Multifocal motor neuropathy: update on clinical characteristics, pathophysiological concepts and therapeutic options.

Fax +41 61 306 12 34
E-Mail karger@karger.ch
www.karger.com
Review
Eur Neurol 2010;63:193–204
DOI: 10.1159/000282734
Multifocal Motor Neuropathy: Update on Clinical
Characteristics, Pathophysiological Concepts and
Therapeutic Options
Sven G. Meuth Christoph Kleinschnitz
Department of Neurology, University of Wuerzburg, Wuerzburg , Germany
Introduction
More than 20 years ago Roth et al. [1] reported a pa-
tient with chronic asymmetric, distal motor neuropathy
without sensory loss. Electrophysiological examination
revealed proximal multifocal persistent conduction
blocks (CBs) outside the common entrapment sites. Soon
afterwards, others described individuals with similar
characteristics [2, 3] . The term ‘multifocal motor neuro p-
athy’ (MMN) was coined in 1988 by Pestronk et al. [4]
who first recognized the association of MMN with anti-
GM1-IgM antibodies and the responsiveness to immune-
modulating therapies. Since then, systematic clinical
and electrophysiological evaluation of larger patient co-
horts increased our pathophysiological understanding of
MMN and paved the way for more effective treatments
[5–10] . Especially the successful application of intrave-
nous immunoglobulins (IVIgs) marked a cornerstone in
MMN therapy and is nowadays regarded as the gold stan-
dard [10–16] . More recently, diagnostic criteria for this
rare neuropathy have been proposed by various Europe-
an and American neurological associations [17, 18] which
help to delineate MMN from other neuropathies such as
chronic inflammatory demyelinating polyneuropathy
(CIDP) or multifocal acquired demyelinating sensory
and motor (MADSAM) neuropathy (Lewis-Sumner syn-
drome) and motor neuron disease (MND).
Although MMN has meanwhile been identified as a
distinct nosological entity and significant success has
Key Words
Multifocal motor neuropathy � Conduction block �
Anti-GM1 antibodies � Ion channels
Abstract
Multifocal motor neuropathy (MMN) is an acquired immune-
mediated neuropathy characterized by chronic or stepwise
progressive asymmetrical limb weakness without sensory
deficits. The upper extremities are more often affected than
the lower extremities with distal paresis dominating over
proximal paresis. Important diagnostic features are persis-
tent multifocal partial conduction blocks (CBs) and the pres-
ence of high-titer anti-GM1 serum antibodies. Motor neuron
disease, other chronic dysimmune neuropathies, such as
chronic inflammatory demyelinating polyneuropathy and
the Lewis-Sumner syndrome (MADSAM neuropathy), are
important differential diagnoses. While corticosteroids and
plasma exchange are largely ineffective, high-dose intrave-
nous immunoglobulins are regarded as first-line treatment.
In spite of significant success in elucidating the underlying
disease mechanisms in MMN during the past few years, im-
portant pathophysiological issues and the optimum long-
term therapy remain to be clarified. The present review sum-
marizes the clinical picture and current pathophysiological
concepts of MMN with a special focus on the molecular and
electrophysiological basis of CBs and highlights established
therapies as well as possible novel treatment options.
Copyright © 2010 S. Karger AG, Basel
Received: November 12, 2009
Accepted: November 17, 2009
Published online: February 11, 2010
Christoph Kleinschnitz, MD
Department of Neurology, University of Wuerzburg
Josef-Schneider-Strasse 11, DE–97080 Wuerzburg (Germany)
Tel. +49 931 2012 3765, Fax +49 931 2012 3488
E-Mail christoph.kleinschnitz @ mail.uni-wuerzburg.de
© 2010 S. Karger AG, Basel
0014–3022/10/0634–0193$26.00/0
Accessible online at:
www.karger.com/ene

Meuth/Kleinschnitz

Eur Neurol 2010;63:193–204194
been made in elucidating important aspects of the dis-
ease, several issues remain to be clarified. For example,
there are still unsettled questions concerning the etiology
of MMN, the biological basis of CBs as well as the opti-
mum long-term therapy [8, 19–21] .
Clinical Features and Disease Course
MMN is a rare disease with an estimated prevalence
of 1–2/100,000 individuals. It is more frequent in men
than women, with an approximate ratio of 3: 1. The mean
age at disease onset is 40 years. Almost 80% of the pa-
tients develop first symptoms between 20 and 50 years of
age [7, 9] . Thus, MMN predominantly affects young peo-
ple. Clinically, MMN is characterized by slowly progres-
sive or stepwise progressive, asymmetric and distally ac-
centuated paresis related to distinct peripheral nerves.
The upper limbs are usually affected earlier and more se-
vere than the lower limbs [6, 7, 22, 23] . In only 5–10% of
all cases MMN manifests with proximal muscle weak-
ness [9, 24] . The most common initial symptom is wrist
drop and impaired grip strength. Muscle atrophy is often
mild in the early stage, but may become prominent dur-
ing the course of the disease when it is usually associated
with a poor response to immunomodulatory therapy [13,
25] . Other symptoms comprise fasciculations and muscle
cramps in about 50% of the patients, while myokymia has
only been reported occasionally [1, 25] . Another charac-
teristic that defines MMN is the absence of sensory symp-
toms. Only a few patients complain of discrete paresthe-
sia or numbness during the course of the disease, and a
minor loss of vibration sense has been documented in
20% of the subjects [7, 9] . Tendon reflexes from the pa-
retic muscles are usually reduced but may be normal or
even, though rarely, brisk. In the latter case, differentia-
tion from amyotrophic lateral sclerosis or lower motor-
neuron disease can be difficult. Cranial nerve involve-
ment is uncommon and, if present, predominantly af-
fects the N. Hypoglossus [26, 27] .
Most patients develop a slowly progressive disease
course in which the degree of disability correlates with
the overall duration of the disease [28, 29] . Besides, re-
lapsing forms of MMN showing acute deterioration, step-
wise progression, as well as spontaneous remissions have
occasionally been described [2, 13, 30] . Anecdotic reports
on subacute monophasic MMN presenting with tetrapa-
resis, preserved tendon reflexes and normal motor and
sensory nerve conduction following Campylobacter jeju-
ni infections [31–33] most likely reflected aberrant forms
of Guillain-Barré syndrome (GBS) [34] . Although the
prognosis ad vitam is favorable and only 2 fatal cases have
been directly ascribed to MMN after several years of dis-
ease [3, 27] , the majority of patients accumulate signifi-
cant disability as a result of severe paresis. Moreover,
pathological fatigue was only recently highlighted in
MMN [28, 35, 36] .
Pathophysiology of MMN
Molecular Basis of Conduction Block
The electrophysiological hallmarks of MMN are CBs
(see ‘Electrophysiological Findings’ below) which are
supposed to be the underlying cause of muscle weakness.
However, patients exist who present with clinical symp-
toms typical for MMN but in whom CBs cannot be de-
tected by routine neurography. Here, very proximal or
distal CBs inaccessible to standard neurography might be
present [16] . Interestingly, the majority of nerve-conduc-
tion studies in MMN demonstrated significant improve-
ment of CBs after treatment with IVIgs, although muscle
strength in these patients rarely recovers to normal [10] .
In general, CB appears when the incoming action cur-
rent at the node of Ranvier is unable to induce sufficient
depolarization at the subsequent node to generate an ac-
tion potential [37, 38] . Experimental paranodal demye-
lination in rodents severely impaired saltatoric nerve
conduction suggesting that focal demyelination is the
pathological basis of CB [23, 39] . This hypothesis was fur-
ther strengthened by morphological nerve studies con-
firming circumscribed demyelination in biopsies from
MMN patients [38–41] . Beyond focal demyelination,
generalized axonal dysfunction might be present in
MMN [42] . Pathological and electrophysiological find-
ings have highlighted the functional role of axonal disin-
tegration and impaired axon-myelin interactions [41, 43–
46] . The question whether axonal degeneration is an in-
trinsic pathophysiological feature of MMN or caused by
persistent CBs is still under debate. Recent studies point
out that activity-dependent processes can induce CB-in-
duced axonal degeneration. According to this, the axonal
membrane hyperpolarizes at the vicinity of a CB due to
altered K + conductivity [42] , but depolarizes at the site of
CB through inhibition of the Na + /K + -ATPase caused e.g.
by edema, reduced oxygen supply or immune-mediated
mechanisms such as binding of autoantibodies [38, 46]
( fig. 1 A, B). This scenario is further aggravated by disrup-
tion of the blood-nerve barrier leading to increased K +
concentrations in the endoneural fluid hence further

Multifocal Motor Neuropathy Eur Neurol 2010;63:193–204 195
supporting accumulation of intracellular Na + ions and
membrane depolarization. Persistent Na + influx can only
be counterbalanced by Na + ions moving intracellularly
along the axon to a site where the Na + /K + pump is still
functional. As a consequence, the Na + gradient from the
lesion site to the distal part of the nerve is decreased [46]
( fig. 1 B). Raised intracellular Na + concentrations are the
‘driving force’ for the Na + /K + -ATPase resulting in in-
creased pump activity and membrane hyperpolarization
[47–49] . According to this pathophysiological concept,
ongoing CB would trigger sustained elevation in intracel-
lular Na + . Under these conditions the activity of the Na + /
Ca 2+ exchanger (3 Na + /1 Ca 2+ ) can be reversed [44–46,
50] leading to intra-axonal accumulation of Ca 2+ and
subsequent axonal degeneration ( fig. 1 C).
Immunopathogenesis of MMN
There are several arguments for MMN being an im-
mune-mediated disease [5, 6, 25] : anti-GM1 antibodies
are found in 20–80% of patients suffering from MMN,
and GM1 is expressed on axons and the myelin sheath.
Interestingly, the molecular composition of GM1 differs
between sensory and motor nerves resulting in different
binding affinities of anti-GM1 antibodies hence, offering
a possible explanation for the selective impairment of
motor fibers in MMN [51–53] . Moreover, many MMN
patients respond to immunomodulatory treatment. Fi-
nally, highly specific therapies that selectively interfere
with the immune system such as TNF � antagonists (in-
fliximab) can induce MMN in rare cases [54] . Similar to
other neurological disorders associated with serum anti-
distal
CB
axon
Site distal to CB r hyperpolarization
Axonal degeneration
axon
axon
3 Na+
3 Na+
3 Na+
3 Na+
3 Na+
3 Na+
Na+
3 Na+
3 Na+
CB
Na+/K+-ATPase
Na+/K+-ATPase, blocked
Na+/K+-ATPase, enhanced
Na+ Ions
Na+/Ca2+ exchanger
Na+/Ca2+ exchanger, reversed
Na+-gradient
 Na+ gradient from CB to vicinity
 increased activity of the Na+/K+ pump
 hyperpolarization
2 K+
2 K+
2 K+
2 K+
2 K+
2 K+ 1 Ca2+
3 Na+1 Ca2+
1 Ca2+
Focal demyelination, CB r depolarization
 block of Na+/K+ pump (edema?, O2f?, GM1-Ab?)
 [Na+] i F
 depolarization
 CB
 long lasting block of Na+/K+ pump CB
 [Na+] i F
 reversal of the Na+/Ca2+ exchanger r [Ca2+] i F
 axonal degeneration
A
B
C
Fig. 1. Possible molecular mechanisms of CB and axonal degen-
eration in MMN. A Edema formation, restricted O 2 diffusion and/
or antibody-mediated processes following focal demyelination
can cause inhibition of Na + /K + pumps. This then leads to an in-
crease in intracellular Na + and, according to the electrogenic na-
ture of the Na + /K + -ATPase, to depolarization of the membrane
potential. Significant depolarization finally presents as CB. B Lo-
cally increased Na + concentrations at the site of CB form a Na +
gradient at the vicinity of CB. This scenario activates Na + /K +
pumps (increased ‘driving force’) accompanied by membrane hy-
perpolarization. C Persistent CB can saturate the activity of the
sodium-transporting machinery, hence reversing the function of
the Na + /Ca 2+ exchanger. As a consequence, Ca 2+ accumulates in
the axon and finally causes axonal degeneration.
Co
lo
r v
er
si
on
a
va
ila
bl
e
on
lin
e

Meuth/Kleinschnitz

Eur Neurol 2010;63:193–204196
bodies the question arises whether anti-GM1 antibodies
in MMN are pathologically relevant or represent a mere
epiphenomenon. Indeed, previous studies reported anti-
GM1-mediated focal demyelination and blockade of volt-
age-dependent Na + channels at the node of Ranvier in
vivo and in vitro [55–57] , but these findings could not be
confirmed by others [58–60] . Although human IgM anti-
GM1 antibodies can act on voltage-gated Ca 2+ channels
in vitro, the significance of this interaction for the patho-
genesis of MMN remains unclear [61] . The remarkable
proportion of anti-GM1 antibody-negative patients (up
to 50%) who, in comparison to anti-GM1-positive indi-
viduals, similarly respond to IVIg argues against an ex-
clusively antibody-mediated disease mechanism [12, 25,
62, 63] . Along these lines motor-nerve conduction in
mice could be blocked by human serum samples devoid
of anti-GM1 antibodies indicating that other soluble me-
diators are pathogenetically relevant in MMN [10, 64] .
Finally, IVIg treatment, although clinically effective,
does not reduce anti-GM1 titers [4, 65] .
Taken together, the exact immune mechanisms opera-
tive during MMN are still unknown and the available
data can currently not prove or disprove a causative
pathogenetic role of anti-GM1 antibodies [64] .
Diagnostics of MMN
Electrophysiological Findings
The most prominent electrophysiological features in
MMN are multifocal, persistent, partial CBs present in
motor but not sensory nerve fibers and located outside
the common entrapment sites [3, 12, 66, 67] (see ‘Molecu-
lar Basis of Conduction Block’ above). In general, CB has
been defined as the reduction of the amplitude or area
under the curve of the compound motor action potential
(CMAP) on proximal compared to distal nerve stimula-
tion. However, consensus on the required magnitude of
amplitude or area reduction that unambiguously defines
partial CB has not yet been reached. This is mainly due
to the fact that besides CB, several other mechanisms can
lead to significant CMAP reduction (‘pseudo CB’). Be-
cause axons of a chronically demyelinated nerve display
different conduction velocities (known as temporal dis-
persion, TD), the positive phase of the fast motor-unit ac-
tion potentials overlaps with the negative phase of slow
motor-unit action potentials (a phenomenon called ‘in-
terphase cancellation’), resulting in a disproportionate
proximal CMAP that can mimic true CB [68–71] . TD is
common in chronic demyelinating disorders of the pe-
ripheral nervous system, such as CIDP or polyneuropa-
thy associated with IgM gammopathy. In addition, col-
lateral nerve sprouting present for example in MND in-
creases the rate of polyphasia and reduces CMAP
amplitudes [72, 73] . Finally, technical limitations such as
insufficient supramaximal stimulation of the proximal
nerve segments can lead to pseudo-CB.
The degree of CMAP reduction necessary to define
partial CB in MMN varies considerably between differ-
ent studies, ranging from 20% to more than 50% with a
maximum admissible TD between 15 and 30% [7, 9] . A
computer simulation study in rats, in which compound
muscle-unit action potentials were reconstructed from
motor-unit action potentials, showed that maximum TD
can result in a decrement in the CMAP area of up to 50%
[70] . A recent retrospective investigation in humans for
the first time established simulation-based thresholds for
CB in the forearm segment of the median nerve [23] .
However, these thresholds still have to be evaluated for
other nerves and validated in a prospective manner. Even
in healthy individuals CMAP amplitudes are commonly
lower after proximal compared to distal nerve stimula-
tion with a range of reduction between 12 and 54% [74] .
Given the results from the above-referenced studies, the
commonly applied cutoff level of 50% CMAP decline
(amplitude or area) is the most validated electrophysio-
logical criterion of partial CB. Consequently, this thresh-
old was chosen for the definition of definite partial CB in
most of the peripheral nerves according to the consensus
criteria of the American Association of Electrodiagnos-
tic Medicine [18] and the European Federation of Neu-
rological Societies/Peripheral Nerve Society ( table 1 ).
However, the 50% limit should not be applied if CMAP
amplitudes are below 20% of the normal value. Then, po-
tentials are often too polyphasic to allow proper quanti-
fication. The relatively restrictive American und Euro-
pean electrophysiological criteria aim to avoid confusion
between real CB and TD, i.e. pseudo-CB. This approach,
however, may lead to the underdiagnosis of MMN which
represents a potentially treatable neuropathy [28, 67, 71,
75, 76] . Hence, it is important to bear in mind that a re-
duction in amplitude or area smaller than 50% might al-
ready represent partial CB, especially because CB, al-
though considered to be persistent [73, 77] , is a dynamic
entity that changes over time [75] , and a CMAP reduction
of 1 50% may be preceded by a smaller decrease underlin-
ing the need for electrophysiological evaluation at regular
intervals. Sometimes subtle focal CB can be detected us-
ing the so-called ‘inching technique’ where several nerve
sites with an interstimulation distance of 10–15 mm are

Multifocal Motor Neuropathy Eur Neurol 2010;63:193–204 197
stimulated sequentially [78, 79] . Here, an abrupt and cir-
cumscribed reduction in CMAP amplitude differentiates
focal CB from pseudo-CB which, in contrast, is charac-
terized by a more gradual CMAP decrease with increas-
ing stimulation distances. At the site of CB, conduction
velocity in motor but not sensory fibers is usually signif-
icantly reduced. Moreover, one should bear in mind that
the same nerve can be blocked at several sites [12, 67,
80] .
It is important to note that CB is not specific for MMN
but can also be found in several other neuropathies. In
contrast to MMN, CB occurring during acute compres-
sive neuropathy or hereditary neuropathy with liability
to pressure palsies are located at common anatomical en-
trapment sites like the ulnar sulcus or caput fibulae while
those present after ischemic nerve injury are transient
and usually reversible [81, 82] . Patients suffering from
GBS or CIDP in addition to CB regularly develop other
electrophysiological signs of severe demyelination like
markedly prolonged distal motor latencies or increased
F-wave latencies [83, 84] .
Another specific problem with the diagnosis of MMN
is the detection of proximally located CB. Due to ana-
tomical restrictions, the routinely used surface electrodes
are not able to stimulate proximal nerve segments (plex-
us, nerve roots). This technical limitation can be over-
come by the application of transcutaneous magnetic coils
or high voltage stimulators [85, 86] which, however, can
often not deliver the focal impulses necessary for the ex-
act calculation of nerve conduction velocities or are inap-
propriate for supramaximal fiber stimulation [87] . Al-
though F waves provide information on the integrity of a
peripheral nerve over its whole course and, at least theo-
retically, should be an ideal tool for the detection of (prox-
imal) CB, F-wave persistency depends on several other
factors such as axonal integrity and its reduction does not
necessarily indicate proximal CB [86] . Whether the re-
cently reported ‘magnetic fatigue test’ in which activity-
dependent CBs are unmasked by serial magnetic stimula-
tion [88] can indeed delineate between true CB and the
reduction of CMAP in MND needs to be further estab-
lished. Finally, a novel approach using a triple stimulation
technique to detect CB proximal to Erb’s point might help
to increase the diagnostic sensitivity in the future [89] .
Although CB clearly is an important hallmark of
MMN, the question whether its presence is mandatory
for the diagnosis of MMN is still under debate [90] . About
30 cases of MMN with typical clinical presentation and
a good response to IVIg but without CB have been re-
ported so far [4, 76, 91, 92] . In a recent retrospective anal-
ysis, patients with and without CB showed similar clini-
cal features and a comparable response to IVIg treatment
after a median follow-up of 7 years, suggesting related, if
not identical disease entities [93] . Final appraisal of the
existence of CB-negative MMN is hampered in that it is
not clear whether these subjects really never had CB or
whether CB merely disappeared over time due to second-
ary axonal degeneration and subsequent reduction also
of the distal CMAP amplitudes [22, 28, 94] .
Other electrophysiological hallmarks in MMN apart
from CB can comprise increased distal CMAP latencies
and prolonged or absent F-waves both of which mainly
result from mild demyelination [13, 67, 80, 95] . The clin-
ical finding of severely paretic muscles but with preserved
bulks and normal neurography from the corresponding
nerves is suspicious for distally located CBs. Those can
occasionally be confirmed by needle electromyography
(EMG) when an increased motor unit discharge rate ( 1 20
Hz) in the absence of spontaneous activity is found [74] .
Laboratory Findings
The most common laboratory findings in MMN are
IgM serum antibodies against the ganglioside GM1 [4]
which can be detected at high titers in 30–80% of the pa-
tients [7, 9] . The reported variations in the incidence
of GM1 antibodies are probably related to the different
ELISA assays used in the different studies as well as het-
erogeneous control populations [27, 96–99] . Besides
Table 1. Electrophysiological criteria of definite CB according to
the EFNS/PNS Joint Task Force [115] and the AAEM [87]
Nerve segment Amplitude
reductiona, %
Area
reduction, %
Median nerve
Elbow/wrist >50 >40 (50)
Axilla/elbow >50 >40 (50)
Ulnar nerve
Elbow/wrist >50 >40 (50)
Elbow prox./dist. >50 >40 (50)
Axilla/prox. elbow >50 >40 (50)
Peroneal nerve
Dist. fibula/ankle >60 >50 (50)
Fibula prox./dist. >50 >40 (50)
Tibial nerve
Knee/ankle >60 >50
Values in brackets according to the EFNS/PNS Joint Task
Force [115].
a Not included in the EFNS/PNS Joint Task Force [115].