Article Text

Research paper
Antiganglioside antibodies are associated with axonal Guillain–Barré syndrome: A Japanese–Italian collaborative study
  1. Yukari Sekiguchi1,
  2. Antonino Uncini2,3,
  3. Nobuhiro Yuki4,5,
  4. Sonoko Misawa1,
  5. Francesca Notturno2,3,
  6. Saiko Nasu1,
  7. Kazuaki Kanai1,
  8. Yu-ichi Noto1,
  9. Yumi Fujimaki1,
  10. Kazumoto Shibuya1,
  11. Shigeki Ohmori1,
  12. Yasunori Sato6,
  13. Satoshi Kuwabara1
  1. 1Department of Neurology, Graduate School of Medicine, Chiba University, Chiba, Japan
  2. 2Department of Neuroscience and Imaging, University ‘G d’Annunzio', Chieti-Pescara, Italy
  3. 3Neurocentre of Southern Switzerland (EOC), Lugano, Switzerland
  4. 4Department of Microbiology, National University of Singapore, Singapore
  5. 5Department of Medicine, National University of Singapore, Singapore
  6. 6Clinical Research Centre, Chiba University Hospital, Chiba, Japan
  1. Correspondence to Dr S Kuwabara, Department of Neurology, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan; kuwabara-s{at}


Background Whether or not antiganglioside antibodies are related to axonal or demyelinating Guillain–Barré syndrome (GBS) is still a matter of controversy, as detailed in previous studies conducted in Western and Asian countries.

Objective To clarify whether antiganglioside antibodies are associated with axonal dysfunction in Japanese and Italian GBS patient cohorts.

Methods Clinical and electrophysiological profiles were reviewed for 156 GBS patients collected from Japan (n=103) and Italy (n=53). Serum IgG antibodies against GM1, GM1b, GD1a and GalNAc-GD1a were measured by ELISA in the same laboratory. Electrodiagnostic criteria and results of serial electrophysiological studies were used for classification of GBS subtypes: acute inflammatory demyelinating polyneuropathy (AIDP) and acute motor axonal neuropathy (AMAN).

Results In both Japanese and Italian cohorts, any of the antibodies were positive in 36% of the patients, and antibody positivity had a significant association with the AMAN electrodiagnosis. Approximately 30% of Japanese and Italian antiganglioside positive patients showed the AIDP pattern at the first examination whereas sequential studies showed that most finally showed the AMAN pattern. Clinically, seropositive patients more frequently had preceding diarrhoea and pure motor neuropathy in both Japanese and Italian cohorts; vibratory sensation was normal in 97% of Japanese and in 94% of Italian seropositive patients.

Conclusions In GBS, clinical and electrophysiological features appear to be determined by antiganglioside antibodies, and the antibodies are associated with motor axonal GBS in both Japan and Italy. Classification of the GBS subtypes as a disease entity should be made, combining the results of antiganglioside assays and serial electrodiagnostic studies.

Statistics from

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.


Guillain–Barré syndrome (GBS) is an acute immune mediated polyneuropathy, currently classified into several subtypes by electrophysiological and pathological criteria.1–3 The two major subtypes are acute inflammatory demyelinating polyneuropathy (AIDP; the classical demyelinating form) and acute motor axonal neuropathy (AMAN; an axonal subtype). Over the past 20 years, major advances have been made in understanding the immunopathogenesis of GBS, particularly in AMAN, and it is now established that in certain patients, AMAN is caused by molecular mimicry of human gangliosides by the Campylobacter jejuni lipo-oligosaccharide, the immunological target of antiganglioside antibodies being the axolemma of motor fibres.4–6

Autopsy studies in AMAN patients revealed Wallerian-like degeneration in motor axons with deposits of IgG and complement at the node of Ranvier without demyelination, suggesting that the disorder primarily involves the axonal membrane.2 7–9 Moreover, in an animal model of AMAN rabbits sensitised with GM1, pathological studies showed findings identical to those in human AMAN patients; again, axonal degeneration without demyelination.5 10–12 Based on these findings, antiganglioside antibody positive GBS is expected to be identical to axonal GBS or AMAN.

Several studies from Japan have suggested that AMAN is associated with antiganglioside antibodies and pure motor axonal damage13 15 but this association was not confirmed in other studies in China and Western countries.1 16 Therefore, it has not yet been concluded whether antiganglioside antibodies cause axonal GBS, and each study has been conducted in a different country with different antiganglioside assays.17

To elucidate the precise relationship between antiganglioside antibodies and GBS electrodiagnoses, we analysed the clinical and electrophysiological data in two different populations of GBS patients, in Japan and Italy, examined with the same electrophysiological protocol and antiganglioside assay. The data suggest that GBS with antiganglioside antibodies is associated with AMAN as a disease.



A total of 156 patients with GBS were collected from two cohorts (103 Japanese and 53 Italian). All patients fulfilled the clinical criteria for GBS. Japanese patients were seen consecutively at Chiba University Hospital between 1998 and 2009. There were 64 men and 39 women, aged 12–81 years (median 39). A total of 53 Italian GBS patients were collected at the University Hospital of Chieti between 1995 and 2009. There were 34 men and 19 women, aged 9–79 years (median 48).

Clinical and electrophysiological data of both cohorts were analysed separately and then combined if they were consistent. Some of data have been published previously.18 19 The first electrophysiological studies were performed within 14 days from the onset of neurological symptoms except in 10 patients examined 15–28 days from onset. Pretreatment serum sample was taken on the day of electrophysiological examination, and frozen at −80°C. The study protocol was approved by the local ethics committees.


In both Japanese and Italian patients, nerve conduction studies were performed using the same conventional procedures and by a Viking 4 electromyography machine. Motor nerve studies were made of the median, ulnar, peroneal and tibial nerves, including F wave analyses. Antidromic sensory nerve conduction studies were done for the median and ulnar nerves. In 81 of the 156 patients, sequential studies were done up to 6 weeks after the onset of GBS, and 50 Japanese patients received three or nerve conduction studies up to 6 months after onset.

Patients were classified as having AIDP or AMAN on the basis of the electrodiagnostic criteria reported by Ho and colleagues.1 AIDP was diagnosed when patients had one of the following in two or more nerves: (1) motor nerve conduction velocity <90% of the lower limit of normal (LLN); (2) distal motor latency >110% of the upper limit of normal (ULN); (3) unequivocal temporal dispersion; and (4) F wave latency >120% ULN. AMAN was diagnosed when distal compound muscle action potential (CMAP) amplitudes were <80% of LLN in at least two nerves without evidence of demyelination. In patients who underwent serial nerve conduction studies, the final diagnosis of AMAN was made: (1) when there was an AMAN pattern 3–6 weeks after onset or (2) when there was evidence of early reversible conduction failure characterised by reduced distal CMAP amplitude with normal or slightly prolonged distal motor latencies and/or partial conduction block (proximal/distal <0.5) in intermediate nerve segments which, at serial recordings, recovered within 6 weeks without developing temporal dispersion (increased duration of negative peak of proximal CMAP >30% compared with distal CMAP) or conduction slowing.13 19 Therefore, the final electrodiagnosis was based on the results of sequential studies and taking into consideration the whole electrophysiological history of each patient. The criteria of Hadden et al16 were also used, and results were compared with those of Ho's criteria.

Patients were classified as acute motor and sensory axonal neuropathy (AMSAN) when motor nerve conduction studies showed the AMAN pattern and sensory nerve action potential amplitude was <50% of LLN in at least two nerves.20

In this paper, because the electrodiagnosis and disease entity may not be equivalent, the electrodiagnoses of AIDP and AMAN were respectively expressed as ‘AIDP’ and ‘AMAN’, and the disease entity as AIDP and AMAN.

Antiganglioside antibody assay

IgG antibodies against gangliosides GM1, GM1b, GD1a and GalNAc-GD1a were measured by ELISA, as described elsewhere,15 by one of the authors (NY) who was blinded to the clinical and electrophysiological data. These gangliosides were chosen as antigens because they are reported to be elevated in sera from GBS patients during the acute phase of the syndrome, and previous studies showed their close correlation with the electrodiagnosis of AMAN.14 15 Therefore, it was supposed that antibodies to these gangliosides should be examined first. When at least one of the antibodies was positive (>1:500), patients were regarded as antiganglioside positive.

Statistical analyses

Patients were initially grouped by positivity of antiganglioside antibodies and then by the electrodiagnoses of AIDP and AMAN. Comparative analyses were made with Fisher's exact test or the χ2 test, and median values were compared with the Student's t test. Summary statistics were constructed using frequencies and proportions for categorical data, and means and SDs for continuous variables. We compared positive and negative groups using Fisher's exact test for categorical outcomes and t tests for continuous variables, as appropriate. A p value <0.05 was considered to be statistically significant. All statistical analyses were performed using the SAS software program, V.9.2 (SAS Institute Inc).


Antiganglioside antibodies and first electrodiagnoses

Table 1 shows the relationship between antiganglioside antibodies and electrodiagnoses at the first nerve conduction study. Any of the IgG antiganglioside antibodies measured were positive in 36% of cases in both the Japanese and Italian cohorts. In 156 patients, the frequency of antibody against GM1 was 26%, GM1b 15%, GD1a 12% and GalNAc-GD1a 16%. The frequency of each antibody was not significantly different in the Japanese and Italian cohorts.

Table 1

Antiganglioside antibodies and electrodiagnosis at the first exam in patients with Guillain–Barré syndrome

Of the 103 Japanese patients, 37% were electrodiagnostically classified as having ‘AIDP’ and 23% as having ‘AMAN’. No Japanese patient had AMSAN. The frequency of the ‘AMAN’ electrodiagnosis was significantly higher in the antiganglioside positive group than in the negative group. Of the 53 Italian patients, 58% had ‘AIDP’ and 17% had ‘AMAN’. The frequency of the ‘AIDP’ electrodiagnosis was significantly higher in the seronegative group, and the frequency of ‘AMAN’ was significantly higher in the seropositive group. In the study period, three Italian patients, in addition to the 53 patients, had ‘AMSAN’, but because of the small number of patients, they were excluded from the later analyses. In both the Japanese and Italian populations, many patients were unclassified at the first electrodiagnostic study because they showed only minor nerve conduction abnormalities which did not meet the criteria for ‘AIDP’ or ‘AMAN’, whereas some of them developed the ‘AIDP’ or ‘AMAN’ pattern in the follow-up studies (see below). There was no patient with inexcitable nerves in either the Japanese or Italian patient cohorts. When the electrodiagnoses based on Hadden's criteria were analysed, the results were very similar to those of Ho's criteria.

Clinical and electrophysiological profiles

Table 2 shows the clinical features in the antiganglioside positive and negative patients for all 156 patients. The findings were similar for the Japanese and Italian cohorts. Clinically, ganglioside positive patients had preceding diarrhoea more frequently, and cranial and sensory nerve involvement less frequently. In particular, vibratory sensation was normal in 97% of Japanese and 94% of Italian antiganglioside positive patients.

Table 2

Clinical and electrophysiological profiles in 103 Japanese and 53 Italian patients

In the first nerve conduction studies, antiganglioside positive patients showed shorter distal motor latencies and higher amplitudes of sensory nerve action potentials in both populations. These analyses were performed including patients who were unclassified by electrodiagnostic criteria. Overall, in the majority of antiganglioside positive patients, there was motor dominant involvement and less prominent nerve conduction slowing and reduced sensory response. As the results of both the Japanese and Italian studies suggested that clinical and electrophysiological profiles were related to the positivity of antiganglioside antibodies, further analyses combined the Japanese and Italian patient data.

Table 3 shows the clinical profiles of the four GBS subgroups according to the electrodiagnosis at the first examination and antibody positivity. The aim of this analysis was to clarify whether antiganglioside antibodies or electrodiagnoses determined clinical features, and therefore electrophysiologically ‘unclassified’ patients were excluded. Features of antiganglioside positive ‘AMAN’ and negative ‘AMAN’ were similar (column A vs B) except age. By contrast, clinical features were significantly different between the antiganglioside positive and negative ‘AIDP’ groups (column C vs D); antiganglioside positive ‘AIDP’ patients showed more frequent preceding diarrhoea and less frequent facial palsy and sensory impairment, suggesting that the antibody positive ‘AIDP’ (column C) shared some common features to those of the ‘AMAN’ groups (columns A and B). This was confirmed by the lack of difference in comparing ‘AMAN’ and ‘AIDP’ with antiganglioside antibodies (columns A and C). Only antiganglioside negative ‘AIDP’ patients (column D) showed frequent facial and sensory nerve involvement, consistent with those of a typical classical AIDP. Overall, these findings indicate that clinical features are determined by antiganglioside antibodies rather than the electrodiagnosis at the first study.

Table 3

Clinical profiles in 102 Guillain–Barré syndrome patients (62 Japanese and 40 Italian), excluding electrodiagnostically ‘unclassified’ patients

Serial changes in the electrodiagnosis

Figure 1 shows the initial and final electrodiagnoses in 81 patients (50 Japanese and 31 Italian) who underwent at least one follow-up study. Many patients changed the electrodiagnosis, and of the 29 antiganglioside positive patients, the final electrodiagnosis was ‘AIDP’ in 17%, ‘AMAN’ in 69% and 14% remained unclassified. Of the 52 antiganglioside negative patients, the final electrodiagnosis was ‘AIDP’ in 60%, ‘AMAN’ in 8% and 33% were unclassified.

Figure 1

Electrodiagnosis of Guillain–Barré syndrome patients in the first study and after serial recordings. The first studies were performed in weeks 1–2 and the follow-up studies were carried out in weeks 3–6. The arrows indicate the directions of changes from one subtype to another and the numbers are the number of subjects changing group. The findings showed close associations of the antiganglioside positivity and ‘AMAN’, and antiganglioside negativity and ‘AIDP’. AIDP, acute inflammatory demyelinating neuropathy; AMAN, acute motor axonal neuropathy.

When excluding electrodiagnostically ‘unclassified’ patients even by the sequential findings, of the 25 antiganglioside positive patients, the final diagnosis was ‘AIDP’ in 20% and ‘AMAN’ in 80% (p<0.001). The five patients were finally classified as having ‘AIDP’ because of mildly prolonged distal latencies (112–125% of the ULN) in the follow-up studies. Of the 38 antiganglioside negative patients, the final diagnosis was AIDP in 89% and AMAN in 11% (p<0.001).

Among the nerve conduction parameters, distal motor latencies showed the most frequent and prominent abnormalities, in agreement with previous results.18 19 Figure 2 shows serial changes in distal latencies of median motor nerve studies in 50 Japanese patients. In the antiganglioside positive group (n=22), the extent of prolongation was minimal, and none did not exceed >150% of the ULN at any point, irrespective of the first electrodiagnosis of ‘AMAN’ or ‘AIDP’.

Figure 2

Sequential changes in motor distal latencies in median nerve studies of antiganglioside positive and negative patients. The electrodiagnosis reported is at the first test. Note the lack of prominent prolongation in the antiganglioside positive patients, and the distinct time course. AIDP, acute inflammatory demyelinating neuropathy; AMAN, acute motor axonal neuropathy.

In contrast, 28 antiganglioside negative patients showed prominent nerve conduction slowing, progressive up to week 25. In particular, the trend was obvious in patients with the initial electrodiagnosis of ‘AIDP’ (figure 2B, bottom). These findings suggest that the majority of GBS patients with antiganglioside antibodies do not show prominent conduction slowing consistent with typical segmental demyelination, as shown in patients without these antibodies, and electrophysiological profiles appeared to be determined by the presence of antiganglioside antibodies rather than by the electrodiagnosis.

Representative waveforms of CMAPs and sensory nerve action potentials in the median nerve of patients with or without antiganglioside antibodies are shown in the supplementary figure (available online only).


Our results show that antiganglioside positive GBS patients constitute a subgroup with no or mild/transient nerve conduction slowing, suggestive of axonal dysfunction. The majority of these patients had motor polyneuropathy, and thereby AMAN as a disease. Some antiganglioside positive patients are classified as having ‘AIDP’ because of mild nerve conduction slowing, but even classified as ‘AIDP’, their clinical features are similar to those of ‘AMAN’. Because the different frequency of ‘AMAN’ was reported between Europe and Asia, this study examined two GBS patient cohorts from Japan and Italy; the findings were consistent for both cohorts. Our findings suggest that in GBS, clinical and electrophysiological profiles are determined by antiganglioside antibodies rather than by electrodiagnosis at the first study in both Europe and Asia. Many AMAN patients show mild and transient nerve conduction slowing, and therefore may be erroneously classified as ‘AIDP’.

Previous electrophysiological studies have suggested that the pathophysiology of antiganglioside positive GBS is not simple axonal degeneration. Rapidly reversible nerve conduction block and slowing resolving within days to a few weeks are frequently found for AMAN patients.13 19 Such a time course suggests functional or microstructural changes at the nodes of Ranvier rather than segmental demyelination and remyelination.13 17 18 21

These findings are consistent with experimental results of microstructural changes in the animal model of AMAN. In AMAN rabbits sensitised with gangliosides, pathological findings showed that in addition to axonal degeneration, nodal sodium channel clusters disappear after complement deposition.12 The disappearance of nodal sodium channel clusters would significantly lower the safety factor of impulse transmission and cause conduction block.22 Another important change in this model is detachment of the paranodal myelin terminal loops. The immune attack also causes the disappearance of paranodal adhesion molecules such as contactin, contactin associated protein and neurofascin-155, indicative of impaired tight paranodal axoglial junctions12; this mimics paranodal demyelination, and is likely to account for nerve conduction slowing/block seen in human AMAN.

The present study confirmed that the frequency of ‘AMAN’ electrodiagnosis was significantly higher in antiganglioside positive patients than in negative patients,13 15 but several large studies failed to show a correlation between anti-GM1 antibody and the electrodiagnosis of ‘AMAN’.1 16 There are some possibilities for this discrepancy. Firstly, the target molecules involved in AMAN are not only GM1 but also GM1b, GD1a, GalNAc-GD1a or ganglioside complexes.23 24 Other unidentified gangliosides could also be targets for autoantibodies in AMAN. Moreover, the sensitivity and specificity of antiganglioside assays could have been different in previous studies. Secondly, with current electrodiagnostic criteria, the cut-off values suggestive of demyelination were determined based on the assumption that AMAN causes simple axonal degeneration, but this is not correct.16 19 Thirdly, the cut-off values were determined by the percentage of the normal value established in each laboratory (eg, motor distal latency, >110% of ULN). Such setting of the cut-off-value is reasonable for multicentre studies, but normal values could be differently set in each laboratory, and this might affect the electrophysiological classification.

In antiganglioside negative GBS, patients with the electrodiagnosis of ‘AIDP’ appear to constitute a uniform subgroup which represents ‘classical’ AIDP. Our data suggest that antiganglioside antibody determines the pathophysiology of the subgroup of GBS, and therefore the majority of antibody positive GBS patients have AMAN. For the diagnosis of AIDP as a disease entity, electrodiagnostic criteria should be combined with the results of the antiganglioside assay. When antibodies are positive, the ‘AMAN’ electrodiagnosis indicates AMAN as a disease entity, and the ‘AIDP’ electrodiagnosis at the first test will change in the majority of cases at follow-up. If antiganglioside antibodies are negative, the electrodiagnosis of AIDP would indicate primary demyelination, with features of a classical demyelinating GBS.

Electrophysiologically, differentiation of AIDP and AMAN is difficult and unreliable based on the single studies in the early phase of the diseases because nerve conduction slowing can occur in AMAN, as well as AIDP. Our results suggest that: (1) in single electrophysiological studies, electrodiagnosis 3–6 weeks after GBS onset is more correct than one done in week 1 or 2; and (2) if sequential studies are performed, early reversible nerve conduction failure (rapid normalisation of conduction slowing/block) supports the diagnosis of axonal GBS.

AMAN was reported to be a rare subtype of GBS in Western countries but it may be underestimated.19 It is reported that AMAN is a major form of GBS in Asia; in Central and South America, the frequency is 38∼65%.1 25–28 In the world as a whole, a considerable number of GBS patients will suffer from AMAN, and therefore it is important to recognise GBS subtypes. Each subtype of GBS presumably has a different immunopathogenesis, and constitutes subgroups with similar but somewhat different features and, most importantly, possibly different responses to treatment.29 30 Future studies should elucidate the optimal treatment for each subtype of GBS. To achieve this, classification of the GBS subtypes requires both sequential electrophysiology and antiganglioside antibody assays.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Funding This work was supported in part by the Health and Labour Sciences Research Grant on Intractable Diseases (Neuroimmunological Diseases) from the Ministry of Health, Labour and Welfare of Japan (SK).

  • Competing interests None.

  • Ethics approval Ethics approval was provided by Chiba University, Chiba, Japan, and University ‘G d’Annunzio' Chieti-Pescara, Italy

  • Provenance and peer review Not commissioned; externally peer reviewed.