Adult Traumatic Brachial Plexus Injuries

Alexander Y. Shin, MD, Robert J. Spinner, MD, Scott P. Steinmann, MD, and Allen T. Bishop, MD

Abstract Adult traumatic brachial plexus injuries are devastating, and they are occurring with increasing frequency. Patient evaluation consists of a focused assessment of upper extremity sensory and motor function, radiologic studies, and, most important, preoperative and intraoperative electrodiagnostic studies. The critical concepts in surgical treatment are patient selection as well as the timing and prioritizing of restoration of function. Surgical techniques include neurolysis, nerve grafting, neurotization, and free muscle transfer. Results are variable, but increased knowledge of nerve injury and repair, as well as advances in microsurgical techniques, allow not only restoration of elbow flexion and shoulder abduction but also of useful prehension of the hand in some patients.

Dr. Shin is Associate Professor, Department of Orthopaedic Surgery, Division of Hand Surgery, Mayo Clinic, Rochester, MN. Dr. Spinner is Associate Professor, Department of Neurosurgery and Department of Orthopaedic Surgery, Division of Hand Surgery, Mayo Clinic. Dr. Steinmann is Assistant Professor, Department of Orthopaedic Surgery, Mayo Clinic. Dr. Bishop is Professor, Department of Orthopaedic Surgery, Mayo Clinic. None of the following authors or the departments with which they are affiliated has received anything of value from or owns stock in a commercial company or institution related directly or indirectly to the subject of this article: Dr. Shin, Dr. Spinner, Dr. Steinmann, and Dr. Bishop. Reprint requests: Dr. Shin, Mayo Clinic, E14A, 200 1st Street SW, Rochester, MN 55905. J Am Acad Orthop Surg 2005;13:382396 Copyright 2005 by the American Academy of Orthopaedic Surgeons.

rachial plexus lesions frequently lead to significant physical disability, psychological distress, and socioeconomic hardship. These lesions can result from a variety of etiologies, including birth injuries, penetrating injuries, falls, and motor vehicle trauma. Most are closed injuries involving the supraclavicular region rather than the retroclavicular or infraclavicular level. The roots and trunks are more commonly affected than the divisions, cords, or terminal branches. Most injuries occur as a result of fracture or compression or as a combination of these. In the supraclavicular region, traction injuries occur when the head and neck are violently moved away from the ipsilateral shoulder, often resulting in an injury to the C5 or C6 roots or upper trunk. Traction to the brachial plexus also can occur secondary to violent arm movement; when the arm is abducted over the head with significant force, traction occurs within the lower elements of the brachial plexus (C8-T1 roots or

lower trunk). Compression injuries to the brachial plexus usually occur between the clavicle and the first rib. Direct blows also may result in injuries to the brachial plexus, especially around the coracoid process of the scapula. The exact number of brachial plexus injuries that occur each year is difficult to ascertain; however, with the advent of increasingly extreme sporting activities and highenergy motor sports, as well as the increasing number of survivors of high-speed motor vehicle accidents, the number of brachial plexus injuries continues to rise throughout the world.1-6 Most of these injuries occur in males aged 15 to 25 years.5,7,8 Based on his experience with 1,068 patients with brachial plexus injuries during an 18-year span, Narakas9 developed his rule of seven seventies. He reported that approximately 70% of traumatic brachial plexus injuries occurred secondary to motor vehicle accidents; of these, approximately 70% involved motorcycles or bicy-

382

Journal of the American Academy of Orthopaedic Surgeons

Alexander Y. Shin, MD, et al

Figure 1

Anatomy of the brachial plexus. A, The brachial plexus has five major segments: roots, trunks, divisions, cords, and branches. The clavicle overlies the divisions. The roots and trunks compose the supraclavicular plexus, and the cords and branches compose the infraclavicular plexus. B, The relationship between the axillary artery and the cords. The cords are named for their anatomic relationship to the axillary artery: lateral, medial, and posterior. LC = lateral cord, LSS = lower subscapular nerve, MABC = medial antebrachial cutaneous nerve, MBC = medial brachial cutaneous nerve, MC = medial cord, PC = posterior cord, TD = thoracodorsal nerve, USS = upper subscapular nerve. (Adapted by permission of Mayo Foundation.)

cles. Of the cycle riders, approximately 70% had multiple injuries. Overall, 70% had supraclavicular lesions; of those, 70% had at least one root avulsed. At least 70% of patients with a root avulsion also have avulsions of the lower roots (C7, C8, or T1). Finally, of patients with lower root avulsion, nearly 70% will experience persistent pain. Treatment recommendations for complete root avulsions have varied widely over the past 50 years. Following World War II, the standard approach was surgical reconstruction by shoulder fusion, elbow bone block, and finger tenodesis.10 Yeoman and Seddon11 noted a tendency among these patients to become one-handed within 2 years of injury, resulting in few successful outcomes regardless of the treatment approach. Their retrospective study revealed no good results from braVolume 13, Number 6, October 2005

chial plexus intervention surgery. However, amputation plus shoulder fusion performed within 24 months of injury resulted in predominantly good and fair outcomes. Consequently, in the 1960s, transhumeral (above-elbow) amputation, combined with shoulder fusion in slight abduction and flexion, was advocated.12 However, loss of glenohumeral motion caused by brachial plexus injuries limited the effectiveness of bodypowered prostheses (eg, figure-of-8 harness with farmers hook). Advances in brachial plexus reconstruction have yielded outcomes superior to historical results. A better understanding of the pathophysiology of nerve injury and repair, as well as recent advances in microsurgical techniques, have allowed reliable restoration of elbow flexion and shoulder abduction, in addition to useful prehension of the hand in some cases.

Anatomy
The brachial plexus is formed from five cervical nerve roots: typically, C5, C6, C7, C8, and T1 (Fig. 1). Additionally, there may be contributions to the brachial plexus from C4, ranging from small branches to larger contributions, and from T2. A plexus with contributions from C4 is called prefixed. The incidence of prefixed plexuses ranges from 28% to 62%. When contributions from T2 occur, the plexus is termed postfixed. The incidence of postfixed plexuses ranges from 16% to 73%.13 The so-called true form of the brachial plexus was described by Kerr,13 who performed detailed anatomic dissections on 175 specimens. In the true form there are five separate sections of the brachial plexus: roots, trunks, divisions, cords, and terminal branches. Formed by the coalescence
383

Adult Traumatic Brachial Plexus Injuries

Figure 2

A, Anatomy of the brachial plexus roots and types of injury. The roots are formed by the coalescence of the ventral (motor) and dorsal (sensory) rootlets as they pass through the spinal foramen (A). The dorsal root ganglion holds the cell bodies of the sensory nerves; the cell bodies for the ventral nerves lie within the spinal cord. Three types of injury can occur: avulsion injuries pull the rootlets out of the spinal cord (B); stretch injuries attenuate the nerve (C); and ruptures result in complete discontinuity of the nerve (D). B, Intraoperative photograph of a preganglionic injury (root avulsion) as well as a postganglionic injury. The C5 root is avulsed with its dorsal and ventral rootlets. The asterisk marks the dorsal root ganglion. The C6 root, which is inferior, demonstrates a rupture at the root level. (Panel A adapted by permission of Mayo Foundation. Panel B reproduced by permission of Mayo Foundation.)

of the ventral and dorsal nerve rootlets, the root passes through the spinal foramen (Fig. 2, A). The dorsal root ganglion holds the cell bodies of the sensory nerves and lies within the confines of the spinal canal and foramen. A preganglionic injury is one in which the spinal roots are avulsed from the spinal cord (Fig. 2, B). Preganglionic injuries can be separated into central avulsions, in which the nerve is avulsed directly from the spinal cord, and intradural ruptures, in which rootlets rupture proximal to the dorsal root ganglion. An injury distal to the dorsal root ganglion is called postganglionic (Fig. 2, B). Distinguishing between a preganglionic and a postganglionic injury is important when considering the possibility of spontaneous recovery and implications for surgical reconstruction because there is little potential recov384

ery at this time for preganglionic injury. The C5 and C6 roots merge to form the upper trunk, and the C8 and T1 roots merge to form the lower trunk. C7 becomes the middle trunk. The point at which C5 and C6 merge (Erbs point) marks the location at which the suprascapular nerve emerges. Each trunk then divides into an anterior and a posterior division and passes beneath the clavicle. The posterior divisions merge to become the posterior cord, and the anterior divisions of the upper and middle trunk merge to form the lateral cord (Fig. 1, B). The anterior division from the lower trunk forms the medial cord. The posterior cord forms the axillary nerve and the radial nerve. The lateral cord splits into two terminal branches: the musculocutaneous nerve and

the lateral cord contribution to the median nerve. The medial cord contributes to the median nerve as well as to the ulnar nerve. A few terminal nerve branches come off the roots, trunks, and cords. The branches off the C5 root include a branch to the phrenic nerve, the dorsal scapular nerve (rhomboid muscles), and the long thoracic nerve (serratus anterior muscle) (Fig. 1, A). The branches off C6 and C7 also contribute to the long thoracic nerve (serratus anterior muscle). The branches off the upper trunk include the suprascapular nerve (supraspinatus and infraspinatus muscles) and the nerve to the subclavius muscle. The lateral cord gives off the lateral pectoral nerve, while the posterior and medial cords each have three branches. The posterior cord gives off branches (proximal to distal) that in-

Journal of the American Academy of Orthopaedic Surgeons

Alexander Y. Shin, MD, et al

clude the upper subscapular nerve, thoracodorsal nerve, and the lower subscapular nerve. The medial cord gives off the medial pectoral nerve, the medial antebrachial cutaneous nerve, and the medial brachial cutaneous nerve. By noting loss of function to these muscles, one can gain knowledge on pinpointing the level of brachial plexus injury. The sympathetic ganglion for T1 lies in close proximity to the T1 root and provides sympathetic outflow to the head and neck. Avulsion of the T1 root (a pre-ganglionic injury) results in interruption of the T1 sympathetic ganglion, resulting in Horner syndrome, which consists of miosis (small pupil), enophthalmos (sinking of the orbit), ptosis (lid droop), and anhydrosis (dry eyes).

Patient Evaluation
Physical Examination Brachial plexus injury is often seen in patients who have sustained polytrauma; thus, diagnosis of the nerve injury necessarily may be delayed until the patient is stabilized and resuscitated. A high index of suspicion for a brachial plexus injury should be maintained when examining a patient with severe shoulder girdle injury. On initial examination, the patient is often obtunded or sedated, and careful observation is needed as the patient becomes more coherent. A detailed examination of the brachial plexus and its terminal branches can be performed in a few minutes on an awake, cooperative patient when the examiner is experienced and systematic. The median, ulnar, and radial nerves are evaluated by examining finger and wrist motion. Elbow flexion and extension are examined to determine musculocutaneous and high radial nerve function. Examination of shoulder abduction can determine the function of the axillary nerve, a branch of the posterior cord. Injury to the posterior cord may affect both deltoid
Volume 13, Number 6, October 2005

function and the muscles innervated by the radial nerve. Examination of wrist extension, elbow extension, and shoulder abduction may help determine the condition of the posterior cord. The latissimus dorsi is innervated by the thoracodorsal nerve, which is also a branch of the posterior cord. This muscle can be palpated in the posterior axillary fold and can be felt to contract when a patient is asked to cough. The pectoralis major is innervated by the medial and lateral pectoral nerves, each a branch of the medial and lateral cords, respectively. The medial pectoral nerve innervates the sternal head of the pectoralis major, and the lateral pectoral nerve innervates the clavicular head. The entire pectoralis major muscle can be palpated from superior to inferior as the patient adducts the arm against resistance. Located proximal to the cord level, the suprascapular nerve is a terminal branch at the trunk level. It can be examined by assessing shoulder external rotation and elevation. Often, in a chronic situation, the posterior aspect of the shoulder demonstrates significant atrophy in the area of the infraspinatus muscle. Supraspinatus muscle atrophy is harder to detect clinically because the trapezius muscle covers most of the supraspinatus muscle. Loss of shoulder flexion, rotation, and abduction also may be caused by a significant rotator cuff or deltoid injury. Both axillary nerve function and rotator cuff integrity should be evaluated when testing shoulder function. Certain findings suggest preganglionic injury on clinical examination. For example, the patient should be examined for the presence of Horner syndrome, which is suggestive of a root avulsion at the C8-T1 level. Injury to the long thoracic nerve or the dorsal scapular nerve suggests a higher (more proximal) level of injury because both nerves originate at the root level. The long thoracic nerve is formed from the roots of C5-

C7 and innervates the serratus anterior muscle. This nerve, >20 cm long, is vulnerable to injury as it descends along the chest wall. Injury to the long thoracic nerve with resultant dysfunction of the serratus anterior muscle causes significant scapular winging as the patient attempts to forward elevate the arm. The dorsal scapular nerve is derived from C4-C5 and innervates the rhomboid muscles, often at a foraminal level. Careful examination demonstrates atrophy of the rhomboids and parascapular muscles when this nerve is injured. The patient must be observed posteriorly to fully evaluate the serratus anterior and rhomboid muscles. Neighboring cranial nerves must be considered during motor testing. The spinal accessory nerve that innervates the trapezius muscle can occasionally be injured with the neck or shoulder trauma that affects the brachial plexus. Its integrity is important because the spinal accessory increasingly is used as a nerve transfer. Careful sensory (and/or autonomic) examination should include various nerve distributions (especially autonomous zones). Sensation of root-level dermatomes can be unreliable because of either overlap from other nerves or anatomic variation. The examiner should record active and passive ranges of motion as well as the presence or absence of reflexes. The presence of concomitant spinal cord injury should be considered by examining for lower limb strength, sensory levels, increased reflexes, and pathologic reflexes. Percussing the nerve is especially helpful. Acutely, pain over a nerve suggests a rupture. Lack of percussion tenderness over the brachial plexus indicates an avulsion. An advancing Tinel sign is sometimes suggestive of a recovering lesion. Because it is possible also to rupture the axillary artery at the time of significant brachial plexus injury, a vascular examination should be per385

Adult Traumatic Brachial Plexus Injuries

Figure 3

ciated inflammation or edema, and it can evaluate mass lesions in the patient with spontaneous nontraumatic neuropathy affecting the brachial plexus or its terminal branches. Despite this, in the acute setting, CT myelography remains the primary mode of radiographic evaluation for nerve root avulsion. Electrodiagnostic Studies Electrodiagnostic studies are integral to both preoperative and intraoperative decision-making. They help in confirming a diagnosis, localizing lesions, defining the severity of axon loss and the completeness of a lesion, eliminating other conditions from the differential diagnosis, and revealing subclinical recovery or unrecognized subclinical disorders. Electrodiagnostic studies are an important adjunct to a thorough history, physical examination, and imaging studies, not a substitute for them. For closed injuries, baseline electromyography (EMG) and nerve conduction velocity (NCV) studies are best performed 3 to 4 weeks after injury because wallerian degeneration will have occurred by then. Serial electrodiagnostic studies can be done every few months in conjunction with a repeat physical examination to document and quantify ongoing reinnervation or denervation. EMG tests muscles at rest and with activity. Denervational changes (ie, fibrillation potentials) in different muscles can be seen in proximal muscles as early as 10 to 14 days after injury (and in 3 to 6 weeks in more distal muscles). Reduced recruitment of motor unit potentials can be demonstrated immediately after weakness occurs from lower motor neuron injury. The presence of active motor units with voluntary effort and few fibrillations at rest offers a good prognosis compared with the absence of motor units and many fibrillations. EMG may help distinguish preganglionic from postganglionic lesions by needle exami-

Presence of a pseudomeningocele (asterisks) indicates greater likelihood of a nerve root avulsion. A, Anteroposterior myelogram demonstrating multiple root avulsions (asterisks). B, Those avulsions (asterisk) are clearly seen on axial CT myelogram. The arrows on the opposite side of the avulsion demonstrate the normal dorsal and ventral rootlet outline of the uninjured side. These outlines are missing on the injured side. (Reproduced by permission of Mayo Foundation.)

formed. Vascular injuries are not infrequent findings with infraclavicular lesions or with even more severe injuries, such as scapulothoracic dissociation. Radiographic Evaluation After a traumatic injury to the neck or shoulder girdle, radiographic examination should include views of the cervical spine, shoulder (anteroposterior and axillary views), and chest. The spine radiographs should determine the presence of any associated cervical fractures that could put the spinal cord at risk. Transverse process fractures in the cervical vertebrae may suggest root avulsion at the same level. Clavicle or rib fractures (first or second rib) may indicate trauma to the brachial plexus. Chest radiographs may reveal old rib fractures, which are important should intercostal nerves be considered for nerve transfer (rib fractures often injure the associated intercostal nerves). Additionally, phrenic nerve injury causes associated paralysis of the hemidiaphragm. When vascular injury is suspected, arteriography or magnetic resonance
386

angiography may be indicated to confirm the patency of a previous vascular repair or reconstruction. Computed tomography (CT) combined with myelography has been instrumental in helping to define the level of nerve root injury.14-16 With an avulsion of a cervical root, the dural sheath heals with development of a pseudomeningocele. Immediately after injury, blood clot is often present in the area of the nerve root avulsion and can displace dye from the myelogram. Therefore, a CT myelogram should be done 3 to 4 weeks after injury to allow time for blood clots to dissipate and for pseudomeningoceles to fully form. A pseudomeningocele on CT myelogram is highly suggestive of a root avulsion (Fig. 3). Magnetic resonance imaging (MRI) may be useful in evaluating patients with a suspected nerve root avulsion,17-19 and it has some advantages over CT myelogram. MRI can visualize much of the brachial plexus, whereas CT myelography demonstrates only nerve root injury. Additionally, MRI can demonstrate large neuromas after trauma or asso-

Journal of the American Academy of Orthopaedic Surgeons

Alexander Y. Shin, MD, et al

nation of proximally innervated muscles that are innervated by root level motor branches (eg, cervical paraspinals, rhomboids, serratus anterior). NCV studies are performed along with EMG. In posttraumatic brachial plexus lesions, the amplitudes of compound muscle action potentials (CMAPs) are generally low. Sensory nerve action potentials (SNAPs) are important in localizing a lesion as preganglionic or postganglionic. SNAPs are preserved in lesions proximal to the dorsal root ganglia. Because the sensory nerve cell body is intact and within the dorsal root ganglion, NCV studies often demonstrate that the SNAP is normal, when clinically the patient is insensate in the associated nerve sensory distribution. SNAPs are absent in a postganglionic or a combined pre- and postganglionic lesion. For example, a patient with a normal SNAP in the ulnar nerve, with an insensate ulnar nerve distribution, has avulsions (preganglionic injury) of the C8-T1 roots. There are limitations to electrodiagnostic studies. The EMG/NCV study is only as good as the experienced physician who is performing the study and interpreting the results. EMG may demonstrate evidence of early recovery in muscles (eg, emergence of nascent potentials, a decreased number of fibrillation potentials, or the appearance of or an increased number of motor unit potentials); these findings may predate clinically apparent recovery by weeks to months. However, EMG recovery does not always equate with clinically relevant recovery either in terms of quality of regeneration or extent of recovery. EMG recovery merely indicates that an unknown number of fibers have reached muscles and have established motor end plate connections. Conversely, evidence of reinnervation may not be detected on EMG in complete lesions, despite ongoing regeneration, when target end organs are more distal.
Volume 13, Number 6, October 2005

Intraoperative electrodiagnostic studies also may play a part in brachial plexus surgery. A combination of intraoperative electrodiagnostic techniques can be used to maximize the information gathered before making a surgical decision. These techniques routinely include nerve action potentials (NAPs) and somatosensory evoked potentials (SSEPs), as well as CMAPs. NAPs allow the surgeon to test a nerve directly across a lesion to detect reinnervation months before conventional EMG techniques would demonstrate activity and to determine whether a lesion is neurapractic (negative NAP) or axonotmetic (positive NAP). The presence of a NAP across a lesion indicates preserved axons or significant regeneration. Primate studies have suggested that the presence of a NAP indicates the viability of thousands of axons rather than the hundreds seen with other techniques.20 The presence of a NAP suggests that recovery will occur after neurolysis alone without the need for additional treatment (eg, neuroma resection and grafting). More than 90% of patients with a preserved NAP will gain clinically useful recovery.20 NAPs indirectly can help distinguish between preand postganglionic injury. A faster conduction velocity with large amplitude and short latency, together with severe neurologic loss, indicate a preganglionic injury. A flat tracing suggests that adequate regeneration is not occurring; this is consistent with either a reparable postganglionic lesion or an irreparable combined pre- and postganglionic lesion. With the latter, sectioning the nerve back to an intraforaminal level would not reveal good fascicular structure.20 Intraoperative somatosensoryevoked potentials (SSEPs) are also used during brachial plexus surgery. The presence of an SSEP suggests continuity between the peripheral nervous system and the central nervous system via a dorsal root. A positive response is determined by the

integrity of a few hundred intact fibers. The actual state of the ventral root is not tested directly with this technique. Instead, it is inferred from the state of the sensory nerve rootlets, even though perfect correlation between dorsal and ventral root avulsions does not always exist. SSEPs are absent in postganglionic or combined pre- and postganglionic lesions. Motor-evoked potentials assess the integrity of the motor pathway via the ventral root. This technique, which uses transcranial electrical stimulation, has recently been approved in the United States.21 CMAPs are not useful intraoperatively in complete distal lesions because of the time required for regeneration to occur into distal muscles. However, CMAPs are useful in partial lesions because the size of the lesion is proportional to the number of functioning axons.

Concepts of Surgical Management

The three most important concepts in the surgical management of brachial plexus injuries are patient selection, the exact timing of surgery, and the prioritization of restoration of function in the upper arm. Surgery should be performed in the absence of clinical or electrical evidence of recovery or when spontaneous recovery is impossible. Despite the improvements in electrodiagnostic studies and imaging, selecting when and on whom to operate remains one of the most difficult decisions in peripheral nerve surgery. During the observation period, physical therapy should be performed to prevent contractures and to strengthen functioning muscles. Timing of surgery or intervention depends on the mechanism of injury as well as the type of injury. Immediate exploration and primary repair of the injured portion of the brachial plexus is indicated in sharp open injuries. This facilitates end-to-end repair of the injured nerves. When
387

Adult Traumatic Brachial Plexus Injuries

Figure 4

duction and stability, hand sensibility, wrist extension and finger flexion, wrist flexion and finger extension, and intrinsic function of the hand.

Surgery
Brachial plexus surgery can be divided into primary and secondary reconstruction. Primary reconstruction is the initial surgical management and may include nerve surgery/reconstruction (eg, direct repair, neurolysis, nerve grafting, nerve transfers) and/or soft-tissue procedures (eg, free functioning muscle transfer). Secondary reconstruction may be necessary to improve function, either to augment partial recovery or to obtain function when none has been achieved. This may include softtissue reconstruction (eg, tendon/ muscle transfer, free muscle transfer) and bony procedures (eg, arthrodesis, osteotomy), but typically not nerve surgery. Often a combination of these techniques can be used, necessitating a broad surgical armamentarium. Primary Reconstruction Direct repair of nerve ends can be done after sharp injuries (eg, lacerations), but it cannot be applied to stretch injuries. External neurolysis is a necessary prerequisite for intraoperative electrical studies. Neurolysis alone may be performed when the nerve is in continuity and a NAP is obtained.22
Intraplexal Nerve Grafting

Intraplexal nerve grafting with donor nerves can be performed in the setting of postganglionic injury with viable nerve root stumps available. With postganglionic injuries on C5, C6, and C7, nerve grafts can be used to target shoulder abduction (C5 to the suprascapular nerve [A] and posterior division of the upper trunk [B]), elbow flexion (C6 to the anterior division of the upper trunk [C]), and wrist extension and elbow extension (C7 to the posterior division of the middle trunk, targeting radial nerve function [D]). SSN = suprascapular nerve. (Adapted by permission of Mayo Foundation.)

the open injury is secondary to a blunt object with avulsion of the nerve, the ends of the lacerated nerve should be tagged and a delayed repair performed 3 to 4 weeks later. By 3 to 4 weeks, the injured nerve ends will have demarcated, enabling better access to the zone of nerve injury. Low-velocity gunshot wounds should be observed because most of these injuries are neurapraxic; however, high-velocity gunshot wounds are associated with significant softtissue damage and usually mandate surgical exploration. For stretch injuries, the exact timing of surgery is more controversial. The timing is determined somewhat by the mechanism and type of injury, physical examination and imaging findings, and surgeon preference. Operating early may not allow sufficient
388

time for spontaneous reinnervation, but waiting too long before operating may unnecessarily lead to failure of the motor end plate and thus failure of reinnervation. Early exploration and reconstruction (between 3 and 6 weeks) is indicated when there is a high suspicion of root avulsion. Routine exploration is performed 3 to 6 months after injury in patients who have not demonstrated adequate reinnervation. Results from delayed (6 to 12 months) or late (>12 months) surgery are poorer because the time for the nerve to regenerate to the target muscles is greater than the survival time of the motor end plate after deenervation. Most surgeons consider elbow flexion the highest priority when restoring function to the flail extremity. Next in priority are shoulder ab-

Nerve grafting can be performed with ruptures or postganglionic neuromas that do not conduct a NAP across the lesion. In such cases, the nerve rootbecause of its connection to the spinal cordhas maintained viable motor axons that can be grafted to specific targets. Interpositional grafts (typically using cable grafts of sural or other cutaneous nerves) are coapted between nerve stumps without undue tension. For example, C5 is targeted for shoulder abduction (suprascapular nerve, axil-

Journal of the American Academy of Orthopaedic Surgeons

Alexander Y. Shin, MD, et al

Figure 5

Neurotization for shoulder abduction with the spinal accessory nerve29 (A) or the phrenic nerve24 (B) can be performed in the supraclavicular exposure. (Adapted by permission of Mayo Foundation.)

lary nerve), C6 for elbow flexion (musculocutaneous nerve), and C7 for elbow extension and wrist extension (radial nerve)22 (Fig. 4).
Nerve Transfer (Neurotization)

Nerve transfer can be performed for preganglionic injury or to accelerate recovery by reducing the time for reinnervation by decreasing the distance between the site of nerve repair and the end organ. A functioning nerve of lesser importance is transferred to the more important denervated distal nerve. Ideally, nerve transfers should be performed within 6 months of injury; however, even after the preferred 6-month time frame, nerve transfers may be more suitable than grafting because nerve transfers have faster recovery than grafting. Several donor nerves are sources for neurotization. Some of the more common include the spinal accessory nerve (cranial nerve XI), intercostal nerves (motor and sensory), and medial pectoral nerve. More recentVolume 13, Number 6, October 2005

ly, using a fascicle of a functioning ulnar nerve (Oberlin transfer) or the median nerve in patients with intact C8 and T1 nerves has allowed rapid and powerful return of elbow flexion, with 94% of patients achieving M4 strength.23 The phrenic nerve24 and the contralateral C7 (or hemicontralateral C7)25 nerve also have been used to expand the pool of extraplexal donors and to improve outcomes. The deep cervical plexus and hypoglossal nerve (cranial nerve XII) have been used, but poor motor recovery has been reported.26 The average number of myelinated axons in these donor nerves varies. The spinal accessory nerve has approximately 1,700 axons; the phrenic nerve, 800 axons; a single intercostal motor nerve, 1,300 axons; and the contralateral C7 nerve, 23,780 axons.26 The goal is to maximize the number of myelinated axons per target function and minimize donor site morbidity. Several series have reported an acceptable morbidity with transfer of the con-

tralateral C7 and phrenic nerves, but long-term studies are not available.25,27 Neurotization for shoulder abduction can be easily obtained by nerve transfer of either the spinal accessory nerve or the phrenic nerve to the suprascapular nerve.24,26,28 The benefit of these two transfers is that no additional interposition nerve grafts are needed, and a direct coaptation of the nerves is possible (Fig. 5). When additional nerve sources are available, neurotization of the axillary nerve (nerve grafting from C5) is recommended to provide further shoulder stability and abduction. Neurotization for elbow flexion can be performed using either intercostal nerves (Fig. 6) directly or the spinal accessory nerve with an interpositional graft29 directly targeting the biceps motor branch (rather than the entire musculocutaneous nerve). Separating the biceps motor branch from the lateral antebrachial cutaneous nerve in a retrograde manner allows the maximum number of mo389

Adult Traumatic Brachial Plexus Injuries

Figure 6

Neurotization for elbow flexion with intercostal nerves. The motor branches from the intercostal nerves can be easily harvested and neurotized to the motor branch of the musculocutaneous nerve to the biceps. (Adapted by permission of Mayo Foundation.)

tor axons to be transferred directly to the biceps muscle. This also helps gain length for the transfer, thus eliminating the need for interpositional grafts in the case of intercostal nerves and shortening the length of the graft for the accessory nerve. Some have advocated using the phrenic nerve with an interpositional graft to the musculocutaneous nerve.24 In the event of an upper trunk avulsion injury, two popular options exist for restoring elbow flexion. The medial pectoral nerve may be transferred to the musculocutaneous nerve or the biceps branch.26 Alternatively, a fascicle from the ulnar nerve (Oberlin transfer) can be transferred to the motor branch of the biceps with excellent results (94% of patients achieved M4 strength)23 (Fig. 7). Before separating the fascicles from the ulnar nerve, they are tested with a nerve stimulator. Fascicles that stimulate the intrinsic muscles of the hand are avoided; those that stimulate wrist flexion (flexor carpi ulnaris) are chosen.

This technique is an excellent alternative to the intercostal neurotizations or spinal accessory nerve with an interpositional graft. The contralateral C7 or a hemicontralateral C7 nerve can be used via a vascularized ulnar nerve graft (in the case of a complete plexus avulsion injury) or via sural nerve grafts to bring a large number of motor axons to the injured side.25,27 When used with the vascularized ulnar nerve graft, the contralateral or hemicontralateral C7 nerve can be used to innervate the median nerve, with the goal of obtaining useful finger flexion (29% of patients achieved M3 or M4 finger flexion) and protective sensation in the median nerve distribution (81% of patients)27 (Fig. 8).
Outcomes of Nerve Transfers

Neurotization for elbow flexion and shoulder stability has been shown to be an effective means of restoring muscle function.28 In a critical meta-analysis of the Englishlanguage literature, Merrell et al28

Figure 7

A, When the ulnar nerve is normal (ie, upper trunk injury sparing C8 and T1), a fascicle can be transferred to the motor branch of the biceps to obtain elbow flexion. Top left: Transection (dashed line) of the motor branch to the biceps muscle. Top center: Fascicle(s) obtained from normal ulnar nerve (dashed line). Top right: Fascicle(s) transferred to the motor branch of the musculocutaneous nerve. B, Intraoperative photograph demonstrating the fascicle from the ulnar nerve transferred to the motor branch of the biceps. MC n = musculocutaneous nerve, Transferred fascicle = portion of ulnar nerve, Ulnar n = ulnar nerve. (Part A adapted by permission of Mayo Foundation. Part B reproduced by permission of Mayo Foundation.) 390
Journal of the American Academy of Orthopaedic Surgeons

Alexander Y. Shin, MD, et al

Figure 8

Contralateral C7 (or a hemicontralateral C7 [A and B]) nerve transfer via a vascularized ulnar nerve graft (in cases of complete C5-T1 avulsions) can be used to bring a large number of motor axons into the injured side. The hemicontralateral C7 transfer can be used effectively with a vascularized ulnar nerve graft to reinnervate the median nerve for finger flexion and sensation. (A) The portion of the C7 (contralateral) that primarily innervates pectoral function is isolated, and half of the nerve is isolated (B). The ipsilateral distal ulnar nerve is coapted with the hemicontralateral portion of C7 (C). The proximal ulnar nerve (*) is divided (dashed line). The injured side median nerve (D) is identified and divided (dashed line). The proximal ulnar nerve is transferred to the distal median nerve stump of the injured side (E). (Adapted by permission of Mayo Foundation.)

evaluated the results of 1,088 nerve transfers in 27 studies to determine the outcome of nerve transfers of the shoulder and elbow. For restoration of elbow flexion, 26 studies with a total of 965 nerve transfers were evaluated. Overall, 71% of transfers to the musculocutaneous nerve achieved M3 (antigravity) flexion on the Medical Research Council grading scale, and 37% achieved M4 (against gravity, not normal) flexion. The two most common donor nerves were the intercostal (54%) and spinal accessory (39%). Overall, the intercostal achieved M3 in 72% of patients. When an interpoVolume 13, Number 6, October 2005

sition nerve graft was used, only 47% achieved M3 strength. When the spinal accessory nerve was transferred to the musculocutaneous nerve, 77% of patients had restoration of elbow flexion M3 and 29% had restoration of function M4. Use of the Oberlin transfer (two fascicles of the ulnar nerve transferred to the musculocutaneous nerve) resulted in 97% M3 flexion and 94% M4 flexion.28 For restoration of shoulder abduction, 8 studies with a total of 123 transfers were evaluated. Overall, 73% of patients achieved M3 shoulder abduction, and 26% achieved M4 abduction. The spinal accessory

nerve was used in 41% of transfers and the intercostal nerves in 26%. The spinal accessory nerve performed significantly (P < 0.001) better than the intercostals in achieving M3 abduction (98% and 56%, respectively). However, even with good results, shoulder abduction reached only 45. Further research is still needed in the field of outcomes analysis of brachial plexus injuries. Unfortunately, it is not known which treatment produces the best outcomes for C5 and C6 ruptures or severe neuromas. To be determined, for example, is whether it is best to graft from C5
391

Adult Traumatic Brachial Plexus Injuries

Figure 9

Free gracilis muscle transfer for elbow flexion. The proximal end of the gracilis is secured to the clavicle (A). The vascular inflow and outflow is via the thoracoacromial trunk, and the muscle is powered by the spinal accessory or intercostal motor nerves (inset). Distally, the gracilis tendon is woven into the biceps tendon (B). (Adapted by permission of Mayo Foundation.)

and C6, or whether nerve transfers should be performed closer to the end organ.
Free Functioning Muscle Transfer

Advances in microsurgical techniques have led to innovations in surgical reconstruction of the upper extremity following brachial plexus injury. Free functioning muscle transfer is the transplantation of a muscle and its neurovascular pedicle to a new location to assume a new function. The muscle is powered by neurotizing the motor nerve by a do392

nor motor nerve; circulation is restored to the transferred muscle via microsurgical anastomosis of the donor and recipient vessels. Within several months, the transferred muscle begins to become reinnervated by the donor nerve; it eventually begins to contract and then gains independent function. Free functioning muscle transfers were first done either in patients who presented late (>12 months after injury) or as a salvage procedure in those who had failed earlier nerve reconstruction.30-32 Based on the success with free muscle transfer in

these late situations, it has been incorporated into a strategy for early reconstruction in patients for whom other donors of tendon transfers are unavailable. Most commonly, free muscle transfer is used to provide reliable elbow flexion.5,33-37 A variety of free functioning muscles can be transferred, including the latissimus dorsi (thoracodorsal nerve), the rectus femoris (femoral nerve), and the gracilis (anterior division of the obturator nerve). The gracilis has become one of the most commonly used muscles in brachial plexus reconstruction because of its proximally based neurovascular pedicle (which allows earlier reinnervation) and its long tendon length (which reaches into the forearm for hand reanimation). The gracilis can be used for restoring biceps function,38,39 wrist extension, or finger flexion, or as a double muscle transfer in the two-stage Doi procedure.5,6,40 For elbow flexion, it is desirable to place the major vascular pedicle of the gracilis in proximity to the thoracoacromial trunk in the infraclavicular fossa. The proximal gracilis tendon is passed beneath the clavicle and secured to its superior border. It is then tunneled subcutaneously to the antecubital fossa in preparation for securing it to the biceps tendon. The obturator nerve branch to the gracilis may be repaired to the spinal accessory nerve or to the two intercostal motor nerves, either of which should be harvested at as great a length as possible to allow direct nerve repair distal to the clavicle. Distally, the gracilis tendon is woven into the biceps tendon (Fig. 9). Doi et al5 recently described using a double free functioning gracilis muscle transfer, which has provided hope of achieving prehension to patients with complete brachial plexus lesions. The goals of this two-stage operation are to restore both elbow flexion and extension as well as wrist extension and finger flexion. In stage I, the brachial plexus is explored and

Journal of the American Academy of Orthopaedic Surgeons

Alexander Y. Shin, MD, et al

a free functioning gracilis muscle is harvested and neurotized by the spinal accessory nerve (Fig. 10). The gracilis is proximally attached to the clavicle and routed distally under the brachioradialis to the radial wrist digit extensors. The gracilis vessels are anastamosed to the thoracoacromial artery and venae comitantes or other available venous outflow. Stage II is performed approximately 6 to 8 weeks after stage I (Fig. 11). The second gracilis muscle and the motor and sensory intercostal nerves from the third to sixth intercostal spaces are harvested. The gracilis is attached proximally to the second rib, routed subcutaneously along the medial side of the arm, and attached to the flexor tendons. It is then neurotized with two of the motor intercostal nerves, while the sensory intercostals are neurotized to the median nerve to provide hand sensation. The gracilis is vascularized by the thoracodorsal vessels. The remaining two motor intercostal nerves are neurotized to the radial nerve innervating the triceps. At our institution, we have slightly modified the double free muscle transfer as originally described by Doi and colleagues.5,6,34,35,37,39-42 We prefer to secure the stage I gracilis muscle to wrist extensors rather than to finger extensors, believing that this helps promote finger flexion through a tenodesis effect. To create a more effective pulley with diminished bowstringing of the tendon, we route the gracilis tendon posterior to the biceps tendon and underneath the brachioradialis muscle. A more effective pulley should improve muscle excursion and strengthen wrist extension. Using a double free muscle transfer, Doi et al5 restored excellent to good elbow flexion in 96% of patients (25/26). At our institution, eight patients have been followed for at least 1 year after the stage II transfer. Transfer for combined elbow flexion and wrist extension lowered the overall results for elbow flexion strength compared with elbow flexVolume 13, Number 6, October 2005

Figure 10

Stage I Doi free gracilis muscle transfer.5 A free functioning gracilis is harvested and neurotized by the spinal accessory nerve and anastamosed to the thoracoacromial trunk. The gracilis is attached proximally to the clavicle (A) and routed distally under the brachioradialis and flexor carpi ulnaris pulley (B), then woven into the wrist extensors (C). (Adapted by permission of Mayo Foundation.)

ion alone. Seventy-nine percent of the free functioning muscle transfers for elbow flexion alone (single transfer) and 63% of similarly innervated muscles transferred for combined motion (stage I Doi procedure) achieved M4 elbow flexion strength (P > 0.05). This is not surprising in that the muscles must use some of their strength and excursion to extend the wrist or digits and invariably lose some effect because of bowstringing at the elbow.43 Grasp function in the double free muscle procedure relies on recovery

of some triceps function to stabilize the elbow during contraction of the gracilis muscle. Grasp function also relies on adequate muscle strength and the absence of significant adhesions. In the series of Doi et al,5 65% (17 patients) achieved >30 of total active motion of the fingers with the second muscle transfer. Such function allows only rudimentary grasp in many patients, but grasp function is difficult to achieve with other methods. Previous efforts at restoring prehension in the setting of a brachial plexus injury have been un393

Adult Traumatic Brachial Plexus Injuries

Figure 11

Stage II Doi free gracilis muscle transfer.5 A, Vascularity is supplied via the thoracodorsal artery and vein. The gracilis muscle is attached proximally to the second rib. B, The gracilis is neurotized with two of the intercostal motor nerves (1), while the sensory intercostal nerves are neurotized to the median nerve for hand sensation (2). The remaining two motor intercostals are used to neurotize the triceps (3). Lat = lateral cord, M = motor intercostal nerve, Med = medial cord, Post = posterior cord, S = sensory intercostal nerve. C, Intraoperative photograph. Motor and sensory intercostal nerves are harvested from beneath the third to sixth ribs. (Parts A and B adapted by permission of Mayo Foundation. Part C reproduced by permission of Mayo Foundation.)

successful because of the long distance between nerve repair and motor end plates and the resultant prolonged reinnervation time. Therefore, these results must still be regarded as an advance in these otherwise irreparable avulsion injuries. The only alternative currently
394

offering the possibility of hand function in root avulsion injuries is combining the contralateral C7 nerve with a vascularized ulnar nerve conduit.27
Postoperative Follow-up

After any surgical intervention of

the brachial plexus, it is imperative that the patient and his or her family understand that the recovery of nerve function is a slow and arduous process. The newly reapproximated nerve graft or nerve transfer grows at rate of 1 mm a day or 1 inch per month. For long transfers, such as a

Journal of the American Academy of Orthopaedic Surgeons

Alexander Y. Shin, MD, et al

hemicontralateral C7 nerve graft, clinical results may not be seen for 2 to 3 years. The shorter the distance to the target muscle, the more rapid the time to reinnervation. It is essential that the patient, while waiting for reinnervation, be enrolled in physical therapy to keep the joints of the upper extremity supple and to prevent joint contractures. This can be done with custom-made plastic resting splints as well as a daily range-of-motion protocol for the shoulder, elbow, wrist, and digits. The efficacy of electrical stimulation in preserving muscle (motor) end plates remains controversial. However, its use is psychologically beneficial to patients who like to see muscles contract during the long recovery period. Dedicated follow-up at 3- to 4-month intervals for a minimum of 2 to 3 years (preferably 5 years) is recommended to assess for full recovery. Secondary Reconstruction Secondary reconstruction should be considered when there has been no further recovery of motor function or when function can be further improved or refined with a relatively minor surgical intervention. Secondary reconstructive options include tendon transfer,9 free functioning muscle transfer,5,36 shoulder arthrodesis,10 and wrist and hand arthrodesis.10 Tendon transfer is commonly delayed until maximal motor recovery has occurred. Such transfers represent a spectrum of procedures that provide a gratifying result to a patient who has made a partial recovery and who would benefit from additional function. Free functioning muscle transfer can be performed as secondary reconstructive surgery to improve the strength of a weakly reinnervated biceps or triceps muscle or of finger flexors (provided that the joints are supple). Finally, arthrodesis may be useful for secondary reconstructive surgery of the shoulder, wrist, and hand.
Volume 13, Number 6, October 2005

Shoulder fusion can be performed as a salvage procedure for the persistently painful subluxating shoulder should the nerve surgery fail to result in shoulder stability. Shoulder fusion as a primary reconstructive technique is less frequently done because recent studies have shown that patients prefer voluntary shoulder abduction when it can be achieved through nerve reconstruction.44 Other bony procedures, such as humeral rotational osteotomy, thumb axis arthrodesis, bone-block opponensplasty, or finger joint arthrodesis, can improve function.

4.

5.

6.

7.

Summary
Injuries to the adult brachial plexus are often intimidating to the orthopaedic surgeon who may be managing concomitant injuries. The injury can be devastating to the patient and is often difficult for the patient and family to comprehend. A thorough understanding of the anatomy, clinical evaluation, radiographic and electrodiagnostic studies, treatment options, and proper timing of surgical intervention will enable the treating surgeon to offer optimal care. Surgical options include neurolysis, nerve grafting, neurotization, and free muscle transfers. These treatment options offer patients with brachial plexus injuries the ability to obtain elbow flexion, limited shoulder abduction with shoulder stability, and hope for limited but potentially useful hand function.
8.

9.

10.

11.

12. 13.

14.

References
1.

15.

Allieu Y, Cenac P: Is surgical intervention justifiable for total paralysis secondary to multiple avulsion injuries of the brachial plexus? Hand Clin 1988;4:609-618. 2. Azze RJ, Mattar Junior J, Ferreira MC, Starck R, Canedo AC: Extraplexual neurotization of brachial plexus. Microsurgery 1994;15:28-32. 3. Brandt KE, Mackinnon SE: A technique for maximizing biceps recovery in brachial plexus reconstruction. J Hand Surg [Am] 1993;18:726-733.

16.

17.

Brunelli G, Monini L: Direct muscular neurotization. J Hand Surg [Am] 1985;10(6 pt 2):993-997. Doi K, Muramatsu K, Hattori Y, et al: Restoration of prehension with the double free muscle technique following complete avulsion of the brachial plexus: Indications and long-term results. J Bone Joint Surg Am 2000;82: 652-666. Doi K, Kuwata N, Muramatsu K, Hottori Y, Kawai S: Double muscle transfer for upper extremity reconstruction following complete avulsion of the brachial plexus. Hand Clin 1999;15: 757-767. Malone JM, Leal JM, Underwood J, Childers SJ: Brachial plexus injury management through upper extremity amputation with immediate postoperative prostheses. Arch Phys Med Rehabil 1982;63:89-91. Allieu Y: Evolution of our indications for neurotization: Our concept of functional restoration of the upper limb after brachial plexus injuries. Chir Main 1999;18:165-166. Narakas AO: The treatment of brachial plexus injuries. Int Orthop 1985;9: 29-36. Hendry AM: The treatment of residual paralysis after brachial plexus injuries. J Bone Joint Surg Br 1949;31:4249. Yeoman PM, Seddon HJ: Brachial plexus injuries: Treatment of the flail arm. J Bone Joint Surg Br 1961;43:493500. Fletcher I: Traction lesions of the brachial plexus. Hand 1969;1:129-136. Kerr AT: The brachial plexus of nerves in man, the variations in its formation and branches. Am J Anat 1918;23:285395. Carvalho GA, Nikkhah G, Matthies C, Penkert G, Samii M: Diagnosis of root avulsions in traumatic brachial plexus injuries: Value of computerized tomography myelography and magnetic resonance imaging. J Neurosurg 1997;86:69-76. Nagano A, Ochiai N, Sugioka H, Hara T, Tsuyama N: Usefulness of myelography in brachial plexus injuries. J Hand Surg [Br] 1989;14:59-64. Walker AT, Chaloupka JC, de Lotbiniere AC, Wolfe SW, Goldman R, Kier EL: Detection of nerve rootlet avulsion on CT myelography in patients with birth palsy and brachial plexus injury after trauma. AJR Am J Roentgenol 1996;167:1283-1287. Doi K, Otsuka K, Okamoto Y, Fujii H, Hattori Y, Baliarsing AS: Cervical nerve root avulsion in brachial plexus

395

Adult Traumatic Brachial Plexus Injuries injuries: Magnetic resonance imaging classification and comparison with myelography and computerized tomography myelography. J Neurosurg 2002;96:277-284. Gupta RK, Mehta VS, Banerji AK, Jain RK: MR evaluation of brachial plexus injuries. Neuroradiology 1989;31: 377-381. Nakamura T, Yabe Y, Horiuchi Y, Takayama S: Magnetic resonance myelography in brachial plexus injury. J Bone Joint Surg Br 1997;79:764-769. Tiel RL, Happel LT Jr, Kline DG: Nerve action potential recording method and equipment. Neurosurgery 1996;39:103-109. Burkholder LM, Houlden DA, Midha R, Weiss E, Vennettilli M: Neurogenic motor evoked potentials: Role in brachial plexus surgery. Case report. J Neurosurg 2003;98:607-610. Kim DH, Cho YJ, Tiel RL, Kline DG: Outcomes of surgery in 1019 brachial plexus lesions treated at Louisiana State University Health Sciences Center. J Neurosurg 2003;98:10051016. Teboul F, Kakkar R, Ameur N, Beualieu JY, Oberlin C: Transfer of fascicles from the ulnar nerve to the nerve to the biceps in the treatment of upper brachial plexus injury. J Bone Joint Surg Am 2004;86:1485-1490. Gu YD, Ma MK: Use of the phrenic nerve for brachial plexus reconstruction. Clin Orthop 1996;323:119-121. Gu YD, Chen DS, Zhang GM, et al: Long-term functional results of contralateral C7 transfer. J Reconstr Microsurg 1998;14:57-59. Chuang DC: Neurotization procedures for brachial plexus injuries. Hand Clin 1995;11:633-645. Songcharoen P, Wongtrakul S, Mahaisavariya B, Spinner RJ: Hemicontralateral C7 transfer to median nerve in the treatment of root avulsion brachial plexus injury. J Hand Surg [Am] 2001;26:1058-1064. Merrell GA, Barrie KA, Katz DL, Wolfe SW: Results of nerve transfer techniques for restoration of shoulder and elbow function in the context of a meta-analysis of the English literature. J Hand Surg [Am] 2001;26:303314. Songcharoen P, Mahaisavariya B, Chotigavanich C: Spinal accessory neurotization for restoration of elbow flexion in avulsion injuries of the brachial plexus. J Hand Surg [Am] 1996; 21:387-390. Ruch DS, Friedman A, Nunley JA: The restoration of elbow flexion with intercostal nerve transfers. Clin Orthop 1995;314:95-103. Mikami Y, Nagano A, Ochiai N, Yamamoto S: Results of nerve grafting for injuries of the axillary and suprascapular nerves. J Bone Joint Surg Br 1997;79:527-531. Chuang DC, Yeh MC, Wei FC: Intercostal nerve transfer of the musculocutaneous nerve in avulsed brachial plexus injuries: Evaluation of 66 patients. J Hand Surg [Am] 1992;17:822828. Akasaka Y, Hara T, Takahashi M: Free muscle transplantation combined with intercostal nerve crossing for reconstruction of elbow flexion and wrist extension in brachial plexus injuries. Microsurgery 1991;12:346-351. Doi K, Sakai K, Fuchigami Y, Kawai S: Reconstruction of irreparable brachial plexus injuries with reinnervated free-muscle transfer: Case report. J Neurosurg 1996;85:174-177. Doi K, Shigetomi M, Kaneko K, et al: Significance of elbow extension in reconstruction of prehension with reinnervated free-muscle transfer following complete brachial plexus avulsion. Plast Reconstr Surg 1997; 100:364-374. Manktelow RT, Zuker RM, McKee NH: Functioning free muscle transplantation. J Hand Surg [Am] 1984; 9A:32-39. Doi K, Sakai K, Kuwata N, Ihara K, Kawai S: Double free-muscle transfer to restore prehension following complete brachial plexus avulsion. J Hand Surg [Am] 1995;20:408-414. Chung DC, Carver N, Wei FC: Results of functioning free muscle transplantation for elbow flexion. J Hand Surg [Am] 1996;21:1071-1077. Doi K, Sakai K, Ihara K, Abe Y, Kawai S, Kurafuji Y: Reinnervated free muscle transplantation for extremity reconstruction. Plast Reconstr Surg 1993;91:872-883. Doi K: New reconstructive procedure for brachial plexus injury. Clin Plast Surg 1997;24:75-85. Doi K, Sakai K, Kuwata N, Ihara K, Kawai S: Reconstruction of finger and elbow function after complete avulsion of the brachial plexus. J Hand Surg [Am] 1991;16:796-803. Doi K, Hattori Y, Kuwata N, et al: Free muscle transfer can restore hand function after injuries of the lower brachial plexus. J Bone Joint Surg Br 1998;80: 117-120. Barrie KA, Steinmann SP, Shin AY, Spinner RJ, Bishop AT: Gracilis free muscle transfer for restoration of function after complete brachial plexus avulsion. Neurosurg Focus 2004;16:E8. Doi K, Hattori Y, Ikeda K, Dhawan V: Significance of shoulder function in the reconstruction of prehension with double free-muscle transfer after complete paralysis of the brachial plexus. Plast Reconstr Surg 2003;112:15961603.

28.

36.

18.

37.

19.

29.

38.

20.

30.

39.

21.

31.

40.

22.

41.

32.

23.

42.

33.

24.

43.

25.

34.

44.

26.

35.

27.

396

Journal of the American Academy of Orthopaedic Surgeons