Historical

The first surgical repair of the BP was attempted in the last decade of the 19th century after successful attempts at nerve suturing,^3,4 experimental work on nerve grafts,⁵ and the introduction of this method to clinical surgery.⁶ The repair techniques commonly applied were neurolysis, direct repair with epineural sutures, and coaptations wrapped in a membrane. Over the years, surgeons have used different kinds of suture material for nerve repair, including animal and human hair, animal tendon, fascia, linen, silk, cotton, and polyester. Various aqueous or gel media have been used for adherence of nerve ends difficult to suture. BP lesions became a separate clinical entity among peripheral nerve lesions in the first half of the 19th century. At autopsy, Flaubert described rupture of the C5 spinal nerve with avulsion of C6 through T1 caused by reduction of a dislocated shoulder.⁷ Subsequently, BP management of 8 cases in wounded soldiers during the American Civil War⁸ and 24 European cases⁹ was reported.

Secondary suture of a traction injury to the BP 7 months after trauma was initially carried out in the beginning of the 20th century.¹⁰ The first nerve transfer consisted of implanting the distal stump of the damaged spinal nerve C5 into the healthy C6.¹¹ Unfortunately, the results of these first surgical attempts at repair are not known. At that time, however, the results of surgery were inconsistent and often poor. BP surgery was performed only by a small group of surgeons,^12–15 and several deaths related to surgery occurred.^14,16 Before and during World War I, nerves were regarded as simple cord-like structures. Severed nerves were repaired by simply restoring continuity in the expectation that nerve regeneration would then restore function. Procedures to close large gaps under considerable tension were still preferred over nerve grafting. Throughout this period, infection remained a problem, and the results of nerve repair continued to be disappointing.

Surgical treatment was more or less abandoned in the 1920s, and during the next several decades, general pessimism was present regarding the value of surgical reconstruction of closed BP traction lesions. No clear distinction had been made between severe BP injuries with root avulsion not amenable to repair, trunk and cord rupture that could benefit from surgery, and milder cases with a lesion in continuity in which at least acceptable spontaneous restoration of function was likely to be achieved. This shortcoming was due to the similarity of the clinical picture in severe and mild cases, and diagnostic techniques were not available to distinguish one from the other. A comparison between patients treated conservatively and those treated by nerve repair was not yet possible. This confusion prevailed until the 1960s and is still not completely eradicated.

During the first half of the 20th century, reconstructive surgical procedures were developed to treat the sequelae of poliomyelitis and could also be applied to BP lesions. These musculotendinous transfers had the approval of surgeons because they produced more predictable results and were more readily available than the demanding, time-consuming nerve repair surgery. BP lesions remained uncommon before World War II because the violent trauma needed to cause avulsion or rupture often resulted in life-threatening lesions as well. During the 1930s, BP lesions became more common, and the role of motorcycle accidents as an important cause of BP lesions was recognized by Bonola.¹⁷

The Second World War produced a large number of patients with BP lesions who were thoroughly studied and operated on by the team of Sir Seddon (Barnes, Bonney, Brooks, and Yeoman). They reported on the use of autologous nerve grafts in five patients with success in two.^18,19 The accumulation of World War II victims stressed the necessity of intensifying the search for better treatment modalities and led to a breakthrough in nerve repair. For the first time these patients were referred to special centers to receive specialist attention and to be available for thorough investigation and study. Antibiotics made effective control of wound infection possible. Studies at a basic level revealed the complexities of the internal structure of nerves. Then came the realization that restoration of continuity is just the first step in restoration of function. Disorderly regeneration and loss of regenerating axons at the suture line were recognized as important factors that complicated and adversely affected recovery. The central objective of nerve repair finally emerged, namely, to reduce the loss of axons that occurs during regeneration and to assist regenerating axons to reestablish useful functional connection with the periphery such that the new pattern of innervation approximated the original as closely as possible. Procedures and techniques were developed to create optimal conditions for regeneration and improve the repair. Among these was the suggestion that microsurgical techniques might be used to improve the repair of severed nerves.²⁰ The significance of the data emerging from basic studies filtered through to the clinical level in the 1960s.

By the late 1950s and into the 1960s, intercostal nerve (ICN)–to–musculocutaneous nerve (MCN) transfers were performed.²¹ Several other surgical teams also contributed to the understanding and treatment of BP lesions.^22–26 Unfortunately, during the 1960s, repair of traction injuries remained disappointing, and direct surgery was discontinued in favor of nonoperative treatment.^27–32 Above- or below-elbow amputation was advised in many cases.³³

Surgeons involved in reconstructive work on the upper extremity often explored the BP mainly to determine the diagnosis and obtain a prognosis. The discouragement widely felt culminated at a meeting of the International Society for Orthopaedic Surgery and Traumatology in Paris in 1966, where the following consensus was reached: (1) surgical exploration of BP injuries, particularly at the supraclavicular level, was of no real benefit to diagnosis and prognosis, and (2) repair was mostly impossible and, when performed, did not achieve substantial results.³⁴ The introduction of microscopic magnification in the same period offered new opportunities in the surgical management of BP injuries.³⁵ In the mid-1960s, a major step forward was made independently by Millesi and Narakas, who used microsurgical technique for nerve grafts coapted to ruptured trunks and cords. Both men were in disagreement with Fletcher, who favored amputation of the limb in severe cases and fitting of a prosthesis.³⁶ Their first reports stirred much interest but were also met with skepticism.^37–40 Gratifying results were achieved in a sufficient number of patients to encourage other surgeons to pursue early surgical reconstruction after injury. The efforts of Millesi and Narakas were soon complemented by those of other surgeons such as Allieu, Alnot, and Sedel in France, Brunelli in Italy, Kline in the United States, and Hudson in Canada.

Knowledge of nerve regeneration broadened as diagnostic tests became available, including electromyography (EMG)⁴¹ and cervical myelography.⁴² Refinements included the use of evoked nerve action potentials (NAPs)⁴³ and the association of computed tomography (CT) and myelography.⁴⁴ Seddon’s¹⁸ and Sunderland’s²⁰ classifications of nerve injury could be applied with more clarity, and the distinction between patients who should or should not undergo surgery became less vague. In the 1970s, survival after motorcycle trauma increased because of the use of helmets and improvement in lifesaving procedures. From 1975 onward, regular exchange of knowledge was established between European teams motivated by the steady increase in the number of patients with BP lesions. In similar fashion, groups of peripheral nerve experts became established in various parts of the world and published articles on a variety of topics related to peripheral nerves. Surgical repair of the peripheral nervous system has become commonplace in most major medical centers. A comment made in 1951 by Sir Sydney Sunderland, “it is no longer a question of what can be done, but of establishing what should be done,”²⁰ would appear to have been addressed.

Anatomy

Although numerous variations in formation of the BP have been reported, the following common form can be taken as a point of departure. The BP is composed of five spinal nerves that join to form three trunks, each of which divides into an anterior and posterior division that lead to the three cords from which all the major nerves to the arm subsequently arise. The five spinal nerves are C5, C6, C7, C8, and T1. Commonly, there is a contribution to the BP from C4. The C5 and C6 spinal nerves merge (Erb’s point) to form the upper trunk. The C8 and T1 nerves merge to form the lower trunk, and the C7 nerve continues into the middle trunk. Each trunk terminates in an anterior and posterior division. The posterior divisions come together into the posterior cord. The anterior divisions of the upper and middle trunks form the lateral cord, and the anterior division of the lower trunk forms the medial cord. Each cord has two main terminal branches: the lateral cord provides the MCN and the lateral component of the median nerve, the medial cord leads to the ulnar nerve and the medial component of the median nerve, and the posterior cord provides the axillary nerve and the radial nerve. Over its entire trajectory, smaller nerves to the shoulder and upper limb muscles leave the plexus. Variations in anatomy may hamper clinical efforts to localize the site of a lesion.⁵⁹

The axons in a peripheral nerve are surrounded by several layers of supporting connective tissue sheaths. Surrounding the axons and their Schwann cells is the basal lamina tube. After axonal injury, the contents of the basal lamina tube distal to the site of the lesion are phagocytosed because of wallerian degeneration. Reestablishment of continuity occurs by elongation of axons proximal to the lesion that are still in continuity with their cell body in the spinal cord’s ventral horn (motor) or spinal ganglion (sensory).

Pathophysiology and Classification

Traumatic injuries to the BP are predominantly stretch or compressive in nature but also include sharp laceration and concussive forces secondary to a high-velocity projectile. Traction injuries occur when the head and neck are moved away from the ipsilateral shoulder, which often results in an injury to the C5 or C6 spinal nerves or upper trunk. Caudal traction on the arm will also usually affect the upper spinal nerves and trunks. When the arm is abducted over the head with significant force, traction with potential for injury occurs within the lower elements of the BP. The more violent the trauma, the greater the risk for avulsion injury. Spinal nerves C5, C6, and C7 have fibrous attachments of the epineurium to the cervical transverse process that provide additional protection. The absence of ligaments at C8 and T1 explains the higher incidence of avulsion of these nerves, whereas C5 and C6 sustain a higher incidence of rupture at the transverse process. C7 occupies an intermediate position. Compression injuries to the BP usually occur between the clavicle and the first rib and are often associated with fracture of the clavicle. Nerve laceration can involve any part of the cross-sectional anatomy of the plexus and its elements and result in separation of the proximal and distal portion of the nerve. Partial nerve injuries are more common than complete lesions and often consist of a spectrum of injury to the nerve.

Despite the numerous ways to injure a nerve, the pathologic reactions are similar. After axonal injury, responses are at first degenerative and later regenerative. An increase in traction force applied to a peripheral nerve results in stepwise rupture of the peripheral nerve elements moving from inside to outside. Anatomic rupture begins with the axon or its coverings and then proceeds to the basal membrane, endoneurium, perineurium, and finally the epineurium. Based on these sequential events, the severity of the nerve lesion is graded in relation to the degree of neural damage. The most frequently applied classification scheme is Seddon’s classification of neurapraxia, axonotmesis, and neurotmesis.¹⁸ Sunderland introduced a classification based on five grades of progressive pathologic events.²⁰

The mildest form of nerve injury occurs when distortion of the myelin overlying the nodes of Ranvier results in a focal conduction block. This type of injury, conduction block but without wallerian degeneration, is referred to as neurapraxia or Sunderland type I injury. There can be some demyelination with this injury that prolongs the time until recovery; however, complete and rapid return of function is anticipated.

Moderate nerve injuries interrupt the axon’s continuity and result in wallerian degeneration but leave the endoneurium intact. These lesions are referred to as axonotmesis or Sunderland type II injury. Because the neural connective tissue structures remain intact, the potential for spontaneous regeneration is retained. In a Sunderland type III lesion, the endoneurium is additionally interrupted and the potential for spontaneous regeneration is reduced. When the perineurium surrounding the fascicle is also disrupted, the lesion is Sunderland type IV and spontaneous regeneration is greatly reduced to absent (neurotmesis with continuity). Finally, complete rupture of the nerve, including the epineurium, results in a Sunderland type V injury (neurotmesis without continuity). Spontaneous recovery is not anticipated with a Sunderland IV or V lesion. It should be emphasized that lesions are often mixed and that neurapraxia and axonotmesis may coexist. Similarly, the severity of injury may vary across fascicles in an injured nerve.

The dorsal root ganglion contains the cell bodies of the sensory neurons. It lies within the neural foramen of the bony spine. Preganglionic injuries involve the nerve roots and include intradural rupture of the nerve or avulsion of the nerve from the spinal cord. In both instances, continuity of the distal sensory axon with the cell body is maintained. Postganglionic lesions separate the cell body from the distal sensory axon and result in wallerian degeneration of the distal axon. The sympathetic component of the T1 spinal nerve provides sympathetic outflow to the head and neck. Injury to the T1 nerve can result in sympathetic ganglion loss and Bernard-Horner syndrome (miosis, ptosis, and anhidrosis).

Clinical

When gathering the history, information germane to dealing with a BP injury includes a complete description of the trauma and associated injuries. The description of the injury will suggest the type and severity of forces that the BP may have experienced. The time course of any recovery thus far can give an indication of the severity of the injury and the possibility for further spontaneous regeneration. Factors that impede nerve regeneration may include the patient’s age, neuropathy, metabolic disturbances, and various systemic diseases.

In many cases, family members or a review of the initial medical records can also provide valuable history concerning the course of events. This information may include details of an initial hospital course. If, for example, the patient required surgical procedures such as vascular repair of the subclavian artery or vein, surgical planning for repair of the BP may include having a vascular surgeon available. Fractures of the bony spine can be associated with spinal nerve or root injury.

The clinical history, physical examination, and adjuvant testing provide insight on localizing the injury and determining the extent of injury. The physical examination requires a detailed motor and sensory examination of the affected limb. The grading system of the Medical Research Council of Great Britain (MRC) dates back to the early treatment of poliomyelitis and war injuries and is at present the most commonly used strength-grading system.^60,61 Evaluation of both passive and active range of motion of the joints may reveal contractures requiring management.

When evaluating for proximal nerve injury, including avulsion, review of the dorsal scapular nerve (rhomboid muscle) and the long thoracic nerve (serratus anterior muscle) is critical. The upper trunk, proximal to the divisions, gives rise to the SSN. Preservation of rhomboid and serratus anterior function but loss of external rotation of the shoulder (infraspinatus muscle) and the first 30 degrees of shoulder abduction (supraspinatus muscle) moves the injury distal to the spinal nerve and into the upper trunk.

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. Atrophy of this muscle indicates a severe pan-plexus injury.

The Bernard-Horner sign is highly indicative of but not conclusive for avulsion of C8 and T1. False negative (absence of the sign with severe injury) is more common than false positive. The finding may not be present during the first 48 hours after injury and can also fade over time.

Regenerating nerve fibers develop mechanosensitivity. Percussion over the course of the nerve produces tingling paresthesia in the sensory distribution of the nerve (Hoffmann-Tinel or Tinel’s sign). Starting distally and progressing toward the lesion site, tingling is perceived as soon as the frontal tip of the down-growing highly sensitive but not yet myelinated fibers is met. Serial examination demonstrating distal shift of this point determines the presence and rate of growth of regenerating axons. However, an advancing Tinel sign does not indicate the number or quality of regenerating axons and does not guarantee functional outcome. Conversely, lack of an advancing Tinel sign is indicative of a poor prognosis.

Electrodiagnostic studies are useful in both the preoperative and intraoperative evaluation of a BP injury. They can help determine the nerves injured, the location of the injury, and the presence of regeneration. Electrodiagnostic studies are an important adjunct to a thorough history, physical examination, and imaging studies, not a substitute for them.

As wallerian degeneration of motor axons proceeds (over an approximately 2-week period), the trophic influence of nerve on muscle is lost and denervation potentials (positive sharp waves and fibrillations on EMG) appear. As a result, shortly after injury the severity of the lesion may be underestimated by needle EMG. Reduced recruitment of motor units occurs with axonal injury; however, the presence of motor units under voluntary control indicates a nerve in continuity and a good prognosis. EMG is also useful to look for early signs of recovery. Reinnervation potentials (nascent motor units, polyphasic action potentials) appear in a muscle several weeks in advance of clinical evidence of recovery. We often wait 2 weeks after injury to perform the first EMG and then, depending on the clinical scenario, will either opt for early surgery or perform serial EMG testing to look for early reinnervation. Recovery of EMG activity (new voluntary motor units) does not always predict a functional outcome.

An injury proximal to the dorsal root ganglion will spare the sensory axons because the cell body remains in continuity with the axons. Sensory nerve action potentials (SNAPs) are therefore preserved in injury that is proximal to the dorsal root ganglion. The presence of an intact SNAP with loss of sensation in the same distribution is highly suggestive of preganglionic injury. The presence of SNAPs plus absence of voluntary motor units is also highly suggestive of preganglionic injury. SNAPs are absent in postganglionic injury and in combined preganglionic and postganglionic injury.

Imaging of the spine and chest can provide valuable insight on the nature and extent of the injury. Preoperative imaging should include plain films of the cervical spine and chest. A clavicular fracture is frequently associated with traumatic BP injury. A spinal transverse process fracture can be associated with nerve root injury or avulsion. A rib fracture can indicate trauma to an ICN and preclude use of the nerve for transfer. An elevated hemidiaphragm suggests a phrenic nerve issue and probably C3, C4, and C5 spinal nerve injury.

Magnetic resonance imaging (MRI), as early as 4 days after injury (roughly 2 weeks before EMG changes become apparent), can document muscle changes caused by denervation. Once reinnervation has occurred, MRI muscle signals revert to normal. As yet, the information gained by MRI over traditional EMG has not changed treatment. In the acute situation, MRI can be used to visualize a hematoma and document vascular injury.

The term magnetic resonance neurography (MRN) applies to novel MRI techniques that have greatly improved the ability to visualize peripheral nerves.⁶² Thus far, this technique has been more valuable in evaluating compression syndromes and tumors than traumatic injury. Notwithstanding, with improvement in the resolution of MRN will come added ability to localize nerve injury and characterize the underlying pathophysiology.

Preoperative diagnosis of a preganglionic or root avulsion injury indicates the need for early surgery and use of nerve transfers. One standard for the diagnosis of root avulsion is laminectomy with direct visualization of the roots. This can be modified to a minimally invasive procedure using an endoscope⁶³; however, it has not become a commonly used technique. Clinical findings consistent with root avulsion include lack of power in the proximal muscles, presence of SNAPs with EMG denervation, Horner’s syndrome, and lack of nerve roots on imaging. When compared with intradural inspection of roots, CT myelography with 1- to 3-mm axial slices enables accurate diagnosis of the integrity of the roots in 75% to 85% of patients.⁶⁴ MRI with 3-mm axial slices provided an accurate diagnosis in just 52% of patients when compared with intradural inspection.⁶⁴ The presence of a pseudomeningocele, although highly suggestive of root avulsion, is not conclusive evidence. In the acute situation, an intradural blood clot can prevent filling of a pseudomeningocele. The resolution of MRI has improved enough in some institutions to trace the presence or absence of intradural nerve roots and thus negate the need for CT myelography. MRI acquisitions timed to the respiratory and cardiac cycle can improve the resolution of fine structures in cerebrospinal fluid by reducing apparent motion, although scan time may be greatly increased.

Based on his experience in treating 1068 patients with BP injuries during an 18-year span, Narakas developed his rule of “seven seventies.” He reported that approximately 70% of traumatic BP injuries occur secondary to motor vehicle accidents; of these, approximately 70% involve motorcycles or bicycles. Of the cycle riders, approximately 70% had multiple injuries. Overall, 70% had supraclavicular lesions; of these, 70% had at least one root avulsed. At least 70% of patients with a root avulsion have avulsions of the lower roots (C7, C8, or T1). Finally, of patients with lower root avulsion, nearly 70% will experience persistent pain.⁶⁵

Therapy/Management

There are five key elements in managing a BP injury: nonoperative care, appropriate selection of surgical candidates, timing of surgery, priorities of the surgical targets, and method of nerve repair. Despite progress in imaging and electrodiagnostics, determining whether a specific nerve injury is going to regenerate and thus the possible need for surgical intervention remains a formidable challenge. Surgery should be performed once it is determined that the risks associated with waiting are greater than the benefits of going ahead with surgery.

Nonoperative care is instituted after any life-threatening injuries have been stabilized. Maintaining range of motion of joints, muscles, and tendons is of utmost importance. Patients should be educated on passive and active range of motion of the paralyzed joint. While awaiting nerve recovery and muscle activity, the patient must be engaged in an aggressive stretching program. Splints and other mechanical appliances, in addition to physical and occupational therapy, can be used to help with maintenance of musculoskeletal integrity.

Neuropathic pain is common with BP injury, and most often the pain can be managed with medications. A combination of an anticonvulsant with a narcotic can be used for severe pain. Fortunately, the pain often resolves as regeneration is completed with innervation of targets.

Avulsion of nerve roots may lead to the development of severe pain in the distribution of the injured nerve root. The pain is often described as a burning or crushing pain with paroxysmal burning or shooting pain. Nonoperative treatment may require polypharmacy under the guidance of a pain center. Fortunately, the natural history of avulsion pain is that about half the patients become pain free or able to cope with their pain within 1 year and the majority are pain free within 3 years. Unfortunately, in some patients avulsion-related pain can become exceedingly severe. Avulsion pain can be surgically managed by making a series of lesions at the dorsal root entry zone of the traumatized spinal cord.⁶⁶

The mechanism of injury can help in determining the timing of surgery. For a sharp laceration with transection, early or immediate exploration and repair if possible are indicated. Delayed surgery is clearly met with a poor prognosis. Early surgery allows identification of the nerve ends and frequently a direct end-to-end repair of Sunderland V injuries. Surgical intervention is performed as soon as a team can be mobilized.

When there is evidence of a Sunderland V (rupture) injury as a result of a stretch mechanism at the site of transection, the nerve ends often need to be resected back to a more healthy region. Demarcation of the injury can be difficult to appreciate acutely, and 2 to 3 weeks is generally required for the injury to declare itself. This delay allows more effective trimming of the stumps to healthy tissue and results in better outcomes.^67,68 An alternative is to explore the wound, tag the nerve ends, and then re-explore in a few weeks. For gunshot wounds, a low-velocity projectile is often associated with Sunderland grade I injuries, whereas high-velocity projectiles cause more soft tissue disruption and higher grade Sunderland injury. Most gunshot wounds leave the plexus element in continuity. If there is no improvement on EMG in the first few months, exploration, NAP recordings, and if indicated, repair are necessary.

The optimal timing for surgical intervention in patients with BP injury secondary to stretch remains controversial, and no clear clinical guidelines exist. Patients are usually observed for a period of 3 to 4 months while undergoing serial examination for signs of regeneration, including an advancing Tinel sign or changes on EMG. Animal research has confirmed the clinical suspicion that delay in surgical repair will degrade the eventual outcome, so a balance must be maintained between waiting to see whether spontaneous regeneration will occur versus opting for surgical intervention. The history may point to a neurotmetic or avulsion lesion that could be operated on within days.

One factor that may limit recovery is the death of neurons after axotomy. Retrograde transport of neurotrophic factors potentially derived from target organs can support the survival and regeneration of both sensory and motor neurons.⁶⁹ Some neuroprotection is afforded by the presence of a nerve graft, but the amount of cell loss already present at the time of repair remains.⁷⁰ It is also apparent that over time, loss of Schwann cells in the distal nerve likewise has a negative impact on regeneration. These findings suggest that earlier repair of nerve injury is likely to be associated with improved outcome.

An argument against early surgery is that it does not allow the possibility of spontaneous recovery, which may bestow a better outcome than achieved with surgery. By waiting 3 months, there is the possibility of performing intraoperative electrophysiologic studies on a lesion to help determine whether regeneration is occurring. In contrast, early surgery is easier to perform because there is less scarring. Early surgery can allow visual inspection of the acute lesion and perhaps some prediction of outcome with or without nerve repair. Although there is some clinical evidence that early surgery may be beneficial to outcome, the debate rages on.⁷¹

Surgical management is divided into the initial, or primary, surgery and delayed, or secondary, surgery. Primary surgery is aimed at repairing the nerves and is time dependent. Examples include end-to-end nerve repair, end-to-side nerve repair, nerve grafting, nerve transfer, and neurolysis. Secondary surgery is performed when primary nerve repair is not likely to result in functional improvement or when recovery after nerve repair reaches a plateau still lacking in function. Secondary surgery is focused on maximizing function by performing muscle and tendon transfers and bone or joint work.⁷²

The priorities for restoration after adult BP injury are well established. Elbow flexion is the most important motion to restore because this alone provides a helper extremity. Shoulder stability with active abduction is often attainable and thus a second priority. Protective sensation of the hand is needed to maintain the extremity. Repeated trauma to an insensate extremity leads to chronic pathology and eventual deterioration. When possible, further targeting of repair includes wrist extension, finger flexion, wrist flexion, finger extension, and finally, intrinsic hand function. Several factors play a role in functional recovery after nerve repair, including delay in repair, length of the graft, scarring, viability of the proximal stump, age and general condition of the patient, and the complexity of functions to be restored.