An improved understanding of the structure and physiology of peripheral nerves has led to great advances in the assessment and management of peripheral nerve injuries over the past few decades. This knowledge has been accumulated over the past 170 years, starting with Schwann's descriptions of the cells named after him in 1839 up to the sophisticated ultrastructural studies of recent years.

Peripheral nerves are unique structures that travel over long distances from the spine to the skin, muscles, and viscera. Because of the elongated course of these structures, they are more susceptible to trauma in many areas along their courses. Nerves anatomically and physiologically have evolved to minimize disruption of their function (i.e. to conduct an impulse to or from the neuron in the spinal cord or nerve root ganglion).

Gross Anatomy:

The cranial nerves and spinal nerves leave the central nervous system in pairs at specific levels of the nervous system, usually in relation to specific anatomic bony structures. The cranial nerves traverse bony foramina in the base of the skull before emerging peripherally. The spinal nerves go through intervertebral foramina. The nerves within the dura are termed nerve roots and vary in structure some­what from the more peripheral nerve. The spinal roots are divided into an anterior motor root and a dorsal sensory root. These coalesce near the point where the root exits through the dura. The roots differ from the more peripheral portions of nerves in that they are not invested with the large amount of connective tissue that is present distally.

After leaving the dura mater, the spinal roots in the cervical and lumbosacral regions join together into plexuses, which rearrange the course of many of the nerve fibers into identifiable peripheral nerves. These nerves then follow well-known anatomic pathways into the extremities. The cranial and thoracic nerves generally do not involve themselves in plexus formation and can be traced from the skull or the spine to their destinations.

The roots, plexuses, and peripheral nerves branch at various levels, sending fibers to specific muscles along their course and receiving sensory fibers from sensory endings in the skin, muscle, and viscera. These branches generally follow a fairly consistent pattern on joining the nerve trunk, but this can be variable. This pattern of branching has been helpful to clinicians assessing nerve function following injury and is one of the anatomic bases for electromyographic evaluation of nerve injuries.

The long course of the peripheral nerves makes them susceptible to damage from movements of the limbs. Areas of greater susceptibility exist in most peripheral nerves, and these areas of entrapment are well known clinically. In the upper extremity, the median nerve is entrapped as it traverses the wrist underneath the transverse carpal ligament. Less known but equally damaging is compression of the nerve at the ligament of Struthers at the distal extent of the humerus. The anterior interosseus branch may be caught in the pronator teres or in the fascia of the flexor muscles in the forearm. The ulnar nerve may be entrapped at the cubital tunnel or in the groove in the elbow, where it is also susceptible to trauma. Another area of entrapment is found at the wrist in Guyon's canal. The radial nerve is most susceptible to injury in the spiral groove of the humerus, where it is in close apposition to the bone. It also may be bound down as it makes a sharply angled dive to become the posterior interosseus nerve just below the elbow. In the lower extremity, the peroneal nerve lies very close to the head of the fibula in a superficial position, allowing it to be traumatized quite easily. It also is bound with fibrous tissue to some extent at this point. The nerve also is bound at the ankle. This is probably of little clinical importance, however. The posterior tibial nerve enters the arch of the foot through the tarsal canal, made up of ligaments of the arch and underlying bone, and is subject to trauma in this region. The sciatic nerve can be fixed in the sciatic notch, especially with marked flexion of the hips when squatting (hunkering). The sciatic nerve also pierces the piriformis muscle in a significant number of persons and may be compressed at that point. The femoral nerve is most susceptible as it enters the femoral triangle in the groin area.

Anatomy of the Nerve Trunk

Nerve trunks are made up of axons, Schwann cells, fibrous tissue, and vascular components. The ratio of neural tissue to supportive tissue is variable. Generally, connective tissue predominates, more so in areas where the nerve is in apposition to bone or joints, in areas of potential entrapment, or where the extremities are most movable.

The axons and their associated Schwann cells are coalesced into fascicles within the connective tissue matrix. The fascicles may be numerous or sparse in a nerve and are arranged variably from one area of the nerve to the next. In addition. the pattern of fascicular arrangement varies from nerve to nerve and also between individuals. Nerve fibers may change from one fascicle to another throughout the length of the nerve trunk.

The connective tissue matrix in which the fascicles lie has been divided into an epineurium and perineurium, within the fascicles, connective tissue is less obvious and is termed the endoneurium. The epineurium is a loosely organized sheath of connective tissue surrounding the nerve that also separates the fascicles within the nerve itself interfascicular epineurium) (Figure 1 A, B, C). The collagen associated with this connective tissue is generally arranged longitudinally, though the interfascicular epineurium may have some collagen fibers that traverse the nerve, This tissue provides protection, tensile strength. and supports the blood supply to the nerve. The outer portion of the sheath is relatively dense compared to the more inner regions, allowing for greater structural support (this is most useful in suturing cut nerves). The major blood vessels supplying the nerve lie in the epineurium.

The perineurium is a thin but dense layer of connective tissue arranged circularly about the nerve fiber fascicles. The cells lie in layers bounded by basal lamina on each side. Cells within the same layer have tight junctions between them and connections between various layers of cells are observed. The perineurium extends to the nerve endings. In the nerve root, the pia-arachnoid invests the fascicles. In this region, it is analogous to the perineurium. The tight junctions and layered structure of the perineurium serve, in part, as a blood-nerve barrier, resisting the penetration of substances through the perineurium.

The endoneurium consists of fibroblasts with processes that disseminate through the fascicles between nerve fibers and Schwann cells. The collagen fibers observed in the endoneurium tend to be longitudinal and often are closely apposed to the Schwann cells. This close relationship of endoneurium and Schwann cells helps form the tube through which regenerating nerve may pass following nerve injury.

These connective tissue structures serve to support and protect the underlying nerve tissue. They provide resistance to stretching, have some elastic properties, provide protection from penetration, and help dissipate compressive forces on the nerve, A nerve may, therefore, be stretched without impairment of axon integrity. Tolerance to stretching may vary, in part due to nerves tested, relationship to points of entrapment, and the condition of nerves studied. Generally, the nerves may be stretched up to about 25% to 30% before the axon is damaged.

Anatomy of the Nerve Fibers

The nerve fibers (axons) are contained in the fascicles. surrounded by the endoneurium and processes of the Schwann cells. Nerve fiber diameters vary from 20 µm down to under 1.5 µm. Fiber diameter diminishes as the nerve proceeds distally and also is variable from point to point along its course. The larger fibers are myelinated, whereas the smallest fibers are embedded in the Schwann cell walls (Figure 1 C). When viewed longitudinally, myelinated fibers have indentations in the myelin (nodes of Ranvier), which are the borders between adjacent Schwann cells. The axon is exposed in this area for a very short distance, but the exposed area is most critical for propagation of a nerve impulse. Schwann cell nuclei and cell bodies cover the myelin and. in turn, are covered by endoneurium. The axon is narrowed at the nodes and occasionally at other areas, such as under Schwann cell nuclei or other intracellular material within the Schwann cell. Unmyelinated fibers do not show the nodal pattern and are invested by Schwann cell processes. One Schwann cell may incorporate one or more small nerve fibers within its endoneurial tube.

Axons may branch along the course of the nerve, usually distally. This allows one neuron to innervate widely separated regions. Axon reflexes, such as the triple-flare response, may be explained by such branching, as might referred pain, though there is also evidence that referred pain may be a more central phenomenon. Nerve fibers lie very loosely within the fascicles. This allows some movement within the fascicle but also allows the nerve trunk to be moved or stretched without stretching the axons significantly. The connective tissue structures also tend to be lax, allowing much of the same protection against stretch injury.

Blood Supply of Nerves

The blood supply of a nerve trunk. consists of a network of longitudinally oriented arteries within the epineurium and over the nerve sheath. These arteries periodically receive branches from arteries in the surrounding tissues. forming an arborization similar to that observed in the mesentery of the bowel. If one of these nutrient arteries is damaged, as happens in surgical mobilization of the nerve, there is still an adequate blood supply in the nerve through these longitudinal anastomoses. Mobilization of a nerve up to 11 cm has not shown significant impairment of circulation.

Some interconnections between the longitudinal arteries then branch to deeper structures, pierce the perineurium in an oblique manner, and enter the endoneurial space. The capillaries in the endoneurium have tight junctions and form the blood-nerve barrier similar to the type of barrier seen within the brain. This blood­nerve barrier is of importance in some of the metabolic neuropathies, and the breakdown of this barrier in nerve injuries may be of some importance during repair. Although the basic metabolic support of an axon comes from the cell body, there is considerable evidence that the endoneurial blood supply is very important to maintain axonal function. In clinical situations where the blood supply to a nerve has been restricted, symptoms have occurred.

The Schwann Cell

The Schwann cells have an intimate relationship with the axons. They probably have a trophic effect on the axons, help nourish the axon, and help form the "tube" through which the axon travels. The origin of these cells is disputed, but most feel that they migrate from the neural crests along with the axons. The Schwann cells are the source of the myelin in peripheral nerves, analogous with the oligodendroglial cells of the central nervous system. Myelinated axons are invested in myelin by a spiraling of a Schwann cell process about them. Nonmyelinated fibers lie embedded within a Schwann cell. Often such a cell may be surrounding several such axons (Figure 1 C). With axonal death, myelin is destroyed, but the Schwann cells survive and frequently increase in numbers. If the axon regenerates, the Schwann cell reinvests the axon, and forms myelin if needed.

Physiology

Transmission of a nerve action potential is dependent on the integrity of the axonal membrane, Damage to this membrane will interfere with normal neural function. In the steady state, this membrane has a transmembrane electrical potential of about -70 to -90 mV with the inside of the axon being negative.

The reason for this potential difference lies in both the structure of the membrane and the distribution of the solutes in the intracellular and extracellular spaces. The cell membrane is composed of a double layer of phospholipids with protein molecules scattered over the surface but also forming transmembrane channels for ions to cross the membrane. The membrane acts as a semipermeable membrane that allows some molecules to cross it while restricting others. Nerve membrane is quite permeable to K⁺ ions, Cl^- ions, and less so to Na⁺ and other larger ions. Intracellular K⁺ concentration is markedly higher than that found outside the cell. If the K⁺ were free to diffuse across the membrane, there would be an efflux of the ion. The high extracellular Na⁺ would tend to try to get into the cell, where Na⁺ is low. The membrane is less permeable to this ion, so less of a flow is present. The negative potential resists these flows and maintains the stability of the membrane. Other ions also participate in vari­ous gradients across the membrane and add their electrotonic forces to the equation, pro­ducing the final resting membrane potential. The transmembrane potential of K⁺ is very close to the actual resting membrane potential. In addition, an energy-dependent Na⁺-K⁺ "pump" moves Na⁺ ions out of the cell and K⁺ into the cell, maintaining the relative concentrations within the cell. When a chemical or electrical stimulus is applied to this system, a series of events occurs that terminates in the generation of a nerve action potential. Such a stimulus needs to reverse (or depolarize) the negative polarization of the membrane in order to develop the action potential. When a critical level of depolarization is reached, there is a sudden reversal of polarity of the membrane to about +30 - +40 mV and an action potential is formed. Each time that threshold is exceeded, the same amplitude of reversal occurs (the "all or none response"). Associated with this event is a sudden, brief change in membrane permeability of Na⁺ that flows into the cell. About 1 millisecond later, a similar but longer-duration change occurs in the K⁺ permeability, which acts to end the action potential and repolarizes the membrane. During these brief periods of increased permeability, very few Na⁺ ions actually enter the cell, but the Na⁺ - K⁺ pump will work to remove those few ions from the internal milieu.

When the action potential is generated, a current flows into the active areas of the membrane of the axon from the extracellular space. This flow then goes down the axon and exits the axon across the normal surrounding areas of the membrane into the extracellular space, completing the circuit. If the electrical changes in these normal regions exceed the threshold levels, then a new action potential is generated and the action potential is propagated down the axons by way of these local circuits. In unmyelinated fibers this process is relatively slow; however, the addition of myelin speeds up this process considerably. With the insulation provided by the myelin sheath not allowing the exit of electrical current except where it is absent (nodes of Ranvier), the flow of electrical current leaves the axon at some distance from the action potential (one to three nodes away). A new action potential is thus generated much farther down the nerve, allowing it to propagate down the nerve at a much faster rate (saltatory conduction). The longer the internode distance, the more rapidly the axon will conduct the action potential.

It should be noted that the metabolism in an axon is greater in the nodal regions. Mitochondria are grouped in these regions, providing for the energy needed to sustain the Na⁺-K⁺ pump. The propagation of an action potential requires no energy, but maintenance of the resting membrane potential does.

Axon metabolism, in part, depends on substances produced in the cell body, which are conveyed distally by axoplasmic flow. Both a slow and a fast transport system occur down the axons, and, in addition, there seems to be a flow in the opposite direction. There probably are some Schwann cells and endoneurium contributions to axonal metabolism. Certainly, oxygen and carbon dioxide gaseous exchange occurs in the nodal areas, as vascular occlusion of the vasa nervorum will cause malfunction of the axon.

Clinical Electrodiagnosis

Electrodiagnostic tests are an extension of the bedside examination of the peripheral nervous system. They add objective data about the function of the peripheral nerve and should provide accurate localizing information if a nerve is damaged. These tests are useful when minor changes are unable to be identified clinically or when the functions tested are in locations that are difficult to examine clinically. They shed light on pathophysiologic mechanisms that otherwise would be difficult to delineate at the bedside (e.g. differentiating neuropraxia from a more severe injury to the axon, or delineating sensory nerve root involvement from a plexus injury).

Clinical electrophysiologists have to be well versed in neuroanatomy, topographic anatomy, and nerve physiology to make meaningful assessments of nerve function. The procedures require discrete placements of the recording electrodes, needles, and stimulating probes to be accurate. Inaccurate placement of either the stimulating or recording electrodes greatly diminishes the value of the studies. In addition, knowledge of the disease processes affecting peripheral nerves is of great importance to the examiner in order for him or her to interpret the test findings in the proper context of the nerve dysfunction. Clinical electrophysiologic testing of the peripheral nervous system can be divided into two broad categories: (1) nerve conduction studies with their related studies, somatosensory evoked responses, and long latency reflexes (H-reflex, F wave); and (2) electromyography (EMG).

Nerve Conduction Studies

The function of the peripheral nerve is to transmit an electrical impulse from one point to another. The electrical stimulus normally comes from the nerve cell body or from receptor structures. In nerve conduction studies, however, the nerve is stimulated by an external electrical source. When the nerve is near the surface of the body, skin electrodes may deliver the shock. Deeper nerves require needle electrodes. With nerves exposed at surgery, stimulating electrodes may be applied directly to the nerves. Stimulation is made with supermaximal shocks to make sure that all nerve fibers are stimulated and that a maximal response is obtained. Less than maximal stimulation may give spurious results.

Recording electrodes may also be surface, needle, or directly applied types. They may be placed over muscle to record the evoked muscle action potential, or they may be applied directly over a nerve to record a nerve action potential. In sensory nerves, the potential is purely a sensory nerve action potential (SNAP), but over a nerve trunk, elements of both motor and sensory nerve action potentials are present (mixed nerve action potential). Conduction velocities measure the fastest conducting fibers of the nerve.

Motor nerve conduction studies are done by stimulating the nerve at two or more points along the course of the nerve and measuring the evoked motor responses from an appropriate muscle. If the nerve length can be measured between the stimulus sites, conduction velocities can be calculated, various segments along the nerve may be tested, allowing for greater precision in identifying an area of dysfunction. Motor nerve conduction velocities vary from nerve to nerve but generally are comparable from side to side: therefore, it is most helpful to have information from the "normal" nerve on the opposite side to compare with the target nerve being evaluated. Exact normal velocities expressed in meters per second vary somewhat from lab to lab but generally are similar.

Sensory nerve conduction studies may be performed in two ways. A stimulus may be applied distally to a pure sensory nerve and recorded proximally (orthodromic) or to a nerve trunk and recorded distally off of the pure sensory branch (antidromic). Both methods achieve comparable results, though antidromic stimulation may elicit motor responses that may obscure the smaller sensory response. Like motor conduction studies, comparison with the other side is often helpful.

Conduction velocities are only part of the information that can be obtained from the test. The amplitude of the response, whether motor or sensory, is a reflection of the numbers of axons that are conducting an impulse. Low­amplitude responses suggest problems with or loss of axons between the nerve cell body and the site of recording. The presence of normal sensory nerve action potentials in the presence of severe sensory loss points to a lesion proximal to the dorsal root ganglion, suggesting an avulsion of a nerve root.

Somatosensory evoked potentials (SEPs) are most helpful in evaluating the proximal segments of a peripheral nerve that normally are inaccessible to conventional nerve conduction studies. A stimulus is usually applied to a nerve peripherally, and recordings of potentials are made from proximal nerve sites, areas of entry into the spinal cord, sites on the spinal cord, and more proximal areas within the brain. SEPs, therefore, allow evaluation of the entire sensory system. Proximal nerve segments, therefore, can be compared with the more peripheral segments. SEPs should be performed unilaterally and also simultaneously for comparison between the two sides.

The H-reflex, first described by Hoffmann, is the electrical evocation of the spinal monosynaptic reflex. It therefore allows for the assessment of both proximal sensory and proximal motor nerve pathways. It is best elicited from the calf muscles but also is seen in the flexor carpi radialis. The stimulus in the leg is applied to the posterior tibial nerve, allowing evaluation of conduction in the sciatic nerve and in the S1 root. In the arm, the median nerve, the lateral cord and upper trunk of the brachial plexus, along with the C6 and C7 root, may be assessed with the H-reflex.

F waves measure the motor conductions along the proximal portions of the nerve. The stimulus impulse travels toward the cord in the motor axon (antidromic). Upon reaching the motor neuron in the anterior horn, it reverses itself and goes peripherally along the same axon to the muscle (orthodromic). Unlike the H-reflex, which can be elicited only in a few nerves, the F wave response may be obtained from any accessible motor nerve.

Nerve conduction studies may be affected by numerous factors. Nerve conduction velocities are faster in larger nerves and those nerves that are myelinated. They tend to be faster in the proximal segments than distally. Higher temperatures may increase conduction velocities. This, in part, may account for the above observation, Conversely, cool temperatures slow conductions, giving the impression that nerve conduction velocities are slower in wintertime when the extremities tend to be colder. Constant temperature conditions in the examining room minimize these effects. Age affects conduction velocities, with infant velocities being low and speeding up to adult levels at about 3 years of age. Ischemia within a limb also may slow conduction.

The greatest slowing in conduction velocities occurs with demyelinization or compression of the nerve, or both. Neuropraxia and nerve lacerations abolish nerve conduction across the lesion; however, after a neuropraxic lesion, the distal segment remains excitable and conduction remains normal. After a transection, the distal nerve may remain excitable for 4-7 days after the injury and then stop functioning. Reports of nerve conduction studies should include (1) distal latency (the time required to elicit a response in the distal most studied segment of a nerve); (2) amplitude of the elicited response (as noted previously, this gives some idea of the numbers of functioning axons within the nerve); (3) conduction velocities (this is the rate of transmission of an impulse between two points on a nerve. The segment being tested should be indicated in the report); and (4) normal ranges for the lab performing the test (standard textbooks of electrodiagnosis often contain tables of normal values for reference where the norms are not otherwise available).

Electromyography

EMG tests the electrical activity of muscles and indirectly the function of both the upper motor neuron system and the lower motor neuron. Defects anywhere in this pathway will alter the EMG findings. The basic unit of muscle activity is the motor unit. This consists of a variable number of muscle fibers innervated by one neuron. When the neuron transmits its impulses, all of its component muscle fibers are activated and an electrical potential is generated. This potential represents the summation of electrical events in the individual muscle fibers within the motor unit and can be recorded by an electrode placed nearby. Needle electrodes are used and multiple locations must be sampled within each muscle in order to assess the numbers of motor units in the target muscle. When a needle is inserted into a normal muscle, a brief burst of electrical activity occurs that subsides immediately. This "insertional activity" may be altered by both denervation and muscle disease. It may be helpful in differentiating between them. The muscle should be observed next in the relaxed state. In normal muscle, no electrical activity occurs at rest. Denervated muscle will demonstrate fibrillations and positive sharp waves as individual muscle fibers become hyperexcitable and discharge spontaneously. The muscle is examined next during increasing volitional movement. Motor unit potentials appear with minimal activity.

As strength increases, new motor units will be recruited until, ultimately, individual motor units cannot be identified (interference pattern). Denervation decreases the numbers of motor units available for recruitment or, if complete. will show no motor unit activity. There also may be changes in the form, amplitude, and duration of individual motor units as the result of denervation. Muscle disease also may alter these parameters of motor unit potentials that are observed. Reports generated by the EMG should reflect information from observations in all four of the preceding areas of assessment.

The EMG requires knowledge of derivation of nerve fibers going to each muscle. Nerve fibers in the nerve roots pass through plexuses and may go to a large number of muscles through various peripheral nerves. When evaluating injury to the peripheral nervous system, muscle should be tested in a logical sequence in order to determine the location of the lesion. Evaluation of a nerve root lesion should include EMG of the paraspinous muscles, as these muscles are innervated by the posterior ramus of the spinal nerve that branches at the nerve root.

Following nerve injury, the EMG changes of denervation will not be present until 2-3 weeks have elapsed. With this in mind, EMG investigation should not be attempted until 3 weeks after an injury if one is to obtain full benefit from the examination. This wait also allows soft tissue changes to resolve in order to better appreciate the location of muscles and the nerves to be tested. EMG should be done with great care in anticoagulated patients and probably should not be done in patients with infections in areas through which the needle electrodes might traverse. No other contraindications to this procedure exist.