata is performed with the patient positioned in the lateral decubitus position for easy access to the anterior and posterior aspects of the shoulder and the lateral aspect of the ipsilateral thigh. The anterior approach is usually performed while a second surgical team is simultaneously harvesting the fascia lata. This approach is made somewhat easier by the lateral position. The anterior incision (4 to 6 cm long) is made almost entirely in the axillary crease and is extended only a centimeter superiorly. This technique results in a well-hidden, cosmetic incision. The sternal head of the pectoralis major is easily identified at the inferior margin of the muscle. It wraps posteriorly under the clavicular head and inserts more medial and superior to it. The sternal head is detached directly off the humerus, with care taken to avoid damage to the biceps tendon. The sternal head is then freed up medially onto the chest wall to allow greater excursion of the muscle. The harvested fascia lata graft should measure approximately 14 × 5 cm. The fascia lata is rolled into a spiral tube, draped around the pectoralis major tendon and muscle, and secured with multiple sutures. A heavy running locking suture is then placed through the fascia lata graft and tagged for transfer.
The posterior incision should be made at the junction of the middle and distal lateral edges of the scapula. It is important that an assistant manually pushes the scapula as far anteriorly and laterally as possible. When this position is achieved, a 4-cm incision is made over the lateral edge of the scapula, and the muscles are cleared off the bone and retracted. An 8- to 10-mm hole is created in the scapula with a bur, just medial to the thick lateral mass of the scapula, to ensure a strong bony bridge during the healing period. A large bone hook is helpful at this stage to secure the scapula. Using a large, blunt clamp, a subcutaneous path is created from the anterior incision along the chest wall to the posterior incision. Much of this dissection can be performed with digital palpation. This plane is safe because the brachial plexus is more superior and abducted with the arm.
Once a clear path has been created for the pectoralis major and graft, the traction suture is passed posteriorly and the graft is pulled through the scapular fenestration in an anterior-to-posterior direction. The graft is then pulled tight, attempting to abut the native pectoralis major tendon to the scapula. The tendon is then folded back and sutured to itself. It is important for the scapula to be held as far anterior on the chest wall as possible while these sutures are secured. No reports of overtight pectoralis major transfers have appeared in the literature. Any excess graft after suture fixation is excised. Drains are not usually needed at closure, and the arm is placed into a sling.
Direct transfer of the bony insertion of the sternal head of the pectoralis major to the scapula is also performed in the lateral position, but the thigh does not need to be included in the surgical field as there is no need for tendon graft. A small anterior incision is performed from the axillary fold proximally for 5 cm. Exposure of the pectoralis and the separation of the sternal head from the clavicular head are similar to the technique described above. However, when the sternal head is ready to be detached, electrocautery is used to define the bony insertion of the sternal head of the pectoralis major. Before performing the bony detachment, the surgeon can either retract the clavicular head and leave it attached to the humerus, or detach it and reinsert it later at the site of the bony defect created at the site of the bony harvesting of the sternal head of the pectoralis major. We prefer the latter because it makes the freeing of the sternal head of the pectoralis major and obtaining full excursion easier. We use a combination of electrical saw and small osteotomes to perform the bony detachment of the insertion of the sternal head of the pectoralis major. Once the tendon is detached with its bony insertion, the tendon is tagged with a nonabsorbable suture to facilitate its mobilization and transfer ( Fig. 18-12 ). The rest of the steps are very similar to those in the description above, except for the attachment of the tendon. The lower scapula is detached and the lateral distal aspect of it is debrided to obtain bleeding bone ( Fig. 18-13 ). The detached tendon is passed from the anterior to posterior using a grasping instrument ( Fig. 18-14 ), and then multiple (at least five) No. 2 nonabsorbable sutures are placed in transosseous fashion to be used for the repair. Then, bone-on-bone repair is performed using the sutures as cerclage sutures. If the clavicular head was detached, it is reattached at this time to the proximal humerus at the level of the bone defect that is created from bone harvesting. The technique of attachment is very similar to that for biceps repair.
Postoperatively, the patient is told to use the sling full time and to avoid any abduction of the shoulder. At 6 weeks after surgery, the sling may be discarded and the patient allowed to resume all normal daily activities with the arm, although lifting of objects heavier than 1 kg is not permitted. A formal physical therapy program is not usually necessary. Patients tend to regain a normal range of motion quite readily. It is assumed that healing takes place during the initial 3- to 6-month period, and therefore return to manual labor or sporting activities is allowed after 6 months. However, early failure has occasionally been reported after pectoralis major transfer, apparently related to a premature return to full function before complete healing. If a patient, especially an athlete, who has undergone direct transfer of the bony insertion of the sternal head of the pectoralis major wishes to return early to sports activities, a CT scan of the scapula is obtained and the patient is allowed to return much more quickly to full sports activities if it shows adequate bone-to-bone healing ( Fig. 18-15 ).
The suprascapular nerve originates from the C5 and C6 nerve roots at the junction of the upper trunk and its divisions. The nerve follows the omohyoid muscle posteriorly and then runs inferiorly through the suprascapular notch—a region where it is relatively fixed in position—bridged by the superior transverse scapular ligament. The suprascapular artery and vein typically pass superior to the ligament. After exiting the suprascapular notch, the nerve gives off branches that innervate the supraspinatus. It then continues medial to the superior edge of the glenoid and enters the spinoglenoid notch at the lateral margin of the scapular spine.
As the nerve travels posteriorly, it may be less than 20 mm from the glenoid edge. In the spinoglenoid notch, the spinoglenoid ligament can potentially impinge on the nerve. A spinoglenoid ligament however is not present in all patients. After exiting the spinoglenoid notch, the nerve divides into two to four branches that enter the infraspinatus. Near the suprascapular notch, the suprascapular nerve also supplies articular branches to the shoulder joints. These fibers might explain the pain experienced by patients with suprascapular nerve lesions at the transverse scapular ligament. In contrast, more distal lesions, such as at the spinoglenoid notch, characteristically produce painless atrophy of the infraspinatus.
The cutaneous supply of the suprascapular nerve is still somewhat debatable. Some have described sensory fibers supplying the skin over the posterior of the shoulder; however, few patients are noted to have loss of sensation after nerve injury.
Suprascapular neuropathy can occur due to a variety of causes. Direct blunt trauma to the shoulder or sudden twisting of the shoulder can acutely injure the nerve. A compressive injury can develop at the suprascapular or spinoglenoid notch.
A common cause of nerve compression in this region is a ganglion cyst ( Fig. 18-16 ). The origin of a cyst in this region is presumed to be from the glenohumeral joint, and it is very often associated with degenerative tears of the glenoid labrum. A rare type of cyst, an intraneural ganglion cyst, can affect the suprascapular nerve. With this lesion, mucin is contained within the epineurium of the peripheral nerve. Suprascapular intraneural ganglion cysts have been shown to originate from the glenohumeral joint and to track along the articular branch into the parent nerve as far as the neck ( Fig. 18-17 ). They have also been associated with superior labrum anterior and posterior tears.
Repetitive rotatory motion of the shoulder, such as occurs during many sporting activities, can cause a chronic traction injury of the nerve. Neuropathy of the suprascapular nerve has been noted in participants of baseball, volleyball, tennis, and weightlifting. Parsonage-Turner syndrome (acute brachial neuritis) can also result in an idiopathic etiology of suprascapular neuropathy. In rare instances a fracture of the scapula that enters the suprascapular notch can damage the nerve.
Clinical Manifestation and Diagnosis
Patients with a suprascapular nerve injury at the suprascapular notch commonly complain of pain over the posterior and lateral aspects of the shoulder. Pain on deep palpation over the suprascapular notch itself might also be present. In addition, patients might have weakness of abduction, external rotation, or both and may be noted to have atrophy of the supraspinatus and infraspinatus. In many individuals the supraspinatus is difficult to visualize because a well-developed trapezius covers the supraspinatus and makes it difficult to see muscle wasting. Loss of muscle bulk in the infraspinatus however is fairly easy to see on clinical examination. Some patients do not demonstrate any atrophy on examination; shoulder pain and mild weakness of abduction may be the only findings, making a correct diagnosis difficult.
An appropriate clinical examination with EMG/NCS often detects suprascapular neuropathy. The diagnosis can easily be made if conduction velocity is delayed (in comparison to the contralateral side) and fibrillation potentials are noted in the supraspinatus or infraspinatus. In some patients EMG/NCS shows only mild involvement. These patients should be examined carefully for other shoulder pathology that could have been missed during the clinical examination. Surgical treatment tends not to be as beneficial in these patients as in those with substantial changes on EMG/NCS.
Shoulder pain can be caused by several conditions, and thus cervical disk disease, rotator cuff tear, impingement syndrome, and acromioclavicular joint degeneration should all be excluded in these patients, especially those with only mild EMG/NCS findings.
Treatment of a patient with suprascapular neuropathy is based on the type of injury and the duration of disability, pain, and atrophy. A patient with a chronic neuropathy from repetitive shoulder motion, such as a baseball or volleyball player, should be placed in therapy and treated conservatively. Similarly, a patient with Parsonage-Turner syndrome should be observed for an extended period. Both groups, if monitored for 2 to 3 months, tend to demonstrate clinical improvement. The results of surgical intervention in these two groups are somewhat unpredictable.
A patient with evidence of compressive neuropathy at either the suprascapular notch or the spinoglenoid notch should be observed for improvement in function over a 3- to 4-month period. If the patient remains symptomatic after that time, surgical decompression can be performed.
Early surgical intervention (i.e., within a few weeks) is rarely needed, except for patients with an obvious ganglion cyst noted on MRI or the rare patient with intractable pain and significant compressive neuropathy noted on EMG/NCS. Patients with Parsonage-Turner syndrome experience considerable pain for the first few weeks of the neuropathy, but typically, they do not require operative intervention because their symptoms tend to improve spontaneously over time.
The suprascapular nerve can be approached from an anterior, superior, or posterior direction. The anterior approach demands exposure of at least a portion of the brachial plexus and does not allow exposure of the spinoglenoid notch. A direct superior approach performed by splitting the trapezius in line with its fibers has commonly been recommended. This technique however results in a deep surgical wound in which the suprascapular notch and superior transverse scapular ligament can be visualized, but only a small distance of the suprascapular nerve itself can be exposed.
The posterior approach detaches the trapezius from the spine of the scapula 6 to 8 cm medial to the lateral edge of the acromion. The supraspinatus is then retracted posteriorly to visualize the suprascapular nerve and the notch. Careful exposure should be performed while in the area of the suprascapular notch because the artery and vein travel above this, and occasionally, a branch of the nerve may be positioned above rather than below the superior transverse scapular ligament. If greater exposure of the nerve is needed with this approach, two additional steps can be performed: the trapezius can be detached laterally and anteriorly onto the acromioclavicular joint to allow wide exposure of the nerve. However, at closure, the trapezius will need to be accurately repaired back to the acromion.
If the suprascapular nerve needs to be explored into the spinoglenoid notch, the posterior deltoid can be detached from the spine of the scapula, and exposure of the nerve from the infraspinatus to the suprascapular notch can then be achieved.
In cases involving a ganglion at the spinoglenoid notch, arthroscopic debridement of the cyst is an excellent technique. Arthroscopy also allows visualization of the joint and the opportunity to debride any associated labral tears. Most of these posterosuperior labral tears are not amenable to surgical reattachment. Recently, arthroscopic and endoscopic approaches to decompress the suprascapular nerve at the suprascapular and spinoglenoid notches have been described.
For patients with suprascapular neuropathy, good or better surgical results have been achieved in the recovery of supraspinatus function. Infraspinatus recovery is more variable. The best results are seen in patients with nerve lesions from masses; the worst are in patients with neuropathy caused by fractures or severe trauma.