Electrodiagnostic evaluation of the neuromuscular junction

INTRODUCTION — Repetitive nerve stimulation (RNS) and single-fiber electromyography (SFEMG) are important confirmatory tests for the diagnosis of disorders of the nicotinic neuromuscular junction, particularly for myasthenia gravis, Lambert-Eaton myasthenic syndrome (LEMS), and botulism. This topic will review the basic principles of neuromuscular transmission and electrodiagnostic testing with RNS and SFEMG.

Other clinical aspects of neuromuscular junction disorders are reviewed separately. (See "Clinical manifestations of myasthenia gravis" and "Diagnosis of myasthenia gravis" and "Lambert-Eaton myasthenic syndrome: Clinical features and diagnosis" and "Botulism" and "Neuromuscular junction disorders in newborns and infants", section on 'Infant botulism'.)

ANATOMY AND PHYSIOLOGY OF THE NEUROMUSCULAR JUNCTION — The nicotinic neuromuscular junction is a complex, specialized structure incorporating the distal axon terminal and the muscle membrane that allows for the unidirectional chemical communication between peripheral nerve and muscle. It consists of the presynaptic nerve terminal, the synaptic cleft, and the postsynaptic endplate region on the muscle fiber. Acetylcholine serves as the neurotransmitter for voluntary striated muscle. The sections that follow provide a simplified summary of the morphology and neurophysiology of the neuromuscular junction that permits an adequate understanding of the electrodiagnostic testing principles (figure 1).

Morphology — The presynaptic region of the nicotinic neuromuscular junction consists of a specialized terminal nerve structure surrounded by Schwann cells. The presynaptic region includes voltage-gated calcium channels (VGCCs) embedded in the nerve membrane, numerous mitochondria, and acetylcholine-containing synaptic vesicles. Each vesicle contains 5000 to 10,000 molecules of acetylcholine, known as a quantum [1,2]. The populations of vesicles behave as if they are divided roughly into three pools [3]:

●The primary pool for immediate release

●The secondary or mobilization storage pool

●The reserve storage pool

The primary pool is located in the vicinity of the "active zones" in the presynaptic nerve terminal and consists of approximately 1000 vesicles that are available for immediate release. The secondary pool consists of approximately 10,000 vesicles that can be mobilized for release within a few seconds. The reserve pool consists of approximately 100,000 to 300,000 vesicles that must be transported from the axon or cell body to replenish the other stores prior to becoming available for release.

The synaptic cleft is the roughly 50 nm space between the nerve terminal and the muscle membrane. Acetylcholinesterase, the enzyme responsible for the degradation of acetylcholine, is released into this area.

The postsynaptic membrane is a specialized structure within the muscle fiber membrane consisting of complex junctional folds that increase the surface area approximately 10-fold [3]. Numerous nicotinic acetylcholine receptors are embedded in the membrane at the crests of the folds [4]. The acetylcholine receptors are complex structures containing four subunits surrounding an ion channel (figure 2). When the two acetylcholine binding sites are filled, there is a conformational change in the receptor structure that allows for an increase in sodium conductance and thus a local change in the muscle membrane potential, as described in the next section below.

Physiology of acetylcholine release — The nerve action potential propagates down the motor axon to the presynaptic nerve terminal. The presynaptic voltage-gated calcium channels are depolarized, allowing the flow of calcium ions into the presynaptic terminal region [5]. In the presence of calcium, the acetylcholine-containing vesicles of the primary pool combine with the terminal membrane at the active zones, releasing their contents into the synaptic cleft. The presynaptic VGCCs serve to terminate calcium entry, and thus halt the release of acetylcholine. The acetylcholine diffuses across the synaptic cleft and combines with the postsynaptic acetylcholine receptors, generating a localized, graded, nonpropagating, depolarizing current with no refractory period known as the endplate potential (EPP). When the EPP reaches threshold, an all-or-none depolarization of the postsynaptic muscle membrane occurs, inducing a muscle fiber action potential and ultimately a muscle fiber contraction. Normally, approximately 10 to 20 percent of the primary pool or 100 to 200 quanta of acetylcholine are released at a given neuromuscular junction in response to an action potential [6,7], creating an EPP that is several-fold greater than the threshold to depolarize the muscle fiber [8,9]. Under normal circumstances, this considerable safety factor ensures that a depolarization of the distal nerve segments will induce reliably a contraction of the innervated muscle fiber.

The probability of acetylcholine release is primarily dependent on two factors:

●The amount of intracellular calcium

●The number of acetylcholine vesicles in the primary pool at the terminal nerve region

It takes approximately one to two seconds for the primary pool of acetylcholine vesicles to be replenished. Furthermore, it takes approximately 100 ms to return calcium concentrations to their pre-depolarization equilibrium. Therefore, rates of repetitive nerve stimulation in the 1 to 5 Hz range or weak voluntary muscle contractions result in the depletion of available acetylcholine vesicles without significantly altering the calcium concentration. This process results in a smaller number of quanta released with each subsequent nerve impulse. At high rates of stimulation (10 to 50 Hz) or following high-intensity voluntary muscle contractions, there is a sustained, significant increase in the intracellular calcium concentration because the influx of new calcium ions occurs before the previously introduced calcium has diffused back out of the presynaptic terminal [10]. In addition, there is a less important increase in the size of the immediately available acetylcholine pool due to augmentation from the secondary storage pool that results in an increase in quantal release following subsequent nerve stimulations.

In normal muscle, these maneuvers have little impact on function, as the safety factor inherent in the neuromuscular junction transmission assures a reliable muscle fiber contraction [3,9]. The compound muscle action potential (CMAP) is essentially unchanged due to the large safety factor that promotes and maintains near maximal muscle fiber activation with supramaximal nerve stimulation despite fluctuations in the EPP.

With a neuromuscular junction disorder of the postsynaptic membrane such as myasthenia gravis, the number of acetylcholine receptors is reduced, and this in turn reduces the safety factor for muscle fiber activation (see "Pathogenesis of myasthenia gravis"). In this situation, some muscle fibers fail to reach threshold due to the decreased presynaptic acetylcholine release that normally occurs at slow rates of repetitive stimulation, thereby leading to a decrease in the CMAP amplitude, a phenomenon known as the decremental response (waveform 1). By contrast, the modest increase in the presynaptic release of acetylcholine that occurs at high rates of stimulation or with maximal intensity voluntary muscle contractions will usually have a minimal effect on the CMAP amplitude in the great majority of patients with myasthenia gravis, since the safety factor for transmission is usually adequate for contraction of all the muscle fibers even before the increased amounts of acetylcholine become available. In the most severe cases of myasthenia gravis, the baseline CMAP amplitude can be reduced at rest due to failure of action potential generation in some fraction of neuromuscular junctions because of incomplete muscle fiber depolarization from the reduced EPPs (ie, reduced safety factor). A modest increase in CMAP amplitude, known as the incremental response, can therefore be seen in this setting as the increased acetylcholine release with tetanic stimulation increases the EPPs in most muscle fibers beyond threshold.

In disorders of the presynaptic terminal, such as Lambert-Eaton myasthenic syndrome (LEMS), the number of released vesicles is markedly reduced at baseline (see "Lambert-Eaton myasthenic syndrome: Clinical features and diagnosis", section on 'Pathophysiology'). Therefore, the pattern of evoked response amplitudes is different from that seen in myasthenia gravis. In most cases of LEMS, the reduced acetylcholine released in response to nerve stimulation results in reduced EPPs and a correspondingly reduced CMAP amplitude. High-frequency stimulation can markedly increase the amount of available presynaptic intracellular calcium, which in turn can augment the number of released quanta of acetylcholine and thus enlarge the size of the resultant CMAP amplitude, consistent with an incremental response. The changes can be quite dramatic with CMAP amplitude increases of several hundredfold (waveform 2) [11]. Following a weak voluntary contraction or at low rates of repetitive stimulation (1 to 5 Hz), small decremental responses can be seen in presynaptic neuromuscular junction disorders due to the reduction in the size of the available acetylcholine vesicle pool with an unaltered calcium concentration (waveform 3).

INITIAL ROUTINE ELECTRODIAGNOSTIC STUDIES — A complete electrodiagnostic study including nerve conduction studies (NCS) and electromyography (EMG) is a standard component of the evaluation of neuromuscular junction disorders. However, additional techniques may be required to assess the fidelity of neuromuscular junction. (See "Overview of nerve conduction studies" and "Overview of electromyography".)

Under most circumstances, these studies are normal in postsynaptic neuromuscular junction disorders, such as myasthenia gravis. Presynaptic neuromuscular junction disorders such as the Lambert-Eaton myasthenic syndrome (LEMS) often have a pattern of low compound muscle action potential (CMAP) amplitudes with normal motor latencies and normal sensory nerve action potentials. It is only with RNS or SFEMG that the true nature of the disorder is demonstrated.

REPETITIVE NERVE STIMULATION — Repetitive nerve stimulation (RNS) represents a modification of conventional motor nerve conduction studies.

Physiologic response to repetitive nerve stimulation — The interval between repeated motor nerve stimulations is varied to manipulate the neuromuscular physiology described above (see 'Physiology of acetylcholine release' above). Slow, repetitive stimulation is designed to maximally stress the neuromuscular junction safety factor. Since it takes approximately one to two seconds to replete the primary pool of presynaptic acetylcholine vesicles, slow rates of repetitive stimulation (1 to 5 Hz) deplete the number of available quanta for release with subsequent depolarizations. The percent of the remaining quanta released with the next stimulation will be determined by the presynaptic calcium concentration. Repetitive stimulation rates of 1 to 5 Hz allow for the re-equilibration of the presynaptic calcium concentration between stimulations so with the smaller primary pool, a smaller number of quanta will be released with each subsequent stimulation for at least the first five or six stimuli in a train. In the normal neuromuscular junction, this reduced amount of acetylcholine is still adequate to reliably depolarize the muscle fiber. However, in pathologic states where the safety factor for neuromuscular junction transmission is considerably reduced, some fibers will fail to depolarize in the later stimuli of a train and the compound muscle action potential (CMAP) amplitude will drop, which is referred to as a decrement.

The presence of this decremental response on RNS has been the neurophysiologic hallmark of myasthenia gravis. Suspicion of myasthenia gravis (MG) remains the major indication for the procedure. (See "Diagnosis of myasthenia gravis", section on 'Nerve conduction testing with repetitive nerve stimulation'.)

Technique — The technique of RNS incorporates the surface stimulation and surface recording procedures used in routine motor nerve conduction studies (see "Overview of nerve conduction studies", section on 'Methodology'). The CMAP obtained after single supramaximal stimulation represents the summation of the electrical activity from the individual muscle fibers of the entire muscle. The negative peak area (or alternatively the amplitude) is a reflection of the number of muscle fibers activated.

Before delivering the RNS train, the joint is immobilized if possible. As an example, taping the fingers together or applying a hand brace can prevent finger movement, and thus movement artifacts, when stimulating the ulnar nerve. Some muscles are not amenable to immobilization, such as the facial or trapezius muscles. The stimulator is taped in position and held to prevent movement of the stimulating electrode, which leads to a submaximal stimulation and a reduction of the CMAP amplitude or area. This problem is one of the major causes of inaccurate data. (See 'Artifacts and pitfalls of slow RNS' below.)

Slow RNS — With slow RNS, a train of five to nine stimuli is delivered to the peripheral nerve and the resultant CMAPs are sequentially recorded (waveform 4). A slow rate (1 to 5 Hz) causes depletion of the immediately available acetylcholine pool without increasing presynaptic calcium. This procedure reduces the neuromuscular junction transmission safety factor and is usually used to study patients suspected of having a postsynaptic neuromuscular junction disorder such as myasthenia gravis. The sensitivity of the procedure is further enhanced by retesting the muscle after a period of tetanic stimulation [12]. Electrical high-frequency (tetanic) stimulation is extremely uncomfortable for the necessary 45 to 60 seconds, but the same effect is obtained in the cooperative patient by one minute of maximal isometric exercise specifically in the muscle being tested. Although initially inducing facilitation (post-tetanic potentiation) by an increase in the presynaptic calcium and increased mobilization of the acetylcholine pool, this effect wanes after two to four minutes, and the depletion of presynaptic acetylcholine reduces the neuromuscular junction safety factor, increasing the decrement in a subsequent train of five to nine stimulations (post-tetanic exhaustion). While several different protocols have been suggested [12,13], we suggest RNS testing with trains of five to nine potentials before exercise and after one minute of maximal isometric activation at one-half, one, two, and four minute intervals. A typical series of RNS stimuli trains is demonstrated for both a normal patient (waveform 5) and a patient with myasthenia gravis (waveform 6).

The degree of decrement is usually expressed as the percent of the decrease in amplitude (A) or area of the lowest response (usually the fourth or fifth) relative to the first potential:

Artifacts and pitfalls of slow RNS — The most common causes of misinterpreted slow RNS data relate to a series of common artifacts. Each must be carefully assessed before a final conclusion is reached.

●The stimulus intensity must be supramaximal for each stimulation. Most commonly, the initial stimulus intensity is supramaximal, but if the stimulator moves during the course of testing, it can deliver a submaximal stimulus and a pseudodecrement may occur. Alternatively, the resulting waveforms may have considerable wave-to-wave variability in a seemingly random pattern (waveform 7) in contrast to the gradual decline in amplitude and area typical of a biologically significant decrement (waveform 1). It is easy to understand how the decrement in the figure showing the resulting waveforms could be reported as significantly abnormal, 45 percent (waveform 7), when in reality the problem is entirely technical. To avoid this artifact, we suggest taping the stimulator in place and holding it throughout the course of testing, even between trains of stimuli, to prevent movement.

●Patient movement or electrode contamination alters the interface between skin and recording electrode and between the electrode and the signal generator (the muscle). This may induce instability in the baseline of the CMAPs. Significant baseline variability impairs the accurate determination of the amplitude and often invalidates the results. The baseline for each waveform should be identical and create a level floor for the determination of amplitude (waveform 4). Therefore, when possible, the limb being studied should be immobilized to minimize movement of the recording electrode. In general, the more distal the muscle, the more easily it is immobilized.

●Acetylcholinesterase inhibitors should be discontinued prior to study as they can theoretically reduce abnormalities, causing a false-negative result [13]. Six hours is probably adequate, but we suggest stopping acetylcholinesterase inhibitors 12 or more hours before the RNS study.

●As is true for all nerve conduction studies, the control of temperature is critical. Reduced temperature increases the safety factor of neuromuscular transmission and thus reduces the decrement, resulting in a false-negative result [14].

The careful evaluation of all waveforms will help prevent inappropriate diagnoses based upon artifact. The physiologic abnormalities in myasthenia gravis tend to occur smoothly among potentials in a series without abrupt or inconsistent fluctuations in the individual waves (waveform 1). The maximal decrement is usually present between the first and either the fourth or fifth potential in a series. In the sixth to ninth potentials (waveform 1), there is either a mild increase in amplitude/area or a leveling off, most likely due to the eventual repletion of the acetylcholine primary pool. This pattern is characteristic of myasthenia gravis, and if it is not present, the data must be viewed with skepticism.

The final test for the accuracy of a decrement is its repair following brief high-intensity isometric exercise. This can be done after a single RNS train demonstrates a decrement or after the occurrence of post-tetanic exhaustion where 15 seconds of a maximal isometric contraction followed by a train of five supramaximal stimuli can demonstrate the resolution of the decrement by increasing the presynaptic calcium concentration and thus the amount of acetylcholine released (waveform 6).

Nerve selection for slow RNS — The selection of nerves for study should be based upon the accessibility of a nerve for study and the distribution of muscle weakness. The likelihood of a positive result increases if a clinically affected nerve is studied, but some nerves are more easily evaluated with RNS than others. The ulnar nerve is most commonly tested, with recordings made from the abductor digiti minimi muscle (ADM) because of the ease of immobilization and stimulation. However, the probability of an abnormal study is higher in more proximal muscles. One study suggested that recording over the extensor indicis proprius muscle, innervated by the radial nerve, is technically straightforward with greater sensitivity compared with the ADM [15]. In another study, abnormal decrement in the median nerve (recorded over the abductor pollicis brevis [APB]) had a frequency of abnormality equivalent to the spinal accessory nerve and was documented in some patients with MG who had normal responses in the spinal accessory and/or facial nerves [16]. Myasthenic patients with abnormal median nerve recording were more likely to have limb weakness than those with normal recording. Any shoulder or leg muscle amenable to study would be the next recommended location if the selected muscle was clinically weak, particularly the axillary nerve (deltoid muscle recording) [17] and the peroneal nerve (tibialis anterior). If not, the spinal accessory nerve recording over the trapezius muscle is particularly uncomplicated and would be recommended [18].

Although technically more difficult, the highest yield with RNS is typically obtained by studying a facial muscle. Thus, RNS of a facial muscle should be performed if the study of more accessible muscles is not diagnostic. The orbicularis oculi muscle is the facial muscle most often examined. The trigeminal nerve (masseter) [19] and other segments of the facial nerve (nasalis [20,21], orbicularis oris) are also useful at times. One report of 45 patients with myasthenia gravis who had RNS found that sensitivity for the diagnosis was maximum with bilateral examination of the orbicularis oculi (facial nerve), trapezius (accessory nerve), and anconeus (radial nerve) muscles [22].

The importance of respiratory muscle involvement in myasthenia gravis led to investigations of phrenic nerve RNS [23-25]. Patients with isolated oropharyngeal and/or diaphragmatic involvement can be particularly difficult diagnostic problems [26]. In one phrenic nerve RNS report, the investigators stressed the importance of having the patients hold their breath during the stimulus train [23], but this can be difficult for some individuals with diaphragmatic weakness. We have had difficulty obtaining reproducible data utilizing these techniques, but they represent the only electrophysiologic procedures available for the evaluation of the neuromuscular junction of the respiratory muscles.

Slow RNS protocol — For suspected myasthenia gravis, slow RNS is performed at 2 or 3 Hz using trains of nine stimuli at rest. If a decrement of ≥20 percent is present, 10 seconds of maximal isometric exercise is performed to see whether the decrement repairs. If it does not, another 10 seconds of exercise is recommended. No further RNS testing of that nerve is necessary if the decrement is >20 percent, repairs with exercise, and contains the other physiologic characteristics of decrement (see 'Artifacts and pitfalls of slow RNS' above). If not, testing for postexercise exhaustion is performed with one minute of maximal isometric exercise, followed by trains of nine stimuli every minute for at least four minutes and at times six to eight minutes if an ambiguous result is seen. Concentric needle examination of several muscles is also suggested at a minimum, including any muscle with an abnormal RNS, since any illness with reinnervation and several different types of muscle disease can have a decrement on slow RNS, as discussed in the next section below.

Interpretation — In the great majority of cases, slow RNS is performed to evaluate the possibility of myasthenia gravis. The same principles of interpretation apply for other postsynaptic neuromuscular junction disorders, with some notable exceptions described below. The demonstration of significant decrement in two or more muscles is preferred, but not required, for the diagnosis of myasthenia gravis in the right clinical setting. (See "Diagnosis of myasthenia gravis".)

The cutoff value for an abnormal RNS decrement is a balance between specificity and sensitivity. A reproducible decrement of >10 percent has been the most quoted determinant for an abnormal test in most muscles [27-30]. A notable exception to a proposed 10 percent cutoff for an abnormal decrement is in the tibialis anterior muscle (peroneal nerve), where a decrement of >20 percent has been recommended as the abnormal cutoff value in one study [31,32].

The sensitivity of RNS in the diagnosis of myasthenia gravis varies widely among the published studies [27,32-34]. The sensitivity is increased by studying multiple nerves, studying nerves in the distribution of weakened and more severely affected muscles, studying nerves to proximal muscles [21,35], and testing after high-frequency (tetanic) stimulation or after maximal isometric exercise.

RNS to the ulnar nerve recording at the abductor digiti minimi is the easiest to perform, but in patients with pure ocular myasthenia or mild generalized disease, the sensitivity is low [27]. By contrast, the yield in the ADM is much higher in patients with severe generalized weakness, especially when intrinsic hand muscle weakness is present [21,27,28]. Although the technical demands are high in the intensive care unit, RNS has high sensitivity in patients with myasthenic crisis, especially when recording in the orbicularis oculi and ADM muscles [36,37].

Subtype and duration of myasthenia may also influence the sensitivity of RNS. In one study in patients with generalized myasthenia gravis, facial RNS was more likely to be abnormal in the subgroup of patients who were seropositive for muscle-specific tyrosine kinase (MuSK) antibodies than for patients who were seropositive for acetylcholine receptor antibodies or seronegative for both MuSK and acetylcholine receptor antibodies [38]. Another study found that RNS was normal in all 17 patients with low-density lipoprotein-related protein 4 (LRP4)-positive myasthenia gravis [39]. In a study that examined the effect of symptom duration, the nine patients with acute-onset generalized myasthenia gravis (defined as a symptom duration of less than four weeks) were less likely to have abnormal RNS studies than those patients with a subacute or chronic presentation (11 versus 84 percent) [40].

Myasthenia gravis is not the only etiology associated with a decremental response, especially if the decrement increases through the nine or more stimuli in a train. The differential diagnosis is not usually clinically difficult, but conventional nerve conduction studies and electromyography are usually necessary. Conditions that have been noted to induce decremental responses include the following:

●Lambert-Eaton myasthenic syndrome (LEMS)

●Radiculopathy

●Motor neuron disease

●Peripheral neuropathy

●Polymyositis

●Some myopathies (McArdle disease, paramyotonia congenita, hyperkalemic periodic paralysis, etc)

●Medication effects (especially d-penicillamine, aminoglycosides, and nondepolarizing neuromuscular blocking agents)

●Botulism

●Organophosphate poisoning

In the past, several methods were used in an attempt to increase the sensitivity of the procedure, including muscle warming, limb ischemia, and curare. However, exercise is the only provocative procedure commonly employed in modern clinical practice. A study that explored the influence of temperature made a compelling case for the use of limb warming combined with exercise in some situations [41].

For a patient with suspected myasthenia gravis who has normal or equivocal RNS studies, single-fiber electromyography (SFEMG) is suggested if the diagnosis is uncertain. (See 'Single-fiber electromyography' below.)

Rapid RNS — In disorders of the presynaptic terminal of the neuromuscular junction, where the number of released vesicles is markedly reduced, high-frequency (tetanic) stimulation can markedly increase the release of acetylcholine and thus the size of the CMAP. The clinical correlate is a corresponding increase in muscle power. Therefore, in disorders of impaired presynaptic acetylcholine release, the defining neurophysiologic pattern is a reduced CMAP amplitude in rested muscle that significantly increases with tetanic stimulation. The reduced baseline CMAP amplitude typically involves the great majority of motor nerves, even recording over strong muscles, and the absolute value is usually well below the normal range. The most frequently encountered clinical disorder of presynaptic function is LEMS.

As discussed previously, tetanic stimulation is most easily accomplished by voluntary maximal isometric muscle contractions, although RNS at 10 to 50 Hz can be employed in obtunded patients. The typical clinical protocol for suspected LEMS is 15 seconds of maximal isometric contraction of the target muscle, preceded and followed by a supramaximal CMAP (waveform 8). The alternative procedure is high-frequency electrical repetitive stimulation (20 to 50 Hz) for 1 to 10 seconds (waveform 2), which is consistently reported to be uncomfortable. (See "Lambert-Eaton myasthenic syndrome: Clinical features and diagnosis", section on 'Repetitive nerve stimulation and exercise testing'.)

Rapid RNS protocol — For suspected LEMS, the abnormalities are usually evident in all muscles, so the abductor digiti minimi (ulnar nerve) is chosen for study. The supramaximal CMAP baseline is recorded, followed by a slow RNS train of five stimuli at 2 to 3 Hz. Next, 10 seconds of maximal isometric exercise is performed, if the patient is cooperative, and the CMAP repeated. For those patients unable to activate the muscle voluntarily, a high-frequency RNS is delivered at 20 Hz for two to three seconds, after preparing the awake patient for the significant expected discomfort. If unclear, a second trial at another readily accessible muscle is recommended, with a preference for a weak muscle if possible. The procedure for suspected botulism is similar. However, the abnormalities are often more subtle with botulism than with LEMS, and the examination of a weak muscle is more important.

SINGLE-FIBER ELECTROMYOGRAPHY — Conventional monopolar or concentric needle electrodes record potentials that represent the summated compilation of individual muscle fibers firing in near synchrony. Single-fiber electromyography (SFEMG) is most commonly performed with a specially configured electrode (figure 3) that allows the recording of a small number of individual muscle fiber potentials within the same motor unit [42]. Although this technique has several applications, the primary purpose is the electrophysiologic evaluation of the neuromuscular junction and the determination of "jitter," which represents the variability of the interval between stimulation and depolarization of a single muscle fiber with stimulated SFEMG (figure 4), or between two single muscle fibers from the same motor unit with conventional SFEMG (figure 5). As described below, SFEMG is the most sensitive electrodiagnostic procedure for the determination of myasthenia gravis. (See 'Utility of SFEMG' below.)

Jitter — When single muscle fibers from the same motor unit are activated or stimulated repetitively, there is a very small variability in the latency of each response. This is due primarily to changes in the amplitude and slope of the endplate potential (EPP) on the postsynaptic membrane; the contributions from the peripheral nerve or muscle fiber components of the motor unit are negligible. This variability in the latency is termed "jitter." To measure jitter in the clinical setting, the recording electrode is positioned to record two or more muscle fibers simultaneously using the techniques described below. In the absence of a disorder of the peripheral motor nerve, any variability in the interpotential latency difference, referred to as the "interpotential interval" (IPI), would be due to changes in the time of neuromuscular junction transmission.

The conventional method of expressing jitter is as the mean value of consecutive differences (MCD) rather than the mean IPI. This calculation compensates for small changes in the electrode position during the study.

When the neuromuscular junction safety factor is reduced, there will be times when the EPP may be inadequate to depolarize the muscle fiber, and its failure to contract is called blocking. The clinical correlation of a significant degree of blocking is muscle weakness. Muscles with increased jitter but little or no blocking usually have normal strength.

Needle electrode — A special concentric needle electrode is usually used for SFEMG recordings. A small recording surface is present 3 mm from the tip of the electrode on the side of the shaft (figure 3) [42]. This allows for a selective extracellular recording field adjacent to a small number of muscle fibers. There is an exponential drop of the recorded motor fiber amplitude with increasing distance from the recording site. This is a critical characteristic since it minimizes the contamination from adjacent muscle fibers. The typical active recording surface of the SFEMG electrode is 25 micron in diameter with a functional recording field of approximately 300 micron from the uptake area and a 200 micron interelectrode distance. In comparison, a conventional concentric needle electrode has an exposed surface of 150 by 600 micron with a recording radius of approximately 1 mm.

The smallest standard concentric needle electrode has an exposed surface of 80 by 300 micron. Studies using these smaller needles require different filter settings, different reference jitter values, and experienced practitioners to be reliable [43,44].

Recording procedures — The ability to isolate a stable pair of single-fiber potentials from the same motor unit requires practice and patience. Use of a triggering device and delay are necessary and are included in the automated programs of the modern digital electromyography machines.

Recording the single-fiber potentials requires that the muscle be activated in a controlled manner. Traditionally, this has been accomplished with a sustained weak voluntary contraction. There are obvious restraints to the use of this procedure in patients who have difficulty performing the necessary contraction or who are unable to cooperate. Stimulated SFEMG has gained acceptance as an accurate alternative that is usually quicker and easier for the patient, but offers some novel technical considerations.

●With voluntary contraction SFEMG, the patient weakly activates the relevant muscle, most often the extensor digitorum communis, and the recording electrode is inserted. The needle is slowly rotated or advanced to maximize the muscle fiber potentials with the trigger set on the initial positivity of the action potential. The key is making small movements of the recording electrode and rotating the shaft until two or more suitable time-locked potentials are present. A single-fiber potential will have an identical repeating waveform since it depolarizes in an all-or-none manner while a waveform with a varying configuration is likely a summated potential of two or more single-fiber potentials and is excluded.

●Stimulated SFEMG is performed by stimulating distal nerve twigs, often intramuscular, with a needle electrode and recording the individual muscle fiber depolarizations with the same recording techniques and recording electrode as described above when utilizing voluntary activation (figure 4).

Stimulated SFEMG is technically more difficult than the conventional voluntary procedure. As only one muscle fiber is studied at a time, the stimulus-to-response interval is measured rather than an IPI. Multiple fibers can be recorded at a time, but each involves a separate analysis. Any waveform peak meeting the appropriate criteria can be used, often allowing for the evaluation of numerous fibers in a single tracing. The normal mean MCD in stimulated SFEMG is approximately two-thirds the value for voluntary activation, since the jitter is occurring in only a single fiber instead of both fibers of a pair. The top normal mean MCD for the extensor digitorum communis (stimulated) is 25 microsecond.

Stimulated SFEMG is useful in many clinical situations, and is particularly appropriate for the study of facial muscles since steady prolonged voluntary activation of these muscles is difficult. Other applications vary with the clinical setting.

In evaluating the results of a SFEMG study, it must be remembered that jitter is altered by many factors, including:

●Temperature

●Firing rate

●Neuromuscular blocking agents

●Acetylcholinesterase inhibitors

●Ischemia

●Reinnervating illnesses

●Neuromuscular junction disorders

Most patients on cholinesterase inhibitors still have abnormal tests, but the sensitivity is reduced, and we recommend the medications be stopped 12 to 24 hours prior to the study [45].

Single-fiber potential — The single-fiber potential is dominated by high-frequency components. This allows the low-frequency filter to be set high, minimizing the volume-conducted components of more distant motor units. The biphasic spike has a rise time range of approximately 75 to 200 microsecond and duration of approximately 1 microsecond. The amplitude varies from 200 microV to 10 mV but it is usually in the 1 to 5 mV range.

Criteria for abnormality — Jitter varies considerably between different muscles. Standard values have been established for many muscles for voluntary muscle-activated SFEMG [46-48]. These values differ from data obtained from testing with concentric needle electrodes [44]. We generally use the SFEMG reference values published in 1994 by the Ad Hoc Committee of the AAEM Single Fiber Special Interest Group [47].

The best studied muscle for SFEMG is the extensor digitorum communis, which is well suited to the difficult process of minimal voluntary activation. Three pieces of data are usually considered in the interpretation of a study:

●The jitter, expressed as the MCD (see 'Jitter' above)

●The percent of pairs whose jitter exceeds the upper limit of normal for the muscle under study

●The percent of blocked fibers

Since blocking represents the most severe form of neuromuscular junction dysfunction, it is rarely seen without a clear increase in jitter. For the extensor digitorum communis muscle, jitter (mean MCD) >35 microsecond is generally considered abnormal, as is >10 percent of pairs with MCD values >55 microsecond or >10 percent of pairs with any fiber blocking. The same values for stimulated SFEMG of the extensor digitorum communis are mean MCD >25 microsecond or >10 percent of pairs with MCD values >40 microsecond [49]. Values for the orbicularis oculi muscle are 20 and 30 microsecond, respectively [50].

Jitter values remain fairly stable until roughly 60 years of age when the normal values in some muscles need to be adjusted [46-48].

Utility of SFEMG — SFEMG is the most sensitive electrodiagnostic procedure for the determination of myasthenia gravis. It is more sensitive than RNS because neuromuscular junction abnormalities that do not induce muscle fiber blocking (decrement) can still demonstrate important increases in jitter. Consequently, there are often significant jitter abnormalities in muscles without weakness.

●In a pattern similar to RNS, the sensitivity of SFEMG is directly related to the severity of myasthenia. As an example, one report found that SFEMG of the extensor digitorum communis muscle was abnormal in 60 percent of patients with ocular myasthenia gravis and 89 percent with generalized myasthenia gravis [51]. These values increased to 97 and 99 percent, respectively, if SFEMG of any of three tested muscles were used as the criteria (extensor digitorum communis, frontalis or orbicularis oculi) [51]. Other studies have reported similar values [52-54]. The yield has been consistently highest in weak muscles and/or facial muscles [55]. Even in pure ocular myasthenia gravis, SFEMG in the extensor digitorum communis (an unaffected muscle) is abnormal at the time of presentation in the majority of patients [56,57].

●Some have suggested the subgroup of patients with MuSK antibody-positive myasthenia gravis may be less likely to have abnormal SFEMG in the extensor digitorum communis muscle than patients with acetylcholine receptor antibody positive myasthenia gravis or other patients with seronegative myasthenia gravis [58-60], but this was not confirmed by others [38,61]. All three groups have had approximately the same frequency of abnormalities with SFEMG in the orbicularis oculi muscle [38,55,58].

●Voluntary SFEMG and stimulated SFEMG have similar sensitivities for the diagnosis of MG.

●In two reports, SFEMG of the extensor digitorum communis in patients with ocular myasthenia gravis did not predict progression to generalized myasthenia gravis [57,62], although in the one prospective study, those patients with a normal SFEMG were less likely to subsequently develop generalized myasthenia [57].

●Although there is a tendency for increased jitter with increasing stimulation rates in some patients with myasthenia gravis, this is not a consistent or predictable finding and it is not useful as a diagnostic test [51,56,63].

Like myasthenia gravis, Lambert-Eaton myasthenic syndrome (LEMS) has a reduced safety factor for neuromuscular junction transmission. Jitter is therefore increased in this condition, often with concomitant blocking. In addition, both jitter and blocking are characteristically improved at higher stimulation rates in patients with LEMS [63-66]. However, the converse is seen often enough that the absence of an inverse rate effect does not exclude the diagnosis [63,67]. Furthermore, an inverse rate effect on jitter and blocking with higher stimulation rates can also be seen in myasthenia gravis, and thus its presence does not exclude myasthenia gravis with or without LEMS [65].

SFEMG has also proven to be useful in suspected cases of botulism where conventional nerve conduction and RNS studies are normal or only mildly abnormal. In one outbreak of food-borne botulism, SFEMG correctly diagnosed all seven cases while rapid RNS was not diagnostic in any [68,69].

SUMMARY

●Neuromuscular junction structure – The nicotinic neuromuscular junction is a complex, specialized structure incorporating the distal axon terminal and the muscle membrane that allows for the unidirectional chemical communication between peripheral nerve and muscle. It consists of the presynaptic nerve terminal, the synaptic cleft, and the postsynaptic endplate region on the muscle fiber. Acetylcholine serves as the neurotransmitter for voluntary striated muscle (figure 1). (See 'Anatomy and physiology of the neuromuscular junction' above.)

●Initial neurodiagnostic testing – Routine electrodiagnostic studies consisting of nerve conduction studies (NCS) and electromyography (EMG) are a standard component of the evaluation of neuromuscular junction disorders. However, additional techniques may be required to assess the fidelity of neuromuscular junction. (See 'Initial routine electrodiagnostic studies' above.)

●Repetitive nerve stimulation – Repetitive nerve stimulation (RNS) represents a modification of conventional motor nerve conduction studies. The interval between repeated motor nerve stimulations is designed to maximally stress the neuromuscular junction safety factor. (See 'Repetitive nerve stimulation' above.)

•Slow RNS – Slow rates of RNS (1 to 5 Hz) deplete the number of quanta of acetylcholine available for release with subsequent depolarizations. In pathologic states where the safety factor for neuromuscular junction transmission is considerably reduced, some muscle fibers will fail to depolarize in the later stimuli of a train and the compound muscle action potential (CMAP) amplitude will drop, which is referred to as a decremental response (waveform 1). A reproducible decrement of >10 percent is considered abnormal in most muscles. (See 'Repetitive nerve stimulation' above and 'Slow RNS' above.)

•Rapid RNS – In disorders of the presynaptic terminal of the neuromuscular junction, where the number of released vesicles is markedly reduced, voluntary maximal isometric muscle contraction or high-frequency RNS can markedly increase the release of acetylcholine and thus enlarge the size of the resultant CMAP amplitude and area, which is referred to as an incremental response (waveform 2). (See 'Repetitive nerve stimulation' above and 'Rapid RNS' above.)

●Single-fiber electromyography – Single-fiber electromyography (SFEMG) is most commonly performed with a specially configured electrode (figure 3) that allows the recording of a small number of individual muscle fiber potentials. The primary purpose is the evaluation of the neuromuscular junction by the determination of "jitter," which represents the variability of the interval between the depolarization in two single muscle fibers from the same motor unit (or in the case of stimulated SFEMG, between the stimulation and depolarization of a single muscle fiber) (figure 5). SFEMG can also be performed with a small concentric needle electrode with added technical challenges and different reference data. (See 'Single-fiber electromyography' above and 'Jitter' above.)