Structure of Peripheral Nerve

The peripheral nerve consists of both unmyelinated and myelinated fibers and their supporting elements. The unmyelinated axons are surrounded only by the plasma membrane of a Schwann cell. The myelinated axons are engulfed by a Schwann cell that wraps around the axons multiple times, thereby insulating the axon with multiple layers of cell membrane, which is rich in lipid sphingomyelin. The myelinated axon is surrounded completely by myelin and Schwann cells, except at certain gaps. In adults, the gaps between myelin segments, called the nodes of Ranvier, measure approximately 1 μm, while myelinated segments between nodes, called the internodal segments, measure approximately 1 mm each.

The relatively thick myelin sheath has a low capacitance and a large resistance to the electrical current that attempts to escape from the axon to the extracellular space. Also, the axons contain a high concentration of voltage-gated sodium channels at the nodes of Ranvier, which are essential for the propagation of action potentials. These characteristics of myelinated fibers (myelin and Na channels at the nodes of Ranvier) result in a rapid saltatory conduction between consecutive nodes of Ranvier.

Pathology of Peripheral Nerve Injury

Transient neurologic symptoms related to minor peripheral nerve compression are extremely common and are rapidly reversible. They probably result from action potential propagation failure caused by ischemia. They are not associated with structural alteration of the axon, myelin, or supporting structure. In contrast, prolonged or severe compression, traction, laceration, thermal, or chemical injury may damage the myelin, axon, or the supporting components of the peripheral nerves and results in significant disability from which the patient may not recover completely.

Nerve injuries that are associated with focal interruption of the continuity of the axons cause significant changes in the structure of the peripheral nerve distal to the lesion (Table C1-1). The distal axons undergo a degenerative process, known as wallerian degeneration. This occurs since all the necessary building blocks needed for maintaining the axon are made in the cell body (peikaryon) and cannot reach the distal stump. The rate at which wallerian degeneration proceeds varies depending on the nerve injured, axon diameter, and the length of distal stump (the larger and the longer the distal stump the more time is needed for wallerian degeneration to be completed). Within hours of most nerve injuries, myelin begins to retract from the axons at the nodes of Ranvier. This is followed by swelling of the distal nerve segment, leakage of axoplasm, and subsequently the disappearance of neurofibrils. Within days, the axon and myelin fragment, and digestion of nerve components starts. By the end of the first week, the axon and myelin become fully digested and Schwann cells start to bridge the gap between the two nerve segments. In chronic nerve lesions, the endoneurial tubes in the distal stump shrink, the nerve fascicles atrophy distal to the lesion, and, in complete nerve transection, the severed ends retract away from each other.

Table C1-1 Consequences of Focal Axonal Injury Distal to the Lesion

In contrast to the severe changes that occur distal to the lesion, only minor changes occur proximally. Though most of the proximal stump survives and maintains its ability to regenerate, there is often a slight retrograde degeneration of axons, up to several centimeters from the site of injury depending on the severity of the lesion. Also, the neuron cell body reacts to the axonal injury, by revealing an eccentric nucleus and marginally placed rough endoplasmic reticulum (Nissl’s substance). These changes are worse with proximal than with distal nerve lesions.

Classification of Peripheral Nerve Injury

Many classifications of peripheral nerve injury have been suggested, but Seddon’s and Sunderland’s classifications are the most widely used in clinical practice. These are based on the functional status of the nerve and on histologic findings. They are shown in Table C1-2 and in Figure C1-2, with their corresponding electrophysiologic findings.

Figure C1-2 Sunderland classification of peripheral nerve injury. 1, First degree: conduction block. 2, Second degree: wallerian degeneration secondary to a lesion confined to the axon, with preservation of the endoneurial sheath. 3, Third degree: disruption of the axon and endoneurial tube with an intact perineurium. 4, Fourth degree: disruption of all neural elements except the epineurium. 5, Fifth degree: transection with complete discontinuity of the entire nerve trunk.

(Reprinted with permission from Sunderland S. Nerve injuries and their repair: a critical appraisal, Edinburgh: Churchill Livingstone, 1991.)

Diagnosis of Peripheral Nerve Injury

Injuries to peripheral nerves are highest in prevalence in young adults between the ages of 18 and 35 years and result in substantial degree of disability. They are often accompanied by other bodily injuries including fractures, dislocations, or soft tissue damage. When associated with head or spine injury, peripheral nerve lesions may be overlooked until late during the rehabilitative phase of treatment. Traumatic nerve injuries may be direct (such as with a stab wound to the sciatic nerve) or indirect (such as with radial neuropathy following humeral fracture). These lesions are much more common during wartime, but they also accompany civilian trauma that results from vehicular accidents, industrial accidents, gunshots, or knife wounds. Also, a significant percentage of peripheral nerve injuries encountered in clinical practice are iatrogenic, occurring in the setting of surgical or radiological procedures, or following needle insertion or medical therapy such as with the use of anticoagulation.

The diagnosis of peripheral nerve injury often requires a detailed history and neurologic examination, with the EDX studies and surgical findings playing important roles in diagnosis and management. The history and physical examination are extremely important in predicting the location, type, and severity of the nerve lesion. For example, a stab wound injury to a nerve is often associated with axonal interruptions and grade three to five nerve injuries, while intraoperative nerve compression distant from the site of surgical field is usually a grade one (neurapraxic and demyelinating) or two (axonal) nerve injury.

Electrodiagnosis of Peripheral Nerve Injury

The EDX studies are the cornerstone in the diagnosis and management of nerve injuries by providing valuable information as to the location of the lesion, and its severity, pathophysiology, and prognosis (Table C1-3). Intraoperatively, the EDX studies guide the surgeon during the procedure and help assess the status of the regenerating axons within the injured nerve segment. During the recovery stage of peripheral nerve injury that may occur spontaneously or after surgical repair, the EDX studies are also essential in the evaluation of remyelination, regeneration, and reinnervation.

Table C1-3 Role of Electrodiagnostic Studies in Peripheral Nerve Injury

In contrast to the anatomical classification of nerve injuries, the pathophysiologic responses to peripheral nerve injuries have a limited repertoire: that is, axon loss, demyelination, or a combination of both. The EDX studies evaluate the integrity of the myelin sheath and the axon exclusively, and can only distinguish a neurapraxic injury (myelin injury) from all other degrees of injury that are associated with axonal damage and wallerian degeneration.

Localization of Nerve Lesions Using Nerve Conduction Studies

There are essentially three electrophysiologic consequences to peripheral nerve injury that can be assessed by nerve conduction studies. Two of them, namely focal slowing of conduction and conduction block, are caused by myelin disruption; the third is a manifestation of axonal loss (conduction failure).

Focal Slowing

Focal slowing in peripheral nerve injuries represents a convenient method of localizing lesions. When focal slowing is an isolated finding such as of the ulnar nerve across the elbow, the patient is not symptomatic and has no weakness or sensory loss. In symptomatic peripheral nerve injuries, focal slowing is associated with conduction block due to internodal demyelination, axon loss, or both.

Focal slowing of conduction usually is caused by widening of the nodes of Ranvier (paranodal demyelination) and, sometimes, focal axonal narrowing. It is evident on NCSs by slowing of conduction of a specific nerve segment, while other segments of the same nerve as well as neighboring nerves remain normal. When the large myelinated fibers are slowed to essentially the same extent, focal slowing across the involved nerve segment is synchronized. This is manifested by either a prolongation of distal latencies (in distal lesions) or slowing in conduction velocities (in proximal lesions), while the CMAP amplitude, duration, and area are not affected and do not change when the nerve is stimulated proximal to the lesion. When variable number of the medium or small nerve fibers (average or slower conducting axons) are affected only, desynchronized (differential) slowing of conduction across the nerve segment is evident. In this situation, the CMAP is dispersed on stimulation proximal to the lesion and has prolonged duration, with normal (nondispersed) response on distal stimulation. If this finding is isolated, the distal latency or conduction velocity, which represent the speed of the largest (fastest) axons, are normal. However, in most clinical situations, the large fibers are often involved also, desynchronized slowing is usually accompanied by slowing at the involved segment, resulting in concomitant slowing of distal latency or conduction velocity.

Conduction Block

Normally, the action potential is generated by sufficient temporal and spatial summation of excitatory inputs to motor or sensory axons. The nerve potential travels a myelinated axon in a saltatory fashion, passing hundreds of nodes of Ranvier without failure. The axonal regions at the site of the nodes of Ranvier are rich in Na channels. An abrupt change in Na conductance forms the basis for the generation of nerve action potential and the maintenance of saltatory conduction. Loss of myelin can involve one or more segments of these axons (segmental demyelination). Segmental demyelination can result from in widening of the nodes (paranodal demyelination) or the loss of one or more internodal segments (internodal demyelination). Both forms of demyelination can result in slowing or block of conduction. However, at least in compressive/entrapment neuropathy, focal slowing of conduction is characteristic of paranodal demyelination, whereas conduction block is a manifestation of internodal demyelination.

Before one can understand the electrophysiologic diagnosis of conduction block, the normal conduction studies of nerves, especially in reference to temporal dispersion and phase cancellation, and, ultimately, conduction block, must be discussed.

Three physiologic facts play a pivotal role in the generation of the CMAP which is obtained with surface recording.

1. The CMAP is produced by supramaximal stimulation of peripheral nerve and represents the summation of all individual muscle fiber action potentials directed to the muscle through the stimulated nerve.

2. The surface-recorded motor units are biphasic, with an initial negative phase followed by a positive phase, and a total duration of 5 to 15 ms in most human muscles.

3. The motor axons are not uniform but differ in size, thickness of myelin, and conduction velocities. The range of conduction velocities of individual human motor axons is 12 to 13 m/s.

Because of this physiologic variability, the CMAP configuration changes according to the site of stimulation. Typically, as the stimulus site moves proximally, the CMAP increases in duration and decreases in amplitude and, to a lesser extent, area. With more proximal stimulation, action potentials generated by motor units of slowly conducting fibers are increasingly dispersed in time with respect to those from fast-conducting fibers. This results in positive/negative phase overlap and cancellation of some components of the motor unit waveforms, thus prolonging its duration and reducing the amplitude and area of the summated response (CMAP).

Temporal dispersion and phase cancellation are more prominent in sensory nerve conduction studies due to (1) the disparity of sensory fiber conduction velocities which are almost double that of the motor axons (25 m/s) and (2) the surface recorded nerve action potentials are triphasic. The SNAP may normally decrease in amplitude and area by 50+ or more and its duration can increase by 100+ or more with proximal stimulation in antidromic studies (or with proximal recording with orthodromic studies). Hence, it is a common practice not to rely on sensory studies in the diagnosis of conduction block. Figures C1-3 and C1-4 depict the concept of temporal dispersion and phase cancellation using computer modeling.

Figure C1-3 Temporal dispersion and phase cancellation of two surface-recorded motor unit potentials at distal and proximal sites. This can be translated into many similar biphasic potentials, which contribute to the compound muscle action potential (CMAP).

Rights were not granted to include this figure in electronic media. Please refer to the printed book.

(Reprinted with permission from Kimura J et al. Relation between size of compound sensory or muscle action potentials, and length of nerve segment. Neurology 1986;36:647–652.)

Figure C1-4 Temporal dispersion and phase cancellation of two surface-recorded single-fiber sensory potentials at distal and proximal sites. This can be translated into many similar biphasic potentials, which contribute to the sensory nerve action potential (SNAP).

Rights were not granted to include this figure in electronic media. Please refer to the printed book.

(Reprinted with permission from Kimura J et al. Relation between size of compound sensory or muscle action potentials, and length of nerve segment. Neurology 1986;36:647–652.)

Impeding transmission of action potentials is the basis of conduction block. This usually results from internodal demyelination, but can occur in axonal loss before wallerian degeneration (“axonal” conduction block). Blockage of the transmission of electrical impulses anywhere throughout the course of motor axons results in motor weakness that is often indistinguishable from weakness that results from loss of motor neurons or motor axons. Experimental evidence on tourniquet paralysis on baboon hind limb showed that conduction block is reversible and the distal nerve remains normal and excitable (Figure C1-5).

Figure C1-5 Evoked compound muscle action potential (CMAP) from the abductor hallucis muscle of a baboon at different intervals after a tourniquet was inflated for 95 minutes to 1000 mmHg around the knee. S1, S2, and S3 are the sites of stimulation, as is shown in the schematic (bottom).

(Reprinted with permission from Fowler CJ, Danta G, Gilliatt RW. Recovery of nerve conduction after a pneumatic tourniquet: observation on the hind-limb of the baboon. J Neurol Neurosurg Psychiatry 1972;35: 638–647. © BMJ Publishing Group.)

In practice, conduction block is defined as a relative decrease in the CMAP amplitude and area with proximal stimulation, when compared with the CMAP on distal stimulation, without significant prolongation of CMAP duration. Conduction block should be distinguished from physiologic or abnormal temporal dispersion. Based on experimental studies, differential slowing along medium and thinly myelinated fibers may result in temporal dispersion and phase cancellation manifesting as significant drop of amplitude that may occasionally reach up to 80+. This is often associated with obvious and marked prolongation of CMAP duration. In contrast to amplitude decay, differential slowing does not drop the area beyond 50+. Hence, in true conduction block a significant drop in amplitude should always be corroborated by a similar drop in CMAP area.

There are no uniformly accepted criteria for the identification of conduction block. Table C1-4 reveals some of the common errors made in the EMG laboratory in the diagnosis of true conduction block. Table C1-5 lists practical criteria for the diagnosis of conduction block. In general, an amplitude change should be always supported by area change, since a significant drop in amplitude up to 50+ or more may occasionally be due solely to abnormal temporal dispersion while an area drop of more than 50+ is always due to conduction block. In clinical practice, the identification of demyelinative conduction block is an excellent tool for precisely localizing peripheral nerve lesions. Conduction block is often caused by acute nerve compression such as peroneal mononeuropathy at the fibular neck or radial mononeuropathy across the spiral groove. It is also a common finding in immune-mediated peripheral neuropathies such as acute inflammatory demyelinating polyneuropathy, chronic inflammatory demyelinating polyneuropathy, or multifocal motor neuropathy. Finally, conduction block usually is reversible and amenable to treatment, by removing the offending compression factor from the injured nerve or immunotherapy.

Table C1-4 Common Errors in the Diagnosis of Conduction Block

Technical Submaximal percutaneous stimulations (proximal sites, obesity, edema) Examination of long peripheral nerves (tibial nerve, tall subjects) Anomalous innervation (e.g., Martin-Gruber anastomosis)

Pathological

Abnormal temporal dispersion with phase cancellation

Table C1-5 Electrodiagnosis of Conduction Block^*

Definite in Any Nerve† ≥50% decrease in CMAP amplitude, with <15% prolongation of compound muscle action potential (CMAP) duration, and ≥50% decrease in CMAP area, or ≥30% decrease in area or amplitude over a short nerve segment (e.g., radial across the spiral groove, ulnar across the elbow, peroneal across the fibular neck)

Remyelination

In patients with a neurapraxic nerve injury (first degree) that is due to segmental demyelination and manifests as conduction block, the process of remyelination is usually rapid and may take up to 2–3 months for completion, provided that the offending cause (such as compression by hematoma or bony structure) is removed. For example, a patient who develops a wrist drop due to a purely demyelinating radial nerve injury at the spiral groove often recovers completely in 2–3 months. Hence, if follow-up NCSs are done after that time, remyelination is confirmed by resolution of the conduction block and restoration of the proximal CMAP. During that phase and for a short period after the reversal of conduction block, NCSs may reveal focal slowing across the injured nerve segment which was not present on the initial EDX study. This is best explained by the presence of newly formed thin myelin that was laid out by the Schwann cells. As the myelin thickens with time, focal slowing also disappears.

Reinnervation by Collateral Sprouting

Collateral sprouting is a process in which the surviving (intact) motor axons send axon terminals (sprouts) to the denervated muscles in an attempt to reinnervate these muscle fibers and restore muscle power. This is a quick and effective method of reinnervation that applies only to partial axon loss lesions where some axons escape injury and wallerian degeneration. Collateral sprouting is clinically effective in restoring function when only a modest number of axons are injured. In practice, it is most effective when less than 80+ of the axons are damaged. In very severe lesions, collateral sprouting may lead to little or no change of motor function.

Collateral sprouting in partial axon loss lesions starts as early as 1–2 days after injury. However, the early signs of reinnervation first become evident on needle EMG by one month, and are usually definite by 2–3 months postinjury. Immediately following nerve injury, there is a decrease in MUAP recruitment in affected muscles that is appropriate to the number of lost axons. In the first few weeks after injury, MUAPs of surviving axons retain their normal morphology. As collateral sprouting proceeds, muscle fibers become progressively incorporated to the territory of the motor unit.

Early on, the collateral axons (sprouts) have thin or incomplete myelin. Hence, action potentials along collateral sprouts conduct slowly. This is often reflected on needle EMG by MUAPs with satellite potentials (linked or parasite potentials), late spikes of the MUAP that are distinct and time-locked with the main potentials. The satellite potential trails the main MUAP because the newly formed nerve terminal may be long, or small and thinly myelinated, or both, resulting in slower conduction. When a satellite potential is suspected on needle EMG, it is useful to use a trigger line to demonstrate that this potential is time-locked to the main potential (Figure C1-7).

Figure C1-7 A motor unit action potential with a satellite potential recorded in a raster mode (arrow). Note that the second potential is always linked to the first potential.

Reinnervation MUAPs, including satellite potentials, may be unstable (Figure C1-8). The MUAPs may show evidence of intermittent nerve conduction blocking or neuromuscular junction blocking due insecure action potential transmission at the sprout or endplate, respectively. This results in individual muscle fibers being either blocked or come to action potential at varying intervals, leading to a MUAP that changes in configuration from impulse to impulse (amplitude or number of phases or both). Over time, the sprout matures and the conduction velocity increases and the satellite potential then fires more closely to the main potential, and ultimately fuses to become an additional phase or serration within the main MUAP complex (Figure C1-9). In general, MUAPs become more stable, more polyphasic, and longer in duration as collateral sprouting continues. In very chronic lesions, MUAPs are typically stable with long duration and high amplitude and little or no polyphasia, reflecting the maturity of all the nerve sprouts (Figure C1-10). Also, as reinnervation proceeds, there is a decline in the number of fibrillation potentials, since reinnervated muscle fibers will cease to generate this spontaneous activity.

Figure C1-8 Unstable motor unit action potential (arrow) preceded by a satellite potential that is also unstable (dashed arrow) from the brachioradialis muscle in a patient with severe radial nerve injury secondary to a compound humeral fracture, recorded four months postinjury. This complex potential is triggered upon using a trigger, delay line, and a raster mode. Note the extreme variation in the morphology of the unit and its satellite potential between subsequent discharges.

Figure C1-9 Polyphasic and stable motor unit action potential with long duration recorded from the tibialis anterior, eight months after a severe axon loss peroneal nerve injury at the fibular neck. This complex potential is triggered upon using a trigger, delay line, and a raster mode. Note that all components of the units are present with every discharge.

Figure C1-10 Large (long-duration and high-amplitude) motor unit action potential from the abductor digiti minimi of a patient with a remote severe ulnar nerve laceration above the elbow. This complex potential is triggered upon using a trigger, delay line, and a raster mode. Note that the main potential is followed by several small polyphasic satellite potentials that trail the main component.

On NCSs, the CMAP and SNAP amplitudes slowly increase in size with time as reinnervation continues. In mild to moderate nerve lesions, effective reinnervation may render the CMAP within normal values, and result in NCSs that do not clearly show evidence of a remote nerve injury. In these situations, however, needle EMG will continue to confirm the old injury by exhibiting large MUAPs that fire rapidly.

Reinnervation by Axon Regeneration

In complete or very severe axon loss peripheral nerve lesions, improvement is dependent solely or primarily on axonal regeneration that may occur spontaneously or following surgical repair. Unfortunately, in most cases of nerve injury, regeneration is slow and incomplete. In more severe axon loss lesions, the regenerating axons may not find intact endoneurial tubes and sometimes form a neuroma with tangled axons at the site of injury. In such lesions as well as lesions with complete transection, surgical repair is often needed.

In humans, the axons have to first traverse the injured segment. This may be achieved in 8–15 days when the endoneurial tubes are intact (second degree nerve lesion). Once the axons cross successfully, they continue to regenerate at a slow rate averaging 1 to 2 mm/day (or about 1 inch/month). Based on this, the timing of repeat EDX studies in complete or severe axon loss lesions depends on the site of injury in relation to the most proximal muscle that is expected to be reinnervated first. For example, following a median nerve injury in the middle of the arm, the first muscle expected to show reinnervation is the pronator teres muscle with its branch arising from the nerve in the antecubital fossa. If the distance between this lesion and the muscle is 5 inches, then the repeat study should be done about five to six months after the injury. The timing of surgical intervention is based on the fact that muscles that do not reinnervate after 18–24 months will undergo atrophy and fibrosis and their muscle fibers will not be more viable. With more proximal severe or complete axon loss nerve lesions, such as those of the lower brachial plexus or sciatic nerve, the target muscles to be reinnervated (hand muscles or leg muscles respectively) are situated far from the site of injury so that early surgical intervention is often necessary.

On needle EMG, the early signs of regeneration can be confirmed by the appearance of small, complex, unstable MUAPs, sometimes referred to as “nascent” MUAPs, that precedes the onset of visible voluntary contraction. These units appear first in muscles nearest to the site of the injury and progress distally, and hence are useful in assessing the advancement of this proximodistal regeneration. Nascent MUAPs are very low in amplitude and extremely polyphasic, with normal or increased duration. These small nascent MUAPs mimic the MUAPs seen with necrotizing myopathies. Nascent MUAPs are often unstable due to conduction or neuromuscular junction blocking and are associated with decreased MUAP recruitment (Figure C1-11). As reinnervation proceeds, nascent MUAPs that are unstable become transformed into stable, long-duration, and polyphasic MUAPs, reflecting increased numbers of muscle fibers per motor unit, full myelination of the regenerating axons, and the maturity of the neuromuscular junctions. Similar to what is seen in reinnervation by collateral sprouting, there is also a decline in the number of fibrillation potentials and the progressive improvement of the SNAP and CMAP on NCS. However, in these severe or complete nerve lesions, it is common that the CMAP and SNAP never return to baseline values and there is often permanent slowing and dispersion of the CMAPs due to the extreme variability in the diameter and myelination of the regenerated axons that results in significant differential slowing of conduction velocities.

Figure C1-11 Nascent motor unit action potentials from the quadriceps in a patient with severe femoral nerve injury following abdominal hysterectomy. This complex potential is triggered upon using a trigger, delay line, and a raster mode. Note the low-amplitude and highly polyphasic potential (dashed arrow), which is fairly stable and is followed by two distinct satellite potentials (arrows).

Aberrant Regeneration

Aberrant regeneration occurs when regenerating axons are misdirected into new end organs and is most common in axon loss nerve injuries that distort the endoneurial tubes (third degree or more) and in proximal peripheral nerve or root injuries. Misdirected fibers may not find endoneurial tubes and generate a neuroma at the site of the lesion. Regenerating motor axons in a mixed sensorimotor nerve may elongate into sensory nerves or vice versa. Motor axons may also get misdirected into the wrong muscles and result in co-contraction of muscles that can interfere with the intended function or cause abnormal movements.

The most common neurologic sequelae of aberrant reinnervation occur after facial nerve injury including after idiopathic Bell’s palsy. Aberrant regeneration between motor axons results in facial synkinesis, mainly contraction of the lower facial muscles on the affected side whenever there is an eye blink or vice versa. Other much less common, yet more publicized, examples of abnormal regeneration patterns are the “crocodile tears,” manifested as lacrimation of the ipsilateral eye during chewing, and the Marin-Amat syndrome, or “jaw-winking,” manifested as closure of the ipsilateral eyelid when the jaw opens.

Another example of aberrant regeneration occurs following injury to the C5 spinal root. Motor axons destined to the diaphragm may get misdirected to one or more shoulder muscles (biceps, deltoid, or spinati) with the result that the shoulder muscles fire in time with the respiratory cycle (breathing arm). Similar phenomena were recently reported form obstetric brachial plexopathies involving the lower plexus and T1 cervical roots and resulting in aberrant reinnervation of hand muscles from axons destined into intercostal muscles (breathing hand).