Article Text


Ambiguous value of Haemophilus influenzae isolation in Guillain-Barré and Fisher syndromes
  1. M Koga1,
  2. S Koike2,
  3. K Hirata1,
  4. N Yuki1
  1. 1Department of Neurology, Dokkyo University School of Medicine, Tochigi, Japan
  2. 2Institute for Medical Science, Dokkyo University School of Medicine
  1. Correspondence to:
 Dr Michiaki Koga
 Department of Neurology, Dokkyo University School of Medicine, Kitakobayashi 880, Mibu, Shimotsuga, Tochigi 321-0293, Japan;


Background:Haemophilus influenzae is considered a causative agent of Guillain-Barré syndrome (GBS) and Fisher syndrome, but the frequency of this infection in GBS is controversial.

Objective: To determine whether isolation of H influenzae indicates it is a causative agent in GBS and Fisher syndrome.

Results: Four (15%) of 27 patients with GBS and Fisher syndrome in whom H influenzae was isolated were also seropositive for Campylobacter jejuni. Antiganglioside IgG antibodies in these four patients did not cross react with their H influenzae lipo-oligosaccharides, whereas antiganglioside antibodies in the four patients with positive serology for H influenzae did.

Conclusions: The findings suggest that H influenzae isolation is not always indicative of the causative agent in these syndromes and that tests for other infections should be made, even in cases of positive culture.

  • GBS, Guillain-Barré syndrome
  • LOS, lipo-oligosaccharide
  • Guillain-Barré syndrome
  • Fisher syndrome
  • Haemophilus influenzae
  • ganglioside

Statistics from

Guillain-Barré syndrome (GBS) is a heterogeneous disorder in which various infections may occur before and after neurological onset. Such preceding infections as Campylobacter jejuni, cytomegalovirus, Epstein-Barr virus, and Mycoplasma pneumoniae are generally accepted as causes of GBS.1 The antecedent infectious agent is usually determined by a serological test or isolation of the agent after GBS onset, sometimes making it difficult to distinguish the antecedent infection from complications of GBS itself. This is a problem, particularly in the isolation of respiratory infectious agents, because nosocomial respiratory tract infections resulting from aspiration pneumonia and intratracheal intubation are common in GBS.2

Haemophilus influenzae, a major pathogen of community acquired respiratory infection, is considered a causative agent of GBS and Fisher syndrome, but the frequency of this infection in GBS is controversial.1,3–6 It is hypothesised that ganglioside epitopes on H influenzae lipo-oligosaccharide (LOS) induce autoantibodies against GM1 and GQ1b, thereby resulting in neuropathy.3,6 Identification of antecedent H influenzae infection is important in clinical terms because it is an indicator of rapid recovery from GBS.7H influenzae isolation is the standard diagnostic procedure for this infection, but it is not clear whether isolation of the organism in cases of GBS or Fisher syndrome necessarily means that it is the cause of those disorders. We investigated other possible antecedent infections in patients with H influenzae isolations and examined whether the isolation indicates that it is a causative agent in GBS and Fisher syndrome.


We examined recent infections (H influenzae, C jejuni, cytomegalovirus, and M pneumoniae) serologically, as described elsewhere,6 in 27 patients with GBS (n = 11), Fisher syndrome (n = 14), and overlapping Fisher syndrome/GBS (n = 2), from whom H influenzae had been isolated in sputum or swab cultures. All the strains were isolated shortly after neurological onset, and artificial ventilation was needed for five patients after culture sampling. Serotypes of the isolates were all non-typable, the biotypes predominantly II and III, like the isolates of uncomplicated respiratory infections. There was a history of antecedent respiratory infection symptoms in 21 patients (78%).

Serum IgG antibodies to GM1, GM1b, GM2, GD1a, GalNAc-GD1a, GD1b, GT1a, GT1b, and GQ1b were assayed by enzyme linked immunosorbent assay (ELISA), as reported elsewhere.8 An absorption study then was done to examine whether these antiganglioside antibodies cross reacted with the LOS on the isolates. Crude LOS was prepared as described previously.6 A 100 μl portion of diluted serum, which gave an optical density of between 1.0 and 2.0 for antiganglioside antibodies, was incubated in LOS precoated wells (LOS lysate 1 μl per well) of microtitre plates at 4°C for 48 hours. The supernatant was used as the primary antibody in the standard ELISA. Absorption rates are expressed as percentages of the optical densities obtained with and without absorption treatment.


Antiganglioside IgG antibodies were detected in 10 of 11 GBS patients (91%) and in all 16 Fisher syndrome patients. Frequent antibodies were anti-GM1 (73%) and anti-GM1b (55%) in the GBS patients, and anti-GQ1b and anti-GT1a (both 100%) in the Fisher syndrome patients.

Four patients (15%: three GBS; one Fisher syndrome) had positive serology for C jejuni and seven (26%: five Fisher syndrome; one GBS; one Fisher syndrome/GBS) had positive serology for H influenzae, but none for both. All the patients were negative for cytomegalovirus and M pneumoniae. Of the four C jejuni seropositive patients, one had a previous history of soft faeces, but the others had no preceding gastrointestinal symptoms (table 1). Follow up serum was available for one C jejuni seropositive patient (patient 3, table 1) who had a significant anti-C jejuni IgA titre decrease during the recovery phase of illness. IgG antibodies to GM1, GM1b, GT1a, or GQ1b were detected in all four patients. In contrast, all seven patients with positive serology for H influenzae infection had a previous history of respiratory infection symptoms. Anti-GQ1b and anti-GT1a IgG antibodies were detected in six of these and anti-GM1 IgG antibody in the seventh.

Table 1

 Patients from whom H influenzae was isolated and who had positive serology for recent C jejuni infection

Cross reactivity of antiganglioside IgG antibodies to the LOSs of their H influenzae isolates was examined in eight patients, four with positive serology for C jejuni and four for H influenzae. Cross reactivity was absent in the four who had positive C jejuni serology, whereas it clearly was present in at least three of the four H influenzae seropositive patients examined (fig 1). The target ganglioside of the antibodies was not related to cross reactivity (data not shown).

Figure 1

 Cross reactivity of antiganglioside IgG antibodies in H influenzae-isolate patients to their isolate lipo-oligosaccharide (LOS). Absorption rates are given as percentages of optical densities obtained with and without absorption. Circles represent the mean absorption rates of five wells, error bars the standard errors. Antiganglioside IgG antibodies did not cross react with H influenzae LOS in patients with serological evidence of recent C jejuni infection, whereas the antibodies did react with the isolate LOS in those with serological evidence of recent H influenzae infection. Ab, antibody.


We found that 15% of our patients from whom H influenzae was isolated had positive serology for recent C jejuni infection. The specificity of the C jejuni serological assay system used is 97%6; moreover, a significant titre decrease during the recovery phase was confirmed in one patient, indicating that there were no false positive results. In GBS and Fisher syndrome after H influenzae and C jejuni infections it is hypothesised that the ganglioside mimics present on the bacterial LOSs produce antiganglioside antibody.9 The antibodies in our H influenzae seropositive patients did cross react with their isolate LOSs, whereas those in the C jejuni seropositive patients did not. This suggests that production of antiganglioside antibodies in the latter patients was not triggered by H influenzae, although H influenzae infection had certainly occurred, either before or after the neurological onset. It also is noteworthy that only one of four C jejuni seropositive patients had a history of preceding gastrointestinal symptoms, but this does not necessarily indicate false positive results in our C jejuni serology assay because C jejuni has been isolated from some GBS/Fisher syndrome patients who experienced only upper respiratory infectious symptoms or fever.10 Our findings suggest that H influenzae isolation does not always indicate that this is the causative agent in GBS and Fisher syndrome. Even when H influenzae is isolated, we should also test for other infections—in particular C jejuni—irrespective of the antecedent symptoms. This also may hold true for cases in which other agents are isolated and serological evidence obtained. We believe that more careful judgment of the causal agent in GBS and Fisher syndrome will lead to a decrease in the numerous case reports of sham causal infection in these syndromes.

Positive H influenzae serology was found in 26% of the tested cases in whom H influenzae had been isolated. This low frequency may partly reflect the assay’s low sensitivity, or it may indicate that H influenzae infection was merely a complication of GBS or Fisher syndrome in some of the patients, because most of the serum samples assayed were obtained soon after neurological onset, at a time when neither antibody response nor H influenzae infection may have occurred. However, anti-GQ1b and anti-GT1a antibodies were detected in most of the H influenzae seropositive patients, all of whom had Fisher syndrome. This supports our previous results showing that H influenzae infection is associated more closely with Fisher syndrome than with GBS and that production of antiganglioside antibody is mediated by a GQ1b/GT1a epitope on the bacterial LOS.4,6 A serological H influenzae test and examination of the cross reactivity of antiganglioside antibodies with the isolate LOS would help in determining the causal role of the bacterium when H influenzae is isolated in such cases.


We thank Ms Maki Okazaki for her excellent support in all the assays. Supported in part by grants from the Ichiro Kanehara Foundation to MK; the Kanae Foundation for Life & Socio-Medical Science to MK; the Japan Intractable Diseases Research Foundation to MK; a grant for Scientific Research (B) (KAKENHI 14370210 to NY) from the Ministry of Education, Culture, Sports, Science and Technology of Japan; a Research Grant for Neuroimmunological Diseases to NY from the Ministry of Health, Labour and Welfare of Japan; a Health Sciences Research Grant (Research on Psychiatric and Neurological Diseases and Mental Health) to NY from the Ministry of Health, Labour, and Welfare of Japan; and a grant from the Human Frontier Science Program (RGP0038/2003-C to NY).


View Abstract


  • Competing interests: none declared.

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.