Background Ataxic Guillain–Barré syndrome is characterised by profound ataxia with negative Romberg sign and no ophthalmoplegia. Its nosological relationship to acute sensory ataxic neuropathy has yet to be discussed.
Methods Medical records were reviewed of patients suffering acute ataxia and reduced muscle stretch reflexes but without external ophthalmoplegia. Clinical features and laboratory findings were analysed. Rat muscle spindles were immunostained by anti-GQ1b and -GD1b antibodies.
Results The Romberg sign was negative in 37 (69%) of 54 patients with acute ataxic neuropathy without ophthalmoplegia, but positive in the other 17 (31%). The negative and positive subgroups had similar features; preceding infectious symptoms (86% vs 83%), distal paraesthesias (70% vs 88%), superficial sense impairment (27% vs 24%), IgG antibodies to GQ1b (65% vs 18%) and GD1b (46% vs 47%) and cerebrospinal fluid albuminocytological dissociation (30% vs 39%). Findings did not differ between the subgroups of 466 patients with Fisher syndrome with and without sensory ataxia. Acute ataxic neuropathy patients more often had anti-GD1b (46% vs 26%) and less often anti-GQ1b (50% vs 83%) antibodies than Fisher syndrome. Anti-GQ1b and -GD1b antibodies strongly stained parvalbumin-positive nerves in rat muscle spindles, indicative that proprioceptive nerves highly express GQ1b and GD1b.
Conclusion Clinical and laboratory features suggest that ataxic Guillain–Barré syndrome and acute sensory ataxic neuropathy form a continuous spectrum. The two conditions could be comprehensively referred to as ‘acute ataxic neuropathy (without ophthalmoplegia)’ to avoid nosological confusion because Fisher syndrome is not classified by the absence or presence of sensory ataxia. That is, acute ataxic neuropathy can be positioned as an incomplete form of Fisher syndrome.
- Acute sensory ataxic neuropathy
- anti-GD1b antibody
- anti-GQ1b antibody
- Fisher syndrome
- Guillain–Barré syndrome
Statistics from Altmetric.com
- Acute sensory ataxic neuropathy
- anti-GD1b antibody
- anti-GQ1b antibody
- Fisher syndrome
- Guillain–Barré syndrome
‘The ataxic form of Guillain–Barré syndrome (GBS)’ is characterised by acute onset of profound ataxia with negative Romberg sign and no (or minimal) ophthalmoplegia.1 Clinical findings of hypo- or areflexia, distal paraesthesias and cerebrospinal fluid (CSF) albuminocytological dissociation suggest that this condition is a GBS variant. In the original description, the type of ataxia was cerebellar rather than sensory, but some investigators have reported a patient who presented with sensory ataxia as having ataxic GBS.2 The syndrome of acute sensory neuropathy is characterised by the absence of sensory nerve action potentials, normal CSF findings, loss of large myelinated fibres and axonal atrophy without inflammation in the sural nerve biopsy.3 The sensory variant of GBS is characterised by acute, monophasic, sensory neuropathy accompanied by reduced muscle stretch reflexes, CSF albuminocytological dissociation and nerve conduction features of demyelination; moreover, some patients with sensory GBS show sensory ataxia.4 The relationships between these three conditions are confusing, and a new disease classification is required.
Both clinical and laboratory findings help in the understanding of the nosological relationships between similar conditions. Fisher syndrome (FS), which is characterised by ophthalmoplegia, ataxia and areflexia, is an established GBS variant.5 Clinical presentations of ataxic GBS are similar to those of FS, except for the absence of ophthalmoplegia, indicating that ataxic GBS is an incomplete form of FS. Supporting this indication, IgG anti-GQ1b antibodies, a serological marker of FS,6 have been identified in ataxic GBS patients.7 In contrast, monospecific IgG anti-GD1b antibodies have been found in a patient with ataxic GBS8 and in two patients with acute sensory ataxic neuropathy.9 The latter two patients could be diagnosed as having acute sensory neuropathy syndrome,3 not sensory GBS4 because their nerve conduction features were axonal, and not demyelinating.
These reports prompted us to investigate the nosological relationship between ataxic GBS and acute sensory ataxic neuropathy in a large series of cases. We reviewed medical records of patients with acute ataxia and reduced muscle stretch reflexes but without ophthalmoplegia during the illness, as well as the records of FS patients. Each condition was divided into two subgroups, patients with and without the Romberg sign, and their clinical features and laboratory findings compared. Moreover, we have shown the presence of GQ1b and GD1b epitopes in rat muscle spindles to be the responsible lesion for ataxic neuropathy associated with autoantibodies against GQ1b and GD1b.
Patients and diagnostic criteria
Between 2000 and 2008, we received about 11 000 requests from Japanese physicians to test serum antiganglioside antibodies from patients with various neurological disorders. Information on age, sex, antecedent infectious symptoms, initial symptoms, neurological signs during the illness and CSF findings was obtained from each primary physician. Clinical features used in the tentative diagnosis of ‘Acute Ataxic Neuropathy without ophthalmoplegia (AAN)’ were (1) progressive ataxia by 4 weeks of onset, (2) hypo- or areflexia and (3) absence of ocular movement limitation and limb weakness during the course of illness. Those for FS were (1) progressive ophthalmoplegia and ataxia by 4 weeks of onset, (2) hypo- or areflexia and (3) absence of limb weakness during the illness course. Both conditions required alert consciousness during illness to exclude Bickerstaff brainstem encephalitis. CSF albuminocytological dissociation was defined as a raised protein concentration (more than 45 mg/dl) associated with a count of ≤10 leucocytes/μl. Written informed consent was obtained from all the patients. The Ethical Committee of Dokkyo Medical University approved the study.
Sera were taken within 2 weeks after the disease onset and before treatment. Serum IgG antibodies to GQ1b and GD1b were measured by ELISA.10 Serum was considered positive for antiganglioside antibodies when the absorbance was 0.1 or more at 1:500 dilution.
Adult female Wistar rats were purchased from Japan SLC (Shizuoka, Japan). Handling and care were in accordance with Dokkyo Medical University's Regulations for Animal Experiments and Japanese Governmental Law (No 105) and approved by Dokkyo Medical University's Animal Experiment Committee.
Immunostaining for rat muscle spindle
Rat masseter was selected as the target tissue because a relatively high number of muscle spindles are reported in this muscle.11 For simultaneous detection of gangliosides and neurofilament 200 kDa (NF200) or type IV collagen, fresh cryosections from two rats were made and double-immunoenzyme-stained. For simultaneous detection of gangliosides and parvalbumin, the rats were perfused with 4% paraformaldehyde, and double immunoenzyme staining was performed after acetone etching of the cryosections as described previously.12 Mouse monoclonal antibody to GQ1b (FS3) was made in house, which cross-react with GT1a,13 as well as serum anti-GQ1b antibodies from FS patients.14 Monoclonal antibody to GD1b (GGR12; Seikagaku Co., Tokyo, Japan), polyclonal antibodies to NF200 (Chemicon, Temecula, California), parvalbumin (Affinity BioReagents, Golden, Colorado) and mouse type IV collagen (Cosmo Bio, Tokyo, Japan) were purchased. The secondary antibodies used were alkaline phosphatase-labelled goat IgG antibodies (A9316; Sigma Chemical Co., St Louis, Missouri) to mouse IgG and horseradish peroxidase-labelled goat F(ab′)2 to rabbit IgG (55693, Cappel, Aurora, Ohio). Gangliosides were stained blue with an alkaline phosphatase substrate kit (Vector Blue, Vector Laboratories, Burlingame, California). Other antigens were stained brown with diaminobenzidine for peroxidase (WAKO Pure Chemical Industries, Osaka, Japan).
Differences in proportions were examined by the χ2 or Fisher exact test where appropriate. A difference was considered significant when the p value was less than 0.05.
Based on the medical records available, AAN was diagnosed in 54 patients. The Romberg sign was negative in 37 (69%) of the 54 and positive in the other 17 (31%) patients (table 1). Onset age distribution was similar in the subgroups. Most patients had antecedent infectious symptoms (86% vs 83% (Romberg sign-negative vs -positive subgroup)) (supplementary table). Common initial symptoms were distal dysaesthesias (51% vs 71%) and gait disturbance (49% vs 35%). At disease onset, 19% of the Romberg sign-negative patients complained of diplopia, and during the illness 49%, but there were no signs of external opthalmoplegia. Distal paraesthesias (70% vs 88%) and superficial sense impairment (27% vs 24%) were present in each subgroup during the illness. The median time to nadir was 4 days (range 2–15 days) in the negative subgroup and 7 days (range 3–13 days) in the positive subgroup. Treatments given were intravenous immunoglobulin (IVIG) (62% vs 58%; Romberg sign-negative vs -positive subgroup), plasmaphaeresis (15% vs 0%), intravenous methylprednisolone (8% vs 25%), oral prednisolone (8% vs 17%) and no specific immunotherapy (23% and 25%). Median hospitalisation was 16 days (range 3–68 days) in the former and 22 days (range 5–150 days) in the latter. No patient died in either subgroup. At discharge, 26% of the negative subgroup and 11% of the positive subgroup of patients showed complete remission with no residual symptoms.
FS was diagnosed in 466 patients and divided into Romberg sign-negative and -positive subgroups (table 1 and supplementary table). Sex ratios, age distributions, antecedent infection symptoms, initial symptoms and neurological signs during the illness were similar in both subgroups.
CSF albuminocytological dissociation occurred in 33% of AAN and 37% of FS patients during the first week of illness (table 2). Dissociation was detected in both the AAN Romberg sign-negative and -positive subgroups (30% vs 39%) and in each FS subgroup (35% vs 63%).
IgG anti-GQ1b antibodies were positive in AAN and FS (50% vs 83%, p<0.0001) (table 2). Anti-GQ1b antibodies were detected in both the AAN-negative and -positive Romberg sign subgroups (65% vs 18%, p=0.034), as well as the FS-negative and -positive Romberg sign subgroups (83% vs 84%). IgG antibodies against GQ1b but without GD1b reactivity were found in the AAN-negative and -positive Romberg sign subgroups (32% vs 6%, p=0.032), and the FS-negative and -positive Romberg sign subgroups (60% vs 51%).
IgG anti-GD1b antibodies were positive in AAN and FS (46% vs 26%, p<0.0028). Anti-GD1b antibodies were detected in AAN Romberg sign-negative and -positive subgroups (46% vs 47%) and in each corresponding FS subgroup (25% vs 35%). IgG antibodies against GD1b but without GQ1b reactivity were detected in the AAN-negative and -positive Romberg sign subgroups (14% vs 35%, p=0.072) but less frequently in the FS-negative and -positive Romberg sign subgroups (2% vs 3%).
Sensory-nerve-conduction-study results were available for 30 AAN patients with negative (n=19) and positive (n=11) Romberg signs. In both subgroups, sensory nerve action potentials were reduced without conduction delay (21% vs 64%) or were absent (11% vs 27%), indicative of neuronal or axonal damage. Sensory nerve conduction was delayed (16% vs 0%), suggestive of demyelination.
Immunostaining of rat muscle spindles
Five to 10 muscle spindles per section were present in the equatorial region (figure 1A,D) containing the periaxial fluid space in which group Ia sensory nerves were abundant.15 Axons of nerve fibres in the muscle spindles stained positively for NF200 (figure 1B). NF200-positive axons mostly were surrounded by ring-like GQ1b (figure 1B) and GD1b staining, indicative of their presence in Schwann cells. In addition, there were focal horseshoe-like stainings of GQ1b (figure 1B) and GD1b on the surfaces of intrafusal muscle fibres that were NF200-negative. Parvalbumin, a marker of proprioceptive neurons, was detected in spindle axons of the formaldehyde-perfused specimens (figure 1C,E). Although rims of those intrafusal muscle fibres were positive for parvalbumin, the test gangliosides again were present on the surfaces of intrafusal fibres. The fact that group Ia sensory nerve terminals encircled the intrafusal fibres in the equatorial region is evidence that group Ia afferent terminals of the rat muscle spindle contained high GQ1b and GD1b levels. Table 3 summarises the immunostaining results.
We reviewed the medical records of 54 AAN patients and classified them into Romberg sign-negative or -positive subgroups. The former subgroup of patients corresponds to ataxic GBS,1 whereas, sensory-nerve-conduction-study results suggest that most patients from the latter subgroup had acute sensory (ataxic) neuropathy rather than sensory GBS.3 4 16 Previously, we assumed that ataxic GBS was distinct from acute sensory ataxic neuropathy. Clinical and laboratory findings, however, showed the Romberg sign-negative and -positive subgroup patients to be very similar, often having had preceding infection symptoms, distal paraesthesias superficial sense impairment, areflexia and CSF albuminocytological dissociation. These findings suggest that ataxic GBS and acute sensory ataxic neuropathy are GBS variants and share a common autoimmune aetiology. We acknowledge that the retrospective nature of this study and the fact that some of the treating physicians may not have been experienced neurologists will raise doubts as to the accuracy of such clinical data as the presence or absence of the Romberg sign. Prospective studies are required to confirm these clinical findings.
Originally, IgG anti-GQ1b antibodies were proposed as a diagnostic marker of FS characterised by acute onset of ophthalmoplegia and ataxia,6 but subsequent studies found that some patients who presented with acute self-limited ataxia without ophthalmoplegia also carried these autoantibodies.7 17 Those studies showed that IgG anti-GQ1b antibodies are associated with ataxic GBS as well as FS, and ataxic GBS should be considered an atypical type of FS. Our study showed that these anti-GQ1b antibodies are associated with acute sensory ataxic neuropathy as reported elsewhere.18 In contrast, IgG anti-GD1b antibodies have been detected in ataxic GBS and acute sensory ataxic neuropathy.8 9 18 19 Our study confirmed that both ataxic GBS and acute sensory ataxic neuropathy are associated with IgG anti-GD1b antibodies without GQ1b reactivity. We also found anti-GD1b antibodies to be positive in some patients with anti-GQ1b antibody-negative FS.
In summary, the immunological profiles seen in the current study support the hypothesis that FS, ataxic GBS and acute sensory ataxic neuropathy are not distinct but form a continuous spectrum. However, IgG antibodies against GQ1b but without GD1b reactivity were more often detected in ataxic GBS, whereas IgG antibodies against GD1b but without GQ1b reactivity were more frequent in acute sensory ataxic neuropathy, although the latter did not reach statistical significance. The findings suggest that these specific antibodies play a role in the development of the different conditions.
One of Fisher's original patients presented with loss of lower-limb vibration sense, whereas two did not.20 Our study showed that FS patients often have a positive Romberg sign and that Romberg sign-negative and -positive subgroups have similar clinical and immunological features. In view of the common clinical and autoantibody profiles for both AAN and FS with and without the Romberg sign, we concluded that AAN should be positioned as an incomplete form of FS. The term AAN may be more useful in the nosological classification of this group of patients as AAN might not always be differentiated into ataxic GBS and acute sensory ataxic neuropathy.
An autopsy study of ataxic GBS, designated AAN in this paper, showed major degeneration of Clarke's column fibre system but no lesions in the cerebellum, which suggests that cerebellar ataxia is a consequence of damage done by afferent fibres to the spinocerebellar nucleus.1 An FS patient whose MR images showed enhanced lesions in the spinocerebellar tracts at the lower medulla level has been reported.21 Postural body sway analysis findings suggested selective involvement of the Ia afferent system group in an ataxic GBS patient who had IgG anti-GQ1b antibodies, as well as in FS patients.22 23 These findings suggest that AAN patients have a dysfunctional proprioceptive afferent system and that special sensory ataxia is caused by the selective involvement of muscle spindle afferents. Because of the electrophysiological time course in acute sensory ataxic neuropathy, it has been hypothesised that immune-mediated reversible conduction is caused by involvement of either paranodal myelin or the nodal axolemma of sensory axons.18 IgG with anti-GQ1b reactivity from a patient with acute sensory ataxic neuropathy could bind to the nodes of Ranvier on frozen human dorsal root.24
Sensitisation of rabbits with GD1b produces experimental ataxic neuropathy, indicating that anti-GD1b antibodies function in the development of ataxia.25 Muscle spindles, proprioceptive transducers within muscles, are an integral part of the γ reflex loop. They contain specialised muscle fibres that have motor innervation and are enriched by sensory endings. Antibodies to GQ1b and GD1b stained neural components and intrafusal muscle fibres of these spindles in humans.15 Our immunohistochemical study further showed that anti-GQ1b and -GD1b antibodies, mostly stained parvalbumin-positive nerves and putative nerve endings in rat muscle spindles, evidence that proprioceptive nerves highly express GQ1b and GD1b. Antibodies to GQ1b and GD1b also stained some large neurons in human dorsal root ganglia.17 Because AAN and FS patients usually show remission with no sequelae, muscle spindles or the dorsal roots, rather than dorsal root ganglia, may be the target site for antiganglioside antibody-related ataxia.
There is no established treatment for AAN, but its aetiology is close to that of FS and GBS. Because randomised controlled trials have established the efficacy of plasma exchange and IVIG for GBS,26 either may provide useful treatment for AAN. Their efficacy in treating FS, however, has yet to be shown, as there have been no randomised controlled trials.27 A retrospective analysis of 92 consecutive FS cases found that IVIG lessened ophthalmoplegia and ataxia but did not affect outcomes, presumably because of the good natural recovery of FS.28 In our retrospective study, the outcomes of all the AAN patients could not be followed up, but both AAN and FS patients appeared to have made good recoveries. IVIG efficacy was unclear, although 6/10 AAN patients received it. Complement activation is an important nerve injury mechanism that produces anti-GQ1b antibody-mediated neuropathy.29 IVIG does have anticomplement activity.30 A more rational treatment uses complement inhibitors such as eculizumab and nafamostat mesilate,31 32 both of which are used clinically to treat other conditions and are suitable candidates for GBS, FS and AAN trials.
In conclusion, clinical and laboratory features suggest that ataxic GBS and acute sensory ataxic neuropathy form a continuous spectrum, being comprehensively referred to here as AAN, and that AAN is an incomplete form of FS. AAN patients more often have IgG anti-GD1b antibodies and less often IgG anti-GQ1b antibodies compared with FS, and this immunophenotyping could add to the differentiation of AAN from FS. The search for both anti-GD1b and -GQ1b antibodies is recommended in AAN patients, as 70% of the patients carried either one of the antibodies.
We thank C Yanaka (Department of Neurology, Dokkyo Medical University), for her technical assistance, and N Shahrizaila (University of Malaya, Kuala Lumpur, Malaysia), for her editing.
See Editorial Commentary, p 239
Linked articles 235440.
The material in this paper was presented in platform format at the Peripheral Nerve Society meeting in Würzburg, Germany, 5 July 2009.
Competing interests None.
Patient consent Obtained.
Ethics approval Ethics approval was provided by the Dokkyo Medical University.
Provenance and peer review Not commissioned; externally peer reviewed.
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.