Multifocal Motor Neuropathy: A Narrative Review

Authors: Benjamin Claytor, David Polston, Yuebing Li

Abstract

Multifocal motor neuropathy (MMN) is an acquired autoimmune polyneuropathy that affects almost exclusively the motor nerve fibers. Typically seen in middle‐aged adults, its predominant clinical feature is a chronically progressive asymmetric weakness that affects the distal upper extremities most significantly. Minor sensory symptoms, sensory examination findings or abnormal sensory nerve conduction studies can be seen in the lower extremities in a minority of patients. Electrodiagnostic studies reveal motor conduction blocks at noncompressible sites, and minor findings of other demyelinating features such as conduction slowing or temporal dispersion. Anti‐GM1 antibody titers are elevated in less than half of MMN patients, and more recent studies suggest mechanisms including antibody‐induced complement attack at the node of Ranvier with resulting ion channel dysfunction. Peripheral nerve magnetic resonance imaging and neuromuscular ultrasound often reveal non‐uniform enlargement of the nerve roots, plexuses, or peripheral nerve segments, thus being useful in assisting diagnosis. The differential diagnosis of MMN mainly includes motor neuron disease or demyelinating sensorimotor polyneuropathies. Immunoglobulin therapy is the first‐line and mainstay of treatment, being effective in maintaining or restoring muscle strength in the majority of patients. However, motor strength often slowly declines over the long term, even with maintenance immunoglobulin treatment. More effective immunotherapy is needed to halt the slow progression of MMN, and complement inhibition appears to be a promising option in the near future.

Introduction

Multifocal motor neuropathy (MMN), first described by several groups independently in the 1980s, is an autoimmune motor neuropathy with a low prevalence of < 1 in 100,000 [1, 2, 3, 4, 5]. It is characterized by insidiously or stepwise progressive weakness that usually occurs over the course of months to years [5, 6, 7, 8, 9]. Most affected patients are male, with a male‐to‐female ratio of 1. The average symptomatic onset age is in the fourth or fifth decade of life; however, onset age may vary from the teenage years to the seventh decade [6, 10, 11, 12, 13]. In this review, we aim to provide an update on its clinical, electrodiagnostic (EDX), laboratory and radiological features, pathogenesis, and treatment options.

Clinical Presentation

Typical Clinical Features

Asymmetric chronic or stepwise progressive weakness is the characteristic feature seen in 80%–100% of patients and leads to accumulation of disability over time. Weakness starts most commonly in the distal upper limb (60%–80%) while a smaller portion of patients have distal lower limb onset (15%–30%) [5, 6, 8, 13, 14]. Almost all patients will eventually have distal upper limb involvement which often becomes the most severely affected region [15]. Within the distal upper limb, finger extensors are affected more commonly than flexors [5, 11]. Weakness usually follows the distribution of specific nerves and muscles within the same nerve distribution can be affected to different degrees, which can lead to the clinical finding of differential finger extension weakness [5, 11].

Worsening weakness with cold exposure termed cold paresis is reported in up to 83% of MMN patients [16, 17]. While this symptom is not specific, it is five to six times more frequent in MMN patients when compared with patients with chronic inflammatory demyelinating polyneuropathy (CIDP) or progressive muscular atrophy (PMA) [17].

Muscle atrophy is eventually seen in nearly all patients, but a potential clue to an MMN diagnosis is weakness out of proportion to degree of atrophy especially in earlier stages [5, 8, 18]. The degree of atrophy often correlates with disease duration [6]. Other lower motor neuron findings including fasciculations and cramps are common, occurring in 20%–60% of patients, and reflexes are typically reduced or normal [5, 6, 8, 11].

Uncommon Clinical Features

Rarely patients with MMN can present with weakness starting in the proximal arm in 2%–5% of cases. Hyperreflexia in the affected segment is also reported in a small percentage of patients (8%) [5, 8]. Other unusual motor features including muscle hypertrophy and myokymia have been described in case reports [1, 6, 19]. The muscle hypertrophy is thought to be due to continuous muscle activation, frequent cramps, or fasciculations in affected muscles. Cranial nerve involvement, typically lower cranial nerves X–XII, has been reported as has ophthalmoplegia although these features are exceptionally rare [13, 20, 21].

Subjective sensory symptoms are reported by a minority of patients but up to 22% of patients can have subtle sensory abnormalities on exam, typically vibratory sensation reduction in the distal lower extremities [5, 8, 22, 23]. Patients with longer disease courses are more likely to have objective sensory findings in the lower limbs [5].

As noted previously MMN is typically a progressive disorder but isolated cases of spontaneous improvement and stabile deficits without immunotherapy have been reported [5, 14, 24]. More generalized and acute onset has been described [25, 26]. Mortality directly related to MMN is a very infrequent occurrence [27].

Associated Conditions

Several autoimmune disorders are seen at higher frequencies in MMN patients compared with controls, including ankylosing spondylitis, Crohn's disease, and celiac disease [28]. This association may be mediated by the HLA‐DRB1‐15 allele, which is seen more commonly in MMN [28, 29]. The presence of the HLA‐DRB1‐15 allele in MMN patients, however, does not seem to be associated with GM1 antibody titers, response to intravenous immunoglobulin (IVIG) or clinical disease severity [29].

EDX Findings and Criteria

Definition of Conduction Block

A key EDX feature for the diagnosis of MMN is conduction block. The most widely utilized criteria for the definition of conduction block in MMN was set forth by the European Federation of Neurological Societies/Peripheral Nerve Society (EFNS/PNS) which is based on the degree of compound muscle action potential (CMAP) area loss with proximal stimulations in comparison to distal ones [30]. The criteria also stipulate the degree of allowable temporal dispersion, as phase cancellation can lead to a reduction in the CMAP area with proximal stimulations, which could be mistaken for condition block. Computer models have demonstrated that phase cancellation could account for up to a 50% drop in CMAP area with proximal stimulations and this helped refine cutoffs for definite conduction block [31]. Accordingly, the EFNS/PNS criteria define conduction block as Definite conduction block as a ≥ 50% reduction in the CMAP area across any nerve segment (median, ulnar, or peroneal) with ≤ 30% increase in CMAP duration, and a distal CMAP amplitude of ≥ 20% the lower limit of normal or 1 mV.Probable conduction block under the following Reduction in CMAP area by 30%–49% across a long nerve segment in the upper extremity with ≤ 30% increase in CMAP duration ORReduction in CMAP area by ≥ 50% (same as definite) with > 30% increase in CMAP duration

The American Association of Neuromuscular and Electrodiagnostic Medicine (AANEM) has also published guidelines on EDX criteria for definite and probable conduction block which differ somewhat from the EFNS/PNS criteria [32]. AANEM criteria allow for characterization of conduction block using various cutoffs of 30%–60% based on a multitude of modifying factors including adoption of CMAP area versus amplitude reduction, presence or absence of temporal dispersion, and nerve or nerve segment of interest. Usage of many different cutoff values for the definition conduction block makes these criteria more time consuming to utilize in routine clinical practice [32].

It should be noted that the criteria are consensus guidelines based on expert opinion and have not been extensively validated. While EDX criteria are useful at assessing the level of confidence in the diagnosis, MMN cannot be solely diagnosed by EDX findings alone. Some have argued that above criteria for conduction block are too stringent and have demonstrated that the AANEM criteria for conduction block do not discriminate non‐responders from responders to IVIG [33, 34].

Frequency and Distribution of Conduction Block in MMN

In large case series of MMN, definite conduction block was seen in 56%–81% of patients [5, 13, 33, 35]. The presence of probable conduction block was more variable but seen in up to 94% of patients [5, 9, 13, 33, 35]. Inching studies demonstrate that most blocks involve short nerve segments typically < 50 mm in length [36]. Conduction blocks, not surprisingly, are more common in the upper extremity and typically seen within the median and ulnar nerve distributions [5, 35, 37]. Longer disease durations without treatment are associated with increased numbers of conduction blocks [38]. Conduction block can improve with treatment, but in some patients, conduction block worsens or appears in new nerve segments despite effective treatment [14, 15, 39, 40, 41, 42]. Thus, conduction block cannot be used as an outcome measure or to monitor disease activity.

The varying detection rate of definite or probable conduction block is at least in part due to utilization of different criteria. In a study comparing the EFNS/PNS and AANEM criteria, which included 84 patients, 61% of subjects met EFNS/PNS criteria for definite or probable MMN whereas 77% of subjects met the AANEM criteria for definite or probable MMN. The higher sensitivity of the AANEM criteria, however, was likely due to missing data on CMAP area in 14 patients which precludes a true comparison between criteria [43].

Patients with an otherwise typical clinical phenotype for MMN may not have any evidence of conduction block on EDX testing [11, 18, 44, 45, 46, 47]. Conduction block that is very proximal or distal can be difficult to demonstrate on routine nerve conduction studies. Over time motor axon loss develops in MMN patients, which if severe, will also make detection of conduction block difficult. Comparing MMN patients with and without conduction block does not seem to reveal a difference in demographics, GM1 antibody titers, or clinical severity, although those without conduction block may have less proximal weakness [11, 44]. Patients who have a pure lower motor neuron syndrome without conduction block and with a positive response to IVIG are more likely to have distal upper extremity weakness at presentation, normal creatine kinase (CK) values, normal needle electromyographic examination of thoracic paraspinal muscles, and lack of EDX abnormalities outside of regions of clinical weakness [28, 46, 48, 49].

Other Methods of Evaluating Conduction Block

Given the vexing problem of differentiating MMN patients without conduction block from other pure lower motor neuron syndromes on conventional EDX studies, some have utilized other EDX techniques to assess for proximal conduction block. These include spinal root stimulation with needle electrode, transcutaneous root stimulation, cervical magnetic root stimulation, and transcranial magnetic stimulation using a triple stimulation technique [50, 51, 52, 53]. These techniques, however, can be painful, do not have widespread availability, and are fraught with technical difficulties including inability to ensure supramaximal stimulation of the motor nerve roots. At present these other EDX techniques have limited clinical utility.

In MMN, conduction block may be activity‐dependent and its detection can be enhanced by brief contractions [45, 53]. Utilizing magnetic root stimulation, brief isometric contractions in MMN patients led to transient conduction block, mild temporal dispersion, and transient waveform changes when compared with patients with amyotrophic lateral sclerosis (ALS) and healthy controls [53]. Other studies utilizing conventional nerve conduction studies have questioned the effect of activity on conduction block [54]. There is no consensus on how to evaluate for the presence of activity‐dependent conduction block, so this too is not yet advisable for routine clinical practice.

Other EDX Features

In addition to conduction block, other demyelinating features can be found in patients with MMN but they are much less prominent when compared with demyelinating sensorimotor polyneuropathies like CIDP [55]. Conduction velocity can be mildly reduced in regions of conduction block, typically in the range of 27–33 m per second, but is usually normal outside these regions [6, 8, 10, 56]. F waves can be absent or demonstrate prolonged latencies, but this is usually restricted to nerves harboring conduction block [6, 8, 10]. Mildly and prominently prolonged distal motor latencies can be seen in 30%–40% and < 5%–10% of patients, respectively [5, 8, 35, 40]. Reported frequencies of temporal dispersion are variable, likely due to the utilization of different diagnostic criteria [8, 10, 35].

Distal CMAP amplitudes are often reduced in MMN which could be due to axon loss, and reductions in CMAP amplitude correlate with disease duration [38]. Several studies have demonstrated that low amplitude or unobtainable CMAP reversed rapidly with immunotherapy, suggesting a reversible distal conduction block [15, 42, 44].

Abnormalities on needle electromyography are ubiquitous and clinically weak muscles often show the presence of both active denervation and chronic reinnervation. In weak muscles with motor unit action potentials of normal configuration, reduced recruitment is a potential clue to the presence of a hidden conduction block without axon loss along the innervating nerve [8, 35]. Axon loss and conduction block are associated with one another, but axon loss is a stronger determinant of weakness than is conduction block [35].

Sensory Nerve Conduction Findings

While minor subjective sensory symptoms and mild objective distal lower extremity vibratory sensation loss are allowable for the diagnosis of MMN, objective sensory nerve action potential (SNAP) abnormalities are characteristically absent including across nerve segments showing motor conduction block [5, 9, 22, 30, 40, 56]. In longitudinal studies, some MMN patients can develop reductions in SNAP amplitudes [23, 57, 58]. Patients with SNAP amplitude reductions have greater number of nerves affected and lower distal CMAP amplitudes, but the reduction in SNAP amplitudes is independent of disease duration and presence of GM1 antibodies [57, 58].

It is unclear if those who develop minor objective sensory abnormalities over time truly have MMN or potentially a separate disease. In one study, the diagnosis of MMN was changed to multifocal CIDP, also known as multifocal acquired demyelinating sensory and motor neuropathy (MADSAM) in 42% of patients who had a positive ganglioside antibody and 7% who were seronegative. Among patients who were re‐classified, the median time for development of sensory symptoms was 30 months [59]. It is therefore important to be vigilant for the development of any objective sensory abnormalities in MMN patients as it could signify an alternative diagnosis.

Laboratory Evaluation

GM1 Antibodies

Antibodies to the GM1 ganglioside are found in roughly 30%–40% of MMN patients in modern cohorts, and typically are of the IgM subclass [5, 8, 35, 60]. The presence of GM1 antibodies at high titers may predict a more severe disease, but this observation remains controversial [5, 13]. Data are also conflicting as to whether GM1 antibody positivity or high GM1 antibody titer is a positive indicator of favorable treatment response [5, 6, 14, 36, 59, 61].

Positive GM1 antibodies at low titers can be seen in patients with ALS, other inflammatory and non‐inflammatory neuropathies such as multifocal CIDP, and in healthy subjects [59, 62, 63, 64]. When utilizing higher titer cutoff values of 4000, GM1 antibodies have a good specificity of 93%, albeit with limited sensitivity of 50% [63]. While a high titer‐positive GM1 antibody may be helpful diagnostically, it does not perfectly discriminate MMN from other demyelinating neuropathies or motor neuron disease. Low titers of GM1 are of limited diagnostic significance [63, 64].

To improve detection rates of GM1 antibodies, an ELISA assay containing an admixture of GM1 and glucocerebroside (GM1/Galc) was utilized [60]. Utilizing this method, 75% of MMN patients were positive including all GM1‐positive patients, and most MMN patients with GM1 antibodies had higher titers to GM1/Galc. GM1/Galc binding antibodies, however, are less specific for an MMN diagnosis. The authors of the study posit that the combination of GM1/Galc can enhance anti‐GM1 antibody binding or lead to recognition of a new antigen [60].

Other Associated Antibodies

Antibodies to other gangliosides including GD1a, GD1b, and GM2 can be found in MMN patients [5, 9, 59, 61, 65]. The presence of these ganglioside antibodies is of likely limited clinical significance, as GM2 antibodies, for instance, are found in similar numbers of control and MMN patients [60]. Antibodies to nodal and paranodal structures including gliomedin and neurofascin‐186 as well as antibodies targeting a disulfate glucosamine‐uronic acid heparin disaccharide moiety NS6S have been putatively associated with MMN, however, follow up studies have failed to confirm these findings [60, 66, 67, 68, 69]. More recently antibodies to myelin oligodendrocyte glycoprotein have also been described in patients with peripheral nervous systems findings compatible with MMN. These patients, however, harbor features of central nervous system demyelination and optic neuritis [70, 71].

Other Laboratory Abnormalities

An IgM monoclonal gammopathy is found in 7% of MMN patients compared with 2% of controls [72]. Two thirds of MMN patients harboring a monoclonal gammopathy are positive for GM1; however, the majority GM1‐positive patients do not have a monoclonal gammopathy [72]. Modest CK elevations of 300–600 IU/L can be seen in up to two thirds of patients [33, 40].

Cerebrospinal Fluid Evaluation

Cerebrospinal fluid (CSF) findings are often normal, but mild protein elevations of 50–80 mg/dL can be detected in up to 20%–40% of patients [6, 8, 9, 40, 56]. Protein elevation in the CSF is typically lacking in MMN patients without conduction block [9, 10].

Radiological Findings

MRI

Between 30% and 50% of MMN patients have been reported to have either T2 hyperintensity or nerve enlargement on magnetic resonance imaging (MRI) of the brachial plexus, and abnormalities can be seen in those with both long and short disease durations (Figure 1) [73, 74, 75, 76]. Contrast enhancement, by comparison, seems to be relatively inconspicuous in MMN [74]. Diffusion tensor imaging of the brachial plexus can also distinguish MMN from patients with other lower motor neuron syndromes [76]. MRI imaging in the forearm can also demonstrate findings of nerve enlargement as well as reduced axonal diffusion [77].

Findings on MRI appear to correlate with the clinical distribution of symptoms. Patients with MMN and multifocal CIDP almost always have asymmetric findings when bilateral cervical plexuses are imaged, as compared to < 10% of patients with CIDP [73, 74]. The degree of nerve hypertrophy in MMN is typically less than what can be appreciated in CIDP [76]. There appears to be no difference in clinical presentation or response to therapy in MMN patients with and without MRI abnormalities [33, 74].

Most of the imaging abnormalities described in MMN are qualitative, and the reproducibility of these qualitative measures has been challenged [76, 78]. Qualitative MRI abnormalities of the plexus such as signal hyperintensity and nerve enlargement have an overall poor to moderate intra‐ and interobserver agreement and widely variable positive and negative predictive values [76]. These qualitative measures could not distinguish MMN from other inflammatory neuropathies, PMA, and control subjects. In one study up to 23% of control subjects were noted to have signal hyperintensity within the plexus [76]. Given the relative lack of specificity of qualitative MRI abnormalities, the presence of signal hyperintensity or nerve enlargement should not be viewed as unequivocal evidence of MMN. Mild MRI abnormalities should not be relied upon solely to diagnose MMN or to differentiate MMN from other lower motor neuron syndromes. Semiquantitative measures including measures of cervical root size appear to have better specificity and reliability but overall poor sensitivity [76, 78].

Ultrasound

Nerve ultrasound has become an inexpensive noninvasive method to evaluate peripheral nerve morphology. Protocols are now available to evaluate MMN that take a mere 15 min to complete [79]. Enlargement of the intra‐scalene roots and brachial plexus trunks on ultrasound has been demonstrated in 68% of MMN cases but up to 90%–100% of patients may have regional or multifocal nerve enlargements distally [75, 80, 81]. Abnormalities are more common in the upper limb, and the median nerve forearm and mid‐arm segments are the most frequently involved (Figure 1) [82]. Enlargement is more common in clinically affected nerves, however, it can be seen in nerves with normal nerve conduction studies [81]. This suggests that a clinically focused ultrasound plus a rapid screening of clinically unaffected nerves likely produces the highest yield. When compared with healthy controls, MMN patients have higher side‐to‐side variability on ultrasound findings, thus side‐to‐side comparison studies may be useful [83].

Ultrasound abnormalities persist in patients who are actively treated and there is a lack of association between ultrasound findings, disease duration, GM1 antibody status, and response to therapy [80, 81, 83, 84]. Over time multi‐focal or regional nerve enlargements can evolve and become more diffuse [81]. Ultrasound findings may not correlate with clinical severity at presentation, however, progressive nerve hypertrophy may correspond to worsening weakness [75, 81, 83]. There is relatively poor correlation between ultrasound findings and sites of conduction block, suggesting that nerve enlargement and conduction block represent different pathological processes [47, 80, 82, 83, 84].

Ultrasound evaluating for nerve enlargement has been shown to reliably distinguish MMN patients from healthy controls with good reproducibility [83]. However, there is no general agreement on the minimal number of regions showing enlargement or the criteria for positive ultrasound findings, as different studies utilized different measures [75, 83, 84]. Ultrasound findings are more useful in differentiating demyelinating neuropathies from diseases predominated by axon or motor neuron loss, but less useful in distinguishing MMN from other demyelinating neuropathies [75, 79]. In a prospective study utilizing 10 different landmarks, ultrasound findings differentiated MMN from ALS with a sensitivity of 87.5% and a specificity of 94%, together with very good interclass and interrater correlation coefficients [84]. In another study, however, by evaluating one nerve innervating the most clinically weak muscle using prespecified cutoff values, authors could differentiate ALS from MMN with a sensitivity and a specificity of 100%, respectively [85]. No single ultrasound finding is specific for MMN, and nerve cross‐sectional area is influenced by age, gender, and BMI [86, 87]. Analogous to MRI abnormalities, minor findings on nerve ultrasound cannot be taken as conclusive evidence defining an MMN diagnosis.

MRI Versus Ultrasound

Studies that have compared MRI and ultrasound findings of the brachial plexus in MMN have generally demonstrated a high degree of concordance when abnormalities are present [74, 75]. Ultrasound, however, may be more sensitive for nerve root thickening, and MRI of the plexus can be normal in 9% of MMN cases with an abnormal ultrasound [75]. Others have demonstrated that MRI measures do not add any diagnostic value to nerve conduction studies and ultrasound [78]. Ultrasound measurement of the cross‐sectional areas of the median nerve in the upper arm and the cervical roots also outperformed qualitative MRI measures of the brachial plexus with much greater reproducibility [76]. Given the low cost, higher sensitivity, high reliability, and limited resources needed to obtain accurate imaging, we advocate for ultrasound, if available, as the first‐line imaging modality in MMN if additional supportive criteria are needed for diagnosis. Greatest attention should be paid to nerves innervating clinically weak muscles, but screening of unaffected nerves is appropriate as is side‐to‐side comparison.

Diagnosis and Differential Diagnoses of MMN

Clinical Diagnostic Criteria

Core clinical criteria for the diagnosis of MMN set forth by the AANEM and EFNS/PNS emphasize the presence of asymmetric progressive or stepwise limb weakness without objective sensory loss beyond minor vibratory loss in the lower extremity [30, 32]. Exclusionary criteria that would preclude an MMN diagnosis include the presence of unequivocal upper motor neuron signs, prominent bulbar involvement, significant sensory deficits, and diffuse symmetric weakness [30, 32]. Table 1 outlines the criteria for the diagnosis of MMN, incorporating contents from both AANEM and EFNS/PNS diagnostic criteria [30, 32].

Differential Diagnoses

Table 2 lists typical clinical findings in MMN, and red flag findings that would suggest an alternative diagnosis. The rarity of MMN and its overlapping features with several different disorders likely account for the high initial misdiagnosis rate, approaching 87% in some case series [8]. Not surprisingly, the most common misdiagnosis is motor neuron disease [8, 33, 88]. Long diagnostic delays are unfortunately common, ranging from 2 to 6 years [5, 11, 13]. Diagnostic delay is decreasing over time, possibly owing to increased recognition of the disorder but it is still unacceptably long [89]. In cases where there is diagnostic uncertainty, we advocate for referral to a specialty neuromuscular center.

The main differential diagnoses for MMN are listed in Table 3. Differentiating MMN from variants of ALS with isolated or predominant lower motor neuron involvement such as the flail arm variant, or PMA requires detailed EDX testing searching for conduction blocks [74, 84]. Serum neurofilament light chain level is more likely elevated in ALS patients with predominant lower motor neuron dysfunction manifestation compared with MMN and could be a helpful adjunct piece of clinical data [90].

Hirayama disease causes progressive and asymmetric distal upper extremity weakness. It typically begins in the teenager years, much earlier than the average age of diagnosis for MMN [91]. Hirayama disease can be differentiated from MMN with the use of a dynamic cervical spine MR (see Table 3). Painless cervical polyradiculopathies could mimic MMN; however, EDX abnormalities in a polyradiculopathy would be expected to follow a root distribution with potential involvement of paraspinal muscles as opposed to nerve distribution abnormalities in MMN [92].

Many demyelinating neuropathies including CIDP, multifocal CIDP, and hereditary neuropathy with liability to pressure palsies can present with significant upper extremity weakness and conduction block. Imaging as well as serologic and CSF testing can be helpful in discriminating these different disorders (see Table 3). Additionally, the specific pattern of EDX abnormalities can also be helpful. In MMN the defining demyelinating feature is conduction block whereas other characteristics of demyelination including conduction slowing, temporal dispersion, and distal latency prolongation are seen less often and are much less pronounced when compared with either multifocal CIDP or typical CIDP [93].

Pathology and Pathophysiology

The origin of conduction block in acquired neuropathies is most often attributed to demyelination [94, 95]. However, conduction block can also arise from other processes that may disrupt saltatory conduction such as metabolic or immune‐mediated disruption of ion channels or axonal transection [96]. The underlying origins of conduction block in MMN have been explored using various methods including pathological assessment of biopsy specimens, serum analysis, nerve hyperexcitability studies as well as animal transfer studies using human sera [97, 98, 99, 100, 101, 102].

Nerve Biopsy

Routine nerve biopsies, most commonly of the sural nerve, have rarely been diagnostically useful in MMN [103]. However, in one series of 11 sural nerve biopsies from MMN patients, similar abnormalities were noted in all cases. These abnormalities were characterized by thinly myelinated large‐caliber fibers and minor onion bulbs detected via electron microscopy, with no evidence of inflammation. These findings suggest MMN is a generalized process with a strong preference for motor fibers [22].

Biopsy of motor nerves at the site of conduction block has been performed in small series and case reports [96, 100, 104, 105]. In the largest series, eight fascicular biopsies were collected from the arms of seven patients at the site of conduction block. In these nerves, the most characteristic findings were fiber degeneration with a significant reduction of large myelinated fibers compared with smaller fibers, and areas of remyelination. Small foci of inflammatory cells were noted in only two nerves. Onion bulbs were not appreciated, and teased fiber analysis did not demonstrate remyelination or demyelination [100].

The above findings contrast with those described in CIDP, which include evidence of demyelination and remyelination on teased nerve fibers, onion bulb formation, and varying degrees of inflammatory infiltrates, but without the predilection for large myelinated fibers involvement found in MMN. Some case reports with single biopsies near sites of block or enlargement in MMN identified small onion bulbs without associated inflammation, similar to findings in sensory nerve biopsies [104, 105]. These pathological differences suggest distinct pathophysiological mechanisms between CIDP and MMN, indicating that MMN is not primarily caused by inflammatory demyelination [96].

Antibodies and Their Nodal Targets

Gangliosides such as GM1 are widely expressed in the myelin sheaths and on axons of both sensory and motor nerves, playing a significant role in maintaining the structure and homeostasis of nerves [106, 107, 108]. However, the density of ganglioside expression varies among different sensory and motor nerves [109]. Some studies have reported a greater abundance of GM1 in motor nerves and noted structural differences between GM1 found in motor and sensory nerves. These findings have led to suggestions that such differences may facilitate preferential binding of GM1 antibodies to motor nerves, contributing to the clinical phenotype in MMN [69, 110, 111, 112, 113, 114, 115]. However, other studies have found the proportion of GM1 in sensory and motor nerves to be similar. These studies suggest that the motor predominant presentation of MMN may result from the influence of other gangliosides or the relative proportion of gangliosides in the motor nerves, which might lead to more avid binding at these sites [109, 116, 117].

At the nodes of Ranvier and the adjacent paranodes GM1 and certain other gangliosides are found in greatest abundance. These highly organized, distinct structures are known targets for immune responses (Figure 2) [109, 118]. Given the mild demyelinating and remyelinating features observed in nerve biopsies, it is possible that initial direct myelin injury exposes the nodal and paranodal regions to immune‐mediated attack [102]. The concentration of gangliosides in these areas has led to the classification of MMN and other acquired neuropathies as nodopathies and paranodopathies [109, 119].

Although GM1 antibodies are associated with MMN, their direct pathogenic role remains uncertain. This uncertainty arises because some studies have successfully reproduced pathology by transferring serum from GM1 antibody‐positive patients into animal models, while others have not [97, 98, 110, 111, 120, 121, 122, 123, 124, 125]. As previously noted, GM1 antibodies are present in only about 40% of MMN patients, and serum transfer studies using serum from GM1 antibody‐negative patients have generated conduction block in some animal models [126]. These findings suggest that other factors contribute to the pathogenesis of MMN. Alternatively, GM1‐negativity might be partially due to limitations in the detection method, as adding galactoctocerebrosidase to the testing array increases sensitivity of detection of GM1 antibodies [60, 127].

Nerve Excitability (Hyperpolarization—Depolarization)

The lack of prominent demyelination on nerve pathology has led to the evaluation of other potential mechanisms including alterations of the axon membrane, which have been investigated through non‐invasive nerve hyperexcitability studies [100, 103, 105, 128, 129]. In a study of six patients with distal conduction block in the forearm, all nerves exhibited findings consistent with hyperpolarization, like that seen in ischemia, distal to the site of conduction block [99, 130]. The authors hypothesized that this distal hyperpolarization may result from focal depolarization at the site of conduction block caused by altered ion accumulation along the membrane. This imbalance could lead to more generalized impairment of the sodium‐potassium pump and ion channel dysfunction [55, 99].

Another study investigated the strength‐duration curves of 18 ulnar nerves of MMN patients, comparing them to those of normal controls and ALS patients. This study revealed that the strength‐duration was abnormally short in MMN, with some reversal of shortening following IVIG treatment. These findings support the concept of hyperpolarization of nerves distal to the site of block and rapidly reversible ion conductance abnormalities playing a role in the pathophysiology of MMN [129]. Conversely, a separate study involving six nerves from five MMN patients found varying results. At the sites of conduction block, hyperpolarization was observed in three nerves, depolarization in two, and mixed features in one. The reasons for these variations were unclear but could be related to the disease severity or chronicity [101].

Despite varying results, these studies collectively suggest that persistently altered conductance due to pump and channel dysfunction contributes to motor impairment, the clinical hallmark of MMN. This is further supported by the high frequency of cold paresis in MMN, indicating functional impairment potentially due to the temperature sensitivity of the sodium‐potassium pump [17, 131]. Ongoing dysfunction of the pump is proposed to cause elevated levels of intracellular sodium, which in turn activates the sodium–calcium exchanger. This leads to abnormal levels of intracellular calcium, potentially causing the axonal injury seen in MMN (Figure 2) [105, 129, 130, 132, 133].

Complement Activation

Complement‐mediated inflammatory damage is a common mechanism in various ganglioside autoantibody‐mediated neurological disorders [134]. Recent studies also support its role in MMN. In a rabbit model of acute motor axonal neuropathy (AMAN), anti‐GM1 antibodies, when administered under specific immunization protocols, lead to complement activation and deposition of membrane attack complex at the nodes of Ranvier in motor nerve fibers. This process disrupts voltage‐gated sodium channel clusters and detaches paranodal myelin terminal loops [135, 136]. In this model, the complement activation property of anti‐GM1 antibodies was found to be crucial for the development of muscle weakness [137].

Complement deposition was observed in motor fibers at the nodes of Ranvier in autopsies of seven AMAN patients. All cases were fulminant, with death occurring 4–12 days after symptomatic onset, so this deposition may not represent the pathologic changes seen in less severe cases [138]. However, a similar mechanism may be involved in the development of MMN, given that GM1 is a common target in both AMAN and MMN. Higher levels of systemic complement activity are associated with MMN disease severity, and a gene promoter region deletion that reduces complement activity is associated with a better prognosis in MMN [139, 140]. Despite this, complement deposition at the nodal region has yet to be shown in MMN patients and, due to varying results among studies, the precise role of GM1 antibody‐mediated complement activation in the development of conduction block remains unclarified [97, 122].

In a human‐induced pluripotent stem cell‐derived motor neuron model, anti‐GM1 IgM antibodies from MMN patients bind directly to motor neurons. This binding alters calcium homeostasis through complement‐dependent and complement‐independent pathways, resulting in marked axonal damage [117]. Another study demonstrated that anti‐GM1 IgM antibodies from MMN patients activate complement exclusively via the classical pathway [140]. However, results are mixed regarding complement activation in MMN patients without anti‐GM1 antibodies [117, 140].

These findings suggest that while the role of complement activation in MMN is not fully clarified, it presents a potentially promising therapeutic target. A study using the previously noted motor neuron model revealed that complement inhibition could rescue motor neurons from structural damage. IVIG, a first‐line treatment of MMN, reduces such complement deposition [140, 141]. Preliminary research exploring the use of complement inhibitors as a treatment option for MMN patients and in vitro MMN models has shown promising results, further supporting the role of complement activation as a part of the pathogenesis in MMN [142, 143, 144].

Treatment

Historically, many therapeutic agents have been used to treat MMN, including prednisone, plasmapheresis, cyclophosphamide, azathioprine, mycophenolate mofetil, methotrexate, cyclosporine, interferon beta 1A, rituximab, IVIG and subcutaneous immunoglobulin (SCIG). However, the majority of these treatments have failed to show notable or consistent efficacy in MMN. At present, IVIG remains the only US Food and Drug Administration‐approved maintenance therapy to improve muscle strength and disability in the adult MMN population [145]. Both the EFNS/PNS Task Force and the AANEM Ad Hoc panel recommended IVIG as the first‐line treatment for MMN [30, 146]. More recently, SCIG has been shown to be an efficacious option for MMN.

Intravenous Immunoglobulin

A positive treatment response of MMN to IVIG was first reported in 1992 [20, 147, 148]. Efficacy of IVIG was previously demonstrated in several randomized trials of mostly small sizes (Table 4) [149, 150, 151, 152, 153]. The largest randomized study included 44 MMN patients and demonstrated increased grip strength and improved measures of disability following IVIG treatment [153]. Keddie et al. [154] performed a meta‐analysis of these trials, and concluded that low‐certainty evidence exists in supporting the efficacy of IVIG for improving muscle strength and disability in MMN patients.

Mechanisms of immunomodulatory action of IVIG could be numerous. IVIG may neutralize or reduce the level of pathogenic antibodies by inhibiting antibody production and/or accelerating antibody catabolism. However, IVIG does not appear to reduce GM1 antibody titers in MMN [4]. IVIG may reduce complement level, inhibit antibody‐mediated complement activation and alleviate alteration of antibody‐induced calcium homeostasis in MMN [117, 141]. Other actions of IVIG may include blockade of Fc receptors on macrophages, modulation of lymphocyte proliferation, inhibition of pro‐inflammatory cytokines, and down‐regulation of cell adhesion [155, 156].

Approximately 5% of MMN patients demonstrate limited weakness without impairing their daily life thus not requiring immunotherapy [5]. Overall 70%–94% of MMN patients respond to IVIG [5, 14, 38]. Following IVIG administration, improvement of muscle strength usually occurs rapidly within the first 7–10 days. Improvement may appear in both distal and proximal muscle groups, and is typically more significant in muscle groups that show greater initial weakness [157].

Several prior studies evaluated possible predictors of response to immunoglobulin therapy in MMN [5, 14, 33, 89, 158]. The initial disease severity may not be a poor prognostic factor as treatment naive patients with severe illness could still show good response to IVIG [38]. However, longer disease duration has been suggested to be a predictor of poor treatment response, likely in relation to the gradually developing axonal degeneration with time, and earlier treatment is of more benefit [5, 9, 38]. Similarly, presence of muscle atrophy, CMAP amplitude reduction, and axonal loss changes on needle electromyography may predict a suboptimal treatment response [33, 38, 158]. Van den Berg‐Vos et al. [33] identified the following predictors of positive treatment responses to IVIG: young onset age, small number of affected limbs, presence of definite conduction block, high distal CMAP amplitudes on nerve conduction studies, abnormal MRI of brachial plexus, and high GM1 antibody titer. Abnormal SNAPs and CSF protein elevation in MMN patients do not correlate with responses to IVIG while CK elevation is more suggestive of IVIG nonresponsiveness [33, 57, 58].

Conduction block has been viewed as a possible predictor of good response to IVIG. Many studies have shown that MMN patients with conduction block responded to IVIG better than those without [14, 15, 149, 151, 152, 159]. The management of MMN patients without conduction block merits a few important considerations, prior to deciding on long‐term maintenance IVIG treatment. First, these patients should have a typical clinical presentation for MMN, and there may be other indicators of demyelination on EDX studies or other supportive clinical findings such as imaging abnormalities on MRI or ultrasound. Second, their response to IVIG may not be as significant or noticeable as that of MMN patients with conduction block, as axonal loss predominantly occurs in nerves without conduction block [15]. Finally, while a trial of IVIG is indicated for such patients, assessment of objective treatment response is preferred. And it is necessary to assess for signs or symptoms of an alternative diagnoses such as motor neuron disease or motor neuropathy of other causes, if typical treatment response is lacking.

The initial dose of IVIG is typically 2 g/kg of body weight given over a period of 2–5 days. Typical clinical improvement following IVIG treatment lasts for 4–6 weeks. In a minority of patients, improvement following one or a few IVIG sessions can persist for years [5, 15, 20, 160]. In the majority of patients, periodic maintenance IVIG infusions, at times indefinitely, are necessary for maintaining strength and activity. An IVIG maintenance regime can be variable in both its dosage and frequency, which may range from 0.4 g/kg every 8 weeks to 1.0 g/kg biweekly. IVIG dosage and dosage interval need to be based on individual response, with a focus on the objective improvement in strength and disability level, as opposed to basing treatment changes solely on patients' subjective perceptions. The benefit/risk ratio of chronic IVIG maintenance therapy warrants periodic revaluation of the need for treatment for each patient in view of its variable efficacy, potential risks, and high cost. There may not be noticeable differences among many different brands of IVIG available [161, 162].

Use of higher maintenance doses of IVIG (such as 5 g/kg/month divided into weekly or biweekly administration) has been described and appeared to be beneficial to improve and maintain strength in a portion of MMN patients [42, 163, 164, 165]. Ultrahigh‐dose IVIG might be more efficacious in preventing axonal degeneration and promoting reinnervation [42]. In MMN patients who fail to show significant treatment response to routine maintenance dosage of IVIG, one or a few sessions of ultrahigh‐dose IVIG can be considered as a rescue therapy to achieve noticeable improvement, before the dose of IVIG is reduced back down. Further studies are necessary to delineate the clinical features of MMN that are predictive of necessity and positive responses to ultrahigh‐dose IVIG.

The benefit of long‐term IVIG treatment has been reported in multiple studies [5, 15, 42, 89, 166]. Despite long‐term IVIG treatment, continued axonal degeneration is observed in the majority of MMN patients. Gradually many MMN patients become progressively less responsive to IVIG, and improvement may become less obvious after months to years of therapy. In the early stage of maintenance treatment, response to IVIG may still be restorable by an increase of IVIG dose and/or frequency. Later in the course, such treatment adjustment may only be partially effective [5, 15, 42, 89, 166].

Adverse reactions associated with IVIG treatment are usually minor and occur in no more than 10% of patients. The most common side effects are headache, chills, nausea, malaise, and myalgia. Rare serious side effects may include acute renal failure, congestive heart failure, aseptic meningitis, and thromboembolic events.

SCIG

SCIG has been used in MMN patients since 2006 [167]. Several studies have utilized SCIG to treat MMN patients, and their main results are summarized in Table 5. In most studies, SCIG was introduced as an IVIG replacement therapy for patients who had stably responded to IVIG for some time [168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178]. The benefits of switching to SCIG include less systemic side effects, more convenience and reduced economic burden, autonomy of administration, and no requirement for venous access. Possible disadvantages of switching to SCIG may include lack of similar efficacy to IVIG in a minority, necessary training of subcutaneous administration by patients or caregivers, and discomfort associated with injection site reactions.

Results from the majority studies show that long‐term SCIG therapy can be an effective alternative approach to IVIG [168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178]. For most patients, switching from IVIG to SCIG maintained strength, with better tolerability and patient satisfaction. Racosta et al. [181]. performed a meta‐analysis of SCIG in MMN and CIDP, and concluded that SCIG is equally effective as IVIG and is associated with a 28% relative risk reduction of moderate and/or systemic adverse effects. However, some patients may develop worsening weakness on SCIG and require a dose increase, periodic IVIG boosting, or a return to the original IVIG treatment within 2 years [175, 179].

SCIG is typically given weekly. Compared with IVIG, SCIG is absorbed and distributed more slowly and steadily, resulting in less fluctuations of serum immunoglobulin level. Trough levels of immunoglobulin tend to be higher with SCIG, which may take several weeks to months to stabilize [182, 183]. The optimal dosing of SCIG remains somewhat unclear. Most studies in MMN used a 1 SCIG to IVIG dose ratio while one study used 1.53:1 (Table 5) [169, 170, 172, 173, 174, 175, 178]. A SCIG to IVIG dose ratio of 0.5 is insufficient to maintain muscle strength [168]. When large volumes are required, SCIG can be administrated multiple times per week or simultaneously at multiple sites once per week. Combination of SCIG with pretreatment using hyaluronidase (often called facilitated SCIG) allows large volume (up to 600 mL) to be given per site, and such products have been successfully administrated in MMN patients. While its efficacy appears to be non‐inferior to traditional SCIG or IVIG, there is a higher frequency of short‐lasting localized side effects that comes with large volume infusions [177, 180, 184].

The most common side effects of SCIG treatment are injection site swelling, erythema, soreness, induration and pruritus, which are usually self‐limiting and improve with repeated administration. Systemic effects such as headache, nausea, and malaise may occur, but mostly are transient. Rare cases of thromboembolic events have been described [178, 180].

Treatment Other Than Immunoglobulin

Limited therapeutic options exist for MMN patients who are non‐immunoglobulin responders. Rituximab and cyclophosphamide have been studied more extensively than others. The efficacy of these treatments, however, has not been demonstrated in randomized trials. At times, these treatments can be used as adjunctive therapy to augment IVIG therapy for poor IVIG responders, to reduce IVIG dosage and/or frequency in IVIG responders or as an alternative treatment for patients who are allergic or could not tolerate immunoglobulin therapy.

Rituximab

Rituximab is a chimeric mouse‐human monoclonal antibody directed against the B lymphocyte marker CD20, and has shown variable results so far in treating MMN. Two studies of MMN patients reported contrasting results. Pestronk et al. [185] described the usage of rituximab in 21 patients with IgM antibody‐associated polyneuropathies. Among them, 14 patients had asymmetric distal predominant motor neuropathies, and 11 patients had motor conduction block on EDX testing. Treatment with rituximab led to a mean increase of 23% in strength measured by hand‐held dynamometry with a corresponding reduction of 43% in serum IgM autoantibody titers. Chaudhry and Cornblath [186] described their open‐label trial of six MMN patients who were already on maintenance IVIG. Addition of rituximab failed to reduce IVIG dose in the next 12 months despite quick CD20 lymphocyte depletion. Similarly, mixed results were also described in several case or case series reports [187, 188, 189, 190, 191, 192]. The mixed response to rituximab indicates that a randomized placebo‐controlled trial to study the benefit of rituximab in MMN is warranted. There is also a need to identify the predictors of a good response to rituximab in MMN.

The most commonly used loading dose of rituximab is four intravenous infusions of rituximab (375 mg/m [2] body surface area) weekly or 1000 mg twice given 2 weeks apart, with subsequent administration based on clinical worsening or at fixed intervals such as 6 months. The usage of rituximab appears safe in MMN patients, and can be considered in patients with severe weakness who do not respond well, are allergic or intolerant to IVIG/SCIG.

Cyclophosphamide

Pestronk et al. [4] first described two MMN patients who failed initial trials of high‐dose prednisone and plasmapheresis, but treatment with intravenous followed by oral cyclophosphamide led to marked improvement in both patients. Krarup et al. [193] treated two MMN patients with cyclophosphamide which failed to lead to improvement in either. Subsequent studies indicated that the use of cyclophosphamide was effective in some MMN patients, resulting in strength improvement, anti‐GM1 antibody reduction, resolution of conduction block, and reduction of IVIG dosage [7, 41, 103, 194]. A combination of monthly intravenous cyclophosphamide and plasma exchange could be suitable for refractory or severe MMN patients. The treatment course consists of 5–7 repeated monthly regimens of plasma exchange on 2 consecutive days, followed by intravenous cyclophosphamide at a dose of 1 g/m^2^ body surface area [195].

The dose of intravenous cyclophosphamide administered was usually 1–3 g/m^2^ body surface area, followed 1 month later with an oral dose of 2 mg/kg/day. There are numerous serious side effects of cyclophosphamide including bone marrow suppression, gonadal damage, hemorrhagic cystitis, and long‐term increased risk of cancer. The risk of these serious side effects and the lack of evidence for cyclophosphamide to reverse established disability restrict its widespread usage in patients with MMN. However, it could serve as an option for patients with severe MMN who are refractory to immunoglobulin therapy and rituximab.

Other Immunotherapies

Two earlier studies reported the use of interferon‐1A to treat MMN patients [196, 197]. Both studies were of small size and performed more than two decades ago, without any recent updates. The limited data on the efficacy of interferon‐beta 1A in MMN and its diminishing usage in treating multiple sclerosis make it a less attractive option for MMN patients.

Immunosuppressive therapies such as azathioprine, mycophenolate mofetil, cyclosporine, and methotrexate have been studied in MMN, mostly as adjunctive therapies in small case series [6, 24, 198, 199, 200, 201, 202]. Piepers et al. [203] conducted a randomized, placebo‐controlled study of mycophenolate mofetil as an add‐on therapy to IVIG treatment in 28 MMN patients. This 1‐year study failed to show benefits in reducing IVIG dose, or improvement in muscle strength.

Corticosteroids are infective when given orally or intravenously at high doses and may worsen patients' symptoms [4, 158, 204, 205]. Plasmapheresis is also ineffective and may be associated with clinical worsening including the appearance of new conduction blocks in previously unaffected nerves [4, 206, 207]. Immunoadsorption appears to be ineffective as well [208].

Complement Inhibitors

Complement inhibition has emerged as a potential newer approach in treating MMN. Several complement inhibitors are being studied as a therapeutic option for MMN. Eculizumab is a monoclonal antibody of complement C5 inhibitor that has been approved for patients with generalized myasthenia gravis. Fitzpatrick et al. [142] conducted an open‐label study of 13 MMN patients who received eculizumab for 14 weeks, 10 of whom were concomitantly receiving IVIG. The use of eculizumab did not modify IVIG dosing but improvements in patient‐rated subjective scores and selected clinical and electrophysiological measurements were observed. Using a human‐on‐a‐chip electrical conduction model, Rumsey et al. [144] demonstrated that MMN patient sera could activate complement pathways and reduce motor neuron action potential frequency and conduction velocity. Application of a monoclonal antibody of C1 inhibitor TNT005 to this model rescued the serum‐induced complement deposition and functional deficits, demonstrating its therapeutic potential for MMN.

Empasiprubart is a humanized monoclonal antibody against C2 generated in mice, and serves as an inhibitor of the classical and lectin pathways while leaving the alternative pathway intact [209]. In an induced pluripotent stem cell‐derived motor neuron model for MMN, Budding et al. [143] demonstrated that empasiprubart inhibits complement activation induced by MMN patient‐derived anti‐GM1 antibodies bound to motor neurons. Interim results from a phase II study to evaluate the safety and efficacy of empasiprubart in 27 adult patients with MMN (NCT05225675) showed positive treatment responses [210].

Summary and Future Directions

Most MMN patients are able to maintain active lives, although many have poor dexterity in manual tasks [8]. Issues with hand dexterity and fine motor skills often impact negatively patients' social and professional functioning in a significant fashion, and distal lower extremity weakness may result in imbalance and falls. The main treatment is IVIG or SCIG which may slow down the gradual decline of motor strength and functioning. However, deterioration of muscle strength continues to occur in those patients receiving immunoglobulin treatment and becomes more evident over time. The deterioration of strength correlates with the gradual emergence of axonal loss on EDX testing [211]. As such, early diagnosis of MMN is important and treatment should be initiated as early as possible.

It is evident that maintenance immunoglobulin therapy is unable to suppress the underlying disease process effectively in all patients, and more potent immunomodulatory agents are needed that could effectively stop the progression of MMN, or better, achieve a complete remission. In addition to the promising preliminary results from complement inhibitor usage, other potentially effective B‐cell depletion treatments (e.g., rituximab, inebilizumab) and chimeric antigen receptor T cell therapies could be explored in prospective investigational trials [212]. There is also a need to conduct controlled trials and evaluate the efficacy of designed therapy and exercises in the context of ongoing immunotherapy to maximize patients' functional capacities and independence.

Author Contributions

Benjamin Claytor: conceptualization, investigation, writing – original draft, writing – review and editing. David Polston: conceptualization, investigation, writing – original draft, writing – review and editing. Yuebing Li: conceptualization, methodology, investigation, supervision, writing – review and editing, writing – original draft.

Ethics Statement

The authors have read and understood Muscle and Nerve Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Conflicts of Interest

Dr. Yuebing Li has consulted for Amgen, Argenx, Catalyst, Immunovant and UCB Pharma, and has received grant support from Argenx. Dr. Benjamin Claytor and Dr. Polston have no disclosures.