Article Text


Axonal neuropathy associated with monoclonal gammopathy of undetermined significance
  1. Kenneth C Gorson,
  2. Allan H Ropper
  1. Neurology Service, St Elizabeth’s Medical Center, Tufts University School of Medicine, Boston, MA, USA
  1. Dr Kenneth C Gorson, Division of Neurology, St Elizabeth’s Medical Center, 736 Cambridge Street, Boston, MA 02135, USA.


OBJECTIVE The neuropathy associated with monoclonal gammopathy of undetermined significance (MGUS) is typically a predominantly demyelinating process that may have additional features of axonal degeneration. Sixteen patients with MGUS and a pure or predominantly axonal neuropathy are reported and compared with 20 consecutive patients with demyelinating neuropathy and MGUS who were seen during the same period.

METHODS Retrospective review of a consecutive series of patients with neuropathy and MGUS evaluated during a five year period.

RESULTS The axonal group had mild, symmetric, slowly progressive, predominantly sensory neuropathy, usually limited to the legs. There were no differences in the age of onset or duration of symptoms at the time of presentation, initial symptoms, or the severity of weakness between the axonal and demyelinating cases. However, the axonal process was associated with less vibration and proprioceptive loss, did not include leg ataxia (present in 55% of patients with demyelinating type), less often had generalised areflexia (19% v 70%), IgM gammopathy (19% v 80%), and anti-MAG antibodies (0% v 40%), and had lower CSF protein concentrations (mean, 49 v 100 mg/dl). The illness was also generally milder with less disability (mean Rankin score 2.1v 2.8). Fewer patients with axonal neuropathy improved with immunomodulating therapy (27% v 75%).

CONCLUSION There is an axonal neuropathy associated with MGUS that is clinically and electrophysiologically distinct from the more typical demyelinating pattern.

Statistics from

Neuropathies associated with monoclonal gammopathy are an important category of acquired chronic polyneuropathy, comprising about 10% of patients referred for evaluation of an idiopathic neuropathy.1 About one third of these patients have multiple myeloma, osteosclerotic myeloma, amyloidosis, lymphoma, or other lymphoproliferative disorder, and the remainder have a monoclonal gammopathy of undetermined significance (MGUS).2-4 Most studies of neuropathies associated with MGUS have focused on the immunoglobulin abnormalities, and several investigators have found clinical and pathological differences between IgM-MGUS and IgG/IgA-MGUS neuropathy.2 5-8 At least half of the patients with IgM-MGUS have anti-MAG antibodies,2 8 9and those with IgM-MGUS neuropathy may have more severe demyelination on nerve conduction studies than patients with IgG/IgA-MGUS, further separating the two types.5 6 There is considerable pathological evidence suggesting that the neuropathy associated with MGUS is primarily demyelinating,2 9 10 and most EMG studies have shown slowing of conduction velocities and other features of demyelination or changes consistent with both demyelination and axonal degeneration.2 3 5 6 10-13 However, a small group have a purely axonal neuropathy that shares similar clinical and electrodiagnostic features with the neuropathy associated with multiple myeloma.2-3 5-6 10 14-17 This heterogeneity of electrophysiological abnormalities may reflect different pathogenic mechanisms of the axonal neuropathy in patients with MGUS. We studied the axonal type of MGUS neuropathy to determine if there were differences in the clinical presentation, response to treatment, or course compared with patients with demyelinating neuropathies.



Clinical, laboratory, and electrophysiological information and nerve biopsies were collected uniformly but reviewed retrospectively for all patients with a monoclonal gammopathy of undetermined significance (MGUS) associated with neuropathy who were seen by the neuromuscular service at St Elizabeth’s Medical Center, Boston, from 1 January 1991 to 1 January 1996. We identified 36 consecutive patients who had no evidence of a haematological malignancy (multiple myeloma, osteosclerotic myeloma, Waldenstrom’s macroglobulinaemia, amyloidosis, or lymphoma) or other diseases that might have explained the polyneuropathy. The amount of the monoclonal protein was <3 g/l in all and remained stable with repeated testing. The patients were classified as having axonal neuropathy (A-MGUS) or a demyelinating neuropathy (D-MGUS) according to the electrophysiological pattern on EMG (see later).


These included complete blood count, routine chemistry, liver and renal function tests, total protein, B12 and folate concentrations, erythrocyte sedimentation rate, antinuclear antibody, thyroid function tests, rapid plasma reagin, and serum glucose. Other tests, such as Lyme serology, hepatitis B or C, SSA, SSB, or HIV titres, rheumatoid factor, angiotensin converting enzyme activity, and cryoglobulins were performed in some patients.

Immunoelectrophoresis or immunofixation electrophoresis with measurement of serum immunoglobulins and κ and λ light chains was obtained in all patients with a high resolution agarose gel technique according to the procedure described by Keren et al.18 Monospecific antisera to IgM, IgG, IgA, κ, and λ were used and immunoglobulins were quantified with nephelometry. Antibodies directed against myelin associated glycoprotein (anti-MAG) were measured using an enzyme linked immunosorbent assay (ELISA; Athena Diagnostics, Worcester, MA, USA). A lymphoproliferative disorder was excluded with skeletal bone survey, bone marrow biopsy, urinary protein electrophoresis, and chest and abdominal CT.


Nerve conduction studies were performed in a standard fashion as previously described.19 The median, ulnar, peroneal, and tibial motor nerves and median, ulnar, and sural sensory nerves were examined. Electromyographic studies were classified as axonal or demyelinating according to established criteria.11 The criteria for an axonal neuropathy associated with MGUS were essentially the converse of those used to establish demyelinating neuropathy: (1) reduced or absent motor or sensory amplitudes in at least two nerves; (2) conduction velocities >70% of the lower limit of normal in all nerves; (3) normal or minimally prolonged distal latencies in proportion to the conduction velocity; (4) normal or minimally prolonged F responses (<120% of the upper limit of normal); (5) no conduction block with proximal nerve stimulation; (6) needle electrode examination showing abnormal spontaneous activity (fibrillation potentials, positive sharp waves, or complex repetitive discharges) in distal muscles or neurogenic motor unit potentials with reduced recruitment. By contrast, all patients with demyelinating neuropathy and MGUS had at least three of the four electrodiagnostic criteria for demyelination established by the ad hoc subcommittee of the American Academy of Neurology20: (1) two nerves with slowing of conduction velocities <70% of the lower limit of normal; (2) two nerves with prolonged distal latencies >150% of the upper limit of normal; (3) two nerves with absent or prolonged F waves (minimum 10 trials) >150% of the upper limit of normal; (4) at least one nerve with conduction block defined by >20% drop in peak to peak amplitude between proximal and distal sites.


Twelve patients had sural nerve biopsies (six A-MGUS, six D-MGUS). One segment was sectioned at 1 μ and stained with toluidine blue. A second segment was embedded in paraffin, cut at 8 μ, and stained with haematoxylin and eosin and with Congo red.


Strength at the time of the maximum deficit and after recovery subsequent to treatment was assessed in a proximal and distal muscle of the arm and leg (deltoid, finger extensors, psoas, tibialis anterior) using the Medical Research Council (MRC) scale of 0–5. An average MRC score was determined from the four muscle groups. Vibration, joint position, and pin prick sensation were graded quantitatively as follows: 3=normal; 2=impaired at the fingertip or big toe; 1=impaired at the proximal interphalangeal joint or at the ankle; 0=impaired proximal to the proximal interphalangeal joint or above the ankle. The modified Rankin disability scale was used to determine functional impairment at the time of the maximal neurological deficit and after treatment: 0=asymptomatic; 1=non-disabling symptoms that do not interfere daily activities; 2=slight disability, unable to carry out all activities but still able to look after themselves; 3=moderate disability, requiring assistance with some activities but able to walk without assistance; 4=moderately severe disability, unable to walk without assistance and unable to attend to own bodily needs without assistance; 5=severe disability, totally dependent, requiring constant nursing care and attention.


All but four patients received treatment with either corticosteroids (about 60 mg/day for a minimum of two months), a trial of IVIg (0.4 g/kg/day for five days), or plasma exchange (15 litres total in five exchanges). Those who failed to improve with the initial treatment received an alternative second or third treatment at the discretion of the treating physician. A treatment response was defined as improvement of at least one grade of the Rankin score for at least two months.

We followed up most patients regularly during the course of their illness. For those who were followed up elsewhere, information was obtained by telephone follow up with the patient and local physician. The median duration of follow up was three years (range one to four years).


The clinical features, EMG abnormalities, and response to treatment in patients with A-MGUS and D-MGUS were compared using Fisher’s exact test to compare categorical responses and the Mann-Whitney independent rank sum test or the Kruskal-Wallis one way analysis of variance (ANOVA) to compare ordinal and continuous variables. A P value of 0.01 was used as the criterion for significance.



Tables 1 and 2 summarise the clinical details of patients with A-MGUS. There were 10 men and six women with a mean age of 70 (range 43–87) years. The average duration of symptoms at the time of our evaluation was 1.5 years (range two months-six years). Distal leg numbness, paraesthesiae, and imbalance were the most frequent complaints, although falling occurred in only three patients. Five complained of pain, usually a burning or aching sensation limited to the distal legs and soles of the feet. Weakness of foot dorsiflexion was mild in four, moderate in three, and severe in one (complete foot drop). Five had mild proximal leg weakness but could walk independently and two had more severe weakness and needed a walker. Hand weakness was slight in seven and severe in one. Vibratory sensation was diminished in the legs in all and impaired in the fingertips in slightly less than half. Sensation of joint position at the big toe was diminished in three. Pin prick sensation was abnormal in the feet in eight, diminished at the ankle in one, and impaired at the mid-calf or more proximally in four. One patient had severe sensory loss in the hands with pseudoathetosis and ataxia, preserved sensation in the legs, and normal strength. Achilles tendon reflexes were absent in all but two, half had absent knee and ankle reflexes, but only three had generalised areflexia. Three patients had a mild postural tremor of the arms and one had upper limb ataxia.

Table 1

Comparison of patients with axonal and demyelinating neuropathy associated with monoclonal gammopathy: symptoms

Table 2

Comparison of patients with axonal and demyelinating neuropathy associated with monoclonal gammopathy: signs

Most had no systemic illnesses. A haematological malignancy was excluded in all patients at the initial evaluation, but one developed a B cell lymphoma three years after the onset of the neuropathy. One had a chronic anaemia and another had temporal arteritis established by biopsy but no evidence of systemic vasculitis.

Twelve patients had an IgG monoclonal gammopathy (nine κ, three λ), three had IgM (two κ and one λ and κ), and one IgA (κ). Urinary protein electrophoresis was normal. None had positive anti-MAG antibodies. Protein in CSF was raised (>50 mg/dl) in two of seven patients. Table 3 provides a summary of the electrophysiological studies.

Table 3

Comparison of patients with axonal and demyelinating neuropathy associated with monoclonal gammopathy: nerve conduction studies

The nerve biopsies in six patients showed varying degrees of axonal degeneration with axonal clusters and loss of large myelinated fibres. In addition, one had a slight increase of thinly myelinated fibres and another had moderate endoneurial fibrosis. None had inflammatory cell infiltration or amyloid, and direct immunofluorescence staining for immunoglobulin deposition was negative in three.

Twelve patients were treated with immunomodulating therapy, with an average follow up of 3.3 (range 1-6) years. Two were treated with plasma exchange, two with IVIg, and one with steroids. Two patients had plasma exchange after steroids were ineffective, and two others received steroids after no improvement with plasma exchange. One patient unresponsive to IVIg was treated with plasma exchange, and two received all three modalities. Two responded to plasma exchange and one improved with steroids. One patient developed a relapsing course and is dependent on monthly cycles of plasma exchange. The patient with B cell lymphoma received cyclophosphamide and several cycles of adriamycin and vincristine without improvement in the neuropathy. Table 4 summarises the response to treatment and clinical course.

Table 4

Comparison of patients with axonal and demyelinating neuropathy associated with monoclonal gammopathy: response to treatment and outcome


There were 15 men and five women with a mean age of 64 (range 42–90) years. The mean duration of symptoms at our initial evaluation was 3.6 years (range 3 months-15 years) with an average follow up of four (range 1-12) years. Ten had symmetric generalised weakness, seven had weakness confined to the legs, and two had only hand and distal leg weakness. One had near normal power with a severe sensory ataxia. Virtually all had large fibre sensation loss in the distal limbs associated with symptoms of numbness (19 patients), paraesthesiae (15 patients), imbalance (17 patients), and falling episodes (13 patients). Seven complained of pain and 14 had generalised areflexia.

Sixteen had IgM (12 κ, four λ) and four had IgG gammopathy (one κ, three λ). Anti-MAG antibodies were raised in eight (range 1:1600–1:53 000) patients with IgM paraproteinaemia. There were no significant differences between the eight MAG positive and eight MAG negative patients in the quantitative neurological examination, mean electrophysiological studies, and number of patients responding to treatment. The mean protein concentration in CSF was 100 mg/dl. Table 3summarises the electrodiagnostic studies.


Tables 1-4 summarise the comparison of the clinical and electrophysiological features, treatment response, and course between patients with A-MGUS and those with D-MGUS. There were no differences in the age at onset, sex, duration of symptoms at initial presentation, and motor or sensory symptoms between the groups. Patients with A-MGUS less often had falling episodes. Patients with A-MGUS had less severe loss of vibration sensation in the hands and sensation of joint position in the feet, and leg ataxia and generalised areflexia were less often present. The mean Rankin disability score at initial evaluation was significantly better in the A-MGUS group.

Three patients with A-MGUS had IgM paraproteinaemia compared with 16 with IgM gammopathy in the D-MGUS group, and none of the patients with A-MGUS had anti-MAG antibodies. Two of seven patients with A-MGUS had raised concentrations of CSF protein compared with 11 of 12 patients with D-MGUS (P=0.01). The mean CSF protein concentration was also considerably lower in the A-MGUS group.

There were no differences between the groups in the proportion of patients responding to IVIg or corticosteroids, but fewer patients with A-MGUS responded to plasma exchange. Eleven of 15 patients with IgM in the D-MGUS group improved with plasma exchange. The mean MRC and Rankin scores were similar between the groups after therapy, but the change in Rankin score was greater in the D-MGUS group.

There was no difference in the age, sex, duration of symptoms, clinical features, nerve conduction studies, frequency of response to therapy, or course between patients with IgG/IgA and IgM-MGUS in the A-MGUS group. Similarly, there were no clinical, electrodiagnostic, or treatment response differences between patients with IgM-MGUS and IgG-MGUS in the D-MGUS group, except that a marginally greater proportion of patients with IgM were men (P=0.03).


Neuropathies associated with MGUS are usually demyelinating or mixed, with features of demyelination and axonal loss, and the axonal changes are usually attributed to severe or chronic demyelination. Pure axonal neuropathies in patients with MGUS are considered rare, comprising only a few cases in large series.1 3 6 10 11 14 15 21 However, a recent prospective study found 40% of patients with polyneuropathies that were associated with MGUS had electrodiagnostic features of a purely axonal process.13 Another study from the same investigators found that almost 20% of patients in their series with chronic axonal neuropathy had MGUS.22 Our experience was similar in that we found that 44% of a consecutive series with MGUS and neuropathy had primarily axonal disease.

These patients with A-MGUS had a relatively homogenous clinical presentation with modest sensory symptoms—distal leg numbness, paraesthesiae, and mild imbalance—with pain affecting one third of patients. Weakness or falling was uncommon. A few patients with A-MGUS had unusual presentations, with severe sensory loss in the upper limbs or moderate leg weakness with bilateral foot drop. Most had an unsteady gait but all were ambulatory and disability was usually mild.

Although the degree of weakness was similar in patients with A-MGUS and D-MGUS, there were several clinical features that distinguished the groups. Falling was more common in patients with D-MGUS, reflecting leg ataxia and impaired sensation of joint position. Patients with D-MGUS also had more severe loss of large fibre sensation and more often had generalised areflexia, typical of patients with chronic inflammatory demyelinating polyneuropathy.20 20a Accordingly, they had greater disability and a worse Rankin score compared with the A-MGUS group.

There was an association between the type of MGUS and the electrophysiological features of the neuropathy. IgM-MGUS and anti-MAG antibodies occurred often in patients with D-MGUS, confirming previous reports that patients with IgM neuropathies have more prominent demyelination on EMG studies5 6 8 11 12 In the three of our patients with A-MGUS and IgM gammopathy, none had anti-MAG antibodies. This supports the notion that anti-MAG antibodies may be relatively specific for demyelinating neuropathies.8 9 23

Patients with IgM-MGUS have been traditionally distinguished from those with IgG/IgA-MGUS, in part because of greater demyelination found on nerve conduction studies in the IgM group.2 5 6 In our patients with MGUS and either axonal or demyelinating neuropathy, there were no important clinical, EMG, or treatment features that separated patients with IgM and IgG-MGUS. Thus in addition to the nature of the immunoglobulin abnormality, the predominant electrodiagnostic pattern (axonal v demyelinating) may be helpful in differentiating the clinical features, course, and response to therapy in patients with MGUS associated neuropathies (see below).

Could the association between patients with axonal neuropathy and MGUS be chance? M proteins are present in up to 3% of the population over the age of 70, and the finding of an M protein in older patients with an “idiopathic” axonal neuropathy may be coincidental and not related to the neuropathy.24 25 None the less, there are several lines of evidence that suggest that the association between MGUS and axonal neuropathy is more than chance. In their study of patients with chronic idiopathic axonal neuropathy, Notermans et al concluded that there were no significant differences between patients with and without MGUS, but the data showed that MGUS patients had greater involvement of the arms, worse disability on the Rankin score, and a higher frequency of denervation compared with patients without MGUS.22 Sherman et al 26and others15 16 27-29 have shown antibodies to chrondroitin sulphate C in patients with IgM-MGUS and axonal neuropathy, and immunostaining of the sural nerve showed IgM deposits in the endoneurium.27 Others have found IgG binding to a low molecular weight neurofilament protein in axons.30 31Lastly, three of our patients with axonal neuropathy improved with immunomodulating therapy, and others have reported a similar experience in a few patients treated with prednisone and cyclophosphamide.32

Most patients with A-MGUS had a progressive course, and only one patient had a relapsing course responsive to plasma exchange. Only three of our patients with A-MGUS responded to immunotherapy, by contrast with 15 of the 20 patients with D-MGUS. There was no difference in the response rates with IVIg or steroids, but patients with D-MGUS had a significantly greater response rate to plasma exchange and a larger change in the mean Rankin score. Most of the D-MGUS group (75%) had IgM-MGUS, and most of these improved with plasma exchange, consistent with the experience of other investigators.20a33-36 The beneficial effects of plasma exchange in our patients with IgM need to be interpreted cautiously due to the limitations of a retrospective analysis; results of the small prospective trial by Dyck et al showed that their patients with IgM had a less favourable response to plasma exchange compared with patients with IgG or IgA-MGUS.37 Most of the patients with A-MGUS had only one series of five exchanges, a single course of IVIg, or a short trial of steroids. The lack of a therapeutic response may indicate that short term immunomodulating therapy may be inadequate in patients with MGUS with prominent axonal loss, but our numbers are too small to draw definitive conclusions.

In summary, there is a subset of patients with MGUS who have a mild, predominantly sensory axonal neuropathy that can be clinically and electrophysiologically distinguished from the demyelinating or mixed neuropathies that are more consistently associated with MGUS. The pattern of electrodiagnostic abnormalities, in addition to the nature of the M protein, may be helpful in defining the course and response to treatment. Some of these patients may improve with immune modulation.

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 20a.
  22. 21.
  23. 22.
  24. 23.
  25. 24.
  26. 25.
  27. 26.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
  34. 33.
  35. 34.
  36. 35.
  37. 36.
  38. 37.