Respiratory failure is the most serious short-term complication of Guillain–Barré syndrome (GBS), requiring invasive mechanical ventilation in 13–25% of patients.1^–8 Moreover, 60% of those who are intubated develop a major complication; pneumonia, sepsis, gastrointestinal bleeding and pulmonary embolism (reviewed in Durand et al.1). Anticipation of respiratory failure is crucial to avoid respiratory distress and aspiration, as well as to triage patients to the appropriate unit (ward versus intensive care). Early predictors of the need for mechanical ventilation include a time between disease onset and hospital admission of less than 7 days, inability to lift the head, presence of bulbar dysfunction, vital capacity less than 60% of that predicted and electrophysiological evidence of demyelination.1 7 8 Autoantibodies to gangliosides also are probable predictors of respiratory failure in GBS. Anti-GQ1b IgG antibodies were more frequently detected in ventilated than in unventilated GBS patients,5 although another study found no difference.1 IgG antibodies to the GD1a/GD1b and GD1b/GT1b complexes, in this paper designated cD1a/D1b and cD1b/T1b, also have been reported as predictors of the need for artificial ventilation.9

In a large study, three of 53 patients with Fisher syndrome (FS), characterised by ophthalmoplegia, ataxia and areflexia, developed GBS.10 That condition is defined as FS overlapping GBS (FS/GBS) in the present study. One author (SK) found four of six FS/GBS patients who required mechanical ventilation. His personal experience led us to investigate whether FS/GBS patients in a large series often experienced respiratory failure. We thereafter analysed the clinical features of, and anti-ganglioside antibodies in, patients with FS/GBS to find clinical and serological predictors of respiratory failure.

METHODS

Patients

From April 1994 to August 2006, physicians throughout Japan sent us 4121 serum samples from GBS patients and 629 samples from FS patients for anti-ganglioside antibody testing. Although diagnostic criteria for GBS require progressive weakness in more than one limb by 4 weeks after onset and hypo- or areflexia,11 our study used stricter criteria; tetraparesis (3 or less on the Medical Research Council scale) and a monophasic course. Diagnostic criteria for FS in this study were acute ophthalmoplegia, ataxia, hypo- or areflexia, and no limb weakness, criteria stricter than previously.12 Diagnostic criteria for FS/GBS were acute ophthalmoplegia, ataxia, and hypo- or areflexia, as well as subsequent tetraparesis (3 or less on the scale). When patients develop leg weakness followed by arm weakness, the condition was called ascending weakness. When patients develop arm weakness followed by leg weakness, the condition was defined as descending weakness. Medical records of each patient with FS and GBS were reviewed to obtain clinical features, and additional questionnaires were obtained from each primary physician. Patients with profound limb weakness at disease onset and altered consciousness during illness course were excluded. Forty-five patients (men/women, 29/16; median age 46) fulfilled the FS/GBS criteria. During the same period, 100 patients with GBS (55/45; age 52) and 100 with FS (54/46; age 43) were selected randomly as the controls. Information on age, sex, antecedent infectious symptoms, initial symptoms, neurological signs during illness and artificial ventilation were obtained. This study was approved by the Ethics Committee of Dokkyo Medical University.

Nerve conduction studies

Systematic electrodiagnostic studies were made on the subgroup of the six FS/GBS patients seen at Chiba University Hospital. Motor nerve conduction studies were made on the median, ulnar, peroneal and tibial nerves, including F-wave analyses. Sensory nerve action potentials were recorded from the median, ulnar, superficial radial, superficial peroneal and sural nerves after anti-dromic stimulation. Somatosensory-evoked potentials were recorded after median or tibial nerve stimulation. Patients were classified as having acute motor axonal neuropathy (AMAN) or acute inflammatory demyelinating polyneuropathy (AIDP) based on the published electrodiagnostic criteria.13

Serological studies

Serum IgG antibodies against individual gangliosides (GM1, GM1b, GD1a, GalNAc-GD1a, GT1a and GQ1b; 5 pmol/well) were measured by ELISA as described elsewhere.14 In this study, serum was considered positive when the optical density was 0.5 or more at a dilution of 1:500, which was stricter than in our usual studies. The present definition is based on our earlier report.15 An absorption study was performed as described elsewhere.16

IgG antibodies to ganglioside complex GM1/GD1a (cM1/D1a) were tested with a mixture of GM1 and GD1a (each 5 pmol/well) as antigen. Anti-cM1/D1a antibodies were judged positive when the optical density was 0.5 greater than the sum of the antibodies against individual GM1 and GD1a. Antibodies to at least one combination of two of the seven gangliosides (GM1, GM2, GD1a, GD1b, GT1a, GT1b and GQ1b) were similarly tested and judged for positivity in the 45 FS/GBS patients. Because anti-cD1a/D1b and anti-cD1b/T1b IgG antibodies are reported to be predictors of the need for assisted ventilation,9 their presence in the 100 GBS patients was investigated.

Statistical analyses

Differences in proportions were examined by the χ² or Fisher exact test where appropriate. Antibody frequency against each ganglioside complex was analysed using a logistic regression model. The dependent variable was artificial ventilation. The independent variables were the IgG antibodies to each ganglioside complex. A difference was considered significant when p<0.05. Analyses were performed with SPSS 14.0J software (SPSS, Chicago).

RESULTS

Clinical features

As the initial symptom, none of the FS/GBS patients had leg weakness, whereas 41% of the GBS patients did (p<0.001) (table 1). During the course of the illness, facial weakness significantly was less common in FS/GBS (9%) than in GBS (38%) (p<0.001) but not in FS (19%). Oropharyngeal palsy more often was present in FS/GBS (42%) than in FS (10%, p<0.001) but not in GBS (33%). Neck weakness was more frequent in FS/GBS (60%) than in FS (13%, p<0.001) but not in GBS (51%). Whereas ascending tetraparesis was less common in FS/GBS than in GBS, descending tetraparesis was more common in FS/GBS than in GBS (31% vs 6%, p<0.001).

As the initial symptom, ventilated FS/GBS patients more frequently showed titubation (cerebellar ataxic gait) than non-ventilated patients (55% vs 18%, p = 0.04) (table 1). FS/GBS patients more frequently required mechanical ventilation than GBS patients (24% vs 10%, p = 0.04). None of the FS patients required it during the course of the illness. Ventilated FS/GBS patients more often had descending tetraparesis than did unventilated patients during the course of the illness (64% vs 21%, p = 0.02). Similarly, ventilated GBS patients more frequently had descending tetraparesis than unventilated ones (40% vs 2%, p<0.001). The FS/GBS patients required mechanical ventilation sooner than the GBS patients (p = 0.03), and those who needed it had been admitted earlier than those who did not (p = 0.008). Median days to admission did not differ for FS/GBS and GBS but were significantly shorter than for FS (p = 0.04). Frequencies of multiple cranial nerve involvement did not differ between 10 ventilated and 90 unventilated patients with GBS (23% vs 50%, p = 0.14). Oropharyngeal palsy, however, was more frequent in the ventilated patients than in the unventilated patients with GBS (70% vs 29%, p = 0.02).

Of the subgroup of FS/GBS patients, in whom detailed nerve conduction study results were available (n = 6), two were classified as having AMAN based on the electrodiagnostic criteria (reduced amplitudes of compound muscle action potentials without nerve conduction slowing). As an isolated nerve conduction abnormality, the remaining four had no F-waves and so were regarded as unclassifiable.17 Therefore, none of the six who underwent serial nerve conduction studies had AIDP, sensory nerve conduction studies and somatosensory-evoked potentials being normal in all of them, indicative of pure motor axonal involvement in FS/GBS. Interestingly, those two AMAN patients needed artificial ventilation, whereas the other four patients did not.

Serological findings

Of the anti-ganglioside antibodies tested, IgG antibodies to GQ1b and GT1a more frequently were detected in FS/GBS patients than in FS (98% vs 79%, p = 0.007; 98% vs 78%, p = 0.005) (table 2). Neither frequency, however, differed significantly for ventilated and unventilated FS/GBS patients. Also, the anti-GQ1b IgG antibody frequencies did not differ between the 10 ventilated and 90 unventilated patients with GBS (10% vs 11%). In 92 GBS patients without external ophthalmoparesis, the presence of autoantibodies was not associated with the need for artificial ventilation (11% vs 7%). Anti-GT1a IgG antibodies in GBS were associated with neither bulbar palsy nor neck weakness. IgG antibodies to GM1, GM1b, GD1a or GalNAc-GD1a, serological markers of AMAN, were less often positive in FS/GBS than in GBS (16% vs 57%, p<0.001): these autoantibodies were more frequent in ventilated patients than in the unventilated FS/GBS patients (27% vs 12%), but the frequency did not reach a significant difference.

Forty-four available serum samples with anti-GT1a IgG antibodies were used in the absorption studies. Anti-GT1a IgG activity decreased to less than 50% after incubation with GT1a, indicative of successful absorption treatment. Twenty-two samples were excluded because GT1a did not effectively absorb the antibody. Absorption rates for the ventilated and unventilated FS/GBS patients did not differ (62% vs 60%, p = 0.7), nor did those for patients with ascending and descending tetraparesis (61% vs 61%, p = 0.45).

Nine of the 11 ventilated FS/GBS (82%) and 29 of the 34 unventilated patients (85%) had antibodies to at least one combination of two of the seven ganglioside antigens (table 3). IgG antibodies to cM1/D1b, cD1a/D1b and cD1a/T1b, respectively, were more frequent in the ventilated patients than in the unventilated FS/GBS patients (45% vs 17%, 27% vs 9% and 18% vs 3%). However, IgG antibodies to each ganglioside complex were not associated with an artificial ventilation requirement or descending limb weakness with a logistic regression model. Frequencies of anti-cD1b/T1b IgG antibodies were similar in the ventilated and unventilated patients (9% vs 12%). Neither the frequencies of anti-cD1a/D1b IgG nor the frequencies of anti-cD1b/T1b antibodies differed between the 10 ventilated and 90 unventilated GBS patients (11% vs 2%, p = 0.70; 0% vs 2%, p = 0.48).

DISCUSSION

Forty-five of 629 FS patients (7.2%) fulfilled the diagnostic criteria for FS/GBS in this study. The frequency is consistent with an earlier report that three of 53 FS patients (5.6%) later developed profound limb weakness.10 Clinical feature results for 45 FS/GBS patients showed that they more frequently required artificial ventilation than the 100 GBS patients. Moreover, those with FS/GBS who developed descending tetraparesis more often needed mechanical ventilation than those with ascending tetraparesis. The ventilated GBS patients also more frequently showed descending quadriparesis. An earlier study reported that GBS patients who had more rapid tetraparesis progression required mechanical ventilation sooner than those who did not.18 Ours is the first study to show that descending limb weakness is a clinical predictor of respiratory failure both in GBS with and without overlapping FS. The time between disease onset and hospital admission was significantly shorter for the FS/GBS patients than for the FS patients, and for the ventilated versus unventilated FS/GBS patients.

Our findings showed that of the six FS/GBS patients who underwent detailed electrophysiological examinations, two had the AMAN pattern, and the other four were unclassifiable because F-wave absence was an isolated nerve conduction abnormality, suggestive of conduction block or axonal dysfunction restricted to the nerve roots.17 Normal sensory nerve conduction supports a clinical diagnosis of AMAN; therefore, none of our six FS/GBS patients had an AIDP electrodiagnosis. Another study reported that all of four FS patients with tetraparesis had normal motor nerve conduction velocities, distal latencies and F-wave latencies, but a slight tendency for lower compound muscle action potential amplitudes than did five FS patients without tetrapraresis.19 None of their FS/GBS patients had AIDP, whereas 10 of their 14 GBS patients did. Taking into account two other case reports,20 21 we suggest that FS/GBS more often is associated with the AMAN than the AIDP pattern, although the number of patients receiving sufficient neurophysiological examination was small. By contrast, another study has reported that AIDP patients often experience respiratory failure.1 These findings suggest that respiratory failure in GBS is divisible into at least two categories; one seen in severe AIDP, the other associated with descending GBS initially showing FS. In the latter, patients frequently have oropharyngeal palsy and proximal arm weakness, and the pattern of descending progression is more likely to involve the phrenic nerves earlier than the ascending progression common to AIDP.

Forty-four ventilated patients, compared with 87 unventilated GBS patients, more frequently had multiple cranial nerve involvements (91% vs 50%) and anti-GQ1b IgG antibodies (27% vs 8%).5 In GBS without ophthalmoparesis, the presence of anti-GQ1b IgG antibodies was associated with respiratory failure (12% vs 0%), but our study, as well as another study,1 does not confirm this earlier one. IgG antibodies to cD1a/D1b and cD1b/T1b are reported serological predictors of the need for artificial ventilation in GBS,9 but this was not confirmed in our 100 GBS and 45 FS/GBS patients.

Our FS/GBS patients had bulbar palsy, neck weakness and descending tetraparesis during the progression of illness. The clinical features are similar to those of pharyngeal–cervical–brachial weakness, a regional variant of GBS.22^–24 Unlike in GBS,7 oropharyngeal palsy was not associated with the need for artificial ventilation in FS/GBS. Monospecific anti-GT1a IgG antibodies were found in pharyngeal–cervical–brachial weakness,25 and GBS patients carrying these autoantibodies had bulbar palsy and neck weakness.16 The presence of monospecific anti-GT1a antibodies was not associated with the need for assisted ventilation in FS/GBS, but we found that the need for it was associated with descending quadriparesis in both GBS and FS/GBS.

As an initial symptom, titubation could be a predictor of the need for assisted ventilation. The need for mechanical ventilation was associated with descending quadriparesis in both GBS and FS/GBS. FS patients usually have a good natural illness course,10 but we suggest that at least some FS patients develop GBS and respiratory failure. FS patients who note titubation at onset need to be carefully monitored during the progression of illness, and the patients who develop significant limb weakness should be treated with intravenous immunoglobulin as soon as possible.