Clinical manifestations

Presentation and course

Antibodies to GD1b/GD1a ganglioside complexes are present in a proportion of Guillain-Barré syndrome (31; 27). Several case reports have suggested the presence of ataxia in patients with Guillain-Barré syndrome and GD1b IgG antibody. Interestingly, both sensory ataxia and cerebellar ataxia have been reported (66; 02; 79; 21; 22; 48). Highly specific, GD1b IgG antibodies have been reported to more likely lead to sensory ataxia (30). However, the clinical features of GD1b-associated Guillain-Barré syndrome can extend to involve cranially innervated muscles (03; 42). Occasionally, the presence of GD1b IgG antibody has been reported in the Miller Fisher syndrome variant, even in the absence of GQ1b antibody (43; 17; 12; 34). Usually, GD1b IgG antibody in Guillain-Barré syndrome is present along with other anti-ganglioside antibodies, but mono-specific presence of GD1b has been rarely reported (66; 77; 79; 30; Bhagat abd Brown 2021). However, overall, the frequency of anti-GD1b antibody in Guillain-Barré syndrome is low, as it is present in approximately 11% to 12% cases of classic Guillain-Barré syndrome and 14.3% cases of Miller Fisher syndrome (62; 64). Anti-GD1b IgG-mediated neurologic symptoms related to COVID have been reported, but clinical phenotype has varied significantly, including cranial neuropathy with meningo-polyradiculitis and brainstem encephalitis (43; 17; 16; 15).

Similar to Guillain-Barré syndrome associated with GD1b IgG antibody, chronic sensory neuropathy associated with anti-GD1b IgM antibodies is a rare disorder, and the majority of the literature is based on case reports. IgM anti-GD1b is reported in about 2.4% cases of Guillain-Barré syndrome (versus 12.1 for IgG GD1b antibodies), 3% cases of monoclonal gammopathy of unknown significance neuropathy (compared to 52/101 MGUS patients with MAG antibody from the same series), and also in one patient with paraneoplastic anti-Hu associated neuropathy (62). To date, 181 cases have been reported with four case series reporting more than 10 patients (69; 13; 38; 49). Seventy-seven percent of reported cases were male, and the age range varied between 30 and 81 years. Disease onset is usually insidious or chronic, but acute presentation is reported in 18% to 22% of patients (69; 38; 49). Interestingly, in some cases, the initial presentation was acute and treatment-responsive, and the relapse or the second event happened years later (54). Several patterns of onset have been described in the chronic form of the illness: insidiously progressive, acute relapsing and remitting, and acute relapses on a chronic background, often extending for more than 10 years. In addition to the chronic syndrome, patients with craniobulbar motor involvement may have these either as a fixed set of symptoms or as relapsing-remitting symptoms. These comprise double vision and eyelid drooping, dysphagia, and dysarthria. The episodes of relapse affecting the craniobulbar motor system may be mistaken for brainstem strokes or central demyelination; however, it should be borne in mind that central nervous system manifestations may also be present in these chronic cases, which are yet to be fully defined (69; 38; 49).

The anti-disialosyl IgM antibodies are usually associated with a clinical syndrome with chronic sensory ataxia without functionally substantial limb weakness in most cases. Large fiber sensory loss to vibration and proprioception is a common finding, and the majority of the patients also have sensory paresthesia (69; 13; 38; 49). Areflexia is usually present in more than 90% cases (38). Motor weakness is rare at the time of diagnosis but may present later in fewer than 30% to 40% cases. Both distal and proximal motor weakness have been reported (69; 13; 38; 49). In some cases, the IgM paraprotein also has cold agglutinating activity with anti-Pr specificity (01; 70; 19). In other cases, craniobulbar motor involvement, including ophthalmoplegia, may also be present (24; 19; 69). To reflect this, in 1996 the acronym CANOMAD was coined to encompass this phenotype: chronic ataxic neuropathy with ophthalmoplegia, M-protein, agglutination, and disialosyl antibodies (68). Although this acronym is useful, not all cases exhibit the full spectrum of clinical features. Considering the limitations of this acronym and potential restriction in terms of diagnosis, an alternative acronym, CANDA (chronic ataxic neuropathy with disialosyl antibodies), has been proposed (80). This acronym recognizes the strong association between ataxia and anti-GD1b antibodies (59).

In the initial case series by Willison and colleagues, all patients had at least cranial nerve 3 involvement, and ataxia was present in all but one patient (69). The clinical features of another series of 11 cases have been reported, adding to the consistent pattern of this presentation (13). In a Japanese series of five cases of sensory ataxic neuropathy with anti-GD1b antibodies and IgM paraproteinemia, all patients had severe sensory ataxia and areflexia, but only one of the five cases manifested a transient diplopia on lateral gaze, and two of the five cases had cold agglutinins (60). In contrast, two cases reported as part of a long-term follow-up of paraproteinemic neuropathy cases had all the features of CANOMAD (50). Two large series of CANOMAD/CANDA patients (at least one serum IgM antibody reacting against disialosyl epitopes among GD1b, GD3, GT1b, and GQ1b) from France and Switzerland confirmed previous findings and reported sensory ataxia in 76% to 85% of patients (38; 49). One case series from the United States reported tremor in more than 90% of cases and pseudoathetosis in 27% of cases (13).

Among other clinical symptoms, motor and sensory involvement of cranial nerves, particularly fifth and seventh nerves, can be present (69; 38). Two cases have reported the concomitant presence of optic neuropathy (55). Unusual manifestations include choreoathetoid movements of the face and tongue and respiratory insufficiency requiring intubation (29). Some cases may have autonomic involvement with pupillary abnormalities and bladder or bowel symptoms; however, these are rarely symptomatically prominent (38).

Prognosis and complications

Typically, CANDA and CANOMAD are treatment responsive with relapsing-remitting features. As mentioned before, relapse after many years of stability is not uncommon. In some cases, the disease can be progressive. Despite relative motor strength sparing, limb ataxia may lead to significant disability, and some patients are unable to walk without assistive devices. About 7% to 27% of patients become wheelchair-dependent (69; 13; 38; 49). Reports of bulbar dysfunction varied between less than 10% to 66% cases and, when present, can lead to swallowing-related complications, including impaired nutrition and aspiration (69; 13; 38; 49).

Clinical vignette

A 67-year-old woman was first noted by her local blood transfusion laboratory to have cold agglutinins when examined at 49 years of age. At 55 years of age, she presented with paraesthesia and numbness first affecting her feet and then her hands. She also had difficulty with fine finger movements and unsteadiness while walking, particularly in the dark. She was only able to walk with the aid of a walking frame. She later developed variable and intermittent symptoms of ptosis and double vision, alterations in the quality of her voice, and difficulty with chewing and swallowing, in addition to perioral paraesthesia. She did not have symptoms or signs of cold agglutinin disease.

On examination, she had an area of perioral sensory loss to pinprick, a mild bilateral ptosis, impaired pupillary light responses, and otherwise normal cranial nerves. There was no wasting or weakness in the limbs. She was areflexic. Glove and stocking sensory loss to light touch and pinprick was present with an additional area of sensory loss over the mid-abdominal region and sternum. Joint position sense was markedly impaired in the digits and vibration sense absent to the elbows and ribs. Deep pain and thermal sensation were normal. She had prominent pseudoathetosis of the outstretched hands and gait ataxia with a positive Romberg test. She was unable to walk unaided.

Electrophysiological examination of motor nerves revealed prolonged distal motor latencies and F wave latencies (right common peroneal distal motor latencies: 6.9 msec, F wave: 76.4 m/sec; right tibial distal motor latencies: 5.8 m/sec, F wave: 83.9 m/sec; right median distal motor latencies: 5.2 m/sec, F wave: 43.2 m/sec). Motor conduction velocities in the leg and arm were reduced (right common peroneal: 34 m/sec; right tibial: 30.5 m/sec; right median: 37.5 m/sec). All sensory nerve action potentials examined (right sural, median, ulnar, and radial) were absent. Needle electromyography revealed borderline neurogenic changes. Visual evoked potentials were normal. CSF examination was normal, with a protein level of 0.36 g/L.

Sural nerve biopsy demonstrated a severe reduction in large myelinated fibers with preservation of the unmyelinated small fibre population.

No active fiber degeneration or demyelination was observed, although there was a degree of axonal regeneration with some isolated fibers surrounded by thin myelin sheaths suggestive of remyelination. There were no hypertrophic changes, and no inflammatory cells or immunoglobulin deposits were noted.

Laboratory tests revealed a normal full blood count, elevated ESR at 96 mm/hr (recorded at room temperature), elevated serum IgM at 6.4 g/L with an IgM-lambda paraprotein at around 4.0 g/L and normal serum IgG and IgA levels. Bone marrow examination, urine analysis for Bence Jones proteinuria, and skeletal survey were normal. Examination of the patient’s serum by ELISA demonstrated IgM antibodies reactive with all gangliosides containing NeuNAc(a2-8)NeuNAc(a2-3) configured disialosyl groups at titers in excess of 1/105. The highest titers were, thus, obtained with GD1b (1/850,000), GQ1b (1/450,000), GT1b (1/300,000), and GD3 (1/260,000). Weak reaction was observed with GD1a (1/570) and GM3 (1/660), which contain a terminal NeuNAc(a2-3)Gal(b1-) structure. The monosialic series gangliosides GM1 and GM2 gave no reaction, neither did asialo-GM1.

Red blood cell studies on the serological reactions of her cold agglutinins demonstrated that they (1) were strongly agglutinated all "common phenotype" red blood cells including cord cells, (2) did not agglutinate neuraminidase-treated human or canine red blood cells, (3) weakly agglutinated papain modified red blood cells, (4) weakly agglutinated MkMk red blood cells; (5) strongly agglutinated "common phenotype" red blood cells modified by mild treatment with sodium metaperiodate, and (6) strongly agglutinated canine red blood cells both before and after modification with papain. These characteristics are typical of anti-Pr2 specificity.

She was initially treated with a single course of plasma exchange (total exchange 15 liters) followed by low-dose oral cyclophosphamide (150 mg/week in three divided doses). She reported an improvement in her sensory symptoms, hand function, and mobility after 8 months of treatment although there was no quantifiable evidence of this, and she remained significantly disabled. She was subsequently treated with 2 g/kg of intravenous human immunoglobulin, again with no definite improvement, either symptomatically or objectively.

Biological basis

Etiology and pathogenesis

Like many other autoimmune diseases, the exact pathophysiology of GD1b-mediated neuropathy remains unclear. However, there is experimental evidence to suggest a direct role of the anti-GD1b in disease pathophysiology. Sensitizations of six rabbits with GD1b led to the production of IgM anti-GD1b antibodies, with a low titer of antibody against GM1, but no cross-reactivity was noted against GD1a or GT1b. Three of six rabbits developed neurologic symptoms with splayed limbs and awkward functioning of the limbs and retained motor power. No definitive cranial nerve dysfunction was present, and no conduction block was detected. Positive immunostaining was noted in dorsal root ganglia but not in ventral roots and paranodal myelin. There was axonal degeneration with loss of cell bodies in dorsal root ganglia, but no demyelinating features or mononuclear cell infiltrations were observed (37). Dorsal root ganglion neurons might be injured by antibody-mediated downregulation of trkC, leading to growth factor deprivation and subsequent neuronal apoptosis (63).

GD1b has been shown to be present on dorsal root ganglion neurons; thus, anti-GD1b antibodies may bind at this site, which is relatively deficient in the blood-nerve barrier and can cause antibody and complement-mediated injury. In fact, whole serum from a patient with CANOMAD only bound to dorsal root ganglia but not to the sciatic nerve, lumbar spine, or paranodal area (46; 36). However, other reports suggested more widespread immunoreactivity of IgM cold agglutinating antibodies from patients with CANOMAD, including sensory and motor nerves, third cranial nerve, with prominent staining in groups of dorsal root ganglia neurons but sparing the myelin or paranodal area in both murine and human tissue (68; 28). Thus, the balance of the existing evidence points to a direct role for anti-GD1b antibodies in causing dorsal root ganglion neuronal injury, but this remains to be fully investigated mechanistically (09).

Motor nerve terminals in selected sites, such as in craniobulbar muscles, may also be affected by antibody binding similar to that which occurs in Miller Fisher syndrome (53). In support of this, an immunohistological study has shown that anti-GD1b antibodies and related antibodies reactive with other disialylated gangliosides bind to motor end plates in human extraocular muscle (39). In contrast, binding to end plates in limb muscles was sparse, but the nerve terminals inside muscle spindles and in direct contact with intrafusal fibers were labeled with anti-GD1b, -GQ1b, and -GT1a ganglioside antibodies. A pathophysiological study in mice has shown that anti-GD1b antibodies, in comparison with anti-GM1/GD1a antibodies, preferentially target nodal axolemma in sensory nerves, leading to nodal disruption (61). An in vitro study on myelinating sensory neuron cultures has demonstrated both complement-dependent demyelination and complement-independent dysmyelination on exposure to anti-GD1b IgM antibody, indicating there may be multiple pathways mediating nerve injury (07).

Although the phenotype described here is distinctive, many cases of chronic sensory ataxic neuropathy do not have antidisialosyl antibodies. These cases, therefore, presumably have a different etiology that may be toxic, degenerative, genetic, or autoimmune. In a series of 17 patients with chronic idiopathic sensory ataxic neuropathy, only 3 of 17 patients had antidisialosyl antibodies (23). In the control group of 93 cases in which the etiology of the sensory ataxic neuropathy was known, no cases had antidisialosyl antibodies. Normal human sera do not contain these antibodies, indicating a high specificity for this distinctive syndrome and a likely causal relationship. In another report of two cases of chronic ataxic neuropathy, one case had antidisialosyl antibodies, whereas the other had anti-GD1a antibodies; the latter was more typically associated with pure motor syndromes (65). This curious case underlies some of the etiologic complexities that remain unresolved in this clinical area.

The anti-disialosyl antibodies found in these disorders may arise through cross-reactivity with bacterial lipopolysaccharides, for example as carried on certain serostrains of Campylobacter jejuni. This molecular mimicry model has been substantiated for the anti-GQ1b antibodies that arise in Miller Fisher syndrome and some of the anti-GQ1b antibodies found in Miller Fisher syndrome also react with GD1b. Similar experiments have been conducted with anti-disialosyl paraproteins and Campylobacter jejuni lipopolysaccharides that support this view (28). Exactly how these B lymphocyte antibody responses make the transition from acute phase immune responses to chronically stable, regulated clones of paraprotein-secreting B cells is unknown.

Detailed postmortem pathological studies on this syndrome are sparse. In one case report, autopsy showed severe dorsal column atrophy and dorsal root ganglionopathy and infiltration of clonal B-lymphocytes within the endoneurium, perineurium, and leptomeninges of peripheral nerve (44). It is curious that sural nerve biopsies from these patients have not shown any IgM deposition (unlike the IgM deposition seen in the anti-MAG paraproteinemic neuropathy), suggesting that the pathophysiology may be complex.

Epidemiology

The incidence and prevalence of anti-GD1b antibody-associated disorders is unknown, and given the difficulty with diagnosis and lack of awareness, a reported prevalence at this stage will likely be lower than the true prevalence. However, it remains clear that this rare disease has a strong male predisposition, and the majority of the patients were between 40 and 70 years of age. Furthermore, four of five relatively bigger case series are from Europe, but it is unclear if there is a reporting bias or if there is a true geographical or genetic influence (69; 13; 38; 49).

Differential diagnosis

Confusing conditions

A clinically similar pattern of sensory neuropathy and ataxia can be seen in a number of other conditions, including sensory variant chronic inflammatory demyelinating neuropathy and chronic immune sensory polyradiculopathy (57), paraneoplastic neuropathy, Sjogren syndrome (09; 51; 47; 74; 57). Furthermore, drug-induced and toxic neuropathies associated with semisynthetic penicillins (58) or neuropathies caused by pyridoxine overuse, cisplatinum, paclitaxel, methyl mercury, or thalidomide, vitamin deficiency syndromes (B12, B6, or E), neuropathies associated with HIV and human T cell lymphotrophic virus, and hereditary conditions including hereditary sensory and autonomic neuropathies, Friedreich ataxia, olivopontocerebellar atrophy, sensory ataxic neuropathy dysarthria with ophthalmoparesis (18), and abetalipoproteinemia may have a similar presentation (18). However, as explained above, the key feature is the presence of anti-GD1b antibody, which is an available commercial test that should be performed to confirm the diagnosis of CANDA or CANOMAD in the proper clinical scenario (72; 05; 13).

Associated or underlying disorders

As clear in the acronym CANOMAD, paraproteinemia is common in this syndrome, and monoclonal gammopathy is usually present in about 65% to 100% of patients, with essentially an equal distribution of Kappa and Lambda light chains. In some cases, Kappa and Lambda light chains may co-exist (69; 05; 13; 49). Of note, paraproteinemia may not be apparent at the time of initial presentation and can develop later in the disease process (44).

Hematological malignancies can co-exist and Waldenström macroglobulinemia has been reported in about 3.6% to 20% of patients from larger studies (38; 49). Different types of lymphoma, including diffuse large B cell lymphoma and other subtypes of B cel lymphoma, marginal zone lymphoma, mantle cell lymphoma, and chronic myeloid cell lymphoma, have been reported in some patients (33; 32; 38; 49).

Diagnostic workup

As for any other neuropathy, basic workup to assess treatable causes of neuropathy per the American Academy of Neurology guideline is recommended, and assessment should include evaluation for vitamin deficiencies, toxic causes, and hereditary and infectious workup as applicable.

The diagnostic workup for the acute presentation of IgG anti-GD1b-mediated neuropathy should be similar to that for Guillain-Barré syndrome. Based on the current practice, electrodiagnostic studies, along with CSF analysis with or without spinal and nerve root imaging, are recommended. Given the uncertain evidence of whether these ganglioside antibodies may change management, routine testing for anti-ganglioside antibodies is not evidence-based.

For IgM anti-GD1b antibody-mediated CANDA/ or CANOMAD, strong clinical suspicion is required, particularly if there is cranial nerve involvement. Blood work to assess paraproteinemia and cold agglutinin can be helpful. However, cold agglutinins are not always present, and wide variations exit in terms of presence of cold agglutinins (0% to 50%) (69; 38; 49).

Electrodiagnostic studies can help to assess the severity and characterize the pattern of neuropathy and can serve as a baseline but may not have specific diagnostic value. Both sensory and motor nerves can be involved. Two large studies from Europe reported about 60% of the cases with demyelinating neuropathy (38; 49), but others have reported both axonal and demyelinating patterns, and sometimes a mixed pattern (69; 13). Motor conduction block in the absence of anti-GM1 antibody has also been reported (33; 41; 26).

Cerebrospinal fluid analysis may show albumin cytological dissociation in some cases and can be helpful to rule out underlying infectious, inflammatory, or other etiologies. Occasional mild elevation of lymphocytes has been observed (69; 13; 38; 49). Similar to CSF studies, nerve biopsies do not have any specific diagnostic values and are not routinely recommended. However, nerve biopsy may be indicated in special circumstances, including rapid progression or suspicion for other etiologies. Reported findings include demyelination, endoneural fibrosis, and IgM deposits (69; 13; 38; 49). Inflammatory infiltrates are usually absent, but rare nonspecific epineurial lymphocytic infiltration can be observed (38).

When a central nervous system lesion is suspected, a CT or MRI of the brain should be performed. MRI brain is usually normal, but in some cases, central demyelination has been reported (69; 49). For patients with myelopathic signs, the spinal cord, including the region of the foramen magnum, should be investigated by myelography or MRI. Furthermore, MRI can show nerve root enhancement and hypertrophy, particularly of lumbosacral and brachial plexus (49). Neuromuscular ultrasound has shown regional nerve enlargement consistent with acquired demyelination in some cases (13).

Management

No randomized clinical trials of treatment have been performed. Conventional immunomodulatory therapy with corticosteroid, plasma exchange, and intravenous immunoglobulin has been used with symptomatic benefit. Even in the absence of systematic studies, corticosteroid appears to be less effective without long-term lasting benefit, and disease progression is common (38). However, there can be reporting bias, and comparative studies are required (49).

More recent reports used immunoglobulin more frequently, and often successfully (32; 35; 50). In about 50% to 60% of cases, IVIg showed partial response and often prevented relapse (13; 38; 49). At present, it remains unclear if IVIg is particularly useful in patients with demyelinating features in electrodiagnostic studies.

However, in some cases there was disease relapse and progression despite immunoglobulin therapy, and B cell depletion therapy with rituximab has been successfully employed with long-lasting benefit (11; 40; 13; 38; 49). In a large study with a French cohort, rituximab either alone or in combination with other immunomodulatory agents showed positive response in 22 of 26 patients (38). Another study reported complete response in 50% of patients treated with rituximab (49). Similarly, rituximab successfully halted disease progression over 2 year in eight of nine patients from a United States cohort (13).

Other immunomodulatory and immunomodulatory therapies, including azathioprine, mycophenolate mofetil, chlorambucil, and intravenous cyclophosphamide have been used intermittently, but there is no definitive pattern in terms of treatment response (13; 38; 49).

In the presence of an IgM paraprotein, care should be taken to fully characterize any underlying hematological abnormality; in some cases, underlying Waldenström macroglobulinemia or different types of lymphoma have been identified, which requires specific treatment (56).

Treatment should always be offered with caution, and the risk-benefit analysis should be explained to the patient, particularly when using cytotoxic therapies or B cell depletion therapies. Pre-infusion tests for certain viral organisms and their prompt treatment is a helpful approach to prevent or limit flare-up of underlying infections. Vigilance for the emergence of infections, including opportunistic infections, is recommended when using immunosuppressive agents. Another point of caution is that the temperature of replacement fluids during plasma exchange, or during other intravenous fluid use, should be carefully controlled in patients with cold agglutinins to avoid intravascular agglutination.

Special considerations

Anesthesia

No information is available on adverse effects of anesthetic agents. Patients with cold agglutinins should not be deliberately or inadvertently cooled during anesthesia as they are likely to agglutinate red cells with serious side effects.