Abstract
Article abstract-We performed detailed electrophysiologic studies on 16 patients with clinically defined multifocal motor neuropathy and found a wide spectrum of demyelinating features. Only five patients (31%) had conduction block in one or more nerves. However, in 15 patients (94%) at least one nerve showed other features of demyelination. We also noted a significant degree of superimposed axonal degeneration in 15 patients. Eight patients (50%) had individual nerves with pure axonal injury, despite the presence of demyelinating features in other nerves. Antiganglioside antibodies were elevated in four of five patients with conduction block and five of 11 patients without conduction block. We conclude that multifocal motor neuropathy is characterized electrophysiologically by a wide spectrum of axonal and demyelinating features. Diagnostic criteria requiring conduction block may lead to underdiagnosis of this potentially treatable neuropathy.
NEUROLOGY 1997;48: 700-707
Multifocal motor neuropathy (MMN) is a potentially treatable condition that can be distinguished from disorders of the lower motor neuron such as spinal muscular atrophy and amyotrophic lateral sclerosis (ALS). The absence of bulbar involvement and upper motor neuron signs, normal bulk in some very weak muscles, slow progression, and weakness that follows the distribution of individual nerves may distinguish MMN from these other conditions. [1-4] Conduction block (CB) in motor nerves is considered the electrodiagnostic hallmark of MMN. [5-10] Strict criteria for CB have been utilized to avoid misinterpretation of pseudoconduction block that may be a feature of other motor neuropathies such as ALS. [11] However, CB is only one of a variety of electrodiagnostic features that suggest the presence of segmental demyelination. These other features include focal slowing of conduction velocity (CV), temporal dispersion (TD), delayed F-wave responses, and prolonged distal latencies. [12] The objective of our study was to determine how often CB and other features of demyelination are present in patients who present with clinical features of MMN.
Methods.
Study population.
We retrospectively reviewed records from patients evaluated in our neuromuscular clinics for lower motor neuron weakness between January 1, 1993 and March 1, 1995. Patients were considered for the study if they were 20 years of age or older at onset and had weakness in the distribution of two or more peripheral motor nerves. To distinguish our patients from those with motor neuron disease, patients were excluded if they had facial weakness, bulbar weakness, respiratory difficulty, upper motor neuron signs (spasticity, hyperreflexia, extensor plantar response), or weakness that could not be localized to individual peripheral nerves. Patients were also excluded if there was a history of severe pain in the affected extremity at the time of onset, or if they no longer met the study criteria after a follow-up period of at least 1 year. Mild sensory symptoms were permitted as long as there were no sensory deficits on examination, and sensory nerve conduction studies (NCS) were normal.
Electrodiagnostic studies.
Motor and sensory NCS were performed on all patients. Studies included median, ulnar, peroneal, and tibial compound muscle action potentials (CMAP), and median, ulnar and sural sensory nerve action potentials. Radial motor and sensory, musculocutaneous motor, femoral motor, and superficial peroneal sensory studies were performed when there was weakness in the distribution of these nerves. Skin temperature was maintained at 32 degrees C. Motor NCS were recorded using surface electrodes as follows: the median nerve was stimulated at the wrist, elbow, axilla, and Erb's point, recording over the abductor pollicis brevis; the ulnar nerve was stimulated at the wrist, distal elbow, proximal elbow, axilla, and Erb's point, recording over the abductor digiti minimi; the radial nerve was stimulated at the proximal elbow, distal spiral groove, and Erb's point, recording over the extensor indicis; the peroneal nerve was stimulated at the ankle, fibular head, and popliteal fossa, recording over the extensor digitorum brevis; and the tibial nerve was stimulated at the ankle and popliteal fossa recording over the abductor hallucis. Measurements included onset latency, amplitude (baseline to negative peak), duration and area of the negative peaks, and conduction velocity for each segment. Cervical root stimulation was performed in four patients and lumbar root stimulation in one patient using monopolar electrodes.
Median and ulnar orthodromic sensory responses were obtained stimulating with ring electrodes placed on the second or third and fifth digits, respectively, and recording at the wrist. Sural responses were recorded at the lateral malleolus with stimulation of the posterolateral calf. If conduction slowing, TD, or CB were noted in the forearm segment of the median or ulnar nerve, appropriate sensory responses were recorded across the same segment to ensure that there was no electrophysiologic evidence of sensory involvement.
Supramaximal percutaneous stimuli were used for all NCS except root stimulation. Care was taken to avoid volume-conducted stimulation of neighboring nerves on distal stimulation, especially in cases where the median and ulnar nerves had large differences in amplitude. Erb's point studies were not accepted if the response was not reproducible or if a contralateral response could not be obtained.
We adapted criteria for demyelination and CB developed by the American Academy of Neurology task force [12,13] that were later modified for MMN. [5,11] CB was diagnosed when the CMAP amplitude and area decreased by more than 50% across a standard segment, and the CMAP duration increased by less than 30%. Possible conduction block (PCB) was diagnosed when the amplitude and area decreased at least 50% across a standard segment and the CMAP duration increased 30% or more. Abnormal TD was diagnosed when duration increased 30% or more without a 50% reduction of amplitude and area across a nerve segment. If CB, PCB, or TD were found across a standard segment, inching was used to localize the site of the lesion. [14] Slowed conduction in the demyelinating range was defined as a CV less than 70% of the lower limit of normal (LLN) in nerves with a distal CMAP amplitude less than 80% of the LLN, or as a CV less than 80% of the LLN when the amplitude was greater than 80% of the LLN. Minimum F-wave latencies and distal latencies greater than 120% of the upper limit of normal were considered in the demyelinating range. A pure axonal injury was defined as the following: (1) the distal motor amplitude was less than 80% of the LLN; (2) there was needle electromyographic evidence of fibrillations or neuropathic motor unit potentials in that distribution; and (3) there was no electrophysiologic evidence of demyelination in that nerve using the above criteria. A mixed axonal-demyelinating injury was present if the first two axonal criteria were met but demyelinating features were also present in that nerve.
Other investigations.
Patients had serum tested for antibodies to GM1 and asialo GM1 (Athena Diagnostics, formerly Genica Laboratories, Worcester, MA). Serologic tests were considered positive only if antibodies were present in high titers (anti-GM1 > 800, anti-asialo GM1 > 1,600).
Treatment.
Nine patients were treated with intravenous immunoglobulin (IVIG) with an induction dose of 2 g/kg given over 2 to 5 days followed monthly by doses of 0.4 g/kg (Table 1). In addition, two patients were treated with intravenous cyclophosphamide [9,15,16] at a dose of 3 g/m2 given in five divided doses on alternate days. Neither patient received follow-up doses of cyclophosphamide. One patient who received cyclophosphamide also received IVIG. Response was assessed by manual muscle testing of 34 muscle groups using a modified Medical Research Council grading system [17] and handgrip dynamometry at biweekly to monthly intervals following each infusion. A response was considered favorable when there was improvement of at least one grade in any muscle group, or if there was a 30% improvement in handgrip dynamometry.
Table 1. Clinical features*
Results.
Clinical features.
Sixteen patients met all clinical criteria for this study (see Table 1). There were 12 men and four women with an average age of 48.1 years. Average age at the time of onset was 40.5 years (range 20-64 years). The average duration of symptoms was 7.6 years (range 2-20 years). Twelve patients felt their symptoms had been static for a period of 1 to 9 years, while weakness was slowly progressive in the remaining four.
By chart review and history, 11 patients had weakness restricted to the distribution of one nerve at the time of onset and they did not develop multifocal involvement for an average of 3.0 years (range 2 months to 10 years). This involved muscles innervated by the ulnar nerve in four patients, the peroneal nerve in two patients, the radial nerve in two patients, and the axillary, median, and tibial nerve in one patient each. Weakness in four patients began in the distribution of more than one peripheral nerve, but involved only one extremity. Only one patient had involvement of more than one limb at onset. Four patients underwent surgical releases for presumed nerve entrapments before developing multifocal involvement.
All 16 patients developed weakness of intrinsic hand muscles. Eight patients also had lower extremity weakness. Overall, weakness was identified in the distribution of 17 ulnar nerves, 13 median nerves, 11 peroneal nerves, six tibial nerves, four radial nerves, three femoral nerves, and one musculocutaneous, axillary, and spinal accessory nerve each.
Electrophysiologic findings.
Fifteen of 16 patients had evidence of demyelination in at least one nerve. CB was found in five patients (Table 2, patients 1-5; Figure 1). Patients with CB had other nerves that demonstrated other features of demyelination. Patients 6 to 11 (see Table 2) had PCB across short nerve segments (Figure 2). Patients 12 and 13 had several nerves with abnormal TD and slowing but without abnormal area reduction (Figure 3). Patients 14 and 15 had no evidence of TD or abnormal amplitude and area reduction across individual segments but had slow CVs and prolonged F waves or distal latencies.
Table 2. Electrophysiologic features
Figure 1. Conduction block on motor nerve conduction study of the left median nerve (patient 5), stimulating at the wrist (A) and elbow (B). There is an amplitude reduction of 75% and an area reduction of 69%. The negative peak duration increases by 23%.
Figure 2. Possible conduction block on motor nerve conduction study of the ulnar nerve (patient 5), stimulating at the wrist (A), distal to the elbow (B), proximal to the elbow (C), the axilla (D) and Erb's point (E). There is a 90.4% amplitude reduction and an 83% area reduction between the axilla and Erb's point. However, the negative peak duration increases by 228%.
Figure 3. Temporal dispersion on motor nerve conduction study of the left median nerve (patient 12), stimulating at the wrist (A) and elbow (B). There is an amplitude reduction of 71%, but area reduction is only 27%, and the duration of the negative peak increases by 138%.
The only patient who did not meet demyelinating criteria (patient 16) was a 30-year-old woman with 10 years of hand weakness, in a median nerve distribution, that had remained stable for 5 years. She had a prolonged median F wave, normal median amplitude on distal stimulation, and a 31% drop in median CMAP area across the forearm without abnormal TD. None of these findings met our demyelinating criteria. She had elevated antibodies to GM1.
(Table 3) summarizes the NCS. PCB and TD were the most common segmental demyelinating abnormalities. PCB was seen in nine patients (56%). The degree of TD in these lesions ranged from 51% to 252%. TD without abnormal amplitude and area reduction, ranging from 36% to 361%, was found in seven patients (44%). A prolonged F-wave latency that was not explained by distal slowing was present in eight nerves, and a prolonged distal latency was found in nine nerves.
Table 3. Summary of abnormal findings
Motor amplitudes were reduced on distal stimulation in at least one nerve in 14 patients. Eight of these 14 patients had at least one individual nerve with pure axonal features. Individual nerves with mixed axonal and demyelinating lesions were noted in 13 patients. When all nerves were considered, no patient in this study had a pure axonal neuropathy. All eight patients with at least one nerve that met our criteria for a pure axonal lesion had other nerves with demyelination.
The ulnar nerve was affected most often on NCS. Segmental demyelination, defined here as CB, PCB, or TD, occurred in seven proximal, two intermediate, and eight distal ulnar sites (see Table 2). These changes were observed only across the forearm in the median nerve, between the elbow and the distal spiral groove in the radial nerve, and below the knee in the peroneal and tibial nerves. We did not find evidence of demyelination at common sites of entrapment with the exception of prolonged median motor distal latencies in three nerves and slowing of CV across the elbow segment of one ulnar nerve. On two other occasions, we noted abnormal TD across the elbow in the ulnar nerve. However, in both cases we localized the lesions across short segments several centimeters proximal to the elbow by inching. Root stimulation did not document CB or TD in any proximal nerve segment. CB, TD, and PCB occurred as a greater percentage of total lesions in the upper extremities, while prolonged F waves and pure axonal lesions were more common in the lower extremities (see Table 3, Seg/Total column).
Relationship of weakness to electrophysiologic findings.
Of the 17 ulnar nerves that innervated weak muscles, 14 had changes of demyelination on NCS. Similarly, nine of 13 median nerves that innervated weak muscles showed demyelination. Lower-extremity weakness was more commonly associated with axonal features. Weakness was present in the distribution of all five nerves with CB, in nine of 11 nerves with PCB, but in only five of 10 nerves with TD as the sole abnormality. In addition, there was weakness in the distribution to all nerves with prolonged distal latencies or F waves.
Serologic findings.
Nine patients had elevated titers of antigangloside antibodies (see Table 1). Anti-GM1 antibodies were present in three patients with CB and four patients without CB. IgM anti-GM1 antibodies were present in all seven patients with anti-GM1 activity. One of these patients also had IgG antibodies to GM1. Anti-asialo GM1 antibodies were present in four patients. Two patients had both anti-asialo GM1 and anti-GM1 antibodies. In total, four of the five patients with CB had antiganglioside antibodies. Of the 11 patients without CB, five had antiganglioside antibodies.
Response to therapy.
Ten patients were treated, eight with IVIG, one with intravenous cyclophosphamide, and one with both. Three of the five patients with CB received IVIG, and two had improvement. Of six patients without CB who received IVIG, four had a response to therapy (see Table 1). Five patients who received IVIG had improvement within the first month and one patient had improvement only after additional treatments. Three patients who received IVIG had no response to therapy. Two of the patients with CB received cyclophosphamide without improvement, but patient 3 had a later response to IVIG.
Discussion.
We found electrophysiologic evidence of demyelination in 94% of patients presenting with clinical features of MMN. However, only 31% of our patients had CB. Although we applied electrophysiologic criteria similar to that used by Chaudry et al., [5] our study differs since we defined entry criteria by clinical presentation and not by the presence of CB. Of note, each of the nine patients in their series had additional evidence of demyelination, including TD, prolonged F waves and distal latencies, and conduction slowing. [5] Recently, other authors have also reported a variety of demyelinating features in addition to CB in MMN. [8,10,15,18] Our study indicates that CB should not be considered a mandatory finding.
The low rate of CB in our patients may reflect the proposed strict criteria for the detection of CB in MMN. [5,11] The purpose of these criteria is to exclude patients with pseudoconduction block. [14] A reduction of CMAP amplitude may occur in motor neuron diseases from phase cancellations in the small number of surviving motor axons. Phase cancellation has produced up to 50% drops in area in computer models of peripheral nerve. [19] As a result, CB criteria for MMN required at least a 50% drop in amplitude and area with less than a 15% change in duration across a small segment of nerve. [5]
Adherence to these rigorous criteria, however, may result in underdiagnosis of treatable motor neuropathies. Earlier studies of demyelinating neuropathy used looser criteria for CB. Brown and Feasby [20] considered amplitude reductions of 20 to 30% significant in studies of Guillain-Barre syndrome (GBS). Published criteria for CB in GBS [12] and chronic inflammatory demyelinating polyneuropathy (CIDP) [13] have also adopted a 20% drop in amplitude and area. In addition, Oh et al. [21] reported that amplitude reductions exceeding 30% should be considered abnormal in the median, ulnar, and peroneal nerves. However, rather than loosening the criteria for CB, and potentially misdiagnosing demyelination in patients with pseudoconduction block, we wish to emphasize the variety of demyelinating features observed in patients with MMN. As our study indicates, other demyelinating findings are far more common than CB in MMN.
TD is an important electrodiagnostic finding in acquired demyelinating neuropathies. [22] Similar to the situation with CB, published criteria for TD have varied. Brown and Feasby [20] proposed abnormal TD be defined as greater than a 15% increase of the negative CMAP peak duration. This became the standard for demyelination criteria in GBS [12] and CIDP. [13] Chaudry et al. [5] used the same TD criteria in their MMN patients, even though they made the diagnosis of CB more stringent by increasing the limit of amplitude and area drop from 20% to 50%. However, Oh et al. [21] later demonstrated a 20 to 30% increase in the proximal CMAP duration in normal nerves, a phenomenon we have also observed. They recommended that pathologic TD should be diagnosed only if it exceeds these normal limits. [21] Lange et al., [11] Comi et al., [10] and Bouche et al. [8] each set TD criteria at 30% in MMN studies and this is the value we employed. The relative nature of these criteria is demonstrated by Figure 1. This waveform would not satisfy CB criteria in earlier MMN papers [4,5,9,23] because the proximal negative peak duration increased by 23% despite an amplitude and area reduction of more than 50%. If we had employed the 15% value for TD, two more of our patients would not meet CB criteria.
Both TD and CB are electrophysiologic correlates of demyelination in teased fiber studies of patients with demyelinating neuropathies. [24] In our series, abnormal TD was nearly always present at sites of segmental demyelination, while CB was relatively uncommon. This study and others have suggested that CB is more often associated with clinical weakness than TD or conduction velocity slowing. [5,20,24,25] Our patients always had weakness in the distribution of nerves with CB. In contrast, at times there was no weakness in the distribution of nerves with abnormal TD. This illustrates that TD and CB have different functional consequences. However, the electrophysiologic demonstration of abnormal segmental TD, with or without abnormal amplitude and area reduction, is in itself a valid indication of an acquired demyelinating neuropathy that is as important as CB in the diagnosis of MMN.
Criteria that recognize all demyelinating abnormalities increase the diagnostic sensitivity of NCS in MMN but risk a reduction in specificity. Our methods were based on electrodiagnostic criteria that were not intended for the diagnosis of demyelination in individual peripheral nerves. [12,13] Slowing of CV, prolonged F-waves, and in particular, prolonged distal latencies can occasionally be seen in individual nerves undergoing axonal degeneration. [26] However, only two patients had these abnormalities without PCB, CB, or TD in other nerves (see Table 2, patients 14 and 15). Confidence that demyelinating abnormalities of any type result from MMN and not from other conditions is increased when lesions are found distant from common sites of entrapment, over short segments, [14] or in nerves with normal distal CMAP amplitudes.
In addition, electrophysiologic evidence of axonal degeneration is a common finding in MMN. Our study demonstrates that a patient may have evidence of severe axonal degeneration in some nerves while demyelination predominates in other nerves. We found a pure axonal lesion in at least one nerve in 50% of our patients and a mixed axonal-demyelinating lesion in at least one nerve in 81%. However, no patient had a purely axonal neuropathy by electrophysiologic criteria. In a recent study, Bouche et al. [8] divided MMN patients into two groups: those that had atrophy in the distribution of involved nerves, suggesting axonal degeneration, and those that did not, implying a primarily demyelinating condition. They did not distinguish the groups electrophysiologically. In our experience, detailed electrophysiologic studies will show both demyelinating and axonal features in most patients.
The distribution of lesions in our patients is consistent with a process wherein focal injuries occur at random sites along the length of individual nerves. We found CB, PCB, and TD with greater frequency in the arms, while prolonged F waves and pure axonal lesions were more common to the legs (see Table 3). This pattern may reflect the arms having a greater proportion of the nerve accessible to stimulation, thereby facilitating the demonstration of segmental demyelination. Proximal nerve segments in the legs are less accessible to direct stimulation and, therefore, a prolonged or absent F wave or a low-amplitude CMAP may be the only evidence of proximal lesion. Perhaps we would have found more nerves with segmental demyelination by routinely performing root stimulation. [11] However, the root stimulation studies that we performed were not revealing, similar to other recent experience. [8]
Antiganglioside antibodies to GM1 are often present in patients with MMN, but they are not required for the diagnosis. [5,16] A number of studies have documented patients with MMN and CB who were seronegative. [4,5,8,16] There are other reports of high antibody titers to GM1 in patients without CB. [2,27] In our study, patients both with and without CB had high anti-GM1 titers. Overall, only 44% had elevated titers, consistent with prior studies showing that anti-GM1 antibodies are not highly sensitive for MMN. [8,28]
Our study is retrospective and we did not treat all of our patients. Some of our patients refused therapy because of the nonprogressive nature of their illness as well as the high cost, inconvenience, and potential side effects of the treatments. Nonetheless, patients both with and without CB had responses to IVIG. The response to immunotherapy in patients with distal lower motor neuron weakness without CB has been variable. [27,29] However, these patients only rarely had other demyelinating features on NCS. Prior studies suggest that weakness caused by CB is reversible with immunosuppressive therapy. [30] Improvement in strength has been accompanied by an increase in the proximal to distal CMAP amplitude ratio in affected nerves. [10,23,30] In contrast, weakness associated with significant muscle atrophy may not be reversible. [8] Improvement in patients without CB may still result from a reversal of the same pathophysiologic processes causing weakness as in those with CB, even when strict criteria are not met.
We conclude that definite CB is an infrequent finding in patients with clinical features of MMN. Our study suggests that MMN results from multifocal lesions that produce varying degrees of demyelination and axonal degeneration. The mechanisms of injury are still not known. Different factors including variations in the nature and intensity of the immune response, [31] the breakdown of the blood-nerve barrier, [32] impairment of sodium channel function, [32,33] and the age of the lesion [8] may contribute to the type of abnormality at each site. Appreciation of the spectrum of electrophysiologic features present in these patients should increase recognition of this potentially treatable neuropathy.
- Copyright 1997 by Advanstar Communications Inc.
REFERENCES
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.