A history of the mechanism of injury can be obtained from the patient or parents. High-energy injuries, for example, from motor vehicle accidents, have a lower chance of spontaneous recovery than low-energy injuries. Sharp injuries from lacerations should be explored acutely or subacutely. Gunshot wounds, in contrast, should be observed as many of these will exhibit spontaneous recovery over time.

As other injuries may be associated with the injury to the brachial plexus, management of these takes precedence in the multitrauma patient. Management follows acute trauma life support (ATLS) principles, with attention to the airway, breathing and circulation taking priority, followed by treatment of other life- and limb-threatening injuries. A detailed examination of the various components of the brachial plexus should be performed when the patient is stable and able to cooperate with the examiner. The exam should be recorded in a manner that allows for comparison of dates (Fig. 3). Serial detailed examinations help determine the presence or absence of nerve recovery and prognosis for spontaneous improvement.

In an older child, a detailed systematic examination of the brachial plexus may be possible, together with documentation of muscle strength following the modified grading system of the British Medical Research Council (BMRC) for adults (M0 to M5) (Mendel and Florence 1990; Table 1). Note that a patient cannot have grade 3 power unless there is full active motion against the existing passive range of motion.

A preganglionic lesion can be diagnosed through the presence of Horner’s syndrome, which suggests a root avulsion at the T1 level. Additional muscles that are innervated close to the spinal cord provide further evidence of a preganglionic lesion. Paralysis of the serratus anterior muscle, manifest through winging of the scapula (may be very difficult to examine secondary to paralysis of other periscapular muscles), suggests a lesion proximal to the long thoracic nerve, formed by the C5, C6, and C7 nerve roots. Atrophy of the rhomboids and parascapular muscles also suggests a preganglionic lesion proximal to the origin of the dorsal scapular nerve. Examination of individual sensory dermatomes may sometimes be unreliable due to overlap from other nerves or anatomical variation.

Different patterns of injury may be predicted based on the mechanism of injury. Upper brachial plexus injuries (Fig. 4) occur in motorcyclists who fall with the shoulder forced downward and the head pushed to the other side. Lower brachial plexus or pan-plexus injuries (Fig. 5) may occur during fall from height through hyperabduction of the injured upper extremity.

Examination of donor nerves should also be performed if nerve transfer is contemplated as a treatment option, such as the spinal accessory nerve, nerve to triceps, and medial pectoral nerve. Presence of a Tinel’s sign and tenderness in the supraclavicular or infraclavicular area suggests a postganglionic lesion, while absence of these suggests a preganglionic lesion. An advancing Tinel’s sign is a prognosticator and suggests a recovering lesion. Minimal preservation of movement in tested muscle groups suggests a partial injury with a greater potential for recovery. Examination of stability, active, and passive range of motion of all joints should also be assessed. Concomitant spinal cord injury should be ruled out by performing a full neurological examination of the upper and lower extremity, testing for power, sensation, and reflexes. Finally, a vascular exam should be performed, as the subclavian or axillary artery can be ruptured or damaged in substantial injury to the brachial plexus.

In younger children and infants, evaluation of single muscles is not easy and the patient often not cooperative or able to tolerate a lengthy physical examination. The muscle grading system of Gilbert and Tassin (1984) has been proposed (M0 = complete paralysis; M1 = perceptible contraction; M2 = weak movements; M3 = normal muscle); however, it is our practice to attempt to use the BMRC grading as not to confuse muscle grading systems. The posture of the child can be used to evaluate the level of injury. Paralysis of the upper roots is suggested by the upper limb being held in internal rotation and pronation, with no abduction possible. Slight flexion of the elbow may suggest involvement of C5, C6, and C7, while full extension of the elbow suggests involvement of only C5 and C6. In complete involvement of the brachial plexus, the entire upper extremity is flail. Horner’s syndrome may also be observed in a preganglionic lesion involving the C8-T1 roots.

Assessment of sensation in younger children and infants is similarly difficult. Often, children will only respond to testing with painful stimuli. Tinel’s sign may be used to assess for the level of the lesion and presence of nerve regeneration. In addition, in children, deafferentation pain is absent; hence, unlike in adults, patients with preganglionic lesions will seldom exhibit pain. Disturbances of the sympathetic system may be observed in the immediate period following injury, such as anhidrosis, cyanosis, and edema.

Radiographs of the cervical spine, shoulder, and chest should be obtained as part of the workup. Transverse fractures of the cervical vertebrae suggest preganglionic injuries with root avulsion. Fractures of the clavicle or ribs (first or second) may be associated with injuries to the brachial plexus. Fractures of other ribs may preclude the use of intercostal nerves as a donor nerve for nerve transfer (Kovachevich et al. 2010). An elevated hemidiaphragm suggests damage to the phrenic nerve and a possible preganglionic injury.

Computed tomography (CT) myelography is very useful in determining the presence of preganglionic injury. This is usually performed at least 3–4 weeks after the injury. This delay allows blood clot in the area of the avulsed cervical root to resorb and for a pseudomeningocele to form. CT myelography has been shown to have a diagnostic accuracy ranging from 70 % to 95 % for detection of nerve root avulsion compared to plain myelography alone (Carvalho et al. 1997; Doi et al. 2002). Presence of a pseudomeningocele is associated with preganglionic injury in 98 % of cases (Nagano et al. 1989; Fig. 6).

Magnetic resonance imaging (MRI) is also useful in evaluating patients with suspected root avulsion (Fig. 7) and allows better visualization of the entire brachial plexus. Large neuromas, abnormalities of the rootlets, inflammation and edema, as well as mass lesions can be visualized using MRI. While CT myelography should still be considered the first-line imaging modality in suspected nerve root avulsion, MRI using an overlapping coronal-oblique slice technique has been shown to be as reliable as CT myelography in detecting nerve root avulsion, with a diagnostic accuracy of 93 % (Doi et al. 2002).

Electrodiagnostic studies are pivotal in localizing and determining severity of injury in the brachial plexus. Baseline nerve conduction studies (NCS) and electromyography (EMG) should be performed 3–6 weeks after injury, after Wallerian degeneration has occurred. Serial follow-up studies can then be performed every 6–10 weeks to assess for recovery, complementing findings on physical examination for the purposes determining if there will be spontaneous recovery or if surgical reconstruction is necessary.

Nerve conduction studies include testing of motor and sensory nerves. Motor nerve testing is useful in detecting more distal injuries and conduction blocks in incomplete injuries. In traumatic brachial plexus injuries, amplitudes of compound muscle action potentials (CMAPs) are in general low. Sensory nerve testing is useful in determining whether a root injury is pre- or postganglionic. In a preganglionic injury, the dorsal root ganglion is spared injury even though it is detached from the spinal cord; hence, sensory nerve action potentials (SNAPs) are preserved. However, the patient is insensate in the associated sensory nerve distribution. There is excellent correlation between C6 (superficial radial nerve), C7 (median sensory to long digit), C8 (ulnar sensory nerve to small digit), and T1 (medial antebrachial cutaneous nerve) nerve root levels and individual peripheral sensory nerves in the upper extremity, aiding in localizing the level of the lesions in the brachial plexus. C5 and C6 innervate the lateral antebrachial cutaneous nerve, and this can be tested as part of sensory nerve evaluation for C5.

EMG provides the most reliable assay of motor nerve injury. Fibrillation potentials and positive sharp waves, indicative of denervation, can be seen in proximal muscles as early as 10–14 days and in distal muscles in 3–6 weeks. Motor unit analysis can determine the presence of injury in individual muscles. Polyphasic motor units occur in the presence of injury or pathology, while nascent potentials indicate axonal regeneration. Reduced recruitment of motor unit potentials can be demonstrated immediately after injury. Testing of individual muscles can help to distinguish preganglionic from postganglionic lesions. For example, denervation of the paraspinal muscles, rhomboids, and serratus anterior is a strong indicator of a preganglionic lesion as these muscles are innervated by branches from the cervical roots. Unfortunately, EMG recovery does not always equate with clinical recovery, and evidence of clinical recovery may not be detected through EMG in complete lesions, despite ongoing regeneration, when target end organs are more distal.

Intraoperative electrodiagnostic studies are useful prior to making a definitive surgical decision. These include the use of nerve action potentials (NAP), somatosensory and motor evoked potentials (SSEPs and MEPs) and CMAPs. The presence of a NAP distal to a lesion indicates preserved axons in an incomplete lesion or significant regeneration, with a correspondingly better prognosis (Kline and Happel 1993). Hence, neurolysis alone without additional treatment may be sufficient. The presence of an SSEP and MEP suggests continuity between the peripheral nervous system and the spinal cord via a dorsal root and ventral root, respectively. Hence, an SSEP and MEP is only present in postganglionic lesions. As lesions are not necessarily all or none, there may be situations where the SSEP is present and MEP absent or vice versa. In such cases, the viability of the root becomes difficult to ascertain, and correlation with EMG, clinical findings, and radiographic findings is necessary. CMAPs are useful in partial lesions where the magnitude of the lesion is proportional to the number of functioning axons.

Brachial plexus injuries are often caused by high-energy trauma; hence, there are often substantial concomitant injuries. When a patient is evaluated in the emergency room, standard ATLS principles should be followed and the airway, breathing, and circulation stabilized prior to a search for other life- and limb-threatening injuries.

Injuries that should be evaluated include traumatic brain injuries and spinal cord injuries as well as vascular injuries, ipsilateral upper extremity musculoskeletal injuries, scapulothoracic dissociation, pneumothorax, hemothorax, and rib fractures. Treatment of these injuries takes precedence initially over the brachial plexus injury.

Brachial plexus injuries can be classified by the cervical/thoracic nerves involved, level of the injury, and severity of each nerve injury according to the Seddon (1975) and Sunderland (1978) classifications.

Determination of the cervical/thoracic nerves involved in the injury as well as level of the injury is based on clinical examination supplemented by imaging studies, as described previously. The level of the injury can be broadly divided into preganglionic root, supraclavicular plexus (roots and trunks), retroclavicular plexus (divisions), and infraclavicular plexus (cords and branches). Different authors have described various classifications for the level of injury, pursuant on their individual surgical strategies. Narakas (1981) describes dividing the level of injury into five levels (supraganglionic root, infraganglionic spinal nerve, infraganglionic trunk, and retroclavicular and terminal branches). Chuang (2010) alternatively divides brachial plexus injuries into four levels (Level 1, preganglionic root injury including spinal cord, rootlets, and root injuries; Level 2, postganglionic spinal nerve injury limiting the lesion to the interscalene space and proximal to the suprascapular nerve; Level 3, preclavicular and retroclavicular BPI including trunks and divisions; Level 4, infraclavicular BPI including cords and terminal branches proximal to the axillary fossa).

The severity of the injury to each nerve is classified according to the Seddon/Sunderland classification, with five levels of nerve injury. Seddon initially described three basic types of peripheral nerve injury, which was expanded to five types by Sunderland. In neurapraxia (first-degree injury), a focal physiologic conduction block exists at the site of nerve injury, with the endoneurium, perineurium, and epineurium remaining intact. This typically recovers spontaneously within days to weeks and does not require surgical intervention. In axonotmesis (second- to fourth-degree injury), the connective tissue framework of the nerve is preserved, but the axon is disrupted to varying extents, resulting in Wallerian degeneration distal to the site of nerve injury with a disruption of nerve conduction. In a second-degree injury, the axon is disrupted, but the endoneurium, perineurium, and epineurium are preserved. Nerve regeneration and recovery of motor and sensory function are expected but may take months to years. In third-degree injury, the axon and endoneurium are disrupted, but the perineurium and epineurium are intact. In a fourth-degree injury, the axon, endoneurium, and perineurium are disrupted, while the epineurium remains intact. In third- and fourth-degree injuries, regeneration occurs variably or may not occur at all; hence, these lesions often require surgical intervention. In neurotmesis (fifth-degree injury), the nerve is completely disrupted; hence, surgical intervention is necessary. Fifth-degree injuries include preganglionic avulsions and postganglionic ruptures of the brachial plexus.

There is no clear consensus on the best instrument to measure outcomes after brachial plexus reconstruction in either the adult or pediatric patient. The BMRC muscle strength scale is the most common tool used, on a scale from M0 to M5. However, this tool has a number of limitations including its lack of precision and inter-rater reliability as well as a wide range of strength covered by grade of M4 (MacAvoy and Green 2007; Shahgholi et al. 2012). The Disabilities of the Arm, Shoulder and Hand (DASH) score is the most validated measure of upper extremity function (Dowrick et al. 2005) and has been useful as an outcome tool following brachial plexus reconstruction; however, it is not validated for the pediatric population.

A number of outcome tools are used to assess outcomes following obstetric brachial plexus injury and reconstruction, and there may be a role to adopting them to infants or young children following traumatic brachial plexus injury. The most widely used tool is the Mallet classification (Abzug and Kozin 2010), which measures integrity of muscles innervated by the upper brachial plexus. The arm is tested in five different movements: abduction, external rotation, hand behind head, hand to back, and hand to mouth. Each movement is classified from grade 0 to V. A recent modification was proposed, the addition of a hand to belly button category, which tests the child’s ability to reach to midline. Other tools used include the Toronto Test Score (Bae et al. 2003), Active Movement Scale, and Gilbert shoulder classification.

In pediatric patients, surgical reconstruction options are affected by the presence of open growth plates and potential growth that may affect the long-term results of certain procedures. The most active physes in the upper extremity include the proximal humerus, which accounts for 80 % of longitudinal growth (Pritchett 1991) and the distal radius. The length of the humerus grows approximately 1.3 cm in boys and 1.2 cm in girls from age seven until skeletal maturity, while the length of the radius grows approximately 1.0 in boys and 0.9 cm in girls from age seven until skeletal maturity (Pritchett 1988). Boys usually reach skeletal maturity between 16 and 17 years old, while girls reach skeletal maturity between 14 and 15 years old.

Therefore, in the pediatric population, the use of functional free muscle transfer should be done so with care as the transferred muscle may not grow in proportion with the rest of the upper extremity and lead to joint contractures. Additionally, secondary procedures that cross joints, such as tendon transfers, tenodesis, and joint fusions, may lead to joint contractures and limitation of function of the affected limb.