Brachial plexus injuries: adult and pediatric

Gross anatomy

The brachial plexus is a large plexus which has the most complex structures in the peripheral nerve system. It locates superficially between two highly mobile structures, neck and arm, causing its susceptibility to injury, especially traction injury.

The brachial plexus is formed by the anterior primary rami of the lower cervical (C5–8) and the first thoracic (T1) spinal nerves, which give motor innervation to muscles of the shoulder, including all anterior and posterior chest muscles related to glenohumeral joint movement, muscles of the entire upper limb, and sensory innervation of the entire upper limb except the skin on some parts of the medial aspect of the upper arm (T2 zone). Not infrequently, the C4 and T2 spinal nerves also contribute nerve fibers to it. Whenever the C4 contribution is large and the T1 contribution is small, the brachial plexus is called a “prefixed brachial plexus”. When the C5 contribution is small and T2 contribution is large, it is termed a “postfixed brachial plexus”. Clinically, the prefixed type is more common than the postfixed. But postfixed plexus in adult BPI has been rarely seen, probably because the injuries are either so extensive that further dissection of C8–T1 is hazardous and unnecessary, or the injuries are diagnosed not to involve C8–T1 so that further identification is also unnecessary. Each spinal nerve is formed by the joining of the ventral root (motor fibers) and the dorsal root (sensory fibers). Each root is formed by a number of rootlets which exit from each spinal cord. These separate rootlets can be defined on magnetic resonance imaging (MRI) with three or four bands in the upper cervical roots (C5–7) and two bands of rootlets in the lower roots (C8–T1). The dorsal roots carry sensory information to the central nervous system (retrograde afferent), while the ventral roots convey motor fibers to the muscles (antegrade efferent). The cell bodies (neurons) of the motor fibers are located in the anterior horn of the spinal cord, while the cell bodies of the sensory fibers reside in the dorsal root ganglion located within the intervertebral foramen, immediately outside the dura mater of the spinal cord. The ventral and dorsal roots carry with them an extension of the arachnoid and dura which forms the root sleeve, where the root sleeve attaches to the ventral root and ganglion to form the sheath of the spinal nerve so that the cerebrospinal fluid space does not extend beyond the intervertebral foramen. The anatomy detailed here describes the location of level I ( Fig. 23.1 ).

The dorsal and ventral roots unite a few millimeters distal from the ganglion to form a spinal nerve, a mixed nerve, which goes through the interscalene space between the scalene anterior and middle muscles. Nerves arising from the anterior primary ramus include branch to scalene muscles (C5–8), branch to longus colli muscle (C5–8), the long thoracic nerve (C5–7), a portion of phrenic nerve (C5), and a portion of the dorsal scapular nerve (C5). The anatomy detailed here describes the location of level II (see Fig. 23.1 ).

Just out of the scalene muscles the five postganglionic spinal nerves make a first union to form the three trunks: upper (formed by C5 and C6), middle (C7 itself), and lower trunk (formed by C8 and T1 spinal nerves). Two branches are given off from the upper trunk: the nerve to the subclavius muscle and the suprascapular nerve. Each trunk divides into anterior and posterior divisions just proximal to or directly under the clavicle, retroclavicular. Lateral pectoral nerve is formed from anterior divisions of the upper and middle trunk to innervate the clavicular part of pectoralis major muscle; medial pectoral nerve is formed from the anterior division of the lower trunk to innervate the sternal part of the pectoralis major muscle. The anatomy detailed here describes the location of level III (see Fig. 23.1 ).

Infraclavicularly the nerves exchange fibers and form the second union, just distal to the clavicle, and are termed “cords”. The lateral cord is formed by the fusion of the two anterior divisions of the upper and middle trunks, containing fibers derived from C5 through C7. The posterior cord joining all three posterior divisions, containing fibers derived from C5 through C8. The medial cord is simply continuation of the anterior division of the lower trunk, containing fibers from C8 and T1. The subclavian artery becomes the axillary artery at the lateral border of the first rib. Names for the cord relationship are based on the axillary artery. Lateral and medial cords pass anterior, and the posterior cord passes posterior to the axillary artery. The cords run anterior to the subscapularis muscle, just distally behind the pectoralis minor muscle. Each cord has two or more terminal branches to the periphery. The medial and lateral cords, giving one or two terminal branches, join in a Y-shape to form the median nerve. Other main terminal branches of the cords contain the musculocutaneous nerve from the lateral cord, the ulnar nerve from the medial cord, the axillary and radial nerve from the posterior cord. The anatomy detailed here describes the location of level IV (see Fig. 23.1 ).

Numerous anatomical variations of the brachial plexus do exist and should be always kept in mind. For example, the musculocutaneous nerve may sometimes arise from the median nerve and not from the lateral cord. In some rare cases of C5–6 root avulsion, the musculocutaneous nerve is still found to be functional because part of the musculocutaneous nerve derives from the median with its origins from C7.

Microanatomy

The microanatomy or internal topography of the brachial plexus has been extensively studied and described. The monofascicular pattern is usually found in regions of: (1) the spinal nerves; (2) anterior and posterior divisions of the upper trunk; and (3) the origin of the suprascapular and musculocutaneous nerve. Marked changes in fascicular topography occur every 10 mm. It is especially true at the level of the upper trunk where interfascicular crossovers are so extensive that direct repair or repair with short nerve grafts will frequently lead to co-contractions due to aberrant regeneration of a group of muscles. This aberrant regeneration is mainly found in obstetrical brachial plexus palsy (OBPP) patients due to incomplete rupture or complete rupture but with small gaps. It is rarely seen in adult BPI, where the ruptures have longer gaps. In addition, plexus connective tissue is more abundant than neural tissue. All these factors are reasons why the results of brachial plexus nerve surgery are so unpredictable. Knowledge of the internal topography of the plexus can be helpful in connecting corresponding nerve fibers with bridging nerve grafts. However, localization within the spinal nerve to define specific axonal groups to supply specific muscles or specific branches is difficult and not practical.

Etiology of adult brachial plexus injury

BPI may be caused by trauma (open or closed type), compression, tumor, infection, inflammation, toxins, and other etiologies.

Patient history

Patient history should include mechanism of injury, conscious level at the time of trauma, associated injury (such as head injury, fracture, open wound, chest injury, vascular injury), kinds of previous surgical intervention (such as chest intubation, cervical spine surgery), and characteristics of pain. This information helps to determine the degree and extent of injury and the need for surgical intervention. Mechanism of injury (e.g., upward or downward traction and with or without rotation) is not easily detected due to the patient’s loss of consciousness or amnesia for the accident. A history of shoulder dislocation or glenoid fracture may have a high incidence of level IV injury, whereas a history of cervical spine injury or fracture may cause a level I root injury. Artery rupture and repair imply the site of nerve injury. For instance, arm traction by rolling machine or conveyor belt often causes an open wound in the axilla, extensive ecchymosis around the shoulder and chest (due to rupture of axillary vessels), and level IV BPI. Segmental thrombosis of the subclavian artery is usually associated with C8–T1 root injury. History of rib fracture and chest intubation may preclude intercostal nerve transfer because of a higher failure rate. Extreme causalgia with or without a phantom limb is often seen in cases of root avulsion in lower-root (C8–T1) avulsion as they contain the richest sympathetic fibers. The pain character, like an electric shooting, continues for short duration for seconds, followed by spontaneous relief and recurrence. Extreme causalgia is also a major factor for poor outcome due to poor rehabilitation. Sometimes a partial Brown–Sequard syndrome (hemitransection of the spinal cord with ipsilateral upper motor neuron lesion below the level of lesion, and contralateral abnormal sensation to pain and temperature which may not be at the same level) is also noted in the level I injury.

Motor examination

Muscle-by-muscle examination should be completed in a distal-to-proximal fashion and recorded, using the British Medical Research Council (MRC) scale (M0–5). We have modified the motor evaluation system, adding more detailed differentiation: M5, strength against four fingers (examiner) resistance; M4, against one finger, resistance for longer than 30 seconds; and M3, against gravity ( Table 23.4 ). M4 is recognized as useful muscle strength. The action of each muscle should be examined separately in relation to the movement of a single joint. Although there is no single muscle innervated by a single spinal nerve, some muscle palsy can give specific information related to the level of the injury. For instance:

• 1.

  Diaphragm palsy implies C4 and very proximal C5 (level I) injury.

• 2.

  The levator scapulae muscle lies anterior to the trapezius muscle in the neck, and can be more easily detected than the rhomboid muscles, which are covered by the trapezius muscle. Both levator scapulae and rhomboid muscles are innervated by the same nerve (dorsal scapular nerve, or C4 and C5). Preservation of its function in upper plexus or total plexus injury may imply C5 is a rupture injury (level II) with an available proximal stump.

• 3.

  Serratus anterior muscle: The long thoracic nerve has two portions: the upper portion originating from C5 and C6, and the lower portion from C7. The upper portion is responsible for scapular protraction, and the lower portion is important for scapular stabilization. Positive anterior traction of the scapula (shoulder protraction test) shows that at least C5 is ruptured after branching to the long thoracic nerve, so the proximal C5 is available for transfer. Scapular winging is observed only when the lower portion is denervated, but isolated C7 root avulsion is rarely seen in adult BPI. In pure C5–6 level I injury, the lower part of the muscle is still functional. The result of spinal accessory nerve transfer to the suprascapular nerve is much superior in the reconstruction of total root avulsion.

• 4.

  Clavicular and sternal portions of the pectoralis major muscle: The major pectoral muscle can be separated into two parts: clavicular and sternal parts. The clavicular part is innervated by upper and middle trunks or its divisions (lateral pectoral nerve), while the sternal part is innervated by the lower trunk (medial pectoral nerve). An incomplete or complete paralysis of the clavicular part of the pectoralis major muscle may imply at least level III or more proximal lesion.

Sensory examination

Sensory evaluation should include sensory tests and elicitation of a Tinel’s sign. Sensibility tests include pain and temperature appreciation, static and moving two-point discrimination, constant touch, and vibration. However, performing complete sensory tests in BPI is both unnecessary and illogical because we are examining the dermatomal distribution from spinal nerves, not the cutaneous distribution. Pinprick test from areas of normal to abnormal sensation to map out the area of sensory disturbance is sufficient for most brachial plexus-injured patients. Sensory grading is based on the British MRC scale (S0–4), modified by adding a grade for sensory overreaction (S2+) (see Table 23.4 ). Such sensory evaluations can give some clues about the level and degree of BPI.

The Hoffman–Tinel’s sign is an important clinical sign to determine the location of a neuroma or to track the regeneration of the injured nerves. Palpation or percussion at the neck, at the supraclavicular Erb’s point (clavicular insertion of the sternocleidomastoid muscle), at the infraclavicular coracoid process of the scapula, or at the route of different nerves along their course may induce an electric current sensation (like pins and needles) running down to the shoulder or the hand (positive Tinel’s sign). If the Tinel’s sign remains at a fixed point, which means retardation of its progression, surgical exploration is warranted. If the Tinel’s sign advances from supraclavicular to infraclavicular and distally to the arm or forearm at successive examinations, observation is recommended, as this may indicate a Sunderland third-degree lesion. A weak or absent Tinel’s sign in the neck region usually indicates total root avulsion.

Horner’s syndrome (miosis, ptosis, enophthalmos, and anhidrosis) is a sign of sympathetic nervous system disturbance. It indirectly implies avulsion of the T1 and C8 roots because the sympathetic fibers from the T1–2 sympathetic ganglia are quite close to the preganglionic fibers of T1 and C8. This syndrome may regress with time. It is a more reliable sign in adult BPI, but less accurate in obstetric traction injury.

Plain X-ray and imaging studies

Plain X-rays of the chest and cervical spine are required. The chest X-ray should include inspiration and expiration views to exclude diaphragmatic palsy. Cervical spine X-rays are evaluated for any fracture of the transverse process, spinous process, or vertebrae body.

Cervical myelography and computed tomography (CT) myelography can provide valuable information related to the level I injury of the brachial plexus. However, in recent years these studies have been gradually replaced by noninvasive MRI. The most useful MRI technique for the evaluation of a possible level I lesion is the three-dimensional (3D) fast imaging employing steady-state acquisition (FIESTA). These 3D source data are reconstructed along the planes of ventral and dorsal rootlets using a curve planar reformat technique to demonstrate the respective rootlets in a better perspective. Other MR techniques help in the imaging of the whole brachial plexus, notably of level II ( Fig. 23.6 ). Abnormal posterior paraspinal muscles on contrast-enhanced MRI is also one of indirect signs of root avulsion.

Electrodiagnostic studies

Electrodiagnostic studies, mainly consisting of nerve conduction studies (NCS) and needle electromyography (EMG), are used to localize the lesion and to assess its severity. For the NCS, only amplitudes of the sensory nerve action potentials (SNAP) and compound muscle action potentials (CMAP) are of value. Both SNAP and CMAP amplitudes provide a good indication of the degree of axon loss, or in contrast, the number of survival axons capable of conducting impulses. Sensory NCSs assess the function of the postganglionic portion of the sensory pathway. Therefore, abnormally low SNAP amplitudes indicate a ganglionic or postganglionic lesion. Conversely, SNAP amplitude that remains normal in a complete paralysis of limb implies pure level I lesion (root avulsion). A combination of un-elicitable CMAPs with abnormal low or absent SNAP amplitudes suggests a combined preganglionic and postganglionic lesion. To assess the major elements of the brachial plexus, NCSs of multiple nerves are usually required. These include sensory NCS of the median, ulnar, radial, musculocutaneous, and axillary nerves, as well as motor NCS of the median, ulnar, radial, musculocutaneous, and axillary nerves. The presence of fibrillation potentials on needle EMG may suggest that the lesion is at least axonotmesis. The reduction of amplitude of compound muscle action potentials during motor NCS is more reliable than the presence of fibrillation potentials to indicate axonal loss rather than a neurapraxia lesion.

Needle EMG can detect a minimal amount of motor axon loss. For a comprehensive evaluation of the brachial plexus, adequate EMG sampling of muscles is critical. In addition to muscles innervated by major terminal nerves, examination of those supplied from or proximal to the brachial plexus can help to localize the lesion. These include the rhomboid major, serratus anterior, pectoralis major (both the clavicular and sternocostal parts), latissimus dorsi, teres major, and cervical paraspinal muscles. Needle EMG can also reveal evidence of early reinnervation and chronicity of the lesion. Presence of denervation potentials (i.e., fibrillation potentials and positive sharp waves) is the most sensitive indicator of motor axon loss. However, they require about 3 weeks to develop after axon injury.

Percutaneous somatosensory evoked potentials (SEPs) provide far less information than comprehensive NCSs and needle EMG studies in most patients with BPI. Thus SEPs are not routinely performed. Intraoperative SEPs may be useful to determine the continuity of individual segments of the brachial plexus and roots.