Brain (1992), 115, 1093-1106

RECURRENT GUILLAIN-BARRE SYNDROME

CLINICAL AND LABORATORY FEATURES

by F. GRAND'MAISON, T. E. FEASBY, A. F. HAHN and

W. J. KOOPMAN

{From the Department of Clinical Neurological Sciences, Victoria Hospital, University of Western
Ontario, London, Ontario, Canada)

SUMMARY

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

The clinical and laboratory features of recurrent Guillain-Barre' syndrome (RGBS) were reviewed in 12
patients in whom a total of 32 episodes fulfilled accepted criteria for Guillain-Barre' syndrome (GBS). All
patients were asymptomatic or only mildly symptomatic between attacks. In a given patient, the time to
reach peak deficit from the onset of symptoms, the functional grade at peak deficit and the duration of
the intervals between episodes varied considerably and unpredictably from one episode to the next. Analysis
of these parameters across the entire group revealed no significant change as the number of attacks increased.
The distribution of weakness varied between episodes with the possible exception of features of the Miller
Fisher variant which were more constant. Tremor was noted in two patients and enlarged nerves in one
patient. There was no evident response to immunosuppressive therapy. Results of cerebrospinal fluid (CSF)
analysis and nerve conduction studies during recurrences were those expected in typical monophasic GBS.
On nerve biopsy, onion bulb formations were sometimes observed after several recurrences.
The following characteristics of RGBS may be sufficiently distinctive from those of chronic relapsing
poly neuropathy to justify their nosological separation: rapid onset of symptoms with subsequent complete
or near complete recovery, high incidence of an antecedent illness, lack of an apparent response to
immunosuppressive therapy and normal CSF protein levels at the onset of a recurrence.

INTRODUCTION
Guillain-Barre" syndrome (GBS) is defined as an acute, largely reversible, idiopathic
monophasic demyelinating polyradiculoneuropathy. There are several reports of patients
with multiple episodes of otherwise typical GBS (Austin, 1958; Castaigne et al., 1966;
Pleasure et al., 1968; Thomas et al, 1969; Bolin and de Carle, 1972; Pollard and Selby,
1978; Seyal et al., 1978; Galvan et al., 1983; Wijdicks and Ropper, 1990). It has been
estimated that GBS recurs in 2—5% of patients (Dyck and Arnason, 1984).
Repeated bouts of demyelinating polyneuropathy have also been described as a variant
of chronic idiopathic demyelinating polyneuropathy (CIDP) (Dyck et al., 1975;
McCombe et al., 1987), which has been referred to as chronic relapsing polyneuro-
pathy (CRP) (Thomas et al., 1969; Prineas and McCleod, 1976; Dalakas and Engel,
1981).
Although GBS and CIDP share many features, certain clinical and immunological
differences may justify their separation into distinct syndromes. An antecedent viral
Correspondence to: Francois Grand'Maison, MD, Centre Hospitalier Universitaire de Sherbrooke, Service de
Neurologic, 3001, 12e Avenue N, Sherbrooke, Quebec, Canada, J1H 5N4.

© Oxford University Press 1992

1094 F. GRAND'MAISON AND OTHERS

infection and weakness of respiratory and facial muscles are frequent in GBS but unusual
in CIDP (Dyck and Arnason, 1984). The two conditions may differ in their response
to immunosuppressive treatments (Toyka et al., 1988). The recent demonstration that
serum of CIDP patients contains significantly fewer soluble interleukin-2 receptors than
those of GBS patients further suggests the possibility of different pathogenic mechanisms
(Hartung et al., 1990). Differentiation of the two disorders relies on arbitrary clinical
criteria: the time to reach maximum disability in GBS is usually less than 4 wks (Asbury
and Cornblath, 1990) compared with more weeks or months in CIDP (Dyck et al., 1975;
McCombe et al., 1987; Cornblath et al., 1991). Hughes (1990) has defined a group
of patients intermediate between GBS and CIDP, with an onset phase of 4 - 1 2 wks,
as 'subacute CIDP'. The difficulty of classifying inflammatory poly neuropathies is
illustrated by the occasional patient with attacks of both acute and more chronic onset.
Nevertheless, for reasons outlined above and for the lack of evidence to the contrary,
GBS and CIDP are generally considered different entities for clinical and research

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

purposes.
Recurrent GBS (RGBS) and CRP, representing the respective recurrent forms of GBS
and CEDP, may similarly correspond to distinct entities. However, in many descrip-
tions of relapsing demyelinating poly neuropathies, no distinctions have been made on
the basis of the time course of individual attacks and the characteristics of RGBS and
CRP remain unclear (Dyck et al., 1975; McCombe et al., 1987). The objective of this
study is to further define the features of RGBS and to compare some of these
characteristics with those described in the literature for CRP.

P A T I E N T S AND M E T H O D S

Recurrent GBS was defined as two or more episodes of acute idopathic demyelinating polyneuropathy
each with an onset phase lasting less than 8 wks followed by complete or near-complete recovery. It was
felt that an onset phase of 4 wks or less may prove too restrictive for the purposes of this study. Individual
attacks fulfilled the clinical criteria for the diagnosis of GBS (Asbury and Cornblath, 1990).
Results of nerve conduction studies were not uniformly recorded because examinations were performed
by different individuals in different laboratories over a period of more than 20 yrs. Electromyographic
(EMG) criteria suggestive of demyelination were judged to recent guidelines (Asbury and Cornblath, 1990).
For analysis of nerve biopsies, fascicles of the superficial and deep peroneal nerves were fixed in buffered
glutaraldehyde, processed and embedded in Epon. Semithin sections were stained with toluidine blue. Thin
sections were stained with uranyl acetate and lead citrate and viewed in a Phillips 410 electron microscope.
Samples of the nerves were teased in Epon and classified according to Dyck et al. (1984).
From a chart review of patients diagnosed with GBS since 1970, 12 cases conforming to the above condi-
tions were compiled from the records of the three teaching hospitals associated with the University of
Western Ontario. Over that time period, a total of 270 patients were diagnosed with GBS.
Functional grade was scored using the following scale: 0 = healthy; 1 = minor symptoms or signs
and not limited in ordinary daily activities; 2 = able to walk 5 m without assistance but unable to do physical
work; 3 = able to walk 5 m with assistance only; 4 = chair- or bed-bound; 5 = assisted ventilation required
(Winer et al., 1988).
Conventional statistical analysis was performed. Differences of sample means were evaluated by
performing Student's t test and one-way analysis of variance was used for comparison or more than two
means.
Two cases will be detailed.
Patient 1
In 1964, at age 15 yrs, this man gradually developed ascending weakness and paraesthesiae 2 wks following
a rubella infection. The neurological deficit reached a peak 4 wks after onset, at which time he had severe
 RECURRENT GUILLAIN-BARRE SYNDROME 1095

weakness in his legs and moderate weakness in his arms. There was no facial or respiratory muscle involve-
ment. The cerebrospinal fluid (CSF) contained 2850 mg/1 of protein 10 d after onset of symptoms. He
slowly improved and was discharged from the hospital 3 mths later. He fully recovered within 6 mths.
In 1967 he noticed mild paraesthesiae and weakness in his fingers and toes 2 wks following an upper
respiratory tract infection (URTT). There was gradual progression of weakness over 2 wks but he remained
ambulatory. He was hospitalized for 1 mth and recovered fully in 2 mths.
In 1971 he had a brief illness with malaise and vomiting followed 2 wks later by tingling in his fingers
and toes. He developed ascending weakness and facial paraesthesiae over the next week. By the 10th d,
he was quadriplegic, had bilateral facial weakness and required assisted ventilation. There was no
ophthalmoplegia. He began to improve 3 wks after onset and was on the ventilator for a total of 3 wks.
He was discharged from the hospital after 3.5 mths and recovered completely by 7 mths. He was treated
with prednisone during his acute illness.
In 1977, 2 wks after a 2-d bout of malaise and vomiting, he noted paraesthesiae in his fingers and toes.
He developed mild leg weakness but his arms and face were not involved. He was admitted to hospital
and treated with prednisone 20 mg daily from the second day for a period of 3 wks. Recovery was again
complete.

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

The fifth episode occurred in 1981, 2.5 wks following a 1-d episode of nausea, vomiting and anorexia.
He developed paraesthesiae in his fingers and toes and mild leg weakness. The weakness gradually ascended
to involve his arms, reached a peak by day 35 and began to improve 4 d later. By day 25, a mild postural
and intention tremor of both hands became apparent and slowly worsened. Examination on day 43 revealed
mild weakness of all four limbs, more severe distally with wasting of extensor digitorum brevis bilaterally.
The tremor became prominent and affected the hands and, to a lesser degree, the legs and head. Tremor
frequency was 4 - 5 Hz. He was areflexic. Sensory examination revealed a mild deficit to pin and vibra-
tion and reduced proprioception distally, The ulnar nerves and dorsal cutaneous nerves of the feet were
enlarged. Light tapping over the ulnar, median, radial, common peroneal and the dorsal cutaneous nerves
of the feet evoked a sensation of intense tingling in the cutaneous territory of those nerves. The CSF contained
1510 mg/1 of protein and one leucocyte per cubic millimetre. The posterior tibia! nerve conduction velocity
was 11.6 m/s and conduction block was found in multiple motor nerves. Biopsy of the right deep and
superficial peroneal nerves revealed acute demyelination, remyelination, onion bulb formation and mild
lymphocytic infiltration (Fig. 1). He was treated with prednisone, 20 mg/day from day 3. Prednisone was
slowly tapered and discontinued on day 78, at which time he felt he had recovered at least 90% of his
strength. The tremor persisted but to a lesser degree. Within days after discontinuing prednisone, he noted
increasing paraesthesiae and weakness. Prednisone was temporarily reinstituted; he was improved by day
98 and felt normal on day 120.
In October 1982 he noticed increasing numbness of his fingers and toes. Examination on day 5 revealed
mild bilateral wasting of intrinsic foot muscles and mild distal reduction in vibration sense but no weakness.
Myotactic reflexes except ankle jerks could be elicited. Treatment with prednisone, 80 mg/day, was initiated
but he rapidly developed ascending weakness, facial diplegia and increasing postural and intention tremor
of arms and legs. He remained ambulatory. He was treated with four plasma exchanges of 4 1 each over
a period of 2 wks. During plasma exchange his strength started to improve but 2 wks following the last
exchange he again developed weakness which gradually started to improve 3 wks later. Despite increasing
strength, the tremor worsened and was often incapacitating. He resorted to ingesting 1.5 oz rye whisky
before meals which improved the tremor and facilitated eating. Propanolol was ineffective. Over the next
6 mths, his strength returned to normal and the termor almost completely disappeared. Myotatic reflexes
were all normal. He went back to work full time as a builder.
The seventh episode began in 1986, 2 wks after a viral illness, when he noted paraesthesiae in his fingers
and toes. Ascending weakness reached a peak on day 45, at which time he required assistance to walk.
Reflexes were absent. Hand tremor slowly increased but remained relatively mild. He did not receive
immunosuppressive treatment. Over the next 5 mths he fully recovered and was back at work.
Patient 2
In 1954 this 32-yr-old woman developed blurred vision, leg weakness and paraesthesiae of her hands
and feet a few days after an upper respiratory tract infection. After several days, she complained of increasing
gait unsteadiness and limb weakness until she could no longer walk. She had bilateral facial weakness,
complained of diplopia and bilateral ptosis. Deep tendon reflexes were absent. She did not suffer signifi-
cant respiratory failure. She made a gradual recovery and was functionally normal after a year.
10% F. G R A N D ' M A I S O N AND O T H E R S

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

FIG. 1. Biopsy of the superficial peroneal nerve of patient 1 demonstrating inflammatory demyelination. Many axons
show disproportionately thin myelin sheaths with early onion bulb formation. A few fibres are segmentally demyelinated
(arrow) with associated macrophages (arrow head). Lymphocyte-like mononuclear cells (asterisk) infiltrate the
endoneurium. Cross-section: 1 /un; toluidine blue; magnification = 25 jim.

In 1981, 2 d after a cold, she noticed weakness in all limbs, paraesthesiae in her hands and feet and
diplopia. She was found to have mild weakness in all limbs, areflexia and weakness of abduction of the
right eye. She did not receive immunosuppressive treatment. After 1 wk, the weakness and paraesthesiae
gradually improved but diplopia lasted 3 mths.
One year later, 1 wk after an upper respiratory tract infection, she developed gait unsteadiness and felt
bilateral leg weakness. The next day, she noted bilateral hand numbness, diplopia and could not walk.
By day 6, bulbar weakness and respiratory failure necessitating ventilatory support had developed. Maximum
disability was reached on day 8 when she had marked ptosis, facia) weakness, complete bilateral external
ophthalmoplegia, quadriplegia, areflexia and severely reduced vibration and position sense in all limbs.
There were no automatic abnormalities. The first sign of improvement, bilateral eyelid movement, was
noted on day 11. She breathed independently by 4 wks and walked 1 wk later. Immunosuppressive treat-
ment was not used. At discharge from hospital, 10 wks after onset, she had mild right ptosis, mild bilateral
weakness of eye abduction, minimal distal weakness, moderate generalized ataxia and reduced vibration
sense distally. On day 15, CSF contained 1440 mg/1 of protein and one leucocyte per cubic millimetre.
By day 60, CSF protein had fallen to 490 mg/1. By 6 mths, neurological examination was normal except
for mild incoordination of fingers movements, mild gait ataxia and areflexia.

RESULTS
Clinical features
Table 1 summarizes the clincal features and the course of the neurological condition
in each case. Among the 12 subjects, seven were male and five were female. The average
 RECURRENT GUILLAIN-BARRE SYNDROME 1097

age at the onset of the initial episode was 28.9 yrs (8—63 yrs) and at the onset of the
last episode 46.1 yrs (16 — 69 yrs). While the inclusion criterion for the time to reach
peak disability from the onset of symptoms was 8 wks, for most episodes this interval
was 2 wks or less. The mean interval between attacks was 9.7 yrs (2 mths to 43 yrs).
Eight patients had a single recurrence, three patients two recurrences and one partient
six recurrences. An episode of isolated unilateral facial palsy in one patient (patient
8) was not included.
As only one patient had more than three attacks, statistical analysis was carried out
only for the first three episodes. From the first to the third episode, there seemed to
be a trend towards shorter intervals from the antecedent illness to symptom onset, towards
longer intervals from symptom onset to peak deficit and towards longer intervals between
episodes (Table 2). However, these changes did not reach statistical significance
(P > 0.05). The mean functional grade at peak deficit was quite similar from one attack
to the next (Table 2).

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

Sixty-seven percent of initial episodes and 85 % of recurrent episodes were preceded
by an infective illness (20 episodes in nine patients), immunization (one episode: tetanus
toxoid) or were associated with pregnancy or with the immediate post-partum period.
In patient 5, two episodes were associated with the first pregnancy (at seventh month
of pregnancy and 4 d post-partum) and one episode began 21 d after delivery of the
second child 5 yrs later. In patient 8, the first attack of GBS appeared 60 d post-partum;
the other two episodes were unrelated to pregnancy. In one patient, serology suggested
a recent herpes simplex infection.
In nine patients, the neurological symptoms and signs were typical of monophasic
GBS. The most common initial symptoms were weakness and paraesthesiae distally
in the limbs. Facial weakness was present in 47% of episodes, bulbar weakness in 30%
and respiratory muscle weakness requiring assisted ventilation in 13%. All patients were
areflexic. Nerves were palpably enlarged only in patient 1, following the fifth episode.
Postural and intention tremor were prominent features in two patients. The characteristics
of the tremor in patient 1 were outlined above. A similar slow postural and intention
tremor was seen in patient 4 at the onset of the second episode. There was no clear
relationship between the degree of weakness and the amplitude of tremor, except for
the absence of tremor when plegic. No cerebellar or extrapyramidal signs were noted.
In three patients (patients 2, 9, 12), features of the Miller Fisher variant of GBS
were prominent (Fisher, 1956). In every episode of these three patients, external
ophthalmoparesis of variable severity was associated with limb weakness and, in four
episodes, with respiratory failure requiring assisted ventilation. Ataxia in the second
bout of patient 9 was marked and associated with only mild weakness of limb muscles.
Ocular signs were reported in two other patients: mild bilateral ptosis during both episodes
in one patient and weakness of both external rectus muscles during one attack in another
patient. Neither presented with limb or gait ataxia.
After the initial episode, symptomatic recovery was complete in eight patients. Among
the other four patients, two described mild proximal weakness, one complained of distal
paraesthesiae in the feet and of gait unsteadiness and one noted mild incoordination
of finger movements. No patient was limited in daily activities. After the second episode,
three patients fully recovered subjectively and objectively. Five were reported to have
mild residual distal sensory or motor signs with depressed ankle jerks after a follow-up
1098 F. GRAND'MAISON AND OTHERS

TABLE 1. CLINICAL FEATURES AND CSF PROTEIN LEVELS IN 12 PATIENTS WITH

RECURRENT GBS

Time to Time• to
Age Age Recurrence onset of peak Grade CSF
Case/ first last interval Antecedent symptoms deficit at peak protein
sex episode episode Episode (mths) condition (days) (wks) deficit mg/l (days)
1M 15 37 1 Rubella 14 4 4 2850 (10)
2 36 URTI 14 2 2 —
3 48 Enteritis 14 2 5 —
4 72 Enteritis 14 2 2 —
5 48 Enteritis 17 7 2 1510 (15)
6 10 Enteritis 10 3 3 —
7 44 Enteritis 14 7 4 2850 (12)
2 F 32 59 1 URTI 5 1 4 —
2 324 Flu-like 2 1 2 —
3 14 URTI 7 1 5 490 (60)

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

3M 32 46 1 Flu-like 25 1 3 1440 (15)
2 175 URTI 14 1 4 1420 (7)
4M 63 69 1 Herpes simplex _ 1 5 870 (3)
2 67 Flu-like 2 1 4 1400 (19)
5 F 25 30 1 Pregnancy+urti 14 2 4 1900 (15)
2 2 P-partum (4 d) — 2 4 5750 (27)
3 65 P-partum (21 d) — 2 2 1320 (12)
6 F 11 16 1 — 2 3 —
2 60 Mononucleosis 5 1 4 570 (5)
7M 28 41 1 URTI 42 1 4 —
2 156 URTI 21 1 2 364(1)
8 F 34 47 1 P-partum (2 mths) — 4 2 1000(60)
2 120 — 1 1 —
3 36 — — 8 2 —
9M 44 51 1 1 4 840 (16)
2 79 URTI 2 1 5 380 (3)
10 F 40 53 1 Tetanus toxoid 17 2 4 390(4)
2 156 URTI 7 1 4 300(2)
11 M 8 51 1 1 4 6000 (12)
2 516 URTI 20 1 4 539 (1)
12 M 27 54 1 1 4
2 336 — — 1 5 500(10)

Case/ Facial Bulbar Extra-ocular

sex Episode weakness weakness muscle weakness Tremor Treatment
1M 1
2
3 + + Prednisone
4 Prednisone
5 ^ Prednisone
6 -i Prednisone+plasmapheresis
7
2 F 1 + + Diplopia, Ptosis
2 Right VI Prednisone
3 + + Total external ophthalmoplegia
3M 1 Bilateral VI
2 + + Prednisone
4M 1 Prednisone
2 -k Prednisone
 RECURRENT GUILLAIN-BARRE SYNDROME 1099

Case/ Facial Bulbar Extra-ocular

sex Episode weakness weakness muscle weakness Tremor Treatment
5 F 1 Prednisone
2
3
6F 1
2 Prednisone
7M 1
2
8 F 1
2
3
9M 1 Total external ophthalmoplegia
2 Total external ophthalmoplegia Plasmapheresis
10 F 1 Mild bilateral ptosis
2 Mild bilateral ptosis Plasmapheresis

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

11 m 1
2 Plasmapheresis
12 M 1 Diplopia
2 Total external ophthalmoplegia Plasmapheresis

TABLE 2. AVERAGE VALUES OF PARAMETERS FOR EACH EPISODE

Episode no.
i ! 2 3 4 5 6 7
Interval from 19.5 (6) 9.7 (9) 10.5 (2) 14.0 (1) 17.0 (1) 10.0 (1) 14.0 (1)
antecedent condition
to onset of symptoms
in days (number of
patients)
Time to reach peak 1.8 (12) 1.2 (12) 3.2 (4) 2.0 (1) 7.0 (1) 3.0(1) 7.0 (1)
deficit in weeks
Grade at peak deficit 3.8 (12) 3.4 (12) 3.5 (4) 2.0 (1) 2.0 (1) 3.0(1) 4.0 (1)
Interval between 169.0 (12) 41.0 (4) 72.0 (1) 48.0 (1) 10.0 (1) 44.0 (1)
episodes in months

period averaging 38 mths. The other myotatic reflexes reappeared. Among the other
three patients, symptomatic proximal weakness persisted in two and sensory ataxia
affecting gait persisted in the other. In three of the four subjects with two or more
recurrences, neurological deficits tended to accumulate with the number of recurrences
but no significant disability resulted. In patient 1, no deficits accumulated and only mild
vibration loss and atrophy in the feet were documented following recovery from the
sixth recurrence.
Nine episodes in six subjects were treated acutely with prednisone, four episodes
in four subjects with plasmapheresis and one episode with both prednisone and
plasmapheresis. Because of the variable modes of therapy, variable dosage schedules
and the uncontrolled nature of treatment, the possible effects of the above therapies
were not analysed. It can be stated, however, that there were no instances of an apparent
dramatic response to treatment.
No patient had a family history of demyelinating polyneuropathy.
1100 F. GRAND'MAISON AND OTHERS

Cerebrospinal fluid
Cerebrospinal fluid was obtained in all patients on at least one occasion, during a
total of 21 episodes. The protein level was elevated (>450 mg/1) in the symptomatic
phase in all attacks in which measurements were made 1 wk or more after the onset
of symptoms: the average CSF protein level was 2480 mg/1 (840-6000 mg/1; n = 11)
and, in all cases, the leucocyte count was normal (less than 5 x 109 cells per litre). In
two patients, during the early recovery period (within 60 d of peak deficit), the protein
level was mildly elevated, 490 and 1000 mg/1. Protein levels in the CSF were not obtained
in any subject during the asymptomatic interval. However, all protein measurements
performed within 1 wk of onset of a recurrence were normal (five episodes in three
subjects) or only mildly elevated (two episodes in two subjects).
Electrophysiological studies
In total, 18 nerve conduction studies were performed on 10 patients during 13 different

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

episodes (Table 3). Only three studies were performed during the first episode and first
interval (most initial episodes occurred prior to 1966), limiting the possibility of
longitudinal analysis. In total, 14 nerve conduction studies were consistent with a
demyelinating polyneuropathy (10 episodes in seven patients), one was normal, two
showed only prolongation of the terminal latency of the median nerve and one provided
insufficient evidence for multifocal demyelination. Eleven of 15 examinations performed
during the acute symptomatic phase revealed evidence of demyelination. Only three
examinations were performed following clinical recovery: two studies showed only mild
abnormalities while the other revealed more severe demyelination (patient 8, 7 mths
following the fourth episode at which time she was mildly symptomatic).
Median nerve conduction velocity during the acute phase averaged 37.4 m/s (14 studies)
and following recovery 34.2 m/s (three studies). In the early days of the fifth attack
of patient 1, the marked prolongation of the distal latencies from stimulation of the median
and ulnar nerves at the wrist suggested very severe demyelination and continuous
(non-saltatory) nerve conduction in those segments. Conduction block or temporal disper-
sion were present in at least one nerve in 10 of 14 examinations (five of eight patients)
during the acute symptomatic phase and in two of four studies (one of three patients)
during asymptomatic interval. Median, ulnar and sural sensory nerve action potentials
were absent or diminished in amplitude in all but one study. Fibrillation potentials of
positive sharp waves in distal muscles were reported in five of 14 needle examinations
(four of nine patients) during the acute phase and in two of four examinations (one of
three patients) in the asymptomatic interval.

Histological observations
Fascicular biopsies of the superficial and deep peroneal nerves (SPN and DPN) were
performed in two subjects (patients 1, 2). In patient 1 (Fig. 1), the biopsies were
performed during the fifth attack. Examination of both nerves under the electron
microscope revealed expansion of the endoneurium by excess deposition of ground
substance and collagen. The overall density of myelinated fibres was reduced. Numerous
fibres were relatively thinly myelinated and showed evidence of recurring demyelina-
tion with onion bulb formation. Numerous nomonuclear cells were scattered within
the endoneurium and around endoneurial blood vessels. Examination of teased fibres
 Downloaded from brain.oxfordjournals.org by guest on January 14, 2011
TABLE 3. NERVE CONDUCTION STUDIES AND ELECTROMYOGRAPHY IN 10 RGBS PATIENTS

Motor conduction Sensory conduction

Time Median Ulnar Common peroneal Median Ulnar Sural
from Latency Velocity Latency Velocity Latency Velocity Conduction Temporal Amplitude Amplitude Amplitude EMC JO
m
Case Episode onset (ms) (ms)' (ms) (ms)' (ms) (ms) block dispersion (uV) (uV) (uV) Fibs/PSW n
1 5 days 16.0 Y Y
c
JO
5 days 58.0 28.0 32.0 6.3 - Y - 73
5 42 d 21.4 14.0 NR NR N m
z
2 3 16 d 3.9 46.0 — — 4.5 45.0 N N 2.5 — 2.5 N H
F = 72.7 O
4 1 42 d 6.2 50.0 4.6 47.0 — 43.0 — Y NR NR NR Y C
F
2 10 d 3.8 39.0 4.0 74.0 41.0 Y NR NR NR N r
6 2 70 d — 40.0 — 50.0 — — — N — — — Y
z
7 2 yrs 4.4 46.7 5.0 45.0 5.0 45.0 N — NR NR NR N CD
8 2 60d 2.9 18.0 25.3 5.3 43.3 Y Y N
4
4
45 d
200 d
4.9
3.6
30.0
25.8
3.2 23.8 5.3 27.0 Y
Y
Y
Y
5.3
E Y
Y
73
73

Z Z >•
9 1 9d 5.0 55.0 3.0 62.0 7.0 47.0 N N
1 300d 4.6 57.0 2.5 65.0 N N NR
2 10 d 4.0 35.7 3.4 57.1 43.0 N N 11.0 6.0 z
10 2 5d 2.9 54.6 2.3 62.7 4.8 44.0 N N 19.0 17.0 4.0 N o
73
11 2 10 d 5.3 35.3 3.6 36.6 5.0 18.3 Y N 2.0 2.0 Y O
2 42 d 16.2 13.2 8.8 8.1 Y Y Y 2
m
12 2 10 d 5.0 40.0 3.6 51.0 6.4 42.6 Y Y 12.1 8.4 9.6 N
d = davs., F = F-wave. NR = no resrx>nse, Y = [>resent, N •3 absent, Fibs/PSW ° fibrillation cotentials or r>ositive shairp waves.
1102 F. GRAND'MAISON AND OTHERS

(Table 4) showed very marked evidence of chronic ongoing segmental demyelination.

Associated axonal degeneration was minimal. On electron microscopy, numerous
formerly myelinated fibres showed a lack of myelin sheath or were thinly myelinated.
In patient 2, the biopsies were performed during the third episode. Changes in both
nerves were identical. Light microscopy revealed a mild reduction of myelinated fibres,
particularly of large fibre size. A number of fibres were undergoing acute Wallerian
degeneration. Occasional regenerative clusters were noted. Rare fibres with segmental
demyelination and a few thinly myelinated axons were observed. A few macrophages
were present throughout the endoneurium and in a perivascular distribution. Examina-
tion of the teased fibre preparations revealed a considerable degree of ongoing Wallerian
degeneration with only very little evidence of paranodal and repaired segmental
demyelination (Table 5).
Electron microscopy revealed a mild reduction of myelinated fibres and numerous
fibres showing early stages of Wallerian degeneration. Only rare fibres with segmental

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

demyelination were observed. No lymphocytic infiltration was noted.

TABLE 4. CASE 1: TEASED FIBRE ANALYSIS (%)

Nerve A and B C D E F G H Total (n)
Deep peroneal nerve 19 7.5 69.9 0.9 2.8 — — 106
Superficial peroneal nerve 10 — 84.4 1.1 1.1 — — 90

TABLE 5. CASE 2. TEASED FIBRE ANALYSIS (%)

Nerve A and B C D E F G H Total (n)
Deep peroneal nerve 75.4 0.9 0 15.4 8.2 — — 110
Superficial peroneal nerve 77.5 — 0 17.4 5.2 — — 115

DISCUSSION
Twelve patients with one to six recurrences of GBS are reported. In a given patient,
the functional grade at peak deficit, the time to reach peak deficit and the time interval
from the antecedent illness to the onset of neurological symptoms varied considerably
and unpredictably from one episode to the next as did the duration of asymptomatic
intervals between episodes. In addition, analysis of these parameters across the entire
group revealed no significant change as the number of episodes increased.
Except in one patient in whom all episodes were related to pregnancy or to the early
post-partum period, the nature of the antecedent illness tended to differ from episode
to episode. This non-specificity in triggering events may explain the variable intervals
to the onset on neurological symptoms. This contrasts with the patient described by
Pollard and Selby (1978) who developed GBS on three occasions following the administra-
tion of tetanus toxoid; the interval from exposure to onset of symptoms tended to decrease
with exposure.
Although the absence (nine subjects) or presence (three subjects) of ophthalmoparesis
with ataxia were constant from one episode to the next in each patient, facial, bulbar
and respiratory involvement varied considerably. This contrasts with previous reports
 RECURRENT GUILLAIN-BARRE SYNDROME 1103

of RGBS in which facial or bulbar weakness seemed consistent from one episode to
the next (Castaigne et al., 1966; Wijdicks and Ropper, 1990) and argues against reac-
tivation of smouldering inflammation at sites specifically affected during the initial attack
as the main mechanism of recurring polyneuritis (Asbury et al., 1969; Wijdicks and
Ropper, 1990). Rather, the changing distribution of weakness suggests a more generalized
reactivation of the disease. Features of the Miller Fisher variant of GBS were much
more constant from episode to episode in patients presenting initially with this syndrome
as was reported elsewhere (Kaplan et al., 1985; Schapira and Thomas, 1986; Kaiser-
Smith, 1989) suggesting the possibility of dysimmune mechanisms different from those
operant in recurrences of classical GBS.
In two of our patients (patient 1, episodes 5, 6, 7; patient 4, episode 2) a postural
and intention tremor was prominent. The characteristics of this tremor were very similar
to those described by Dalakas et al. (1984) in seven patients with CRP. We also observed
that tremor was not associated with cerebellar or central nervous system signs and that

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

it correlated poorly with the degree of weakness and loss of proprioception. The above
authors noted that tremor developed preferentially during later relapses of CRP. Similarly,
tremor is unusual in monophasic GBS but has been reported during recurrences of GBS
(Pollard and Selby, 1978; Galvan et al., 1983). Tremor, which is quite uncommon in
axonal poly neuropathies, has also been described in the demyelinating form of hereditary
motor and sensory neuropathy (type 1) and in the demyelinating neuropathy associated
with IgM paraproteinaemia (Thomas, 1984). The mechanism of this tremor and its
relationship to demyelination and remyelination remains unclear.
From a review of the literature, it is difficult to ascertain the clinical characteristics
of CRP because results have usually been pooled with the chronic progressive form
of CIDP (Dyck et al., 1975; Dalakas and Engel, 1981; McCombe et al., 1987).
However, analysis of the detailed report by Prineas and McCleod in 1976 allows the
identification of 19 patients with CRP defined as two or more episodes each with an
onset phase of more than 3 wks following which both symptoms and signs improved.
Therefore there is some overlap in time-course with our patients. Only 15 of approxi-
mately 54 episodes (28%) were associated with an antecedent illness, mostly infection.
It is possible, although unlikely, that the much higher incidence of potential triggering
events in our patients (78%) represents ascertainment bias related to the more acute
onset of symptoms, whereby patients are more likely to recall a recent infectious episode.
In Prineas and McCleod's series, five patients (26%) had facial weakness, one (5%)
had ophthalmoparesis and two (11%) required ventilatory assistance during their most
severe attack; these percentages would likely be much lower if computed as a function
of the total number of episodes, as was done in the present study. Neurological examina-
tion was normal between relapses in three of their patients (subjects 1, 3, 7), who were
among those with the shortest intervals from symptom onset to peak deficit. It is possible
that the condition in these three patients corresponds to RGBS, as defined by our clinical
criteria, rather than to CRP. It is noteworthy that many of their subjects showed a rapid
apparent response to treatment with prednisone or with adrenocorticotropic hormone.
Although the present study was not designed to assess properly the efficacy of
corticotherapy in RGBS, no significant response was noted in our patients.
Results of CSF studies in our patients during recurrences and asymptomatic intervals
are those expected in classical monophasic GBS (Asbury and Cornblath, 1990). Early
1104 F. GRAND'MAISON AND OTHERS

during a recurrence and following recovery, CSF protein levels are normal or minimally
elevated but increase by 1 wk after the onset of an attack. Serial measurements of CSF
protein levels in CRP are reported in two studies (Prineas and McCleod, 1976; Dalakas
and Engel, 1981). In both series, it is mentioned that CSF protein levels tend to remain
elevated during remissions, albeit at levels below those during exacerbations. These
observations suggest quiescence or resolution of the inflammatory response between
episodes in RGBS and ongoing, though subclinical, inflammatory activity during
remissions in CRP.
Results of nerve conduction studies in our patients are consistent with those reported
in the literature for acute monophasic GBS (Brown and Feasby, 1984; Albers, 1987).
Slowing of conduction velocities, prolonged distal latencies, conduction block or
dispersion, which represent the cardinal features of acquired demyelinating poly-
neuropathies, were seen in most patients. Evidence of axonal degeneration in the form
of fibrillation potentials and positive sharp waves, observed in some of our patients,

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

has frequently been reported in GBS (Brown and Feasby, 1984; Albers, 1987). It is
difficult to compare electrophysiological findings between RGBS and CRP since, to
our knowledge, no systematic longitudinal studies have been performed in either case.
In our two patients in whom nerve conduction studies were obtained both during the
acute phase and during the asymptomatic interval, no significant improvement in
conduction velocities or distal latencies was noted. This does not necessarily indicate
ongoing disease activity during asymptomatic intervals in RGBS since nerve conduction
abnormalities may persist following full functional recovery from acute monophasic
GBS (Albers, 1987). It is therefore unlikely that nerve conduction studies will provide
a strong basis to distinguish RGBS from CRP.
The nerve biopsy in patient 1, obtained in the acute phase of the fifth episode, revealed
segmental demyelination, onion bulb formations and substantial endoneurial infiltra-
tion by inflammatory cells. These findings are relatively non-specific and indicate repeated
episodes of demyelination and remyelination. Similar morphological observations have
been made in CIDP (Dyck et al., 1975), CRP (Prineas and McCleod, 1976) and in
another case of RGBS (Pollard and Selby, 1978). Biopsies of SPN and DPN from patient
2, although performed during the third episode, showed no evidence of chronic disease.
The predominant axonal changes may be explained by a focus or foci of inflammation
and demyelination proximal to the biopsy site with secondary axonal degeneration. This
explanation is strongly supported by the prior demonstration of a much delayed common
peroneal F-response (72.7 ms) despite normal conduction velocity in the foreleg indicating
demyelination proximal to the knee.
In conclusion, RGBS consists of multiple episodes of typical acute GBS, each bout
varying considerably and unpredictably in severity and in the distribution of involve-
ment with the possible exception of features of the Miller Fisher variant which seem
more constant from one episode to the next. Tremor and enlarged nerves, unusual in
GBS, followed multiple attacks. The CSF findings and EMG abnormalities during each
individual episode are similar to those described in acute monophasic GBS. Nerve conduc-
tion abnormalities sometimes persisted between episodes. Nerve biopsy revealed onion
bulb formations, indicating repeated demyelination and remyelination. Whether RGBS
and CRP represent different conditions remains unclear. The distinction is blurred by
the fact that the time course of the onset phase of individual attacks varies considerably
 RECURRENT GUILLAIN-BARRE SYNDROME 1105

in few patients (Prineas, 1971; McCombe et al., 1987). Nevertheless, our observa-
tions on RGBS suggest that the rapid onset of symptoms with subsequent complete or
near complete recovery, the high incidence of an antecedent illness, the apparently poor
response to immunosuppressive therapy and the normal CSF protein levels at the onset
of recurrences may be sufficiently distinctive from CRP to justify their separation until
more definite evidence proves differently.

ACKNOWLEDGEMENT
Presented in part at the Peripheral Neuropathy Association Meeting in Oxford, UK, August 15, 1990.

REFERENCES

ALBERS JW (1987) Inflammatory demyelinating polyneuropathy. In: Clinical Electromyography. Edited

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

by W. F. Brown and C. F. Bolton. Boston: Butterworths, pp. 209-244.
ASBURY AK, CORNBLATH DR (1990) Assessment of current diagnostic criteria for Guillain-Barre' syndrome.
Annals of Neurology, 27, Supplement, S21-S24.
ASBURY AK, ARNASON BG, ADAMS RD (1969) The inflammatory lesion in idiopathic polyneuritis: its
role in pathogenesis. Medicine, Baltimore, 48, 173—215.
AUSTIN JH (1958) Recurrent polyneuropathies and their corticosteriod treatment: with five-year observa-
tions of a placebo-controlled case treated with corticotrophin, cortisone, and prednisone. Brain, 81,
157-192.
BOUN TD, DE CARLE DJ (1972) Recurrent Guillain-Barr6 polyneuritis. Medical Journal of Australia, 2,
262-264.
BROWN WF, FEASBY TE (1984) Conduction block and denervation in Guillain-Barre' polyneuropathy. Brain,
107, 219-239.
CASTAIGNE P, BRUNET P, NOUAILHAT F (1966) Enquete clinique sur les polyradiculoneVrites inflammatoires
en France. Revue Neurologique, 115, 849—872.
CORNBLATH DR, ASBURY AK, ALBERS JW, FEASBY TE, HAHN AF, MCLEOD JG et al. (1991) Research
criteria for diagnosis of chronic inflammatory demylinating polyneuropathy (CIDP): report from an
Ad Hoc Subcommittee of the American Academy of Neurology AIDS Task Force. Neurology,
Cleveland, 41, 617-618.
DALAKAS MC, ENGEL WK (1981) Chronic relapsing (dysimmune) polyneuropathy: pathogenesis and
treatment. Annals of Neurology, 9, Supplement, 134—145.
DALAKAS MC, TERAVAINEN H, ENGEL WK (1984) Tremor as a feature of chronic relapsing and
dysgammaglobulinemic polyneuropathies: incidence and management. Archieves of Neurology,
Chicago, 41, 711-714.
DYCK PJ, ARNASON B (1984) Chronic inflammatory demyelinating polyradiculoneuropathy. In: Peripheral
Neuropathy. Second edition. Edited by P. J. Dyck, P. K. Thomas, E. H. Lambert and R. Bunge.
Philadelphia: W. B. Saunders, pp. 2101-2114.
DYCK PJ, LAIS AC, OHTA M, BASTRON JA, OKAZAKI H, GROOVER RV (1975) Chronic inflammatory
polyradiculoneuropathy. Mayo Clinic Proceedings, 50, 621—637.
DYCK PJ, KARNES J, LAIS A, LOFGRHN EP, STEVENS JC (1984) Pathologic alterations of the peripheral
nervous system of humans. In: Peripheral Neuropathy. Second edition. Edited by P. J. Dyck,
P. K. Thomas, E. H. Lambert and R. Bunge. Philadelphia: W. B. Saunders, pp. 760-870.
FISHER M (1956) An unusual variant of acute idiopathic polyneuritis (syndrome of ophthalmoplegia, ataxia
and areflexia). New England Journal of Medicine, 255, 57—65.
GALVAN P, REIXACH R, POTES E, QUERA MA, SUBIRANA A, FERRER I et al. (1983) Polirradiculoneuritis
recidivante. Aportaci6n de un caso anatomoclfnico. Consideraciortes etiopatogeneticas. Archives de
Neurobiologia Madrid, 46, 257-264.
HARTUNG HP, HUGHES RAC, TAYLOR WA, HEININGER K, REINERS K, TOYKA KV (1990) T cell activa-
tion in Guillain-Barre' syndrome and in MS: elevated serum levels of soluble IL-2 receptors. Neurology,
Cleveland, 40, 215-218.
1106 F. GRAND'MAISON AND OTHERS

HUGHES RAC (1990) Chronic idiopathic demyelinating polyradiculoneuropathy. In: Guillain-Barri

Syndrome. Edited by R. A. C. Hughes. London: Springer-Verlag, pp. 205-246.
KAISER-SMITH, BARONE (1989) Relapsing polyneuritis following classic Miller Fisher syndrome. Journal
of Neurology, Neurosurgery, and Psychiatry, 52, 802—803.
KAPLAN JG, SCHAUMBURG HH, SUMNER A (1985) Relapsing ophthalmoparesis-sensory neuropathy
syndrome. Neurology, Cleveland, 35, 595—596.
MCCOMBE PA, POLLARD JD, MCLEOD JG (1987) Chronic inflammatory demyelinating polyradiculo-
neuropathy: a clinical electrophysiological study of 92 cases. Brain, 110, 1617 — 1630.
PLEASURE DE, LOVELACE RE, DUVOISIN RC (1968) The prognosis of acute polyradiculoneuritis. Neurology,
Minneapolis, 18, 1143-1148.
POLLARD JD, SELBY G (1978) Relapsing neuropathy due to tetanus toxoid. Journal of the Neurological
Sciences, 37, 113-125.
PRINEAS JW (1971) Demyelination and remyelination in recurrent idiopathic polyneuropathy. Acta
Neuropathologica, 18, 34—57.
PRINEAS JW, MCCLEOD JG (1976) Chronic relapsing polyneuritis. Journal of the Neurological Sciences,
27, 427-458.

Downloaded from brain.oxfordjournals.org by guest on January 14, 2011

SCHAPIRA AHV, THOMAS PK (1986) A case of recurrent idiopathic ophthalmoplegic neuropathy (Miller
Fisher syndrome). Journal of Neurology, Neurosurgery, and Psychiatry, 49, 463—464.
SEYAL M, ZIEGLER DK, COUCH JR (1978) Recurrent Guillain-Barre' syndrome following influenza vaccine.
Neurology, New York, 28, 725-726.
THOMAS PK (1984) Clinical features and differential diagnosis. In: Peripheral Neuropathy. Second edition.
Edited by P. J. Dyck, P. K. Thomas, E. H. Lambert and R. Bunge. Philadelphia: W. B. Saunders,
pp. 1169-1190.
THOMAS PK, LASCELLES RG, HALLPIKE JF, HEWER RL (1969) Recurrent and chronic relapsing Guillain-
Barre' polyneuritis. Brain, 92, 589-606.
TOYKA KV, HARTUNG HP, SCHAFER B.'HEININGER K, FIERZ W (1988) Immune mechanisms in acute
and chronic inflammatory poly neuropathies. Journal of Neuroimmunology, 20, 277—281.
WUDICKS EFM, ROPPER AH (1990) Acute relapsing Guillain-Barre' syndrome after long asymptomatic
intervals. Archives of Neurology, Chicago, 47, 82—84.
WINER JB, HUGHES RAC, OSMOND C (1988) A prospective study of acute idiopathic neuropathy. I. Clinical
features and their prognostic value. Journal of Neurology, Neurosurgery, and Psychiatry, 51, 605 —612.

(Received August 20, 1991. Revised February 2, 1992. Accepted March 12, 1992)