Axonal degeneration results from disorders of the nerve cell (neuronopathy) or axon (axonopathy). Pathophysiologic findings resemble those associated with nerve transection, varying only in degree (3). Separation of the distal axon from the nutritive cell body produces distal axonal degeneration and breakdown of the myelin sheath (Wallerian degeneration). Landau reported that muscle contraction produced by nerve stimulation distal to the lesion persisted for several days after transection, but then disappeared (4). Sensory and motor electrical responses similarly remain normal for several days, with stimulation distal to the transection. Within days, sensory nerve action potential (SNAP) and compound muscle action potential (CMAP) motor amplitudes diminish and ultimately disappear, although conduction along individual axons remains relatively normal before disappearance.

Following nerve transection, the needle EMG examination initially shows absent voluntary activity but no other findings. Gilliatt and Taylor (5) demonstrated spontaneous discharge of individual muscle fibers (i.e., fibrillation potentials) beginning 1 to 4 weeks after axonal degeneration, depending upon proximity to the transection (first appearing in muscles closest to the lesion). Fibrillation potentials reflect muscle fiber hypersensitivity to acetylcholine (ACh) and are associated with proliferation and migration of extrajunctional acetylcholine receptors (AChRs) on the muscle membrane (6). Initial findings include increased muscle fiber sensitivity to mechanical stimulation (typically in the form of positive waves), followed by sustained spontaneous activity (fibrillation potentials) at rest. The amplitude of fibrillation potentials and positive waves diminishes over time, proportional to muscle fiber atrophy, and provides a useful marker for assessing the duration of partial denervation.

Findings with partial or incomplete axonal lesions are similar, with decreased (instead of absent) sensory and motor responses, normal conduction along surviving axons, and reduced voluntary motor unit action potential (MUAP) recruitment (3). Abnormal spontaneous activity appears in denervated muscle fibers. After partial denervation, some denervated muscle fibers eventually are reinnervated by collateral sprouts from surviving axons, resulting in large motor units when the total number of motor units remains reduced (7). ACh hypersensitivity resolves once muscle fibers are reinnervated, and abnormal spontaneous activity disappears. Ballantyne and Hansen demonstrated that regenerating axons produced new motor units by reinnervating muscle fibers shed by abnormally large motor units (8).

The most common morphologic response to a variety of disorders producing neuropathy is a distal axonopathy (3). A variety of mechanisms exist to explain distal axonal degeneration in neuropathy, including failure of axonal transport of some nutrient required for maintenance of the distal axon, as proposed by Schaumburg et al (9). The concept of axonal atrophy is controversial, but it may represent a form of incomplete axonal lesion appearing before axonal degeneration. It is described at the
terminal axon as a reduced diameter of the distal axon (axonal stenosis). Because conduction velocity is proportional to axonal diameter, action potential propagation is reduced proportional to the size of the atrophic distal axon. Unlike axonal degeneration, sensory and motor amplitudes are not substantially reduced in axonal atrophy, and muscle membrane excitability is unaffected.

Disorders of the myelin sheath produce conduction abnormalities similar to those associated with focal nerve compression, whereby conduction may be slowed or blocked across the site of compression without producing axonal degeneration (10,11). In experimental models using a compressing tourniquet, Ochoa et al identified localized defects beneath the edges of the tourniquet, with telescoping of one myelin segment beneath the next at the node of Ranvier (12,13). This structural abnormality is thought to reduce or block ionic current flow by occluding the node, thereby slowing or preventing action potential propagation (10). Paranodal demyelination is associated with a variety of conditions, including focal compression and neuropathy, and conduction block is attributed to localized demyelination and intramyelin edema. Ochoa and Marotte demonstrated that chronic nerve compression produces similar findings with distorted myelin segments (segmental demyelination), exposed axon membrane, and paranodal remyelination with short internodal distances (11). Membrane excitability does not increase substantially with demyelinating lesions. Conduction slowing or block across a compressive lesion is relevant to the evaluation of generalized neuropathy, in that findings attributed to acquired demyelinating neuropathies have similar myelin abnormalities distributed throughout the peripheral nervous system, with combined demyelination and remyelination, short internodes, and reduced conduction velocity.

Reduced conduction velocity does not always indicate histologic abnormality such as axonal stenosis or demyelination, because metabolic disorders may produce conduction slowing without identifiable structural abnormalities (3). Among patients with hyperglycemia, increased conduction velocity 6 hours after normalizing glucose levels suggests that nonstructural changes account partially for conduction slowing (14). In association with the neuropathy attributed to chronic diabetes mellitus, the metabolic pathophysiology is poorly understood. Hyperglycemia is believed to produce a decrease in nerve myo-inositol and increased polyol pathway activity related to the increased conversion of glucose to sorbitol by aldose reductase. The reduced nerve myo-inositol leads to reduced Na⁺/K⁺-ATPase activity and a resultant increase in intracellular Na⁺ (15). In isolation, the resultant mild depolarization of the resting membrane potential decreases conduction velocity independent of structural alteration. Additional changes, including inactivation of sodium channels and axoglial disjunction, may also contribute to conduction velocity abnormalities (16). Finally, coexisting microvascular injury to the vasa nervorum and diminished production of nitrous oxide contribute to axonal ischemia and cellular damage, as may possible autoimmune damage mediated by antineuronal antibodies (17).

SNAPs and CMAPs are recorded using surface electrodes and percutaneous electrical stimulation (Figs. 11-1 and 11-2) (3). Response amplitudes and latencies are measured and conduction velocities are calculated as part of the evaluation. Conduction over an entire motor nerve is evaluated by F wave latency (Fig. 11-3). F wave measures accentuate mild generalized slowing because of the long conduction distances. Most normal values are age-dependent and some vary according to patient size and other factors (18,19,20,21,22,23). Improper electrode placement, inaccurate measurements, and failure to monitor and control limb temperature influence the results (3). Limb temperature is particularly important in the evaluation of neuropathy. Cooling decreases conduction velocity and increases amplitude, a combination of findings atypical for most pathologic processes. Limb temperature should be monitored, and cool limbs should be warmed to a surface temperature of between 32° and 36°C (23,24). In the context of the electrodiagnostic evaluation of neuropathy, failure to record limb temperature and warm cool limbs limits the ability to interpret the nerve conduction study results.

The needle EMG examination plays a limited but important role in suspected neuropathy (3). The needle EMG examination evaluates insertional activity (i.e., positive waves and fibrillation potentials) and volitional MUAP recruitment, size, and configuration. As a sensitive indicator of denervation, the needle EMG examination documents the distribution of axonal lesions, providing information from muscles inaccessible to nerve conduction study (e.g., the paraspinal muscles).
Recruitment refers to the sequential introduction of additional MUAPs into the interference pattern as force is increased. In the evaluation of neuropathy, the needle EMG examination provides an indication of ongoing or previous denervation. It also is used to define the distribution of axonal lesions, identifying disorders sometimes confused with or superimposed upon a generalized neuropathy. The amplitude of positive waves or fibrillation potentials and the configuration of MUAPs are used to distinguish acute from chronic disorders and provide an estimate of the rate of progression of axonal loss.

The initial goals are to determine the presence and location of sensory or motor involvement (see Table 11-2). Clinically apparent sensory loss may reflect a lesion proximal to the dorsal root ganglia, whereas abnormal SNAPs document peripheral involvement at or distal to the dorsal ganglia (3). Weakness and atrophy in combination with low CMAP amplitudes reflect abnormality of the lower motor neurons or axons. When present in isolation, these findings cannot localize the lesion more precisely (25,26). The distribution of needle EMG examination abnormalities is helpful in further localizing the abnormality.

The next goal is to identify to the extent possible the primary pathophysiology, such as axonal degeneration or demyelination (3). Axonal neuropathies usually are easily identified. They are characterized by reduced amplitudes, little evidence of conduction slowing, and neurogenic changes on needle EMG with evidence of active denervation and reinnervation (i.e., decreased MUAP recruitment and increased MUAP amplitude, duration, and polyphasia). Using a computerized model of the peripheral nerve, expected CMAP responses for a normal nerve are shown in Figure 11-4. Motor conduction abnormalities associated with axonal degeneration are shown in Figure 11-5. With a loss of
75% of the axons, the CMAP amplitude is markedly diminished, but conduction velocity is reduced only to the extent that is associated with the loss of the largest myelinated axons. There is no evidence of abnormal temporal dispersion.

Although primary demyelination is characterized by conduction slowing, overemphasis on mild or focal slowing is a common error in establishing the presence of demyelination (3). Important differences exist between hereditary and acquired demyelination. Hereditary disorders of peripheral myelin (e.g., hereditary motor sensory neuropathy type I) have uniform involvement of all myelinated fibers. Conduction along individual
fibers may be greatly reduced, but slowing is uniform; abnormal temporal dispersion is present only if conduction velocities are markedly slowed and there is substantial phase cancellation. Conduction slowing is disproportionate to the relatively preserved response amplitudes following distal and proximal stimulation (27). Evidence of conduction block is not a typical feature of hereditary demyelination unless conduction velocities are very slow and produce substantial phase cancellation.

Acquired demyelination is characterized by multifocal, nonuniform abnormalities. The disproportionate involvement of some myelinated fibers compared to others produces increased CMAP duration and a characteristic temporal dispersion of the CMAP (Fig. 11-6) (28). Partial conduction block results from transmission failure along the axon. Because the likelihood of conduction block in any given fiber is length-dependent, there is abnormal dispersion of the response and evidence of partial conduction block when the results of proximal stimulation are compared to those from distal stimulation. This is demonstrated in the model in Figure 11-7 after random, multifocal demyelination, in which propagation is slowed across a single demyelinated internode and blocked if two adjacent internodes are demyelinated (shown by the absence of the myelin sheath). Proximal stimulation produces a CMAP of slightly reduced amplitude and increased duration because of increased dispersion. The area beneath the negative phase of the CMAP is only slightly reduced. Proximal stimulation produces a low-amplitude, highly dispersed CMAP because of the variable amounts of demyelination in some fibers compared to others, producing an increased range of axonal conduction velocities. The initial component of the CMAP is greatly separated from the trailing portion of the CMAP, representing the fastest- and slowest-conducting axons respectively. Phase cancellation of the dispersed responses produces an additional reduction in the proximal CMAP amplitude. Conduction block in two of the axons further reduces the CMAP amplitude to a greater extent than could be explained by abnormal temporal dispersion alone.

Numerous criteria exist to identify acquired demyelination (Table 11-4), but all of the criteria have limitations (28,29,30,31,32). In some criteria, conduction velocity and distal latency thresholds are amplitude dependent. In general, conduction velocities that are less than 70% of the lower limit of the normal range cannot be attributed to axonal loss alone (28). The criteria are not intended to provide strict cutoffs and a “yes-or-no” response. Rather, the results should be used to raise or reduce suspicion for a disorder associated with substantial conduction slowing. The presence of fibrillation potentials and neurogenic MUAP findings does not exclude the diagnosis of demyelinating neuropathy because most hereditary and acquired demyelinating neuropathies have some superimposed axonal degeneration.

The remaining goals of the electrodiagnostic examination are to characterize the neuropathy’s distribution, severity, rate of progression, and
prognosis. The distribution of neuropathy is usually defined clinically, not electrodiagnostically, except for the needle EMG examination of paraspinal muscles in polyradiculopathy or the intrinsic foot muscles in mild axonal neuropathy. Neuropathic severity is best related to CMAP and SNAP amplitudes, because they are proportional to the extent of abnormality, particularly in axonal disorders. The needle EMG examination is useful in documenting very mild axonal neuropathies, as noted above. Defining severity in demyelinating neuropathy is difficult because conduction slowing is not usually associated with functional impairment. Conduction block results in weakness, but conduction block may be difficult to distinguish from abnormal temporal dispersion. Serial electrophysiologic studies are used to estimate the rate of progression, although a mixture of large- and small-amplitude fibrillation potentials can suggest active and chronic denervation.

Neuropathy is classified by a variety of means, including clinical, biochemical, pathologic, electrodiagnostic, or a combination thereof. The electrodiagnostic results provide information additional to that obtained from the clinical evaluation and are used to assign patients with suspected neuropathy to general categories, thereby directing the subsequent clinical and laboratory evaluations. The classification that follows separates these disorders into broad categories based on electrodiagnostic evidence of sensory or motor involvement combined with conduction slowing suggesting the presence of uniform or multifocal demyelination versus pure axonal loss lesions (2). This classification scheme is not inclusive and there is substantial overlap between categories. This overlap will be manifest by several forms of neuropathy appearing in more than one of the tables that follow. There also is some subjective component to the categorization scheme, requiring that the electromyographer use substantial clinical common sense. Nevertheless, when combined with the clinical history and examination, the electrodiagnostic results may suggest a specific diagnosis or direct the subsequent evaluation. Selected examples are provided in each category; a more extensive discussion exists elsewhere (2).

One issue not addressed in this classification scheme is symmetry. Most forms of peripheral neuropathy are relatively symmetric. There are exceptions, however, as noted earlier. The most common exceptions involve the different forms of
vasculitis producing multifocal abnormalities that can become confluent. Other examples of asymmetric diseases include multifocal motor neuropathy, hereditary neuropathy with liability to pressure palsy (HNLPP), and the neuropathies associated with dapsone, porphyria, leprosy, sarcoidosis, eosinophilia syndromes, and sensory ganglionopathies.