The pathogenesis of multifocal motor neuropathy and an update on current management options

Abstract

Multifocal motor neuropathy (MMN) is a rare and disabling disease. Several experimental studies and clinical data are strongly suggestive of an immune-mediated pathogenesis, although underlying mechanisms in MMN seem to be very specific, mainly because of the presence of IgM anti-GM1 serum antibodies and the dramatic response to intravenous immunoglobulins (IVIg). The origin of antiganglioside antibodies and the way in which they act at the molecular level remain unclear. Several studies have demonstrated the key role of complement activation in the underlying mechanisms of MMN, as well as in animal models of acute motor axonal neuropathy (AMAN). Deposition of the membrane attack complex may disrupt the architecture of the nodes of Ranvier and paranodal areas, causing local disruption of nodal sodium-channel clusters. In patients with MMN, muscle weakness is the consequence of conduction blocks (CB), which leads to secondary axonal degeneration, consequently the aim of the treatment is to reverse CB at early stages of the disease. High-dose immunoglobulin is to date the only therapy which has proven efficacy in MMN patients in providing transient improvement of muscle strength, but long-term follow-up studies show a progressive motor decline. Therefore, other therapies are needed to improve the conduction nerve properties in long-term design. The reduction of complement activation and more generally the gain in paranodal stabilization could be directions for future therapeutic strategies.

Introduction

Multifocal motor neuropathy (MMN), which was first described in 1986 in original reports coming from two groups of authors [Roth et al. 1986; Chad et al. 1986], is a rare disease, with a prevalence of around 0.6 per 100,000 individuals. It is a purely motor neuropathy, characterized by progressive distal asymmetric limb weakness that usually starts and predominates in the upper limbs, with minimal or no sensory impairment. Nerve conduction studies have found multifocal persistent conduction blocks (CB) that distinguish MMN from motor neuron disease (MND). The association of MMN with high serum levels of IgM antibodies against the ganglioside GM1 were then reported, together with the positive effects of immunomodulatory treatments [Pestronk et al. 1988; Feldman et al. 1991]. These initial reports were followed by larger case series that described the clinical, electrophysiological and immunochemical features of patients with MMN.

High-dose intravenous immunoglobulin (IVIg) and subcutaneous immunoglobulin (SCIg) have been proven by randomized, controlled trials (RCT) to improve weakness in patients with MMN, and therefore are now considered to be the gold standard treatment of this disabling disease [Eftimov and van Schaik, 2011; Guimaraes-Costa et al. 2013a]. Consequently, MMN is considered a treatable disease, with regards to other chronic immune-mediated neuropathies, such as chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) or IgM anti-myelin-associated glycoprotein (anti-MAG) neuropathy. However, MMN does not respond to some immunomodulatory treatments that are effective in CIDP, and the effect of IVIg/SCIg on motor symptoms and signs may decline after several years.

The two main challenges in MMN are first to better understand the underlying pathogenic mechanisms and secondly to look for alternative disease-modifying treatments that prevent permanent weakness.

In this review, we discuss the different issues in the pathogenesis of MMN, and give an update on current management options, in providing the best strategies in differential diagnosis, diagnostic criteria and long-term treatment.

Pathophysiology

The two main issues for a better understanding of pathophysiology in MMN are to improve our knowledge both in the mechanism of CB, and in the role of antiganglioside antibodies.

CB as a key for explaining motor weakness

Motor CB is the characteristic, although shared with other dysimmune neuropathies, electrophysiological finding in MMN. Several authors showed that nerve dysfunction may be located at the nodes of Ranvier, with either hyperpolarization or depolarization, which both lead to failure of action potential conduction. In the years 2000–2006, several original articles studied the effect of maximum voluntary contraction (MVC), and hyperpolarization/depolarization at the site of CB in patients with MMN. In a first study [Kaji et al. 2000], the authors examined changes in muscle force during MVC and monitored CB before and after MVC in five patients with MMN, and compared the results with those of contralateral unaffected homonymous muscles, and those from patients with amyotrophic lateral sclerosis (ALS), with similar compound muscle action potential (CMAP) amplitudes after proximal stimulation. During MVC for 60 s, the affected muscles developed prominent fatigue, with a diminution of the force at the end of the contraction significantly lower than the initial force, when compared with controls. In addition, after MVC, the amplitude ratio of CMAPs after proximal versus distal stimulation transiently decreased when compared with that before MVC in the affected muscles, but not in controls. The authors concluded that activity-dependent CB may play a role in MMN, by causing muscle fatigue. The group of Bostock [Kiernan et al. 2002] reported a second study 2 years later, showing features of abnormalities in axonal membrane hyperexcitability in MMN patients, closely resembled those in normal axons hyperpolarized following release from ischaemia. To test for axonal hyperpolarization, depolarizing currents were applied to the nerves of MMN patients, and all of the excitability parameters were normalized by depolarization. The authors therefore suggested that the distal hyperpolarization is probably linked to focal depolarization and that the clinical features of MMN are consistent with a depolarizing/hyperpolarizing lesion. Another group [Priori et al. 2005] studied the effects induced by polarizing direct currents on motor conduction along forearm nerves in normal nerves, versus nerves at the site of CB in MMN patients. In controls, depolarization failed to change the CMAP, while hyperpolarization elicited a significant, charge-dependent decrease in the conditioned CMAP size. On the other hand, analysis of individual nerves in MMN patients showed that polarizing currents elicited markedly heterogeneous effects. In summary, pathophysiological abnormalities were consistent with either a depolarizing, a hyperpolarizing or a mixed block. Lastly, in a more recently reported study [Straver et al. 2011b], the authors aimed to confirm these previous data concerning the relationship between activity-dependent CB and weakness in pa-tients with MMN. They consequently employed supramaximal electrical stimulation in nerve segments of MMN patients, excluded nerves with marked axonal loss and adopted criteria for activity-dependent CB. The authors failed to find significant changes in mean areas ratios after MCV, that induced no activity-dependent CB. In segments with CB before MCV, the MCV induced increased duration prolongation. The authors concluded that MCV induced temporal dispersion but no activity-dependent CB.

Several mechanisms have been suggested to underlie membrane abnormalities and CB, including paranodal demyelination, disruption of nodal sodium-channel clusters, dysfunction of nodal sodium channels, and sodium–potassium pump hyperactivity.

The role of antiganglioside antibodies

A number of research articles have been published in the recent years on the role of gangliosides at the nodes of Ranvier, as potential target antigens in motor neuropathies, mainly axonal variants of Guillain-Barré syndrome, acute motor axonal neuropathy (AMAN), and acute motor-sensory neuropathy (AMSAN), and MMN [Yuki, 2013].

The antiganglioside antibodies and the role they play in the pathogenesis of AMAN, AMSAN, and most likely MMN are keys to understand the pathophysiology of these diseases.

GM1 is ubiquitously expressed, but has been found most abundantly in peripheral motor rather than sensory nerves. GM1 localizes to both the axolemma and myelin of the peripheral nerves, being found in greatest abundance at the nodes of Ranvier and adjacent paranodes [Willison and Yuki, 2002]. It concentrates in cholesterol-enriched domains of the plasma membrane. The functions of these domains are not fully understood, the most relevant being paranodal stabilization and clustering of ion channels [Susuki et al. 2007]. GM1 is thought to facilitate the maintenance of tight junctions through this paranodal stabilization as it also provides an anchor for potassium channels and concentrates sodium channels. These functions are necessary for action potential propagation and maintenance of conduction velocity. Disruption of these functions would be expected to induce failure of conduction across paranodal areas.

There has been much debate as to whether the antiganglioside antibodies are primary autoantibodies that cause neuropathies or are only biomarkers, resulting from the cascade leading to nerve injury. A number of data suggest the antiganglioside antibodies to have a pathogenic role in MMN. In 2011, the group of Yuki showed that anti-GM1 IgM antibodies coming from sera of MMN patients bound to GM1 and active complement in vitro [Yuki et al. 2011]. The depositions of complement component deposits were highly correlated with anti-GM1 antibody titre. In addition, IVIg reduced the deposition of these complement components dose-dependently. The authors postulate that IgM-induced, complement-mediated injury, occurs at the nodes of Ranvier in peripheral motor nerves and generates CB and motor weakness. In vitro IVIg inhibited the complement activation, suggesting that in vivo, the resulting reduction in membrane attack complex (MAC)-mediated damage leads to improvement of muscle strength. Thus, complement activation may be considered a key effect in the immunologic cascade producing pathological effects.

More recently, other authors wished to ascertain whether sera from MMN patients may open the blood–nerve barrier (BNB) [Shimizu et al. 2014]. They mainly found that these sera decreased the claudin-5 protein expression and the transendothelial electrical resistance in human peripheral nerve microvascular endothelial cells (PnMECs). This effect was reversed after application of anti-vascular endothelial growth factor (anti-VEGF) neutralizing antibody. The authors concluded that sera from MMN patients may disrupt the BNB function via the secretion of VEGF in PnMECs, or the exposure to autoantibodies against PnMECs that are contained in the MMN sera. Consequently, the disruption of the BNB would allow for the entry of a number of circulating inflammatory cells in the peripheral nerve.

To summarize, although an animal model of MMN does not exist, the mechanism of damage from IgM GM1 could be similar to that observed in the rabbit model of AMAN, in which IgM GM1 activates the complement cascade leading to the production of a MAC. MAC compromises membrane integrity, then spreads to the internodal regions of the nerve, leading to disruption of sodium channels, allowing further binding of antibodies to the axolemma [O’Hanlon et al. 2003]. The use of complement inhibitors abrogates the antiganglioside antibody mediated damage in animal models, providing further evidence for complement dependent pathology.

The presence of anti-GM1 IgM antibodies has been documented in the earliest descriptions of MMN, but their prevalence is not similar in the reported series, probably linked to methodological differences and the lack of a gold standard to measure the titres of these antibodies. In a retrospective study of 40 MMN patients [Léger et al. 2008], we found that 21 (64%) of patients had IgM anti-GM1 antibodies, while this percentage was 43% in a recently reported series of 88 MMN patients [Cats et al. 2010a]. Moreover, in this series, the presence of anti-GM1 IgM antibodies was significantly correlated with severe weakness, disability and axonal loss, while no correlation was found between the presence of anti-GM1 IgM antibodies and response to IVIg therapy in our study. In a previously reported RCT, we also found that anti-GM1 antibody titres remained relatively stable during follow up in both placebo and IVIg-treated patients [Léger et al. 2001]. Other rare MMN patients may have anti-GM2 and anti-GD1b IgM antibodies, both shown to be cross-reactive with GM1 in absorption studies [Cats et al. 2010a].

A large proportion of MMN patients do not have significant titres of anti-GM1 antibodies in enzyme-linked immunosorbent assay (ELISA). It is unknown whether these patients have low titres of these serum antibodies that are undetectable with ELISA, or whether they have other antibodies directed against antigen targets. The clinical characteristics of MMN patients with and without anti-GM1 antibodies do not differ, and the treatment response is similar in both groups [Léger et al. 2008]. However, the findings of antiganglioside complexes, which have been shown to enhance or reduce GM1 binding, may provide an explanation [Nobile-Orazio et al. 2010]. Several original articles aimed to increase the detection of auto-antibodies in MMN. The group of Pestronk reported results in IgM binding to NS6S heparin disaccharide, which was associated with motor neuropathy syndromes and occurred with similar frequency to IgM binding to GM1 [Pestronk et al. 2010]. The authors concluded that testing for IgM binding to NS6S in addition to GM1 may increase the frequency of finding IgM antibodies in motor neuropathies from 43% to 64%. A newly developed combinatorial gycoarray able to identify antibodies to 45 different heteromeric glycolipid complexes and their 10 individuals glycolipids components was then reported [Galban-Horcajo et al. 2012]: by ELISA, 22/33 MMN cases had detectable anti-GM1 IgM antibodies, whereas 19/33 MMN samples were positive for anti-GM1 antibodies by glycoarray. Antibodies to the GM1/galactocerebroside (GM1/GalC) complex were more significantly associated with MMN, returning 33/33 MMN samples as positive with glycoarray and 29/33 positive by ELISA. In a more recent study, measurement of IgM antibodies to GM1, GM2, galactocerebroside, GM1/GalC and NS6S was compared in 40 consecutive patients coming from two centres including ours [Nobile-Orazio et al. 2013]. Significant titres of IgM anti-GM1 antibodies were found in 48% of MMN patients. Anti-GM1/GalC antibodies were present in all anti-GM1 positive MMN patients and in 11 additional patients (28%) with MMN raising the sensitivity of antibody testing to 75%. The specificity for MMN was, however, moderately reduced compared with anti-GM1 IgM, even if they rose with increasing anti-GM1/GalC titres. Our conclusion was that testing for anti-GM1/GalC IgM significantly increased the sensitivity of antibody testing in MMN, when compared with anti-GM1 alone, and may represent a preferred option for GM1 reactivity testing in MMN.

Update on current management options

Update in diagnostic features

Clinical aspects and electrophysiological abnormalities mainly support the diagnosis of MMN, both of them having been recently reviewed by a Task Force of both European Federation of Neurological Societies (EFNS) and Peripheral Nerve Society (PNS), which edited a Guideline on management of multifocal motor neuropathy, firstly revised in 2010 [van Schaik et al. 2010]. In addition, the diagnosis may be supported by laboratory and imaging data, as well as previously described immunological features [Guimaraes-Costa et al. 2013b].

Clinical characteristics

The main clinical diagnostic criteria (Table 1) are weakness without objective sensory loss, asymmetric motor deficit in the territory of two or more motor nerves together with cramps and fasciculations, absence of upper motor neuron signs, and slowly or stepwise progressive course. Several reported series aimed to precisely identify the frequency of both classical and unusual symptoms and signs [Slee et al. 2007; Cats et al. 2010b]. In the second one, conducted within a national cross-sectional description study in 88 patients with MMN, onset of weakness was in distal arm (61%) or distal leg (34%), and occasionally in the upper arm, but never in the upper leg. Symptom onset was more frequent in the dominant hand. At neurological examination, reflexes were usually diminished or abolished in the affected territories, but brisk, unless not pathologic reflexes, were found in 8%. In the first one regrouping 46 patients with MMN [Slee et al. 2007], differential motor deficit across muscles supplied by a common terminal motor nerve was seen in about 54% of patients and differential finger extension weakness was a frequent early manifestation, likely reflecting vulnerabilities of terminal branches of the posterior interosseous nerve.

Table 1.

Clinical criteria for multifocal motor neuropathy [Van Schaik et al. 2010].

Core criteria (both must be present)

1. Slowly progressive or stepwise progressive, focal, asymmetric* limb weakness, that is, motor involvement in the motor nerve distribution of at least two nerves, for more than 1 month.† If symptoms and signs are present only in the distribution of one nerve only a possible diagnosis can be made (Table 4)

2. No objective sensory abnormalities except for minor vibration sense abnormalities in the lower limbs‡

Supportive clinical criteria

3. Predominant upper limb involvement§

4. Decreased or absent tendon reflexes in the affected limb¶

5. Absence of cranial nerve involvement**

6. Cramps and fasciculations in the affected limb

7. Response in terms of disability or muscle strength to immunomodulatory treatment

Exclusion criteria

8. Upper motor neuron signs

9. Marked bulbar involvement

10. Sensory impairment more marked than minor vibration loss in the lower limbs

11. Diffuse symmetric weakness during the initial weeks

^*Asymmetric = a difference of 1 MRC grade if strength is MRC >3 and 2 MRC grades if strength is MRC ⩽ 3.

^†Usually more than 6 months.

^‡Sensory signs and symptoms may develop over the course of MMN.

^§At onset, predominantly lower limb involvement account for nearly 10% of the cases.

^¶Slightly increased tendon reflexes, in particular in the affected arm, have been reported and do not exclude the diagnosis of MMN provided criterion 8 is met.

^**Twelfth nerve palsy has been reported.

Patients frequently experience an exacerbation of weakness with cold, in line with the disruption of nodal sodium-channel clusters and dysfunction of nodal sodium channels, suggested above as possible underlying mechanisms [Straver et al. 2011a]. In addition, a marked motor deficit with no or slight amyotrophy may probably be considered a hallmark of the disease and the clinical expression of CB [van Schaik et al. 2010]

Interestingly, the extent of sensory signs and symptoms has been reconsidered and the development of electrophysiological sensory changes, with or without sensory signs and symptoms over the course of MMN, has been reported (see below) [Lievens et al. 2009; Delmont et al. 2009]. Transient paresthesia may be reported by some patients, although they are not usually accompanied by objective loss of sensation, except for a slight diminution of vibration sense in the only distal legs. Abnormal vibration sense in distal leg was found in 22% of MMN patients in one series [Cats et al. 2010b], and was linked to a longer median disease duration, when compared with patients without sensory involvement. The explanation for this sensory involvement could be due to the fact that sensory nerves are less vulnerable to damage following anti-GM1 IgM antibody-binding [Harschnitz et al. 2014]. Cranial nerve involvement is rare and has been scarcely reported for the hypoglossal and the abducens nerves. The correlation could be considered when cranial involvement stands or relapses in association with other features of MMN [Galassi et al. 2012].

Finally, in one study [Cats et al. 2010b], 18% of patients reported minimal or no disability of the arms (Overall Disability Sum Score: ODSS = 0 and 1), 61% had moderate impairment of the arms (ODSS = 2), whereas 21% had severe disability (ODSS = 3). The mean response to the fatigue severity scale (FSS) was 4.7 with severe fatigue (FSS > 5) present in 51% of patients. Interestingly, univariate analysis suggested that more severe disability was associated with more axonal loss, years untreated, symptom onset in a leg, and presence of IgM anti-GM1 antibodies; multivariate analysis identified only axonal loss as an independent determinant of more severe disability.

Electrophysiological features

The presence of CB in the only motor nerves, outside the usual sites of nerve compression, is the hallmark of the disease. However, some patients with otherwise typical MMN have no detectable CB, probably because these blocks are activity-dependent (see above) or are located in segments which cannot be assessed by routine electrophysiological examination [Delmont et al. 2006]. Other techniques with restricted availability, such as transcranial magnetic stimulation, triple-stimulation technique, and transcutaneous cervical root stimulation have been used to identify CBs with greater sensitivity. These techniques may be useful, especially where CBs are proximally situated [Attarian et al. 2005; Deroide et al. 2007].

The first articles defined CB as a 20–30% amplitude or area reduction if the distal CMAP duration did not exceed 15% greater than normal. Computer modelling of CB and temporal dispersion in an animal model has demonstrated that up to 50% area reduction of the proximal to distal CMAP can be due entirely to interphase cancellation [Rhee et al. 1990]. Similar studies in man have shown that distal CMAP duration and proximal CMAP duration prolongation are important factors for the definition of CB in the median nerve segment over the forearm: the shorter the distal duration and proximal duration prolongation, the less CMAP amplitude reduction is needed to diagnose a CB [van Asseldonk et al. 2006]. In the EFNS/PNS Guideline [van Schaik et al. 2010], grading of CB was defined as definite or probable (Table 2). In their retrospective study of 88 patients [Cats et al. 2010b], the authors found that 81% had at least one definite CB, and 18% had no definite but at least one probable CB. CB was most often detected in the ulnar (80%) and median (77%) nerves. Other electrophysiological alterations may be found such as prolonged distal motor latencies, prolonged or absent F-waves, minor reduction of motor nerve conduction velocities and temporal dispersion [Guimaraes-Costa et al. 2013b]. Electromyography may disclose fasciculations and fibrillation potentials, usually detected in territories with marked amyotrophy as they are consistent with significant axonal degeneration.

Table 2.

Electrophysiological criteria for conduction block^* [Van Schaik et al. 2010].

1. Definite motor CB

Negative peak CMAP area reduction on proximal vs. distal stimulation of at least 50% whatever the nerve segment length (median, ulnar, and peroneal). Negative peak CMAP amplitude on stimulation of the distal part of the segment with motor CB must be >20% of the lower limit of normal and >1 mV and increase of proximal to distal negative peak CMAP duration must be ⩽30%

2. Probable motor CB*

Negative peak CMAP area reduction of at least 30% over a long segment (e.g. wrist to elbow or elbow to axilla) of an upper limb nerve with increase of proximal to distal negative peak CMAP duration ⩽30%

OR

Negative peak CMAP area reduction of at least 50% (same as definite) with an increase of proximal to distal negative peak CMAP duration >30%

3. Normal sensory nerve conduction in upper limb segments with CB (see exclusion criteria)

CB, conduction block; CMAP, compound muscle action potential.

^*Evidence for CB must be found at sites distinct from common entrapment or compression syndromes.

Sensory potentials are usually normal, but have been scarcely studied in long-term follow up. In a retrospective study of 21 patients with MMN with a mean follow up of 7 ± 2.5 years [Lievens et al. 2009], we found that 13 patients (62%) had a reduction of the amplitude of at least one sensory potential, of whom four patients had abnormalities of two or more sensory potentials, while eight patients had no abnormality. No significant difference was found for gender, age at onset, number of involved motor nerves, presence/absence of IgM anti-GM1 antibodies and response to IVIg between the two groups.

Supportive criteria and diagnostic categories

Supportive criteria, including detection of IgM anti-GM1 antibodies (see above) are listed in Table 3. MRI may disclose hyperintense signals on T2-weighted sequences or contrasted-enhanced T1 sequences in the brachial plexus. Alterations usually correlate with the distribution of muscle weakness and eventually with CB [van Es et al. 1997; Echaniz-Laguna and Dieteman, 2011].

Table 3.

Supportive criteria [Van Schaik et al. 2010].

1. Elevated IgM anti-ganglioside GM1 antibodies

2. Laboratory: increased cerebrospinal fluid protein (<1 g/l)

3. Magnetic resonance imaging showing increased signal intensity on T2-weighted imaging associated with a diffuse nerve swelling of the brachial plexus

4. Objective clinical improvement following IVIg treatment

IVIg, intravenous immunoglobulin.

Summarizing all of the above diagnostic criteria, together with clinical response to IVIg, the Guideline [van Schaik et al. 2010] agreed on defining diagnostic categories for MMN (Table 4).

Table 4.

Diagnostic categories [Van Schaik et al. 2010].

Definite MMN

Clinical criteria 1, 2, and 8–11 (Table 1) AND electrophysiological criteria 1 and 3 in one nerve (Table 2)

Probable MMN

Clinical criteria 1, 2, and 8–11 AND electrophysiological criteria 2 and 3 in two nerves

Clinical criteria 1, 2, and 8–11 AND electrophysiological criteria 2 and 3 in one nerve AND at least

two supportive criteria 1–4 (Table 3)

Possible MMN

Clinical criteria 1, 2, and 8–11 AND normal sensory nerve conduction studies AND supportive criteria 4

Clinical criteria 1 with clinical signs present in only one nerve, clinical criteria, 2, and 8–11 AND electrophysiological criteria 1 or 2 and 3 in one nerve

MMN, multifocal motor neuropathy.

Other features

Associated autoimmune conditions to MMN, such as coeliac disease and Hashimoto’s thyroid disease were recently reported [Cats et al. 2012]. Moreover, first-degree family members also apparently have a higher incidence of diabetes type 1, coeliac disease and Hashimoto’s thyroid disease. These associations may suggest a common pathogenic mechanism and underline the importance of searching for associated conditions that may contribute to impair patient’s quality of life. In the same way, a higher expression of HLA-DRB1*15 was found in patients with MMN [Sutedja et al. 2010], which may suggest a similar pathogenic pathway to other diseases such as in multiple sclerosis and CIDP, which share a higher frequency of this specific human leukocyte antigen (HLA) than in controls. Unfortunately, until the present, no correlation with this HLA and age of onset, clinical course or disease severity had been found.

Update in treatment options

MMN mostly has a chronic slowly or stepwise progressive course. Therapies aim to reduce motor deficit along with CB, slow down ongoing axonal degeneration or even promote reinnervation and remyelination. On the other hand, a minority of patients achieve prolonged remissions upon treatment, but most patients need treatment for years. In a retrospective study in 40 patients with MMN [Léger et al. 2008], we found that, within a mean follow up of 2.2 ± 2.0 years, only eight patients (22%) had significant remission, defined as lasting stabilization of clinical improvement >6 months, without further treatment, after initial IVIg therapy during at least 6 months, whereas 25 patients (68%) were dependent on maintenance IVIg infusions to stabilize their motor condition. Among these 25 patients, eight were given additional immunosuppressive agents during various periods. However, the recommendations from the Guideline [van Schaik et al. 2010] concluded that if IVIg is not sufficiently effective, then immunosuppressive treatment may be considered, but no agent has shown to be beneficial in a clinical trial (see below). In another retrospective study of 88 MMN patients [Cats et al. 2010b], within a mean follow-up of 6 years (range 0–17), 67 patients (76%) received IVIg maintenance therapy at time of the study, while 17 patients (19%) did not use maintenance treatment for several reasons, including lack of beneficial effect (five patients), and stable disease course without treatment (eight patients). A total 35 (40%) patients had unsuccessfully used other immunomodulators, mainly interferon-beta and mycophenolate mofetil.

The crucial point in treating patients with MMN is that therapies are usually accompanied by a considerable social and financial burden. Some goals have been recently achieved in (1) the confirmation that IVIg therapy not only improves muscle strength but also limits disability, (2) the confirmation that SCIg is a proven alternative treatment, and (3) a better definition of specific outcome measures.

Intravenous immunoglobulin

IVIg remains the treatment of choice for patients with MMN [Léger et al. 2007; Donofrio et al. 2009; van Schaik et al. 2010]. IVIg is recommended as a first-line therapy based on a meta-analysis of four randomized, double-blind, placebo-controlled, trials [Azulay et al. 1994; van den Berg et al. 1995; Federico et al. 2000; Léger et al. 2001], involving a total of 34 patients: 78% of study subjects had a significant short term improvement in strength, selected as primary outcome measure, following IVIg treatment, when compared to 4% following placebo [van Schaik et al. 2005]. Despite that, the mechanisms of action are not clearly understood. IVIg neutralizes pathogenic antibodies and superantigens, inhibits antibody production by B cells and accelerates catabolism of antibodies, suppresses pro-inflammatory mediators produced by T cells, inhibits complement-mediated inflammation and damage, induces blockade of Fc receptors on macrophages and regulates proliferation and adhesion of T cells [Kazatchkine and Kaveri, 2001; Dalakas, 2002; Jacob and Rajabally, 2009; Yuki et al. 2011]. A recently published controlled trial aimed to assess the efficacy, safety and tolerability of 10% liquid IVIg in 44 MMN patients [Hahn et al. 2013]. Patients were randomized 1:1 to either double-blind treatment of IVIg followed by placebo for 12 weeks each or the reverse. Open-label IVIg was administered for 12 weeks at the beginning and end of the study for clinical stabilization, and between double-blind periods to prevent a carry-over effect. Mean maximal grip strength of the more affected hand declined 31.38% during placebo and increased 3.75% during IVIg (p = 0.005). In addition, in 37.5% of participants, the disability scale (Guy’s Neurological Disability score [GNDS] for upper limbs) worsened during placebo and not during IVIG, whereas the converse was observed in 11.9% (p = 0.021); 69% switched prematurely from placebo to open label, because of significant deterioration, and 2.4% switched from blinded to open-label IVIg (p < 0.001). The authors concluded that IVIG was effective in improving both muscle strength and disability in MMN patients.

IVIg, at a cumulative dose of 2 g/kg, was efficacious as well in 70% of 22 treatment-naïve MMN patients in our retrospective study [Léger et al. 2008], and in 94% of 84 MMN patients in another retrospective study [Cats et al. 2010b], both based on an increase of at least one Medical Research Council (MRC) grade in at least two muscle groups, without a decrease in other muscle groups. Analysis of predictive criteria in our study revealed that the two best predictive factors for response to IVIg (although not significant) were female gender (p = 0.08) and lower MRC score at inclusion (p = 0.07). In addition, among the 22 treatment-naïve patients, the number of CBs decreased for eight patients, with complete disappearance of CB for two patients, remained stable for four patients and increased for two patients. Consequently, no pertinent correlation was found in this study between clinical improvement and electrophysiological features, as it was the case for anti-GM1 antibody titres in another study (see above). Good practice points for treatment [van Schaik et al. 2010] recommend an initial dose of 2 g/kg given over 2–5 consecutive days, and if effective followed by maintenance therapy, ideally 1 g/kg every 2–4 weeks or 2 g/kg every 1–2 months (Table 5).

Table 5.

Recommendations and good practice points [Van Schaik et al. 2010].

Good practice points for diagnostic criteria

1. Clinical: the two core criteria and all exclusion criteria should be met (Table 1).

2. Electrodiagnostic: definite or probable CB in at least one nerve (Table 2).

3. Supportive: anti-GM1 antibodies, MRI, CSF, and treatment response (Table 3).

4. Categories: definite and probable MMN (Table 4).

Good practice points for diagnostic tests

1. Clinical examination and electrodiagnostic tests should be done in all patients.

2. Antiganglioside GM1 antibody testing, MRI of the brachial plexus, and CSF examination should be considered in selected patients.

3. Investigations to discover concomitant disease or exclude other possible causes should be considered, but the choice of tests will depend on the individual circumstances.

Good practice points for treatment

1. IVIg (2 g/kg given over 2–5 days) should be the first line treatment (level A) when disability is sufficiently severe to warrant treatment.

2. Corticosteroids are not recommended.

3. If an initial treatment with IVIg is effective, repeated IVIg treatment should be considered in selected patients (level C). The frequency of IVIg maintenance therapy should be guided by the response. Typical treatment regimens are 1 g/kg every 2–4 weeks or 2 g/kg every 1–2 months.

4. If IVIg is not sufficiently effective then immunosuppressive treatment may be considered. However, no agent has shown to be beneficial in a clinical trial and data from case series are conflicting.

5. Toxicity makes cyclophosphamide a less desirable option.

CB, conduction block; MRI, magnetic resonance imaging; CSF, cerebrospinal fluid; MMN, multifocal motor neuropathy; IVIg, intravenous immunoglobulin.

Although standard IVIg therapy is effective in almost all patients with MMN, the treatment does not prevent a gradual decline in muscle strength over time. Several long-term studies have assessed this issue in recent years. In a first study [Van den Berg-Vos et al. 2002], the authors performed a long-term follow-up of 11 patients with MMN, who were treated initially with one full course of IVIg (2 g/kg), then 0.4 g/kg every week, followed by maintenance therapy ranging from one infusion every 1–7 weeks. During the 4–8-year follow-up period, patients were assessed by MRC sumscore of 20 muscle groups, handheld dynamometry on a selection of weak muscle groups, electrophysiological studies and GNDS. Muscle strength improved significantly within 3 weeks of the start of IVIg treatment and was significantly better at the last follow-up examination than before treatment, even though it decreased slightly and significantly during the follow-up period. CB disappeared in six nerve segments but new CB appeared in eight nerve segments during the follow-up period. Changes consistent with improvement (demyelination or reinnervation) occurred in 13 nerves during follow-up and changes consistent with worsening (demyelination or axonal loss) occurred in 14 nerves. The authors concluded that IVIg maintenance therapy had a beneficial long-term effect on muscle strength and upper limb disability, but may not prevent a slight decrease in muscle strength. The electrophysiological findings imply that IVIg treatment favourably influenced the mechanisms of remyelination or reinnervation, but that axon loss cannot be prevented. Similar results were found by other authors [Terenghi et al. 2004] who reported follow up in 10 MMN patients responding to an initial course of IVIg with periodic infusion for 5–12 years (mean 8.2 years). At last follow up, only two patients had maintained the maximal improvement achieved during therapy, while eight worsened despite Ig dosage. This decline started after 3–7 years (mean 4.8 years) of therapy and correlated with a reduction of distal CMAP amplitudes. On the other hand, a third study [Vucic et al. 2004] reported significant and sustained improvement in muscle strength (assessed by MRC in eight muscle groups), disability (assessed by modified Rankin score), CB and signs of axonal degeneration on electrophysiological studies, in 10 MMN patients, with a follow up of 3.5–12 years (mean 7.25 years). The authors concluded that long-term IVIg therapy improved muscle strength and functional disability, decreased the number of CBs and the extent of axonal degeneration, and promoted reinnervation. The difference from previous findings may be explained by the different regimens in giving IVIg, the patients in this last study being treated with significantly higher IVIg maintenance doses.

Limitations in long-term IVIg therapy are adverse side effects, mainly thromboembolic events, anaphylactic reactions which may occur in patients with selective IgA deficiency and anti-IgA antibodies, renal tubular necrosis which may be prevented by detection of pre-existent renal disease and correct hydration, and possible difficulties with venous access.

Subcutaneous immunoglobulin

The last above limitation have led to investigate Ig maintenance therapy by subcutaneous administration. Weekly SCIg has been proven to be a good alternative in terms of similar efficacy and patients’ quality of life, as SCIg may be given in self-infusions at home. A randomized, single-blinded cross-over study was conducted in nine IVIg responsive patients [Harbo et al. 2009] and assessed by dynamometric strength of affected muscles and the SF-36 quality of life questionnaire. IVIg and SCIg were equally effective, the mean change in muscle strength after SCIg being 3.6%, versus 4.3% after IVIg. One patient presented sustained erythema and oedema at the injection sites for a few weeks, but all other adverse events with SCIg were mild and transient. No difference was significant between the two treatments for the SF-36 score. After the study, five out of nine patients preferred continuation with SCIg. Another open-label study [Eftimov et al. 2009] conducted in 10 patients similarly found that SCIg therapy was feasible and safe and maintained strength as well as IVIg. Finally, a 2-year follow-up study was reported by the authors of the first study [Harbo et al. 2010] in six IVIg-responsive MMN patients. The dosage of SCIg varied between 13 and 51 g per week, corresponding to a volume of 80 to 320 ml, infused twice or thrice weekly. No major side effects were reported, including local skin reactions being mild and transient. The impairment and disability scores remained stable.

Other immunosuppressant and immunomodulatory agents

Corticosteroids and plasma exchanges, which are effective in CIDP, are not recommended in MMN, and may lead to worsening of the motor condition [van Schaik et al. 2010]. Consequently, they should not be considered a therapeutical option in patients with MMN.

On the other hand, some patients with MMN do not respond to IVIg, and some others require progressively more frequent doses to maintain remission, or have an involvement of new motor nerves, despite periodic IVIg infusions. Therefore, there is a need for searching for alternative or adjunctive immunosuppressive therapies [Umapathi et al. 2012].

Among possible alternative/adjunctive therapies, cyclophosphamide and mycophenolate mofetil have been to date the more frequently proposed. In addition, eculizumab has been tried as a specific immunomodulatory agent.

Cyclophosphamide

Several uncontrolled studies, including the historical one [Pestronk et al. 1988] have suggested the efficacy of cyclophosphamide as alternative therapy. However, the Guideline [van Schaik et al. 2010] concluded in considering this immunosuppressor a less desirable therapeutical option, mainly because of its toxicity and the lack of evidence of efficacy. As an adjuctive treatment, only one uncontrolled study showed a reduction of the frequency of IVIg infusions in six patients [Meucci et al. 1997], but three patients presented severe side effects.

Mycophenolate mofetil

A randomized, single-centre, placebo-controlled, ‘add-on’ study of mycophenolate mofetil 1 g twice a day for 1 year [Piepers et al. 2007] was conducted in 28 MMN patients. The results failed in showing significant IVIg dose reduction, nor difference in muscle strength, functional scores and IgM anti-GM1 antibody titres, between patients having received mycophenolate mofetil or placebo.

Eculizumab

The rationale of the treatment with this original immunomodulator is linked to the fact that several experimental studies showed that the pathogenic effect of anti-GM1 antibodies is complement mediated, and consequently inhibition of complement factors may prevent nerve damage (see above). The monoclonal antibody eculizumab binds and neutralizes human complement factor C5 preventing terminal complement activation and membrane lysis via MAC. Its safety and efficacy has been proven in complement-mediated diseases, mainly paroxysmal nocturnal haemoglobinuria. An open-label clinical trial was conducted in 13 MMN patients with eculizumab for 14 weeks, 10 of whom in association with IVIg [Fitzpatrick et al. 2011]. The results disclosed a trend towards an improvement in patient-rated subjective scores and increased muscle strength as measured by myometry. In electrophysiological studies, there was a small yet significant net decrease in the median percentage CB across all nerves studied. The authors concluded in a small treatment effect occurring in some patients that appeared supplementary to an independent of the IVIg treatment effect, and occurred more frequently in patients with higher baseline motor function. This very original trial raises new horizons for further research concerning the use of complement inhibitors in the treatment of MMN.

Outcome measures for MMN

This review clearly shows that different subsets of outcome measures have been used for the evaluation of motor deficit and disability in MMN patients, but in the same way there is no consensus concerning a primary outcome measure both in clinical trials and in clinical practice.

Several outcome measures have been proposed at the impairment and activity and participation levels for MMN, to be used in future trials. However, some original outcome measures were recently outlined from the results of the PeriNoms Study (PNS), which was conducted between 2007 and 2012, on the auspice of the Inflammatory Neuropathy Consortium, a special interest group of the PNS [van Nes et al. 2011; Vanhoutte et al. 2012; Merkies et al. 2012]. Recommendations were edited by a group of experts who met at the 196th ENMC International workshop in February 2013 [Vanhoutte et al. 2013]. The primary outcome in MMN in future trials should, at the activity and participation level, be measured by the disease specific Rasch-built Overall Disability Scale (R-ODS). The minimal core set for future MMN studies was defined as following: at the impairment level, the Martin Vigorimeter should be used; in addition, patient-specific affected muscles should be incorporated and future studies are needed to redetermine these patient-specific MRC sum scores that may vary from patient to patient; the use of Rasch-transformed MRC scores (RT-MRC) was proposed. At the activity and participation level, the R-ODS MMN was suggested. At the quality of life level, a Rasch-transformed quality of life scale based on the future findings of the PNS was suggested. Concerning future needs in MMN, the experts concluded that the R-ODS could be expanded.

Conclusion

MMN is a rare but treatable disease. It is considered as an immune-mediated disease, however underlying mechanisms seem to be very specific. The two main issues for a better understanding of the pathophysiology in MMN are to improve our knowledge both in the mechanism of CB, and in the role of antiganglioside antibodies. Several mechanisms have been suggested to underlie membrane abnormalities and CB, including paranodal demyelination, disruption of nodal sodium-channel clusters, dysfunction of nodal sodium channels, and sodium–potassium pump hyperactivity. In addition, other studies have demonstrated the key role of complement activation in the pathophysiology of MMN as well as in animal models of AMAN. Future research will aim to explore the binding epitope of anti-GM1 antibodies, and if heteromeric complexes containing GM1 and other components may influence the binding affinity of these and other antibodies. IVIg/SCIg are to date the only therapy which has proven efficacy in MMN patients in providing transient improvement of muscle strength, but long-term follow-up studies with IVIg show a progressive motor decline. Thus, other therapies are needed to improve the conduction nerve properties in long-term design, as IVIg/SCIg periodic infusions are constraining and expensive. The reduction of complement activation and more generally the gain in paranodal stabilization could be directions for future therapeutical strategies.