Clinical manifestations

Presentation and course

Patients typically have an insidiously progressive, distal and symmetrical, sensory or sensorimotor neuropathy. The disorder usually afflicts men over the age of 60 years. Initial features include acral numbness, paresthesias, reduced proprioception, and Romberg sign. Sensory ataxia occurs in half of the cases and often is the most disabling symptom (15; 67; 68; 30). Neuropathic pain is rarely a presenting feature but develops in a minority of patients as the condition advances (30). Sensory disturbances invariably begin in the feet in a stocking distribution and frequently involve all modalities, but vibration sensation and proprioception are preferentially affected. Strength is usually normal or only mildly affected at the outset, although as many as 50% of patients develop substantial weakness (22; 92; 30). Deep tendon reflexes are absent or diminished (22). An intention tremor is common, may occur at any stage of the illness, and frequently becomes a disabling feature (67; 73; 15; 68). Cranial nerves and autonomic functions are rarely affected. A minority of patients have a fulminant, predominantly motor neuropathy with rapid progression, profound weakness, and fatal outcome (02; 92).

The classic sensory-ataxic form was observed in 20 of 34 patients with IgM gammopathy related neuropathy, whereas multiple mononeuropathy, polyneuropathy with predominant motor impairment, and painful small fiber neuropathy were identified in the remaining 14 cases (5, 8, and 1, respectively) (49).

In another study on phenotypic variability, 60 cases of neuropathy associated with MAG antibodies were described (53). Of those, 19 (32%) patients had the distal acquired demyelinating symmetrical (DADS) phenotype (36), with 11 of 19 having marked ataxia. Seventeen (28%) patients had isolated distal numbness or pain without ataxia. Thirty-two (53%) patients had sensorimotor neuropathy with variable degrees of motor weakness, leading to severe impairment in a minority of them. We think that eight out of these 32 cases overlapped with the DADS group. Weakness was usually distal, although two patients had prominent proximal weakness.

Most of the focus in MAG neuropathy reports has been on strength, sensation, and gait balance, but one study directed toward pain and cramps (78). In a cross-sectional study of 55 clinically stable patients with anti-MAG neuropathy, cramps were reported by more than 60% of patients and affected the legs more frequently than the arms. When involving the arms, cramps interfered with subject performance on measures of function. Frequent pain was reported by 80% of subjects and was independently associated with scores of the Inflammatory Rasch-built Overall Disability Scale. Hence, optimizing treatments of pain and cramps maybe a low-lying fruit to improve function and quality of life in anti-MAG neuropathy.

Prognosis and complications

Anti-MAG associated neuropathy is typically a slowly progressive disease. The condition frequently improves with therapy, but there is no cure (41; 84; 68; 30; 37). Several studies have shown that approximately 50% of patients improved after immune therapy, but the benefit was sustained in less than 20% after 5 years (30) and in only 5% after 10 years (68). It should be recognized that many patients have minor sensory symptoms that are not disabling and generally do not require therapy. In contrast, the majority have moderate functional disability (Modified Rankin disability score greater than or equal to 3) prior to therapy. Approximately 25% of patients worsen by one or more points on functional disability scales after 5 years and 60% worsen by 10 years (68; 30). Rarely, anti-MAG neuropathy can be fatal in patients who have a predominantly motor neuropathy (02).

Those with initial sensory symptoms may have slower progression and less disability, whereas prominent demyelinating features on the initial EMG study are predictive of weakness and upper limb involvement after 4 years (24). Interestingly, in several large cohorts the presence or the level of anti-MAG antibody titers had no prognostic value (92; 68; 24; 30). Some investigators have correlated reduced anti-MAG titers by 60% or more and response to treatment, but this observation has been limited to case reports, small series, and expert opinion (64; 06; 44; 37). However, others have suggested little or no relationship between the initial, follow-up, or absolute change of the anti-MAG antibody titer and response to therapy (68; 30). Therefore, the role of anti-MAG antibodies in predicting long-term prognosis or treatment response remains uncertain.

Furthermore, several investigators have indicated that lowering the IgM level by 50% or more also may be associated with a response to therapy (67; 84; 06; 44), but this may be difficult to achieve and sustain; reducing the IgM level by 25% was associated with a clinical response in one study (30).

The risk of progression from monoclonal gammopathy of undetermined significance to a malignant plasma cell dyscrasia, usually multiple myeloma, is approximately 1% per year (39). Data indicate that the risk may be as high as 2.7/100 patient years in polyneuropathy with monoclonal gammopathy of undetermined significance (23). Unexplained weight loss, progression of polyneuropathy, unexplained fever or night sweats, and M-protein level are independent predictors of malignant transformation. Accordingly, long-term hematologic follow-up is essential for all patients with monoclonal gammopathy of undetermined significance, irrespective of the presence of anti-MAG antibodies. Approximately 20% of patients with anti-MAG neuropathy developed a malignant plasma cell disorder after 10 years (68). In contrast, two of 38 patients with IgM-monoclonal gammopathy of undetermined significance neuropathy developed a malignant disorder, 17 and 25 years after onset of the neuropathy, respectively; it was not reported if these patients had anti-MAG antibodies (76).

Asymptomatic patients with IgM monoclonal protein were found to have a higher incidence of subclinical neuropathy if anti-MAG antibodies were present. The detection of anti-MAG antibodies in asymptomatic patients with an IgM monoclonal gammopathy was also found to predict development of neuropathy (60).

In a large prospective cohort from the Netherlands, 140 patients with IgM monoclonal gammopathy and peripheral neuropathy were followed up for up to 23 years (62). The authors identified three independent risk factors that influenced disability as measured by a modified Rankin score of three (requiring some help but able to walk without assistance) or worse. Higher age at onset and demyelinating electrophysiology (as in 72% of cases) increased the likelihood of disability. Interestingly, the presence of anti-MAG antibodies (as in 44% of cases) was associated with a lower disability risk. Overall, the disability risk seemed to plateau between 10 and 15 years from onset, especially in MAG negative patients. The authors offer a web-based prognostic model of disability over time-based on these factors.

In an Italian study, several outcome measures were assessed prospectively for more than a year and half in 32 subjects with anti-MAG neuropathy. Only two measures, the Sensory Modality Sum score at four limbs and the Visual Analogue Scale for pain in the upper limbs, demonstrated significant worsening over the mean follow up time of 18.6 months (12). It has been reported that anti-phosphacan titer to anti-MAG titer (P/M ratio) correlated with disease course (58). P/M ratio has been increased in patients with progressive clinical course. Furthermore, the P/M ratios were significantly increased in worsened cases as compared to stable or improved cases.

Clinical vignette

A 64-year-old man had a 9-year history of slowly progressive distal leg sensory loss and gait disorder. Symptoms began insidiously with paresthesias in the toes and feet, followed by progressive distal leg numbness and imbalance; he noted that at times he would "stagger like a drunk," and began using a cane 3 years prior. There was mild weakness (MRC 4/5) and atrophy of the toe and foot extensors. Deep tendon reflexes were absent and plantar responses were flexor. Vibration sensation was absent in the legs and first detected at the knees, and was reduced in the fingertips. Joint position was absent at the great toe and Romberg sign was present. Pin prick sensation was mildly impaired in the feet. There was a high frequency tremor of the outstretched hands that worsened with finger-to-nose testing. Gait was wide-based and ataxic.

Electrophysiological studies showed prolonged distal motor latencies in the median, tibial, and peroneal nerves, severe slowing of motor conduction velocities, and absent sural and median sensory potentials, consistent with a generalized, symmetric, demyelinating sensorimotor polyneuropathy. There was no conduction block. The cerebrospinal fluid was acellular with a protein level of 121 mg/dL. Serum immunofixation showed an IgMk monoclonal protein with an IgM level of 425 mg/dL (normal, 50 to 320 mg/dL). Anti-MAG antibody was elevated with a titer of 1:102,400. Serum viscosity, B-2 microglobulin, and urine immunofixation were normal, and cryoglobulins were absent. Bone scan, skeletal bone survey, and bone marrow and fat pad biopsy were normal. Sural nerve biopsy showed reduced axonal density with numerous thinly myelinated nerve fibers. Electron microscopy showed widely spaced myelin, and immunofluorescent stains demonstrated deposits of IgM and complement on myelin sheaths.

He was treated with intravenous rituximab (375 mg/m2 weekly for 4 weeks). Gait, balance, and leg weakness improved, and Romberg sign was absent after 4 months. The serum IgM level decreased to 190 mg/dL and the anti-MAG antibody titer was 1:51,200.

Biological basis

Etiology and pathogenesis

The etiology of the neuropathy has not been fully elucidated; many investigators believe the neuropathy is directly related to the presence of elevated anti-MAG antibodies associated with an IgM monoclonal protein.

MAG is a minor constituent of myelin, and is concentrated at the periaxonal region and other compacted regions of peripheral nervous syndrome myelin including Schmidt-Lanterman incisures, lateral loops, and outer mesaxon. MAG has been postulated to play a critical role in the myelination of peripheral nerve fibers, probably by acting as an adhesion molecule that modulates neurofilament spacing, axon caliber, and the formation and maintenance of myelin structure (45; 11; 50). The antibodies bind to an oligosaccharide determinant that is shared by MAG and the glycolipid sulfoglucuronyl paragloboside. Antibody binding to MAG and other HNK-1 bearing glycoproteins and glycolipids may disrupt their function.

The presence of anti-MAG antibodies has been associated with a chronic demyelinating peripheral neuropathy (41; 65; 92). Anti-MAG antibodies were first demonstrated to bind to peripheral nerve myelin by immunochemistry and complement fixation assays (42). The antibodies also bound to myelin associated glycoprotein by the Western blot technique, and to the sulfated glucuronic-acid containing glycolipids, sulfoglucuronyl paragloboside and sulfonyl glucuronyl lactosaminyl paragloboside, by immunostaining on thin layer chromatography. The epitope that is recognized by the human antibodies is shared by the peripheral myelin proteins P0 and P22, but the antibodies bind to these proteins less avidly than to MAG (85; 98). MAG is also expressed in the perisynaptic Schwann cells, which unlike myelinating Schwann cells, do not form multiple myelin wrappings around nerve terminals. In MAG-deficient mice maintenance of myelin is impaired (83).

In approximately 85% to 90% of cases, the anti-MAG antibodies occur with an IgM monoclonal protein, and the light chain is usually kappa (75; 22; 15; 30). Elevated anti-MAG antibody titers are found rarely in the absence of a monoclonal gammopathy, or the M-spike is detected later (66). This monoclonal protein, commonly termed monoclonal gammopathy of undetermined significance or MGUS, is not associated with a malignant plasma cell dyscrasia (multiple myeloma, osteosclerotic myeloma or the POEMS syndrome, systemic amyloidosis) in the majority of patients (41; 82; 27). However, approximately 25% of patients with an IgM monoclonal protein and anti-MAG antibodies have Waldenstrom macroglobulinemia or rarely, B-cell lymphoma or chronic lymphocytic leukemia (41).

There is compelling pathologic evidence that anti-MAG antibodies have a causative role in the pathogenesis of the neuropathy. Intraneural injection of serum from patients with demyelinating neuropathy and anti-MAG antibodies has induced nerve demyelination in animal models (32; 99; 90). In chickens, systemic administration of serum from patients with anti-MAG antibodies has produced a demyelinating neuropathy with separation of the myelin lamella at the minor dense line, similar to pathologic findings in humans (88).

Studies of nerve biopsy specimens have shown loss of myelinated fibers, thinned myelin sheaths, segmental demyelination, and occasionally tomacula and onion bulbs (94; 81). Two pathologic features are characteristic of the neuropathy associated with anti-MAG antibodies: (1) the presence of deposits of the monoclonal IgM and complement on myelin sheaths (87; 33; 94; 61; 81), and (2) the widening of the myelin lamella at the minor dense line with IgM deposits in the areas of the widened myelin loops (59; 94; 40; 50). Widening or irregular uncompacting of the myelin lamella has been shown to occur more frequently and to a greater extent when anti-MAG IgM antibodies were associated with Waldenstrom macroglobulinemia compared to monoclonal gammopathy of undetermined significance (96; 95).

Epidemiology

Although published population-based studies on MAG neuropathy are scant, the originally estimated incidence of anti-myelin-associated glycoprotein neuropathy is between 1 and 5 per 10,000 in the adult population (41). Men are more frequently affected than are women. Data, however, suggest a much lower frequency of MAG antibodies (72). MAG seropositivity was present in 5.8% of all chronic inflammatory demyelinating polyradiculoneuropathy cases without IgM monoclonal gammopathy; 2.9% had monoclonal IgM antibodies and were MAG seronegative.

Differential diagnosis

Confusing conditions

Anti-MAG associated neuropathy shares common features with numerous inherited and acquired neuropathies. The initial features of slowly progressive, symmetric, distal sensory loss and paresthesias may be clinically indistinguishable from patients with numerous conditions associated with sensory neuropathy, including diabetes mellitus, vitamin deficiencies, alcohol abuse, medication- and toxin-induced neuropathy, neuropathies associated with kidney, liver, and thyroid disease, Sjögren syndrome and other connective tissue disorders, and malignancy. Most of these latter conditions can be excluded by history and appropriate laboratory studies. In most cases, serum immunofixation readily identifies patients with an IgM monoclonal protein and directs further evaluation (ie, testing for anti-MAG antibodies).

The hereditary demyelinating neuropathies, Charcot-Marie-Tooth disease Type 1a, 1b, the x-linked form, and hereditary liability to pressure palsies, usually can be distinguished from patients with anti-MAG neuropathy by the relatively early age of onset, positive family history, motor predominant pattern, and other characteristic neurologic signs such as enlarged nerves, high arched feet, hammer toes, stork legs, and the lack of a monoclonal protein in the serum. Furthermore, nerve conduction studies in the hereditary demyelinating neuropathies typically show homogenous slowing of nerve conductions without conduction block or temporal dispersion (89). Genetic studies are now widely commercially available, and the detection of mutations of the myelin proteins P0, PMP22, or connexin confirm the diagnosis of hereditary demyelinating neuropathy in most cases. With the wide availability of genetic testing, conduction block and temporal dispersion have been reported in hereditary demyelinating neuropathies.

Other acquired demyelinating neuropathies can be confused with the anti-MAG associated neuropathy. Chronic inflammatory demyelinating polyneuropathy usually involves motor and sensory nerves, and patients often have generalized, proximal, and distal limb weakness. The tempo is subacute with progression over months, rather than the slower evolution over years that is characteristic of patients with anti-MAG associated neuropathy. The cerebrospinal fluid protein is elevated in both conditions and is not useful to distinguish the disorders. Electrodiagnostic studies of patients with chronic inflammatory demyelinating polyneuropathy show widespread demyelination in proximal, midsegment, and distal motor nerves, and conduction block is common (28). In contrast, there is disproportionate demyelination in the distal motor nerve segments in patients with anti-MAG associated neuropathy (35; 16; 13). Some investigators have stressed the importance of the distal predominance in anti-MAG neuropathy to distinguish these cases from chronic inflammatory demyelinating polyneuropathy (36).

Neuropathies associated with antisulfatide or anti-GD1b IgM antibodies also may have features of a large-fiber sensory loss and ataxia associated with a demyelinating pattern on electrodiagnostic studies (93; 67).

Motor neuropathies are easily distinguished from patients with anti-MAG associated neuropathy. Patients have progressive weakness, wasting, and fasciculations, with few or no sensory symptoms or findings. Anti-GM1 antibody has been reported to be elevated in up to half of patients with multifocal motor neuropathy. The clinical picture sometimes resembles motor neuron disease, but nerve conduction studies show multifocal demyelination and conduction block limited to motor nerves.

Diagnostic workup

Serum immunoelectrophoresis and immunofixation are necessary to confirm the presence of a monoclonal protein. In patients with a neuropathy and an IgM monoclonal protein, anti-MAG antibodies should be measured. The Western blot system, using myelin or MAG as antigen, is the most reliable method for detecting anti-MAG antibodies. Antibody titers are usually measured by the ELISA system, using MAG or sulfoglucuronyl paragloboside as antigen. Most normal individuals have absent or low antibody titers, but higher titers (usually greater than 1:6400) are closely associated with the presence of neuropathy. In the MAG ELISA system, however, impurities in the antigen preparation were thought to produce false positive results, and the ELISA assay can miss patients who are positive by Western blot (92). A new ELISA technique was reported to be more sensitive than Western blot, with good specificity (38). In a study of the optimal MAG antibody titer using the commercial ELISA method, the combination of sensitivity/specificity was found at a threshold > 1500 Bühlmann Titer Units (BTU) rather than the manufacturer recommended greater than 1000. Specificity was best at > 7000 BTU with 92.5% specificity (46).

Based on research studies, more specific and highly sensitive novel antibodies are suggested. In a French report consisting of 41 anti-MAG neuropathies and 118 controls, 40 out of 41 anti-MAG neuropathies had antibodies recognizing a specific epitope called human natural killer 1 (HNK1) shared by NK lymphocytes and several components of the peripheral nerve myelin, whereas only one out of 118 controls had that antibody (20). In anti-MAG neuropathy cases, titers of anti-HNK1 correlated with the INCAT sensory sum score, the Rasch-built Overall Disability Scale, and Overall Neuropathy Limitation Scale. In an Italian study, sera from 40 anti-MAG EIA positive neuropathy patients were tested for anti-MAG by Western blot (WB), for antiperipheral nerve myelin (PNM) on monkey nerve by immunofluorescence assay, and for anti-HNK1 on rat CNS slices by immunofluorescence assay. Antisulfatide antibodies, for comparison, were also tested by EIA (57). Three subgroups of patients were identified. Anti-HNK1 positive patients had the classical DADS phenotype, whereas patients without anti-PNM or anti-HNK1 IFA reactivity had a CIDP-like polyneuropathy. The third group of patients consisted of those who were anti-PNM positive and anti-HNK1 negative, had predominantly axonal neuropathy with a high frequency of antisulfatide antibodies, and had worst prognosis.

It is important to test for Waldenstrom macroglobulinemia because most anti-MAG neuropathy cases also have an IgM kappa monoclonal gammopathy. This is done through consultation with a hematologist/oncologist to evaluate patients for Waldenstrom and the rare association with B-cell lymphoma or chronic lymphocytic leukemia. The detection of a monoclonal protein requires further evaluation for a malignant plasma cell disorder. An M-protein greater than 3.0 g/dL suggests a plasma cell dyscrasia, whereas an amount less than 1.5 g/dL is characteristic of patients who have monoclonal gammopathy of undetermined significance. The serum also should be assessed for cryoglobulins, and serum viscosity should be measured when more than 3.0 g/dL of IgM protein is detected. A bone scan and bone marrow biopsy are usually indicated for patients who have a large amount (3.0 g or more) of monoclonal protein and systemic features suggesting multiple myeloma. A skeletal bone survey also is required in most cases to exclude osteosclerotic myeloma and POEMS syndrome, although this condition is more commonly associated with IgG, IgA, or lambda monoclonal proteins. In patients suspected of having lymphoma or chronic lymphocytic leukemia, chest and abdominal CT scanning and bone marrow biopsy should be performed.

Nerve conduction studies show polyneuropathy with some demyelinating features in over 90% of patients (especially those with anti-MAG titers greater than 5000), although most have evidence of axonal degeneration as well. It has a distinctive pattern with disproportionate prolongation of distal motor latencies in up to 75% of patients, indicating a length-dependent, distal demyelinating process, as previously noted (35; 91; 69; 16; 30; 13). This feature is perhaps best demonstrated by an abnormally short terminal latency index (distal conduction distance [mm]/ (conduction velocity [m/s] x distal motor latency [ms])) (35; 16). The terminal latency index is also useful to distinguish anti-MAG neuropathy from other acquired demyelinating neuropathies (35; 16), although others have suggested that a prolonged distal motor latency (greater than 7.0 msec) or residual latency (greater than 4.0 msec) may be more sensitive (77). In a comparative study, all patients with a terminal latency index of less than 0.26 had MAG neuropathy and all patients with terminal latency index ≥ 0.32 had HMSN1 (52). In the remaining patients with intermediate terminal latency index values, an ulnar distal motor latency value of < 7.4 (at 6 cm distance) had an overall sensitivity of 100% and specificity of 98%.

Other demyelinating abnormalities are also common in anti-MAG neuropathy, including slowed conduction velocities and prolonged or absent F-responses. The finding of conduction block is more variable; most studies have indicated that conduction block is absent or distinctly uncommon (35; 22; 76; 13), but others have demonstrated that it may be observed with some frequency (69; 30). Disproportionate distal slowing with normal distal CMAP duration in the arms may be useful to differentiate chronic inflammatory demyelinating polyneuropathy from MAG neuropathy (26). The lack of temporal dispersion at the distal site suggests a uniform involvement of nerve fibers. One study has suggested a correlation between the anti-MAG titer and the degree of conduction slowing and prolongation of the distal motor latency (91). Others have demonstrated that the relative accentuation of distal motor latencies was no longer evident on serial nerve conduction studies after prolonged follow-up, indicating a centripetal progression of the demyelination or secondary axonal degeneration, leading to normalization of the terminal latency index (76). However, another study suggested stability of median and ulnar motor responses over a mean duration of 6.5 years (07). Lower extremity NCS responses and arm sensory responses declined in amplitude over time.

On nerve conduction studies in 56 cases of neuropathy associated with MAG antibodies, heterogeneity was manifest as two out of three cases had typical anti-MAG pattern (53). The later consisted of a grossly symmetric demyelinating neuropathy with distally predominant demyelination based on low terminal latency index in at least one nerve, a severe decrease of sensory nerve action potential (SNAP) amplitudes in the lower limbs, and no conduction block or temporal dispersion. Thirty percent of patients had nerve conduction studies consistent with chronic inflammatory demyelinating polyneuropathy (CIDP) based on the European Federation of Neurological Societies (EFNS) 2010 criteria; most had conduction block, or marked temporal dispersion, or both, and some had normal or near normal SNAPs in the lower limbs. Though phenotypically 17 patients had clinically isolated distal numbness or pain without ataxia, only two of 56 cases (4%) had electrophysiological evidence of sensory axonal neuropathy.

As stated earlier, the classic sensory-ataxic form was observed in 20 of 34 patients with IgM gammopathy, whereas multiple mononeuropathy (5), polyneuropathy with predominant motor impairment (8) and painful small fiber neuropathy (1) was identified in the remaining 14 cases (49). Nerve conduction studies revealed a pattern of definite demyelination in 30 of 34 patients and possible demyelination in two cases, and one case each had an axonal neuropathy and small fiber neuropathy pattern. Reduced terminal latency index, prolonged distal motor latency and reduced motor conduction velocity were significantly associated with the classic sensory-ataxic phenotype, whereas reduced compound muscle action potential amplitude of the peroneal nerve from the tibialis anterior was mostly associated with atypical forms.

The indications for nerve biopsy in clinical practice have not been defined. In the majority of patients with anti-MAG neuropathy, the diagnosis can be established by the clinical features, presence of an elevated anti-MAG antibody titer and IgM monoclonal protein, and appropriate electrodiagnostic studies; nerve biopsy is usually not necessary. However, nerve biopsy may provide important information in cases where amyloidosis or even vasculitis are considered to be part of the differential diagnosis. Pathologic findings include a decrease in the number of myelinated fibers, thin myelin sheaths, and segmental demyelination. Deposits of monoclonal IgM and complement on myelin sheath as well as widening of myelin lamella are characteristic features (87; 33; 61; 40). In some patients, however, findings of axonal degeneration may predominate in both electrophysiological and pathologic studies (22; 92).

Investigators have demonstrated that distal limb skin biopsy can be helpful in the diagnosis of MAG neuropathy (47). The authors detected diffuse deposits of IgM antibodies in subpapillary nerve plexus of dermis in hairy and in glabrous skin containing several MAG- or MBP-positive small nerve fibers. This was not present in chronic inflammatory demyelinating polyradiculoneuropathy or IgM paraproteinemic neuropathy controls. However, another study showed less promising results. Examination of skin biopsies for immunoglobulin deposition does not add significant diagnostic value in the evaluation of neuropathies suspected to be caused by monoclonal gammopathy of undetermined significance (MGUS) or Waldenström macroglobulinemia (WM) (01). Of the 35 patients with MGUS or WM and peripheral neuropathy, four had immunoglobulin deposition in the skin biopsy, all of whom had an IgM gammopathy. In the control group of 19 MGUS cases without peripheral neuropathy, three had immunoglobulin deposition in the skin biopsy, all of whom had IgM-MGUS. Hence, IgM immunoglobulin deposition in skin biopsy might merely be an epiphenomenon secondary to high IgM blood levels.

Nerve ultrasound is an emerging technology in neuromuscular disease. On ultrasound, the frequency of nerves with enlarged cross-sectional area is referred to as the enlarged nerves sum score (48). It was found to be the most useful ultrasound parameter in differentiating three groups of patients with demyelinating neuropathies. The enlarged nerves sum score in anti-MAG neuropathy patients was greater than in MAG-negative paraproteinemic neuropathies and lower than in CIDP, and it was, therefore, the most useful ultrasound parameter in differentiating these three groups of patients with demyelinating neuropathies. Enlargement of cervical nerve roots and, at common entrapment sites, of peripheral nerves does not distinguish anti-MAG neuropathy from typical CIDP (70). However, the degree nerve enlargement at entrapment sites as compared to non-entrapment sites is significantly higher in anti-MAG neuropathy than in typical CIDP.

Management

Supportive measures and rehabilitation are discussed in another MedLink article: peripheral neuropathies.

No therapy has been definitively proven to be effective in patients with anti-MAG neuropathy. There have been seven published, randomized clinical trials for the treatment of this condition (71; 18; 55; 54; 17; 19; 43). Some studies indicated that intravenous immune globulin was effective in 15% to 20% of patients (18; 68; 30). However, a prospective study indicated that IVIG induced a short-term benefit in 50% of patients with IgM paraprotein-associated neuropathies (11 of 19 patients had elevated anti-MAG antibodies) (17). In a metaanalysis of the Cochrane database, IVIG was found to be relatively safe and may produce modest short-term improvement in some patients (51). Part of the difficulty may be that measuring gait impairment in MAG antibody-associated neuropathy is a challenge given the subjectivity in clinical gait ataxia assessment. In a case report using captured by GAITRite, a computerized walkway with embedded pressure sensors, the spatiotemporal gait parameters demonstrated an improvement in response to intravenous immunoglobulin but not to rituximab (97). Further studies assessing this instrument’s reliability, validity, and degree of variability in MAG antibody-associated neuropathy are needed to determine future use in research studies and in the clinical arena.

Retrospective studies have indicated that plasma exchange may be beneficial in 30% to 50% of cases (31; 22; 68; 30), but improvement rarely is sustained and most patients required periodic exchanges to maintain stability. One randomized, unblinded trial showed that combined therapy with plasma exchange and chlorambucil was no better than chlorambucil alone (71). Corticosteroids alone are ineffective; however, their combination with other therapies might exert additional efficacy (68). An unblinded study initially indicated interferon-alpha 2a was helpful (55), but this was not substantiated in a randomized, blinded, controlled trial (54). Others have suggested that treatment with oral or intravenous pulse cyclophosphamide for 6 months (68; 30; 37), chlorambucil (64; 67; 68), fludarabine (84; 100), or combined treatment with chemotherapy and plasma exchange (41; 64; 06; 22; 44; 100; 68; 30; 37) have been effective in individual patients.

Several reports have indicated the potential benefits of rituximab, a monoclonal antibody to the CD20 antigen on B-lymphocytes (44; 74; 80). Rituximab was well-tolerated, and patients apparently improved within 3 to 12 months after therapy (44; 80). Raising the rituximab dosage from 375 to 750 mg/m2 clinically and electrophysiologically improved four out of eight anti-MAG neuropathy cases (79). An efficacy and safety open-label, uncontrolled trial of rituximab in 13 patients with polyneuropathy associated with antibodies to MAG suggests that rituximab may be particularly efficacious on sensory impairment in patients with moderately elevated antibody titers (04). A randomized, controlled trial of rituximab in 26 MAG neuropathy patients showed a trend toward improvement of the INCAT score in four out of 13 rituximab-treated cases, whereas none of the placebo recipients got better (19).

A rituximab randomized, double-blind, placebo-controlled trial enrolled 54 patients with IgM anti-MAG demyelinating neuropathy and inflammatory neuropathy cause and treatment sensory score (ISS) of 4 or more, and either visual analog pain scale greater than 4 or ataxia score of 2 or more (43). In this study comparing rituximab to placebo, the primary outcome was mean change in ISS at 12 months. Although there were improvements in secondary outcomes, this study provided class I evidence that rituximab is ineffective in improving ISS in patients with IgM anti-MAG demyelinating neuropathy. However, statistically significant changes were observed in the number of patients (4 out of 20 vs. 0 out of 27) with an improvement of at least 2 points in the INCAT (inflammatory neuropathy cause and treatment) disability scale (p = 0.027) but not in the number of cases improving by 1 point (8 out of 20 vs. 4 out of 27; p = 0.0503). In a retrospective study of 45 cases, improvement was defined as a 1 or more point decrease on the 5-point Rankin scale. The review suggested that rituximab combined with purine analogs or cyclophosphamide is effective in severe anti-MAG neuropathy and a Rankin score of 3 (moderate disability/ requires some help but able to walk unassisted) or worse and that responses are more rapid than with rituximab monotherapy (34). This approach has yet to be evaluated in a prospective controlled trial.

Patients with neuropathy and anti-MAG antibodies may respond to rituximab if they have associated proximal muscle weakness or rapid progression, particularly in the first few years of the disease. In attempt to identify the subgroup of responders to rituximab, 33 anti-MAG neuropathy cases treated with rituximab were retrospectively evaluated for predictors of treatment response, as defined by an improvement of at least one point of the Overall Neuropathy Limitation Scale (ONLS) at 6 months or at last follow-up (25). In the later study, response to rituximab was significantly associated with proximal weakness of the lower limbs at the onset of disease and subacute evolution, but not with anti-MAG antibody titers (21). A rare manifestation of anti-MAG neuropathy is acute deterioration, which can be challenging to manage. Four anti-MAG neuropathy cases experiencing such acute deterioration were treated with plasma exchange (03). Plasma exchange resulted in rapid neurologic improvement in all patients and was found to be safe. In a study of 21 MAG neuropathy cases who initially responded to rituximab, patients were followed up for 11 years and were retreated if they relapsed (05). Besides disease severity, high baseline serum B-cell-activating factor levels at onset predicted worse prognosis. Interest in rituximab has continued in a study that associates anti-MAG antibodies with the clinical presentation of cerebellar ataxia as supported by magnetic resonance spectroscopy (101). Out of five cases, one did not have neuropathy, but all had improvement in cerebellar signal magnetic resonance spectroscopy following treatment with rituximab.

Potential monoclonal antibody therapeutic approaches in anti-MAG neuropathy have been reviewed (09). Reports on next generation CD20 monoclonal antibodies with different CD20 binding site (ofatumumab) or higher CD20 affinity via enhanced binding to the FcγRIII receptor on immune cells (obinutuzumab) are either inexistent (ofatumumab) or very limited (obinutuzumab), with concern for severe pneumonia in the two chronic lymphocytic leukemia with anti-MAG neuropathy treated cases.

Four patients diagnosed with MAG neuropathy developed Waldenstrom macroglobulinemia 1 to 8 years later (56). The treatment regimen consisted of a 6-months combined bendamustine (a nitrogen mustard compound)-rituximab. Most responders improved within the first 2 months as evidenced by a decrease in the INCAT score. Two cases relapsed at 36 to 40 months and were treated with rituximab only and were relapse-free for another 12 months. In a retrospective study of MAG neuropathy, 34 were treated with rituximab and 30 with immunochemotherapy (ICT), which consisted of rituximab plus one or two chemotherapeutic agents (63). As assessed by the modified Rankin scale, there were more responders in the immunochemotherapy group than in the rituximab only group. However, adverse effects occurred more frequently in the immunochemotherapy group than in the rituximab only group (62% vs. 15%). A patient developed secondary acute leukemia 5 years after receiving immunochemotherapy.

MYD88 L265P mutation is most common in Waldenstrom macroglobulinemia (with CXCR4 modulating disease severity) and a smaller proportion of IgM MGUS harbor this same change (09). This led to interest in drugs that inhibit Bruton tyrosine kinase (BTK) in treating Waldenstrom macroglobulinemia. Pooled data from two reports indicate that 10% to 15% of cases treated with BTK inhibition to control Waldenstrom macroglobulinemia had neuropathy (n=13), and three of 13 harbored MAG antibodies. Stability of neuropathy was described in half of the cases and improvement on the other half. More systematic neuropathy scales were performed in a report of three anti-MAG neuropathy patients with Waldenstrom macroglobulinemia carrying the MYD88 mutation and wild type CXCR4 (14). All three cases were treated with a BTK inhibitor and had both subjective and objective treatment response. Finally, inhibiting BCL2 in combination with rituximab improved severe neuropathy in a patient with chronic lymphocytic leukemia and anti-MAG polyneuropathy with MYD88 wild-type (10). A listing of studies exploring drugs with a variety of mechanisms of action in anti-MAG neuropathy was published in 2023 (86).

Gorson and colleagues reviewed the modest efficacy of mycophenolate mofetil in 13 chronic inflammatory demyelinating polyradiculoneuropathy and eight IgM-monoclonal gammopathy of undetermined significance demyelinating neuropathy patients. Only one of the later eight cases (13%) responded. The patient had ataxic neuropathy with a treatment response consisting of improved balance and sensory scores (29).